Remote Sensing Techniques for Assessment of Mangrove Forest Structure, Species Composition and Biomass, and Response to Environmental Change. By Anthea Lauren Mitchell.
School of Biological, Earth & Environmental Sciences, The University of New South Wales, High Street, Kensington, NSW 2052, Australia.
REMOTE SENSING TECHNIQUES FOR ASSESSMENT OF MANGROVE FOREST STRUCTURE, SPECIES COMPOSITION AND BIOMASS, AND RESPONSE TO ENVIRONMENTAL CHANGE.
ANTHEA LAUREN MITCHELL.
A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy.
Postgraduate Board
University of New South Wales December, 2003
2
ABSTRACT Through the integration of fine resolution stereo aerial photography, hyperspectral Compact Airborne Spectrographic Imager (CASI) data, and NASA JPL polarimetric Airborne Synthetic Aperture Radar (AIRSAR) data, spatial datasets relating to the extent, height, density, species/community composition and biomass of mangroves within the Alligator Rivers Region, Kakadu National Park (KNP), Northern Australia were generated. Using stereo photography, 1 m spatial resolution orthomosaics and accompanying Digital Elevation Models (DEMs) of canopy height for the main areas of mangrove in the ARR were produced. Tree density was retrieved from the stereo photography through a tree top delineation and counting algorithm developed using eCognition software and an adjustment for sub-canopy density based on height class distributions.
The species and community composition of the mangroves was
discriminated using hyperspectral CASI data. Spatial modeling techniques were then used to integrate the fine resolution datasets of height, density and species, as well as allometric equations to provide spatial estimates of total above ground and component biomass for the mangroves. The total biomass was also derived through empirical relationships with low frequency AIRSAR data. The mapping techniques developed for mangrove characterization in this project are applicable to mangroves of both tropical and temperate affinities. The derived datasets represent important baselines for a pristine area of mangroves within the World Heritage listed KNP, and through their integration, provided a unique insight into the structure and dynamics of the mangroves and their coastal environment. The processes of change within the mangroves were linked to both rapid and longerterm changes in the coastal environment, and also climate. The combined use of airborne remote sensing datasets and satellite SAR for local to global survey of mangrove environments within a suitable monitoring framework was also discussed. It was concluded that such a program is essential for one critical reason, that being, the mangroves are important barometers of the greenhouse effect given their sensitivity to changing climate and rising sea levels, and may provide early warning of adverse change within the coastal environment.
i
EXECUTIVE SUMMARY This project considered the use of airborne optical and microwave remote sensing technologies for tropical mangrove characterization, and the capacity to monitor the response of mangroves to coastal environmental change.
Relatively undisturbed,
pristine areas of mangrove within the World Heritage listed Kakadu NP in Australia’s Northern Territory were selected, as the changing mangrove distributions could be related to changes in the immediate environment, with limited interference from human activity. Mangroves respond to both rapid and episodic changes in the near-shore environment, and their response is typically manifested in changing species composition, extent, structure and condition. The objectives of the project therefore were to establish and demonstrate the use of historical and more recent airborne remote sensing data for baseline characterization of mangroves, and to use the datasets to identify changes in the mangroves over time, and interpret these in the context of previous coastal environmental change in the region. Kakadu NP is located in the Top End Coastal region of northern Australia and constitutes an area of around 20,000 km² within the Alligator Rivers Region (ARR) of the Northern Territory.
The Park is significant on an international scale, widely
recognized as a major centre for biodiversity, with a number of wetlands designated as Ramsar sites of international importance, and is an area of outstanding natural and cultural value. The Park is jointly owned and managed by the traditional Aboriginal landowners and Parks Australia North (Commonwealth Department of Environment and Heritage). Major landforms within Kakadu vary between the western Arnhem Land escarpment and the extensive coastal floodplains, with a diverse array of habitats ranging from savanna woodland, open Eucalypt forest and mosaics of interconnected wetlands, billabongs, tidal creeks, mangroves and saline mudflats. A vast diversity of fauna utilize the many niches of the Park, and include over a third of Australia’s birds, a quarter of its land mammals, and the highest diversity of freshwater fish in the country. International and national tourists alike are drawn to Kakadu to witness the stunning beauty and diversity within its borders.
ii
The mangroves of the ARR occur as narrow fringes along the Wildman, West, South and East Alligator Rivers and also along the coastline. Field and Barron Islands situated directly north of the ARR coastline also contain significant stands. Species diversity is relatively high, with twenty-one of Australia’s twenty-nine mangrove species occurring within the region. Located at the land-sea interface, the mangroves form an important buffer zone, which protects the coastline from erosion. The mangroves also provide an important habitat for a wide variety of fish, crustaceans, lizards, snakes, turtles, estuarine crocodiles, small mammals, and migratory waterbirds. Aboriginal inhabitants of the Park also forage for food in the mangroves and have strong cultural links to these environments. Coastal and estuarine processes largely determine the extent and distribution of mangrove communities in the ARR.
The prevailing climate, with monsoonal wet
season and lengthy dry spell, as well as the occasional cyclone and storm surge, contribute to their changing extent and condition as well. Significant changes over the past half-century have involved the expansion and retraction of mangroves along different sections of the coastline, the inland extension of tidal creeks, and subsequent saltwater intrusion and mangrove colonisation. As well, the more subtle changes in species composition, inter-zonal changes in distribution, and the landward expansion of certain species have been observed. A number of causal factors have been suggested, including changing tidal movements, long-term sea level rise and subsequent readjustment of species to changing environmental gradients across the landward to seaward edge of communities.
The changes occurring within the mangrove
communities, notably the saltwater intrusion, are of concern, as the impacts are widespread and have already affected some of the adjoining freshwater wetlands and water bodies within the Park. In order for the extent and rates of change over time to be assessed, baseline datasets that characterize the extent and distribution of mangroves are required. Changes in the extent of mangroves are indicative of the changing conditions that either favour colonization and hence expansion, or that lead to retraction of communities as a result of storm damage, changing tidal regimes, or other naturally induced or anthropogenic impacts. Over time however, the extent of communities may remain unchanged, and
iii
inter-zonal changes in species composition, structure and biomass may occur. Hence, other types of baseline data may be important, relating to forest structural attributes (height and density), species/community composition, and total and component biomass. For monitoring purposes, baseline datasets need be established for reference years, against which the response to future change can be assessed. An extensive set of true colour stereo aerial photographs (66 stereo pairs) acquired in 1991 and 2 black and white (B&W) stereo photographs acquired in 1950 were used to generate 1 m spatial resolution orthomosaics from which the extent of mangroves could be mapped, and digital elevation models (DEMs) from which canopy height was determined.
The 1991 photographic coverage included the coastal and estuarine
sections of the Park, including all of the Alligator Rivers, Wildman River, Field and Barron Islands, while only the lower West Alligator River photographs were digitized for 1950.
Using a digital photogrammetric workstation (DPW), running LH Systems
Socet Software, the scanned photographs were triangulated and DEMs over mangrove areas were generated with horizontal and vertical resolutions of 0.37 m and 0.5 m respectively. Orthomosaics were then generated for the coastal regions and for each river system and island community. The availability of only true colour imagery and the low dynamic range of the spectral data limited the use of aerial photography for baseline mapping of the distribution of species and communities.
For this reason, 1 m spatial resolution
Compact Airborne Spectrographic Imager (CASI) data was acquired over the West Alligator River mangroves in July 2002. The dataset was comprised of 14 bands, encompassing the visible to near infrared wavelength region, and using standard spectral classification procedures, provided superior discrimination of the mangrove species and major zones. Using the stereo photography, it was possible to identify the sunlit crowns of individual treetops, and map the location of the large upper canopy trees from which stem density was obtained. This was achieved through a semi-automated crown top mapping procedure, developed using eCognition software. Following generation of the tree maps, adjustments were applied based on field measurements for sub-canopy
iv
density, and estimates of density were derived for each major mangrove zone. The fine spatial resolution of the photography provided the optimal medium for use of the mapping techniques and subsequent discrimination of tree crowns in mangroves of low to moderate stem densities. Where higher stem densities were observed (typically >1 stem m-2) in regrowth forest, alternative methods relating to spectral indices (e.g., the Normalized Difference Vegetation Index, NDVI) derived from hyperspectral data were required. The integration of the resulting spatial datasets on tree height and density (derived from stereo photography) and species composition (derived from CASI), with available allometric equations, enabled the generation of spatial datasets of total and component (leaf, branch, trunk and root) biomass of the mangroves. Spatial modeling techniques were used to integrate the datasets, and field measurements were used for verification. Polarimetric SAR data was also investigated for the capacity to provide spatial estimates of biomass.
By relating field measurements of mangrove structural
parameters and biomass to SAR backscatter, empirical relationships were derived and used to retrieve reasonable estimates of the total biomass. Analysis of the SAR data through extraction of profiles, scattering classifications, and coherence and phase measurements, also revealed information on the influence of mangrove forest structure and biomass on the microwave scattering response. The project also demonstrated the use of integrating the datasets acquired by a range of sensors for the detection of changes within the mangrove environment.
By
comparing the 1991 baseline of mangrove extent for the West Alligator River with that generated in 1950 from stereo photography, and also the 2002 CASI data, significant changes within the mangrove communities were observed. The comparison revealed the steady upstream colonization of mangroves, the formation of cut-offs and narrowing of some tidal channels, the expansion of mangroves on the upper western bank, including the formation of a new zone of Sonneratia species, extensive erosion of coastal mangroves on the eastern bank of the river, and inter-zonal changes in species composition and distribution. These changes were related to both historical and more recent trends in climate and coastal environmental change.
v
By integrating the orthomosaics (mangrove extent), DEMs (tree height) and CASI (species) data, a unique insight into the structure and dynamics of the West Alligator mangroves was achieved, which could not have been obtained through analysis of the orthomosaics alone. Analysis of the DEMs revealed significant areas of growth and also dieback of mangroves; in particular, substantial mangroves were observed colonizing the extending tidal creeks on the eastern bank of the West Alligator River. The change over the years was indicative of the response of the mangroves to the longterm effects of saltwater intrusion. The mangrove species most tolerant to the changing conditions were identified from the CASI data. The DEMs also confirmed considerable erosion of the taller mangroves at the mouth of the Wildman and East Alligator Rivers, in response to changing coastal conditions. An abrupt change in canopy height and substantial movements between the years (i.e., having advanced or retreated) within the landward Avicennia and Rhizophora zones on the West Alligator suggested changing environmental gradients (salinity and nutrient availability) in response to changing tidal conditions and a possible sea level rise. A further benefit of integrating the information content from all three datasets was that areas subject to more rapid change (associated with storm activity) were revealed. The datasets provide a visual record of the plant distributions and the impacts of storm activity, increases in the intensity of which have been associated with global warming and climate change. The techniques for baseline mapping and characterization of the ARR mangroves developed in this project are applicable to other coastal regions of the world, given that the appropriate remote sensing data is available or can be acquired. In generating baseline datasets of mangrove extent, species distribution, structure and biomass, a mechanism for assessing the impacts of changes in the coastal environment, including those induced by global warming, has been provided. Aerial photography is available for many areas and more importantly, provides a source of historical data from which to assess change. New baseline datasets should be generated using currently operational hyperspectral and SAR sensors, which could be compared against those generated using historical photography. The datasets could also be used to support the regional mapping and characterization of mangroves using satellite remote sensing data (e.g., Landsat and SPOT sensors), and spaceborne SAR (e.g., JERS-1, SIR-C, ALOS).
vi
ACKNOWLEDGEMENTS The author wishes to express her gratitude to the timeless efforts of Dr Richard Lucas (University of Wales), whose continued support and dedication helped make the following work a success.
Your ongoing commitment to achieving the highest
standards in research has been an inspiration. I have thoroughly enjoyed working with you over the past few years. A warm thankyou also to Prof. Tony Milne (University of New South Wales) whose keen interest in all things ‘Kakadu’ and continual guidance was greatly appreciated over the years. Many thanks also to Brian Donnelly (University of New South Wales) who assisted with the many months of photogrammetric processing, and whose company always kept things lively. In October of 2002, the opportunity to conduct fieldwork in KNP arose, and was a most memorable experience. Even tramping around in the thigh deep mud, in 30 degree heat, with swarming biting insects did not and could not detract from my enjoyment of this amazingly diverse and pristine environment. A huge thankyou to my supervisors and everyone at ERISS who supported the campaign: Max Finlayson, Gary and Therese Fox, Kirrilly Pfitzner (from ERISS), and also Peter Brocklehurst and Chris Mandium from DIPE. The assistance of John Grice in the field is gratefully acknowledged as well. My appreciation is also extended to the Traditional landowners who gave us permission to undertake research on their land. Thanks also to Rod Kennett from PAN and everyone else who assisted in our being able to undertake a successful and safe field campaign. Thanks also to my family and friends who endured the ride with me. The funding for this project was supported by an Australian Postgraduate Award (APA) and through ERISS.
"Where else can one find trees which walk out over the water and grow roots upwards, seeds which germinate before they fall from the parent tree, fish which hop about in the mud and climb trees, and monkeys which eat crabs?" (the Thai mangroves as described by Thom Henley, 1998).
vii
TABLE OF CONTENTS Page number ABSTRACT
i
EXECUTIVE SUMMARY
ii
1.
Acknowledgements
vii
Table of Contents
viii
List of Tables
xvii
List of Illustrations
xxi
INTRODUCTION
1
1.1
Mangroves: a definition
1
1.2
The global distribution of mangroves
2
1.2.1
Old World mangroves
4
1.2.2
New World mangroves
4
1.3
Eco-physiology of mangroves
4
1.4
Importance of mangroves
5
1.5
Natural and anthropogenic change
7
1.5.1
Environmental and climate change
7
1.5.2
Anthropogenic impacts
8
1.6
Identifying change in mangrove environments
11
1.6.1
Extent
11
1.6.2
Structure and growth stage
12
1.6.3
Biomass and productivity
12
1.6.4
Species/community composition
13
1.6.5
Zonation patterns
14
1.6.6
Overview of monitoring change
15
1.7
Distribution and characteristics of mangroves in Australia
15
1.8
Research objectives
17
1.9
Format of thesis
18
viii
2.
THE IMPACT OF ENVIRONMENTAL AND CLIMATE CHANGE ON
21
MANGROVES – A REVIEW 2.1
Introduction
21
2.2
Current evidence of mangrove response to climate change
24
2.2.1
26
2.3
2.4
2.5 3.
Climate change: observation and prediction
29
2.3.1
Direct anthropogenic change
30
2.3.2
Indirect anthropogenic change
31
Expected response of mangroves to future climate change
35
2.4.1
Physical effect
35
2.4.2
Biological effects
37
Concluding summary
40
REMOTE SENSING OF MANGROVES – A REVIEW
42
3.1
Spatial, temporal and spectral resolution
42
3.1.1
Spatial resolution
42
3.1.2
Temporal resolution
45
3.1.3
Spectral resolution
47
3.2
3.3 3.4 4.
Historical evidence
Baseline datasets for mangroves
48
3.2.1
Mangrove extent
49
3.2.2
Species/community composition
53
3.2.3
Mangrove structure
56
3.2.4
Biomass and productivity
62
Comparing baselines
63
3.3.1
63
Integration of remote sensing data
Concluding summary
STUDY SITE: THE ALLIGATOR RIVERS REGION, NORTHERN
66 67
AUSTRALIA 4.1
Kakadu National Park, Northern Territory Australia
67
4.2
Physiography and soils
71
4.3
Climate and hydrology – the Alligator Rivers Region
74
4.4
Plant communities of the ARR
78
ix
4.5
Sandstone plateau and lowlands
79
4.4.2
Freshwater floodplains
80
4.4.3
Salt marshes and salt flats
82
4.4.4
Coastal and estuarine mangroves
83
Fauna of the ARR
86
4.5.1
Birds
86
4.5.2
Fish
88
4.5.3
Reptiles and amphibians
90
4.5.4
Invertebrates
91
4.5.5
Mammals
92
Threats to the ARR
93
4.6.1
Introduced animals
93
4.6.2
Invasive weeds
95
4.6.3
Fire/burning regime
96
4.6.4
Salinization
97
4.6.5
Mining impacts, contamination/pollution
98
4.6.6
Tourism
99
4.7
The dynamic environment
100
4.8
Concluding summary
100
REMOTE SENSING DATASETS
102
5.1
Remote sensing data: availability and acquisition
102
5.1.1
Stereo aerial photography
102
5.1.2
Compact Airborne Spectographic Imager data
104
5.1.3
NASA JPL Polarimetric Airborne Synthetic Aperture Radar 107
4.6
5.
4.4.1
(AIRSAR) data 5.2
5.3
Stereo photography: pre-processing
110
5.2.1
Scanning and calibration
110
5.2.2
Aerial triangulation
110
5.2.3
Image resampling
115
CASI data: pre-processing
116
5.3.1
Data calibration
116
5.3.2
Rectification
117
x
5.4
5.5 6.
AIRSAR data: pre-processing
117
5.4.1
Data calibration
117
5.4.2
Rectification
119
Concluding summary
119
FIELD DATA COLLECTION AND ANALYSIS
121
PART I
122
6.1
6.2
West Alligator River (East bank), 1999
122
6.1.1
Transect and sample locations
122
6.1.2
Standard plot measurements
123
6.1.3
Transect descriptions
124
West Alligator River (West bank), 2002
129
6.2.1
Sampling locations
129
6.2.2
Plot descriptions
130
6.2.3
Standard plot measurements
131
6.2.4
Standard tree measurements
134
6.2.5
Tree structural measurements
135
PART II 6.3
6.4
6.5
DATA COLLECTION
DATA ANALYSIS
138
Species/community composition and structure
138
6.3.1
Species, zonation and growth stage
138
6.3.2
Percentage foliage cover (PFC)
141
6.3.3
Height and diameter
143
6.3.4
Canopy depth
149
6.3.5
Crown area and volume
152
6.3.6
Density
158
Above ground biomass
161
6.4.1
General trends in height and biomass across communities
163
6.4.2
Leaf biomass
167
6.4.3
Branch/Trunk biomass
167
6.4.4
Above ground root biomass
168
6.4.5
Reliability of allometrics
170
Relationships between structural parameters
171
6.5.1
171
Relationships between tree height and diameter
xi
6.6
6.7 7.
6.5.2
Relationships between diameter and crown dimensions
175
6.5.3
Relationships between height and crown dimensions
177
Three-dimensional canopy structure
178
6.6.1
Leaf measurements
179
6.6.2
Leaf spectra
180
6.6.3
Branch measurements
181
Concluding summary
GENERATION OF TREE HEIGHT AND DENSITY MAPS USING
185 186
STEREO AERIAL PHOTOGRAPHY 7.1
Reasons for and approaches to height estimation
186
7.2
Canopy height estimation from stereo aerial photography
188
7.2.1
Using stereo aerial photography
188
7.2.2
Digital elevation modelling
189
7.3
Generation of DEMs for the mangroves of Kakadu NP
192
7.4
Technical assessment
201
7.4.1
Comparing DEMs of varying spatial resolution
201
7.4.2
Height extraction for vegetated and non-vegetated areas
204
7.4.3
Errors associated with open canopies
204
7.4.4
Validation with ground truth
205
7.5
DEM generation using historical aerial photographs
209
7.6
Tree density and crown mapping using optical data
212
7.7
Overview of crown structure and density mapping
212
7.8
Tree density estimation
213
7.8.1
Method I: Identifying and counting tree tops
213
7.8.2
Method II: Empirical relationships
220
7.9 8.
Concluding summary
223
GENERATION OF ORTHOMOSAICS
225
8.1
Mapping mangrove extent using stereo aerial photography
225
8.2
Generation of orthomosaics
226
8.3
Spatial baseline of the extent of ARR mangroves
229
8.4
Technical assessment
238
xii
8.5 9.
Concluding summary
239
SPECIES / COMMUNITY DISCRIMINATION
240
9.1
Review of species mapping
240
9.2
Discriminating mangrove species and communities using true colour 242
aerialphotography 9.3
Species mapping with hyperspectral CASI data
244
9.3.1
Mapping mangrove extent from CASI data
244
9.3.2
Spectral reflectance characteristics
245
9.3.3
Classification of mangroves
250
9.4
Accuracy assessment
253
9.5
Observations on the distribution of mangrove species within the
254
West Alligator River
10.
9.6
Links between mangrove canopy height and species composition
258
9.7
Concluding summary
259
BIOMASS ESTIMATION USING OPTICAL DATASETS
261
10.1 Importance of structural indicators and biomass
261
10.2 Biomass estimation using fine spatial resolution datasets
263
10.2.1 The distribution of biomass within the West Alligator
273
mangroves 10.3 Accuracy assessment 10.3.1
11.
274
Accuracy of height, diameter and density inputs
274
10.3.2 Limitations of equations for height-diameter and biomass
277
10.3.3 Biomass estimates: ground validation
278
10.4 Concluding summary
280
BIOMASS ESTIMATION AND ANALYSIS OF FOREST
282
STRUCTURE USING SAR 11.1 Sensitivity of SAR backscatter to biomass and structure of
282
mangrove forest 11.1.1
Forest backscatter and scattering mechanisms: theory
283
11.1.2
Radar response to mangrove forest
292
xiii
11.1.3
SAR regression analysis and biomass estimation
11.2 Application of techniques to West Alligator SAR data 11.2.1 SAR decomposition: scattering mechanisms and their
297 298 299
classification 11.2.2
SAR decomposition: phase and coherence
302
11.2.3 Relationships between SAR backscatter and stand
305
structure: analysis of backscatter profiles 11.2.4 Empirical relationships between SAR backscatter and
310
stand structural attributes
12.
11.2.5 Relationships between SAR backscatter and biomass
316
11.2.6 Relationships between Polarization ratios and biomass
321
11.2.7 SAR retrieval of biomass
324
11.3 Optical and SAR derived biomass: accuracy assessment
328
11.4 Comparative analysis: French Guiana mangroves
330
11.4.1
Forest structure: Crique Fouillée, French Guiana
330
11.4.2
Radar signatures of mangrove stands
335
11.5 Concluding summary
340
MANGROVE DYNAMICS IN KAKADU NATIONAL PARK
345
12.1 Historical trends in estuarine development and mangrove
345
establishment 12.2 Interpretation of mangrove DEMs and orthomosaics
348
12.2.1
The Wildman River
348
12.2.2
The West Alligator River
352
12.2.3
The South Alligator River and Barron Island
357
12.2.4
The East Alligator River
363
12.2.5
Field Island
369
12.3 Change detection
371
12.3.1
Changes in height
372
12.3.2
Changes in extent
373
12.4 Dynamics of mangroves
384
12.4.1
Erosion of coastal mangroves
384
12.4.2
Expansion of coastal mangroves
386
xiv
13.
12.4.3
Inland encroachment of mangroves along tidal creeks
387
12.4.4
Saltwater intrusion into freshwater areas
387
12.4.5
Changing channel morphology and mangrove distribution
389
12.4.6
Cyclone and storm damage
390
12.5 Implications for the coastal environment
391
12.6 Concluding summary
392
DISCUSSION
394
13.1 Requirement of baseline information
395
13.2 Remote sensing for baseline information in the ARR
397
13.2.1
Mangrove extent
398
13.2.2
Mangrove canopy height
400
13.2.3
Tree density and crown delineation
402
13.2.4
Tree species discrimination
404
13.2.5
Above ground and component biomass
406
13.2.6 Change detection
410
13.2.7 Integration of datasets
411
13.3 Environmental change in the ARR 13.3.1 Remote sensing and change detection
415
13.3.2 Mangrove response along the West Alligator River
415
13.3.3 Response of the ARR mangroves to coastal change
424
13.3.4 Future scenarios of climate change
427
13.4 Opportunities for spaceborne remote sensing
14.
412
430
13.4.1 Multispectral satellites
430
13.4.2 Landsat and SPOT sensors
431
13.4.3 Spaceborne SAR
433
13.5 Potential for local to global scale monitoring
437
13.6 Concluding summary
440
CONCLUDING REMARKS
443
14.1 Remote sensing for baseline inventory of mangroves
443
14.1.1
Mangrove extent
444
14.1.2
Mangrove height
445
xv
14.1.3
Tree density
445
14.1.4
Species/community composition
446
14.1.5
Total and component biomass
447
14.2 The ARR and mangrove response to change 14.2.1
448
DEMs and orthomosaics
449
14.2.2 Changes within the ARR
449
14.3 The requirement for baseline datasets
451
14.4 A final thought
453
BIBLIOGRAPHY
454
APPENDICES
477
A
Triangulation reports: stereo aerial photography
477
B
GCPs and registration report: CASI
478
C
GCPs and registration report: AIRSAR
479
xvi
LIST OF TABLES
Page No.
CHAPTER 1 1.1
Examples of mangrove species distributions under different tidal
14
regimes in northern Australia. 1.2
Typical zonation patterns associated with Australian mangal vegetation 17
(Source: Saenger et al., 1977). CHAPTER 2 2.1
Estimates of current and future global sea level rise (cm).
34
2.2
Estimated rates of past, present and future global sea level rise.
34
CHAPTER 3 3.1
A summary of image processing techniques applied to remote sensing
50
data acquired over mangroves (Green et al., 1998). CHAPTER 4 4.1
Average climate data for Jabiru airport, NT.
75
4.2
Vegetation communities of the Arnhem Land Plateau.
79
4.3
Vegetation communities of the ARR lowlands.
80
4.4
Vegetation communities of the ARR floodplains.
82
4.5
Mangrove communities of the West Alligator River.
85
4.6
Common birds of the ARR wetlands.
88
4.7
Common fish of the ARR.
89
4.8
Reptiles and amphibians of the ARR.
90
4.9
Mammals of the ARR.
92
4.10
Feral animals of the ARR, their history of invasion, and current
94
control measures. 4.11
Noxious weeds of the ARR wetlands.
95
CHAPTER 5 5.1
Available aerial photography for the Kakadu region.
104
5.2
CASI-2 sensor specifications.
105
xvii
5.3
CASI wavelength regions and FWHM values.
105
5.4
AIRSAR data acquisition parameters (1996).
107
5.5
AIRSAR instrument specifications.
108
5.6
Map sheets used for geometric rectification of the B&W and colour
114
aerial photographs. 5.7
Statistical outcomes of raw (left) and filtered (right) imagery.
119
CHAPTER 6 6.1
Transect and plot description.
124
6.2
The dimensions of plots used during the 2002 campaign and the
132
measurements recorded. 6.3
Structural parameters measured within each stand.
136
6.4
Percentage foliage cover (PFC) for the East bank mangroves.
143
6.5
Percentage foliage cover (PFC) for the West bank mangroves.
143
6.6
Mean tree height measurements for the East bank mangroves.
145
6.7
Mean tree height measurements for the West bank mangroves.
145
6.8
Minimum and maximum heights and diameters by species and growth
146
stage, West bank mangroves. 6.9
Canopy depths, west bank mangroves.
151
6.10
Mean crown dimensions, area and volume, West bank mangroves.
155
6.11 6.12
-1
160
-1
160
Tree density (N ha ) for various growth stages, East bank mangroves. Tree density (N ha ) for the various growth stages, West bank
mangroves. 6.13
Available allometric equations used in the estimation of total above
163
ground and component biomass. 6.14
Total above ground and component biomass (t DM ha-1) for the East
164
bank mangroves. 6.15
Total above ground and component biomass (t DM ha-1) for the West
164
bank mangroves. 6.16
Height-diameter relationships for the dominant species.
172
6.17
Mean leaf dimensions and density (leaves m-3) for the dominant species. 180
6.18
Mean branch dimensions and density (branches m-3) for the dominant
184
species and growth forms.
xviii
CHAPTER 7 7.1
Estimates of tree density derived through application of the eCognition
214
tree top mapping algorithm to stereo aerial photography. 7.2
Comparison of field based (2002) and image derived (Aerial
217
photography, 1991) estimates of stem density. CHAPTER 9 9.1
The area of the West Alligator River mangroves (north of UTM
251
coordinate 8639000) dominated by different species. 9.2
Confusion matrix for assessment of maximum likelihood classification. 253
9.3
Percentage of species occurring within particular height zones.
258
CHAPTER 10 10.1
Stem densities used in the calculation of biomass.
267
10.2
Total and component biomass derived from optical datasets: estimated
272
ranges within observed size classes for a sample area on the west bank. 10.3
Total and component biomass for the eastern bank mangroves based
273
on average trunk diameters derived from optical datasets. 10.4
Comparison between field and image derived total and component
280
biomass (t 400 m-2). Field averages are in bold type. CHAPTER 11 11.1
Coefficients of determination (r²) between the logarithm of stand
312
structural attributes and backscattering coefficients (dB) at C-, L-, and P-band. 11.2
Coefficients of determination (r²) between the logarithm of total above
321
ground and component biomass and backscattering coefficients at C-, L-, and P-band. 11.3
Regression coefficients for Log pr = (dB + a) / b, where pr is the
327
forest parameter, and dB denotes the backscattering coefficient of the desired frequency. 11.4
Comparison between field based estimates of total biomass for the east
327
bank mangroves, and SAR derived AGB using P-HV data. 11.5
Comparison of SAR and optically derived above ground biomass.
329
The minimum and maximum field plot biomass is included for reference. 11.6
Structural characteristics of the 12 CF stands.
332
xix
11.7
Component biomass of the 12 mangrove stands observed at Crique
332
Coefficients of determination (r²) between forest structural parameters
333
Fouillée. 11.8
for mangroves at Crique Fouillée and West Alligator (in italics). 11.9
Coefficients of determination (r²) between forest structural parameters
337
for mangroves at Crique Fouillée and West Alligator (in italics) and backscattering coefficients at C-, L-, and P-bands. CHAPTER 12 12.1
Major environmental change associated with the Top End estuaries.
347
CHAPTER 13 13.1
Landsat and SPOT satellite specifications.
432
13.2
Sensor specifications: past and current spaceborne SAR sensors
434
(Source: adapted from Baltzer, 2001). 13.3
Optimal SAR sensor parameters for forest mapping (adapted from:
435
Kasischke et al., 1997). 13.4
Accuracy associated with interferometric DEM generation over mixed
436
relief terrain (source: Toutin & Gray, 2000). 13.5
Suggested framework for monitoring and inventory of mangrove
439
ecosystems using remote sensing data.
xx
Page No.
LIST OF ILLUSTRATIONS CHAPTER 1 1.1
The global distribution of mangroves (dark green).
2
1.2
New World (Indo-West Pacific) and Old World (Atlantic East Pacific)
3
mangroves and their distribution (solid lines around coast) within six biogeographic zones (source: Duke et al., 1998). 1.3
The location (above green line) of Zones 1N (largely northern Australia) 16
and 1NE (largely north-eastern Australia). CHAPTER 2 2.1
Simulated global surface temperature change (Source: IPCC, 2001)
32
based on the outcome of several climate models and using a number of different illustrative scenarios. CHAPTER 3 3.1
Areas of mangroves (red and dark purple) on the east coast of
44
Kalimantan, as identified within the GRFM mosaic (1998, 1996 and 1994 in RGB) generated using JERS-1 SAR data from the Japanese Aerospace Exploration Agency (JAXA). 3.2
The conversion of coastal mangroves for aquaculture in Perak,
46
Malaysia as observed using time-series of JERS-1 SAR data. 3.3
a) Fine (2.5 m) spatial resolution colour composite mosaic of the
55
Daintree River mangroves, Queensland, derived from nine CASI scanner flightlines, and b) Spectral Angle (SAM) classification of the broad mangrove associations of the Daintree River (Held et al., 2003). 3.4
SRTM derived (90 m) DEM of estuarine mangroves, Gabon. The use
59
of moderate (30 m) spatial resolution data would provide better resolving of mangrove canopy height (Source: S. Saatchi, NASA JPL). 3.5
The extensive prop roots of Rhizophora stylosa as observed on the West 61
Alligator River, Kakadu NP, Northern Australia.
The size and density of roots
contributed to the low L- and P-band backscatter (through reduced double bounce and surface interaction) observed on AIRSAR data.
xxi
3.6
False color composite of high resolution spaceborne remote sensing
64
data: a) Landsat TM (red=band 5, Green=band 4, blue=band 3), and b) SIR-C (red=Lband HH, green=L-band HV, blue=C-band HH) showing low mangroves dominated by Rhizophora species (1), open mangroves dominated by Avicennia species (2) and also saltflats (3) and transitional flooded forest (4). CHAPTER 4 4.1
The Alligator Rivers Region in northern Australia.
68
4.2
An example of Aboriginal rock art.
69
4.3
Sandstone escarpments of the Arnhem Land Plateau.
71
4.4
The extensive floodplains of the East Alligator River.
72
4.5
Landforms and soils of the ARR.
74
4.6
Mean monthly rainfall and temperature for Jabiru, NT.
75
4.7
Plots of mean monthly sea level for Darwin: a) 1958 – 2000; b) 1990 –
77
2003 (Source: archive data from the National Tide Facility, Adelaide, Australia). 4.8
The major vegetation types within the ARR.
78
4.9
Vegetation of the Arnhem Land plateau (from left to right): Pandanus
79
species, Native Ginger, and Spinifex grass. 4.10
Vegetation of the KNP lowlands (from left to right): Swamp Banksia,
80
Grevillia and Kapok Bush. 4.11
Floodplain vegetation (from left to right): Melaleuca forest, Melaleuca
82
(Paperbark) swamp with Yellow Lilies, and Water Lily. 4.12
Vegetation of the salt flats on the landward edge of the mangrove
83
community: Samphire (Salicornia australis) and Halorisia in foreground. Scattered seedlings of Avicennia marina are also present. 4.13
Homogeneous mangrove stands observed on the West Alligator River
86
(from top clockwise): Avicennia marina, Rhizophora stylosa, Sonneratia alba. 4.14
Birdlife within the ARR.
87
4.15
Common fish found in the ARR (from left to right): the Sooty Grunter,
89
and Chequered Rainbow fish. 4.16
Reptiles and Amphibians of the ARR.
91
4.17
Invertebrates of the ARR (from left to right): a) Termites and their
92
characteristic mounds; b) Leichhardt’s Grasshopper.
xxii
4.18
Introduced animals of the ARR: the Water Buffalo and Cane Toad.
93
4.19
Invasive species of the floodplains and swamps: Eichhornia crassipes
96
(Water Hyacinth), Salvinia molesta (floating weed) and Mimosa pigra (prickly shrub). 4.20
Salinisation of floodplain areas.
98
4.21
Aerial view of Ranger mine, ARR.
98
CHAPTER 5 5.1
1950 Black and white stereo aerial photograph of the mangroves at
103
the mouth of the West Alligator River. 5.2
1991 True colour stereo aerial photograph of the mangroves at the
103
mouth of the West Alligator River 5.3
CASI-2 image of the West Alligator River mangroves: Bands 14
106
(838 nm), 9 (714 nm) and 1 (447 nm) in R, G, B. 5.4
AIRSAR images of the ARR coastal strip (1996): C-, L- and P-bands
109
(HH, VV and HV in R, G, B respectively. 5.5
An overview of the procedure for determining the exterior orientation
112
for a stereo photographic pair in the point measurement phase of aerial triangulation. Ground Control Points (GCPs) and image points (Tie points) common to overlapping images are identified, and their ground coordinates and accuracies recorded. 5.6
Topographic map of the ARR with the blocks of photographs
113
highlighted. 5.7
The image rectification process involving rotation along the epipolar
116
plane and resampling of stereo pairs. CHAPTER 6 6.1
Field sampling locations, East bank, October 1999: a) Transect 1
122
upstream, and b) Transect 2 towards river mouth. 6.2
Approximate location of the three sampling sites (Transect 2) on the
126
East bank, October 1999, overlain on an aerial photograph. 6.3
Schematic diagram of successive mangrove stands observed along
127
Transect 2, East bank. 6.4
Field sampling area (October 2002) on the west bank of the West
130
Alligator River, and the main zones, as observed using the CASI data (inset).
xxiii
6.5
Location of the thirteen sample plots on the West bank, Sept/Oct, 2002, 131
using the CASI scene as a backdrop. 6.6
Examples of tree layering and structural measurements within different
133
mangrove zones. 6.7
Estimating the leaf density within crowns of A.marina using a quadrat
136
a) Profile diagram illustrating the transitions between zones along a
138
cube. 6.8
tidal creek community (Transect 1); and b) Profile diagram along a seaward facing community towards the West Alligator River mouth (Transect 2). 6.9
Profile diagram illustrating the transitions between zones along a
140
seaward facing community, West bank. 6.10
Percentage frequency distributions for tree height, West bank
147
mangroves. 6.11
Percentage frequency distributions for trunk diameter, West bank
148
mangroves. 6.12
Percentage frequency distributions of crown depth, West bank
151
mangroves. 6.13
a) Percentage frequency distribution of crown area and volume, for
156
A.marina stands. 6.13
b) Percentage frequency distribution of crown area and volume, for
157
R.stylosa stands. 6.13
c) Percentage frequency distribution of crown area and volume, for
158
S.alba stands. 6.14
Variation in total and component biomass with tree height, for the east
166
(top) and west (bottom) bank mangroves. 6.15
Biomass proportioning for the dominant species, based on 2002 data.
169
6.16
Error margin associated with R.stylosa allometric; and extrapolation
170
based on linear regression beyond DBH 13.5 cm. 6.17
Relationship between tree height and trunk diameter for A.marina.
172
6.18
Relationship between tree height and trunk diameter for R.stylosa.
173
6.19
Relationship between tree height and trunk diameter for S.alba: a) all
174
data included; and b) 6 outliers removed. 6.20
Growth trajectories for the main species.
175
xxiv
6.21
Tree crown size and canopy distributions: the variation in trunk
176
diameter with a) crown radius, and b) canopy top height for A.marina (Am), R.stylosa (Rs), and S.alba (Sa). 6.22
Tree crown and canopy dimensions: the variation in tree height with a)
178
crown radius; and b) canopy depth for A.marina (Am), R.stylosa (Rs), and S.alba (Sa). 6.23
Calculation of leaf length and width using ENVI’s measurement tool.
179
6.24
Spectral reflectance curves for the main mangrove species, recorded
181
using an ASD Field spectroradiometer. 6.25
Image based calculation of branch length and diameter using ENVI.
182
6.26
Calculation of branch orientation angles: a) delineating polygons in
183
ENVI; b) method of calculation using Pythagora’s Theorem; c) Probability distribution. CHAPTER 7 7.1
Flow chart outlining the DEM generation process whereby ATE is
190
performed in the area of overlap of a stereo pair. 7.2
DEM of mangrove canopy height for the Wildman River.
194
7.3
DEM of mangrove canopy height for the West Alligator River,
195
containing some of the most productive mangrove communities within Kakadu NP. The tall forest (up to 25 m in height) on the lower east bank consisted largely of Eucalyptus species. 7.4
a) DEM of mangrove canopy height for the South Alligator River, west 196
bank; b) DEM of mangrove canopy height for the South Alligator River, east bank. Quite tall mangroves (up to 14 m in height) can be observed on the inland river bend. 7.5
DEM of mangrove canopy height for Field Island.
198
7.6
a) DEM of mangrove canopy height for the East Alligator River,
199
coastal section. The tall forest (up to 24 m in height) on the inland lower bank consisted largely of Eucalyptus species; b) DEM of mangrove canopy height for the East Alligator River, lower section. 7.7
Comparison of DEMs generated at a post spacing of: a) 20 m, b) 5 m,
202
c) 1 m, and d) Interpolated 1 m DEM (from 5 m). 7.8
Height profiles extracted from the DEMs generated at a post spacing
203
of: a) 20 m, b) 5 m, c) 1 m, and d) Interpolated 1 m DEM (from 5 m). 7.9
Inconsistencies in height estimation associated with DEM generation
205
xxv
for open forest: a) 5 m DEM, and b) Profile extracted through mangroves.
The
mangroves (green) represent coastal stands of mature S.alba (averaging ~12 m height), which were under-estimated by around 8 m. Where patches of denser trees were observed (yellow – orange) the canopy heights were better estimated in the DEM. 7.10
West bank mangroves: comparison of DEM derived heights based on
207
stereo aerial photography (1991) and tree height as measured in the field (2002). 7.11
East bank mangroves: comparison of DEM derived heights based on
207
stereo aerial photography (1991) and tree height as measured in the field (1999). 7.12
DEM generated using: a) 1950 black and white stereo aerial
209
photography; and b) 1991 true colour stereo aerial photography. 7.13
a) Height profiles extracted from the DEM generated with stereo
211
aerial photography acquired in 1950 and 1991; and b) Subsets taken from aerial photographs and DEMs showing approximate position of transects. 7.14
50 x 50 m plots in mangroves dominated by a) A.marina and b)
215
R.stylosa showing the results of the crown top mapping algorithm developed within eCognition and applied to aerial photography (green band). 50 x 50 m plots in c) mixed mangrove forest dominated by R.stylosa and d) forests dominated by S.alba showing the locations of trees mapped using eCognition. 7.15
Crown tops mapped using eCognition and the slope of the red edge
220
image, as derived using a combination of CASI bands centred on 680 nm and 740 nm. 7.16
NDVI profile across the mangrove community as related to a) stem
222
density; and b) tree height, percentage foliage cover (PFC) and crown area. Profile coordinates: 202740 E, 8646837 S (landward), 203655 E, 8647018 S (seaward). CHAPTER 8 8.1
Projecting ground space to image space in the mosaic process.
227
8.2
Flow chart outlining the generation of an orthomosaic.
227
8.3
Orthomosaic for the Wildman River, ARR, generated using stereo
231
aerial photography. 8.4
Orthomosaic for the West Alligator River, ARR, generated using stereo 232
aerial photography. 8.5
a) Orthomosaic for the South Alligator River (west bank), ARR,
233
xxvi
generated using stereo aerial photography; b) Orthomosaic for the South Alligator River (east bank), ARR, generated using stereo aerial photography. 8.6
Orthomosaic for Field Island, ARR, generated using stereo aerial
235
photography. 8.7
a) Orthomosaic for the East Alligator River (coastal section), ARR,
236
generated using stereo aerial photography; b) Orthomosaic for the East Alligator River (lower section), ARR, generated using stereo aerial photography. CHAPTER 9 9.1
a) Subset taken from the West Alligator orthomosaic; and b)
243
Unsupervised ISODATA classification of the subset, illustrating the boundary between mangrove (green) and bare mudflat areas or water (white). The areas of mangrove in yellow represent A.marina species, which were discriminated from the others. Areas of confusion were associated with the mudflats and other vegetation on the eastern side of the image. 9.2
Distribution of vegetated and non-vegetated surfaces as mapped by
245
thresholding the NDVI calculated from 1 m spatial resolution CASI data. 9.3
Reflectance spectra (image derived) and illustrations of observed
246
surface covers from CASI data. 9.4
Image derived (solid lines) and field derived (dotted lines) spectra for
248
the dominant mangrove species and two unknowns (Sp1 and Sp2) identified in the CASI. 9.5
Leaves of the dominant mangrove species used for spectral
249
measurements. 9.6
Spectral reflectance for the main mangrove species acquired in 2002
250
using an ASD Field spectrometer. 9.7
Maximum likelihood classification of mangroves using CASI data.
252
9.8
Species distribution and zonation of mangroves on the west bank.
255
9.9
Species distribution and zonation of mangroves on the east bank.
255
9.10
Distribution of species along the tidal creeks draining into the main
256
channel of the West Alligator River. 9.11
Distribution of species along an intruding tidal creek on the lower east
257
bank of the West Alligator River.
xxvii
CHAPTER 10 10.1
Flow chart illustrating the use of conditional models for determination
266
of trunk diameter and component biomass. The first model uses an equation with height as the independent variable (height = a.ln (DBH)+b) to calculate diameter directly from the DEM. The second model applies allometric equations and an adjustment factor (N/m²) to the diameter and species information assigned to each pixel to calculate biomass. 10.2
Spatial dataset of trunk diameter (cm) generated with optical datasets.
269
Inset and detail provided for West and East bank mangroves. 10.3
Spatial datasets of total above ground and component biomass
270
-1
(t DM ha ) generated using optical data. Detail illustrated for West bank. 10.4 10.5
Comparing field and image (from the DEM) derived trunk diameters. -2
Comparing actual and predicated above ground biomass (t 400 m ) for
277 279
the different species. CHAPTER 11 11.1
Scattering from: a) specular (smooth) surfaces which generally scatters
285
in one direction away from the sensor; and b) diffuse (rougher) surfaces, which scatter in all directions, with some radiation being received by the sensor. 11.2
Sample L-band polarization signatures extracted over coastal S.alba
289
forest, West Alligator River mangroves: a) co-polarized and b) cross-polarized response. 11.3
Backscattering pathways in the mangrove forest, including the trunk-
290
ground (double bounce) interaction (Tr-grd), trunk scattering (Tr), ground scattering (Grd), canopy (volume) scattering (Cpy), and canopy-ground interaction (Cpy-grd). 11.4
Sample C-band coherence and phase images (HHVV, HHHV and
291
VVHV in R, G, B respectively). In the coherence image, mangrove forest is located in the bright blue areas, mud flats are predominantly red, and water is dark, almost black. 11.5
Identification of odd and double bounce and volume scattering in
301
mangrove forest. Bright areas represent the dominance of a particular mechanism. Subset photograph and DEM are provided for reference. 11.6
a) C-band HH-VV coherence, and b) C-band HH-HV coherence,
303
xxviii
overlain on subset with C-HH, VV and HV in R, G, B respectively. Values 0 – 1 represent low – high coherence. 11.7
a) L-band HH-VV coherence, and b) L-band HH-HV coherence,
304
overlain on subset with L-HH, VV and HV in R, G, B respectively. Values 0 – 1 represent low – high coherence. 11.8
a) C-band HH-VV phase; b) L-band HH-VV phase; and c) P-band
304
HH-VV phase, overlain on subset with HH, VV, HV in R, G, B respectively for all bands. 11.9
a-c) SAR backscatter profiles, at C-, L- and P-bands, from the landward 308
to seaward edge across the mangrove community; d) field based estimates of height and density across transect; and e) subset from aerial photograph showing location of transect on west bank of West Alligator River. 11.10
a-c) SAR backscatter profiles, at C-, L- and P-bands, from the landward 309
to seaward edge across the mangrove community; d) DEM derived tree heights across transect; and e) subset from aerial photograph showing location of transect at Point Farewell, East Alligator River. 11.11
Relationships between C-, L-, and P-band backscatter at HH, VV and
311
HV polarizations and the forest structural attributes of height, density and diameter. 11.12
C-band signatures vs. stem density for each diameter class. Species
313
codes: A.marina in blue, R.stylosa in red, R.stylosa regrowth in orange, and S.alba in green. 11.13
L-band signatures vs. stem density for each diameter class. Species
314
codes: A.marina in blue, R.stylosa in red, R.stylosa regrowth in orange, and S.alba in green. 11.14
P-band signatures vs. stem density for each diameter class. Species
315
codes: A.marina in blue, R.stylosa in red, R.stylosa regrowth in orange, and S.alba in green. 11.15
Relationships between C-, L-, and P-band backscatter at HH, VV and
317
HV polarizations and total and component (leaf, branch, trunk and root) biomass. 11.16
Above ground root structures for the dominant mangrove species. The
320
increase in the magnitude of P-band HH return with total biomass is most likely linked to the development of extensive above ground root systems, as for R.stylosa forest 11.17
Copolarization ratio (σºVV/HH) in relation to total biomass at C, L and P
322
xxix
bands. Tendency lines are indicated for each frequency. 11.18
Depolarization ratios: a) σºVV/HV and b) σºHH/HV in relation to total
324
biomass at C, L and P bands. 11.19
SAR derived biomass (t DM ha-1) derived using an empirical
326
relationship with P-HV for a) the whole ARR strip; and b) the West Alligator River. The aerial photograph (c) and DEM (d) of the West Alligator are given for reference. 11.20
SAR backscatter profile (P-HV) from landward to seaward edge across
328
east bank mangroves, compared with DEM derived canopy height and SAR derived AGB (P-HV; t DM ha-1). 11.21
Site comparison: relationships between mangrove structural attributes
334
and biomass. 11.22
Comparison of allometrics used to determine total above ground
335
biomass (kg) for Avicennia and Rhizophora stands at CF and WA. 11.23
C-, L-, and P-band multipolarized SAR signatures vs. forest structural
336
parameters at CF. 11.24
Backscattering coefficients vs. total above ground biomass at both
337
sites considering the whole biomass range. 11.25
Backscattering coefficients vs. total above ground biomass at both
338
-1
sites considering the biomass range up to 250 t DM ha . CHAPTER 12 12.1
The distribution of mangroves along the Wildman River. Points of
349
interest are marked with crosses and mentioned in the text. 12.2
a) Erosion of channel banks and loss of mangroves, West bank,
350
Wildman River; and b) Corresponding area on DEM. 12.3
Young growth associated with mangroves on the coastal edge.
351
12.4
Canopy height profiles associated with mangroves inhabiting
351
downstream meander loops. 12.5
An example of tidal creek incision across the coastal plain, with
352
subsequent colonization by mangroves. 12.6
The distribution of mangroves along the West Alligator River. Points
353
of interest are marked with crosses and mentioned in the text. 12.7
The abrupt change from low open A.marina to tall mature R.stylosa
354
xxx
forest on the landward edge of the community. Samphires and other salt tolerant species located in foreground. 12.8
The Rhizophora “ridge” on both east and west banks and corresponding 355
area from DEM. 12.9
Section of tidal creek extending across salt flats on east bank of river,
356
and corresponding area on DEM. 12.10
Extension of tidal creek, west bank, and formation of swamp adjacent
356
to upland Eucalypt forest. 12.11
The distribution of mangroves along the South Alligator River and
358
Barron Island. Points of interest are marked with crosses and mentioned in the text. 12.12
a) Significant mangrove stands located on the west bank of the South
359
Alligator River; and b) Height profiles extracted from DEM. 12.13
a) Backwater mangroves associated with a meander loop on Brook
360
creek, east bank; and b) corresponding area on DEM. 12.14
a) Mangrove zonation along the South Alligator River coastline, east
360
bank; and b) Canopy profile from DEM. 12.15
Fringing mangroves surrounding Barron Island.
362
12.16
Mangroves of the northern section of Barron Island, and formation of
362
an isle off the coast. To the north-west, episodic coastal erosion is removing sections of mangrove from the seaward edge community. 12.17
The distribution of mangroves along the East Alligator River. Points of 363
interest are marked with crosses and mentioned in the text. 12.18
a) Mangroves surrounding Point Farewell, East Alligator River. Inset
364
shows tidal creek section taken from DEM; and b) Canopy profiles extracted from mangroves on the headland and along coast. 12.19
Landward extension of tidal creeks and intrusion into Paperbark swamps. 365
12.20
a) The spread of mangroves along a tidal creek, upper East Alligator
366
River, with evidence of alluvium deposits; and b) the inland extension and narrowing of a tidal creek, lower East Alligator River. 12.21
Formation of point bars and islands within main river channel and
367
subsequent establishment of mangroves. 12.22
Sedimentary deposit on convex bank and subsequent ribbon–like
367
mangrove development.
xxxi
12.23
Pockets of expansive mangroves along a tidal creek and channel
368
shoreline, upper East Alligator River. 12.24
The distribution of mangroves along Field Island. Points of interest are 369
marked with crosses and mentioned in the text. 12.25
Extensive mangroves on south-east corner of Field Island (landward to
370
seaward across community), and corresponding areas on DEM. 12.26
Extensive mangrove forest along an inland tidal creek, north-west
371
section of Field Island and corresponding area taken from DEM. 12.27
Mangrove regrowth on northern section of Field Island, and
371
corresponding area taken from DEM. 12.28
The retreat (red) and expansion (blue) of mangroves along the West
374
Alligator River between 1950 and 1991, derived from stereo photography. 12.29
Channel development, formation of cutoffs and subsequent colonisation 375
by mangroves, West bank. Areas of retreat and expansion are illustrated in red and blue respectively. 12.30
Tidal creek extension and subsequent spread of mangroves, East bank.
375
Areas of retreat and expansion are illustrated in red and blue respectively. 12.31
The retreat / erosion (red) and expansion (blue) of mangroves along the 376
West Alligator River between 1991 and 2002, derived from stereo aerial photography and CASI data. 12.32
Formation of new S.alba zone on upper west bank: a) 1991 air
377
photograph; b) 2002 CASI; c) Areas of new growth illustrated in blue. 12.33
Substantial growth of A.marina along creeklines draining towards the
379
coast, on upper west bank: a) 1991 aerial photograph; b) 2002 CASI image, where larger crown sizes of A.marina and new growth along and within the channel is observed (A.marina trees identified in yellow, R.stylosa in red, and S.alba in green). 12.34
Changes in the extent of landward edge of R.stylosa (> 18-24 m in
383
height) between 1991 and 2002: a) The landward edge delineated (in white) on the 2002 CASI image, and b) the same edge superimposed on the 1991 true colour stereo aerial photograph. Over this 11-year period, movements over 50 m are commonplace and can be as high as 150 m.
xxxii
CHAPTER 13 13.1
Rate of mean sea level change (estimated simultaneously with annual
414
and semi-annual variations) as measured using TOPEX/POSEIDON Altimeter data. The trends were calculated for the period from December 1992 to June 2002.
xxxiii
CHAPTER ONE
1.1
INTRODUCTION.
Mangroves: a definition
Mangroves are defined as “an ecological group or complex of halophytic tree and shrub species exceeding 1 m in height and inhabiting tidal lands, mainly in the tropical and subtropical regions, although extending to temperate regions”.
The term
‘mangrove’ is generally applied to both individual genera and species but also to the ecosystem as a whole.
However, mangrove formations dominating a shoreline
ecosystem (and associated with mud flats) are also commonly referred to as mangal (Chapman, 1977). Mangal is restricted largely to saltwater or brackish areas, with the growth of some species being optimal under certain salinities. Mangroves establish within a range of coastal environments but are found most frequently along sheltered coastlines where wave action is sufficiently low to allow seedlings to establish and survive, and are particularly widespread where land is encroaching on the sea. Mangroves are also commonplace within estuaries, where they favour areas where there is an abundance of silt, and may penetrate further inland. Mangroves also occur on stable sandy shores and coral reefs where little accretion occurs. The extent and vertical range of mangroves is determined largely by the range and frequency of the tides. Where both are large, extensive mangrove communities may establish and the diversity of communities may be greater. Fringe communities develop more along coastlines where the shelving is steep.
1
1.2
The global distribution of mangroves
Mangroves are a floristically diverse assemblage of plants, widespread throughout the tropics and subtropics, and also extending to more temperate zones in both the northern and southern hemispheres (32o N and 38o S respectively; Figure 1.1). Mangroves flourish particularly well in the equatorial regions and in the tropical summer-rainfall regions to the north and south, but also occur in some subtropical dry regions, reaching their geographic limit in regions described as warm tropical (e.g., Japan and New Zealand). Mangroves are mostly absent from the Central Pacific. The restriction of mangroves to these regions is associated with the sensitivity of mangrove species to temperature and particularly frost.
For example, the lack of
mangroves on the west coast of North America and Africa is attributed to intolerance of the low water temperatures associated with cold ocean currents. In general, mangroves flourish where the temperature in the coldest month does not fall below 20oC and where the range is about 10oC (Chapman, 1977).
Figure 1.1 The global distribution of mangroves (dark green) (Source: Chapman, 1977).
2
A marked difference in the diversity, productivity and structure of mangroves occurs throughout their distributional range. Specifically, trees as high as 30 m and with 75 – 100 % canopy cover are common in equatorial regions, but become shorter and shrubbier at the northern and southern limits. Species diversity also decreases from a maximum in the Malayan region to a minimum in the temperate regions (e.g., New Zealand, Japan). Zonation patterns, which are a common characteristic of mangroves, vary as a function of climate and also species composition. Within their range, two distinct mangrove formations associated with the eastern hemisphere (Atlantic East Pacific) and western hemisphere (Indo-West Pacific) have been recognised (Duke et al., 1998; Figure 1.2). These formations are referred to as the Old and New Worlds respectively (Taal, 1994). Although the formations are similar in terms of their physiognomy and ecological relationships, the former contains a greater diversity of species. The prevailing climate (particularly temperature) and geographic suitability are key determining factors in the worldwide distribution of mangroves in these regions.
Figure 1.2 New World (Indo-West Pacific) and Old World (Atlantic East Pacific) mangroves and their distribution (solid lines around coast) within six biogeographic zones (source: Duke et al., 1998). Their poleward extent is restricted by the winter position of the 20ºC isotherm for seawater (grey areas), which is under the direct influence of oceanic currents (arrows).
3
1.2.1 Old World Mangroves These mangroves occur along the tropical and subtropical coasts of the Indian and West Pacific oceans (Chapman, 1977). Within this formation, those within the IndoMalaysian Group (which encompasses mangroves of south and south-east Asia, including northern Australia) are the most diverse and complex. Within the BurmaIndo-Malaysian Subgroup (particularly along the Malay Peninsular and neighbouring the Greater Sunda Islands), the greatest number of species of any mangal is found. The Australian mangrove differs in some respect from the rest of the eastern formation and can be considered as a distinct sub-formation. 1.2.2 New World Mangroves The New World mangroves occur along the coasts of North, Central and South America, the West Indes and West Africa. Within this formation, species diversity is low with Rhizophora mangle, Avicennia germinans, Laguncularia racemosa and Conocarpus erectus being dominant in most regions, although other species may occur. All of the genera in the New World are found in the Old World formation, but the species differ. Fiji and the Tonga islands in the Pacific are alone in containing species of Rhizophora (i.e., R.mucronata and R.mangle respectively) from both formations.
1.3
Eco-physiology of mangroves
Mangroves are halophytes, a group of plants specially adapted to inhabiting saline environments, including the world’s deserts, salt marshes and coastlines (Raven et al., 1992). As such, they can withstand highly variable salinity regimes, with accumulating salt levels in their tissues often comparable to that of seawater. This tolerance to changing salinity is species specific, and is an important contributing factor in defining the habitat range of a particular species. Mangroves are most prolific in the intertidal zone, which is considered to be the transitional zone between the distinctly terrestrial and marine ecosystems. Communities form above the mean sea level, the landward edge of which is controlled by mainly climatic (rainfall and temperature) and physiographic factors (soil type and salinity).
4
The seaward edge of mangroves is restricted to those areas where trees are not submerged for greater than 30 % of the time (Saenger, 1994). Within this environment, the distribution of mangroves is a function of the sensitivity and adaptability of different species to the prevailing environmental conditions (Blasco et al., 1996).
Key elements influencing mangrove distribution include drainage
patterns, frequency of tidal inundation, underlying soil type and salinity (Semeniuk et al., 1978). Many of these factors interplay to generate gradients from seaward to inland areas, thereby establishing the dynamics of the intertidal zone. Within such a setting, the restricted range of conditions over which different mangrove species can survive typically leads to the formation of distinct mangrove zones that parallel the seaward margins (Lear & Turner, 1994). Mangrove zones are distinctive vegetation bands, comprised of one or more species that occur parallel to the shoreline. Zonation is not strict in the sense that one zone defines the habitat range of a species.
Rather it defines either a pure or mixed
assemblage wherein one or two species may be dominant. Any one particular species may occur throughout the entire community, depending on its sensitivity to key environmental factors. Due to the sensitivity of mangroves to salinity, soil type and inundation, and their vulnerability to, for example, storm surges, these ecosystems may be useful biological indicators of coastal environmental change (Ellison & Farnsworth, 1997; Field, 1995). As different species demonstrate a restricted range in tolerance to external conditions, any changes in these conditions will lead to alterations in mangrove extent, floristic composition and structure, and hence zonation. Such change may be rapid or more gradual depending on site physiography and the resilience of species.
1.4
Importance of mangroves
Mangroves are an important component of coastal ecosystems in many tropical and subtropical regions.
As mangroves occur at the land-sea interface, they create a
shoreline buffer zone that protects the coast from the ravages of erosion and flooding.
5
Mangroves are also important contributors to primary production in the near-shore environment, efficiently recycling nutrients and providing a continual supply of detritus to the marine community (Bandaranayake, 1994). Where seagrass meadows occur on the seaward edge of the mangrove forest, this function provides a valuable nutrient source. Communities also provide nurseries and shelter for a variety of commercially important fish and crustaceans (O’Grady et al., 1996), and act as a seasonal base for a variety of migratory species. They also contribute to groundwater recharge (through filtering of pollutants), nutrient and sediment retention, and shoreline stabilisation (Saenger, 1994; Blasco et al., 1996). For centuries, primary materials from mangroves have been utilized by humans. For many local populations, mangroves have provided a means of livelihood and a valuable source of food and natural products. Early uses included the construction of sea-going vessels, buildings and furniture, with the wood being held in particularly high regard because of its resistance to termite attacks (Walsh, 1977). Over time, mangroves have been exploited increasingly for timber, firewood, charcoal and paper pulp. Products from mangroves have also included tannin, dyebark, plywood adhesive and high alpha (dissolving) pulps for the manufacture of rayon, cellophane, lacquers, cellulose acetate and other cellulose derivatives. Some genera (e.g., Nypa) are used for thatch, alcohol and sugar, and charcoal from others has been used as a substitute for petroleum coke and in the manufacture of calcium carbide and ferro alloys which, in turn, are used in the chemical, plastics and metal industries. Plant oils and chemical extracts for medical products have also originated from mangroves (Primavera, 1997). Some species have edible plant products, providing a food source to local populations. Increasingly, the value of mangroves both intrinsically and as a viable resource, is being recognised by the global community. The Mangrove forest is most certainly a key contributor to the biodiveristy and productivity in the near-shore environment. Mangrove forests are also considered to be renewable, an indication of their inherent resilience to external influences over long time-scales.
6
1.5
Natural and anthropogenic change
Over time, mangroves have been subjected to significant change, either through natural events (changes in the coastal environment or climate) or direct/indirect anthropogenic effects. 1.5.1 Environmental and climate change Historically, the extent of mangroves has changed significantly, largely due to changing climatic conditions. Mangroves have responded to such change, surviving and rapidly colonising areas where suitable conditions prevailed but disappearing or degrading when conditions became adverse.
Within Australia, for example, fossil
pollen and wood evidence suggests a wider distribution of mangroves than present and a loss of tropical mangal from southern coasts since the Eocene (Saenger et al., 1977). The recent response of mangroves to changes associated with global climate has received little attention and yet, as Ellison & Farnsworth (1997) suggest, mangroves will be particularly sensitive to the associated alterations in sea level as induced by global warming. As a result of a doubling of carbon dioxide and other greenhouse gases, otherwise known as the ‘greenhouse effect’ (Pittock, 1988), a rise in sea levels worldwide is anticipated. Global sea levels are predicted to rise by 3 – 10 cm per decade over the next century (IPCC, 1997), and the past 50 years have already recorded an average rise in sea level of 10 cm (Henderson-Sellers & Blong, 1989). Quite extensive changes at the ecosystem level are anticipated as a result of future climate change. The increase in extreme events (e.g., tropical storms and cyclones), which has been associated with global warming, is also expected to have a major impact on the vegetation of coastal areas, including mangroves. On a regional scale, mangroves along tropical shorelines may be subject to temporary flooding (Pittock, 1988), the effects of shoreline change, including increased erosion, changing sediment supply and tidal ranges, and saltwater intrusion (Henderson-Sellers & Blong, 1989; Field, 1995). Changes in annual temperatures and rainfall regimes will have varying influence as well. Other potential impacts include increased severity of frosts at the northern (e.g.,
7
Florida) and southern limits and associated mortality of species, increases in the hydroperiods and depths of inundation preventing establishment of mangroves, and changing biogeochemical characteristics. Blasco et al. (1996), based on observations in several regions, concluded that increases in mangrove mortality would be apparent with even minor variations in tidal or hydrological regimes. The impact of such change on the mangroves is often difficult to detect because a) large areas are being rapidly destroyed and b) there is a lack of baseline information against which to assess change. In the former case, the impacts of sea level rise and other coastal changes are best identified for areas where there has been little or no anthropogenic disturbance (e.g., northern Australia).
As such, only those changes
attributed to climate-induced phenomena may be identified. Regional effects of mangrove adjustment to climate change may involve either the expansion or reduction in community range. If suitable substrate exists, and the rate of sea level rise does not exceed the rate of vertical accretion, then communities may actually expand on coastal shores. The inland encroachment of mangroves on adjacent habitats may also occur. Where changes in the coastal environment lead to the partial modification or total destruction of mangroves, the most likely impacts will involve a loss of stability in shoreline areas and a significant increase in coastal erosion. Changes in mangrove distribution will lead to changes in habitat quality for species associated with these habitats as well as those that are migrating.
Furthermore, saltwater
inundation into adjacent freshwater wetlands may also lead to loss or alteration of biodiversity. 1.5.2 Anthropogenic impacts Despite their ecological importance, mangroves have been exploited at an extraordinary rate. Mangroves have become increasingly threatened worldwide, due largely to their clearance, degradation and retraction (Ramirez-Garcia et al., 1998; Finlayson et al., 1999). In 1980, the global estimate of mangroves was 19.8 million ha. The IUCN World Mangrove Atlas (IUCN) estimated that the global area of mangroves at the end of the 20th century was ~18 million ha. However, the ITTO estimated that 100,000 ha have been destroyed annually and few active areas of expansion are
8
occurring. The FAO in 2003 estimated that, by the end of 2000, only 15 million ha of mangroves remained. Although deforestation of mangroves is ongoing, the rate has slowed from 1.7 % (1980-1990) to 1 % (1990 to 2000) per year. With the traditional viewpoint that mangroves are wastelands (Franks & Falconer, 1999) and perpetual breeding grounds for mosquitoes, extensive areas have been cleared, drained, or otherwise modified. The majority of wetlands in the Indian River Lagoon have been drained, filled or impounded since the 1940s, and large quantities of pesticides including DDT were applied to control breeding areas (Schmalzer, 1995). Chemical contamination of extensive wetland and mangrove areas has also occurred. Chemical toxicants released by the US between 1962 and 1971, destroyed over 104,123 ha of the 400,000 ha mangrove forest in Vietnam (Granich, 1993: Global warming and Vietnam). Many areas continue to be used as a dumping ground for urban waste and sewage sludge (e.g., in the Pacific Islands and Venezuela). Other threats arise from oil spillage and draining of cooling water from power stations (Walsh, 1977). Despite the growing interest in conservation of biological diversity and awareness of the scale of human intervention in the latter part of this century, extensive areas of mangrove worldwide are still being cleared for timber harvesting, conversion to agriculture, mariculture and urban development. Mangroves provide a valuable source of timber and timber products, although with poor management, this use often becomes non-sustainable. In many regions, the land on which mangroves establish is considered more productive if used for agriculture. For example, mangal soils are highly fertile and clearance has occurred to allow, for example, rice production or establishment of coconut plantations. Even in Australia, some areas of mangrove were replaced to support the production of sugar cane (Walsh, 1977). In Guinea, West Africa, increased targeting of mangrove forest for rice farming over the last 15 years, has devastated mangrove areas through intense habitat modification (Wolanski & Cassagne, 2000). Following cropping, the mangrove forest is, in general, unable to re-establish in those areas where the tidal drainage has not been restored. Considerable damage and further losses of mangrove have also occurred through coastal development (e.g., in Australia, Malaysia, central America and South Africa), particularly for tourism.
9
The degradation of wetland areas including mangroves, still continues with little public recognition of the causes or consequences (Davis & Froend, 1999). Fertilisers, pesticides, heavy metals, storm water runoff, effluent and industrial discharges, are typical contaminants that make their way either overland or underground to the intertidal/subtidal zones. Mangrove communities can, to a degree, act as efficient filters of certain nutrient and chemical concentrations, however the far-reaching effects of contamination (bio-magnification and bio-accumulation within wetland food chains) are largely unknown (Davis & Froend, 1999). Mangroves are also a vital resource for many coastal peoples, particularly in developing nations of the Asia-Pacific region.
The main human disturbance to
communities is their intense harvesting for fuelwood and charcoal, a practice that has led to large-scale forest structural changes. As a result of repeat cutting of the most sizeable trees in the Haitian mangroves (largely due to clearing, heavy fishing and conversion to residential sites), communities have been reduced to a state of regeneration/regrowth (Aubé & Caron, 2001). Consequently, mangrove surface area and tree size is declining, fish catches are diminishing, fauna is less abundant, and some species of birds and oysters are becoming rare. Similarly on the Pacific Island of Kosrae (Federated States of Micronesia), the capacity for regeneration of mangroves may eventually be severely compromised if intense harvesting continues (Allen et al., 2001). This picture is only too common in those areas with heavy dependence on a particular resource, and where little information exists on the extent of depletion/degradation, and the prevailing socio-economic and environmental conditions. By far the greatest threat to mangrove forests in recent time has been their extensive clearing for construction of artificial shrimp and fishponds with the rapidly expanding aquaculture and mariculture (commercial production of fish and invertebrates) industries. These industries have grown exponentially over the last 15 years, and globally as much as 50 % of mangrove destruction has been due to clear cutting for shrimp farms. Thousands of hectares have been destroyed with the greatest losses observed in South East Asia and Latin America (Panapitukkul et al., 1998; Primavera, 1997). In general, these industries provide a greater net income compared to other uses. Some fisheries (e.g., oysters) have, however, developed without the need for extensive
10
clearance. Extensive culture is the most widely practiced system in shrimp production and is the main culprit in mangrove conversion (Talaue-McManus, 2000). The system is largely dependent on tidal regimes for water and food inputs; hence coastal mangroves provide the best location. Aside from the mass destruction of mangrove areas for private uni-functional ponds, this type of aquaculture has the potential to impact on adjacent ecosystems through eutrophication and effluent discharge and the introduction of diseases and chemicals into the marine environment (Edyvane, 1999). The arguments for conserving mangroves are compelling. In particular, mangroves are an important protective habitat for many juvenile fish and hence the long-term survival of coastal (estuarine and offshore) fishing industries (for game and food) is dependent upon their preservation. The large quantities of organic material delivered by mangroves also support detritus-based food webs, which are essential for long-term survival of fisheries. With rising sea levels, also, the requirement for coastal protection is likely to increase in many regions (e.g., Pacific Rim islands) following the removal of shoreline mangroves. With these factors in mind, a more sustainable use of mangroves is desirable and possible.
1.6
Identifying change in mangrove environments
To assess both the short and longer-term impacts of coastal environmental change and the impact of anthropogenic activities on mangrove ecosystems, the establishment of baseline datasets relating to key attributes which can be detected and monitored over time are essential.
Such attributes include the extent, structure, biomass and
productivity, and species/community composition of mangroves. 1.6.1 Extent Mangroves are naturally dynamic and are highly responsive to environmental change (e.g., storm damage, increased sedimentation). Mangroves can expand or contract in response to changing coastal conditions, and as such, directional effects can be observed associated with prograding or receding coastlines. Measuring the areal extent of change and calculating the rate at which it occurs, may help determine the magnitude and relative importance of influential factors. Mangroves are opportunistic species, in that
11
they will colonize and even expand their distribution when there is available substrate and optimal conditions for growth (i.e., saline, inundated, with some freshwater input). Where mangroves occur in an estuarine system, the sheltered conditions provided by the channel network provide favourable habitat for mangroves. Where creeks extend inland and where no other vegetation community dominates the river banks, mangroves will most likely spread along the length of the creek. Where other vegetation is present, mangroves may invade and compete for space, thereby increasing their range and dominance within the community. 1.6.2 Structure and growth stage Structure can be quantified by considering attributes such as the density of trees within different size classes and also the height of the canopy. Different patterns in branching (size, density and orientation), crown structure (volume), and also root development (buttress or stilt) are also evident and vary by species and growing conditions. Layering within the canopy is also indicative of growth stage, with different size classes and densities of leaves and branches allocated to each layer. Three layers, for example, with the top as the canopy, the middle consisting of the trunk and mostly branches, and the bottom as the root layer, may define a mature Rhizophora forest, whereas only two layers (canopy and trunk/root) may occur within regenerating Rhizophora forests. Interspecific competition between the individuals present in a canopy may lead to size constraints in branch length and crown area. Where space is limiting, crowns tend to be more compact and either cylindrical or circular in shape, whereas in more open forest, canopies may consist of several small crowns that increase the total crown area. As mangroves respond to change, either through regeneration, degradation or changes in species, an alteration of the overall structure of mangroves may occur although the extent may remain the same. 1.6.3 Biomass and productivity Mangroves are perhaps one of the more productive of estuarine systems (Kuenzler, 1974; Golley et al., 1962), with productivity even exceeding that of tropical forests.
12
Factors that influence mangrove productivity include tidal regimes, as these dictate the delivery of macronutrients, the supply of oxygen to roots, the removal of toxins and the moderation of soil water salinity. The productivity of mangroves is typically reflected in the biomass (or carbon) content. Productivity and also the biomass content also varies depending upon the growth stage (i.e., regenerating, mature or senescent) of the trees. During successive growth stages, carbon may be preferentially allocated to certain plant parts (i.e., the stem or leaves). Young saplings grow rapidly in the initial stages, with the rate of increase in tree height generally exceeding that of trunk diameter (or girth). As the tree matures and finds its place in the canopy, vertical growth slows and the focus shifts to increasing girth to support the mature tree structure. Upon reaching the canopy, there is also the necessary space to allocate growth resources to lateral crown spread. All of these features are a function of age and are reflected in the changing biomass. Changes in biomass or productivity over time are indicative of the overall health and condition of stands. A change in the external factors (e.g., salinity, soil moisture, temperature) that promote plant growth may lead to a decline in total biomass. Such change may be episodic, in response to a rapid change in the conditions (e.g., a flood event), or revealed through longer-term changes (e.g., reduced growth in low rainfall years). 1.6.4 Species/community composition The ability of different mangrove species to establish, compete and survive under the conditions imposed by tidal regimes, climate and sediment movements is the key determinant to the overall composition of the communities.
Certain species are
indicative of a particular environment and their appearance or disappearance may signal a response to a changing coastal environment (e.g., favourable or less favourable saline or tidal conditions, increased availability of suitable substrata for colonisation). As an example, the raising of the ground level and the consequent decrease in the frequency of tidal flooding can lead to a succession of different mangrove species (Table 1.1). Community composition may consist of a single species in pure homogeneous forest, or
13
one dominant species with several understorey species, in mixed heterogeneous forest. Some species are typically found in association with another (e.g., mixed stands of R.stylosa and B.parviflora in tropical northern Australia). Table 1.1 Examples of mangrove species distributions under different tidal regimes in northern Australia. Tidal flooding conditions All high tides All medium high tides All normal high tides Spring tides only
Examples R. mucronata (exceptionally) Avicennia. alba, A. intermedia and Sonneratia. griffithii, R. mucronata Rhizophora species, Sonneratia griffithii. Bruguiera species.
1.6.5 Zonation patterns Within many mangroves, communities often develop parallel to the shoreline or river channel and distinct zonation patterns are typically displayed. Factors that influence the zonation patterns include the frequency of tidal immersion, the nature of the substratum, levels of salinity and the relative rates of erosion and accretion. The development of zones is often species dependent. Zonation patterns are generally less distinct in regions where either there is less variability in climate (e.g., in the tropics where rainfall is relatively even throughout the year) or where the species diversity is lower (e.g., in the more temperate regions of southern Australia where only one species occurs). The influence of climate on zonation patterns is highlighted as follows by Walters (1977). In more humid climates, the transition in the salinity from the outer edge of the mangal to the landward side (not flooded) is more gradual due to regular leaching of salt by rain between two successive flood tides. The most salt resistant mangroves are therefore found on the seaward edge and the least resistant on the landward edge. In more arid climates, evaporation from the soil surface occurs between the two successive inundations, leading to an increase in the salt concentration of water on the landward side, particularly on coastal mangroves. Hence, the most salt resistant mangroves are on the landward edge, which is bounded commonly by a sand surface without vegetation. These differences in salinity, which are partly climate dependent, lead to variations in the ability of species to colonise. Such variations are less evident in estuarine systems. 14
A change in zonation patterns might be associated with a change in extent but regeneration of different species within an area of mangrove may occur as a result of changing environmental conditions, even though a change in extent might not be evident. 1.6.6 Overview of monitoring change Anthropogenic disturbance of mangroves is typically associated with conversion to another land use such as mariculture (shrimp/fish farms) or agriculture (paddy fields). Alternatively, mangroves are logged through clear felling. In these cases, the loss of mangroves is complete and detection of change (e.g., using data acquired by satellite sensors) is relatively straightforward to detect. In other cases, mangroves are simply degraded (e.g., through mine or urban waste) or lost through selective removal of trees for fuel or timber. These latter cases are more difficult to identify. Furthermore, the response of mangroves to coastal environmental change caused by natural processes or occurring as a result of anthropogenically-induced climate change (e.g., sea level rise) is difficult to identify. For this reason, such detection requires monitoring of mangroves which have remained relatively undisturbed. In many countries, such areas are difficult to locate because of the impacts of a large population. However, within Australia, these processes can be studied because of the existence of large areas of mangrove along the coast, and their limited or lack of use or degradation by the population.
1.7
Distribution and characteristics of mangroves in Australia
The Australian coastline totals about 30,266 km in length, of which approximately 6,089 km (or 20 %) represents a mangrove shoreline (Zann, 1995). However, 70.5 % of the 11,617 km2 area of mangroves (Galloway et al., 1984) occurs on saline muds deposited in estuarine deltas, whilst the remaining 29.5 % (Bucher & Saenger, 1994) occur on open shorelines. The greatest development of mangroves is on the northern coasts of Australia (referred to as the Northern-Australian-Papuan Subgroup) where mangrove extent and species diversity are greatest. Within this northern region, two distinct biogeographic regions (Zone 1N and 1NE) were distinguished by Saenger et al. (1977; Figure 1.3). Zone 1N (in Australia’s
15
Northern Territory) contains 20 tree species, 6 understorey species and 6 salt-marsh species whilst Zone 1NE (in northeast Queensland) contains 27 tree species (belonging to 14 families of angiosperms), 10 understorey species and 6 salt-marsh species. The greater diversity of species in northern Australia (and particularly in north east Queensland) is attributable to the close connection with South East Asia during the various changes in palaeo-sea levels, and the similarity in the climate region when mangal first developed. Indeed, the Malaysian region appears to have been the centre of dispersal of modern-day mangrove floras in Australia and little change or movement in species has occurred. The coastline configuration, whereby numerous estuaries were sheltered by the Great Barrier Reef (in Queensland), also provided numerous areas for mangrove colonisation.
Progressively fewer species are found in the subtropical
regions and these mangroves also tend to be lesser in extent. Within the more temperate regions of Australia (and also New Zealand), which form the Australasian Group, only one species (Avicennia marina) typically occurs.
Figure 1.3 The location (above green line) of Zones 1N (largely northern Australia) and 1NE (largely north-eastern Australia).
16
Around the coastline of Australia, variations in the relative width, height and densities of the various zones, and species composition (and hence zonation patterns) occur. A typical zonation pattern of Australian mangal coasts is listed in Table 1.2 (Saenger et al., 1977).
Differences in zonation patterns within mangroves are
attributable partly to species composition and climate.
For example, the mangals
contained within Zone 1N and 1NE, which are relatively species rich, show distinct zonation patterns.
Mangroves in Zone 1N typically contain Sonneratia, although
sometimes Rhizophora species, on the seaward margins but Avicennia species on the landward margins.
The climate of these two zones is, according to Kőppen’s
classification, hot where the temperature does not fall below 18oC in any one month and in both Zone 1N and 1NE, the winter months are relatively dry (Awi; see Figure 1.3). Zone 1NE also experiences climates with very short dry season (Am) and with uniform rain (Af). Table 1.2 Typical zonation patterns associated with Australian mangal vegetation (Source: Saenger et al., 1977). 1 1a 2 2a 3 4 5
1.8
Landward fringe including salt marsh Landward Avicennia zone (Avicennia parkland) Ceriops thickets Vegetation free high tidal flats Bruguiera forests Rhizophora forests Seaward fringe
Research objectives
The primary objective of this thesis was to establish and demonstrate the use of remote sensing data from a variety of airborne sensors for establishing baseline datasets of mangrove extent, structure, biomass (productivity) and species/community composition against which to assess change.
The study focused primarily on the
mangroves of the Alligator Rivers Region in Kakadu National Park in Australia’s Northern Territory for which a range of remote sensing data were available and also acquired as part of this project. These mangroves were selected, as they have remained relatively undisturbed from anthropogenic disturbance. A secondary objective of the thesis was to identify changes in the mangroves, as observed by comparing baseline
17
datasets of selected attributes for different years, and interpret these in relation to previous studies of coastal environmental change in the region.
1.9
Format of thesis
Chapter 1 has introduced the research by providing an overview of the global distributions of mangroves, their importance as a sustainable resource and both the natural and anthropogenic impacts on this ecosystem.
Their potential role as an
indicator of sea level change associated with global warming has been highlighted. The requirement to establish baseline datasets relating to their extent, structure, biomass (productivity) and species/community composition has also been highlighted. The chapter following (Chapter 2) provides an overview of the response of mangroves to climate change. Past geomorphic studies are included and so provide a context for environmental change within the region. The response of mangroves (as observed through changes in extent and structure) to predicted climate change (including sea level rise, localised increases in rainfall and temperature) is discussed. Chapter 3 then reviews previous studies relating to the remote sensing of mangroves. A range of both airborne and satellite sensors and their application to mapping tropical forests including mangroves is discussed. The utility of certain sensors in mapping tropical mangroves is demonstrated through case studies from South America, South East Asia, Africa and Australia. Chapter 4 provides a background to the Alligator Rivers Region (ARR) of Kakadu National Park, with particular focus on the climate, soils, topography and also the vegetation and fauna. Chapter 5 outlines the remote sensing datasets (aerial photography, CASI and AIRSAR) that were available or acquired to support this research.
The initial
calibration and pre-processing steps (georectification) relevant to each sensor are also provided.
18
Chapter 6 gives an overview of the field data that were collected in order to facilitate subsequent interpretation of remote sensing data acquired over mangrove areas. The structural characteristics of the dominant zonations are described, and the total and component biomass within the sample sites is quantified. Analyses undertaken to provide three-dimensional canopy data (leaf and branch structure and densities) to support biomass estimation from aerial photography and interpretation of SAR data are also described.
Structural relationships between certain canopy components (e.g.,
height and diameter, and height and crown dimensions) are derived for each species. Chapter 7 outlines the use of stereo aerial photography for generating baseline datasets of mangrove canopy height and stem density. The process of generating a digital elevation model (DEM) that depicts tree top height is outlined. The use of historical aerial photography for DEM generation is also demonstrated.
DEMs
generated for the ARR mangroves are presented. Methods of retrieving stem densities from aerial photography and CASI data are then illustrated and results are validated using ground data acquired in 2002. Chapter 8 outlines the use of stereo aerial photography for generating baseline datasets of mangrove community extent. The procedure for generating an orthomosaic from a set of orthorectified photographic images and a DEM is outlined. The resulting orthomosaics for the coastal section of the ARR are presented. Chapter 9 describes the use of hyperspectral Compact Airborne Spectrographic Imager (CASI) data, acquired over the West Alligator River, for discriminating different mangrove species. The discrimination of mangrove and non-mangrove vegetation using the aerial photography and through derivation of vegetation indices using the CASI data is demonstrated. Through the analysis, the extent of communities in 1991 and 2002 are mapped.
The effective use of the near infrared wavelength region for species
discrimination is highlighted. Chapter 10 describes the integration of fine spatial resolution datasets for quantifying the total and component biomass of mangroves. The process of establishing a spatial
19
model that incorporates the DEM, species classification, and allometric equations is outlined. Results are validated using available ground truth. Chapter 11 describes an alternative approach to biomass estimation based on the use of polarimetric airborne Synthetic Aperture Radar (AIRSAR) data.
Empirical
relationships between mangrove forest structure and biomass and SAR backscatter are presented. The effective use of SAR in identifying scattering mechanisms in mangrove forest is also highlighted. A comparative analysis with mangrove of the French Guiana coastline is presented. Chapter 12 provides a comparison of baseline datasets of mangrove extent for the years 1950, 1991 and 2002 as generated using a combination of stereo aerial photography and CASI data. Changes in mangroves, as interpreted using the DEMs generated for all mangroves within Kakadu NP, are also interpreted with reference to previous studies undertaken within the region. Chapter 13 discusses the use of airborne and also spaceborne remote sensing data for generating baseline datasets of mangroves against which to map, quantify and monitor change. The advantages and inherent limitations to using certain sensor types are discussed. The optimal approach to discriminating the biophysical characteristics of mangroves and a framework for monitoring is presented. Chapter 14 then concludes the study and makes recommendations for further research.
20
CHAPTER TWO
THE IMPACT OF ENVIRONMENTAL AND CLIMATE
CHANGE ON MANGROVES – A REVIEW.
In tropical and subtropical regions, mangroves are particularly susceptible to coastal environmental changes, including those induced by global warming.
Specifically,
mangroves are responding to changes in rainfall and thermal regimes and rises in sea level through alterations in, for example, extent and species composition, whilst structural alteration is increasing through increased cyclone frequency and intensity. In the recent past, mangroves have responded to change, surviving and rapidly colonizing areas where favourable conditions prevail but disappearing or degrading where conditions become adverse.
However, as the rates of change (particularly those
associated with global warming) increase, mangroves might be expected to respond more rapidly.
This sensitivity to change therefore renders them useful biological
indicators of both short to long-term environmental change, particularly if left undisturbed by human activity.
This chapter therefore explores the response of
mangroves to past, present and future environmental and climate change, and attempts to identify their role as indicators of these processes.
2.1
Introduction
Where mangroves occur, their distribution is generally a function of sensitivity to the prevailing environmental conditions.
As stated previously, the key elements
influencing this distribution include the frequency, range and magnitude of tidal inundation, salinity, oceanic currents, patterns of drainage, underlying soil type (substratum), and both air and water temperature (Chapman, 1977; Semeniuk et al., 1978; Alongi et al., 2002). These factors often interplay to generate gradients across the
21
landward to seaward edge of the community, and are associated with the formation of distinct mangrove zones.
As different species demonstrate a tolerance to external
conditions within a narrow range, any change in these conditions will lead to an associated alteration in mangrove extent, floristic composition and structure, and hence zonation. As such, mangrove ecosystems (particularly when left undisturbed) may be useful indicators of coastal environmental change (Blasco et al., 1996). Even minor changes in coastal conditions, including those that result from altered drainage patterns, saltwater intrusion, and accretion or erosion in response to variations in sea level, are often manifested through a change in the composition, structure and functioning of mangroves (Blasco et al., 1996). Nevertheless, mangrove species display a variety of adaptations and may even exhibit high resilience to naturally induced change, particularly in monsoonal regions where intra and interannual climatic variations may be extreme (Eliot et al., 1999). In many tropical and subtropical regions also, distinct seasons are commonplace such that mangroves are subjected to high mean annual precipitation, temperatures and humidity combined with a greater frequency of storms. Accordingly, communities and their constituent parts become adaptable and resilient.
Evidence has also shown that
mangroves can extend or contract rapidly in response to topographic and climatic changes (Field, 1995), including sea level rise. These forces may act at different rates and intensities on the mangrove environment, and are reflected in both structural and functional changes within communities. The sensitivity of mangroves to coastal environmental change is indicated through long-term (over several centuries or millennia) historical records (Wolanski & Chappell, 1996; Woodroffe & Mulrennan, 1993) and through shorter-term observations (e.g., Lucas et al., 2002; Ramirez-Garcia et al., 1998; Saintilan & Wilton, 2001). Regional geomorphological and stratigraphic investigations have provided evidence for long-term variations in climate and sea level in mangrove-dominated environments around the world. Such records also provide a context for the environmental changes occurring within regions, for the distinction between high and low frequency changes, and identification of those factors that may recur in the near future (Eliot et al., 1999). In the shorter term, observations of coastal systems over time or reference to tidal
22
hydrological records enables an immediate picture of recent environmental change and its impacts to be established in a given area. These records can potentially provide us with important insights into regional climate variation, rates of change, and the likely response of mangroves and ecosystem functioning to such change. Of particular concern to coastal ecosystems is the worldwide rise in sea levels which can be attributed largely to global warming that results from increased levels of atmospheric carbon dioxide (CO2) and other greenhouse gases and leads to the ‘greenhouse effect’ (Pittock, 1988). The associated warming causes both an expansion of the upper layers of the ocean and partial melting of the world’s ice caps and glaciers and hence a rise in sea level. Climatic change is likely to be an ongoing process throughout this century (Eliot et al., 1999), and in association with concurrent sea level rise, may cause irreversible change at the ecosystem level. By 2025, global sea level is predicted to rise by 3 – 12 cm (IPCC, 1997) and even since the 1940s, an average rise in sea level of 10 cm has been recorded (Henderson et al., 1989). The rate of sea level rise is, however, highly variable worldwide. For example, over the past half century, parts of tropical South East Asia have experienced sea levels ranging from a fall of 1.33 mm yr-1 to a rise of 2.27 mm yr-1 (Watson et al., 1998). On the American Gulf Coast, sea level has risen 2.5 – 3 mm yr-1 over the past century, whilst the opposite effect has been recorded on the Canadian/Alaskan coasts where sea level has fallen or remained stable (Watson et al., 1998). In many cases, these rises cannot be associated entirely with global warming as differences in centuryscale sea level trends can be attributed also to vertical land movements and regional climate variability.
Superimposed on this, are the effects of changes in ocean
characteristics (circulation, wind/pressure patterns, ocean water density), which themselves may be induced or influence by global warming. With such subregional differences occurring along coastlines, regional adjustments to sea level rise are anticipated. Mangroves may respond favourably or otherwise to sea level rise and their response will depend upon future tidal ranges, sediment supplies and the composition of tree species (Alongi, 2002). For loss of mangroves to occur, the rate of sea level rise is
23
perhaps more important than the absolute rise. Some communities will be able to respond by vertical accretion with the input of sediment and organic matter (Nicholls, 1999).
Other communities may be able to migrate to proximal low lying terrain
adjacent to the coast. As a result, some areas may experience the inland spread of mangroves and their encroachment into other habitats (Alongi, 2002), while others may suffer total destruction or partial modification of communities over time. While a rise in mean sea level may be the most important factor influencing the future distribution of mangroves (Field, 1995), the greater intensity of rainfall events and cyclones, storm surges and increased temperatures, will also exert an influence although to varying degrees (Alongi, 2002). On a regional scale, mangroves along tropical shorelines may be subject to temporary flooding (Pittock, 1988) and the effects of shoreline change, including increased erosion, changing sediment supply and tidal ranges, and saltwater intrusion (Henderson-Sellers & Blong, 1989; Field, 1995; Alongi, 2002). Local to regional climate changes have the potential to enhance the effect of a sea level rise on those naturally occurring processes within coastal environments. The impact of sea level rise on the coastal environment is evidenced by changes in coastal geomorphology, hydrological flows and also the distribution, structure and functioning of vegetation. However, the changes attributed to sea level rise need to be clearly separated from those resulting from episodic events, such as the El Niño Southern Oscillation (ENSO) phenomenon (Emery et al., 1991) in the Pacific Region. For this reason, the assessment of climate change in tropical regions requires a thorough understanding of the natural variability that prevails (Mitchell, 1992). Only when the functional attributes of a coastal system are understood, can we predict how the coastal environment is responding to the conditions resulting from global warming and climate change.
2.2
Current evidence of mangrove response to climate change
The response of mangroves to climate change is difficult to ascertain for two main reasons.
First, these need to be separated from natural changes, including those
associated with regional climate impacts (e.g., El Niño). Second, many mangroves have
24
been disturbed and it is difficult therefore to establish whether changes are induced by these processes alone. The structural and functional response of mangroves to predicted climate change is largely unknown (Ellison & Farnsworth, 1997). As a basis for predicting the most likely response, current research on mangroves attempts to relate past and present patterns of climate change to aspects of community structure.
Normal ecosystem
functioning (as a result of tidal inundation, sedimentation and succession) will continue regardless. However, the rate at which these processes occur will vary regionally and under different scenarios of future climate change. It is a reasonable assumption that similar changes in mangrove forest structure, as illustrated in historical records, will occur over the next century. However, the long-term impact may be more devastating due to the higher frequency climatic changes in some regions. For example, natural disturbance in the form of tropical cyclones, lightning strikes and sea level changes affects mangroves on variable spatial and temporal scales (Allen et al., 2001). However, with the growing body of knowledge on ‘mangrove disturbance ecology’, and renewed commitment to further research and documentation, a better understanding of the relationship between climatic change and mangrove community structure should be obtained. The response of mangroves to climate change has also been difficult to ascertain as, in many tropical and subtropical regions, mangroves have been severely degraded or destroyed through anthropogenic activity.
However, in some areas, anthropogenic
disturbance is absent or minimal and changes can be better attributed, or otherwise, to climate-induced phenomena. Even in these areas, the natural processes of change, and their rates, need to be established such that anomalies can be identified (Bird, 1993). In particular, El Niño events that typically occur 1 – 2 times per decade (Whetton et al., 1990) need to be isolated over short time scales. During such events, unusually warm water off the Peruvian coast extends into the equatorial central Pacific. Subsequently, affected regions experience low atmospheric pressure, increased sea surface temperature (SST) and regional variation in rainfall and storm patterns (Emery et al., 1991).
25
The strength and direction of winds that drive oceanographic movements and control east-west sea level differences also vary with ENSO events (Godfrey & Love, 1992). Hence, in certain locations, short-term sea level rise and higher rainfall may augment changes in the tidal regime and initiate or accelerate shoreline erosion. Currently our knowledge of ENSO behaviour and associated impacts on coastal ecosystems is limited. As well, its probable response to changing global climate, particularly temperature, is largely unknown. Where mangroves inhabit estuarine coasts, the impact of ENSO on climate may well contribute to the observed distribution and range of species, particularly in the upstream environment. Where droughts occur as a result of El Niño events, structural changes to the mangroves may ensue. As an example, on the Pacific island of Kosrae, diameter growth of certain mangrove species declined dramatically during the 1997/1998 El Niño, where rainfall was well below average for a six month period (Allen et al., 2001). El Niño chronologies have been compiled for various regions (Whetton et al., 1990) in attempts to understand the magnitude and frequency of such events and further investigation regarding the impacts of this phenomenon are necessary. To support these investigations, regional monitoring of mangroves or any coastal ecosystem for sea level and climate change effects should encompass a timescale of at least twenty years, such that the effects of phenomena such as ENSO can be isolated. 2.2.1 Historical evidence An investigation into the geomorphic history of mangrove environments is an important first step in any monitoring program, as a record of system dynamics under conditions of sea level rise (i.e., post-glacial era) is provided and important insights into how the system might respond to future environmental change is revealed (Woodroffe, 2000). Even so, there is no guarantee that past changes will represent either the current or future changes. Within Australia, evidence from the geomorphological evolution of the South Alligator River of Kakadu National Park, indicated that sedimentation on the mangrove tidal plain kept pace with the rising sea level from 7000 to 6000 years B.P, enabling widespread expansion and growth of mangrove communities (Wolanski & Chappell, 1996). Shortly after this period, sea level rose to 1-2 m above present levels for the
26
north Australian coast, before gradually retreating to its present position. During the phase of sea level rise, the rate of sea level increase was similar to the future rise predicted under the greenhouse effect (Chappell, 1990). As mangroves were able to persist and actually extend their distribution in the face of historical sea level rise, their future survival under enhanced greenhouse conditions seems assured in this area. A similar record of estuarine infill and coastal progradation exists for South East Asia, but there are significant variations with location. Extensive mangrove forests persisted throughout many of the deltaic areas until 4000 years ago, shortly after which began a phase of peat swamp development (Woodroffe, 2000). Both northern Australia and South East Asia have experienced the gradual succession of mangrove forest to freshwater wetland vegetation, as sea level stabilized several thousand years ago. Subsequent progradation during the regressive phase (sea level stable or gradually falling) enabled the spread of mangroves along those shorelines where sediment input and the tidal/riverine influence was favourable. Shoreline communities have the potential to respond quickly to change and have evolved under such conditions as prograding or regressing shorelines, recurrent wildfire and unusually extreme temperate or soil moisture fluctuations (Gillison, 2001). The impact of past climatic change has not always had negative repercussions for mangrove ecosystems, and some communities have undergone readjustment in order to survive. In particular, sedimentation and erosion rates and the ability to migrate inland (Granich, 1993) are three important factors in determining the response of mangrove ecosystems to changes in sea level. The physical effects of climate change acting over a range of time scales and intensities have had variable consequences for mangrove ecosystems worldwide. Hsueh & Lee (2000) found a direct correlation between temperature and the diversity and distribution of mangroves in North East Asia. Air and sea surface temperatures (for 1961-1990 and 1994 respectively) for the west coast of Taiwan and its neighbouring countries (Japan, China), indicated an increasing winter coldness from south to north across all regions. This trend corresponded to shifts in the species composition of mangroves and a decline in species diversity with lower minimum winter air
27
temperatures and higher seasonal temperature variations. The interaction of warm and cold oceanic currents determines the sea surface temperature, which in turn affects the air temperature (or climate), which was found to be the primary limiting environmental factor for the extent and species composition of mangroves in North East Asia. Aside from the obvious penchant for a favourable climate, a suitable physiographic setting is required for the growth and establishment of a mangrove community. On a regional scale, high rainfall areas with river-fed estuaries that provide freshwater and terrigenous sediments to the intertidal zone tend to support more diverse communities than areas of low rainfall and limited runoff (Woodroffe & Grindrod, 1991). Regional climate variability, including extreme weather events and its direct influence on active processes (tidal levels, sedimentation, runoff) across the intertidal area, lead to the formation of distinctive zones of mangroves. Hence the climate-landscape interaction is an important determinant of the viability and distributional range of mangrove species and communities. In recent times, mangroves in different parts of the world have been subject to the regional effects of climate change. However, given the diversity of geomorphic settings and differing levels of human intervention, the impact of climate on mangrove communities cannot always be considered in isolation. Where changes in the extent or viability of mangroves have occurred, a number of causes may be identified, including altered tidal regime, sedimentation, subsidence and catchment modifications (Saintilan & Williams, 1999). The significance of these factors and their association with climate should be considered on an estuary-by-estuary basis. For example, the past decade has seen a declining trend in average rainfall over the coast of Senegal, resulting in increased salinity and regional decline of mangroves that were once invaluable to local fisheries (Dennis & Nicholls, 1995). In contrast, the trend of increasing rainfall in South East Australia over the last five decades has led to the landward incursion of mangroves into a number of estuarine saltmarsh environments (Saintilan & Williams, 1999).
Photographic observations from Jervis Bay, Eastern Australia (Saintilan &
Wilton, 2001), show that an increase in mean precipitation has increased freshwater flow and nutrients to the intertidal area, contributing to the landward encroachment of mangroves.
28
Historical and more recent evidence of mangrove community change with respect to global climate change, emphasizes our need for ongoing commitment to documenting and better understanding the changes already in effect. There is still a wealth of knowledge to gain, but the present lack of continuity in data collection means that significant changes may be occurring largely unnoticed. Spatial baselines that provide reliable and reproducible maps of the extent, structure and floristic composition of mangroves are therefore required for reference years to effectively monitor and map the long-term impacts of coastal environmental change on these ecosystems.
By
identifying deviations from these baselines, the extent of impacts can be deduced, monitored and better understood (Lucas et al., 2000). Knowledge of the processes observed can be used in combination with the region’s geomorphological history, thereby providing a context for environmental change within the area. With impending sea level rise and climate change, predictions can be compared against the natural variability in the system, and the probable response of mangrove communities inferred. This information is of relevance to ecologists and landowners who seek to design and implement management programs for those areas regarded as most vulnerable to future climate change.
2.3
Climate change: observation and prediction
Climate change analysis requires the distinction between natural climate variability and changes related to anthropogenic activity. The former relates to the natural trends, oscillations and more random perturbations in climate (Eliot et al., 1999) whilst the latter, to a number of fluctuations in climatic factors beyond the range considered normal. Both natural and anthropogenic changes may occur over a range of time scales. However, anthropogenic change tends to be more rapid and may be irreversible. One of the most profound changes has been the increase in atmospheric carbon dioxide (CO2) in the atmosphere.
Caused largely by fossil fuel burning and
deforestation, CO2 levels have doubled over the past few centuries. In recent years, evidence that the enhanced levels of CO2 are leading to global warming, climate change and sea level rise is mounting (Field, 1995). Establishing whether observed changes are within the realms of natural variability or are heightened, as an indirect consequence of
29
human activity is, however, difficult. Disentangling the relative contribution of natural and anthropogenic contributors to change is particularly difficult in environments where the natural variability in climate is already high (e.g., Australia) or where episodic events (e.g., ENSO) contribute to extremes in climate (Mitchell, 1992). Globally, mangroves are already impacted by direct anthropogenic change (i.e., land clearance, land reclamation and pollution). While the time scale within which human land use changes directly impact communities is 1 – 5 years, the effects of land use changes on climate may operate on longer time frame (e.g., 100 – 500 years, Eddy & Faud, 1996). For example, the higher rates predicted for future sea level rise may present difficulties considering past rates of adaptation by mangrove communities. In the tropics, the area directly landward of the mangroves is one of the most intensively used resources, so survival of communities depends on their ability to move inland, and if this land is available for them to move into. Landward migration may, however, be impeded by human infrastructure and urban/agricultural development, leading to a reduction in the total area of mangrove (IPCC, 1997). This section therefore outlines those changes in mangrove area both directly and indirectly affected by anthropogenic change. 2.3.1 Direct anthropogenic change In many regions of the world, mangroves are being cleared at unprecedented rates. Many areas are being reclaimed and the land used instead for urban development, agriculture or aquaculture. Pollution is also of consequence to the long-term viability and condition of mangroves.
Through human modification of mangrove area,
significant changes have occurred at the ecosystem level. For example, conversion of large areas of mangrove in Vietnam for aquaculture and agriculture has restricted the surface of tidewater and is contributing to current sea level rise in the region (Granich, 1993).
Various hydrological changes, augmented by the construction of dams,
embankments and hydrological stations, have also had variable consequences. Consideration of the changes to landscape patterns and processes at the ecosystem level prior to habitat modification is not always apparent.
30
Coastal infrastructure and dredging practices have had variable effects on the distribution of mangroves and estuarine hydrodynamics along the South East Australian coast (Saintilan & Williams, 1999). Changes to tidal range and frequency have led to the upstream colonization of mangroves and displacement of upland plant communities. While the impact of vegetation changes may not be considered negative, but rather successional in this case, it is the loss of biodiversity, particularly in freshwater wetlands that may be significant. Saintilan & Williams (1999) also describe the dieback of mangroves through regional subsidence and exposure to acidic soils as a result of agricultural drainage practices along the East Australian coast.
There are many
instances of the destruction and degradation of mangroves worldwide. Following the removal of shoreline mangroves, increased coastal protection will most likely be required, in view of the predicted rising sea levels and other greenhouse effects. 2.3.2 Indirect anthropogenic change Although the impact of climate change on mangroves is not well understood, a number of reasonable predictions can be made. In particular, greenhouse gas (CO2, methane and nitrous oxide) concentrations in the atmosphere, which have already doubled since the 1800s, are predicted to increase exponentially from the present rate. Such increases are expected to result in an average global warming of ~2.5 oC, ranging between 1.5 and 4.5 oC (Figure 2.1; Pittock, 1988; Wigley & Raper; 1992, Mitchell, 1992; Field, 1995) and associated changes in humidity (Bennett et al., 2000) over the period 1990 to 2100. Surface temperature will be raised and it is expected that climatic zones will migrate, with tropical regions extending poleward (Bird, 1993). However, compared with temperate zones, warming in the tropics will be both smaller and vary less with season (Field, 1995). In Australia, for example, the northern coastal areas is expected to be, on average 1 – 2 °C, warmer, while further inland, temperature will increase by 2 – 4 °C (Eliot et al., 1999) by the year 2100. With these new regimes, an expansion of the area of mangroves would be expected.
31
Figure 2.1 Simulated global surface temperature change (Source: IPCC, 1997) based on the outcome of several climate models and using a number of different illustrative scenarios: the three A1 groups are linked to rapid economic growth and three main directions for resource use including fossil intensive (A1FI), non-fossil energy (A1T), and a balance between the two (A1B). The A2 group is linked to regional economic growth and self-reliance, with less rapid technological change. The B1 group refers to a shift towards the service and information economy and the use of more sustainable and resource efficient technologies. The B2 group emphasizes a more local to regional economy, with less rapid but more diverse technological change.
The frequency and intensity of tropical cyclones is also expected to increase. Cyclones are generated when Sea Surface Temperature (SST) exceeds 26 °C (Mitchell, 1992). With the anticipated increase in SST, the southern limit of tropical cyclone generation could extend its range southwards by 200 – 400 km (Pittock, 1988). In systems where intense disturbances are common, or become more common under global change, there will be greater likelihood of mortality and replacement of existing trees, and changes in forest structure and composition may be more rapid (Gillison, 2001).
32
Rainfall intensity and frequency is also likely to increase systematically in a warmer world (Mitchell, 1992), with marked regional variations expected with season and location. With such changes, the water balance of regions will inevitably change (Pittock, 1988). Under enhanced greenhouse conditions, precipitation is predicted to increase on average in the high latitudes, in the monsoon region of Asia, and in the midlatitudes during winter (Mitchell, 1992).
Tropical zones may experience both an
increase and decrease in mean annual precipitation depending on locality. The changes in precipitation are anticipated to lead to fluctuations in the salinity of coastal environments that may affect plant growth. Over the past century sea level has risen and is expected to rise at an accelerated rate during the 21st Century (Bird, 1993; Field, 1993), attributable largely to thermal expansion of the oceans, melting of ice caps and, to a lesser extent, mountain glaciers and snowfields (Henderson-Sellers & Blong, 1989). Forecasting the rates and future level of sea level rise is fraught with uncertainty, particularly when regional levels are sought. Depending on regional changes to wind patterns, ocean circulation and other localized phenomena that affect sea level, some parts of the world may experience lower sea level due to global warming. Climate itself is hard to predict, hence the secondary impacts including sea level rise, are even harder to predict with accuracy (Emery et al., 1991). Estimates range between a 7 cm sea level drop to a rise of over 3.6 m (Table 2.1). Future rates of sea level rise due to global warming are also highly variable (Table 2.2). The varying estimations are difficult to assess due to the numerous methodologies and assumptions from which they were derived. There is, however, some consensus amongst authors as to the past and present rates of sea level change.
33
Table 2.1 Estimates of current and future global sea level rise (cm). Authors 2000 2025 – 2030 2050 2085 – 2100 Hoffman, 1984 4.8 - 17.1 13 - 55 23 - 117 56 - 345 Nat. Acad. Sciences, 1985 10 - 32 Polar Res. Board, 1985 10 - 160 US Dept. Energy, 1985 10 - 90 Hoffman et al., 1986 3.5 - 5.5 10 - 21 20 - 55 44 - 368 Thomas, 1986 64 - 230 UNEP/WMO/ICSU, 1986 20 - 140 Clim. Res. Unit, 1988 6 - 34 Jeager, 1988 -5 - 19 -4 - 52 -7 - 138 Villach Conf., 1988 20 - 140 Henderson et al., 1989 20 - 140 IPCC, 1990 20 65 Wigley & Raper, 1992 48 Warrick et al., 1993 13 61 Field, 1995 18 44 Table 2.2 Estimated rates of past, present and future global sea level rise. Authors PAST 100 YEARS Henderson et al., 1989 Henderson et al., 1989 Godfrey & Love, 1992 PRESENT Gornitz & Lebedeff, 1987 Pirazzoli, 1990 Bird, 1993 FUTURE Nat. Acad. Sciences, 1985 Wigley & Raper, 1987 Oerlemans, 1989 Henderson et al., 1989 IPCC, 1991
Year
Rate
1900 - 2000 1950 - 2000 1900 - 2000
1 - 1.5 mm/yr > 2 mm/yr 2 mm/yr 1.2 +/- 0.3 mm/yr 1 - 1.5 mm/yr 1.2 mm/yr
1985 - 2030 1985 - 2025 2025 2030 2000 - 2100
2.2 - 7.1 mm/yr 1 - 2 mm/yr 2 mm/yr 5 - 35 mm/yr 3 - 10 mm/yr
Global climatic change over the following century will exert a variable impact on existing natural processes and system functioning.
It is likely that changes in
greenhouse gas concentrations, temperature, precipitation and sea level will alter the scale and rate of naturally occurring change in some ecosystems.
The effects of
34
greenhouse-induced climatic change on mangrove ecosystems will be discussed in the following section.
2.4
Expected response of mangroves to future climate change
With a predicted sea level rise of between 10 cm and 370 cm over the next 100 years (Pirazzoli, 1990), the coastal environment and proximal inland areas are likely to experience change in the form of temporary flooding, altered estuarine hydrodynamics, increased erosion, sedimentation and deposition, and saltwater intrusion. Such changes will similarly impact on the mangroves and will be manifested largely through changes in extent, species composition and zonation, structure and productivity. The following sections outline these impacts. 2.4.1 Physical effect With sea level rise, localized effects in low-lying coastal areas may involve extensive marine submergence. High and low tide lines will advance landward and at least part of the present intertidal zone may become permanently submerged (Bird, 1993). Flood peaks may increase by an amount corresponding to that of sea level rise, largely affecting the lower reaches of tidal rivers (Henderson-Sellers & Blong, 1989). Furthermore, a rise in sea level may increase the rate of shoreline erosion in vulnerable areas and also rates of deposition in others (Henderson-Sellers & Blong, 1989), with both processes resulting in changing patterns of shoreline retreat and advance. The extent of shoreline change is difficult to predict as local topographic variations and feedback mechanisms that restore the dynamic equilibrium are often poorly understood or quantified. Some ecosystems will have greater resilience to the physical forces that attempt to shape them, while others may have lower resilience, and irreversible changes may result. Coastal areas with beach fringes, deltas, estuaries, intertidal and nearshore areas, salt marshes and mangrove swamps (Bird, 1993) will all be affected to some extent. Estuarine hydrodynamics and processes of sedimentation will most likely be modified should sea level rise occur in the future. Tidal sections of rivers will be subject to change, as channels adapt to increased flows and changing tidal regimes
35
(Chappell, 1990).
In particular, widening and deepening of estuaries, increased
upstream penetration during high tides, increases in tidal ranges and changes in patterns of shoal deposition are expected to occur, although the rates and scales of change will vary depending on their existing configuration and flow dynamics (Bird, 1993). However, if sediment supply from fluvial or marine sources and the tidal range is sufficient, the deepening and enlargement of estuaries may be counteracted despite sea level rise (Jelgersma et al., 1993). As stated before, it is essential to understand the natural variability of a system and its responsiveness to past climate change, as this often reveals important insights into the probable impact of future sea level rise. These broad effects may be evident through changes in mangrove distribution. With the loss of physical stability and nutrient abatement provided by the shoreline buffer zone of mangroves (Eddy & Faud, 1996), some coastlines may become more vulnerable to erosion.
Any changes that affect shoreline processes and hydrology are also
dependent on whether the area experiences a relative sea level rise and whether climatic conditions become adverse.
In particular, with the predicted increase in storm
frequency and severity for tropical regions, there will most likely be a decline in the extent of mangroves as the coastal zone becomes ‘squeezed’ between the sea and inland agriculture/development. Where freshwater wetland or other vegetation dominates the upstream environment there will be increased competition for space, with one community likely to overtake the habitat of the other. It is thereby suggested that the ‘artificial’ succession of plant communities will occur as a result of human influence on global climate and sea level. Extensive saltwater intrusion into coastal lowlands is also expected as a result of sea level rise (Knighton et al., 1991). The process will largely be augmented through rapid extension of tidal creek systems and upstream movement of tidal flows and sediment. Increased penetration of saltwater into estuaries may cause the upstream migration of ‘salt wedges’, unless compensated by increased freshwater discharge, leading to salt contamination of freshwater resources (Emery et al., 1991). In tropical regions where two seasons, the wet and the dry prevail, such changes to fundamental estuarine dynamics may increase the duration and timing of one season over another. In this way,
36
changes in species composition and distribution will occur. These changes are outlined in the next section. 2.4.2 Biological effects Mangroves flourish in warm, wet humid conditions where there is continual input of freshwater into their normal saline environment (Field, 1995).
Communities are
established with the trapping of sediment to build a depositional terrace in the upper intertidal zone (Bird, 1993), thereby enabling successive zones of species to migrate seaward. At the high tide level, other vegetation communities displace the inner edge of mangroves. This gradual process is dependent on a continual sediment supply (Emery et al., 1991). However, if the rate of sea level rise exceeds the rate of sedimentation, then some restructuring of existing vegetation will take place, and loss of mangroves may occur. Most established mangrove ecosystems can, at most tolerate, an average rise in sea level of about 1.3 cm per decade before they degrade. The seaward margin, subject to the increased erosive power of the rising sea and tides, will recede, causing mangroves to spread landward into available low-lying alluvial areas, or displace existing freshwater communities (Bird, 1993). Traditional methods of coastal protection (e.g., sea walls) are ineffective against these rising seas, as they cut off the circulation of nutrients and seawater essential to survival. Hence, severe disruption to shoreline communities is forecast, with only limited options for protection. The most destructive impacts on mangrove ecosystems will likely stem from sea level rise but also from cyclonic activity and episodic storm events. The effects of sea level rise were discussed in the previous section, with hydrodynamics and sedimentation being the main factors controlling the survival and distribution of mangroves. The severity and frequency of storms, which lead to the destruction of mangroves or disruption of zones, may also be as important in certain areas as the rise in mean sea level itself (Mitchell, 1992). The three main hazards associated with tropical cyclones are high wind speeds, intense rainfall and storm surges (HendersonSellers & Blong, 1989), all of which could cause unpredictable partial or mass
37
destruction of mangrove shorelines. With rapid sea level rise over the next century, these changes could be the cause of irreversible damage. Following the predicted changes in sea level, changes in the species composition and community structure of mangroves can be expected (Ellison & Farnsworth, 1997; Pittock, 1988). The distinction between marine and freshwater wetlands is likely to become blurred, as the range distributions of species and functional groups readjust to new environments (Gillison, 2001). The nature and rates of these changes remains uncertain at this stage. Changes in the species composition will also occur as increased saltwater intrusion into the upstream sections of tidal rivers leads to changes in soil salinity and water availability (Field, 1995). These factors are critical for internal cellular functioning and hence the survival and growth rates of mangroves. As soil salinity increases, mangroves must face greater salt levels in their plant tissue and less available water, which may lessen the net assimilation rate per unit leaf area, thereby reducing growth (Field, 1995). Alternatively, if increased rainfall occurs as a result of global warming, and in so doing, reduces soil salinity on a seasonal basis, improved growth rates may be observed for certain species. Simulation experiments based on the mangroves of South Florida as outlined in Eddy & Faud (1996), provided some interesting insights into the sensitivity of mangroves under different tidal regimes to climate change. Slight changes in hydrology separated the mangrove communities and strongly induced fragmentation. Such a response was generally specific to a particular community and based on the variety of adaptations and resilience to environmental variables. Simulated results suggested that mangroves may be strongly influenced by a keystone ecosystem process, the hydroperiod (duration and timing of inundation), as opposed to a keystone species that is more prevalent in terrestrial communities. Increasing the hydroperiod and depth of inundation physically limited the establishment of mangroves and changed the surrounding biogeochemical characteristics. The study concluded that there may be regional changes in habitat quality and temporal availability which would largely affect estuarine-dependent species and migratory birds.
38
Ecosystem functions (primary production, biodiversity, decomposition, export of organic matter and biogeochemical cycling) are therefore most sensitive to potential changes in climate. However, for extensive loss of wetland functions/types to occur, extreme changes in the climate would have to occur (Eddy & Faud, 1996). If through climate change, drier conditions ensued, there would most likely be a shift towards semi-permanent/temporary wetlands, with changing proportions of forest types in a particular area. What may be more significant to the distribution of mangroves in certain coastal settings, however, are the changes augmented through climate, including changes in the physical environment (e.g., salinity and turbidity) and ecology (gap dynamics associated with lightning strikes and insect herbivory) that affect the balance of plant-animal associations. One example (Micheli, 1993) describes the influence of sesarmid crabs on the zonation and productivity of Australian mangroves. Here, mangrove productivity is strongly related to nutrient remineralisation which is influenced by the crabs, so any changes in the hydroperiod that modify crab life cycles therefore indirectly affect mangrove structure/production. Changes in climate parameters that ultimately affect trophic interactions thereby have strong implications for the demographic/functional properties of mangrove ecosystems. Regarding the growth and productivity of mangroves, there is certain debate as to the response of species to elevated temperatures and increasing levels of atmospheric CO2. There is no conclusive evidence that the growth and structure of mangroves will be affected by these factors (Field, 1995; Farnsworth et al., 1996), and further research is required to assess the long-term effects under enhanced greenhouse conditions. Not all mangrove species will respond similarly to elevated temperatures and CO2. Vegetation growth could be impeded by higher temperatures and greater ultra-violet radiation (Bird, 1993). On the other hand, Field (1995) suggests that increasing air temperature may cause the distribution of mangroves to extend their range further north and south. Regardless, these factors will only see a slight increase over the following century and their effects on coastal mangroves will most likely be negligible.
The impact of
elevated CO2 on mangrove growth and production is largely unknown, although some studies (e.g., Ellison & Farnsworth, 1997) have reported enhanced photosynthetic rates
39
and tissue production in some species but not in others. Hence, higher CO2 levels could favour some species over others, and combined with climate change, may lead to changes in species composition in natural environments (Pittock, 1988).
2.5
Concluding summary
Historical and more recent evidence of mangrove community change with respect to global climate change, emphasizes our need for ongoing commitment to documenting and better understanding the changes already in effect. There is still a wealth of knowledge to gain, but the present lack of continuity in data collection means that significant changes may be occurring most of which are largely unnoticed. Spatial baselines that provide reliable and reproducible maps of the extent, structure and floristic composition of mangroves are required for reference years to effectively monitor and map the long-term impacts of coastal environmental change on these ecosystems. By identifying deviations from these baselines, the extent of impacts can be deduced, monitored and better understood (Lucas et al., 2000). Knowledge of the processes observed can then be used in combination with the region’s geomorphological history, thereby providing a context for environmental change within the area. With impending sea level rise and climate change, predictions can be compared against the natural variability in the system, and the probable response of mangrove communities inferred. This information is of particular relevance to ecologists and landowners who seek to design and implement management programs for those areas regarded as most vulnerable to future climate change. Climate change, particularly sea level rise, poses a real threat to the survival of mangrove ecosystems. The impact of changes in climate and weather pattern will be revealed through alterations in the extent, species composition and structure of vulnerable and fragile vegetation communities such as mangroves. Many of these changes will take place over periods ranging from decades to centuries. Establishing baseline information and implementing an effective monitoring program is therefore essential if present and future generations are to observe and understand the consequences of climate change for any given natural system. To effectively monitor
40
the impact of both natural variability and human-induced change on mangroves, will most likely require the use of remotely sensed data.
41
CHAPTER THREE REMOTE SENSING OF MANGROVES – A REVIEW.
Monitoring the impact of climate and other environmental changes on mangroves can be achieved by establishing baseline datasets relating to the extent, structure, species/community composition and biomass of mangroves using remote sensing data acquired by airborne and/or spaceborne remote sensing instruments.
This chapter
reviews the requirement for these baselines and identifies the remote sensing data that are considered most suited for their generation. Using examples from selected regions, the use of remote sensing data for generating baseline datasets is outlined.
3.1 Spatial, temporal and spectral resolution Knowledge of the spatial, temporal and spectral resolutions of remote sensing data is fundamental for understanding the use of these data for establishing baselines and detecting change. 3.1.1 Spatial resolution The spatial resolution of an observing sensor refers to a distance between the nearest objects that can be resolved, is given in units of length (e.g., m) and depends upon the instantaneous field of view (IFOV) of the observing sensor (Lucas et al., 2003). In this study, coarse (> 100 m), moderate (10-100 m) and fine resolution (< 10 m) are considered. For regional baseline mapping of mangroves, coarse spatial resolution data, as provided by sensors such as the NOAA Advanced Very High Resolution Radiometer (AVHRR), the European Resources Satellite (ERS-1/2), Along Track Scanning Radiometer (ATSR), the Système Pour l’Observation de la Terre (SPOT) 42
VEGETATION (VGT) and the TERRA-1 MODIS (Moderate Resolution Imaging Spectrometer), are only appropriate for mapping contiguous areas of mangroves that occupy extensive tracts of coastline (e.g., the mouth of Amazon). However, as many mangroves cover small areas area and often fringe coastlines or estuarine areas, the use of coarser resolution data for mapping is generally limited and, generally, only the presence or absence of mangroves can be recorded (Green et al., 1998). Moderate spatial resolution data, such as provided by the Landsat and SPOT series of sensors and also spaceborne SAR, are more useful for providing a regional or subregional overview of mangroves, although Malthus & Mumby (2003) suggested that these data are still really only useful for broad descriptive level mapping. Even so, optical data at this resolution can be used to provide reasonable maps of mangrove extent and broad zonation patterns can be discerned. However, the mapping of change within mangroves that are relatively undisturbed is often limited, although anthropogenic disturbance (i.e., deforestation) can often be detected and mapped using time-series datasets or inferred from single-date imagery (e.g., by identifying areas of recent deforestation or regrowth). The potential of using Global Rain Forest Mapping (GRFM) multi-temporal 100 m JERS-1 SAR data for establishing baseline datasets of mangroves is illustrated in Figure 3.1. To increase the confidence in the baseline datasets, remote sensing data acquired at fine spatial resolution data are optimal.
Data include aerial photographs or those
acquired by multi-spectral/hyperspectral airborne (e.g., digital videos, CASI, HYMAP; Lucas et al., 2003) or spaceborne (e.g., IKONOS or Quickbird; Andréfouet et al., 2003) sensors. At this spatial resolution, detailed mapping of mangrove extent, structure, species/community composition and biomass can be achieved (Malthus & Mumby, 2003). These data can also provide an alternative means for undertaking detailed investigations of plant communities without the risk and cost associated with field surveys. The limitation of fine spatial resolution datasets, however, is that area of coverage is often restricted and data acquisition costs (particularly for repeat passes) are high (Green et al. 1998).
Even so, selected monitoring of key mangrove areas
vulnerable or experiencing change can be undertaken routinely using such data, with
43
moderate to coarse spatial resolution data then used subsequently to provide a landscape or regional view of ongoing processes (e.g., natural or anthropogenic disturbance).
Figure 3.1 Areas of mangroves (red and dark purple) on the east coast of Kalimantan, as identified within the GRFM mosaic (1998, 1996 and 1994 in RGB) generated using JERS-1 SAR data from the Japanese Aerospace Exploration Agency (JAXA).
44
3.1.2 Temporal resolution The importance of temporal frequency of observation depends upon the rates of change that are occurring. In areas where change is rapid, which largely includes those that are affected by anthropogenic disturbance or severe storms (i.e., cyclones or hurricanes), data acquired several times within in a year and preferably before and after the periods of likely disturbance (if seasonal) are preferred. Such areas include South East Asia, West Africa and East Africa, and Southern Brazil, which are being lost or degraded as a result of aquaculture, agriculture and industrialization respectively (Blasco et al., 1998). An example is given for Perak, Malaysia, in Figure 3.2, which shows the loss of mangroves to aquaculture between 1992 and 1998, as observed using the JERS-1 SAR time-series (Ake Rosenqvist pers. Comm.). The gradual conversion of mangroves to shrimp farms is particularly evident. Areas subject to cyclonic or hurricane activities include northern Australia, the South Pacific Islands and the Caribbean. Although Landsat and SPOT sensor data are often appropriate for assessing change, their use in many tropical and subtropical regions is limited due to the prevalence of cloud cover.
For this reason, data acquired by
Synthetic Aperture Radar (SAR) are often required (Simard et al., 2002) as these provide all-weather viewing capability, day or night. Airborne sensors (e.g., aerial photography or hyperspectral) can also be commissioned to acquire data where cloud cover is minimal. In less disturbed areas and where mangroves respond more slowly to natural change, observations on longer time frames (i.e., several years or decades apart) are often sufficient.
45
Figure 3.2 The conversion of coastal mangroves for aquaculture in Perak, Malaysia as observed using time-series of JERS-1 SAR data.
For this latter purpose, archives of remotely sensed data are often beneficial in establishing historical baselines against which to chronicle and quantify change. A key dataset is aerial photography, which has been acquired for many areas during and immediately after World War II. The archives of Landsat and SPOT sensor data, which extend back to 1972 and 1986 respectively, also provide data with which to establish baseline datasets (Riaza et al., 1998). Other archives include the JERS-1 SAR data (acquired between 1992 and 1998), which provide all weather and historical observations of tropical and subtropical landscapes.
However, for many areas of
mangrove (particularly those that are undisturbed), the extent of change between observations may be so small that it is unable to be resolved at the resolution of the observing sensor. The response of mangroves may also not necessarily be in the form of a change in extent and other attributes (e.g., height, species/community composition,
46
structure and biomass) may have to be considered.
However, estimates of such
attributes are often difficult to obtain due to the generally low quality of the data, both in terms of radiometric and geometric accuracy. 3.1.3 Spectral resolution The spectral region in which observations occur largely dictates the level of information that can be extracted from remotely sensed data. For mapping mangrove extent, data acquired in the spectral, thermal and microwave regions can be used (Green et al., 1998; Gao, 1999; Simard et al., 2002). Using spectral information, from different sensors, mangrove vegetation can be better discriminated from non-mangrove surfaces (e.g., landward forest areas or mudflats) using selected wavebands or derived products (e.g., ratios, texture measures). Discrimination of mangrove species, communities and hence zonation patterns can also benefit, although is generally reliant (but not exclusively so) upon the use of optical (visible, near infrared NIR, and/or shortwave infrared SWIR) reflectance data (Green et al., 1998; Held et al., 2003). The importance of using SWIR data for discriminating mangrove types was illustrated by Green et al. (1995) who noted that the absence of this channel on the SPOT HRVIR rendered this sensor less suitable for mangrove mapping compared to the Landsat TM. For quantifying the three-dimensional structure and biomass of mangroves, data from sensors that are capable of delivering digital elevation models (DEMs) of mangroves or record energy from wavelengths (e.g., microwaves, laser) that penetrate and interact with the internal structures of vegetation are essential.
A number of sensors are
available to provide such information but it is the integration of such data with other sources (e.g., optical) that is likely to generate the required baseline datasets for change assessment.
Malthus & Mumby (2003), particularly, stressed the importance of
integrating SAR and optical data, particularly as accuracies in the classification of the coastal zone are likely to be increased, due partly to the greater range of algorithms (e.g., contextual and neural networks) that are available for classification. For the most part, the integration of optical and radar sensors arguably provides the optimal configuration for mapping mangrove communities, although Gao (1999) suggested that higher accuracies of classification could be achieved where thermal data (e.g., as acquired by Landsat TM) were included. The integration of data was reinforced by
47
Kondratyev et al. (1996) who stated that available datasets should be integrated to better interpret and understand the processes, scales and rates of change, both naturally occurring and human-induced.
3.2 Baseline datasets for mangroves The establishment of baseline datasets is fundamental to the long-term monitoring of mangrove response to coastal environmental change and for assessing and monitoring the impact of anthropogenic disturbance. Where possible, baseline datasets should be established from historical datasets as this provides a longer time-frame with which to observe change, whether anthropogenic or natural. This is particularly relevant as Alongi (2002) pointed out that mangroves are likely to be exploited at current rates until about 2025 but the impacts may be moderated thereafter due to advances in technology and ecological understanding. Establishing baselines for the early 2000s will therefore not allow the true extent of change to be assessed. Even so, the establishment of baseline datasets from present day observations is still critical, particularly where no previous data are available, but in monitoring programs within which baseline datasets from several years are compared routinely. Where data from several sources (e.g., SAR, optical) are required to generate these baselines, these should ideally be acquired at a similar time such that integration can be optimized. An important consideration when establishing baseline datasets is that the capacity exists for quantifying change (e.g., remote sensing data are of comparable spatial resolution) and that the methods used to generate and compare the baseline datasets are consistent within and between regions and datasets. There is also a need to acquire ecoregional data that can be used to characterize and better understand the ecosystem response to environmental change such that differences due to climate, other naturally occurring events and anthropogenically induced impacts can be identified (IOC, 2001). Furthermore, observations should take into account lags characterizing ecosystem processes and recognize temporal scale variations. In the following subsections, the use of both airborne and spaceborne remote sensing data for generating baseline datasets of mangroves is outlined. Specifically, baseline datasets of mangrove extent, structure, species/community composition, biomass and productivity are addressed.
48
3.2.1 Mangrove extent Changes in the extent of mangroves are perhaps the clearest indication of a response to changing environmental conditions, whether natural or anthropogenically induced. Such changes may result from the colonization of new areas that are suitable as a result of, for example, changing sediment patterns or tidal regimes (Walters, 2003). Erosion of mangroves may also occur due to increased wave action or storm activity (e.g., lightning strikes). Mangroves may be deforested (e.g., for aquaculture or timber), degrade as a result of thinning or pollution, or regenerate as a result of natural processes or direct planting (Alongi, 2002; Walters, 2003). For many areas of mangrove, old maps of mangrove extent generated through coastal surveys or aerial photography are available to establish an initial baseline. As indicated above, black and white (B & W) photographs were acquired (often in stereo) for many areas of mangrove since the 1940s and 1950s and their relative age renders these as being of particular value for detecting long-term patterns and rates of vegetation change, despite the spectral information being limited to grey levels (Kadmon & Harari-Kremer, 1999). Subsequent photography was in true colour or colour infrared and the greater spectral range provides the capacity to differentiate species and communities as well as map extent. A particular advantage of images of their fine spatial resolution is that differentiation of intermingled and often small communities of coastal vegetation (e.g., sea grasses, saltmarshes and mangrove zones) can be achieved (Chauvaud et al., 1998). Furthermore, stereo viewing allows the three-dimensional structure of mangroves to be observed (Lucas et al., 2002). For mapping the extent of mangroves, data acquired by moderate spatial resolution sensors (e.g., the Landsat ETM+ and SPOT HRVIR) have often been used. Techniques for mapping the extent of mangroves, which are based mainly on the use of Landsat and SPOT sensor data, have been reviewed recently by Green et al. (1998) and a summary table of studies is included in this chapter (Table 3.1). Techniques for mapping include visual interpretation and the use of vegetation ratios such as the Normalised Difference Vegetation Index (NDVI) and SWIR indices based on TM bands 5 and 7.
49
Table 3.1 A summary of image processing techniques applied to remote sensing data acquired over mangroves (Green et al., 1998). Processing method Visual interpretation Gang and Agatsiva (1992) Roy (1989)
Sensor
Validation datasets
Accuracy Assessment
Level of discrimination achieved.
Field data Field data
No No
Paterson and Rehder (1985) Untawale, A. G. et al. (1982)
SPOT XS MK6 KATE-140 Aerial photos Aerial photos
Field data Field data
No No
Vegetation index image Blasco et al. (1986) Jensen et al. (1991) Claudhury (1990)
SPOT* SPOT XS Landsat TM
Aerial photographs Field data Aerial photographs
No No No
Two classes (fringing and cleared mangrove). Percentage canopy closure. Two classes (labeled according to dominant species).
Field data Field data
95% 97% No
Four classes (2 fringing, mixed, shrub and logged mangrove). Three classes (dense, less dense and cleared coastal vegetation). Failed to distinguish mangrove and forest satisfactorily. Four classes (labeled according to dominant species).
Woodfine (1991) Claudhury (1990)
SPOT XS Landsat TM Landsat TM SPOT XS Landsat TM SPOT XS
Field data Aerial photos
N/A No
Supervised classification Dutriex et al. (1990)
SPOT XS
Field data
No
Four classes (labeled according to dominant species and species associations).
SPOT XS SPOT XS Landsat TM MOS-1 MESSR JERS-1 ERS-1 SAR Landsat TM Landsat TM
Field data Field data, maps
91% No
Four classes (2 fringing, mixed, shrub and logged mangrove). Four qualitative density classes (dense and medium, low and very low density).
Field data Field data
No No
Two classes of wetland vegetation. Three classes (2 labeled according to dominant species, cleared mangroves).
Unsupervised classification Vits and Tack (1995) Loo et al. (1992)
Vits and Tack (1995) Aschbacher et al. (1995)
Mohamed et al. (1992) Eong et al. (1992)
Five classes (labeled after dominant species or associations of species) Seven classes (labeled after associations of dominant mangrove species) Four classes (fringing, black, mixed and riverine mangrove). Ten classes (labeled by species or genera).
50
Palaganas (1992) Vibulsresh et al. (1990)
SPOT XS Landsat TM
Field data Aerial photographs
81% No
Two classes (primary and secondary mangrove). Six classes (4 labeled according to dominant species, 2 mixed mangrove).
Biña et al. (1980)
Field data, aerial Photos, maps Field data Field data
85%
Mangrove (as separate non-mangrove vegetation).
No 78%
Mangrove (as separate from non-mangrove vegetation). Six classes (defined from hierarchical cluster analysis of field data).
Woodfine (1991)
SPOT XP Landsat MSS MOS-1 MESSR Landsat MSS CASI Landsat TM
Field data
No
Five classes (mixed community, complex community, transitional to freshwater, transitional to upland vegetation, cleared mangrove with secondary invasion).
Gray et al. (1990)
Landsat TM
Field data
No
Kay et al. (1991) Long et al. (1994) Populus and Lantieri (1991)
Landsat TM Landsat TM Landsat TM SPOT XS Landsat TM
Field data Aerial photos Field data
N/A No No
Three height classes (tall [>10m]), medium [4-10m] and dwarf [ 5 % (Christian & Aldrick, 1977). Other areas are more or less devoid of soil and bare or rocky areas are common. Plant growth is thus limited and highly seasonal in the upland areas. Within the sand plains of the plateau, deeper permeable, siliceous sands occur. They function as temporary reservoirs from which some ground water aquifers are replenished (Christian & Aldrick, 1977), the water of which contributes to dry season flow through seepage points. Towards the lowlands and floodplain areas, soil type ranges from sandy to clay loams, with fine cracking clays underlying the wetland areas. Soils underlying drainage lines are formed on sandy freshwater alluvium (Christian & Aldrick, 1977), and are seasonally waterlogged. With dry season conditions, most of the predominantly clay soils of the floodplain dry out to a depth of ~80 cm (Christian & Aldrick, 1977), and then harden and crack. Those soils of the tidally inundated coastal plain and mudflats consist largely of marine muds and shell material (Christian & Aldrick, 1977).
73
Figure 4.5 Landforms and soils of the ARR (Source: Tropical Savannas CRC, 1998: Savannah Arc Maps).
4.3
Climate and hydrology - the Alligator Rivers Region
The ARR has a monsoonal tropical climate, with an average annual precipitation (at Jabiru) of 1483 mm (Bureau of Meteorology, 1999), mostly falling in the wet season between November and March (Figure 4.6). Annual temperatures generally have low variability, with mean minimum and maximum annual temperatures of ~21° and 33° respectively (Figure 4.6; Bureau of Meteorology, 2003, Climate averages). October and April tend to be transitional months, the pre and post monsoon periods (Finlayson & Moser, 1991), with occasional storm activity. Humidity increases to about 80 % in the wet season, and is less variable during the dry, ranging between 30 – 55 % (Table 4.1). The prevailing winds are from the E to SE in the dry season and N to NW in the wet season (DEH, 3/9/03: Introduction to the ARR). Tropical cyclones and depressions typically occur in the wet season, with up to 2 cyclones and 7 depressions per year affecting the wider NT region (Butterworth,
74
1995). The wet season has the highest storm frequency, and is when the majority of the rainfall over northern areas is received.
Figure 4.6 Mean monthly rainfall and temperature for Jabiru, NT (Adapted from: Bureau of Meteorology, 2003, Climate averages).
Table 4.1 Average climate data for Jabiru airport, NT (Source: Bureau of Meteorology, 2003: Climate averages). Mean of element
Annual mean
January mean
July mean
min air temp (°C)
22.5
24.5
18.4
max air temp (°C)
34.1
33.5
31.7
monthly precipitation (mm)
1485.3
347
2.6
daily evaporation (mm)
7.1
5.9
7
9 am relative humidity (%)
68
82
57
3 pm relative humidity (%)
43
67
31
3 pm wind speed (km/h)
9.7
9.1
10.9
# of Thunder days
60
75
The dominant river systems, including the Wildman River, West, South and East Alligator Rivers, all have their headwaters in the Arnhem Plateau to the south and east of the ARR. The channels extend out across the lowlands in a north-westerly direction, draining the floodplains, and eventually discharging into Van Diemen Gulf. In their upstream reaches, the channels typically have rocky or sandy beds, and dense vegetation often lines their banks (DEH, 3/9/03: Introduction to the ARR), while in the lower reaches, channel morphology is more variable and interconnecting networks of channels and billabongs form, modified through wet season flooding. All of the main rivers are subject to annual flooding during the wet season and, for this reason, both freshwater and estuarine sediments have accumulated to form extensive floodplains in the lower reaches. During the dry season, a steady flow is maintained within the main river channels, while in some of the smaller tributaries (including Cooper, Magela and Nourlangie Creeks), flow all but ceases towards the end of the season (DEH, 3/9/03: Introduction to the ARR). Levees form at the channel mouths of some of these tributaries and require the wet season downpour to re-establish flow with the main river channel. Once the floodwaters have receded, a few permanent freshwater billabongs and waterholes remain. The high seasonality, and in particular, the annual rainfall distribution, is the driving force for the hydrological regime within the ARR. These factors combine to dictate the magnitude of the total annual stream discharge from the main river channels. The total discharge can vary considerably on a seasonal and annual basis, depending on local climatic conditions. The wet season flow generally consists of a series of peak flows alongside the base flow, and typically occurs in the period December – June (DEH, 3/9/03: Introduction to the ARR). There may be localised variations in stream flow, however, attributed to dry or wet years and the influence of El Nino on rainfall availability. Areas that lie within 30 km of the coast are characterized by extensive freshwater wetlands, while mangrove forest dominates the seaward edge and tidal creek fringes, where tidal inundation has favoured the establishment of mangroves. Coastal and estuarine environments experience a macro tidal regime (Brennan, 1996), with spring
76
tide ranges of 5 – 6 m, affecting upstream riverine flows to a distance of around 100 km inland.
Sea level rise in the wider NT region While there is no conclusive evidence to support a sea level rise over the past half century having affected the ARR, the general trends for Darwin indicate a slight rise in the latter part of the century (Figure 4.7).
Figure 4.7 Plots of mean monthly sea level for Darwin: a) 1958 – 2000; b) 1990 – 2003 (Source: archive data from the National Tide Facility, Adelaide, Australia). 77
As the records only cover a relatively short timeframe and are for the Darwin locality, the trends should be interpreted cautiously. Inter-annual variations in climate, including those attributed to ENSO activity, may be responsible for the shorter-term fluctuations in sea level. It is likely however, that the variation in sea level would have affected coastal processes and tidal activity, particularly along many of the northern estuaries (including the ARR) due to the low elevation and undulating floodplain environments (Eliot et al., 1999).
4.4
Plant communities of the ARR
The wetland vegetation of the ARR can be divided into a number of sub categories ranging from the freshwater ponds/swamps and sedgelands of the floodplains, to the harsh environments of the salt marsh and salt flats, and towards the coast, the dense mangrove communities. Each environment is characterized by its own distinct flora and fauna, and is affected to varying degree by the prevailing climate and high seasonality. A number of invasive weed species and feral animals have been introduced into the region, and will be discussed in Section 4.6. Around 1700 plant species have been recorded within the Park itself, with up to 97 labeled as being vulnerable or with limited knowledge of occurrence (DEH, 23/9/03: Plants of the Kakadu region).
Figure 4.8 The major vegetation types within the ARR (Source: Tropical Savannas CRC, 1998: Savannah Arc Maps).
78
4.4.1 Sandstone plateau and lowlands Within the sandstone escarpments of the Arnhem Land Plateau, vegetation must endure the harsh conditions of extreme heat and drought throughout most of the year. Tall open eucalypt woodlands and more mixed species woodlands, sandstone heaths and hummock grasses are characteristic of the vegetation on the plateau (Christian & Aldrick, 1977; DEH, 23/9/03: Plants of the Kakadu region; Table 4.2; Figure 4.9). Table 4.2 Vegetation communities of the Arnhem Land Plateau. Vegetation type
Species
Latin name
Eucalypt woodlands
Eucalyptus
Eucalyptus dichromoploia E.phoenicea E.miniata E.tetrodonta
Woollybutt Darwin Stringybark Mixed species woodlands / understorey species
Rainforest Grasses
Pandanus sp. Native ginger Acacia Xanthostemon Tropical savannah palm Cypress Pine
Ptychosperma macarthuri Callitris intratropica
Monsoon Vine thicket Allosyncarpia
Allosyncarpia ternata
Pandanus basedowii Curcuma australasica
Spinifex Resurrection grasses
Spinifex sp.
Trigger plants
Pityrodia jamesii Stylidium sp.
Flowering shrubs
Figure 4.9 Vegetation of the Arnhem Land plateau (from left to right): Pandanus species, Native Ginger, and Spinifex grass.
Beyond the escarpments of the eastern section lie the lowlands, dominated by open eucalypt woodlands and seasonal distributions of grasses and other small shrubs (DEH, 23/9/03: Plants of the Kakadu region; Table 4.3; Figure 4.10). Spear grass tends to
79
dominate the understorey towards the end of the wet season. The highest diversity is actually seen in the ground layer, where a variety of grasses, sedges and wildflowers abound. Table 4.3 Vegetation communities of the ARR lowlands. Vegetation type Open Eucalypt woodland
Species Bloodwood Red Box Darwin Woollybutt Cooktown Ironwood Billygoat Plum Green Plum Fern-leafed Grevilliea Kapok bush Turkey bush Spiral Pandanus Swamp Banksia
Latin name Corymbia sp. E.ployanthemos E.miniata Erythrophleum chlorostchys Terminalia ferdinandiana Buchanania obovata Grevillea pteridifolia Cochloprerum fraseri Calytrix exstipulata Pandanus spiralis Banksia dentata
Grasses
Spear grass
Sorghum sp.
Sedges
Eleocharis
Eleocharis sp.
Figure 4.10 Vegetation of the KNP lowlands (from left to right): Swamp Banksia, Grevillia and Kapok Bush (Source: DEH, 23/9/03: Plants of the Kakadu region).
4.4.2 Freshwater floodplains A number of factors interplay that directly influence the extent and survival of the floodplain species, including the degree and frequency of waterlogging, fire, and tree density (Finlayson & Moser, 1991). Vegetation is highly seasonal on the floodplain,
80
and all species must cope, in particular, with the inter-annual conditions of drought and flood. Seasonal flooding actually enhances species diversity and richness, through the succession and replacement of species that succumb to desiccation following the change from flooded to drought conditions. There are two cases for lower species diversity however, that occur as a result of an area being either permanently flooded or infrequently inundated throughout the year. Only a limited number of species can endure such conditions. Common species of the freshwater wetlands are listed in Table 4.4 and illustrated in Figure 4.11. Water depth, duration and periodicity of wet season flooding (Finlayson et al., 1989) are major determinants of the changing vegetation distribution on the floodplains. The depth and duration of flooding during the wet season determines the amount of soil moisture available during the dry season to support plant growth. The greater amount of time spent under flooded conditions, the better the soil moisture content following evaporation and recession of floodwaters in the dry season. The period of time during which the floodplain is inundated also affects the life cycle of certain aquatic species. The majority of floodplain species are annual (Finlayson et al., 1989), but some species are able to germinate, grow and reproduce at a faster rate than others, and hence are not restricted to areas that are flooded for a longer period during the year. Some species display certain adaptations or assume certain growth forms in order to survive. For example, the grass species Pseudoraphis spinescens, grows in short tufts when conditions are dry, and develops long trailing stems that can cope with the floodwaters (Finlayson & Moser, 1991). Other species rely on seeds and bulbs (e.g., Eleocharis sp.) that persist in the dry season. Whatever the mechanism, a high diversity is maintained on the floodplains, and is most prolific during the wet season when the plants are flowering. With the return of drier conditions, the colour and variety of the landscape is replaced by more subtle tones and parched surfaces that persist until the end of the dry season.
81
Table 4.4 Vegetation communities of the ARR floodplains (Source: Storrs & Finlayson, 1997; Finlayson & Moser, 1991; Bayliss et al., 1997). Vegetation type Woodland
Species Paperbark Silver-leafed Paperbark Water Pandanus
Latin name Melaleuca quinqueneriva Melaleuca argentea Pandanus aquaticus
Habitat Low-lying, frequently waterlogged areas
Sedges
Eleocharis Spike rush
Eleocharis sp. Eleocharis dulcis
Low-lying, frequently waterlogged areas
Phragmites karka
Low-lying, frequently waterlogged areas
Reeds Shrubs/flowering plants
Blue water lilies Red water lilies
Nymphaea violacea Nelumbo nucifera
Banks of freshwater creeks & swamp areas
Freshwater mangrove
Itchy tree
Abrringtonia actuangula
Low-lying wetlands
Grasses
Pseudoraphis spinescens Oryza meridionalis Hymenachne
Low-lying wetlands
Herbs
Mimulus uvedaliae Goodenia neglecta Lindernia plantaginea
Low-lying wetlands
Figure 4.11 Floodplain vegetation (from left to right): Melaleuca forest, Melaleuca (Paperbark) swamp with Yellow Lilies, and Water Lily (Source: DEH, 23/9/03: Plants of the Kakadu region).
4.4.3 Salt marshes and salt flats Salt marshes typically occur on the landward side of the mangrove community, adjacent to the salt flats (Finlayson & Moser, 1991). Shrubs (on the more well-drained soils that are hypersaline between tides), rushes/sedges (in permanent and brackish waterholes, e.g., Juncus kraussii) and grass species are common in these areas
82
(Finlayson & Moser, 1991).
Communities must withstand relatively high and
fluctuating salinity levels and the effects of seasonal drought/flooding. Up to 10 species of salt marsh plants have been recorded along the northern coast (Storrs & Finlayson, 1997), with 2 of the species (Batis argillicola and Tecticornia australasica) confined to tropical regions. Salt flats are generally very dry exposed areas, often hypersaline if not frequently inundated, and with a salt-encrusted surface. Salt flat vegetation generally occurs towards the landward edge of the mangroves on adjacent mudflats or coastal plain. Vegetation is scarce and usually seasonal, and dominated by low hardy shrubs. Samphire vegetation is common on the salt flats of the West Alligator River (Figure 4.12). Along the South and East Alligator Rivers, salt flats devoid of vegetation are more common (Storrs & Finlayson, 1997). The extent of communities is dictated by local site conditions, including salinity and tidal flows.
Figure 4.12 Vegetation of the salt flats on the landward edge of the mangrove community: Samphire (Salicornia australis) and Halorisia in foreground. Scattered seedlings of Avicennia marina are also present.
4.4.4 Coastal and estuarine mangroves Mangrove forests are prolific along the top end coastline, with the greatest concentration occurring in the Park’s coastal north. Here mangroves occur as narrow fringes along the channel banks and edges of tidal creeks, and in some areas, extend up
83
to 40 km inland. Species diversity within the ARR and the NT, in general, is relatively low, compared to mangroves occupying similar latitudes (e.g., in Queensland), which is attributable partly to the extremes of the wet and dry season and the associated variability in rainfall and evapotranspiration (Woodroffe, 1995).
Around 14 – 21
species of mangrove are found along the northern coastline (Saenger et al., 1977; Storrs & Finlayson, 1997). Forests are highly productive and through tidal flushing, supply an abundance of nutrients and organic material to the marine ecosystem. Distinct zonation patterns are observed across the mangrove community, due largely to the differential ecological performance of species across environmental gradients (Saintilan, 1998), and differential response of species to tidal inundation (frequency and quantity), freshwater flow, soil type, salinity, and wave action (Chapman, 1977; Semeniuk, 1985; Blasco et al., 1996; Storrs & Finlayson, 1997). Both pure and mixed species stands are common within the ARR.
The most frequent species include
Avicennia marina, with its high salinity tolerance, and Rhizophora stylosa, which also tolerates a wide range of salinities but favours those areas more frequently inundated (Storrs & Finlayson, 1997). Ceriops tagal is also relatively common on the landward, unconsolidated clay edge of the mangrove community (Storrs & Finlayson, 1997). Communities at the seaward edge are adapted to withstand salinity levels comparable to that of seawater, while those at the landward edge cope with low soil moisture, high salinity levels and often anaerobic conditions (Lear & Turner, 1994). A variety of useful organic products can be derived from the mangrove forest, including
tannin,
chemical
compounds,
dyes
and
as
viscose-rayon
fibre
(Bandaranayake, 1994). Certain species of mangrove are also used in bush medicine, including A.marina, Camptostemon schultzii, Excoecaria parviflora, and Hibiscus tiliaceus (Bandaranayake, 1994) to name a few, and certain plant parts also provide a source of food. There is a wide range of other uses of mangrove plants and their components, including a source of timber, pulp and charcoal, green manure, pesticides/insecticides and some as poisons. Commercial exploitation of mangroves in the region and in Australia as a whole has been limited.
84
West Alligator River A typical zonation of coastal fringing mangroves on the West Alligator River is listed in Table 4.5. The dominant species within each zone are illustrated graphically in Figure 4.13. Table 4.5 Mangrove communities of the West Alligator River. Species
Latin name
Habitat
Features
Grey mangrove
Avicennia marina
Landward edge
Low shrubs/trees (< 5 m tall), high stem density, low biomass
Red mangrove Orange mangrove Yellow mangrove Grey mangrove
Rhizophora stylosa Bruguiera parviflora Ceriops tagal A.marina
Intermediate zone – landward
Regrowth stands (up to x m tall), low biomass
Red mangrove
R.stylosa
Intermediate zone
Tall homogeneous mature forest (up to 18 m), high biomass
Red mangrove Mangrove apple Grey mangrove
R.stylosa Sonneratia alba A.marina
Mixed zone seaward
R.stylosa dominant. Tall mature heterogeneous forest (up to x m tall), high biomass
Mangrove apple Kapok mangrove River mangrove
S.alba Camptostemon schultzii Aegiceras corniculatum
Seaward
Mature open forest (up to x m tall), medium biomass
Mangrove apple
S.alba
Seaward edge
Dense younger margin (~ 10m tall), present on west bank only
85
Figure 4.13 Homogeneous mangrove stands observed on the West Alligator River (from top clockwise): Avicennia marina, Rhizophora stylosa, Sonneratia alba.
4.5
Fauna of the ARR
Within the floodplain environment and mangrove swamps of the ARR, bird-life, fish, amphibians and reptiles abound. While the majority of these species are not restricted to any particular environment, many use the mangroves for shelter from predators, as a resting-place and nursery, and as a food source. The floodplains provide important breeding grounds for a number of species, including Magpie Geese, herons and the saltwater crocodile (Storrs & Finlayson, 1997). 4.5.1 Birds Birds are prolific throughout the wetlands of the ARR (Figure 4.14; Table 4.6) with around 280 species recorded, and up to 68 species within the floodplain environment alone (Storrs & Finlayson, 1997; Finlayson & Moser, 1991). At least 30 species of migratory birds visit the wetlands during the wet season each year (DEH, 23/9/03:
86
Animals of the floodplains and wetlands). Birds are generally not confined to one particular environment, however their range is typically seasonally and habitat dependent. With the onset of the wet season, the birds disperse and nest elsewhere. It has been observed however, that every seventh year or so, quite devastating losses of nests occur as a result of rapid flooding, while every third or fourth year, the losses are less severe (Finlayson & Moser, 1991). During the dry season, conditions are also harsh and heavy concentrations of birds at limited freshwater sites may disrupt breeding success and adversely affect population densities (Finlayson & Moser, 1991). The South and East Alligator Rivers have been identified as important dry season feeding areas, while the Mary River, to the west of the ARR, is an important breeding area (Storrs & Finlayson, 1997).
Figure 4.14 Birdlife within the ARR (Letters a-j refer to species names in Table 4.6 below; Source: DEH, 23/9/03: Kakadu Image Gallery).
87
Table 4.6 Common birds of the ARR wetlands. Species Brolga (h) Magpie Goose (a) Egrets (b) Straw-necked Ibis Heron (e) Australian Darter Pied cormorants
Latin name Grus rubicundus Anseranas semipalmata Ardea sp. Threskiornis spinicollis Auhinga melanogaster Auhinga melanogaster Phalacrocax melanoleucos
Habitat Floodplain
Wandering whistling duck Plumed whistling duck Burdekin duck (c) Peaceful dove (i) Little curlew Pin-tailed Snipe Bar-tailed Godwit
Dendrocygna arcuata Dendrocygna eytoni Tadorna radjah Geopelia striata Numenius rninutus Gallinago stenura Limosa lapponica
Jabiru White-bellied sea eagle Black Kite Brush cuckoo Paperbark Flycatcher Azure Kingfisher (d) Rainbow Lorikeet Bar-breasted Honeyeater (g) White-lined Honeyeater Yellow Chat (f) Black-banded pigeon Partrdige pigeon Gouldian finch (j) Red goshawk
Xenorhynchus asiaticus Haliaeetus leucogaster Milvus migrans Cuculus variolosus Myiagra nana Alcedo axurrea Trichoglossus haematodus Ramsayornis fasciatus Meliphaga albilineata Epthianura crocea tunneyi Geophaps smithii Chloebia gouldiae Erythrotriorchis radiarus
Paperbark forests
Red winged Parrot Lorikeets Rosella Cockatoo Rufous-banded Honeyeater
Aprosmictus erythropterus Trichoglossus sp. Platycercus sp. Cacatua sp. Conopophila albogularis
Lowlands, water courses
Chestnut rail Mangrove Kingfisher Broad-billed Flycatcher Black butcherbird Mangrove warbler Red-headed honeyeater Mangrove Golden whistler Mangrove robin
Eulabeornis castaneorentris Todiramphus chloris Myiagra ruficollis Cracticus quoyi alecto Gerygone levigaster Myzomela erythrocephala Pachycephala melanura Eopsattria pulverulenta
Estuarine, tidal areas, mangroves
Status/habits
Nest in mangrove trees during wet “ “ “
migratory migratory migratory
Restricted range endangered “ threatened endangered vulnerable
Restricted range “ “ “ “
Coastal fringe
4.5.2 Fish Fish are found in the creeks that drain the mangrove swamps and in enclosed lagoons and floodplain billabongs further inland (Table 4.7; Figure 4.15; Storrs & Finlayson, 1997; Finlayson & Moser, 1991). At high tide, numerous fish invade the mangroves in
88
search of food, and retreat to deeper waters as the tide recedes. The diversity of freshwater fish is high (Finlayson & Moser, 1991) with around 50 species (8 of which have restricted ranges) occurring in the ARR. No fish have been introduced into the Park. Table 4.7 Common fish of the ARR (Letters a-b refer to illustrations in Figure 4.15 below). Species
Latin name
Habitat/habits
Black anal-finned Grunter Arnhem Land Blue-eye
Pingalla sp. Pseudomugil tennellus
Freshwater sp., restricted range “
Archer Primitive Archer Sharp-nosed Grunter Black Bream, Sooty Grunter (a) Black striped Grunter
Toxotes chatareus Toxotes lorentzi Syncomistes butleri butleri Hephaestus fuliginosus Amniataba percoides
specialised feeders; clear waters of escarpment
Forktailed Catfish Toothless catfish Salmon tailed catfish Banded Rainbow Black-lined Rainbow Chequered Rainbow (b) Saratoga Freshwater Longtoms Silver Barramundi
Arius leptaspis Anodoutiglanis dahli Neosilurus sp. Malanotaenia trifasciata Melanotaenia maccullochi Melanotaenia splendida Scleropages jardini Strongylura kreffti Lates calcarifer
surface feeders in creeks
Mud skippers
Periophthalmus Periophthalmodon
scavenger; near escarpment
creeks / billabongs “ “ “ (also escarpment waters) “ (also escarpment waters) “ larger, mobile species, breeds in marine waters, travels upstream during development exposed mudflats
Figure 4.15 Common fish found in the ARR (from left to right): the Sooty Grunter, and Chequered Rainbow fish (Source: DEH, 23/9/03: Animals of the Kakadu region).
89
4.5.3 Reptiles and Amphibians Within the ARR, approximately 117 species of reptiles, 11 types of turtle and tortoise, 25 species of frogs, 11 species of goanna or monitor lizards, 37 skink species and 36 species of snakes, have been observed (DEH, 23/9/03: Animals of the Kakadu region; Storrs & Finlayson, 1997; Table 4.8; Figure 4.16). Of the reptiles, the saltwater or estuarine crocodile is the most well known. The species grows to about 4 – 5 m, and uses the mangrove forest as a food source for crabs, fish, birds and other reptiles. Commercial hunting of crocodiles within Australia was banned in 1971 (Finlayson & Moser, 1991) and since then, the population has increased rapidly to around 40,000. Table 4.8 Reptiles and amphibians of the ARR (Letters refer to Figure 4.16 below). Species Saltwater crocodile (d) Freshwater crocodile
Latin name Crocodylus porosus Crocodylus johnstoni
Pig-nosed turtle Flat-back turtle (a) Loggerhead turtle Long-necked turtle Green turtle Pacific/Olive Ridley turtle
Carettochleys insculpta Chelonia depressa Caretta caretta Chelodina rugusa Chelonia mydas Lepiochelys loivacea
Northern Bullfrog Marbled frog Green tree frog (g) Spadefoot toad Masked frog (f)
Limnodynastes terraereginae Limnodynastes convexiusculus Litoira caerulea Notaden melanoscaphus Litoria personata
billabongs/swamps “ lowland forests “
All generally inactive during dry season
Arafura file snake White-bellied mangrove snake Freshwater snake Macleay’s water snake Olive python Oenpelli python Water python Taipan Death adder King brown snake Western brown snake (e)
Acrochordus arafurae Fordonia leucobalia
Wetlands
Some sp. breed less frequently & have lower metabolic rates to survive
Gould’s goanna Fire tailed Skink Northern Blue-tongue Giant cave gecko (b) Mitchell’s water monitor (c) Mangrove monitor Merton’s water monitor
Tropidonophis mairii Enhydris polylepis Liasis olivaceus Morelia oenpelliensis Liasis fuscus Oxyuranus scutellatus Acanthophis antarcticus Pseudechis australis Pseudonaja nuchalis Varanus gouldii Proablepharus tenuis Tiliqua scincoides intermedia Pseudothecadactylus lindneri cavaticus Varanus mitchelli
Habitat estuarine areas
Status / Habits Tolerates salinities up to 35 % Reduces food intake Eats terrestrial plants Endangered vulnerable vulnerable
Sandstone plateau Lethal to humans Lethal to humans Lethal to humans Lethal to humans Lowland woodlands
Varauns indicus Varanus mertensi
90
Figure 4.16 Reptiles and Amphibians of the ARR (Source: DEH, 23/9/03: Animals of the Kakadu region).
4.5.4 Invertebrates Approximately 10,000 species of insects are found within the Park, including grasshoppers, beetles, flies, termites, butterflies and moths, bees, wasps, ants, dragonflies, non-biting midges and mayflies. This great diversity is attributed to the diversity of habitats supported within KNP.
The relatively high temperatures and
humidity throughout the year also favour large populations of insects (DEH, 23/9/03: Animals of the Kakadu region). Termite mounds are typically found in open forest, with some mounts growing up to several metres tall (Figure 4.17a). One of the most striking insects found within the Park is Leichhardt's grasshopper (Petasida ephippigera; Figure 4.17b), with its brilliant orange, blue and black colours, and is typically found on the Arnhem Land Plateau.
91
Figure 4.17 Invertebrates of the ARR (from left to right): a) Termites and their characteristic mounds; b) Leichhardt’s Grasshopper (Source: DEH, 23/9/03: Animals of the Kakadu region).
4.5.5 Mammals Around 60 mammal species inhabit the Park, including marsupials and placental mammals. The majority are nocturnal and inhabit the open forest and woodlands (DEH, 23/9/03: Animals of the Kakadu region), which provide shelter and a food source (roots and insects). Of the marsupials, there are 8 species of macropods and a variety of flying foxes, possums, bandicoots, quolls, phascogales and the antechinus. Of the placental mammals, there are 26 species of bats, 15 native rodents, one species of dog (dingo), and one dugong species. Table 4.9 Mammals of the ARR. Species Agile wallaby Antilopine wallaby Short-eared rock wallaby Black wallaroo Brush-tailed phascogale Brown bandicoot Northern quoll
Latin name Macropus agilis Macropus antilopinus Petrogale brachyotis Macropus bernardus Phascogale tapoatafa Isoodon macrourus Dasyurus hallucatus
Black footed tree rat False water rat Golden-backed tree rat Dingo Dugong Black fruit bat
Mesembriomys gouldii Xeromys myoides Mesembriomys macrurus Canis familiaris Dugong dugong Pteropus alecto
Status
Vulnerable Vulnerable
92
4.6
Threats to the ARR
A number of anthropogenic threats and natural disturbances have been identified within the Top End region. The TEC region (as described in Section 4.2) is under the greatest pressure from overuse and conflicting land use, and as such has been identified as bioregion under the greatest pressure (Storrs & Finlayson, 1997) in the NT. Tourism, contamination, overgrazing and poor land management (in the uplands), introduced pasture species, weeds and feral animals and changed fire regimes (Storrs & Finlayson, 1997), all contribute to the complex nature of the problem. Naturally induced processes of change, including salinization, dieback and deterioration of freshwater wetlands, storm/cyclone activity, are typical of the region, and the impacts of these are more widespread in some areas than others. 4.6.1 Introduced animals Asian water buffalo, cattle, pigs, horses, cats, dogs, mice, European bees and the cane toad were all introduced at different times into the Park. Over the years, many have contributed to the deterioration of the floodplain and wetland environment, with the greatest damage caused by the water buffalo and wild pigs (Figure 4.18; Table 4.10).
Figure 4.18 Introduced animals of the ARR: the Water Buffalo and Cane Toad (Source: DEH, 23/9/03: Management of Feral Animals).
93
Table 4.10 Feral animals of the ARR, their history of invasion, and current control measures. Species / Latin name
Invasion history / Damage
Control measures and current status Eradication program began in 1979. Scattered herds remain in Arnhem land. Total eradication unlikely due to remote terrain and cost. Benefits of reduced #s: Domestic livestock protected, clearer water in billabongs, floodplain plants re-established Negatives: rapid expansion of weeds into former bare areas, overtaking riverbanks.
Asian Water buffalo
Introduced into NT 1820s for farming; rapidly spread across ARR lowlands. Damage wetlands through sheer size, weight & hard hooves. Compacted soil inhibits plant growth, causes erosion. By 1960s, ~4,000 ha of reed swamp destroyed. In-stream wallowing, creation of swim channels: erodes riverbanks, destroys levees, increases turbidity, enables saltwater intrusion, degrade freshwater habitats. Compete with native wildlife for food/grasses.
Wild pig Sus scrofa
Cause widespread damage to wetland areas through digging for roots. Exposed areas are prone to weed invasion (particularly by Mimosa), & scouring/erosion during wet season. Compete with native animals for food, e.g. bulbs along wetland shores sought by Magpie geese.
Control is difficult but undertaken regularly (by shooting) in known areas by Park staff.
Cane toad Bufo marinus
Introduced to Queensland ~50 years ago; subsequently spread to NT. First sighting in KNP 12/3/01. Toad is poisonous throughout life cycle, so presents danger to animals that might prey on it (e.g. goannas, snakes, quolls).
Limited knowledge of current distribution. No control measure in place. Likely to become pressing management issue over next decade.
Horses
Typically found in southern woodlands. Spread weeds and damage waterholes.
Cats
Introduced from urban areas. Prey on native wildlife, compete for food.
Limited numbers present in Park. Shot on sight by Park staff.
Dogs
Interbreed with native dingoes, subsequently change dingo gene pool.
Limited numbers present in Park. Shot on sight by Park staff.
94
4.6.2 Invasive weeds The main species causing widespread deterioration of wetland environments in the ARR include Salvinia molesta and Mimosa pigra (Storrs & Finlayson, 1997; Table 4.11; Figure 4.19).
Through choking of waterholes, water quality is subsequently
degraded and unsuitable for use by other aquatic plants and animals of the area (Robert Dyason, 2003: Agfact: water hyacinth). Currently, alien species account for less than 5 % of the floodplain flora within the coastal north (Storrs & Finlayson, 1997). The majority of impacts of the various weed species are not well known, with some species having been partially controlled, while others continue to invade wetland areas. Table 4.11 Noxious weeds of the ARR wetlands. Species / Latin name Prickly shrub Mimosa pigra Floating weed Salvinia molesta
Water hyacinth Eichhornia crassipes
Para grass Brachiara mutica Aleman grass Echinocloa polystachya Olive hymenachne Hymenachne amplexicaulis Mission grass Pennisetum polystachion Gamba grass Andropogon gayanus
Habits
Damage
Overtake surrounds by forming dense mats over water surface, & outcompete native plants. Easily dispersed: rapid expansion.
Reduction in water flow & quality: degraded habitat for wildlife & recreational users. Creates monocultures. ~80,000 ha within NT top end coastal region is affected by Mimosa.
Forms dense mat over water surface
Degraded water quality: reduces dissolved oxygen and carbon dioxide levels.
Invasive behaviour “
Transform wetlands into monocultures, causing reduced biodiversity.
“
“
Control measures and current status Mimosa: under control within Park by ‘search and destroy’ campaign. Salvinia: partially eradicated with introduction of a weevil.
Extent of impacts unknown.
Invade open Eucalypt forests, increase fuel loads, leading to more destructive fires.
95
Figure 4.19 Invasive species of the floodplains and swamps: Eichhornia crassipes (Water Hyacinth), Salvinia molesta (floating weed) and Mimosa pigra (prickly shrub) (Source: Dyason, 2003, Agfact: water hyacinth; DEH, 23/9/03: Plants of the Kakadu region).
4.6.3 Fire/burning regime Fire is an accepted part of the northern landscape, but limited information exists on the ecological impacts to the wetland environment. Within KNP, there has been a deliberate campaign to restore the burning regime used in Aboriginal culture (Storrs & Finlayson, 1997). Traditional fire regimes involved more localized and ‘cooler’ burns within the dry season, as opposed to the extensive late dry season burns that are currently destroying large tracts of land. Around half of the region is burnt each year, with notable destruction of large areas of mature Melalueca forest. Some would argue that firing frequency has increased over the past 20 years, with the impacts as far reaching as the coastal floodplains and some riparian wetlands (Storrs & Finlayson, 1997). Spot fires also occur as a result of lightning strikes during the build up to the wet season storms.
96
Of the more long-term effects of increased firing frequency and widespread burns, the most destructive are those that modify the naturally resilient or fire-sensitive species. For example, the scattered patches of monsoonal rainforest and the heathlands of the upper plateau region are highly sensitive to fire, and require around 4 – 5 years between fires (Woinarski, 2001: Arnhem land tropical savanna) to maintain sustainable communities.
As well, changes in the structural and floristic composition of the
Eucalyptus forests of the plateau and surrounding lowlands, are affecting the wildlife that use these areas, with some species having more restricted ranges as a result. 4.6.4 Salinization The potential threat of saline intrusion into freshwater wetland areas is ever increasing, particularly with the predicted climate changes and sea level rise over this century. At least 240 km² of floodplain in the ARR has already been significantly degraded by saltwater intrusion over the last 50 years (NT government, 1998). The landward extension of many tidal creeks and subsequent movement of saltwater into former freshwater areas is also widespread in the ARR (Figure 4.20). Along many of the major river systems (including the East, South and West Alligator), the expansion of the tidal creek networks in the downstream reaches of channels has lead to the penetration of saltwater into important freshwater habitats such as the Paperbark swamps, which provide essential habitat for a variety of native species. The relatively even terrain and low elevations across the floodplain surface are prone to tidal invasion and channel cutting.
With even a minor rise in sea levels, these areas and their
respective plant communities will succumb to even further intrusions and subsequent habitat modification. Introduced animals including the water buffalo and wild pigs are also contributing to localized salinization of wetland areas and floodplains by inadvertently scouring out areas which become prone to erosion particularly with the onset of wet season flooding, and destroying natural levees that keep the saltwater confined to the channel.
97
Figure 4.20 Salinization of floodplain areas (Source: DEH, 3/9/03, Introduction to the ARR).
4.6.5 Mining impacts, contamination/pollution In such a diverse setting as the ARR, and with multiple users of the environment (e.g., tourism, mining, agriculture), the far-reaching effects of pollution and deterioration of natural resources presents a considerable threat. The potential impacts from mining in the ARR (Figure 4.21) include decreased water quality, interrupted natural flows, accumulation and transport of heavy metal and radioactive elements, and contamination of downstream wetland environments and fish. Mining activity within the Park however, is under strict regulation, with stringent water quality assessment and rehabilitation programs for discontinued mine sites already underway. There is limited knowledge of the impact of pesticides and fertilisers used within the wider NT area. Some pesticides have been used within the Park to eradicate invasive weeds, but no detrimental effects have been observed.
Figure 4.21 Aerial view of Ranger mine, ARR (Source: DEH, 3/9/03: Ranger Uranium Mine).
98
The effects of chemical pollution from sunscreens, soaps and insect repellents (Storrs & Finlayson, 1997) used by Park visitors should not be overlooked. Where recreational use of small freshwater ponds is allowed, the accumulation of chemicals may have adverse affects on the natural biota and water quality. The potential problems of fuel spillage and pollution by boats may also contribute to degrading water quality. Camping is under strict regulation and is only permitted in designated areas, but should continue to be monitored. 4.6.6 Tourism From a survey undertaken in 2000-2001, it was found that approximately 200,000 people visit the Park per year, mostly staying up to a maximum of 3 days (DEH, 23/9/03: KNP visitor survey results). The ever-increasing numbers of visitors to KNP will continue to cause localized impacts to certain environments. Tourism requires certain infrastructure and amenities, and with restricted access to many areas of the Park in some months of the year, over-use is the greatest threat. Recreational fishing for barramundi is a popular activity within the Park, and is already under strict regulation. In 1978, the impact of over-fishing of barramundi in some northern rivers meant that stocks had surpassed their maximum sustainable yield (Storrs & Finlayson, 1997). It is also anticipated that further pressure will be felt on black bream populations, as access to upland gorges becomes less restrictive. Hunting of geese and other waterfowl is yet another activity subject to the effects of over-use by recreational groups. Recent regulation has been enforced (restricted use of lead shot in some areas), and further research into the impacts of hunting activities on population distributions is underway. Important breeding sites have been identified and are subject to intense regulation. Hunting of waterfowl may also cause lead poisoning through ingestion of lead pellets (Storrs & Finlayson, 1997). A 5-year program is currently being developed to completely phase out the use of lead shot in waterfowl hunting (Storrs & Finlayson, 1997).
99
4.7
The dynamic environment
The ARR presents a complex environment and one that requires a multi-faceted approach to management.
The complex interrelationships that exist between the
wetland and floodplain environments at the local and landscape scale, requires catchment-wide strategies. The wetlands cannot be considered in isolation as discrete land units (Storrs & Finlayson, 1997); rather, the region should be managed in its entirety, to maintain its complex physical and biological linkages. The mangrove communities within KNP have been largely immune to human disturbance. However, in recent years, changes in their height and extent have become particularly noticeable and are believed to result from a changing coastal environment. In particular, many are colonizing the tidal creeks that are gradually extending inland whilst expansion and retraction of mangroves is occurring along the coastal margin. The inland extension of creeks is of particular concern as the associated intrusion of saltwater is leading to deterioration of the proximal freshwater environments, including the Paperbark (Melaleuca) swamps. With future predictions of global climate change and sea level rise, the situation may very well be exacerbated, with further expansion of the mangroves and other saline wetland communities (Bayliss et al., 1997; Eliot et al., 1999).
4.8
Concluding summary
This chapter has provided an overview of Kakadu NP within the Alligator Rivers Region within Australia’s Northern Territory. Key conclusions are: •
The Park is diverse and contains a wealth of natural environments.
•
Topography varies across from the Park, from high plateau (rising 200 – 300 m) to lowlands, which are only a few metres above sea level.
•
Cyclones, lightning strikes and storm surges are typical of the region’s weather, all of which impact the mangrove communities.
The coastline is relatively
protected however, which favours the establishment of mangroves. •
Water flows are dictated by the seasonal rainfall variability (the majority of rainfall occurring in the wet season) and also by the tides. A macrotidal range
100
within estuaries (ranging 6 – 8 m) ensures strong bi-directional currents, and penetration of saltwater up to 100 km inland. •
Mangroves are confined to the coastal regions, but also the edge of tidal creeks.
•
Key habitats inland from the mangroves are the salt marshes and freshwater wetlands, some of which have been affected by saltwater intrusion.
•
Human interference to mangroves has been relatively minimal because of their isolation and difficulty of access.
•
Mangrove communities are important centres of biological diversity, with a widely variable flora and fauna. They play a key role in the ecosystem as a haven for wildlife, particularly migratory birds and juvenile fish stocks that shelter in their waters, and are also important for coastal protection.
101
CHAPTER FIVE
REMOTE SENSING DATASETS.
For the project, stereo aerial photography, hyperspectral Compact Airborne Spectrographic Imager (CASI) data, and NASA JPL Synthetic Aperture Radar (AIRSAR) data were acquired. This chapter therefore provides the details of these airborne remote sensing datasets, which were investigated for their potential to generate baseline datasets of the extent, height, density, species/community composition and biomass of mangroves within KNP. The pre-processing of these datasets, in terms of calibration and georectification is also outlined.
5.1
Remote sensing data: availability and acquisition
5.1.1 Stereo aerial photography For the coastal fringes of Kakadu NP, stereo aerial photographs were acquired in both 1950 and 1991 and these represented the main remote sensing datasets used in the study. The photographs were acquired early in the dry season in July, 1950 and also May, 1991 (Table 5.1). Although black and white (B&W) photographs were acquired in 1950 for the entire region, only two covering the northern section of West Alligator River were analyzed in the study (Figure 5.1). These photographs were taken at a flying height of 25,000 ft (7622 m) and with a camera with a focal length of 152 mm. True colour stereo aerial photographs acquired in 1991 were also obtained for the mouth of the West Alligator (Figure 5.2) and were subsequently used in a pilot study to assess their potential for generating digital elevation models (DEMs) of mangrove canopy height, and in conjunction with the 1950 B&W photographs, for quantifying changes in mangrove height and extent over time (Lucas et al., 2002).
102
Figure 5.1 1950 Black and white stereo aerial photograph of the mangroves at the mouth of the West Alligator River (UTM Zone 53, Datum: WGS84).
Figure 5.2 1991 True colour stereo aerial photograph of the mangroves at the mouth of the West Alligator River (UTM Zone 53, Datum: WGS84).
103
Following the success of this venture, the full set (~66 stereo pairs) of true colour photographs acquired in 1991 over Kakadu NP were obtained. These provided near complete coverage for the West, South and East Alligator Rivers as well as the Wildman River and both Field and Barron Islands (Table 5.1). These photographs were taken at a flying height of 13,000 ft (3960 m), using a camera with a focal length of 152 mm. The overlap between the colour photographs was 60 %, as is typical for stereo pairs, and specular reflectance was not prominent within these images. Table 5.1 Available aerial photography for the Kakadu region. Photo date
Film No.
Region
July, 1950
SVY1023
West Alligator River
No. of photographs 2
May, 1991
KNP800
Wildman R. West Alligator R. South Alligator R. East Alligator R. Field Island Barron Island
16 21 29 50 13 4
5.1.2 Compact Airborne Spectrographic Imager data Although the study by Lucas et al. (2002) indicated that the height and extent of mangroves can potentially be mapped using stereo aerial photographs, the ability to discriminate mangrove species and communities was limited primarily by the availability of only true colour imagery. For this reason, hyperspectral (visible to near infrared)
Compact Airborne Spectrographic Imager (CASI-2) data were acquired
over the mangroves of the West Alligator River in July, 2002, as part of this project. The CASI-2 is a CCD pushbroom imaging spectrograph, which had been modified for atmospheric research (ITRES, 28/10/03: CASI 2; Table 5.2) but is also commonly used in terrestrial applications. The CASI data were acquired over the mangroves of the West Alligator River in fourteen adjoining 1 km x ~15 km strips by Ball AIMS (Adelaide, Australia; Figure 5.3). These data were obtained at 1 m spatial resolution in the 446 nm – 838 nm wavelength region (Table 5.3), with a FWHM varying from 8 to 23.4 nm. The wavelength regions were selected to characterize the main reflectance
104
features of the vegetation spectral reflectance curve.
The data were flown at
approximately midday, flying in a north-south direction.
An inertial Navigation
Systems (INS) was on board to allow georegistration of the data.
Table 5.2 CASI-2 sensor specifications (Source: ITRES, 28/10/03: CASI 2). Instrument Type Instantaneous Field of View (IFOV) Max. No. Spatial Pixels Max. No. Spectral Pixels Spectral Range Lowest Resolution Flying Height (for 0.8 m resol.) Operating Modes Dynamic Range
Pushbroom Imaging Spectrograph 37.8° across track (mode dependent), 0.077° along track 512 (across track) 288 per spatial pixel (at ~1.9 nm intervals) 413 – 958 nm 0.6 m 2,000 feet above ground level Spatial, hyperspectral, full frame 4096:1 (12 bits)
Table 5.3 CASI wavelength regions and FWHM values. Band 1 2 3 4 5 6 7
Wavelength region 446.6 530.4 550.1 569.8 598.2 634.2 680.8
FWHM
Band
9.6 13.4 11.6 9.8 13.6 9.8 8.0
8 9 10 11 12 13 14
Wavelength region 696.1 714.3 732.5 741.1 752.6 800.6 838
FWHM 11.8 9.8 8.0 9.8 13.8 9.8 23.4
105
Figure 5.3 CASI-2 image of the West Alligator River mangroves: Bands 14 (838 nm), 9 (714 nm) and 1 (447 nm) in R, G, B. (Projection: UTM Zone 53, Datum: WGS84).
106
5.1.3 NASA JPL Polarimetric Airborne Synthetic Aperture Radar (AIRSAR) data Although the upper canopy structure (density and crown area) can be characterized using fine resolution optical data, the underlying canopy structure is better analyzed using radar frequencies and polarizations that penetrate the canopy. As well, studies (e.g., Proisy et al., 2002) have indicated that the above ground and component biomass can potentially be quantified using empirical relationships. On November 22nd, 1996, full polarimetric AIRSAR data were acquired simultaneously for the ARR mangroves as part of the PACRIM Mission to Australia. Data were acquired at P- (0.44 GHz), L- (1.25 GHz) and C- (5.3 GHz) bands at a mean incidence angle of 42° (Figure 5.4; Table 5.4). Furthermore, an additional over-flight occurred on September 9th, 2000, and full polarimetric L- and P-band data and also CVV data were acquired. These datasets were of limited use however, as a result of high image speckle, particularly at P-band, which could not be sufficiently filtered. As well, a drop out during acquisition had occurred, and coverage had been lost over the lower part of the West Alligator River. As such, only the 1996 AIRSAR data was processed for this project. The AIRSAR instrument is a left-looking Synthetic Aperture Radar (SAR) that operates within NASA’s DC-8 airborne research laboratory (Evans et al., 1995; Table 5.5).
Table 5.4 AIRSAR data acquisition parameters (1996). Acquisition date Strip id Centre latitude Centre longitude Bandwidth Cross-track swath (km) No. of samples Reduction ratio Along-track swath (km) No. of lines Reduction ratio Near look angle Far look angle
22 November, 1996 cm5177, Pt.Farewell 246-2 -12.18 132.46 40.0 8.5 2560 0.2 60.7 6628 (C band), 6570 (L/P bands) 0.5 22.4 61.6
107
Table 5.5 AIRSAR instrument specifications (source: Evans et al., 1995). Radar Band (wavelengths, cm) Polarization Incidence Angle Resolution (m) Swath Width (km) System Sensitivity (dB) Altitude (km) Simultaneous Frequencies Simultaneous Polarizations Orbit inclination Bandwidth (MHz) Data rate (Mbps) Operating Modes
C (5.6), L (25), P (68) All 17 – 60 5 10 – 15 -40 7.3 3 4 aircraft 20, 40 256 Polsar XTI or TOPSAR (cross-track interferometric) ATI (along-track interferometric)
108
Figure 5.4 AIRSAR images of the ARR coastal strip (1996): C-, L- and P-bands (HH, VV and HV in R, G, B respectively). 109
5.2
Stereo photography: pre-processing
The following section outlines the steps undertaken in the processing of stereo aerial photography prior to the generation of baseline datasets of the extent and also height and density of mangroves. 5.2.1 Scanning and calibration Pre-processing of all aerial photographs was undertaken in several stages, namely scanning, calibration and image resampling. For the project, all of the 133 photographs were scanned by Airesearch (Darwin) such that digital images with a pixel resolution of between 12 - 15 µm were generated. At this resolution, the uncompressed image files were of the order of 800 Mb in size. Images were provided in jpeg compressed tiff format, with a compression factor of the order of 4 times, so that little or no detail was lost in the decompression of the image. Colour differences were evident between many of the 1991 images, which was attributed to the lack of (or inadequate) colour balancing applied to the raw scanned images, although corrections were applied using available software with reasonable results. Following scanning and calibration, the images were imported into a Helava digital photogrammetric workstation (DPW) and displayed using LH Systems Socet software (Marconi Integrated Systems Inc., 1998). Each image was imported into Vitec format, the default format for Socet, and an image pyramid was generated commencing with the smallest scale (1:1024 minification; a term used to describe the generation of image pyramids) and finishing with the largest scale (1:1 minification). The approximate position of the camera during acquisition was scaled from 1:100,000 topographic maps of the area and defined in the import properties dialog. This was necessary for the successful computation of the bundle adjustment described below. 5.2.2 Aerial triangulation For mapping purposes, aerial photography is typically acquired in strips with 60 % overlap between successive photographs in the strip and 20 % overlap between adjoining strips. To generate orthomosaics and also derive elevation, both the exterior and interior orientation of the camera needed to be quantified. The term “exterior
110
orientation of a camera” is used to describe the position (in three dimensions, 3D) and the direction of pointing when a photograph is exposed (Slama, 1980). The term “interior orientation of a camera” refers to the internal optical parameters of a camera (Slama, 1980). As aerial survey cameras are calibrated at regular intervals, parameters such as focal length and lens distortion were available for determination of exterior and interior orientation.
In the interior orientation process, calibrated marks on the
photograph, which are generally referred to as fiducial marks, were measured thereby allowing the calibration parameters of the camera for each photograph to be extracted. To determine the exterior orientation of a strip or block of photographs, a small number of ground control points (GCPs) were used in a process known as aerial triangulation (Figure 5.5). This process consisted of two phases, the point measurement and bundle adjustment phase. In the point measurement phase, the image coordinates of common points in the overlap area (Tie points) and GCPs were measured. In a DPW, this is accomplished semi-automatically using image-matching techniques. The ground coordinates of the GCPs together with their accuracy were also recorded. Once the measurement phase was complete, all exterior orientation parameters were computed simultaneously in a least squares solution using a rigorous mathematical model called a bundle adjustment. The term “bundle” refers to the bundle of rays that are modelled for each photograph and join the object points, the perspective centre of the camera and the corresponding image points. In this process, image parameters are adjusted iteratively until the points in adjoining images are similarly positioned, thereby reducing the line and sample differences between points measured in overlapping images.
111
Figure 5.5 An overview of the procedure for determining the exterior orientation for a stereo photographic pair in the point measurement phase of aerial triangulation. Ground Control Points (GCPs) and image points (Tie points) common to overlapping images are identified, and their ground coordinates and accuracies recorded (Source: Marconi Integrated Systems Inc., 1998). Following determination of the interior and exterior orientation of overlapping photographs, stereo images were viewed in a stereoplotter such that 3D measurement of objects in the stereo overlap area could be undertaken.
Within a DPW, the
measurement process automatically generates Digital Elevation Models (DEMs) of the ground surface in the overlap area. The KNP photographs were orientated using standard collinearity equations (Slama, 1980) and resampled to 0.85 m and 0.37 m spatial resolution for B&W and colour photographs respectively. To assist this process, the study area was broken down into more manageable blocks, with each block containing a set of photographs for a particular river system (Figure 5.6). This was also undertaken to minimise the disk space requirements, as each image and corresponding image minifications required approximately 1.2 Gb of disc space. As only two adjoining B&W photographs were processed for the West Alligator River, these were incorporated into the one block. Fiducial marks were then measured semi-automatically in each image, with the resulting residuals being less than one pixel.
112
Field and Barron Islands
Wildman River
West Alligator River
South Alligator River
East Alligator River
Figure 5.6 Topographic map of the ARR with the blocks of photographs highlighted.
For the exterior orientation of both the B&W and colour aerial photographs respectively, GCPs observed within 1989 1:50,000 and 1971 1:100,000 scale topographic maps of the ARR (Table 5.6), were digitized. Most GCPs were associated with permanent features as considerable movement in the extent of vegetation covers between the different years was evident. Scanned images of the maps produced in 1971 were supplied by ERISS for this purpose. The photographs were then georeferenced using polynomial transformations to Universal Transverse Mercator (UTM) coordinates (UTM Zone 53, Datum: AGD66). In conventional mapping projects, GCPs are measured to an accuracy of better than the resolution of the photography. This would imply accuracy in this study of the order of 0.85 m and 0.37 m for the B&W and colour photographs respectively. Also, the GCPs should be able to be identified on the photographs to this accuracy. Neither of these criteria could be met in this study as the scale of the maps was insufficient and the resolution of detail on the maps was far lower than that of the photographs. Therefore, common control points were not easily identified. Furthermore, digitising points from a 1:100,000 scale map implied an accuracy of 0.5 mm on the map and 50 m on the ground.
113
Table 5.6 Map sheets used for geometric rectification of the B&W and colour aerial photographs. Map date Map sheet
1989 Field island South Alligator River
Sheet #
5373 (sub 1) 5373 (sub 2)
Edition Scale Reliability
1-AAS 1:50,000 ± 25 m positional ± 10 m height
1971 Field Island Pt. Farewell West Alligator River East Alligator River 5373 5473 5372 5472 1:100,000 ± 25 m positional ± 10 m height
GCPs were therefore distributed broadly across the images, and assigned a horizontal accuracy of 50 m and a vertical accuracy of 2 m.
Positional accuracy was also
determined through identification of height points, which were accurate to within 1 – 2 m (vertical resolution) on the topographic map and could subsequently be used as control on the images. Numerous tie points were located in the overlapping areas of images to help stabilise adjoining images. The vertical datum of the coordinate system was established by identifying areas along the coastline, and assigning an arbitrary height value of 1 m above mean sea level. The accuracy of this measurement was set as 2 m so as to establish a base level datum. Spot heights within tidal flat and upland forest areas were also retrieved from the topographic maps and assigned an accuracy of 1 m.
The actual surface height of the mudflats is slightly variable beneath the
mangroves and towards their landward and seaward margins (± 1 m), but few height data were available to achieve the necessary ground control. Having established a set of GCPs for a block of photographs, the triangulation was solved. GCPs were then edited, where necessary, to reduce the root mean square (R.M.S) errors to less than one pixel. Horizontal accuracies of less than 50 m were considered acceptable (refer to Appendix A: Triangulation reports: stereo aerial photography).
Following the successful triangulation of a block of photographs,
adjoining strips were added using the same GCP file. GCPs that occurred in the
114
overlapping region between strips were assigned the coordinates from the triangulated images, so as to maintain the stability between strips. The output from the bundle adjustment gave residuals to all measurements for each point in image coordinates. These residuals were checked and if necessary points reobserved until the residuals were less than 1 pixel, thus ensuring that the relative positioning within the block is of the order of 1 pixel or 0.85 m (B&W photos) and 0.37 m (colour photos) on the ground. The residuals on the GCPs were checked to ensure that they were within the predicted horizontal accuracy (i.e., ± 50 m) and vertical accuracy (i.e., ± 2 m). Due to the low horizontal accuracy of the GCPs, the absolute accuracy of the position of any object point on the orthophoto-image is of the order of 50 m. The accuracy of the scale of the orthophoto-image is higher as it depends on how far apart the GCPs are; the further apart, the more accurate the scale.
Hence, relative
measurements between points on the orthophoto-image are more accurate than the absolute accuracy of their positions. 5.2.3 Image resampling Following triangulation, stereo model images were generated through the pairwise rectification process (Figure 5.7). In this process, a stereo (overlapping) pair is used to generate two new images that are rotated and scaled such that stereo visualisation is possible (Marconi Integrated Systems, 1998). With stereo visualisation, surface heights can be extracted, thereby allowing the generation of DEMs of surface objects and terrain of greater accuracy. Additionally, image rectification removes distortions related to camera obliquity and topographic effects, or allows rotation of an image so that north faces upward. Three options are available for the rotation of images: epipolar (rotation so that the epipolar direction is horizontal); relative to north (rotation to an angle relative to north, with a default angle of zero degrees); and oblique (images are rotated obliquely assuming that the ground lies in the YZ plane). Each stereo image is rotated and resampled to the same ground sample distance using one of three methods: nearest
115
neighbour, bilinear or cubic interpolation. The nearest neighbour method simply uses the closest pixel to match the output pixel, without any interpolation.
Bilinear
interpolation uses a weighted average of up to four adjoining pixels to create the output pixel. Lastly, cubic polynomials are used to calculate the output pixel from sixteen neighbouring pixels in a cubic interpolation. In this study, stereo pairs were rotated such that the epipolar direction was horizontal, and resampled to the same ground sample distance (0.85 m and 0.37 m for B&W and colour images respectively).
Figure 5.7 The image rectification process involving rotation along the epipolar plane and resampling of stereo pairs (Source: Marconi Integrated Systems Inc., 1998).
5.3
CASI data: pre-processing
5.3.1 Data calibration As data for calibrating the CASI were not available, the use of a coefficient file for calibration derived for Injune, in central Queensland (Paterson et al., 2001) was used as the bandset was identical and furthermore, similar conditions were experienced at the times of the overflights. Although it was recognized that absolute calibration could not be achieved, relative calibration was reasonable.
116
At Injune, and using ENVI’s empirical line calibration, image spectra were matched to selected field reflectance spectra of both black and white tarpaulins or Pseudo Invariant Targets (PITs) laid out during the time of the CASI-s overpass. The process uses a linear regression for each band to equate the new digital numbers and the surface reflectance (resampled to the CASI bandwidths), effectively removing solar irradiance and atmospheric effects. The empirical line is based on Equation 5.1 (RSI, 2001): Reflectance (field spectrum) = gain * radiance (input data) + offset
Equation 5.1
For the West Alligator CASI data, the Injune calibration file was applied using ENVI’s empirical line calibration utility. Reference spectra for different mangrove species were collected with an ASD field spectroradiometer in September, 2002 (refer to Chapter 6: Field data collection) and these were compared with the spectra obtained by the CASI. A reasonable correspondence between image and field derived spectra was obtained, suggesting that the use of the Injune calibration files was acceptable. 5.3.2 Rectification Initially, the fourteen strips of CASI data had been mosaicked and subsequently registered to a 20 m digital elevation model (DEM) over the region by Ball AIMS. The mosaic was then re-registered to the stereo aerial photographs (1991, covering the West Alligator River) using common GCPs such that a close match could be achieved. RST warping (rotation, scaling and translation) with a second degree polynomial and nearest neighbour resampling was applied to the image, with RMS errors of < 2 m being obtained for the registration (refer to Appendix B: GCPs and registration report: CASI). GCPs were distributed evenly over the image and were located where objects (e.g., individual trees and creek intersections) were visible in both. Around thirty-three GCPs were used in the registration process.
5.4
AIRSAR data: pre-processing
5.4.1 Data calibration With SAR processing, calibration is the process whereby raw digital numbers (in intensity values, ranging 0 to 1) are converted to useable radar brightness or backscatter
117
coefficients (σ°). From the data, the backscattering coefficients σ ° HH, σ ° VV and σ ° HV (expressed in decibels, dB) were calculated for each frequency. This was achieved through an automated C-processing function that initially converted the data from slant to ground range format and generated associated coherence and phase information (provided by C.Proisy). Resampling from slant to ground range is an essential first step as it removes distortions in the imagery caused by side-viewing of terrain by the radar instrument (Elachi, 1987). Image speckle, an artefact of radar image geometry and acquisition, was reduced through application of a Lee filter in ENVI. The Lee filter is a standard deviation based (sigma) filter that filters data based on statistics calculated within individual filter windows (RSI, 2001). The pixel being filtered is substituted by a value calculated using the surrounding pixels. As an adaptive filter, the Lee and other similar sigma filters preserve image sharpness and detail while suppressing noise.
The most effective
adaptive filter will minimise the variation of the standard deviation over the mean of the image (Wakebayashi & Arai, 1996), and hence maintain the original image values to within an acceptable range. A 5x5 multiplicative filter was applied to the data with selected thresholds (noise mean 1, noise variance 0.25). The Coefficient of variation (COV) was calculated in both the raw and filtered image to determine the effective use of the filter (Table 5.7). The COV was equated by dividing the mean of the digital numbers by the standard deviation.
An acceptable outcome would be where the least change is observed
between the COV of the raw and filtered imagery. After comparing the COV obtained using different filters or with variable filter parameters, the optimal approach may then be determined.
118
Table 5.7 Statistical outcomes of raw (left) and filtered (right) imagery. Band C-HH C-VV C-HV L-HH L-VV L-HV P-HH P-VV P-HV
Mean -12.09 -12.33 -19.65 -13.82 -15.96 -22.83 -14.26 -16.23 -24.65
S.D 10.36 10.16 17.24 11.36 14.5 20.44 10.94 13.41 21.78
COV 1.17 1.21 1.14 1.22 1.1 1.12 1.3 1.21 1.13
Band C-HH C-VV C-HV L-HH L-VV L-HV P-HH P-VV P-HV
Mean -12.10 -12.33 -19.66 -13.83 -15.97 -22.84 -14.27 -16.25 -24.66
S.D 10.76 10.5 17.57 11.73 14.87 20.76 11.36 13.81 22.18
COV 1.13 1.17 1.12 1.18 1.07 1.10 1.26 1.18 1.11
5.4.2 Rectification Following calibration, the entire SAR mosaic was registered to the 1971 1:100,000 UTM maps (UTM Zone 53, Datum: AGD66) of the study area used for rectification of the stereo aerial photography (Table 5.6). The registration of AIRSAR imagery was essentially the same as for optical datasets, with the exception of the use of higher order polynomials in the transformation to maintain the image’s geometric integrity. GCPs were selected quite liberally over the images and a 3rd degree polynomial with nearest neighbour resampling was applied. R.M.S errors of 0.72 and 1.68 were achieved for Cband and L- and P- bands respectively (refer to Appendix C: GCPs and registration report: AIRSAR). The C-band data was registered separately to the L- and P-band data, due to the different spatial size. Following the initial registration to the topographic maps, the mosaic was subsetted for individual mangrove systems and registered to the corresponding ortho-images, generated through triangulation and resampling as outlined above.
As such, the
accuracy of the registration was improved through closer matching of points common to both images.
A third degree polynomial with nearest neighbour resampling was
applied, and R.M.S errors of < 2 m were achieved for all bands.
5.5
Concluding summary
For the project, remote sensing datasets that were already available for all mangroves of Kakadu NP included 1950 B&W and 1991 true colour stereo aerial photography and 1996 and 2000 NASA JPL polarimetric AIRSAR data.
As the stereo aerial
119
photography was limited for species discrimination, 1 m spatial resolution CASI data were acquired in 2002 specifically to fill this gap, although only for the West Alligator River. Each set of data was georeferenced to an acceptable level of accuracy such that comparison between datasets could be undertaken. The CASI and SAR data were calibrated to surface reflectance (%) and the backscatter coefficient (dB) respectively. The following chapters focus on the subsequent processing and analysis of these datasets for generating baseline datasets of
mangrove height and extent,
species/community composition and biomass.
120
CHAPTER 6
FIELD DATA COLLECTION AND ANALYSIS.
To support the interpretation of remote sensing data and validation of data products (e.g., canopy top height estimates from aerial photography), field measurements relating to the structure, biomass and species/community composition of mangrove forest were required. The following chapter outlines the collection of these data and is divided into two parts. Part I outlines the methods used in the collection of such data from the West Alligator River in both 1999 and 2002. For each campaign, methods of field data collection and analysis are described. Part II integrates the datasets collected in 1999 and 2002 to provide an overview of mangrove community structure, species composition and biomass at the field level. The first field campaign was undertaken in October, 1999, by Joanna Ellison (University of Tasmania), Max Finlayson (Environmental Research Institute of the Supervising Scientist, ERISS), Peter Brocklehurst and Bart Edmeades (Department of Infrastructure, Planning and Environment, DIPE) and also David Klye, Buck Salau, and rangers Garry Lindler and Victor Cooper. During this campaign, two line transects were established on the north-east bank of the West Alligator River, and standard inventory measurements (e.g., tree height, diameter and species) were recorded within each, largely to support the validation of digital elevation models (DEMs) generated using stereo aerial photography. During the second campaign in October, 2002, which was supported by ERISS, thirteen plots were established on the west bank of the West Alligator River. The data collection was undertaken by myself, Kirrilly Pfitzner and Gary Fox (ERISS), and also Peter Brocklehurst and Chris Mandium (DIPE). This campaign was undertaken to further support the calibration and validation of remote sensing products (e.g., DEMs, biomass and species maps). Measurements were also
121
recorded to facilitate subsequent on-site parameterisation of the SAR backscatter model of Karam et al. (1995), although this analysis was not undertaken as part of this thesis. PART I
6.1
DATA COLLECTION.
West Alligator River (East bank), 1999
The first field campaign focused on the mangroves of the east bank of the West Alligator River, which were accessed by vehicle and helicopter. 6.1.1 Transect and sample locations Two line transects were established through representative and accessible stands of mangroves on the east bank of the river mouth and facing Van Diemen Gulf (Figure 6.1). The position of the transects was determined on-site using an Ashtech differential Global Positioning System (dGPS), and by establishing a base station at 132° 19.140’ E, 12° 21.762’ S, approximately 7 km NE of Transect 1. A non-differential GPS unit was also used.
Figure 6.1 Field sampling locations, East bank, October 1999: a) Transect 1 upstream, and b) Transect 2 towards river mouth.
122
Transect 1 was located on the upper east bank, and perpendicular to the shoreline, and was accessible by vehicle and walking. The starting point was located at 132° 16.31’ E, 12° 23.38’ S and the end point was on the riverbank at 132° 16.26’ E, 12° 23.34 S. The transect extended along a bearing of 310°. Transect 2 was accessible only by helicopter to the landward side of the transect owing to the inability to use vehicles or boats to gain access.
The transect commenced on the landward side of the
mangroves (132° 17.36’ E, 12° 13.27’ S) and was completed on the seaward margin (132° 17.30’ E, 12° 13.07’ S). This transect was perpendicular to both the shore and mangrove zonation, and extended from an inner flat of samphire at a magnetic bearing of 339°. 6.1.2 Standard plot measurements Within each distinct mangrove zone, and at regular intervals along each of the transect lines, three replicate plots (5 x 5 m) were established. Plot size was kept constant at each site and covered an area of 25 m². From the starting point of Transect 2, a 100 m measuring tape was extended and continually repositioned as the field team progressed.
This was useful for measuring zonation changes and site locations,
particularly as the GPS was restricted to open canopies where, unlike in closed forest, a signal could be retrieved. The species composition was recorded in each plot and at certain positions along the length of the transect where there was notable change in diversity. Standard plot measurements included the basal area, density, the diameter, height and species of each tree, and crown cover and density. The basal area, or amount of woody tissue (m² ha-1) of each stand was estimated using the prism sweep method or Angle Count Cruising Method (J. Ellison, Pers. Comm.). The method uses a critical angle from a central location to determine the inclusion or exclusion of individual trees within a sample.
The critical angle is
determined using wedge prisms of variable size (a 1 factor prism was used in this case), with trees sighted in a 360° sweep. For each tree sampled within each plot, trunk diameter (at breast height, 130 cm) was measured and a trained botanist identified the
123
species. Three replicate basal area surveys were undertaken at each site along the transect. The height of trees was measured directly using an extension pole and the mud surface as the reference datum. The profile generated from these measurements was to be compared against the height profile extracted from the equivalent area of a digital elevation model (DEM) of mangrove canopy height generated using 1991 stereo aerial photographs (refer to Chapter Seven). Crown cover, or the degree of canopy closure, was also estimated visually in each plot. Where tree crowns overlapped, the crown cover was approximately 80 %. Crown density, or the density (percentage) of leaves in the crown was also estimated. 6.1.3 Transect descriptions Along each transect, and as presented in the following sections, the mangrove forests were also described in terms of their height, density and species composition.
Transect 1 The starting point was on the landward edge of the mangroves adjacent to the floodplain.
The floodplain was approximately 1.5 km wide and vegetated
predominantly by 50 cm grass, possibly Sporobolus. To the east and landward side of the floodplain was 10 – 12 m tall paperbark woodland (Melaleuca cajaputi). The transect commenced at dGPS position 132° 16 31.14912, 12° 23 38.64433 and extended along a bearing of 310°. Table 6.1 summarises the location and species composition at each site along the transect. Table 6.1 Transect and plot description. Site (Plots)
Location
Dominant Sp.
Other Sp.
Height (m)
1 (1,2,3)
132° 16 30.65282 12° 23 38.26309
Avicennia marina
Ceriops tagal Lumnitzera sp.
up to 4 m
Transition
132° 16 29.75599 12° 23 37.76291
A.marina A.marina
2 (1,2,3)
Xylocarpus moluccensis Bruguiera exaristata Camptostemon schultzii A.marina
2m up to 10 m Bruguiera parviflora C.tagal Sonneratia alba Excoecaria agallocha Rhizophora stylosa
15 m
124
Toward the landward edge of the mangroves (Site 1), Plot 1 was located on the transect line, Plot 2 was located to the SW, and Plot 3 further to the SW and adjacent to Plot 2. All plots were equidistant from the mangrove edge and perpendicular to the transect line. The transitional zone, composed largely of A.marina, was noted for the abrupt change in canopy height between young and mature stands. Due to the height and density of canopy cover, the dGPS could not be used after this point. Toward the riverbank, mangrove species diversity increased. Site 2 plots were located in this tidal creek (mixed) community, where freshwater input from both the river and floodplain leads to a greater number of species and no significant zonation as is usually observed in mangrove communities. This community extended all the way to the riverbank and overhung the water in some locations. A steep drop was observed from the level of the mangrove mud into the river, which caused small creeks to downcut as they drained the mangrove area. Plot 1 was located on the transect line, Plot 2 was located to the SE, at the same distance from the river bank as Plot 1, and Plot 3 was located east of Plot 2, within 20 m of the river bank.
Transect 2 Transect 2 commenced on the landward edge of the mangroves, bordering the salt flats and Samphire communities, and extended along a bearing of 339°. Figure 6.2 illustrates the transect established through the mangroves and the equivalent area taken from an aerial photograph. Figure 6.3 presents a flow diagram for the successive mangrove zones observed along Transect 2. Up to five distinct mangrove communities were encountered from the landward to seaward edge across the transect. Extensive A.marina (in juvenile and intermediate forms) formed the landward zone, within which Site 1 plots were located. Beyond this zone, there was an abrupt change in tree height (8 – 10 m to a maximum of 15 m) where the mature R.stylosa was encountered.
The mixed zone in between these two
communities was composed largely of R.stylosa, with patches of A.marina, C.tagal and B.parviflora. Site 2 plots were located in these tall R.stylosa and also mixed species communities.
125
Beyond the Rhizophora zone was a transitional coastal community, consisting largely of R.stylosa, S.alba and occasional A.marina trees.
S.alba dominated the
seaward edge, with tall open forest and scattered individuals of the species Aegialitis corniculatum, Camptostemon schultzii, and A.marina occurring.
Site 3 plots were
located within that mature S.alba stands, which were either pure or contained a mix of other species.
Figure 6.2 Approximate location of the three sampling sites (Transect 2) on the East bank, October 1999, overlain on an aerial photograph.
126
Salt flats with predominantly Samphires: Halorisia and Bartis (~30 cm tall).
LANDWARD
132° 17.36’ E, 12° 13.27’ S.
0 m Narrow hedge of A.marina (~1-2 m) enclosing open glade. A.marina seedlings (up to 50 cm) and Samphires in glade. Start of transect.
11 m Edge of glade. Dense, tall A.marina (~4 m) with sparse covering of pneumatophores (~5-10 cm). 12.8 m No Samphires in understorey.
34 m Open glade visible 10 m W of transect. Predominantly A.marina forest. Location of Site 1 plots. PLOT 1 On transect. Pure A.marina (~4.8 m). Stem density ~10,575.8 ha-1, DBH < 15 cm. PLOT 2 10 m E of transect. Pure A.marina (~4.8 m). Stem density ~14,795.1 ha-1, DBH < 15 cm. PLOT 3 10 m E of Plot 2. Pure A.marina (~4.8 m). Stem density ~16,298 ha-1, DBH < 10 cm. 57 m Open glade with A.marina seedlings (~75 cm) and Samphires. Shallow creek observed draining at an angle of 330°, S-N in central east of glade. Within glade: low hedge of A.marina (2-3 m) followed by large open area of A.marina seedlings and Samphires. Scattered taller A.marina trees throughout.
100 m tall).
Edge of glade, with closed A.marina forest (~2 m
107 m Clearing with patches of 75 cm tall A.marina seedlings.
113 m Dense closed A.marina forest (up to 4 m in height), with understorey of seedlings and 5 cm pneumatophores.
150 m Extensive A.marina forest (2 m tall) to E and W of transect.
163 m
A.marina forest (~3-4 m). First R.stylosa seedling.
127
168 m Open clearing with 1 m A.marina, and a patch of 1-3 m tall C.tagal and 50 cm tall Aegialitis annulata seedlings. Patch of C.tagal (4 m) to W of transect.
174 m Mixed tall C.tagal and R.stylosa (~10 m) to W of transect.
195 m Mixed tall R.stylosa and B.parviflora forest. Scattered B.exaristata seedlings.
200 m Pure tall R.stylosa (8-10 m), with occasional dead trees standing throughout zone.
214 m Tall R.stylosa (up to 15 m), with understorey of R.stylosa saplings (3-4 m).
238 m Tall open R.stylosa forest. No understorey saplings observed. Location of Site 2 plots. PLOT 1 On transect, 251.1m. Pure tall R.stylosa forest. Stem density ~2,210.1. ha-1, DBH < 30 cm. PLOT 2 307 m. Pure A.marina forest. Stem density ~1,693.9 ha-1, DBH < 40 cm. PLOT 3 343.1 m. Mixed A.marina forest. Transitional zone with mixed forest on seaward edge of the R.stylosa zone. Stem density ~2,008.5 ha-1 (A.marina), DBH < 30 cm; ~356.3 ha-1 (C.schultzii), DBH 10-25 cm; and 25.2 ha1 (S.alba), DBH 20-25 cm. 323 m Mixed forest, with predominantly R.stylosa (~13 m tall) and occasional C.schultzii trees and seedlings. Dead S.alba trees also observed in this zone. Tidal height as measured on the trees was 1.2 m above level of substrate.
400 m Mixed S.alba and C.schultzii forest (~13 m tall). Scattered A.corniculatum trees throughout zone. Location of Site 3 plots. PLOT 1 Mixed open S.alba forest. Stem density ~54.1 ha-1 (A.marina), DBH 20-35 cm; ~349.45 ha-1 (C.schultzii), DBH 15-52.2 cm; ~117.25 ha-1 (S.alba), DBH 20-59 cm. 128
PLOT 2 Mixed coastal S.alba forest. Stem density ~314.9 ha-1 (A.marina), DBH 15-50 cm; ~363.7 ha-1 (C.schultzii), DBH 5-35 cm; 380.44 ha-1 (S.alba), DBH 5-60 cm, ~226.4 ha-1 (A.corniculatum), DBH 510 cm. PLOT 3 18 m W of transect, 488 m. Pure open S.alba forest. 20 m from seaward edge of mangrove forest. Stem density ~44.3 ha-1 (C.schultzii), DBH 245 cm; 310.7 ha-1 (S.alba), DBH 20-45 cm; ~81.5 ha-1 (A.corniculatum), DBH 10-15 cm.
SEAWARD 132° 17.30’ E, 12° 13.07’ S.
Figure 6.3 Schematic diagram for successive mangrove stands observed along Transect 2, East bank. Plot summaries with stem density and diameter data included also.
6.2
West Alligator River (West bank), 2002
The second field campaign focused on the mangroves that occurred on the west bank of the West Alligator River, which was accessed by vehicle and on foot. 6.2.1 Sampling locations The second field campaign was undertaken primarily to provide additional ground truth information to support a) the classification of mangrove species and communities from hyperspectral CASI data (acquired in July, 2002) and b) the further validation of DEMs derived using stereo aerial photography. Furthermore, the campaign was also undertaken to obtain structural measurements which could be used to support the interpretation of the AIRSAR data acquired over the area, and also to enable future parameterization of the SAR backscatter model of Karam et al. (1995), which had been adapted specifically to simulate the interaction of microwaves with mangroves. In preparation for the field campaign, photographic maps based on available fine spatial resolution CASI imagery were produced. These images suggested the existence of five major zones of mangroves in the West Alligator region distributed from the
129
inland to the seaward edge. Based on the 1999 transect, these zones (from the landward to seaward edge) were assumed (correctly) to be dominated by A.marina, R.stylosa (mature), mixed species, and S.alba (mature with younger fringing margin). During the field campaign, thirteen sample plots were established within relatively homogenous areas representing each of the main mangrove zones observed within the imagery. Where a GPS could be used beneath the canopy, plot locations were recorded. The west bank of the river channel was selected as there was better access to communities. Similarly, the sampling area was agreed upon in the field as there was easy access across the tidal flats, and the mangrove stands appeared quite extensive. The sampling area (Figure 6.4) was bounded by latitude and longitude coordinates of 132° 16’ 4.66” E, 12° 13’ 44.63” S (lower left corner) and 132° 16’ 38.58 E, 12° 13 26.24” (upper right corner).
Figure 6.4 Field sampling area (October, 2002) on the west bank of the West Alligator River, and the main zones, as observed using the CASI data (inset).
6.2.2 Plot descriptions Thirteen sample plots were established in successive mangrove zones on the west bank of the river (Figure 6.5).
Plots WA1 – 3 were established in the landward
A.marina zone, with Plot 1 consisting largely of juvenile A.marina, and Plots WA2 – 3 representing intermediate stages of growth. Plots WA4, WA5 and also WA10 were
130
established within the pure adjoining zone of R.stylosa. Plots WA7 and 11 were established within a transitional zone of mixed species, consisting of A.marina, S.alba and the dominant R.stylosa. Plots WA12 and WA13 consisted of R.stylosa regrowth, occurring toward the landward edges of the mature R.stylosa zone. Plots WA8 and WA9 were established in the seaward S.alba zone, with a younger coastal fringe of S.alba (WA6).
Figure 6.5 Location of the thirteen sample plots on the West bank, Sept/Oct, 2002, using the CASI scene as a backdrop.
6.2.3 Standard plot measurements Plot dimensions Standard vegetation plots of fixed area were established to sample the variation in the density and sizes of trees occurring within and between the mangrove zones. The plot size was selected based on the diameter (at breast height, 130 cm) of individuals within each given zone (Table 6.2). Specifically, 5 x 5 m plots were established to sample more expansive areas of younger regeneration and saplings (< 5 cm diameter) with a
131
high density of trees. 10 x 10 m plots were used to sample trees in intermediate stages of growth (5 – 10 cm diameter) and those on the coastal edge, whilst the larger 20 x 20 m plots were used to sample mature trees (> 10 cm diameter). Each plot was marked with flagging tape and individual trees were tagged temporarily once measured.
Table 6.2 The dimensions of plots used during the 2002 campaign and the measurements recorded. DBH range (cm) < 5 cm
Plot dimensions 5x5m
5-10 cm > 10 cm
10 x 10 m 20 x 20 m
Measurements Tree count, average height, diameter and crown dimensions, growth stage and species. Standard tree measurements (see below) Standard tree measurements (see below)
Stand description Prior to measurement, the stand was described and an indication of the number of structural layers was given (Figure 6.6). Stands with noticeable one or two layers were generally regarded as pioneer, while those with three layers were regarded as mature. Where there was limited distinction between the canopy and trunk layers in the pioneer or younger regeneration, these stands were described as having one layer.
The
intermediate and mature stands however, were of sufficient size and height to have an upper layer consisting largely of the leaves and twigs of the canopy and a lower layer consisting of mainly branches and the trunk. In mature R.stylosa stands, a significant amount of woody material was contained within the above ground roots, and hence this component was described as a separate (third) layer.
132
Figure 6.6 Examples of tree layering and structural measurements within different mangrove zones.
133
6.2.4 Standard tree measurements Each tree measured was identified to species and standard inventory measurements of tree height and diameter, crown and root dimensions, were taken to characterize and compare the structural diversity between and within stands. Within each stand, and for each tree, measurements of the height (m) to the canopy top (HT_TOP, height to highest point of canopy), canopy base (HT_CPYB, distance between the ground surface and lowest part of canopy) and first leafing branch (HT_FLB, distance between the ground surface and first leafing branch) were recorded using height poles or a clinometer. For the tallest trees, the distance was measured to the base of the tree using measuring tape and then, using the clinometer, the top of the tree was sighted and the angle read from the instrument scale. Tree height (i.e., HTTOP, HT_CPYB and HT_FLB) was then calculated from the angle and distance measurements using standard trigonometry. When using the clinometer, the height of the observer was also considered. HT_TOP was measured to validate the DEMs derived from API and to allow relationships between canopy height and diameter to be established as the latter, which was the prime input to allometric equations for biomass estimation, was unable to be estimated remotely. The additional height measures of HT_CPYB and HT_FLB were recorded to give an indication of the growth structure, maturity of the stand and the structural succession within the communities (i.e., whether advancing or regressing) and to establish the distribution of leaves, branches and stems within the mangrove canopies such that the layering within the forests could be better quantified. Trunk diameter was measured with a diameter tape or calipers at 130 cm (DBH). For multi-stemmed trees, the diameter of the base (30 cm) and of each separate stem (at 130 cm) was recorded. For R.stylosa species, diameters were measured at 130 cm and 20 cm above the location where the uppermost root was joined to the stem. Trunk diameter was recorded at several locations to facilitate application of the available allometric equations, which required either DBH_30cm or DBH_130cm as
134
input. Measurements were taken above the root for R.stylosa as the main stem was often difficult to access for measurements due to the root system. Crown dimensions were recorded primarily to verify crown density and area measurements retrieved through analysis of stereo aerial photographs and CASI data. Specifically, measurements of the crown diameter in the east-west (CRN_EW) and north-south (CRN_NS) directions were recorded using tapes stretched out below the crowns or through visual estimation. Measurements of the prop roots for R.stylosa species were also undertaken to assist understanding of biomass partitioning in the above ground woody component. Measurements (using a measuring tape) included the height of the top root (HT_ROOT), the number of visible roots (ROOT#) surrounding the base, and the root length (ROOT_LGTH), which was recorded as the distance of the longest root from its apex on the trunk to the ground surface. 6.2.5 Tree structural measurements Within each identified layer, structural measurements, relating to component tree parts (leaves, branches and trunks), were recorded, primarily to assist the interpretation of the SAR data, and ultimately, parameterization of the SAR backscatter model of Karam et al. (1995). The main parameters measured are listed in Table 6.3. Digital photographs of individual trees were also taken looking vertically upwards (to record crown information) and horizontally (to record vertical structure), and included a reference such as a height pole to establish scale.
Leaf and branch
measurements were recorded for a limited number of individuals in each plot, so that a range in tree size distribution was accounted for.
135
Table 6.3 Structural parameters measured within each stand. Layer Top
Bottom
Parameter Length (cm) Width (cm) Thickness (mm) Density (N m-3) Diameter (cm) Length (cm) Etc….
Leaves X X X X
Branches X
X
Trunks X
X X
X X
X
X
Leaf measurements For each species, leaf length and width (cm) were measured with a ruler, while their corresponding thickness (mm) was measured using a thickness gauge. Leaves were also photographed with a ruler for scale. To give a general indication of leaf density within crowns of younger regeneration stands, a plastic quadrat cube (50 cm x 50 cm x 50 cm = 0.125 m³) was arranged in the canopy (Figure 6.7) and the branches and leaves contained were cut and counted. Indicative densities (N m-³) were then obtained. To estimate the leaf density of taller trees (e.g., R.stylosa and S.alba), digital photographs of the canopies were taken using height poles for reference.
Figure 6.7 Estimating the leaf density within crowns of A.marina using a quadrat cube.
136
For different species, leaf spectra were also collected to initiate a library of spectra for the dominant species, including A.marina, R.stylosa, S.alba, A.corniculatum and C.schultzii, which could be used subsequently to assist classification of the hyperspectral data. A minimal number of leaf samples of the dominant species were collected and packaged on ice for spectral analysis back at base camp. Approximately fifty leaves were included in each sample, and between 2 - 4 samples were obtained for each species. Leaves were retrieved from the outer most part of the canopy, where they had been exposed to full sunlight and not shaded by leaves above. An ASD Field spectroradiometer (available through the UNSW Centre for Remote Sensing and GIS) was used to record the percentage reflectance in the visible to near shortwave infrared wavelength region (VIS-SWIR, 350 – 2500 nm). Leaves were laid out on a tray with a dark surface underneath, so that there were several layers of leaves. A torch was used to provide uniform illumination over the leaves and a fibre optical cable used to take the measurements.
Prior to measurement, the instrument was
calibrated using a white reference panel. Each reading took a series of ten successive measurements, after which the average was recorded.
Branch measurements For each species, branch measurements relating to the length, diameter, orientation and density for each diameter class ( 20 m tall) provided a boundary between the landward Avicennia communities. Tree height declined away from this edge, with the majority of the Rhizophora between 17 – 20 m tall. Towards the seaward edge, mixed transitional forest of variable height (between 12 – 16 m), with patches of taller mangrove occurred. The confinement of species to
411
particular height zones was therefore suggestive of a similar response by species to the environmental conditions in the West Alligator. The integration of datasets also provided a mechanism for better interpreting the information content of different forms of remote sensing data.
For example, the
empirical relationships established between SAR backscatter and forest structure revealed definite correlations with changing tree structural parameters including height and biomass. Without the fine scale DEMs of canopy height (generated from stereo aerial photography) and knowledge of the location and extent of species (through interpretation of hyperspectral data and photographs), the influence of structure on the backscatter response would have been difficult to understand.
This is further
exemplified by the fact that the backscatter response at different frequencies and polarizations was indeed a function of both ‘within’ and ‘between’ zone structures. Between successive zones, the radar response was predominantly a function of tree height and spacing, biomass, and the underlying surface (roughness and dielectric constant), while within each zone, there were differences relating to tree height and diameter, three-dimensional structures, canopy volume and layering. The DEM also provided tree height information on a 1 m x 1 m scale, which could be scaled up to provide the mean height at the level of SAR pixels (around 10 m x 10 m). As the tree height was related to trunk diameter and hence total and component biomass, the resulting datasets provided a powerful interpretative tool for evaluating the SAR response to the changing structure of mangroves which was, in part, a function of the species composition.
13.3 Environmental change in the ARR Over the past fifty years, mangroves have responded to coastal environmental change, including tidal reworking, alternating patterns in progradation and erosion of the shoreline, tidal creek incision and channel migration and habitat expansion. In particular, the invasion of the tidally dominated coastal plains by expanding tidal creeks and subsequent saltwater intrusion in the upstream reaches has augmented a series of changes in mangrove species distribution. Over the past ten years, tidal creek incision
412
and mangrove expansion have continued their landward invasion of the tidal flats, and widespread movement of mangroves on both coastal and estuarine shores has occurred. While there has been no appreciable increase in sea levels over the past fifty years, tide gauge records from Darwin Harbour (collated by the National Tide Facility, Adelaide) indicate a slight increase between the years 1972 – 1975 and 1996 – 2001. The records only cover a relatively short time period however and should be interpreted with caution. Most likely, climatic phenomena including ENSO activity and seasonal variability will have contributed to the trends in tropical Northern Australia. Prior to 1975, there were a greater number of La Niña years (BOM, 2003: Australian climate variability and change), typically associated with wetter conditions and increased storminess. A rise in mean sea levels around Darwin from ~3875 to 4125 mm was observed between 1972 – 1975 (Eliot et al., 1999). During the period 1910 to 1995, total annual rainfall in the NT had increased by around 15 % (BOM, 2003: Australian climate variability and change) with an associated 20 – 30 % increase in heavy rainfall during the Autumnal months. At the other extreme, El Niño years, associated with relatively dry conditions and reduced storminess, were more frequent during 1975 – 1995. Sporadic declines in mean sea level around Darwin over the past decade are most likely associated with these events.
Tide gauge records from Darwin indicated the occurrence of sea level
anomalies (typically the result of climatic phenomena), for the years 1991, 1997 and 2002, wherein a drop in mean sea levels was observed. However, these observations conflict with TOPEX POSEIDON satellite sensor data for the seas north of Australia, which suggest an increase in sea level averaging ~1.5 cm per year for the period December 1992 to June, 2002 (Figure 13.1). The changing patterns in climate and sea level phenomena over the past half-decade may well, therefore, have contributed to changing coastal processes and tidal inundation and frequency in the estuarine environments of the ARR (Eliot et al., 1999).
413
Figure 13.1. Rate of mean sea level change (estimated simultaneously with annual and semi-annual variations) as measured using TOPEX/POSEIDON Altimeter data. The trends were calculated for the period from December 1992 to June 2002.
While the majority of movements are related to naturally induced processes of change, the future holds a markedly different and adverse set of bounding conditions and probable outcomes to change. With rapid sea level rise and anthropogenically induced ‘greenhouse’ conditions, the natural processes and rates of change are predicted to increase on an order of magnitude that may result in a complete reversion to the mangrove swamp forests that existed following the marine transgression, ~6,800 years B.P. The loss of the freshwater dominated wetlands and seasonal habitat for a variety of species could be catastrophic from the point of view that this change is anthropogenically driven, rather than successional, with far-reaching consequences for the adjacent and interconnected wetland systems. The following section reiterates the use of available photography and CASI data for the detection of changes in the extent and species distribution of mangroves between
414
1950 and 2002 for the West Alligator River. The changes within the ARR are then discussed with reference to historical and current data, and interpreted in the context of environmental change in the region.
Future scenarios of climate change are then
provided and the vulnerability of the ARR mangroves and wetland environment assessed. 13.3.1 Remote sensing and change detection The key attributes that are most likely to change with climate and coastal/estuarine development over time are primarily the extent of the mangroves, their structure (height and density), species composition and biomass. For this reason, spatial datasets of these attributes at similar spatial resolution need to be generated on a routine basis. Each of the datasets used to generate these products needs to be georeferenced with a high degree of accuracy such that, ideally, changes at the tree or tree cluster level can be detected. The generation of the baselines will also require data to be acquired by, as a minimum, two different sensors. To generate the baselines of height and density, stereo aerial photography or lidar are most appropriate. However, hyperspectral data at fine spatial resolution are required for species/community discrimination. Changes in the mangroves can then be detected by comparing classifications (i.e., baseline products) or a change algorithm that subtracts the image space of one from another (more appropriate for detecting changes in species or productivity). 13.3.2 Mangrove response along the West Alligator River The analysis of change between the dates 1950, 1991 and 2002 revealed that quite substantial alterations in mangrove distribution had occurred. The latter analysis was limited to the upper section of the West Alligator River with coastal communities near the mouth, extending upstream to a distance of ~8.7 km. The photography provided a broader expanse over which to assess change, as coverage included a larger area on the western bank. In particular, the landward intrusion and spread of mangroves along a tidal creek on the lower part of the western bank was visible on both sets of photographs. General changes over the fifty-four year period included the expansion and erosion of mangroves along coastal facing shores, the inland extension of tidal creeks and subsequent upstream invasion of mangroves on both banks, the substantial
415
seaward advance of the central zone of R.stylosa, and the more subtle changes in species composition within otherwise unchanged zones.
Coastal expansion and retraction The extent of mangroves along coastal facing shores of the West Alligator River has experienced significant change over the fifty-four year period. During this time, and through continual aggradation of sediment along the mud banks flanking the western shore, the seaward zone dominated by S.alba and transitional forest has had the opportunity to expand seawards and increase markedly in area. Through longshore drift and deposition of sediment from predominantly marine sources, the coastline has gradually been prograding, aided by the colonisation of mangroves, which assist in sediment trapping and hence ‘land building’. The mouth of the West Alligator River has a relatively narrow funnel-shape, with the western bank situated lower than the eastern bank. Diurnal flood and ebb tides and longshore currents have favoured the gradation of depositional environments on coastal mangrove shores, upon exiting the river mouth. Towards the upper funnel segment, the mud bank is fairly narrow, but fans out further along the western coast, creating a relatively wide beach area. The larger expanse of mangroves on the western bank has contributed to the greater stability of the near-shore mangrove zone and the area immediately seaward of this zone. As such wave erosion is reduced across this area, and continual aggradation of sediment has occurred within and on the seaward edge of the forest. The mangroves, now with available substrate in which to lodge and being opportunistic species, subsequently colonised the area and continued their seaward expansion over the years. Once established, the mangroves assisted in accumulating sediments and organic material, and maintaining the gradual vertical accretion of the mud bank in which they were lodged. The coastal zone dominated by S.alba incurred an increase of around 250 m on the upper west bank between the years 1950 and 1991. A reduced expansion was evident on the lower bank towards the river funnel, of less than 50 m. This led to the formation of extensive transitional forest (with mixed species composition) and a zone of homogeneous mature S.alba on the seaward edge. During the period 1991 to 2002,
416
further expansion of these zones had occurred, and a younger fringing margin had developed seaward of the mature forest. The decline in tree height on the coastal margin, as indicated by the DEM, is typical of a prograding coast. Around 70 – 180 m of additional growth had established seaward of the mature forest, and was most extensive on the lower – mid western coast, declining away from and into the river mouth. As the river mouth narrows, the tidal velocities increase and hence the erosivity of the banks maintains a lower rate of progradation. Some species (e.g., A.marina and S.alba) produce dense underground root networks consisting of cable roots that extend out from the trunk, from which anchor roots protrude downwards and vertical aerial roots (pneumatophores) protrude upwards (Augustinus, 1995).
By assisting the trapping of fine sediments, the dense root
networks increase substrate and soil stability, provide the necessary support for above ground structures, act as a defence against tidal currents and are used for aeration at lower tides. The roots of both Sonneratia and Avicennia species grow faster and denser than those of Rhizophora species (Augustinus, 1995), facilitating greater consolidation of the soil, and increasing its resistance to erosion. Furthermore, pneumatophores are considered to be the most effective of the above ground root structures in terms of their sediment trapping capability (Augustinus, 1995). As such, these species are typical colonisers of prograding shorelines, being suitably adapted to resist both tidal and wave action.
Sediment transport declines across the landward to seaward edge of the
mangroves and the quieter depositional environment on the landward edge enables fine sediments to accumulate and increases the rate of accretion (Saintilan & Williams, 1999). Over time, a new generation of mangroves may establish on the outer edge of the seaward zone and further contribute to sediment trapping.
Some genera (e.g.,
Rhizophora and Bruguiera) reproduce by vivipary, whereby an almost complete tree is formed within the seed coat (the propagule) and once established, grows into a young upright tree in a few days, as compared to a seed which has to germinate first (e.g., Sonneratia species; Augustinus, 1995). All mangrove propagules have the ability to float with the tides (Knox, 1986), and given a continual supply, rapid colonization of the pioneer species is possible.
Furthermore, the dispersal capabilities of mangrove
417
propagules means, that in the event of a gradual sea level rise, mangroves will most likely persist and even increase their distribution (Panapitukkul et al., 1998). In the case of the West Alligator, a younger margin of S.alba had developed on the western bank over the twelve-year period. The growth and maturity of young S.alba plants requires reduced salinity levels (maximal levels around 5 – 50 % of seawater), either though freshwater input or wet season rainfall (Ball & Pidsley, 1995). The expansion of the Sonneratia zone therefore occurred not only as a result of increased sedimentation on the lower intertidal shore, but also the flushing effect associated with wet season rainfall. The seasonal fluctuations in salinity levels (Ball & Pidsley, 1995) and changes in flooding frequency and rainfall are determining factors in the growth and survival of S.alba. On the eastern bank, mangroves have responded to a predominantly receding coastline and processes associated with coastal erosion and reduced sedimentation. Over the period 1950 to 1991, quite extensive erosion of the eastern coastline had occurred, with the loss of around 100 m of Sonneratia forest. In 1991, the rapid decline in height of trees at the seaward edge was indicative of the receding coast. It could not be discerned from the black and white photography whether the fringing trees along the upper eastern bank were of juvenile status. However, their removal over the forty-two year period presents a significant response to coastal environmental change. For the past twelve years, further losses along the coastal edge forest have occurred, particularly along in-cutting tidal creek mouths and towards the river mouth, where the erosive power is greatest.
Seaward movement of Rhizophora In association with the seaward expansion of mangroves on the western bank, simultaneous change has occurred within the Rhizophora forest over the past fifty-four years. The landward extent of the Rhizophora edge on the western bank has undergone significant alteration during this time, and to a lesser extent on the eastern bank. This edge represents the boundary between the central Rhizophora and landward Avicennia communities, evidenced by an abrupt change in canopy height and productivity (biomass). The shape of the edge has been continually modified over the years through
418
storm activity and changing environmental gradients, creating the gaps and incuts that maintain the irregular edge along its length. By 1991, this edge had advanced seawards by ~140 m on the upper western bank, whilst lower on the bank towards the mouth, the increase was around 240 m. Over the past twelve years, further seaward advance had occurred of the order of ~40 m and 70 m on the upper and lower banks respectively. On the eastern bank, the change was less marked, with the edge having advanced seaward by a maximum of ~40 m in 2002. The drop off in canopy height between the tall Rhizophora (~19 m) and landward Avicennia was different for the western and eastern banks, with heights in the Avicennia of around 6 – 8 m and up to 6 m respectively. Areas of dieback of the Rhizophora were evident along the edge on both banks in 1991, the majority of which had been replaced with younger Rhizophora regrowth by 2002. Both the younger Rhizophora and landward Avicennia forest had begun to expand seaward as the edge advanced. The landward extent of the young Rhizophora however had not changed between the two dates, while Avicennia was observed to be expanding in both directions, encroaching on available habitat. The changing position of the edge is an important indicator of the changing conditions in the external environment that affect plant growth. Zonation patterns in mangroves generally correspond to tidal fluctuations in both duration and frequency (Ball, 1988). One possible explanation for the change therefore, is related to changing tidal movements within Van Diemen Gulf that have affected inland tidal flows. With progradation of the western bank and the seaward expansion of Sonneratia, the area of mangroves had increased by 2002, creating a further barrier to tidal flows through the community. As well, the increased growth of the mature Rhizophora and transitional forest, and the spread of Avicennia along the upper reaches of tidal creeks over the twelve-year period, added to the reduction in tidal amplitudes with distance inland. The reduced flows towards the landward edge favoured sedimentation of the upper intertidal zone and hence vertical accretion in these areas. During the dry season, in particular, it is possible that these factors combined to produce elevated soil salinities and modified nutrient availability along the edge, and Rhizophora began to die back along its landward extent. The loss of degrading Rhizophora may also have been promoted by
419
storm activity, particularly as mangroves on the edge of the zone would receive greater exposure to wind and also high temperatures. Small differences in elevation across the intertidal area can greatly influence the frequency and duration of tidal inundation in the upper reaches (O’Grady et al., 1996). Changing environmental conditions (nutrient and salinity levels) across the landward to seaward edge of the mangroves largely determine their composition, structure and growth form.
For example, hypersalinity, reduced nutrient availability and soil
compaction (Lin & Sternberg, 1991) were contributing factors to the dry season distribution of dwarf and tall Rhizophora mangle L. in Florida. Salinity may reach levels higher than seawater during the dry season when evaporation is at its greatest (Lin & Sternberg, 1991). Where tidal flows have been less restricted, as on the eastern bank (with less expansive mangrove), sediments have been transported landward through the tidal creeks, and long-term sedimentation of the upper intertidal zone has occurred. It is suggested however, that vertical accretion of the upper intertidal zone has not occurred at the same rate as on the western bank, due largely to the counter-balancing effect of uninhibited tidal flows and saltwater intrusion into the upper reaches. As such, a dynamic equilibrium between tidal ebb and flow and processes of sedimentation and accretion seems to have been maintained on the eastern bank, and there was limited difference in elevation across the seaward to landward margins. The processes affecting the western bank Rhizophora were not as significant on this side.
Tidal creek extension and spread of mangroves The rapidity of change associated with tidal creek extension and subsequent saltwater intrusion is largely related to the macrotidal range within Van Diemen Gulf (around 5 – 6 m), bi-directional currents with high potential for channel cutting (Knighton et al., 1991), and the small elevation differences across the coastal plains which are easily exploited by the invading channels (Eliot et al., 1999; Bayliss et al., 1997). Mangroves are opportunistic species and following the inland intrusion of creeks and saltwater, there is high potential for their subsequent growth and expansion into the landward environs.
420
All areas of the West Alligator have extensive salt flats to the landward edge of the mangroves, and a few species have extended their range along the tidal creeks that have carved their way through these flats over the past few decades. Over the past fifty-four years, the tidal creek network has expanded in the downstream reaches of the West Alligator through channel cutting, widening and subsequent saltwater intrusion. Tributary extension was most noticeable on the eastern bank, where one of the larger creeks had extended ~1.6 km inland between 1950 – 1991. The observed extension of tidal creeks was associated with the inland spread of mangroves, and using the latter time-series data (1991 – 2002) and the spectral information, the species Rhizophora and Avicennia were typically identified as the main colonisers. By 2002, a number of creeklines toward the landward margins were no longer visible in their upper reaches. As the creeks intrude inland, they generate large quantities of fine sediment, which is subsequently trapped by invading mangroves along the channel banks, and hence, the upper sections of creeks are being overtaken. Bayliss et al. (1997) recognised the elimination of the smallest tidal creeks within the ARR through the spread of mangroves (typically Avicennia species), following the release of sediment by creek extension and incision. Tree heights on the West Alligator DEM revealed mangroves of up to 14 m along invading tidal creeks, indicative of the process having occurred over a long timescale. As well, relatively large Avicennia canopies could be discerned on the landward margins from the orthomosaic, further evidence of the ongoing process.
Expansion of Avicennia The landward intrusion of tidal creeks and associated advance of the Rhizophora zone has led to the expansion of Avicennia on the landward margins of the community. On both banks, A.marina colonises the space in between the tidal creek communities, and forms a narrow fringe on the landward extent of the mangroves. Over the years, Avicennia has expanded in the landward and seaward direction, overtaking bare or exposed areas, and invading the younger Rhizophora communities. A.marina is one of the more salt tolerant species, and frequently occurs in areas which are seasonally hypersaline (Tuffers et al., 2001).
421
The landward expansion was greatest on the upper western bank, with a 60 – 100 m increase in twelve years. Further upstream, the landward expansion was more subtle, typified by larger canopies and fringing Avicennia dominating the landward edge. Similar change had occurred on the eastern bank, but less extensive movements had occurred. Fairly sizeable Avicennia canopies were observed having encroached on the Rhizophora regrowth areas, which is indicative of long-term movements.
Large
Avicennia trees were observed along tidal creek channels, forming mixed communities with R.stylosa. Between 1991 and 2002, there was evidence for their continued growth, with notable increases in canopy size observed on the CASI image. The landward Avicennia forms extensive homogeneous stands on both banks, beyond which (on the seaward edge) pockets of Rhizophora regrowth are encountered, and then a sudden change in height to the taller Rhizophora forest. A slight difference in canopy height of the landward Avicennia on both banks was indicated on the DEM. On the western bank, mean tree height throughout this zone was observed to be 2 m higher than on the eastern bank. As growth rates and the time of colonization were assumed to be the same, this suggested a change in the immediate environment, in particular the underlying substrate and its elevation. As suggested in a previous section, sedimentation of the upper intertidal zone had assisted vertical accretion in these areas, and as such, the Avicennia community had continued to develop on slightly higher ground on the western bank. There are a number of plausible reasons for the expansion of Avicennia on the landward margins of the mangrove community. On the landward edge, higher salinities prevail as a result of reduced tidal inundation and prolonged exposure during the dry season. A.marina is well adapted to those areas with high salinity ranges, as the species actively excludes some salt at the root level, while excretes the majority of absorbed salts metabolically through glands in the leaves. The expansion of mangroves has been observed to occur simultaneously with progradation of the seaward edge (Saintilan & Williams, 1999) through greater rates of vertical accretion. It is possible that the progradation and seaward expansion of S.alba on the western coast of the West Alligator River, which contributed to greater sedimentation and the seaward movement of mangroves, also assisted long-term vertical accretion on the landward edge. The
422
newly deposited sediment favoured the establishment of mangrove propagules (A.marina) and hence, the community began its landward expansion. During the wet season, tidal flushing and greater rainfall would have assisted the process by freshening the salt flat environment, renewing nutrient levels and promoting the spread of mangroves in these areas (Saintilan & Wilton, 2001). In broader terms, an increase in mean annual rainfall, or greater frequency of heavy rainfall years (possible effect of ENSO activity), may dilute the salts accumulated within salt marsh and salt flat environments (Saintilan & Williams, 1999), thereby encouraging mangrove colonization and expansion on the upper intertidal zone. With the widespread tidal creek intrusion, atypical of the ARR in general, saltwater has penetrated further upstream and A.marina has been quick to colonise.
The
frequency of tidal flushing and seasonally variable soil moisture, salinity and nutrient levels, contributes to the changing physiological conditions and growth of the landward species. Keene & Melville (1999) observed the increased growth of A.marina in highly saline areas through the reduction of salinity-induced potassium deficiency in the soil. Areas that approach almost hypersaline conditions (e.g., the soils underlying the sedges and grasses of salt marsh communities) consist of more oxidised, highly saline sediments with lower available potassium. The landward expansion of A.marina may occur following the re-introduction of potassium into the upper intertidal zone (Keene & Melville, 1999), through wet seasonal flooding or saltwater intrusion. This elevates the potassium potentials (or potassium availability) in the topsoil, which encourages growth of the specially adapted A.marina in these highly saline areas.
Lightning strikes and storm activity Evidence of storm activity and in particular, lightning strikes, was observed on the orthomosaic for the West Alligator. Mangroves are generally well adapted to surviving and regenerating from these natural disturbances, and regrowth is fairly rapid. Bayliss et al. (1997) observed however, that where the small branches were removed following the impact of a cyclone, those species of Rhizophoraceae did not regenerate. Other impacts of storm activity involve destructive winds, increased tidal amplitudes and wave heights, and often flooding of the lower intertidal shore. Lightning tends to strike
423
the tall trunks of the Rhizophora, creating small circular gaps (up to 40 m wide) in the canopy.
These were quite numerous in the central Rhizophora zone on both the
orthomosaic and CASI image of the West Alligator.
Regrowth areas were also
observed, the species of which and their condition could be confirmed through spectral analysis using CASI data. Of interest was that the number of gaps associated with lightning strikes was significantly greater in the 2002 CASI image compared to the 1991 aerial photograph. Information regarding the cyclone history of the ARR is very limited, and the canopy gaps are attributed largely to lightning strikes. The abrupt change in height from the landward Avicennia to the central Rhizophora zone could be identified as a possible impact of cyclone activity. The edge between the two zones is irregularly shaped along its length, and occurs on both sides of the main river channel. Areas of regrowth in between the two zones may represent areas of recovery from a large-scale storm. The edge was already present in 1950 and its changing position was investigated using the three datasets spanning 1950, 1991 and 2002. It is possible that a destructive cyclone occurred prior to 1950 that led to the formation of the edge, and subsequent environmental changes have led to its seaward advance. Alternatively, as described in a previous section, the dieback of the Rhizophora and expansion of the Avicennia through changes in the physiological conditions that affect plant growth are suggested as the main cause for this abrupt change. 13.3.3 Response of the ARR mangroves to coastal change Mangrove communities of the ARR are responding to widespread change augmented through tidal inundation, intermittent and long-term sea level rise, storm activity, and seasonal changes in rainfall and flooding regimes.
The seasonal and inter-annual
variations in these factors (Eliot et al., 1999) create the complex and dynamic environment of the ARR. According to Eliot et al. (1999), the coast progrades when sea levels are declining, precipitation is high and fluvial forces dominate, whereas the coastline retreats and tidal creeks extend landward when sea levels are rising and coastal processes dominate. The impacts of such change on mangrove community structure and composition include the loss or expansion of coastal fringing mangroves, the spread of mangroves along tidal creeks, the encroachment of certain species on neighbouring
424
zones, their landward expansion onto adjacent saline mudflats, and localized losses due to storm activity or dieback. Mangroves of the coastal sections of the ARR are responding to long-term changes in sea levels and tidal activity. When conditions are favourable (i.e., declining sea level, reduced tidal amplitude and inundation frequency), areas of expansion may be observed on many coastal shores. The seaward movement of mangroves assists progradation of the coastline and vertical accretion of the upper intertidal zone. Where this occurs, expansion of both the landward and seaward fringing communities may be observed. Particularly where there are bare or exposed areas, or otherwise available space for their landward migration, mangroves will opportunistically expand in range. Furthermore, a changing species composition may be incurred as a result of changing substrate conditions, salinity and nutrient levels in the areas to be colonized. On more receding coastlines, the erosive force of the wind and waves generally leads to losses of coastal or fringing mangroves, or changes in species composition as the community readjusts. It appears as though mangroves of the West Alligator are constantly readjusting to alternating patterns (largely meteorologically driven) in rising and falling sea levels (with eroding and prograding coasts respectively) over time, as there is no conclusive evidence for a widespread sea level rise in the region. Island communities are responding to changing patterns in sedimentation and erosion, as a result of tidal variations in Van Diemen Gulf. The extent of fringing mangroves around Barron and Field Islands is dependent on the capacity for vertical accretion of the coastal perimeter.
Where sedimentation had kept pace with tidal
inundation and wave erosion, mangroves had expanded beyond their fringing zone. Further ‘land-building’ is possible, pending similar patterns in tidal activity and nearshore currents on these sections of coast. Where the coastline was buffered by wave action and erosion was commonplace, mangroves were more restricted and had not extended beyond the seaward fringe. Patches of isolated mangroves on the seaward edge were indicative of high-energy sections of coast where sedimentation was limiting. In response to tidal currents and long-term sea level changes in Van Diemen Gulf, tidal creek development has expanded throughout the entire ARR coastal network. The
425
high tidal range within the Gulf ensures the predominance of strong bi-directional currents and tidal velocities, which promote the headward extension of tidal creeks through channel cutting and incision. These processes are responsible for the intrusion of numerous tidal creeks in the downstream reaches of the estuarine systems in the ARR. Following periods of sea level rise, increased rainfall or storm surges, these processes are enhanced, and rapid expansion proceeds onto the extensive low-lying salt flats. The typically flat or only slightly undulating terrain across the coastal plains means that they are highly exploitable by the invading channels. Changes in mangrove species composition and rates of colonization and growth along the length of channels are indicative of the long-term intrusion into these areas. The overriding tidal influence, even with distance upstream, means that mangroves will continue to colonize the invading creeks in the upper meandering channel segments. Tidal creek extension can be partially attributed to long-term changes in high tide levels as a result of sea level rise (Eliot et al., 1999). The process is also a likely response to intermittent ENSO activity, with wetter, stormier years associated with short-term sea level and tidal fluctuations, which would favour rapid channel cutting and bank erosion. The subsequent impacts are generally widespread, and in the ARR, have included the damaging effects of saltwater intrusion into formerly freshwater areas or the Paperbark (Melaleuca) swamps that typically occur on the landward margins of the floodplains. As well, there are areas upstream already threatened by saltwater intrusion including Magela creek and Boggy Plains in the upper East and South Alligator catchments respectively. Further upstream, bank erosion, sedimentation and channel infilling (Augustinus, 1995) are responsible for channel morphological changes and variable mangrove distributions in the meandering river segments. These processes are driven by tidal activity and seasonal changes in river discharge and flow.
Changing species
distributions reflect the susceptibility of some species over others, when certain conditions are faced. For example, along the funnel segment of the South Alligator River, B.parviflora is more susceptible to wave erosion than Rhizophora species (Augustinus, 1995), leading to changes in species composition along the steep banks of the main channel.
426
Seasonal changes in rainfall and flooding also have far-reaching effects on species composition and zonation patterns.
The tidal range, particularly the duration and
frequency, is a major determinant of mangrove zonation. Higher species diversity is generally found on coasts with a large tidal range. Further upstream, the changing species composition reflects the changing edaphic and physiological conditions that affect plant growth. Changes in tidal flows, both long-term and rapid associated with meteorological events, result in significant movements in the spatial extent and boundaries between adjacent mangrove zones.
These movements typically reflect
succession of the species and are related to natural processes of change. The high seasonality of the ARR, with monsoonal rainfall and wet season runoff, also has a significant effect on the estuarine and coastal plant distributions. Changes in the wet season flooding regime, either through increased rainfall or drought conditions (possibly related to ENSO activity) affects processes of tidal flushing, nutrient availability and salinity levels, which will lead to visible reductions of certain species or areas of expansion. With a high incidence of tropical storms along the northern coastline, including cyclones, mangrove communities are vulnerable to occasional disruption and damage. Sporadic effects on the mangroves include lightning strikes, which remove small areas of trees, stripping of leaves and branches, and dieback of mangroves in areas subsequently weakened by strong winds and heavy rainfall. The regenerative capacity of most mangrove species is quite high, and areas of regrowth are quick to establish following storm activity. Throughout the ARR mangroves, storms will continue to cause localized damage to plant communities, which will respond through processes of regeneration and succession. 13.3.4 Future scenarios of climate change With the predicted rise in sea levels and global scale climate change as induced by the greenhouse effect, mangrove communities of the ARR, which are already undergoing extensive change, will be subjected to a new range of conditions and indeed threats to their survival.
Of direct relevance to the mangroves, are the changes
augmented through a probable sea level rise and changes in the tidal range within Van Diemen Gulf, the greater storm frequency including cyclones, and higher summer
427
rainfall. As such, the physical threats to the mangroves and wetland environment may include increased flooding, seasonal changes in plant distributions, more frequent storm damage and loss of coastal fringing mangroves. Saltwater intrusion may however favour mangroves, and the landward incursion of mangroves into along tidal creeks and onto salt flats may be commonplace. The floodplains and coastal environ are situated between 3 – 4 m AHD, which means they are only about 0.2 – 1.2 m above the mean high water level (Eliot et al., 1999). With an estimated rise of ~20 ±10 cm for the year 2030 (Eliot et al., 1999) for the Australian coastline, the extensive lowlands of the northern coast will most likely be vulnerable to change. The increased saltwater intrusion as a result of sea level rise and further tidal creek penetration will have far-reaching consequences for the freshwater floodplains and plant communities in the upper reaches of the ARR catchments. Some areas are already under threat, particularly in the South and East Alligator Rivers, and with seasonal changes in rainfall and flooding regimes, a wide-scale reversion to the early mangrove swamps of the Marine Transgression era could result. The processes of tidal creek extension and saltwater intrusion over the coastal plains could be enhanced following an accelerated sea level rise and increased tidal range. This is expected to lead to further channel cutting and expansion of the tidal creek network within the coastal lowlands, and changes in mangrove distributions further upstream. Certain species of mangrove will be more suited to the changing conditions, and hence dieback and replacement of some species will occur. Mangroves are generally situated between the mean sea level and high water spring tide mark (Augustinus, 1995), and as such, will survive a sea level rise if vertical accretion can keep pace and also if landward migration is possible. On the coastal edge, established mangroves assist in sediment accretion, often accelerating the process, and should therefore continue to protect the coastline given a sea level rise. However, accretion of the mud flats situated at the seaward edge of the mangrove community is not uniform over time (Panapitukkul et al., 1998). The process is largely determined by annual and decadal changes in sediment deposition, flooding and the extent of the mangroves situated directly behind. A greater frequency of floods and tropical storms,
428
more intense rainfall and rising sea levels may augment changes in sediment delivery and progradation, leading to reduced rates of vertical accretion within the mangrove community as well as erosion of the coastline. Adverse conditions on the seaward edge may gradually lead to losses in mature fringing mangroves and successive communities may not reach the height and status of those of the present day (Bayliss et al., 1997). With fluctuating tide levels and annual increases in sea level, mangroves, which are usually situated in areas close to their range in tolerance, may not be able to sufficiently recover and mature before conditions change again. Landward migration and recruitment are two of the main strategies that will ensure mangrove survival and most likely expansion in the ARR lowlands. Rising tides will ensure a continual supply of propagules to the landward edge, where fresh sediment and favourable conditions will promote growth and establishment. The longterm re-adjustment of communities to coastal change will involve successive recruitment on the landward edge, relative to present day parent populations (Bayliss et al., 1997). The ability to predict the landward migration of mangroves will depend on the difference in slope and elevation across the landward to seaward edge (Doyle, 2003) and the changing extent of the tidal range. In the ARR, the land situated directly behind the mangroves consists largely of extensive and even salt flats that border the freshwater floodplains toward their landward margins. With the greenhouse induced changes, the landward migration of mangroves, as well as further displacement of salt flat environments and contamination of freshwater communities, particularly the Paperbark swamps on the edges of floodplains is expected to occur. On the Mary River, the catchment to the west of the ARR, engineering works have been undertaken in some freshwater billabongs and other important freshwater bodies, to counter the effects of saltwater intrusion, but to large expense and not always with complete success. The movements of saltwater in upstream environments and associated mangrove expansion will be prolific and only accelerate with the effects of climate change. The signs of this change are already evident from this study, and include the continued spread of mangroves along intruding tidal channels, subsequent colonization of the landward margins, and the ongoing invasion of salt flat and freshwater environments.
429
13.4 Opportunities for spaceborne remote sensing Although stereo photography or hyperspectral airborne digital scanners provide the optimal data source for retrieving structural information of relevance to mangrove and other coastal ecosystems, a variety of spaceborne sensors are available with the capacity to provide fine to moderate spatial resolution data products with the added benefit of repeat coverage over extensive areas. Ultimately, airborne data becomes increasingly less cost-effective when repeat coverage at the landscape to regional scale is required (Sader, 1987). Both Ikonos and Quickbird satellites can provide sub-metre resolutions and data is less costly to acquire than, for example, air photography.
Moderate
resolution Landsat and SPOT sensors have been used extensively for coarse habitat mapping, where accuracies in the classification of around 70 % may be achieved (Mumby et al., 1999).
However, despite the increased cost, finer resolution data
acquired using airborne sensors should be used where available for more detailed investigations where high accuracy is required. Furthermore, a range of satellite SARs are also available with both polarimetric and interferometric capabilities, for analysis of forest structure and biomass.
The following sections outline the sensors that are
available, both currently and in the future, for mangrove assessment and monitoring. 13.4.1 Multispectral satellites Ikonos The IKONOS satellite, operational since September 1999, provides fine spatial resolution multispectral imagery in the visible to NIR wavelength range. The satellite combines a panchromatic sensor with 82 cm resolution providing B&W imagery, and a multispectral sensor with 3.28 m resolution, providing spectral-radiometric data for land cover and vegetation analysis (Dial & Grodecki, 2003). By integrating the two data sources, enhanced colour images can be generated at 1 m resolution with the advantages of both visible and multispectral information. Both natural and colour infrared images may be generated for remote sensing applications. Stereo imagery can also be acquired for 3D feature extraction. Geo images may be acquired by IKONOS on a 3-day repeat cycle (Dial & Grodecki, 2003), and are georectified at a constant elevation.
430
Quickbird The Quickbird satellite was launched in October, 2001, and currently provides the highest resolution commercially available data. Operating in two modes, panchromatic (61 – 72 cm) and multispectral (2.44 – 2.88 m, visible to near infrared, 0.45 – 0.9 µm), the sensors can acquire data every 1 to 3.5 days depending on latitude (Toutin & Cheng, 2003: Quickbird – A Milestone for High Resolution Mapping), and also in stereo. Across-track coverage is 16.5 – 19 km, and image products are available in 16.5 x 16.5 km format. The Quickbird sensor is a step towards bridging the gap between the resolutions afforded by air photography and that of satellite acquisition.
Aerial
photographs can provide resolutions of around 0.2 – 0.3 m, and while Quickbird approaches ~0.6 m, it may prove beneficial for some applications (e.g., vegetation discrimination and mapping) with the added multispectral component. 13.4.2 Landsat and SPOT sensors Landsat and SPOT images have been used widely in landscape and regional scale coastal mapping projects for assessment of community structure and extent. With relatively coarse spatial resolutions, ranging from 10 – 120 m, the data typically requires finer spatial resolution imagery (including aerial photographs or hyperspectral data) for ground truthing and image analysis (e.g., training data for supervised classifications). Currently, a wide range of Landsat and SPOT sensors are operational (Table 13.1). The repeat coverage afforded by these sensors (e.g., 250 images/day by Landsat-7) is useful for a wide range of applications for which moderate resolution data are suited. The bandwidth of SPOT XS bands ranges between 0.07 – 0.1 μm with spectral information acquired in the 0.5 – 0.89 μm region, while Landsat TM bands vary in bandwidth between 0.06 – 0.14 μm, but cover a wider spectral range between 0.45 – 2.35 μm (Gao, 1999).
Only the reflected radiation from the upper canopies of
communities is recorded, and as such, the underlying 3D structure can only be inferred. Time-series of these data are available, but historical data is limited to 1972 (Landsat MSS) and 1986 (SPOT HRV). For mapping of small coastal areas (< 60 km wide), SPOT multispectral (XS) were found to be the most cost-effective by Mumby et al. (1999), while for larger areas Landsat TM was better suited and more accurate. Data
431
have already been acquired over mangrove ecosystems in a number of locations for mapping the broad extent and structure of communities, including northern and northeastern Australia (Long & Skewes, 1996; Danaher & Luck, 1991; Danaher, 1994); New Zealand (Gao, 1999); Madagascar (Rasolofoharinoro et al., 1998) and Japan (Dwi et al., 1997). Table 13.1 Landsat and SPOT satellite specifications. Sensor
Satellite
Channels
Spatial Resolution (m) 30 120 (TIR)
Landsat Thematic Mapper (TM)
Landsat-4 Landsat-5
3 VIS 1 NIR 2 MIR 1 TIR
Multispectral Scanner (MSS)
Landsat 1-5
2 VIS 2 NIR
79
Enhanced Thematic Mapper Plus (ETM+)
Landsat-7
30 15 (PAN) 60 (TIR)
SPOT Panchromatic (PAN)
SPOT 1-4
4 VIS 1 NIR 2 MIR 1 TIR VIS
SPOT Multispectral (XS)
SPOT 1-3
3 VIS 1 NIR
20
SPOT High Resolution Visible (HRV)
SPOT 1-2
2 VIS 1 NIR
20 10 (PAN)
10
Detailed community structure cannot be discerned at the resolutions afforded by the Landsat and SPOT satellites. In particular, the complex zonation patterns and species composition, changing stem densities and canopy cover, cannot be investigated using this type of imagery. Landsat sensors generally support a SWIR channel however, which may assist in discriminating between neighbouring zones of mangroves. Confusion generally occurs however, when trying to discriminate between submerged and partially submerged plants (e.g., seagrasses and aquatic plants), because of a mixed signal response in image pixels. As such, it may not be possible to discriminate between mangrove and non-mangrove vegetation at this resolution, without ancillary data or extensive field knowledge prior to analysis. Leaf reflectance for the different species is also similar at this resolution, and discrimination to species level would be
432
difficult if not impossible. Both Landsat and SPOT sensor data are also affected by cloud cover, which poses a significant disadvantage in tropical coastal areas where humidity is quite high, and it may take a period of several months before images of suitable quality are acquired (Rasolofoharinoro et al., 1998). For broad scale spatial detection of mangroves and other coastal communities however, Landsat and SPOT sensors provide relatively quick and straightforward processing routines for generating maps of spatial extent and distribution. Landsat and SPOT sensor data can be merged through colour transforms to improve spatial resolution and assist the visual discrimination of mangrove and non-mangrove vegetation, although spectral information may be compromised as a result (Green et al., 1998). To discriminate between the observed species and cover types, ancillary data and decision rules (e.g., ground elevation, distance from water) may be used to improve classification results (Long & Skewes, 1996). Vegetation indices and band ratios can also be used to transform multispectral data into a single index for separation of vegetation and non-vegetated surfaces (Green et al., 1998). 13.4.3 Spaceborne SAR SAR sensors have the potential to provide information relating to mangrove forest structure (including height, density and biomass) and community extent. SAR imagery has been acquired specifically over mangroves and other coastal ecosystems using a range of spaceborne sensors including ERS-1 and 2, JERS-1, SIR-C, RADARSAT, ENVISAT, SRTM, and ALOS (Table 13.2).
These sensors acquire coarse –
intermediate spatial resolution data between 10 – 100 m. The majority of satellite SARs are unimodal (with the exception of SIR-C), operating at either C- or L-bands and acquiring like-polarized or quad-polarized data. Both ERS-1 and ERS-2 SAR have been used on a number of one-day repeat missions to acquire interferometric data, wherein a pair of images is acquired within a short time interval (1 day), and may be used to detect changes in topography (Baltzer, 2001). The advanced C-band satellites, including ENVISAT and RADARSAT have DEM generation capabilities through stereoscopic data acquisition.
433
Table 13.2 Sensor specifications: past and current spaceborne SAR sensors (Source: adapted from Baltzer, 2001). Satellite
Launch date
Freq (GHz)
Band
Polariz.
Resol. (m) 30
Swath width (km) 100
Look angle (deg.) 20
Incid. angle (deg.) 23-35
Repeat cycle (days) 3, 35, 168
ERS-1
June 1991
5.25
C
VV
JERS-1
1992
1.28
L
HH
18
75
35
38
44
SIR-C
April & Jan 1994
1.25, 5.3
L, C
Quad
10-40
15-90
15-55
17-60
1
ERS-2
1995
5.25
C
VV
30
100
20
23-35
1, 35
RADARSAT
Oct 1995
5.3
C
HH
10100
45500
20-59
17-50
2-3
ENVISAT
1998
5.25
C
HH, VV
30
50400
14-45
20-45
35
SRTM
2000
C, X
VV
30
48
52
57
11
Expt. Quad mode
10
5.3, 9.25
46 8-60 (standard) 18-43 250100 (scansar) 350 (Note: Radarsat repeat cycle is 24 days, but in wide swath mode is 1 day (Arctic) and 2-3 days (Equator). ALOS
2002
1.28
L
70
35
SAR data may be deliberately acquired at different frequencies and in different polarization states to interact with certain properties of the target feature, and so enhance their detection. Low frequency (L- and P-band), co-polarized SARs have been used successfully to detect flooding under vegetation (Hess et al., 1995; Krohn et al., 1983), discriminate between wetland types and determine above ground biomass and structure (Tanaka et al., 1998; Luckman et al., 1998), while higher frequency (C-band) SARs have been applied to mapping of herbaceous vegetation (Kasischke et al., 1997), deforestation (Hoekman & Quiñones, 2000), flood detection and mapping of swamps and wetlands (Tanis et al., 1994; Pope et al., 1997; Wang et al., 1995). Data over coastal ecosystems worldwide has already been acquired in a number of past missions and demonstrate the effective use of satellite SAR in tropical forest mapping. The integrated use of JERS-1 and ERS-1 (L- and C-band data respectively)
434
provided high accuracy mapping of tropical coastal habitat in Central Africa (Simard et al., 2002), while JERS-1 data were used by Milne & Dong (2002) to demonstrate the potential generation of regional or global scale biomass maps. ERS-1 data was used to map the spatial extent of wetlands in the Sundarban delta, India (Kushwaha et al., 2000; Dwivedi et al., 1999). C-HH RADARSAT data was also found to be a useful tool for monitoring the extent of inundation in forested wetlands of Northern Australia (Milne et al., 1998), the Amazon (Costa et al., 1997), and North Carolina, USA (Townsend, 1998). A study by Townsend (2002) concluded that Radarsat (C-HH) data was superior to ERS-1 and ERS-2 data (C-VV) in accurately mapping flooded forest. The use of satellite sensor data in mapping wetland vegetation and seasonal inundation is gaining increased interest with the predicted threats of sea level rise and greenhouse induced change.
The optimal parameters for monitoring of forest
characteristics in relation to wetlands are as follows (Table 13.3). Table 13.3 Optimal SAR sensor parameters for forest mapping (adapted from: Kasischke et al., 1997). Application
Frequency
Polarization
Incidence angle Low and high
Spatial Resolution Both (size dependent)
Sampling Rate Low
Suggested satellite SIR-C
Vegetation mapping
Multiple (opt)
Multiple
Biomass estimation
L or P (opt) L and C
HV or multiple
Low
Both
Low
JERS-1 SIR-C ALOS
Monitoring flooded forests
L or P (opt) C - some sensitivity
HH (opt) VV
Low
Both
High
JERS-1 SIR-C ALOS
ERS-1, 2 SIR-C RADARSAT ENVISAT (Note: Incidence angles – ‘low’ 20 – 40°, ‘high’ 40 – 60°; Resolution – Both implies ‘high’ 20 – 40 m for Monitoring low coastal wetlands
X or C
HH or VV
Low
Both
High
small areas, ‘low’ > 100 m for large areas; Sampling frequency – ‘high’ once every 2 weeks, ‘low’ once per year).
Satellite SAR platforms also provide global scale and repetitive interferometric data retrieval (Baltzer, 2001; Wegmüller et al., 1995) of relevance to forest biophysical
435
modeling (three dimensional structure), global carbon cycles and climate change effects. Fine scale DEM extraction from SAR data can be undertaken at low cost and in much less time than conventional stereo photography and for vast areas of the earth’s surface. Using interferometric techniques, DEMs can be generated over mixed terrain with the following levels of accuracy (Table 13.4). The atmospheric component and image coherence are main the limitations to interferometric DEM generation (due to temporal decorrelation of phase measurements). For successful and accurate DEM generation, specifically there should be minimal displacement and consistent radar-scattering mechanisms and response within a pixel in two successive images (Dixon, 1995: Basic Principles of SAR Interferometry). Interferometric SAR (InSAR) techniques are still in their infancy, but their application to tropical forest mapping and monitoring is of great significance. Table 13.4 Accuracy associated with interferometric DEM generation over mixed relief terrain (source: Toutin & Gray, 2000). Satellite ERS 1-2
Spatial Resolution (m) 24
DEM Accuracy (m) 3 – 20
Comments
JERS-1
18
10 – 20
L-band > coherence than C-band.
RADARSAT (standard)
20 – 29
10 – 20
Dry terrain preferred due to potential loss of coherence.
RADARSAT (fine)
7–9
3 – 10
Dry terrain preferred.
Lower accuracy assoc. with tropical forests & areas of dense vegetation or moisture variability.
Advanced Land Observation Satellite (ALOS) The Japanese ALOS instrument consists of a phased-array type of L-band SAR (PALSAR), which will also potentially provide data for repeat-pass interferometry (Baltzer, 2001). With the benefits of satellite technology and the capacity to acquire polarimetric and interferometric data, the ALOS system will undoubtedly increase the potential for global tropical forest and land cover mapping and biomass estimation. The ALOS instrument supports a Synthetic Aperture Radar (PALSAR), a high resolution stereo panchromatic imager (PRISM), and a medium resolution multispectral imager (AVNIR-2; Benner, 2001). The PALSAR instrument operates in standard (10 m spatial
436
resolution) or scansar (100 m) mode, acquiring single (HH or VV) or dual (HH/HV or VV/VH) polarization data. The PRISM instrument operates in the 0.52 – 0.77 µm wavelength range, with stereo mapping and DEM generation (3 – 5 m) capabilities. The AVNIR-2 instrument acquires visible to near infrared wavelength data in four image bands (between 0.42 – 0.89 µm).
Future SAR missions A number of future missions are still in the planning stages, including the X-band TerraSAR (a German-British collaborative project). The TerraSAR is a proposed high resolution (1 m) X-band radar system, capable of operating in single, dual and quad polarization modes, and with potential for along-track and repeat-pass interferometry (revisiting every 11 days) and multitemporal imaging (Arbinger, 2003: TerraSAR-X Flight Dynamics). Its applications are wide ranging, and include forestry, inventory update, topographic mapping and risk assessments. It is possible that the satellite may be used synergistically with ALOS (with the proposed TerraSAR-L), and with Radarsat or ASAR (with TerraSAR-X).
13.5 Potential for local to global scale monitoring The dynamic nature of the mangrove environment at the land-sea interface generates continual variation in both the physical and biological components that have a tendency to change both in the short (diurnal and seasonal) and long term (annually). In most cases, however, the rates of change within mangrove ecosystems are relatively slow and observations at five or ten-year time intervals may be adequate for monitoring purposes. Fine spatial resolution data acquired through airborne sensors are optimal for local to sub-regional mapping, whereas coarse level data using both airborne and satellite systems may be used on a regional to global scale. At the local to sub-regional level of mapping, where forest to stand level structural attributes and land cover differentiation is required, the following provides a suggested framework for inventory and monitoring with remote sensing data (Table 13.5). The integration of fine resolution datasets in the visible (e.g., stereo photography) and visible to near infrared wavelength range (e.g., airborne hyperspectral sensors) for land
437
cover mapping and change detection, analysis of mangrove structure and estimation of above ground biomass is advocated. SAR data (either airborne or satellite) should be acquired simultaneously for further analysis of 3D forest structure and biomass and also scattering response. The advantage of API is that fine resolution (1 m) DEMs of mangrove canopy height can be generated, of use for forest inventory, assessment of mangrove condition and for estimating biomass. Furthermore, large-scale API (1:25,000 – 50,000) with slightly higher resolution than some airborne hyperspectral scanners, provides better discrimination of tree crowns and hence density estimation in dense mangrove forest. Hyperspectral data is required for species discrimination, which would otherwise be limited using API (where spectral information is confined to the visible wavelengths). For analysis of mangrove forest structure and determination of above ground biomass, remote sensing data acquired on a 5-year basis should be adequate. The assessment of environmental change in mangrove environments often involves detection of fine-scale changes in estuarine or coastal morphology and species composition. Remote sensing data collected on a 5 to 10 year basis should include a hyperpsectral component as a reference for changes at the species level. Low frequency SAR data has the capacity to differentiate between inundated and non-inundated vegetation, and hence in tidally dominated or highly seasonal areas, the extent of wet/dry season flooding and saltwater movements can be mapped. To meet such an objective, data would need to be acquired on a monthly basis, whereas for the determination of vegetation structure and scattering response, annual acquisition would be sufficient.
The interpretation of SAR imagery generally requires either field
knowledge or vegetation descriptors retrieved from optical datasets. SAR backscatter models also provide an additional means of investigating the complex information contained within the data.
438
Table 13.5 Suggested framework for monitoring and inventory of mangrove ecosystems using remote sensing data. Sensor
Sensor specifications
Stereo photography
1:25,000 to 1:50,000 true colour imagery (0.3 – 0.4 m resolution)
Frequency of Acquisition Every 10 years
Application area
Derived products and accuracy levels
Land cover mapping Vegetation structure
Every 5 years
Biomass estimation
Orthomosaic (1 m) – map of extent DEMs (1 m) – tree height Density maps (upper canopy) Crown area maps (upper canopy) Spatial biomass map (t DM /ha) – total and component (leaf, branch, trunk, root) biomass Change detection between georeferenced datasets
Land cover monitoring CASI or other hyperspectral sensor
SAR airborne/satellite e.g., SIR-C, ALOS, AIRSAR JERS-1 (also for flood detection)
Every 5 years
Species mapping
Every 10 years
Land cover monitoring
High spatial resol. (10 – 40 m) L or P band HH or VV pol
Every month
Mapping flooded forest
Multifreq / Multipol
Annual
Land cover mapping
L or P band multipol or L & C band multipol
Annual to five years
Biomass estimation
1 m spatial imaging mode Wavelengths: visible to NIR (446 – 838 nm)
Supervised classification (high Acc. >70%) – species maps, boundaries between zones Change detection between georeferenced datasets Extent of inundation, dry/wet season boundaries Unsup/Superv. classification of tall or low, dense or open forest classes Spatial biomass map (t DM /ha) – total and component biomass
439
On a global scale, satellite sensor data from both optical and microwave sensors should be acquired on an annual or decadal basis to account for changing mangrove distribution and extent.
Spaceborne hyperspectral sensors, including Ikonos and
Quickbird, could potentially be used to retrieve data annually to map and monitor changes in species composition and community extent. Landsat and SPOT data could be used to map the broad vegetation patterns for comparison between sites, and provide regional estimates of vegetative cover through calculation of spectral indices. Satellite SAR data with fine to coarse resolution could be acquired on a monthly and annual basis to map the seasonal inundation patterns and vegetation cover and biomass respectively. SAR systems also show potential in mapping the spatial and temporal changes in soil salinity (Metternicht & Zinck, 2003), as a result of differential sensitivity of the real and imaginary parts of the dielectric constant to microwave radiation.
13.6 Concluding summary This chapter has summarized the reasons for generating baseline datasets of mangrove height, extent, species composition, structure and biomass using both optical and microwave remote sensing data. The datasets were then interpreted and a detailed look at the environmental changes occurring within the ARR and the short and longterm response of the mangroves was undertaken. The opportunities for spaceborne remote sensing were then outlined. The key findings were as follows: •
To assess both the short and long term changes in mangrove extent, structure, species/community composition and biomass, the generation of baseline datasets using a combination of optical and microwave remote sensing data is essential.
•
Stereo aerial photography can be used to generate baseline datasets of mangrove extent and of further benefit, is the capacity to generate DEMs of canopy height, for assessment of structure and condition, and also for biomass estimation.
•
Tree height is also a useful indicator of the long term movements associated with expanding communities.
In conjunction with visual data of the extent of
mangroves, DEMs of tree height can be used to identify areas of dieback, regeneration and growth, providing an added dimension to temporal change analysis.
440
•
Airborne hyperspectral data should be acquired to account for the mangrove species distribution, and can be used as reference in temporal change analysis with API. The response of mangroves to changing environmental gradients (e.g., salinity, inundation frequency and duration) is augmented through changes in species composition, and hence the information need be acquired by hyperspectral means.
•
Analysis of mangrove forest structure can be enhanced using airborne SAR and also backscatter models that simulate the scattering response and help identify the main scattering mechanisms in radar-forest interactions.
Through empirical
relationships, tree structure, notably height, was found to be a major determining factor of the scattering response from the West Alligator mangroves. Tree height was then related to biomass, and it was possible to generate a coarse level dataset of total biomass using the SAR data. •
Through the generation of baseline datasets and change detection using optical data, several key processes of change within the ARR were identified. Significant movements within the coastal and estuarine mangrove communities had occurred over the past half century, and largely in response to coastal environmental change.
Widespread phenomena included the progradation and erosion of
coastlines, expansion of the tidal creek network and subsequent saltwater intrusion, the spread of mangroves along creeklines, species compositional change (within and between zones) and channel morphological changes. •
The expansion and erosion of coastal mangroves on alternate banks of the West Alligator River was attributed to changing patterns of tidal flows within Van Diemen Gulf, and variable patterns in erosion and sedimentation as a consequence bi-directional currents and long-shore deposition.
The long-term effect of a
gradual sea level rise over the past half-century was also identified as a causal factor for these processes and also for the expansion of the tidal creek network throughout the estuarine reaches of the ARR systems. •
With the gradual progradation of the western bank of the West Alligator and subsequent coastal stability and wave dampening, a new zone of young S.alba was able to establish in front of its seaward parent zone. The expansion of the mangrove community led to the substantial seaward advance of the mature Rhizophora zone and changing position of the boundary between Rhizophora and 441
the landward Avicennia forest. These movements were related to the changing environmental conditions that were incurred as a result of reduced tidal extent and tidal creek penetration into the upper reaches of the coastal mangroves. •
The highly salt tolerant mangrove A.marina was identified as the predominant species spreading along the invading tidal creeks of the mud flats. Long-term growth of the species in these areas was indicated through the larger canopy size and replacement of other species (largely Rhizophora) on the landward edge.
•
Spaceborne remote sensing using hyperspectral, multispectral and SAR satellite platforms has the capacity for coarse resolution mapping of mangrove ecosystems worldwide.
These systems have the benefit of increased temporal data
acquisition, which may be advantageous when assessing the predicted greenhouse-induced effects on mangroves. The rates of change within mangrove ecosystems are likely to accelerate as a result of climate change, and it is suggested that the optimal means of monitoring such change will involve optical airborne remote sensing (API and hyperspectral imagers) and low frequency, multipolarized satellite SAR.
442
CHAPTER FOURTEEN
CONCLUDING REMARKS.
Focusing on mangroves of the ARR, Kakadu NP, and specifically the West Alligator River, this thesis has demonstrated the use of airborne remote sensing data for generating baseline datasets of mangrove extent, height, density, species/community composition and both total and component biomass against which to assess change. The limitations associated with each approach have been presented and, where possible, alternative solutions have been provided. From individual baselines and time-series comparisons of baselines (specifically mangrove extent and also height), a significant and continuing response of mangrove within the ARR to coastal environmental change has been observed, which may partly be the result of rising sea levels induced by global climate change. In the discussion, future options for monitoring and assessing this change were presented. This chapter now presents the conclusions of the project and recommendations for further research focusing on a) remote sensing for baseline inventory of mangroves, b) the ARR and mangrove responses to change and c) future considerations for research and observation.
14.1 Remote sensing for baseline inventory of mangroves Until recently, the capacity to generate detailed baseline datasets of, for example, the height and species/community composition of mangroves, has been limited as the available remote sensing data were generally of medium to coarse spatial resolution and of limited spectral resolution. Although finer spatial resolution aerial photographs were available for many sites, the benefits of using these data for mangrove characterization and change detection have barely been recognized. Furthermore, most analyses of these
443
data
have
relied
upon
manual
interpretation
by
skilled
interpreters
and
photogrammetrists. However, with the considerable advances in remote sensing in the past decade, enormous potential now exists for generating the baseline data required for understanding and monitoring the dynamics of mangroves. Specifically, remote sensing data are now able to be acquired routinely at fine to coarse spatial resolutions, in broad to narrow (hyperspectral) spectral regions from the visible through to the microwave, and by sensors that can observe in both two and three dimensions (e.g., LIDAR and SAR). Advances in image processing and computer storage and processing capacity also allow these datasets to be fully and routinely integrated and analyzed.
For
specialist fields such as photogrammetry, new tools for digital processing and product generation have become available. During this time of active development, rapid changes in mangroves have also continued to occur worldwide as a result of natural (e.g., sedimentation) and both direct (e.g., clearance) and indirect (e.g., sea level rise and climate change) anthropogenic processes. The requirement to understand, map and monitor these changes has become increasingly important.
This study has addressed and considered the advances in
remote sensing data and processing technologies, by demonstrating their use in generating baseline datasets of key mangrove attributes, and also the changing dynamics of mangroves within an area vulnerable to change, the ARR of northern Australia. In the following sections, the conclusions arising from the coupling of the baseline datasets with the dynamics of mangroves are presented. 14.1.1 Mangrove extent As anticipated, the extent of mangroves was able to be mapped to a high level of accuracy using both the 1 m spatial resolution orthomosaics and hypespectral CASI data. Although the extent of mangroves could also be mapped using the 10 m spatial resolution AIRSAR data, the same level of detail could not be resolved. The actual extent of mangroves could be mapped (using API and CASI) through standard density slicing or standard supervised or unsupervised classification routines. The level of detail able to be mapped was, however, far superior to that provided by moderate spatial
444
resolution sensors (e.g., Landsat and SPOT). A further benefit was that, as in this study, the techniques could be applied to historical datasets thereby providing an opportunity to quantify change. 14.1.2 Mangrove height The ability to retrieve the height of mangroves from stereo aerial photography was first piloted by Lucas et al. (2002), and the research was part of this study. As an extension to this work, this study has demonstrated the application of the technique to mangroves throughout Kakadu NP. However, the study also revealed two limitations of the technique. First, heights were less successfully retrieved where sparse or relatively open canopies existed, such as was the case with the seaward communities of S.alba. Second, differences in the elevation of the mudflats underlying the mangrove canopies were not compensated for, which led to overestimation of mangrove height, such as on the west bank where the height of A.marina communities was exaggerated by several metres. Compensating for these height differences is difficult given the lack of spatial data relating to the height of the underlying surface, particularly as the photographs used were historical. Future campaigns of aerial photography over the mangroves of the ARR should therefore be undertaken in conjunction with either LIDAR or interferometric SAR so that the true height of the surface can be retrieved and the information content of (and errors within) the DEMs derived from photography can be better understood. 14.1.3 Tree density Stem density was best retrieved using the true colour stereo aerial photography and a tree top delineation and counting algorithm developed using eCognition. However, this approach necessarily requires adjustments (based on field data) for the density of trees in the understorey and sub canopy which are not observed by optical sensors. The estimates of density could, however, be bettered by establishing the heights (from the DEM) of trees located using eCognition, thereby allowing a refinement of density adjustment factors within specific height classes. For mangroves where more than one tree occupies the space of a pixel, including areas of regrowth with several thousand stems per hectare, the use of tree counting algorithms was considered inappropriate. However, this research suggested that empirical relationships could be established
445
between the NDVI and tree density for generating baseline datasets. The field data collected in this study were insufficient to generate the required relationship, but this was identified as a relatively simple topic to address in future research. The alternative approach to estimating tree density from SAR was limited by the lack of a strong empirical relationship at any frequency or polarization. 14.1.4 Species/community composition The use of true colour stereo aerial photography for discriminating mangrove species and communities is limited, largely by the low spectral range and radiometric quality of the data. For the ARR, only areas of mangroves dominated by A.marina were able to be distinguished from all other mangrove communities using photography, although some confusion with S.alba was still evident. Through this research, the requirement of hyperspectral data for discriminating species and communities was recognized and resulted in the successful acquisition of CASI data in July 2002 over the West Alligator River. These data illustrated the enormous benefits of acquiring data in the visible and NIR region for species/community discrimination.
Specifically, using these data,
mangroves dominated by A.marina were able to be distinguished from those dominated by S.alba and R.stylosa. Even so, S.alba (although only regrowth) and R.stylosa were spectrally confused.
Field spectra acquired during the field campaign of 2002
suggested, however, that greater separation of these species could be achieved if the SWIR reflectance data were used. This agrees with other studies (e.g., Green et al., 1998) which confirm that the inclusion of the SWIR channel in the remote sensing dataset significantly improves the classification of mangrove zones. The application of a supervised maximum likelihood algorithm to the CASI data provided the best discrimination of mangrove species and communities. However, a limitation of the approach was that image spectra were only able to be extracted for the dominant species within the communities, as rare or poorly represented species could not be identified within the image. Field spectra were also only able to be extracted for five species. Discrimination might not therefore be so straightforward where a greater diversity of species dominates.
These limitations lend support to the concept of
generating spectral libraries for mangrove species, not only for those in Australia but worldwide. These libraries could be generated from field spectroradiometer data or
446
from fully and consistently calibrated imagery (e.g., CASI). Ideally, both field and image (e.g., HYMAP) spectra from the SWIR regions should be included. When collecting spectra, the background should also be considered given its known impact on the spectral reflectance of mangroves (Meza Diaz & Blackburn, 2003). 14.1.5 Total and component biomass The study considered two approaches to the estimation of total and component biomass. The first involved the integration of spatial datasets relating to mangrove canopy height and density (as derived from stereo aerial photography) and species/community composition with species specific allometric equations for estimating the total and component biomass of mangroves.
Although providing
reasonable estimates of biomass amount and distribution, the lack of accurate tree density estimates, errors in the estimation of height and the use of allometric equations beyond the range for which they were intended led to low levels of accuracy in all communities. These errors could be overcome by better mapping of the surface terrain (see section 14.1.2), providing better adjustments for tree density estimates (see section 14.1.3) and harvesting more individuals, particularly at the higher end of the size range. By minimizing these errors, it is anticipated that a higher and more acceptable level of accuracy in the estimation of total and component biomass can be obtained. The alternative approach of using polarimetric AIRSAR data to estimate the total and component biomass of the mangroves based on empirical relationships was also considered. Relationships with the logarithm of total above ground biomass and the backscattering coefficients were generally stronger at the lower frequency L and P band, but P-HV was considered optimal due partly to the larger dynamic range of the data. However, above the level of saturation (~180 t DM ha-1), estimates of biomass could not be retrieved.
Furthermore, within the higher biomass R.stylosa mangroves, SAR
backscatter at all frequencies and polarizations was reduced, which was attributable to differences in the leaf size, density and orientation (at C-band) and to the disruption of the double bounce effect by the prop roots (at lower frequencies), as suggested by Simard et al. (2002) and Held et al. (2003). Even so, using AIRSAR P-HV data acquired over the mangroves of the ARR, a spatial dataset of total above ground
447
biomass was generated, which demonstrated a reasonable correspondence with ground measurements. Although the spatial datasets relating to mangrove height, density and species/community composition provided an insight into the SAR response of mangroves, an investigation into the behaviour of microwaves using SAR backscatter models (e.g., Proisy et al., 2000) is needed to advance the retrieval of biomass. As an example, the model of Karam et al. (1995) has been developed with wide ranging applicability to mangrove and other multi-layered forest mediums. Using the datasets generated during this project, future work will concentrate on modelling as part of a future collaborative research with scientists in French Guiana. With the assistance of the model, the influence of mangrove forest structure and biomass on scattering mechanisms and polarization response will be investigated. Further comparative studies between French Guiana and Northern Australian mangroves will be undertaken using the results of the simulations, and new information relating to SAR backscatter modelling and the influence of structure on microwave behaviour released through publications in the scientific literature. A particular advantage of using SAR data is that estimates of biomass (total and/or component) can potentially be generated over large areas, particularly as spaceborne SAR data are becoming increasingly available. Comparison with studies undertaken in French Guiana (Proisy et al., 2000) demonstrated that the SAR response from mangroves is related reasonably consistently to structural parameters, including biomass, although variability in some relationships was observed. Further analysis of airborne and spaceborne SAR data from sites additional to northern Australia (e.g., the mangroves of the Daintree, Queensland) should also be undertaken if a regional or global algorithm is to be generated.
14.2 The ARR and mangrove response to change One of the major benefits of this research has been the generation of DEMs of mangrove canopy height and orthomosaics from which mangrove extent can be mapped. Changes in the extent of mangroves and the communities contained have also
448
been observed. The following sections provide an overview of the products generated and their use in assessing change. 14.2.1 DEMS and orthomosaics The DEMs and orthomosaics for the ARR, generated through this research, arguably represent one of the most detailed and comprehensive available for mangroves worldwide. Orthomosaics were generated for approximately 750 km2 of the ARR, whilst DEMs of canopy height were generated for approximately 86 km of coastline. The majority of mangroves have been mapped along the Wildman River, the West, South and East Alligator Rivers and on Field and Barron Islands. These orthomosaics allow large scale mapping of mangrove extent and associated features, both of which are critical for the detection of change. These datasets have been made available to scientists and land managers charged with understanding the dynamics of the coastal environment in the ARR and monitoring future changes. 14.2.2 Changes within the ARR By simply viewing the DEMs and orthomosaics themselves, considerable insight into the dynamics of mangroves and their response to change has been provided. Specifically, abrupt height boundaries on the seaward margins have indicated areas of erosion, which were particularly noticeable near the mouth of the Wildman River. Areas protected from erosion (e.g., on Barron and Field Island) were also noticeable by their more coherent structure and greater height. Areas of lower or gradually decreasing height observed within the DEM were associated with areas of regeneration or where the growth of mangroves was restricted. In the tidal creeks, the existence of high mangroves confirmed longer term colonization in areas that are experiencing saltwater intrusion. Freshwater environments affected by saltwater intrusion (e.g., Paperbark swamps) were also evident within the orthomosaics. Comparing maps of mangrove extent over time provided an even greater insight into the temporal dynamics of mangroves. For the West Alligator River, the comparison of mangrove extent between 1950 and 1991 revealed an expansion of mangroves on the west bank facing the coast and an erosion of mangroves on the east bank. Colonization by mangroves of the upper reaches of tidal creeks on both banks was also
449
evident from the comparison, as was the formation and loss of estuarine features (e.g., creek bends and inlets) and their associated mangrove communities. Comparison of the extent of mangroves between 1991 and 2002 revealed a further seaward expansion on the west bank and erosion on the east bank. Expansion of mangroves on the landward margins was also evident and an acceleration of upstream colonization of tidal creeks by mangroves was apparent, particularly away from the river mouth. The integration of species/community maps generated using hyperspectral CASI data provided even further information on the dynamics of mangroves.
These maps
confirmed that the mangroves expanding on the west bank and retracting on the east bank were dominated by S.alba. The central zone of high mangroves was dominated by R.stylosa and the intermediate zone contained a mix of R.stylosa, A.marina and S.alba. Within creeks experiencing saltwater intrusion, A.marina was observed to be one of the first species to colonize. However, in some creeks, R.stylosa was establishing and replacing the A.marina communities, suggesting a continuation of the process.
A
noticeable increase in the crown area of A.marina was also evident throughout the mangroves of the West Alligator and the number of gaps in the canopy of mature R.stylosa (presumably from lightning strikes) was also greater compared to 1991. An unexpected finding was that the edge of the high R.stylosa zone had actually been eroded by 50 – 150 m, resulting in a significant loss of biomass. These observations of the tidal creeks and the R.stylosa zone, in particular, confirmed that changes in species/community distributions may occur, even though a change in mangrove extent may not. This observation therefore further highlighted the need to generate additional baseline datasets of species and communities. The reasons for the observed changes include an alteration in the coastal environment and particularly in fluvial processes and tidal inundation.
However,
satellite-based (i.e., TOPEX POSEIDON) measurements suggest a significant rise (~1.5 cm over 10 years) in sea level in the north of Australia and through South East Asia in general. Whether this rise is permanent or just temporary (e.g., as a result of ENSO related phenomena) is uncertain.
Establishing whether the observed change in
mangroves is a result of their response to rising sea levels or to naturally occurring events (e.g., changes in flooding patterns etc.) is also an issue of contention which can
450
only be resolved through further observation and research. Certainly, the continued establishment of baseline datasets for the ARR is advocated, particularly as this region is vulnerable to sea level rise and also changes in climate. A further option for understanding the response of mangroves to environmental and climate change is to combine the information on mangrove dynamics, as observed using remote sensing data, with meteorological data and other information relating to tidal activity or coastal processes that affect the mangrove environment. Building on this information, spatial models can potentially be generated that simulate, for example, the transfer and accumulation of sediments, the rises in sea level (e.g., from TOPEX POSEIDON) and the influence of flooding on varying spatial and temporal scales. Many studies have already published on the impacts of greenhouse-induced change on coastal ecosystems and this information provides vital input to the modelling. Through this process, the response and redistribution of mangroves of different species can be simulated. Such models that simulate the future extent and structure of mangrove communities under scenarios of climate change would seem to be the next logical step in evaluating the future viability of these ecosystems worldwide.
Basic models
employing topological data are currently available but are conceptually simple at this stage. By integrating remote sensing datasets and 3D visualizations, a powerful tool for simulating and interpreting future changes could be generated.
14.3 The requirement for baseline datasets The mangroves of the ARR represent a fraction of those worldwide and yet the changes that are occurring are significant. In many ways, it could be argued that the mangroves are actually benefiting from the changes in the coastal environment, which may or may not be the result of rising sea level or global climate change. However, it is not so much the mangroves that are under threat but rather the proximal coastal environments, notably the World Heritage Listed wetlands of Kakadu NP. If such events are occurring in the ARR, it is presumed that they are also occurring elsewhere in the world but have not been recognized. This may, in part, be the result of a lack of records (in the form of maps or remotely sensed data) for these mangroves or
451
the fact that large areas of mangroves are being disturbed and the impacts are therefore being masked. This study has therefore highlighted that mangroves are responding to change in the coastal environment and has also demonstrated how such changes can be assessed, mapped and monitored. The study further emphasizes the need for development of baselines elsewhere. Specifically, our current knowledge of the extent and condition of the World’s remaining mangrove forest is scattered, with inconsistent records having been derived through many sources and at varying spatial and temporal scales. Baseline datasets are rarely available for many mangrove areas and there is little inventory data to support monitoring studies. Even so, in many of these areas, aerial photography exists and the potential for establishing historical baseline datasets (as for the ARR) is enormous. The initiative needs to be taken, however, to undertake the establishment of baseline datasets although it is vital that this initiative is coordinated, both within countries and internationally.
Specifically, international research
organizations, universities and government departments need to foster a collaborative approach to mangrove inventory and assessment and collate and organize data and data products in a way that makes it accessible by the scientific and public community in a format that is user-friendly.
Remote sensing and ancillary data acquired using
standardized techniques and used for the establishment of baselines should also be databased within a GIS and updated on a timeframe that depends upon user requirements. A global survey of key tropical, subtropical and temperate mangroves could also potentially be undertaken using satellite sensor data to assess the past and current status of these ecosystems worldwide. Key candidates include the global mosaics of JERS-1 SAR data (within the period 1992-1998), SRTM (2002) and the forthcoming ALOS PALSAR. A particular advantage of these SAR data and the derived datasets is that they are unaffected by cloud-cover. Therefore, global inventory over a short period (e.g., a year) is possible. The main limitation of these datasets, however, is that the spatial resolution may be too coarse (50 – 100 m) and therefore, for key regions, finer spatial resolution optical sensor (e.g., Landsat ETM+ or SPOT HRVIR) data should be supplemented.
Optical datasets are particularly useful for mapping mangrove
452
communities, and the global mosaic (http://www.landsat.org/) of Landsat sensor data represents a good source of consistently generated data. Even so, for a number of mangrove ‘supersites’ vulnerable to change, fine spatial resolution baseline datasets of the key mangrove attributes of extent, height, species/community composition and biomass need to be generated (using remote sensing data that is existing or acquired specifically for that purpose) and updated on a regular basis. Sites that are relatively unaffected by anthropogenic activity and sites that are actively disturbed should be selected. Mangroves recognized as contributing significantly to global biodiversity or regarded to be of important biological and/or economic potential should also be surveyed. Only through this procedure can a true understanding and overview of the response of mangroves to the changing coastal environment and climate be obtained.
14.4 A final thought Throughout the world, mangroves are regarded as peripheral ecosystems, partly because of their inaccessibility, or a general lack of understanding of their functionality. Yet, these mangroves are critical to the global ecosystem – they harbour a wealth of plant and animal species and are important centres of biodiversity, provide the basis for many inshore and also offshore fisheries, contain significant quantities of carbon within the biomass, and protect the coastline and proximal ecosystems from damage by storms. These factors alone provide justification for their preservation and the development and implementation of a global inventory and monitoring system, particularly as extensive tracts of mangrove have been cleared in the past century and continue to be degraded in the present. The establishment of a global inventory and monitoring system is also essential for a more critical reason. Mangroves are important barometers of the greenhouse effect, particularly given their sensitivity to changing climate and rising sea levels, a problem that is likely to be exacerbated over this century. Particularly when left undisturbed, they represent one of the few ecosystems able to provide humans with an early warning of trends that might ultimately lead to long-term and perhaps irreparable damage to the global environment, either in part or as a whole.
453
BIBLIOGRAPHY
Alongi, D.M. 2002. Present state and future of the world’s mangrove forests. Environmental Conservation, 29, 3: 331-349. Allen, J.A., Ewel, K.C. and Jack, J. 2001. Patterns of natural and anthropogenic disturbance of the mangroves on the Pacific island of Kosrae. Wetl. Ecol. & Mngt., 9: 279-289. Arbinger, C. Page: TerraSAR-X Flight Dynamics. 8/11/2003. Weblab, Germany. Accessed: 22/11/03. http://www.weblab.dlr.de/rbrt/FlightDyn/FutureProjects/TSX/TSX _FD.html Aschbacher, J., Ofren, R.S., Delsol, J. P., Suselo, T. B., Vibulsresth, S. and Charrupat, T. 1995. An integrated comparative approach to mangrove vegetation mapping
using
remote
sensing
and
GIS
technologies:
preliminary
results.
Hydrobiologia, 295: 285-294. Aspinall, R., Miller, D., Finch, P and Fon, T.C. 1994. A model of DEM and orthophotograph quality using aerial photograph. GIS/LIS, p.24-33. Attema, E.P.W. and Ulaby, F.T. 1978. Vegetation modeled as a water cloud. Radio Sci., 13: 357-364. Aubé, M. and Caron, L. 2001. The mangroves of the North coast of Haiti. Wetl. Ecol. & Mngt., 9: 271-278. Augustinus, P.G.E.F. 1995. Geomorphology and sedimentology of mangroves. In Perillo, G.M.E (ed.), Geomorphology and sedimentology of estuaries. Elsevier Science, Netherlands. Avery, T.E. and Berlin, G.L. 1992. Fundamentals of remote sensing and airphoto interpretation. Macmillan, New York. Baatz, M., Benz, U., Dehghani, S., Heynen, M., Holtje, A., Hofman, P., Lingenfelder, I., Mimier, M., Sohlbachj, M., Weber, M. and Willhauck, G. 2001. eCognition:
Object Oriented Image Analysis.
User Guide,
Definiens Imaging,
Munich, Germany. Ball, M.C. and Pidsley, S.M. 1995. Growth responses to salinity in relation to distribution of two mangrove species, Sonneratia alba and S.lanceolata, in northern Australia. Functional Ecology, 9: 77-85.
454
Ball, M.C. 1988. Ecophysiology of mangroves. Trees, 2: 129-142. Ball, M.C. 1988. Organisation of mangrove forests along natural salinity gradients in the Northern Territory: an ecophysiological perspective. In: Marshall, D.W. and Loveday, P. (eds). Northern Australia: Progress and Prospects. Floodplains Research. Australian National University. North Australian Research Unit, Darwin. Baltzer, H. 2001. Forest mapping and monitoring with interferometric synthetic aperture radar (InSAR). Progress in Physical Geography, 25, 2: 159-177. Band, L.E. 1993. Extraction of channel networks and topographic parameters from digital elevation data.
In: Beven, K. and Kirkby, M.J. (eds.), Channel Network
Hydrology. John Wiley & Sons, USA. Bandaranayake, W.M.
1994.
Phyto-chemical constituents and pigments in
mangrove species and mangal associates of Northern Australia. Australian Institute of Marine Science (AIMS) Report, 19: 1-21. AIMS, Townsville, Australia. Bayliss, B., Brennan, K., Eliot, I., Finlayson, M., Hall, R., House, T., Pidgeon, B., Walden, D. and Waterman, P. 1997. Vulnerability assessment of predicted climate change and sea level rise in the Alligator Rivers Region, N.T., Australia. Supervising Scientist Report No. 123. Supervising Scientist, Environment Australia. Canberra, Australia. Beaudoin, A., Le Toan, T., Goze, S., Nezry, E., Lopes E. and Mougin, E. 1994. Retrieval of forest biomass from SAR data. Int. J. Remote Sens., 15, 14: 2777-2796. Benner, D. 2001. Advanced Land Observing Satellite (ALOS). US Government SAR Symposium, March 28-29, 2001, Canadian Embassy, Washington DC. Bennett, K.D. and Willis, J. 2000. Effect of global atmospheric carbon dioxide on glacial-interglacial vegetation change. Global Ecol. & Biogeog., 9: 355-361. Biña, R.T., Jara, R.B., and Roque, C.R. 1980. Application of multi-level remote sensing survey to mangrove forest resource management in the Philippines. Proceedings of the Asian Symposium on Mangrove Development, Research and Management, University of Malaya, Kuala Lumpar, Malaysia, 28-29 August. Bird, E.C.F. 1993. Submerging coasts – the effects of a rising sea level on coastal environments. John Wiley & Sons, England. Blasco, F., Aizpuru, M. and Gers, C.
2001.
Depletion of the mangroves of
Continental Asia. Wetl. Ecol. & Mngt., 9: 245-256.
455
Blasco, F., Gauquelin, T., Rasolofoharinoro, M., Denis, J., Aizpuru, M. and Caldairou, V. 1998. Recent advanced in mangrove studies using remote sensing data. Mar. Freshw. Res., 49: 287-96. Blasco, F., Saenger, P. and Janodet, E. 1996. Mangroves as indicators of coastal change. Catena, 27: 167-178. Blasco, F., Lavenu, F., and Baraza, J. mangroves of Kenya coast.
1986.
Remote sensing data applied to
Proceedings of the 20th International Symposium on
Remote Sensing of the Environment, vol. 3, pp. 1465-1480. Brennan, K. 1996. An annotated checklist of the vascular plants of the Alligator Rivers Region, NT, Australia. Supervising Scientist Report, No. 109. Supervising Scientist, Environment Australia, Canberra, Australia. Brocklehurst, P., Edmeades, B. and Van Kerckhof, D. 1995. Mangrove survey of Darwin Harbour, Northern Territory.
Technical Report No. 96/9.
Department of
Lands, Planning and Environment, Northern Territory. Bucher, D. and Saenger, P. 1994. A classification of tropical and subtropical Australian estuaries. Aquatic Conservation, 4:1-19. Bureau of Meteorology. 2003. Australian climate variability and change. BOM, Commonwealth of Australia. Accessed: 16/11/03. http://www.bom.gov.au/climate/cha nge/auscc.shtml Bureau of Meteorology. 2003. Climate averages. BOM, Commonwealth Australia. Accessed: 23/10/03. http://www.bom.gov.au/climate/averages/ Bureau of Meteorology. 1999. Hydrometeorological analyses relevant to Jabiluka. Supervising Scientist Report, No. 140. Supervising Scientist, Environment Australia, Canberra, Australia. Burrough, P.A. 1986. Principles of Geographical Information Systems for land resources assessment. Clarendon Press, Oxford, New York, USA. Butterworth, I. projected.
A review of Northern Territory wetland rainfall: current and
In: Finlayson, C.M. (ed.), Wetland research in the wet-dry tropics of
Australia. Supervising Scientist Report, No. 101. Commonwealth of Australia. Campbell, J. 1987. Introduction to Remote Sensing., pp. 384-85. Guilford Press, New York. Chapman, V.J. (ed.). 1977. Wet coastal ecosystems. In: Ecosystems of the World, vol 1. Elsevier Scientific Pub. Co., New York.
456
Chappell, J. 1990. Some effects of sea level rise on riverine and coastal lowlands. Geol. Soc. Aust. Symp. Proc., 1: 37-49. Chappell, J. 1988. In: Marshall, D.W. and Loveday, P. (eds). Northern Australia: Progress and Prospects. Floodplains Research. Australian National University. North Australia Research Unit, Darwin. Chaudhury, M.U. 1990. Digital analysis of remote sensing data for monitoring the ecological status of the mangrove forests of Sunderbans in Bangladesh. Proceedings of the 23rd International Symposium on Remote Sensing of the Environment, vol. 1, pp. 493-497. Chauvaud, S., Bouchon, C. and Manière, R. 1998. Remote sensing techniques adapted to high resolution mapping of tropical coastal marine ecosystems (coral reefs, seagrass beds and mangrove). Int. J. Remote Sensing., 19, no. 18: 3625-3639. Christian, C.S. and Aldrick, J.M. 1977. Alligator Rivers study: A Review Report of the Alligator Rivers Region environmental fact-finding study. Aust. Gov. Pub. Service, Canberra, Australia. Clark, R.L. and Guppy, J.C. 1988. A transition from mangrove forest to freshwater wetland in the monsoon tropics of Australia. Journal of Biogeography, 15: 665-684. Clough, B.F., Dixon, P. and Dalhaus, O.
1997.
Allometric relationships for
estimating biomass in multi-stemmed mangrove trees. Aust. J. Bot.,45: 1023-1031. Coates, B. and Hanslow, D. 1998. Monitoring coastal wetland ecosystems. In: Brady, A. (ed.), Wetlands – Conference & Workshop. Proceedings of DLWC Wetlands Conference & Workshop, held 28-30 April, 1997, Myall Shores, Myall Lakes N.P. DLWC Centre for Nat. Res. Ecosyst. Managment Branch, Parramatta, Australia. Cobb, S.M., Saynor, M.J., Eliot, I. and Hall, R. 2000. Saltwater intrusion in the Alligator Rivers Region, northern Australia. In: Eliot, I., Saynor, M., Eliot, M. and Finlayson, M. (eds). Assessment and monitoring of coastal change in the Alligators Rivers Region, Northern Australia. Supervising Scientist Report, 157, Supervising Scientist, Canberra, Australia, 118-131. Comley, B.T.W. 2002. Coastal mangrove wetland productivity and management, Darwin Harbour, Northern Territory, Australia.
Unpublished Masters Thesis, NT
University, Darwin. Costa, M.P.F., Nova, E.M.L.M., Mitsuo, II.F., Mantovani, J.E., Ballester, M.V. and Ahern, F.J.
1997.
Classification of floodplain habitats (Lago Grande, Brazilian
457
Amazon) with RADARSAT and JERS-1 data. In Ahern, F.J. (ed.), RADARSAT for Amazonia: Results of ProRADAR Investigations. Ottawa, Canada Centre for Remote Sensing, Natural Resources Canada, pp. 149-161. Cronin, N.L.R., Lucas, R.M., Witte, C. and Milne, A.K. 2001. Synthetic Aperture Radar for AGB estimation in Australia’s woodlands. Proc. IGRARSS Conference, held in Sydney, Australia July, 2001. Crowley, G.M. 1996. Late Quaternary mangrove distribution in Northern Australia. Australian Systematic Botany, 9: 219-225. Culvenor, D.S. 2000. TIDA: an algorithm for the delineation of tree crowns in high spatial resolution remotely sensed imagery. Computers and Geosciences, 28: 33-44. Danaher, K. and Luck, P. 1991. Mapping mangrove communities using Landsat Thematic Mapper Imagery.
Networking for Sustainability, Proceedings of the
Conference on Remote Sensing and GIS for Coastal Catchment Management, Centre for Coastal Management, University of Lismore, NSW. Danaher, K. 1994. Mapping mangrove communities in Cape York Peninsula using Landsat TM data. Proceedings of the 7th Australasian Remote Sensing Conference, March 1994, Melbourne, Australia. Danson, F.M. and Plummer, S.E. (eds.). 1995. Advances in Environmental Remote Sensing. John Wiley & Sons, Ltd. Davie, J.D.S.
1984.
The mangrove vegetation of the South Alligator River,
Northern Australia. In: Bardsley, K.N., Davie, J.D.S. and Woodroffe, C.D. (eds). Coasts and tidal wetlands of the Australian monsoon region. A collection of papers presented at a Conference held in Darwin, 4-11 Novermber, 1984. Australian National University. North Australian Research Unit, Mangrove Monograph No. 1. Darwin, Australia. Davis, J.A. and Froend, R. 1999. Loss and degradation of wetlands in southwestern Australia: underlying causes, consequences and solutions. Wetl. Ecol. & Mngt., 7: 1323. Daya Sagar, B.S., Venu, M. and Srinivas, D. 2000. Morphological operators to extract channel networks from digital elevation models. Int. J. Remote Sensing, 21, 1: 21-29. Dennis, K.C. and Nicholls, R.J. 1995. Senegal: Senegal’s shoreline is already receding and salinity threatens wetlands. Climate Alert, vol. 8, no. 2.
458
Department of the Environment and Heritage. Pages: Introduction to the Alligator Rivers Region; Alligator Rivers Region; Extent and distribution of wetlands in northern Australia; Uranium mining in the ARR – Intro; KNP World heritage values; Australia’s Ramsar sites. 3-19/9/03. Supervising Scientist Division, Commonwealth Australia. Accessed: 23/10/03. http://www.deh.gov.au/ssd/about/arr/index.html Department of the Environment and Heritage. Pages: Welcome to Kakadu National Park; Kakadu NP; KNP visitor survey results 2000-2001; Kakadu image gallery; Animals of the Kakadu region; Management of feral animals; Plants of the Kakadu region; Geology of the Kakadu region; Landforms of the Kakadu region. 23/9/03. Department of the Environment and Heritage, Commonwealth Australia. Accessed: 23/9/03. http:// www.ea.gov.au/parks/kakadu/index.html Dial, G. and Grodecki, J. 2003. Applications of IKONOS Imagery. Proceedings of ASPRS 2003 Conference, Anchorage, Alaska, May 5-9, 2003. Digby, M.J., Saenger, P., Whelan, M.B., McConchie, D., Eyre, B., Holmes, N. and Bucher, D. 1998. A physical classification of Australian estuaries. Report prepared for the Urban Water Research Association of Australia by the Centre for Coastal Management, Southern Cross Univervsity, Lismore, NSW. Report No. 9, LWRRDC Occasional Paper 16/99. Dixon, T. Page: Basic Principles of SAR Interferometry. Report of a Workshop held in Boulder, CA, Feb 3-4, 1994. July, 1995. University of Miami. Accessed: 23/11/03. http://southport.jpl.nasa.gov/scienceapps/dixon/report2.html Dobson, M.C., Ulaby, F.T. and Pierce, L.E. 1995. Land-cover classification and estimation of terrain attributes using SAR. Remote Sens. Environ., 51: 199-214. Dong, Y., Forster, B.C. and Ticehurst, C. 1998. A new decomposition of radar polarization signatures. IEEE Trans. Geosc. & Remote Sens., 36, 3: 933-939. Donoghue, D.N.M. 2000. Remote sensing: sensors and applications. Progress in Physical Geography., 24, 3: 407-414. Doyle, T. Page name: Effects of hydrology on red mangrove recruits. March, 2003. US Geological Survey. Accessed: 9/6/03. www.nwrc.usgs.gov/factshts/RedMangrove/ FactSheet.pdf Drake, J. B., Dubayah, R.O., Clark, D.B., Knox, R.G., Blair, J.B., Hofton, M.A., Chazdon, R.L., Weishampel, J.F. and Prince, S. 2001. Estimation of tropical forest
459
structural characteristics using large-footprint lidar. Remote Sensing of Environment (in press). Duke, N.C, Ball, M. and Ellison, J.C. 1998. Factors influencing biodiversity of and distributional gradients in mangroves. Global Ecol. Biogeogr. Letts. 7: 27-47. Durden, S.L., Van Zyl, J.J. and Zebker, H.A. 1989. Modeling and observation of the radar polarization signature of forested areas. IEEE Trans. Geosc. & Remote Sens., 27, 3: 290-301. Dutrieux, E., Denis, J., and Populus, J. 1990. Application of SPOT data to a baseline ecological study the Mahakam Delta mangroves East Kalimantan, Indonesia. Oceanologica Acta, 13, 317-326. Dwi, S., Sato, K. and Kohda, Y. 1997. Relationships between TM data of Landsat 5 and stand parameters on mangrove forest in Okinawa. Jnl. of Japan Soc. of Photog. and RS., 36, 2: 4-12. Dwivedi, R.S., Rao, B.R.M. and Bhattacharya, S.
Mapping wetlands of the
Sundarban Delta and it’s environs using ERS-1 SAR data. Int. J. Remote Sensing, 20, 11: 2235-2247. Dyason, R. 2003. Page: NSW Agriculture Agfact: water hyacinth. Accessed: 23/10/03.
NSW
agriculture.
http://www.agric.nsw.gov.au/reader/aquatic-
weeds/p7643.htm Earickson, R.J. and Harlin, J.M. 1994. Geographic measurement and quantitative analysis. Macmillan College Publishing Co., New York, USA. Eddy, J.O.F. and Faud, T.D. 1996. Global climate change impacts on habitats: assessing ecological implications of changes in climate. Bull. Ecol. Soc. America, 77: 109-112. Edyvane, K.S. 1999. Coastal and marine wetlands in Gulf St. Vincent, South Australia: understanding their loss and degradation. Wetl. Ecol. & Mngt., 7: 83-104. Elachi, C. 1987. Introduction to the physics and techniques of remote sensing. John Wiley & Sons Inc., Canada. Eliot, I., Finlayson, C.M. and Waterman, P. 1999. Predicted climate change, sea level rise, and wetland management in the Australian wet-dry tropics. Wetl. Ecol. Management, 7: 63-81. Ellison, A. and Farnsworth, E. 1997. Simulated sea level change alters anatomy, physiology, growth and reproduction of red mangroves. Oecologica, 112: 435-446.
460
Ellison, J.C. and Stoddart, D.R.
1991.
Mangrove ecosystem collapse during
predicted sea level rise: Holocene analogues and implications. Journal of Coastal Research, 7,1: 151-165. Emery, K.O. and Aubrey, D.G. 1991. Sea levels, land levels and tide guages. Springer-Verlag, New York. Engheta, N. and Elachi, C. 1982. Radar scattering from a diffuse vegetation layer over a smooth surface. IEEE Trans. Geosci. Remote Sens., 20: 212-216. Eong, O. J., Khoon, G.W., Ping, W.Y. and Kheng, W.H. 1992. Identification of mangrove vegetation zones using MicroBRIAN and Landsat imagery.
In Third
ASEAN Science and Technology Week Conference Proceedings, vol. 6 (Singapore: University of Singapore) pp. 383-389. Evans, D.L. (ed.). 1995. Spaceborne Synthetic Aperture Radar: Current Status and Future Directions. A Report to the Committee on Earth Sciences, Space Studies Board, National Research Council.
NASA Technical Memorandum 4679.
National
Aeronautics and Space Administration, USA. Farnsworth, E.J., Ellison, A.M. and Gong, W.K.
1996.
Elevated CO2 alters
anatomy, physiology, growth, and reproduction of red mangrove (Rhizophera mangle L.). Oecologia, 108: 599-609. Ferrazzoli, P. and Guerriero, L. 1995. Radar sensitivity to tree geometry and woody volume: a model analysis. IEEE Trans. Geosci. Remote Sens., 33, 2: 360-371. Field, C.D.
1995.
Impact of expected climate change on mangroves.
Hydrobiologia, 295: 75-81. Finlayson, C.M., Davidson, N.C., Spiers, A.G. and Stevenson, N.J. 1999. Global wetland inventory – current status and future priorities. Mar. Freshw. Res., 50: 717-27. Finlayson, C.M. and Bayliss, B. (eds.).
1997.
Data management systems for
environmental research in northern Australia. Proceedings of a Workshop held in Jabiru, Northern Territory, 22 July, 1995. Supervising Scientist Report, 124. Finlayson, C.M. and Moser, M. 1991. Australasia and Oceania. In: Finlayson, C.M. and Moser, M. (eds).
Wetlands. International Waterfowl and Wetlands Research
Bureau, United Kingdom. Finlayson, C.M., Bailey, B.G. and Cowie, I.D. 1989. Macrophyte vegetation of the Magela Creek floodplain, Alligator Rivers Region, NT. Supervising Scientist Research Report, No. 5. Supervising Scientist, Aust. Govt. Pub. Service, Canberra, Australia.
461
Franks, T. and Falconer, R. 1999. Developing procedures for the sustainable use of mangrove systems. Agricultural Water Management, 40: 59-64. Fransson, J.E.S. and Israelsson, H. 1999. Estimation of stem volume in boreal forests using ERS-1 C- and JERS-1 L-band SAR data. Int. J. Remote Sens., 20, 1: 123137. Freeman, A. and Durden, S.L. 1998. A three-component scattering model for polarimetric SAR data. IEEE Trans. Geosci. Remote Sens., 36, 3: 963-973. Fromard, F., Puig, H., Mougin, E., Marty, G., Betoulle, J.L. and Cadamuro, L. 1998. Structure, above-ground biomass and dynamics of the French Guyana mangrove forest. Oecologia., 115: 39-53. Fung, A. K., Li, Z. and Chen, K.S. 1992. Backscattering from a randomly rough dielectric surface. IEEE Trans. Geosci. Remote Sens., 30, 6: 799-808. Gang, P.O. and Agatsiva, J. L. 1992. The current status of mangroves along the Kenyan coast, a case study of Mida Creek mangroves based on remote sensing. Hydrobiologia, 247, 29-36. Gao, J. 1999. A comparative study on spatial and spectral resolutions of satellite data in mapping mangrove forests. Int. J. Remote Sensing., 20, No. 14: 2823-2833. Gillison, A.N. 2001. A review of the impact of climate change on forest biological diversity. Convention on Biological Diversity. UNEP, Canada. Godfrey, J.S. and Love, G. 1992. Assessment of sealevel rise, specific to the South Asian and Australian situations. In: Woodworth, P.L., Pugh, D.T., DeRonde, J.G., Warrick, R.G. and Hannah, J.
Sea Level changes: Determination and effects.
Geophysical monograph, 69, IUGG, v.11, USA. Galloway, R.W., Story, R., Cooper, R. and Yapp, G.A. (1984). Coastal Lands of Australia. CSIRO Division of Water and Land Resources, Natural Resource Series No. 1. Gao, J. 1999. A comparative study on spatial and spectral resolutions of satellite data in mapping mangrove forests. Int. J. Remote Sensing, 20, 14: 2823-2833. Gao, J.
1998.
A hybrid method toward accurate mapping of mangroves in a
marginal habitat from SPOT multispectral data. Int. J. Remote Sensing., 19, no. 10: 1887-1899. Golley, F.B., Odum, H.T. and Wilson, A.F. 1962. The structure and metabolism of a Puerto Rico mangrove forest in May. Ecology, 43: 9-19.
462
Gower, S.T., Kucharik, C.J. and Norman, J.M. 1999. Direct and indirect estimation of leaf area index, ƒAPAR, and net primary production of terrestrial ecosystems. Remote Sens. Environ., 70: 29-51. Granich, S. Page name: Global warming and Vietnam, Impact on Vietnam: The Mangrove ecosystem. 1993. Climate Research Unit, The University of East Anglia, UK. Accessed: 22/9/01. http://www.cru.uea.ac.uk/tiempo/flooro/briefing/vietnam/impa cts.htm Gray, D., Zisman, S. and Corves, C. 1990. Mapping of the mangroves of Belize. Technical Report, University of Edinburgh. Green, R.M. 1998. Relationships between polarimetric SAR backscatter and forest canopy and sub-canopy biophysical properties. Int. J. Remote Sens., 19, 12: 2395-2412. Green, E.P., Clark, C.D., Mumby, P.J., Edwards, A.J. and Ellis, A.C. 1998. Remote sensing techniques for mangrove mapping. Int. J. Remote Sensing, 19, 5: 935-956. Green, E.P., Mumby, P.J., Edwards, A.J., Clark, C.D. and Ellis, A.C.
1997.
Estimating leaf area index of mangroves from satellite data. Aquatic Botany, 58: 11-19. Green, E.P., Mumby, P.J., Edwards, A.J. and Clark, C.D. 1996. A review of remote sensing for the assessment and management of tropical coastal resources. Coastal Management, 24, 1, 1-40. Hallikainen, M.T., Ulaby, F.T., Dobson, M.C., El-Rayes, M.A. and Wu, L-K. 1985. Microwave dielectric behaviour of wet soil – Part I: empirical models and experimental observation. IEEE Trans. Geosci. Remote Sens., 23: 25-34. Held, A., Ticehurst, C., Lymburner, L. and Williams, N. 2003. High resolution mapping of tropical mangrove ecosystems using hyperspectral and radar remote sensing. Int. J. Remote Sensing, 24: 2739-3759. Henderson-Sellers, A. and Blong, R. 1989. The Greenhouse effect: living in a warmer Australia. NSW Uni. Press, Australia. Henry, H.A.L. and Aarssen, L.W. 1999. The interpretation of stem diameter-height allometry in trees: biomechanical constraints, neighbour effects, or biased regressions? Ecology Letters, 2: 89-97. Herwitz, S.R., Sandler, B. and Slye, R.E. 2000. Twenty-one years of crown area change in the Jasper Ridge Biological Preserve based on georeferenced multitemporal aerial photographs. Int. J. Remote Sens., 21, 1: 45-60.
463
Herwitz, S.R., Slye, R.E. and Turton, M. 1998. Co-registered aerial stereopairs from low-flying aircraft for the analysis of long-term tropical rainforest canopy dynamics. Photogrammetric Engineering and Remote Sensing, 64: 397-406. Herz, R. 1991. Manguezais do Brasil. Remote Sensing Laboratory, Sao Paulo, Brazil, 200pp. Hess, L.L., Melack, J.M., Filoso, S. and Wang, Y. 1995. Delineation of inundated area and vegetation along the Amazon floodplain with SIR-C Synthetic Aperture Radar. IEEE Trans. Geosci. Remote Sens., 33: 896-903. Hoekman, D.H. and Quiñones, M.J.
2000.
Land cover type and biomass
classification using AIRSAR data for evaluation of monitoring scenarios in the Colombian Amazon. IEEE Trans. Geosci. Remote Sens.,38: 685-696. Hsueh, M. and Lee, H. 2000. Diversity and distribution of the mangrove forests in Taiwan. Wetlands Ecol. & Mngt., 8: 233-242. Imhoff, M.L. 1995. Radar backscatter and biomass saturation: ramifications for global biomass inventory. IEEE Trans. Geosci. Remote Sens., 33: 511-518. Imhoff, M., Story, M., Vermillion, C., Khan, F. and Polcyn, F. 1986. Forest canopy characterization and vegetation penetration assessment with space-borne radar. IEEE Trans. Geosci. Remote Sens., GE-24, 4: 535-542. Intergovernmental Oceanographic Commission (IOC). 2001. Strategic Design Plan for the IOC-WMO-UNEP-ICSU-FAO Living Marine Resources Panel of the Global Ocean Observing System (GOOS). Tracking Change in Marine Ecosystems. GOOS Report, no. 94. UNESCO, France. Intergovernmental Panel on Climate Change (IPCC). 1997. The regional impacts of climate change: an assessment of vulnerability. A special report of IPCC Working Group II. ITRES. Page name: CASI 2. 28/10/03. ITRES Research Limited. Accessed: 30/11/03. http://www.itres.com/docs/casi2.html Jelgersma, S., Van der Zijp, M. and Brinkman, R. 1993. Sealevel rise and the coastal lowlands in the developing world. Jnl. Coastal Res., 9 (4): 958-972. Jensen, J.R. 1986. Introductory Digital Image Processing. Englewood Cliffs, NJ, Prentice Hall.
464
Kadmon, R. and Harari-Kremer, R. 1999. Studying long-term vegetation dynamics using digital processing of historical aerial photographs. Remote Sens. Environ., 68: 164-176. Karam, M.A., Amar, F., Fung, A.K., Mougin, E., Lopes, A., Le Vine, D.M. and Beaudoin, A. 1995. A microwave polarimetric scattering model for forest canopies based on vector radiative transfer theories. Remote Sens. Environ., 53: 16-30. Karam, M.A., Fung, A.K., Lang, R.H. and Chauhan, S.
1992.
A microwave
scattering model for layered vegetation. IEEE Trans. Geosci. Remote Sens., 30, 4: 767783. Kasischke, E.S., Melack, J.M. and Dobson, M.C. 1997. The use of imaging radars for ecological applications – a review. Remote Sens. Environ., 59: 141-156. Kay, R.J., Hick, P.T. and Houghton, H.J. 1991. Remote sensing of Kimberley rainforests. In: McKenzie, N.I., Johnston, R.B. and Kendrick, P.G. (eds.), Kimberley Rainforests, Chipping Norton, Surrey Beatty & Sons, pp.41-51. Keene, A.F. and Melville, M.D. 1999. Role of nutrients in the delimitation of the Grey mangrove, Avicennia marina, in estuarine sediments. In: Kesby, J.A., Stanley, J.M., McLean, R.F. and Olive, L.J. (eds.), Geodiversity: Readings in Australian geography at the close of the 20th century. Special Publication Series No. 6, Canberra, ACT, School of Geography and Oceanography, University College, Australian Defence Force Academy. Knighton, A.D., Woodroffe, C.D. and Mills, K. 1992. The evolution of tidal creek networks, Mary River, Northern Australia. Earth Surface Processes and Landforms, 17: 167-190. Knighton, D.A., Mills, K. and Woodroffe, C.D. 1991. Tidal-creek extension and saltwater intrusion in northern Australia. Geology, 19: 831-834. Knox, G.A. 1986. Estuarine ecosystems: a systems approach. Vol. 1, CRC Press Inc., Florida, USA. Kondratyev, K.Ya., Buznikov, A.A. and Pokrovsky, O.M. 1996. Global Change and Remote Sensing. John Wiley & Sons, England. Krohn, M.D., Milton, N.M. and Segal, D.B. 1983. Seasat synthetic aperture radar (SAR) response to lowland vegetation types in eastern Maryland and Virginia. J. Geophys. Res., 88, 1937-1952.
465
Kushwaha, S.P.S., Dwivedi, R.S. and Rao, B.R.M. 2000. Evaluation of various digital image processing techniques for detection of coastal wetlands using ERS-1 SAR data. Int. J. Remote Sensing, 21, 3: 565-579. Kuenzler, E.J. 1974. Mangrove swamp systems. In: Odum, H.T., Copeland, B.J. and WcMahan, E.A. (eds.), Coastal ecological systems of the United States, Vol. 1, The Conservation Foundation, Washington, D.C., pp.346-371. Lear, R. and Turner, T. 1994. Mangroves of Australia. University of Queensland Press, Brisbane, Australia. Lefsky, M., Cohen, W., Acker, S., Parker, G., Spies, T. and Harding, D. 1999. Lidar remote sensing of biophysical properties and canopy structure of forests of Douglas-fir and western hemlock. Remote Sensing of Environment, 70: 339¯36. Le Toan, T., Beaudoin, A., Riom, J. Guyon, D. 1992. Relating forest biomass to SAR data. IEEE Trans. Geosci. and Remote Sens., 30: 403-411. Lillesand, T.M. and R.W. Kiefer. 1994. Remote Sensing and Photo Interpretation, 3rd ed. John Wiley & Sons: New York. Lin, G. and Sternberg, L. da S.L. 1992. Effect of growth form, salinity, nutrient and sulfide on photosynthesis, carbon isotope discrimination and growth of Red Mangrove (Rhizophora mangle L.). Aust. J. Plant Physiol., 19: 509-17. Long, B.G. and Skewes, T.D. 1996. A technique for mapping mangroves with Landsat TM satellite data and Geographic Information System. Estuarine, Coastal and Shelf Science, 43: 373-381. Long, B.G. and Skewes, T.D. 1994. GIS and remote sensing improves mangrove mapping. Proceedings of the 7th Australian Remote Sensing Conference, Melbourne, March, pp. 545-550. Loo, M.G.K., Lim, T.M. and Chou, L.M. 1992. Land use changes of a recreational island as observed by satellite imagery. Third ASEAN Science and Technology Week Conference Proceedings, vol. 6, Singapore: University of Singapore, pp. 401-405. Lopes, A. and Mougin, E. 1990. Microwave coherent propagation in cylindricalshaped forest components: interpretation of attenuation observations.
IEEE Trans.
Geosci. Remote Sens., 28, 3: 315-324. Lorenzo, R., de Jesus, B.R. and Jara, R.B. 1979. Assessment of mangroves forest deterioration in Zamboanga Peninsula, Philippines, using Landsat MSS data.
466
Proceedings of the 13th International Symposium on Remote Sensing of the Environment, vol. 2, pp. 1737-1745. Lucas, R.M., Held, A., Phinn, S. and Saatchi, S. 2004. Remote sensing of Tropical Forests. Manual of Remote Sensing, Natural Resources, Vol. 4 (in press). Lucas, R.M., Rowlands, A.R. and Niemann, O. 2004. Hyperspectral sensors: past, present and future. In: Pramod, K., Varsheny and Manoj, K. (ed.s), Advanced Image Processing Techniques for Remotely Sensed Hyperspectral Data., (in preparation). Lucas, R.M., Ellison, J.C., Mitchell, A., Donnelly, B., Finlayson, M., and Milne, A.K. 2002. Use of stereo aerial photography for quantifying changes in the extent and height of mangroves in tropical Australia. Wetlands Ecology and Management, 10: 161-175. Lucas, R.M., Milne, A.K., Mitchell, A., Donnelly, B. and Ellison, J. 2000. Use of stereo aerial photography for assessing changes in the extent and height of mangrove canopies in tropical Australia. Proceedings of the IGARSS Conference, Hawaii. Luckman, A., Baker, J., Honzak, M. and Lucas, R. 1998. Tropical forest biomass density estimation using JERS-1 SAR:
seasonal variation, confidence limits, and
application to image mosaics. Remote Sens. Environ., 63: 126-139. Magnussen, S., Gougeon, F., Leckie, D. and Wulder, M. 1999. Predicting tree heights from a combination of LIDAR canopy heights and digital stem counts. Remote Sensing and Forest Monitoring, Rogow, Poland, Conference Proceedings, 1999. Malthus, T.J. and Mumby, P.J. 2003. Remote sensing of the coastal zone: an overview and priorities of the coastal zone: an overview and priorities for future research. International Journal of Remote Sensing, 24, 13: 2805-2815. Marconi Integrated Systems Inc. 1998. Socet Set User’s Manual. LH Systems. San Diego, USA. McTainsh, G. and Boughton, W.C. 1993. Land degradation processes in Australia. Addison Wesley Longman Pub., Australia. Medina, E., Fonseca, H., Barboza, F. and Francisco, M. 2001. Natural and maninduced changes in a tidal channel mangrove system under tropical semiarid climate at the entrance of the Maracaibo lake (Western Venezuela). Wetl. Ecol. & Mngt., 9: 233243.
467
Merton, R.N. 1999. Multi-temporal analysis of community scale vegetation stress with imaging spectroscopy.
Ph.D. Thesis, Geography Department, University of
Auckland, New Zealand. 492p. Metternicht, G.I. and Zinck, J.A. 2003. Remote sensing of soil salinity: potentials and constraints. Remote Sensing of Environment, 85: 1-20. Micheli, F. 1993. Feeding ecology of mangrove crabs in North Eastern Australia: mangrove litter consumption by Sesarma messa and Sesarma smithii. J. Exp. Mar. Biol. Ecol., 171: 164-86. Milne, A.K. and Dong, Y. 2002. Vegetation mapping using JERS-1 SAR mosaic for northern Australia. Int. J. Remote Sensing, 23, 7: 1475-1486. Milne, A.K., Horn, G. and Finlayson, M. 2000. Monitoring wetlands inundation patterns using RADARSAT multi-temporal data. Canadian Jnl. Remote Sens., 26, 2: 133-141. Milne, A.K., Horne, G. and Finlayson, C.M. 1998. Monitoring wetlands inundation patterns using RADARSAT multitemporal data.
Proceedings of the RADARSAT
ADRO Final Symposium, Montreal, Quebec, 13-15 October, 1998 (Saint-Hubert, Quebec: Canadian Space Agency), CD-ROM. Mitchell, C.D. 1992. Grappling with Greenhouse – Understanding the Science of Climate Change. Comm. Aust. Report. Mitchell, A. and Lucas, R.M. 2001. Integration of aerial photography, hyperspectral and SAR data for mangrove characterisation. IGARSS Conference Proceedings, July, 2001. Mohamed, M.I.H., Hussein, N. A., Ibrahim, M. I. and Kuchler, D.A.. 1992. Coastal resources mapping of Besut (Terrengganu coast), Malaysia, using microBRIAN digital mapping technology and Landsat MSS satellite data. Proceedings of the Regional Symposium on Living Resources in Coastal Areas, Manila, Philippines. Manila Marine Science Institute, University of the Philippines, pp. 545-555. Mougin, E., Proisy, C., Marty, G., Fromard, F., Puig, H., Betoulle, J.L. and Rudant, J.P. 1999. Multifrequency and multipolarisation radar backscattering from mangrove forests. IEEE Trans. Geosc. & Remote Sens., 37, 1: 94-102. Mumby, P.J., Green, E.P., Edwards, A.J. and Clark, C.D.
1999.
The cost-
effectiveness of remote sensing for tropical coastal resources assessment and management. Journal of Environmental Management, 55: 157-166.
468
Mummery, D., Battaglia, M., Beadle, C.L., Turnbull, C.R.A. and McLeod, R. 1999. An application of terrain and environmental modelling in a large-scale forestry experiment. Forest Ecology and Management, 118: 149-159. NASA JPL. Page name: Tech briefs – Mapveg. 2003. NASA JPL, CA. Accessed: 10/12/03. http://www.nasatech.com/software/na.html Nicholls, R. Page name: The impacts of sea level rise. 1999. Middlesex University. Accessed: 13/6/02. http://www.met-office.gov.uk/sec5/CR_div/pubs/brochures/B1999/ imp_sea_rise.html. NT Government. 1998. Mary River Integrated Management Plan. NT government, Darwin, Australia. O’Grady, A.P., McGuinness, K.A. and Eamus, D. 1996. The abundance and growth of Avicennia marina and Rhizophora stylosa in the low shore zone of Darwin Harbour, Northern Territory. Aust. J. Ecol., 21: 272-279. Ormsby, J.P., Blanchard, B.J. and Blanchard, A.J. 1985. Detection of lowland flooding using active microwave systems. Photogramm. Eng. Remote Sens., 51: 317328. Palaganas, V.P.
1992.
Assessing changes in mangrove forest of Infanta-Real,
Quezon Province, Philippines, using remote sensing. Msc dissertation, University of Newcastle upon Tyne. Panapitukkul, N., Duarte, C.M., Thampanya, U., Kheowvongsri, P., Srichai, N., Geertz-Hansen, O., Terrados, J. and Boromthanarath, S. 1998. Mangrove colonization: mangrove progression over the growing Pak Phanang (SE Thailand) mud flat. Estuarine, Coastal and Shelf Science, 47: 51-61. Paterson, M., Chisholm, M. and Lucas, R.M. 2001. Differentiation of Selected Australian Woodland Species Using CASI Data. Proc. IGARSS Conference, 2001. Patterson, S. G. and Rehder, J.B. 1985. An assessment of conversion and loss of mangroves using remote sensing imagery on Marco Island, Florida. Proceedings of the American Society for Photogrammetry and Remote sensing, Reston, VA:ASPRS, pp. 728-735. Petrie, G. 1990. Photogrammetric methods of data acquisition for terrain modelling. In: Petrie, G. and Kennie, T.J.M. (eds.), Terrain Modelling in Surveying and Civil Engineering. Whittles Pub. Services, London.
469
Phinn, S. and Dale, P.
Development of an operational approach for mapping
mosquito breeding sites from airborne synthetic aperture radar. 2003. University of Queensland. Accessed: 4/6/02. http://www.geosp.uq.edu.au/research/proposal.htm. Pirazzoli, P.A. 1990. Present and near-future sea-level changes: an assessment. In: Paepe, R., Fairbridge, R.W. and Jelgersma, S. (eds.), Greenhouse Effect, Sea Level and Drought. Kluwer, Netherlands. Pittock, A.B. 1988. The greenhouse effect and future climatic change. In: Gregory, S. (Ed.), Recent Climatic Change: A Regional Approach. Belhaven press, London. Pope, K.O., Rejmankova, E., Paris, J.F. and Woodruff, R. 1997. Detecting seasonal flooding cycles in marshes of the Yucatan peninsula with SIR-C polarimetric radar imagery. Remote Sens. Environ., 59, 157-166. Populus, J., and Lantieri, D. 1991. Use of high resolution satellite data for coastal fisheries. Report 58, Food & Agricultural Organization, Rome. Primavera, J.H. 1997. Socio-economic impacts of shrimp culture. Aquaculture Res., 28: 815-827. Proisy, C., Mitchell, A., Lucas, R.M., Fromard, F. and Mougin, E. 2003. Estimation of mangrove biomass using multifrequency radar data. Application to mangroves of French Guiana and Northern Australia. Proceedings Mangroves 2003 Conference, held in Bahia, Brazil, South America, May 20-24. Proisy, C., Mougin, E., Fromard, F., Trichon, V. and Karam, M.A. 2002. On the influence of canopy structure on the radar backscattering of mangrove forests. Int. J. Remote Sens., 23, 20: 4197-4210. Proisy, C., Mougin, E., Fromard, F. and Karam, M.A. 2000. Interpretation of polarimetric radar signatures of mangrove forests. Remote Sens. Environ., 71: 56-66. Ramirez-Garcia, P., Lopez-Blanco, J. and Ocana, D. 1998. Mangrove vegetation assessment in the Santiago River mouth, Mexico, by means of supervised classification using Landsat TM imagery. Forest Ecol. Manag., 105: 217-229. Ranganath, B.K., Dutt, C.B.S. and Manikan, B. 1989. Digital mapping of mangrove in middle Andamans of India. Proceedings of the 6th Symposium on Coastal and Ocean Management, vol. 1, pp. 741-750. Ranson, K.J. and Sun, G. 1997. An evaluation of AIRSAR and SIR-C/X-SAR images for mapping Northern forest attributes in Maine, USA. Remote Sens. Environ., 59: 203-222.
470
Ranson, K.J. and Sun, G. 1994. Mapping biomass of a northern forest using multifrequency SAR data. IEEE Trans. Geosc. & Remote Sens., 32: 388-396. Rasolofoharinoro, M., Blasco, F., Bellan, M.F., Aizpuru, M., Gauquelin, T. and Denis, J.
1998.
A remote sensing based methodology for mangrove studies in
Madagascar. Int. J. Remote Sensing, 19, 10: 1873-1886. Raven, P.H., Evert, R.F. and Eichhorn, S.E. 1992. Biology of Plants. Worth Publishers, New York. Research Systems Inc. 2001. ENVI User’s Manual. RSI, USA. Riaza, A., Martinez-Torres, M.L., Ramon-Lluch, R., Alonso, J. and Heras, P. 1998. Evolution of equatorial vegetation communities mapped using Thematic Mapper images through a geographical information system (Guinea, Equatorial Africa). Int. J. Remote Sensing., 19, No. 1: 43-54. Richards, J.A., Sun, G-Q. and Simonett, D.S. 1987. L-band radar backscatter modelling of forest stands. IEEE Trans. Geosc. Remote Sens., GE-25, 4: 487-497. Richards, J.A. Remote sensing digital image analysis. 1986. Springer-Verlag, Berlin. Rosenqvist, A., Imhoff, M., Milne, A.K. and Dobson, C. 2000. Remote sensing and the Kyoto Protocol: a review of available and future technology for monitoring treaty compliance. Report of Remote Sensing and the Kyoto Protocol Workshop, 20-22 October, 1999, Michigan. Published by Centre for Remote Sensing & GIS, UNSW, p.152. Roy, P. S. 1989. Mangrove vegetation stratification using Salyut 7 photographs. Geocarthography International, 3, 31-47. Saatchi, S., Nelson, B., Podest, E. and Holt, J. 2000. Mapping land cover types in the Amazon Basin using 1 km JERS-1 mosaic.
International Journal of Remote
Sensing, 21, 6&7: 1201-1234. Sabins, F. 1978. Remote sensing: principles and interpretation. W.H Freeman, San Francisco. Sader, S.A.
1987.
Airborne remote sensing of forest biomes.
Geocarto
International, 1: 9-15. Saenger, P.E. 1994. Mangroves and saltmarshes. In: Hammond, L.S. and Synnot, R.N. (Eds.), Marine Biology. Addison Wesley Longman, Melbourne, Australia.
471
Saenger, P., Specht, M.M., Specht, R.L. and Chapman, V.J. 1977. Mangrove and coastal saltmarsh communities in Australia. In: Chapman, V.J. (ed.), Wet Coastal Ecosystems. Elsevier, Amsterdam, pp 293-345. Saintilan, N. and Wilton, K. 2001. Changes in the distribution of mangroves and saltmarshes in Jervis Bay, Australia. Wetlands Ecol. & Mngt., 9: 409-420. Saintilan, N. and Williams, R.J.
1999.
Mangrove transgression into saltmarsh
environments in south-east Australia. Global Ecology and Biogeography., 8: 117-124. Saintilan, N. 1998. Relationships between height and girth of mangroves and soilwater conditions in the Mary and Hawkesbury River estuaries, eastern Australia. Aust. J. Ecol., 23: 322-328. Saintilan, N. 1997. Above- and below-ground biomass of mangroves in a subtropical estuary. Mar. Freshwater Res., 48: 601-604. Schmalzer, P.A. 1995. Biodiversity of saline and brackish marshes of the Indian River Lagoon: Historic and current patterns. Bull. Mar. Sci., 57: 37-48. Semeniuk, V. 1985. Mangrove environments of Port Darwin, Northern Territory: The physical framework and habitats. J. Roy. Soc. West. Aust., 67: 81-97. Semeniuk, V., Kenneally, K.F. and Wilson, P.G. 1978. Mangroves of Western Australia. Western Australian Naturalists Club, Perth, Australia. Simard, M., De Grandi, G., Saatchi, S. and Mayaux, P. 2002. Mapping tropical coastal vegetation using JERS-1 and ERS-1 radar data with a decision tree classifier. Int. J. Remote Sensing, 23, 7: 1461-1474. Slama, C.C. 1980. Manual of Photogrrammetry, 4th Edn. American Society of Photogrammetry, Bethseda, Maryland, USA. Spalding, M., Blasco, F. and Field, C. 1997. World Mangrove Atlas, International Society for Mangrove Ecosystems, Okinawa, Japan, 180pp. Storrs, M.J. and Finlayson, M.
1997.
Overview of the conservation status of
wetlands of the Northern Territory. Supervising Scientist Report No. 116. Supervising Scientist, Environment Australia, Canberra, Australia. Taal, M.D.
1994.
Flow and salt transport in mangrove swamps.
Communications on hydraulic and geotechnical engineering, 94-6.
In:
Fac. of Civil
Engineering, Delft University of Technology. Talaue-McManus, L. 2000. Transboundary Diagnostic Analysis for the South China Sea. EAS/RCU Technical Report Series No. 14. UNEP, Bangkok, Thailand.
472
Tanaka, S., Sugimura, T., Senoo, T. and Nishi, T. 1998. Forest stand structure analysis in Mt Fuji area with OPS and SAR data. Adv. Space Res., 22, 5: 685-688. Tanis, F.J., Bourgeau-Chavez, L.L. and Dobson, M.C. 1994. Application of ERS-1 SAR for coastal inundation. International Geoscience and Remote Sensing Symposium, Pasadena, California, August 8-12. Thom, B.G. and Chappell, J. 1975. Holocene sea levels relative to Australia. Search, 6: 90-93. Thomson, A.G., Fuller, R.M., Sparks, T.H., Yates, M.G. and Eastwood, J.A. 1998. Ground and airborne radiometry over intertidal surfaces: waveband selection for cover classification. Int. J. Remote Sensing., 19, No. 6: 1189-1205. Toutin, T. and Cheng, P. Page name: Quickbird – a Milestone for High Resolution Mapping. 2003. DigitalGlobe. Accessed: 21 Nov, 2003. http://www.ccrs.nrcan.gc.ca/c crs/rs/sci_pub/bibpdf/13212.pdf Toutin, T. and Gray, L. 2000. State-of-the-art of elevation extraction from satellite SAR data. ISPRS Journal of Photogrammetry and Remote Sensing, 55: 13-33. Townsend, P.A. 2002. Relationships between forest structure and the detection of flood inundation in forested wetlands using C-band SAR. Int. J. Remote Sens., 23, 3: 443-460. Townsend, P.A. 1998. Detection of hydroperiod in forested wetlands using multitemporal Radarsat imagery. Proceedings of the RADARSAT ADRO Final Symposium, Montreal, Quebec, 13-15 October, 1998 (Saint Hubert, Quebec: Canadian Space Agency), CD-ROM. Tropical Savannas CRC. Page name: Savannah Arc Maps – The vegetation of the Australian Tropical Savannas. 1998. Tropical Savannas CRC. Accessed: 23/9/03. http://eon.ntu.edu.au/tscrc/navegHTML/viewer.htm Tuffers, A., Naidoo, G. and Von Willert, D.J. 2001. Low salinities adversely affect photosynthetic performance of the mangrove, Avicennia marina. Wetl. Ecol. & Mngt., 9: 225-232. Ulaby, F.T., Sarabandi, K., McDonald, K., Whitt, M. and Dobson, M.C. 1990. Michigan microwave canopy scattering model. Int. J. Remote Sens., 11, 7: 1223-1253. Ulaby, F.T. and El-Rayes, M.A. 1987. Microwave dielectric spectrum of vegetation – Part II: Dual dispersion model. IEEE Trans. Geosc. Remote Sens., 22: 550-557.
473
Untawale, A.G., Sayeeda, W. and Jagtap, T.G. 1982. Application of remote sensing techniques to study the distribution of mangroves along the estuaries of Goa. Wetlands Ecology and Management, Lucknow Publishing House (India), pp. 52-67. Vibulsresth, S., Downreang, D., Ratanasermpong, S. and Silapathong, C. 1990. Mangrove forest zonation by using high resolution satellite data. Proceedings of the 11th Asian Conference on Remote Sensing, D-1-6. Vits, C, and Tack, J. 1995. The use of remote sensing as information source for environmental accounting of coastal areas in Kenja. Feasibility Study Reference No. T3/02/603, University of Ghent, 1-45. Wakebayashi, H. and Arai, K. 1996. A new method for speckle noise reduction (CST filter). Canadian Jnl. of Remote Sensing, 22: 190-197. Walsh, G.E. 1977. Exploitation of mangal. In: Chapman, V.J. (ed.), Ecosystems of the World, Vol. 1, Wet coastal ecosystems, pp. 347-362. Elsevier Scientific Publishing Company, New York. Wang, Y., Hess, L.L., Filoso, S. and Melack, J.M. 1995. Understanding the radar backscattering from flooded and nonflooded Amazonian forests: results from canopy backscatter modelling. Remote Sens. Environ., 54: 324-332. Wang, Y. and Imhoff, M.L. 1993. Simulated and observed L-HH radar backscatter from tropical mangrove forests. Int. J. Remote Sens., 14, 15: 2819-2828. Watson, R.T., Zinyowera, M.C. and Moss, R.H. (eds.) 1998. The Regional Impacts of Climate Change: An Assessment of Vulnerability. Cambridge University Press, Cambridge. Wegmüller, U., Werner, C.L., Nüesch, D. and Borgeaud, M. 1995. Forest mapping using ERS repeat-pass SAR interferometry. ESA Bulletin, No. 4, February 1995, The Netherlands. Whetton, P., Adamson, D. and Williams, M.
1990.
Rainfall and river flow
variability in Africa, Australia and East Asia linked to El Niño – Southern Oscillation events. Geol. Soc. Aust. Symp. Proc., 1: 71-82. Wigley, T.M.L. and Raper, S.C.B. 1992. Implications for climate and sea-level of revised IPCC emissions scenarios. Nature, 357: 293-300. Wolanski, E. and Cassagne, B. 2000. Salinity intrusion and rice farming in the mangrove-fringed Konkoure River delta, Guinea. Wetl. Ecol. & Mngt., 8: 29-36.
474
Wolanski, E. and Chappell, J. 1996. The response of tropical Australian estuaries to a sea level rise. Jnl. Mar. Systems, 7: 267-279. Woinarski, J. Page name: Arnhem Land Tropical Savanna. 2001. World Wildlife Fund. Accessed: 1/11/03. http://www.worldwildlife.org/wildworld/profiles/terrestrial/a a/aa0701_full.html. Wong, A., Chapman, B., Richardson, A. and Freeman, A. Page name: How imaging radar works. 2003. NASA JPL. Accessed: 7/12/03. http://southport.jpl.nasa.gov/cdro m/sirced03/cdrom/DOCUMENT/HTML/TEACHERS/MODULE02/MOD2SECE.HTM Woodfine, A.C. component.
1991.
North east Sumatra prawn project, remote sensing
Final Report to Natural Resources Institute/Overseas Development
Agency, December 1991. Woodroffe, C.D.
2000.
Deltaic and estuarine environments and their Late
Quaternary dynamics on the Sunda and Sahul shelves. Journal of Asian Earth Sciences, 18: 393-413. Woodroffe, C.D.
1995.
Response of tide dominated mangrove shorelines in
Northern Australia to anticipated sea-level rise. Earth Surf. Proc. Landforms, 20: 6585. Woodroffe, C.D. and Mulrennan, M.E. 1993. Geomorphology of the lower Mary River plains, Northern Territory, 152pp. North Australian Research Unit, Darwin. Woodroffe, C.D.
1990.
The impact of sea-level rise on mangrove shoreline.
Progress in Physical Geography, 14: 483-502. Woodroffe, C.D. and Grindrod, J. 1991. Mangrove biogeography: the role of Quaternary environmental and sea level change. Jnl. Biogeography., 18: 479-492. Woodroffe, C.D. 1988. Changing mangrove and wetlands habitats over the last 8000 years, Northern Australia and Southeast Asia. In: Marshall, D.W. and Loveday, P. (eds). Northern Australia: Progress and Prospects. Floodplains Research. Australian National University. North Australian Research Unit, Darwin. Woodroffe,
C.D.,
Chappell,
J.,
Thom,
B.
and
Wallensky,
E.
1986.
Geomorphological dynamics and evolution of the South Alligator tidal river and plains, Northern Territory. North Australian Research Unit, Darwin. Wu, S-T. and Sader, S.A. 1987. Multipolarization SAR data for surface feature delineation and forest vegetation characterization. IEEE Trans. Geosc. Remote Sens., GE-25, 1: 67-76.
475
Zann, L.P. 1995. Our Sea, Our Future. Major findings of the State of the Marine Environment Report for Australia.
Great Barrier Reef Marine Park Authority,
Townsville, Australia.
476
APPENDIX A TRIANGULATION REPORTS: STEREO AERIAL PHOTOGRAPHY
Table A1 Ground point parameters for registration of aerial photographs. Region
Block id
RMS x
RMS y
RMS z
1
No. of photographs 15
56.058
79.099
0.828
Total RMSE 66.509
Wildman R. West Alligator R.
1 2
3 18
17.721 43.923
7.131 46.179
3.172 1.779
15.916 52.744
South Alligator R. (west bank) (east bank)
1 2 3 4
9 5 12 7
8.811 10.343 38.965 11.887
15.36 6.161 43.068 19.368
1.022 1.632 1.35 1.271
16.914 8.279 33.555 22.727
Field Island
1
13
13.406
18.671
0.775
13.288
East Alligator R.
1 2 3
24 17 10
29.380 23.074 24.458
26.899 19.878 21.684
0.981 2.722 1.327
37.355 27.376 34.541
477
APPENDIX B
Table B1 Base (x) Base (y) 506 667.25 2090.25 1108.75 3414.33 1420.33 4246 847.67 5147.67 478.67 6220.67 186.33 37.33 1928.67 1668.67 2004 3328.33 1945.67 4945 1784.67 6001 2170.33 273.33 3442 1750 3017 3080 3300.33 4826 2764.33 6299 2947.67 260.33 5122.67 1850 4844.67 2501.67 4808 4125.33 4552.33 6468 4219.33 252.67 6871 1257.67 6577.33 1790 6860.67 3251.67 6712.67 4647 6616 6584.67 6488.33 95.67 8611 440.67 8868.67 1852.67 8794.67 3609.67 9048 4838.67 8735.67 5764.33 9173.33 Total RMSE 1.846
GCPs AND REGISTRATION REPORT: CASI
Ground point parameters for registration of CASI data. Warp (x) 5333.75 6918 8244.67 9071.33 9972 11046.67 4867 6499 8157.33 9771.67 10828 5111.33 6580 7912.33 9655 11131.67 5098.67 6686.67 7336 8958.67 11304.33 5096 6101.33 6633.33 8091 9490.33 11428 4945 5288 6700 8461 9686.33 10620.33
Warp (y) 3397.75 3835.67 4143.67 3569.33 3196 2899 4660.33 4732 4665.67 4501 4882 6174 5742.67 6021 5479.67 5659 7855.67 7567.67 7530.33 7268.33 6928.67 9601.33 9303.67 9583.67 9431 9329.33 9195.67 11341.67 11593.67 11520.33 11770 11449.33 11875.33
Predict (x) Predict (y) Error (x) 5333.34 3397.47 -0.41 6917.84 3836.04 -0.16 8241.84 4142.97 -2.83 9072.37 3568.89 1.04 9973.29 3196.62 1.29 11045.45 2897.77 -1.22 4868.84 4660.98 1.84 6498.42 4730.99 -0.58 8156.79 4667.32 -0.54 9772.8 4501.26 1.13 10829.84 4883.05 1.84 5108.77 6174.26 -2.56 6581.95 5742.55 1.95 7911.37 6020.4 -0.96 9655.83 5479.75 0.83 11130.07 5659.3 -1.6 5100.23 7854.71 1.56 6686.3 7568.66 -0.37 7337.18 7529.12 1.18 8959.58 7267.65 0.91 11303.23 6929.7 -1.1 5096.86 9601.58 0.86 6099.41 9304 -1.92 6631.92 9585.27 -1.41 8092.52 9431.24 1.52 9488.32 9328.89 -2.01 11428.87 9193.98 0.87 4943.94 11338.8 -1.06 5289.16 11596.14 1.16 6700.47 11521.01 0.47 8459.64 11768.51 -1.36 9689.2 11448 2.87 10619.11 11877.52 -1.22
Error (y) -0.28 0.37 -0.7 -0.44 0.62 -1.23 0.65 -1.01 1.65 0.26 1.05 0.26 -0.12 -0.6 0.08 0.3 -0.96 0.99 -1.21 -0.68 1.03 0.25 0.33 1.6 0.24 -0.44 -1.69 -2.87 2.47 0.68 -1.49 -1.33 2.19
RMS error 0.5 0.4 2.92 1.13 1.43 1.73 1.95 1.16 1.74 1.16 2.12 2.57 1.95 1.13 0.83 1.63 1.83 1.05 1.69 1.14 1.51 0.9 1.95 2.14 1.54 2.05 1.9 3.06 2.73 0.83 2.02 3.16 2.51
478
APPENDIX C GCPs AND REGISTRATION REPORT: AIRSAR
Table C1
Ground point parameters for registration of AIRSAR data, C-band.
Map (x) Map (y) 201650 8648746 199766 8645746 203774 8646118 202754 8642686 201542 8639038 205982 8640622 207866 8644726 205514 8647726 210842 8649370 210782 8652754 214466 8646934 214478 8642986 219134 8647174 218678 8649454 219482 8651662 220802 8654638 224570 8647462 232311 8650894 232719 8654398 229083 8657782 233775 8661298 238395 8657038 236727 8654986 237423 8659462 243687 8658010 247407 8660098 244275 8661562 241383 8666110 245583 8668090 251691 8663374 252075 8660842 252159 8666998 253743 8663686 Total RMSE 0.723
Image (x) 511.25 290.67 717.33 584 422.67 915.67 1146.67 916.33 1506 1533.33 1879.33 1859 2390.33 2356.33 2464.33 2636.33 2985.67 3850.33 3919.33 3556.33 4088 4551.67 4357 4463.67 5128 5542.33 5216.33 4950 5426 6020.67 6046 6110.67 6238.67
Image (y) 2133 2447.67 2437.33 2789.33 3177.33 3041 2613 2278 2145.67 1769.33 2433 2866 2448.33 2199.33 1970.33 1660.33 2468.67 2161.33 1784.33 1390.67 1049 1548.33 1754.33 1280 1491.33 1303.33 1115.67 592.67 412 991.67 1279.33 590 982
Predict (x) Predict (y) Error (x) 510.93 2133.27 -0.32 290.45 2448.03 -0.22 717.78 2436.19 0.45 583.99 2789.42 -0.01 422.71 3177.12 0.04 915.62 3041.52 -0.05 1146.73 2613.01 0.06 916.63 2278.26 0.3 1505.86 2145.34 -0.14 1533.85 1769.04 0.52 1878.95 2433.21 -0.38 1858.87 2865.64 -0.13 2390.07 2448.17 -0.26 2356.95 2199.84 0.62 2463.75 1971.14 -0.58 2635.73 1660.19 -0.6 2986.22 2468.54 0.55 3850.21 2161.72 -0.12 3919.29 1783.59 -0.04 3556.04 1390.76 -0.29 4088.58 1049.09 0.58 4551.39 1547.65 -0.28 4356.79 1754.05 -0.21 4465.02 1280.4 1.35 5128.42 1491.97 0.42 5541.58 1304.15 -0.75 5216.2 1114.79 -0.13 4948.59 592.93 -1.41 5427.12 411.7 1.12 6020.71 992.11 0.04 6045.77 1279.02 -0.23 6109.95 590.26 -0.72 6239.48 981.53 0.81
Error (y) 0.27 0.36 -1.14 0.09 -0.21 0.52 0.01 0.26 -0.33 -0.29 0.21 -0.36 -0.16 0.51 0.81 -0.14 -0.13 0.39 -0.74 0.09 0.09 -0.68 -0.28 0.4 0.64 0.82 -0.88 0.26 -0.3 0.44 -0.31 0.26 -0.47
RMS Error 0.41 0.43 1.23 0.09 0.21 0.52 0.06 0.4 0.36 0.6 0.43 0.38 0.3 0.8 1 0.61 0.57 0.41 0.74 0.31 0.59 0.74 0.36 1.41 0.76 1.11 0.88 1.44 1.16 0.44 0.39 0.77 0.94
479
Table C2 Map (x) 201626.02 201194.02 202694.02 203258.02 203618.02 203282.02 199730.02 202766.02 201614.02 202706.02 208358.02 207770.02 203738.02 204674.02 204590.02 205502.02 207722.02 210266.02 211658.02 212486.02 211538.02 212570.02 213734.02 214250.02 215786.02 214502.02 218942.02 218666.02 224402.02 225554.02 220826.02 223226.02 232431.58 234219.58 232275.58 235179.58 236931.58 229311.58 233439.58 233655.58 234627.58 239223.58 243483.58 247743.58 251175.58 250515.58 250383.58 253539.58
Ground point parameters for registration of AIRSAR data, L- and P-bands. Map (y) 8648734 8646814 8648278 8647378 8646094 8645254 8645770 8642686 8640574 8636830 8641318 8645266 8643850 8646154 8647066 8647714 8648062 8649310 8650978 8652682 8649886 8648986 8648062 8646322 8645614 8643010 8644054 8649442 8647498 8651014 8654518 8654482 8654518 8653486 8650882 8652046 8655850 8657974 8661262 8662522 8659582 8658598 8656090 8659582 8662906 8665426 8666350 8663794
Image (x) 457.5 391.75 567 619.5 645.25 602.75 221.25 530.25 390.75 470.25 1122.75 1087.5 645.5 761.5 762.25 864.25 1107 1396.5 1561.5 1666.5 1536.25 1639.5 1756.75 1796.75 1961.5 1800.75 2290.25 2301 2909.5 3055.5 2577.25 2832 3827.75 4014.25 3790.5 4111.75 4323.75 3516 3992.5 4027.75 4105.25 4596 5042.75 5521 5914.25 5865.5 5862.25 6166.25
Image (y) 2135.75 2336.75 2189.5 2292.5 2435.25 2522.75 2439 2791 3004.5 3418.75 2989.25 2558.25 2676.25 2437.75 2336.25 2277.25 2255.25 2141.25 1971.75 1793 2088.75 2197 2305.25 2498.5 2587.75 2853.25 2778.25 2201.5 2460 2091.25 1671.5 1693.75 1772.25 1898.75 2161 2058.75 1662.75 1366.5 1046.75 913 1238.75 1385 1683.5 1347.75 1032.75 756.5 655 969.5
Predict (x) Predict (y) Error (x) 457.03 2134.61 -0.47 390.21 2338.2 -1.54 567.35 2191.85 0.35 619.34 2293.37 -0.16 647.11 2434.46 1.86 604.05 2521.88 1.3 222.04 2438.12 0.79 529.75 2791.85 -0.5 390.64 3004.71 -0.11 470.63 3418.6 0.38 1121.72 2988.94 -1.03 1089.99 2558.06 2.49 643.1 2676.22 -2.4 762.04 2436.57 0.54 760.72 2337.52 -1.53 865.29 2274.74 1.04 1108.25 2255.14 1.25 1394.08 2141.08 -2.42 1560.3 1971.59 -1.2 1667.91 1792.8 1.41 1536.37 2089.49 0.12 1639.07 2196.33 -0.43 1756.75 2306.6 0 1798.82 2498.65 2.07 1959.75 2587.39 -1.75 1800.48 2854.78 -0.27 2291.33 2777.45 1.08 2300.03 2201.32 -0.97 2909.28 2458.51 -0.22 3056.68 2093.01 1.18 2575.9 1669.49 -1.35 2832.45 1695.53 0.45 3826.91 1771.57 -0.84 4015.14 1897.42 0.89 3788.68 2162.15 -1.82 4112.69 2059.19 0.94 4325.44 1662.5 1.69 3517.23 1368.38 1.23 3992.64 1046.26 0.14 4028.74 912.25 0.99 4104.87 1238.44 -0.38 4594.42 1383.1 -1.58 5041.54 1683.79 -1.21 5523.12 1348.02 2.12 5912.53 1033.8 -1.72 5867.03 756.65 1.53 5863.11 654.12 0.86 6165.89 968.48 -0.36
Error (y) -1.14 1.45 2.35 0.87 -0.79 -0.87 -0.88 0.85 0.21 -0.15 -0.31 -0.19 -0.03 -1.18 1.27 -2.51 -0.11 -0.17 -0.16 -0.2 0.74 -0.67 1.35 0.15 -0.36 1.53 -0.8 -0.18 -1.49 1.76 -2.01 1.78 -0.68 -1.33 1.15 0.44 -0.25 1.88 -0.49 -0.75 -0.31 -1.9 0.29 0.27 1.05 0.15 -0.88 -1.02
RMS Error 1.23 2.12 2.37 0.88 2.02 1.56 1.19 0.98 0.24 0.41 1.07 2.49 2.4 1.3 1.99 2.71 1.25 2.43 1.21 1.42 0.75 0.79 1.35 2.07 1.78 1.55 1.35 0.99 1.5 2.11 2.42 1.83 1.08 1.61 2.15 1.04 1.71 2.25 0.51 1.24 0.49 2.48 1.24 2.14 2.01 1.54 1.23 1.08
480
247923.58 8664850 245403.58 8663950 241275.58 8663458 241395.58 8666110 246915.58 8667550 251175.58 8668510 Total RMSE 1.68
5588 5309.25 4862.75 4902.75 5509.75 5971.5
790 860.75 876.25 592 489.25 419.75
5587.86 5309.79 4860.56 4901.96 5510.99 5970.37
790.48 861.91 876.61 593.59 486.64 421.05
-0.14 0.54 -2.19 -0.79 1.24 -1.13
0.48 1.16 0.36 1.59 -2.61 1.3
0.5 1.28 2.22 1.77 2.89 1.72
481
