...

Document 1911736

by user

on
Category: Documents
1

views

Report

Comments

Transcript

Document 1911736
Evaluation of remote sensing sensors for
monitoring of rehabilitated wetlands
by
Althea Theresa Grundling
Submitted in partial fulfillment of the requirements for the degree MAGISTER SCIENTIAE In the Faculty of Natural & Agricultural Science African Vegetation and Plant Diversity Research Center Department of Botany
,
University of Pretoria
Pretoria
Supervisor: Prof Dr. G.J. Bredenkamp
March 2004
© University of Pretoria
ACKNOWLEDGEM
research was financially supported by the Department of Agriculture: Land and
Management. Thanks to Mr. Chris
for his valuable
guidance and Mr. Philip
Beukes and Mr. Eric Economon for all their
Special thanks to Mrs.
van den Berg for all
pre-processing, prc}ceSSI
and
classification of image
I wish to thank
Bredenkamp (African Vegetation and
Centre, Department of Botany,
masters thesis, Mr.
Grundling (Working
!\lell (Institute for
Hennie van den
Chair of Vegetation
shared valuable
, Climate and Water,
(Consultant) and Dr. Jan
who spend much
Pretoria) the supervisor for this
Wetlands,
and Soil
Mr.
Mr.
(Technische Universitat MOnchen,
helping me, offered information,
gave crucial comments on the text of the thesis.
Thank you to all the governmental and
and help regarding
Diversity
data for each
organisations for their support
CONTENTS Acknowledgements 1
ummary
CHAPTER 1: INTRODUCTION
1.1
INTRODUCTION..........................................................................
3
1
1
4
1.2.2
NATIONAL AND INTERNATIONAL
Ramsar
on wetlands.
Agenda
1.3
RESEARCH OBJECTiVES................. ..........................................
7
4
5
CHAPTER 2: BACKGROUND STUDY
2.1 2.1.1
2.1.2
1 .1
1 .2
2.1.2.3
2.1.2.4
2.1.2.5
WETLAND FUNCTIONS AND INFLUENCING FACTORS .................. .
Overview of important wetland
Overview of the important
Climate
Drainage
Hydrology
Geomorphology
Wetland
REHABILITATION AND RESTORATION WETLANDS ................. .
2.2.1 Aims
of wetland rehabilitation ...................................... .
2 1.1 Rehabilitation
................... " ..................................... .
2.2.2 Rehabilitation measures.............................................................
2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.4.2
2.3.5 ENVIRONM
INDiCATION....................................................
Background ..............................................................................
Indication selection criteria .........................................................
The DPSIR framework ................................................................
Studies on environmental indication in South Africa ......................
Inland
systems.................................................................
Indication of wetlands using
sensing data ..........................
2
1 Image
techniques known to
suitable for wetland
monitoring .................................................................................. 2.3.5.2 Wetland inventories .................................................................... ..
...................................................................... .. 2.3.5.3 Wetland
2.3.6 Selection of indicators................................................................
8
8
8
8
8
8
9
11
11 11 11 12 13 13 15 17 20 20 20 21 21
24 28 CHAPTER 3: MATERIAL AN
3.1.
3.1.1
3.1.1.1
3.1.1.2
1.1.3
1.1.4
3.1.2
3.1
3.1
3.1
1.2.4
1 .5 3.1
3.1
3.1 1 1.3.2
1.3.3
3.1.3.4
3.1
3.1
3.1.4
1.4.1
1.4.2
3.1
3.1
3.1.4.5
3.1
1.5 .6 1 1
3.1
3.1.7
3.1.7.1
1.7.2 1
3.1.8
3.1.8.1
METHODS MATERIAL AND METHODS........................................................
Study areas.............................................................................. Selection criteria for the different wetlands .................................... .. lal"'tarl rehabilitated wetlands
2) .................................... .. Rehabilitated wetlands site description (Table 5) .............................. .
..........................................................
Identification of test
General literature, maps and other baseline data .......................... .
Kromme
Wetland ............................................................... .
Mbongolwane Wetland ................................................................. . Wilge River Wetland .................................................................. ..
Wetland .....................................................................
Wetland .............................................................................
Rietvlei Wetland ...........................................................................
Identification
suitable indicators............................................ ..
Kromme River Wetland ............................................................... .
Mbongolwane Wetland............. .................. .............................. .
Wetland .....................................................................
Seekoeivlei Wetland ................................................................ ..
Zoar Wetland ............................................................................
Rietvlei Wetland .......................................................................
Remote
sensors
.............................................
Kromme River Wetland .............................................................. .
Mbongolwane Wetland ................................................................ .
Wilge
Wetland ................................................................ ..
Wetland ....................................................................
Wetland .........................................................................
Wetland ...........................................................................
Identification of optimal time frame for data acquisition ..................
Image processing ......................................................................
Classification and identification
ind
classes........................... .
.......... ..... .......... . ..................... . ......... .
Description
Calculation
the efficiency........................................................
data collection ................................................................... .
Cost estimation of all evaluated sensors .......................................... .
Remote sensor data .................................................................. .
Validation of data.......................................................................
Field assessment. ......................................................................
35 38 40 41 41 69 76 89 100 119 119 119 119 120 120 1
121 121 122 122 1
1
1
123 124 124 1
127 1
127 130 131 131 HAPT R 4: RESU
4.1
4.1.1
4.1.1.1
RESULTS ..................................................................................
Which indicators were detectable with which sensor type ...............
................................ '" ......... .
ETM and Landsat TM
1
133 133 1
SPOT ...............................................................................
EROS .......................................................................................
KODAC
(Near infrared) ...........................................................
Duncan
CIR........................................................... .. ....... .
Comparison of the various
sensors ..................................
420 .................................. ..
Comparison of SPOT 5 and Kodak
Comparison of
and DuncanTech
.....................................
155 4.2
LIMITATIONS AND RESTRiCTIONS ...............................................
157 1.1.2
1.1.3
1.1.4
4.1.1.5
4.1.2
1.2.1
134 137 142 146 152 1
CHAPTER 5: DISCUSSION
1
5.1.1
DiSCUSSiON...... .......... ............. .................................................
Description of selected indicators................................................
158 158 RECOMMENDATION.............. ........ .............................................
159 CONClUSiON...........................................................................
161 REFERENCES 162 APPENDIX 1: Abbreviations.
of terms.
APPENDIX 2:
average rainfall
for wetland
APPENDIX 3: Rehabilitation measures.
APPENDIX 4:
visit and progress
ronr'rT
APPENDIX Interpretation of aerial photographs depicting
changes at the
Hudsonvale peat basin in the Kromme River Wetland.
APPENDIX 6:
I )ifforant
remote sensor imagery for the wetland study areas
maps).
LIST OF TABLES Table 1:
Table 2:
Table 3:
Table
Table 5:
Table 6:
Table
8:
Table 9:
Table 10:
Table 11:
Table 12:
13:
Table 14:
Table 15:
Table 16:
Tahlp- 17:
18:
Table 1
Driving Force-Pressu
framework for rehabilitated
wetland
(adopted from Nell et , 2001).
South African national environmental indicators
inland
(Muller
Pretorius,
Summary of indicators for the monitoring of rehabilitated wetlands
applications.
with the use of remote
Summary
Vegetative Indicators by
Zone (Land-use and
Wetland 1
rian Habitat Working Group, 2001).
mmary information of each wetland chosen.
Kromme River
Summary
the problems, rehabilitation
results to
attained
actions taken in 2001/2002 and the
after the rehabilitation.
Mbongolwane Wetland: Summary of the problems, rehabilitation
results to
attained
actions taken in 2002 and
bilitation.
River Wetland: Summary
the problems, rehabilitation
taken and
desired
to be
after
rehabllitation (Collins
Thompson,
Diagnostic
of the
plana Agrostis lachnantha
Wetlands.
Wetland: Summary
problems, rehabilitation
actions taken in 1
and the
results to attain after
rehabilitation (Collins
Thompson, 1
Zoar Wetland: Summary
the problems, rehabilitation actions
and the
results attain after
rehabilitation.
problems, rehabilitation actions
Wetland: Summary of
and the desired results to attain after the rehabilitation.
"Window of opportunity" for image acquisition.
Remote sensor imagery covering the rehabilitated area around the
structures.
estimation
evaluated sensors.
acquisition
18nris8t TM 8nd
ETM+
SPOT characteristics.
Evaluation summary
various
sensing sensors using
indicating user
categories of Good, Medium and Poor
friendliness,
Medium,
processing time as
as High, Medium and Low o11"'::'''ti\
indication.
19
20
31
38
49
74
78
82
109
124
1
1
130
133
1
1
RES
LIST
Figure 1:
Figure 2:
Figure 3:
"Driving
framework.
study
The location
in South
Africa.
The location of the
rehabilitated
within the different
units and the wetland types (Adapted from Kotze et
wetland.
18
41
of the Kromme valley with 1954 aerial
fan at
Figure
Figure 7:
8: Figure 10:
11 :
12: 13: 14: 15:
Figure 16:
Figure 17:
18:
Figure 19:
20:
Figure 21:
Figure 22:
43
erosion exposing
Horizons of dark
organic,
with
peat-like material,
material) reveal
coloured sands (relatively poor in
tributary catchments.
episodic nature of the sediment supply from
off to prevent over­
The area around
structure must be
grazing and trampling by cattle.
Topographical
(3324 CD,
DO) showing
structures 1- 5 as well as the headcut
locations of
of the Kromme Wetland in the Eastern
Landscape around
headcut at a tributary
Kromme River.
This rehabilitation was done in a first
and not in a
wetland.
Close-up photo
head cut rehabilitation structure at a tributary
the Kromme
This rehabilitation
will not form
the Wilge
Wetland study but
rehabilitation measures
in the catchment.
1: Concrete Weir being constructed.
Rehabilitation
2:
being constructed.
Rehabilitation
3: Concrete Weir
constructed
Rehabilitation structure 4: Gabion weir Kompanjiesdrif.
Rehabilitation structure 5: Gabion weir Hudsonvale. Note
revegetation on the sides.
Part
the Mbongolwane wetland with cattle grazing on the wet
grassland
harvested sugar cane in the foreground.
A
(Colocasia esculenta) referred to by the
people as
Ikhwane for mats and
depicting
Site 1: Amatigulu. Aerial
"""'......"'v changes.
Site
Aerial photographs depicting landscape
activities which impact on Mbongolwane Wetland: Sugar
, cash crops, washing clothes.
47
48
50
50
51
51
52
52
55
58
59
Figure 23: Topographical maps (2831 CC &
CD) showing
locations of
the two rehabilitation structures (Uvova and Amatigulu).
24: Amatigulu
July 2002
View of the
rains.
floods. Note the lush green
Figure downstream on 4 July
at the
Before flooding. Photo
Amatigulu
26: After flooding. Photo taken downstream after July 2002 at the
(Photo: M.
Amatigulu
Geocell chute preparation at 1
Astrup ).
Before
Photo taken upstream on 4 July 2002 at the
Amatigulu
After flooding.
taken
after July 2002
Amatigulu
site.
preparation
gradient (Photo: M. Astrup).
29: Before flooding. Photo taken on 4 July 2002 at
site.
re 30: Uvova
damage to
preparation
July 2002 floods
(Photo: M Astrup).
31: Reparation work on the Uvova site was done during September
October
(Photo: M. Astrup).
32: on the Uvova
The stream
headcut
Photo taken on 4 July 2002.
Mbongolwane
33: The
River Wetland
rehabilitation
the head cut.
ure 34: Wilge
Wetland is one of the sites were
Whitewinged
(Taylor
Flufftail occurs regularly
in significant
Grundling, 2003).
Figure 35: A
(consisting of a headcut and a channel) were found in
Wilge
The remote
imagery will focus on the area
indicated.
ure 36: Road
downstream
rehabilitation
ure 37: Topographical map
showing the
of
rehabilitation structure.
Figure 38:
up of the headcut rehabilitation structure.
Figure 39:
wetland
in the foreground.
40:
wetland near Memel.
Figure 41:
Merel's
This erosion
forms the main flow
Klip River.
Figure 42: Topographical map
showing
locations of
rehabilitation structures.
ure 43:
no. 1. Area north
confluence of the drainage channel that
across the width
the floodplain.
from east to
Figure 44:
no. 1. Concrete
Figure 45:
no.
Figure 46:
no. 2. Bank
62
63
64
64
68
72
74
78
81
84
85
85
Figure
no. 3.
1997/8.
erosion.
. The gabion
were put in place to try
was constructed in
stabilise the bank
Figure
the
vlei. A small weir
Site no. 4.
surface.
no. 5. Gabion
no. 6.
work involved lifting the
structure higher,
strengthening the structure and closing
drain holes (1997). Note
most of
have been washed out.
Figure
border of the Seekoeivlei Nature Reserve. A
no. 7. The
gabion structure was constructed in the flow of the Klip River to
of the floodplain
to
water
promote the
Figure
The Zoar wetland in a Mondi Forestry area.
taken on 3 July
2002 after a veld
F
m
Drain in the wetland before rehabilitation (80 m long and
deep). Photo
from Mondi Wetlands
report (2000).
Artificially
channel before rehabilitation. Photo
from Mondi Wetlands Project report (2000).
Photo of
wetland after the
trees were removed
from the wetland
Lindley).
56: View of a flooded
wetland.
plugs
walls (Photo:
Lindley).
ure
Topographical map (2630CD) showing
locations
rehabilitation structures.
ure
Site no 1:
plugs in a badly
channel at the
border with
owners.
natural channel.
59: Site no
controlled by Mondi Forestry.
60: Site no
61: Site no
draining the
the desiccation
wetland.
Figure 62: Site no. 5:
with cUlverts.
culverts concentrate the flow
and cause eroded channels.
of the
Figure 63:
act
Figure 64:
Figure 65:
Figure 66:
Main drain with smaller drains along the
Clay plugs
walls.
the drain in the burned tributary area that feeds into
Site no
Headcut
The
wetland. Previous
the foreground of the wetland.
with
mining activities can
88
89
90
90
91
96
96
97
97
98
98
99
owners.
seen in
102
Figure 67:
Figure 68:
Figure
Figure 70:
Figure 71:
Figure
Figure
Figure
Figure 76:
re
ure 80:
81:
Figure 82:
83:
84:
photograph (2000) ind
remnants of a
responsible for the desiccation of the wetland. Transects
were
through the Phragmites
to delineate the wetland
rehabilitation could commence. Note the headcuts (C) & (D),
irrigation furrow (E), shallow
feature (F) and donga (G).
no. 5.
area between site no. 4
the area before rehabilitation was
photograph (2000)
no. 4 - 5.2.
the
Topographical map
showing the
rehabilitation structures.
was built across the
no. 1. A man-made
Rietvlei Nature Reserve.
to
the floW of
concentrated.
drift to
were constructed
no. 1. Energy
spread the flow and d
the energy.
with grass
At
no. 2.1 the
low erosion gully was
no.
damming behind the earth
Re-vegetation of
areas is
end of a shallow
taking place down-and
the structure.
1. The
Headcut and
drainage channel)
donga sides
to
by a
on the road crossing were
Site no. 4. Eroded cu
concrete weir.
furrow used during
Site no. 4. A
causeway was blocked to
concrete weir in
erosion.
Site no. 4. A causeway was applied as a structure to disperse water
and to raise the
in the wetland. Downstream of the
structure the erosion was stabilised by using rock
constructed to allow the
1. A
Site no.
construction
weir to
of the
table and to
water.
Construction
a gabion in progress at
no. 1.1. Flooding is
taking place.
ind
structure.
Site no. 5.1.2. Sedimentation taking place
in the water table
Gabion structure
no. 5.2. Note
upstream
downstream of the structure.
the higher water
behind the structure.
bank of the channel
a previous flooding
event.
A drowned
structure between
increase in
flooding from the
no. 6
103
106
107
108
111
111
112
112
113
113
114
114
115
115
116
116
117
117
Figure 85:
Figure 86:
Figures
87 - 90:
Figures
91 - 96:
Figures
97 - 101:
Figures
102 - 107:
Figure 108:
Figures
109 - 112:
Figures
113 - 117:
Figure 118:
Figures
119 - 124:
Figure
125:
Figures
126 - 129:
Figures
130 - 135:
Figures
136 - 139:
Figures
140 - 146:
Site no. 6.3. The bypass furrow must be closed and the energy must
be dissipated to prevent the concentration of water resulting in
channel erosion.
Spectral profile for SPOT 5 Imagery.
Landscape 1 (Folder: PHASE 5 A3 size layouts)
SPOT: Kromme River wetland - site 5.
Landscape 2 (Folder: PHASE 5 A3 size layouts)
EROS example 1: Mbongolwane wetland - Uvova.
Landscape 3: (Folder: PHASE 5 A3 size layouts)
EROS example 2: Seekoeivlei wetland - site 3 and site 4.
Landscape 4: (Folder: PHASE 5 A3 size layouts) (A3).
EROS example 3: Seekoeivlei wetland - site 5.
Reflectance values of KODAK 240 (Near infrared).
Landscape 5: (Folder: PHASE 5 A3 size layouts)
KODAK 240 (Near infrared) example 1:
Kromme River wetland - site 5.
Landscape 6: (Folder: PHASE 5 A3 size layouts)
KODAK 240 (Near infrared) example 2 :
Kromme River wetland - site 1 and site 4.
Spectral Profile for DuncanTech CIR.
Landscape 7a: (Folder: PHASE 5 A3 size layouts)
DuncanTech CIR example 1: Mbongolwane wetland - Uvova.
Landscape 7b: (Folder: PHASE 5 A3 size layouts)
DuncanTech True colour example 1: Mbongolwane wetland - Uvova
Landscape 8: (Folder: PHASE 5 A3 size layouts)
DuncanTech CIR I True colour example 2: Zoar wetland.
Landscape 9: (Folder: PHASE 5 A3 size layouts)
DuncanTech CIR I True colour example 3: Rietvlei wetland.
Landscape 10: (Folder: PHASE 5 A3 size layouts)
Comparison of SPOT 5 and Kodak DCS 420 images.
Kromme River wetland - site 5.
Landscape 11 :(Folder: PHASE 5 A3 size layouts)
Comparison of EROS and DuncanTech CIR images.
Mbongolwane wetland - Amatigulu.
118
134
136
138
139
141
143
144
145
146
148
149
150
151
154
156
Summary
Title:
Evaluation of remote sensing sensors for auditing and monitoring of
rehabilitated wetlands.
Student:
Althea Theresa Grundling
Supervisor:
Prof. G.J. Bredenkamp
Department:
Department of Botany
University of Pretoria
Degree:
MAGISTER SCI
Date:
10 March 2004
This study contributed to the wetlands.
7303)
a procedure for monitoring rehabilitated
measurable
were identified that with the
application of remote sensing techniques to monitor the impacts
measures on selected wetlands,
erosion, sedimentation,
area, water quality, wetland
terrestrial vegetation, alien vegetation,
, disturbances (e.g. cultivation)
rehabilitation structures. A
use of different
The high resolution
•
•
rOrnnl"o
sensing sensors
Kodak DCS 420
recorded sensors
was recoded into seven classes by using image
indicators identified, namely:
Erosion /
soil I
were:
Infrared)
5).
TM and Landsat
was done for all
ized land cover
1: overview
their capabilities, limitations as well as
Airborne sensors
A
water, wet
wetland vegetation,
sedimentation,
open water,
1
study areas. The
...,....+,.... h,.,....
and the
Class 4:
wetland vegetation that
Class 5:
terrestrial vegetation / bum Class 6:
alien vegetation and Class 7:
cultivation. hydrological conditions of the wetland, The exact location of the rehabilitation structures was recorded using a GPS. Ten
in the wetland study areas, the the eleven selected indicators were
exception being
mapping these indicators are the quality.
resolution to produce accurate optimum time
maps versus
data. The resolution of the data plays a vital
on what the objective of the mapping
the images, but the best results were from The
the Kodak DCS 420 Near
CIR images. rehabilitation
To map and monitor the
resolution 1 m or
it
erosion activity,
structure and
gully erosion. For mapping
the data should
of structural damage, movement of headcuts and multispectral data with band width 0.52 to of ground resolution 1.8 m or better. However, !Am is of great
indicators must be monitored over time. In order to monitor rehabilitated wetland vegetation over a longterm
season but from different
the analysis of
in
sensing sensors will
rehabilitation
, the compatible images must represent the same It is recommended that future possible studies include dynamics linked with the hydrology to investigate the change rehabilitation. The choice between the different remote depend on the application of the sensor, or the vegetation response to the rehabilitation measures. 2
Fly UP