assessment of temporal and spatial effects of land use...

Helmschrot, FSU Jena, Germany W3M Conference for Wetlands, Wierzba, Poland, 22 - 25 September 2005

Assessment of temporal and spatial

effects of land use changes on wetland hydrology:

A case study from South Africa

Department of Geoinformatics,

Geohydrology and Modelling

FSU Jena, Germany

Jörg Helmschrot

The project has been funded by the DFG.


• Introduction

• Study Area

• Base Studies

• Wetland Classification

• Modeling Approach

• Model Results

• Conclusion

• Acknowledgements

OU

TL

INE

OU

TL

INE

[ 2/24 ]


1989 2010

BACKGROUND

� Study on landscape dynamics due to the temporal and spatial

impact of large scale afforestation on different wetland types

within the semi-arid headwaters of the Umzimvubu catchment,

South Africa.

WETLANDS IN SOUTH AFRICA

� are different and specific in terms of extent, functions, age,

biodiversity, genesis etc.

� need to be sustainable managed and conserved (National

Water Act 1998)

� key role within the hydrological system due to their importance for

water retention (summer) and low flow conditions (winter).

Integrated approaches are a prerequisite to understand

and assess wetland functioning within a landscape

perspective.

INT

RO

DU

CT

ION

INT

RO

DU

CT

ION

[ 3/24 ]


ST

UD

Y A

RE

AS

TU

DY

AR

EA

Nested Catchment ApproachUmzimvubu: 20000km²

Mooi: 306 km²

Weatherley: 1.2 km²

UpperCatchment

LowerCatchment

LEGEND:

TESTCATCHMENT WEATHERLEYNorth East Cape - Province, South Africa

Catchment boundary

Stream or seep

Crump weir

TransectsRock outcrop

Additional piezometer

Lower catchmenttensiometer nest andpiezometer

Full weather station andfog interceptor

Neutron Probe Tube

Upper catchmenttensiometer nest andpiezometer

Rain gauge andfog interceptor

W2

3

3

13

W2

W1

109

4321

8

6

5

7

123456

78

9

10

RO14L

3/4R

3R

4RR

3/4138A

14

15

121110R

10A

200 m

Scale

[ 4/24 ]


ST

UD

Y A

RE

AS

TU

DY

AR

EA

[ 5/24 ]

1200

1050

900

750

600

450

3001895 1920 1945 1970 1995

Jah

resnie

ders

ch

lag

[mm

]

Jahr

Maclear Mittlerer Jahresdurchschnitt

Mean

An

nu

alP

recip

itati

on

[mm

]

Year

Longterm averageMaclear

semi-arid conditions, high temporal and spatial

variability of rainfall, summer rainfall area, MAP: 750mm, MAT: 15,6 °CC

lim

at

e


ST

UD

Y A

RE

AS

TU

DY

AR

EA

triassic sediments of the Karroo Sequence, mainly

changing layers of sand- and mudstone, dolerite

dykes

Geo

log

y

high variability due to parent material, relief and

local climateSo

ils

626000

626000

626500

626500

627000

627000

627500

627500

655

750

0

65

575

00

655

800

0

65

580

00

655

850

0

65

585

00

655

900

0

65

590

00

#####

#####

# ## ###

###

###########

##

##

###

###

####

#########

##

######

###

########

# #

##########

###

#

#

Soils interpretation

map of the research

catchment Weatherley

Reference: ISCW, 1996 (changed)Projection: UTM, WGS 84

200 0 200 400 m

Duplex

Freely drainedFreely drained - signs of wetness

Hydromorphic

Marsh

Pale topsoilsands

Shallow

vegetationsampling plots

#

river network

catchment boundary

[ 6/24 ]

Lan

du

se

90.0

00 h

a a

ffo

resta

tio

n

� pinus and eucalyptus species

� short term rotations (8-24 y)

� est. water reduction by

higher interception (- 25 %)

Afforestation (since 1989)

� sweet and sour grass species

� extensively used for

stockfarming

� annual burning, drainage

Grassland (before 1989)


BA

SE

ST

UD

IES

BA

SE

ST

UD

IES

[ 7/24 ]

Intensive research has been done on several scales(since 1997)

• Rainfall dynamics by analysis of

rainfall data of 55 stations from 1970-

2003 (Bäse & Helmschrot 2003)Maclear

Port

St. Johns

30°28°

32°

EASTERN CAPE

Kokstad

Ell iot

30°

KWAZULUNATAL

LESOTHO

Mean annual rainfall*

• Land use change analysisutilizing Landsat TM/ETM, forestry

data, ECDB, field mapping between

1989-2002, wetland loss since 1989:

ca.17% (Helmschrot 2000, 2002)

• Landscape dynamics using

geophysical and sedimentological

methods (Hilbich & Helmschrot

2005)


BA

SE

ST

UD

IES

BA

SE

ST

UD

IES


• GIS-based relief analysis by proces-

sing DEMs to develop wetland distri-

bution and needed model parameters

(Dahlke & Helmschrot 2004)

[ 8/24 ]

• Plant growth modeling for pine/

eucalyptus plantation using 3PG-

model (Landsberg & Waring, 1997)

and GGLawn (Gruszszynski 2004)

for grassland areas (Helmschrot

2001, 2004)


BA

SE

ST

UD

IES

BA

SE

ST

UD

IES


• Soil mapping and analysis to

provide distribution and soil para-

meters such as grain size, pH,

hydraulic conductivity, pF, etc.

(ISCW 1999, Lorentz & Helmschrot

2000)

626000

626000

626500

626500

627000

627000

627500

627500

655

750

0

65

575

00

655

800

0

65

580

00

655

850

0

65

585

00

655

900

0

65

590

00

#####

####

#

# ## ###

###

###########

##

##

###

##

#####

#####

####

##

######

###

########

##

##########

###

#

#

Soils interpretation

map of the research

catchment Weatherley

Reference: ISCW, 1996 (changed)Projection: UTM, WGS 84

200 0 200 400 m

Duplex

Freely drainedFreely drained - signs of wetness

Hydromorphic

Marsh

Pale topsoilsands

Shallow

vegetationsampling plots

#

river network

catchment boundary

Gatberg Vlei, GV 1

0

1

2

3

4

5

6

7

8

0 10 20 30 40 50 60

Wassergehalt [%]

Sa

ug

sp

an

nu

ng

[p

F]

32cm

57cm

78cm

82cm

103cm

107cm

128cm

132cm

153cm

157cm

178cm

182cm

203cm

207cm

228cm

232cm

• Hillslope process studies to

identify, characterize and quantify

event-based dynamics of water

fluxes on slopes and within valley

cross profiles using HILLS9 and

HYDRUS2D (Lorentz 2002) 0

0,5

1

1,5

2

2,5

Ru

no

ff (

mm

/h)

0

10

20

30

40

50

Cu

mu

lative

Ra

infa

ll (m

m)

15:00 21:00 03:00 09:00

Observed HILLS9 Simulation Rain

L

1 m

h(x)

Qw

ku

Lower soil

Upper soil

kL

kh

[ 9/24 ]


BA

SE

ST

UD

IES

BA

SE

ST

UD

IES


• Vegetation studies in wetland

/non-wetland areas to provide

vegetation distribution patterns

and model parameters (Dahlke

& Helmschrot 2003, Helmschrot

2004)excessively well moderately poor very poor

0

20

40

60

80

100

so

il m

ois

ture

[v

ol

%]

Leersia hexandra

Monocymbiumceresiiforme

Pennisetumthunbergii

Juncus exsertusJuncus oxycarpus

grassland wetland

saturated

wet

Bulbostylisscleropus

Panicumnatalense

Alloteropsissemialata

Scleria welwitschii

Pycreus uniolides

Aristidajunciformis

shallow soils Cambisols aquic Cambisols gleyic soilsaquic leached

soils

dry

plant community

soil type

soil drainage

• Wetland characterization and classification by combination of

intensive field work in selected

reference wetlands,results of the

base studies and several classifi-

cation approaches (Helmschrot

2004)T1P1 T1P2 T1P3 T1P4 T1P5 T1P6 T1P7 T1P8 T1P9 T1P10

0

10

20

30

40

50

Tra

nsm

issio

n[%

]

0

20

40

60

80

100

Bed

eckung

sgra

d[%

]

0

20

40

60

80

100B

iom

asse

(o.L

.)T

rocken

masse

(u.L

.)[g

/0,0

4m

²]

0

10

20

30

40

50

Vege

tationsh

öhe

[cm

]

0

20

40

60

80

Bo

denfe

uch

te[v

ol%

]

100

80

60

40

20

0

Wassers

ättig

un

gdes

Bod

ens

[cm

Tie

fe]

1287

1288

1289

1290

1291

1292

1293

Ge

lände

höh

e[m

ü.N

N]

0 25 50 75 100 125 150 175m

WETLAND TZUPLAND UPL

WETLANDWETLANDUPL UPLAND

T4P1 T4P3 T4P5 T4P7 T4P9 T4P11 T4P13 T4P15 T4P17 T4P19

0

10

20

30

40

Tra

nsm

issio

n

[%]

50

60

70

80

90

100

Bedeckungsgra

d

[%]

0

20

40

60

80

100

120

Bio

masse

(o.L

.)

Tro

ckenm

asse

(u.L

.)

[g/0

,04m

²]

20

30

40

50

60

70

Vegeta

tionshöhe

[cm

]

10

20

30

40

50

60

Bodenfe

uchte

[vol%

]

100

80

60

40

20

0

Wassers

ättig

ung

des

Bodens

[cm

Tie

fe]

1292

1296

1300

1304

1308

1312

Gelä

ndehöhe

[mü.N

N]

0 50 100 150 200 250 300 350 400 m

TZ TZ TZWETL TZ TZTZ

[ 10/24 ]


WE

T C

LA

SS

WE

T C

LA

SS

[ 11/25 ]


WE

T C

LA

SS

WE

T C

LA

SS

[ 11/25 ]

Zo ne (ii) Zone (i) Zone (iii)


MO

DE

L A

PP

RM

OD

EL

AP

PR

[ 12/24 ]

PHYSICALLY-BASEDMODEL

(MMS/PRMS)

CLIMATE INPUT DATA

HRU(Scenarios)

Parameters(empirical, literature)

CATCHMENT WATER BUDGET

WETLAND WATER BUDGET


Land use

Geology

Soil

DEM

GIS Analysis

HR

UH

RU

-- AP

PR

AP

PR

GIS GIS –– ProcessProcess LevelLevel

shape, slope, aspect, length, area, topology, relief types, river network

vegetation, aET, sealed area, LAI,roots, interception, wetlands

pore volume, field capacity,

texture, unsat. conductivity

storage capacity, texture, Darcy-coefficent, sink

RU RU –– Model ParameterModel Parameter

HRUs as homogeneousprocess-related modelentities

[ 13/24 ]


HR

UH

RU

-- AP

PR

AP

PR

Wetlands (VBW)

Bare rock/soil

Eucalyptus [ 14/24 ]

4..wetland

2..valley

3.. clay

2.. slope 0-5%

1..aspect N


TopologicalTopological HRUHRU--ROUTING (nROUTING (n��1)1)H

RU

HR

U-- A

PP

RA

PP

R

[ 15/24 ]


MO

DE

L A

PP

RM

OD

EL

AP

PR

c

MMS/PRMS (Leavesley, 1982,

USGS)• physically-based• modular structured• distributed function model

(HRU)

INPUT DATAINPUT DATA

PROCESSING and PROCESSING and

STORAGE FUNCTIONSSTORAGE FUNCTIONS

MODEL OUTPUTMODEL OUTPUT

[ 16/24 ]


0

1000

2000

3000

4000

5000

6000

7000

88/1

0/01

88/1

1/01

88/1

2/01

89/0

1/01

89/0

2/01

89/0

3/01

89/0

4/01

89/0

5/01

89/0

6/01

89/0

7/01

89/0

8/01

89/0

9/01

89/1

0/01

89/1

1/01

89/1

2/01

90/0

1/01

90/0

2/01

90/0

3/01

90/0

4/01

90/0

5/01

90/0

6/01

90/0

7/01

90/0

8/01

90/0

9/01

basin_cfs.basin runoff.obs

MO

DE

L R

ES

MO

DE

L R

ES

Mooi Catchment (305 km²) – grassland

1.10.1988 - 30.09.1990

[ 17/24 ]

Mean

Dail

y R

un

off

[cfs

]


MO

DE

L R

ES

MO

DE

L R

ES

Mooi Catchment (305 km²) – with plantations

1.1.1995 - 22.1.2002

[ 18/24 ]

10/1/95 2/1/96 6/1/96 10/1/96 2/1/97 6/1/97 10/1/97 2/1/98 6/1/98 10/1/98 2/1/99 6/1/99

0

40

80

120

160

200

100

80

60

40

20

0

runoff_obs runoff_sim

Mean

Dail

y R

un

off

[m

³/s

-1] D

aily

Pre

cip

itatio

n [m

m]


MO

DE

L R

ES

MO

DE

L R

ES

Weatherley Catchment (1.2 km²) – grassland

1.10.2000 - 30.09.2001

[ 19/24 ]

0

10

20

30

40

50

00/1

0/01

00/1

0/15

00/1

0/29

00/1

1/12

00/1

1/26

00/1

2/10

00/1

2/24

01/0

1/07

01/0

1/21

01/0

2/04

01/0

2/18

01/0

3/04

01/0

3/18

01/0

4/01

01/0

4/15

01/0

4/29

01/0

5/13

01/0

5/27

01/0

6/10

01/0

6/24

01/0

7/08

01/0

7/22

01/0

8/05

01/0

8/19

01/0

9/02

01/0

9/16

01/0

9/30

obs sim

Mean

Dail

y R

un

off

[cfs

]


MO

DE

L R

ES

MO

DE

L R

ES

Weatherley Catchment (1.2 km²) – with pine/eucalypte

1.10.1999 - 30.09.2001

0

5

10

15

20

25

30

35

40

99/1

0/01

99/1

1/01

99/1

2/01

00/0

1/01

00/0

2/01

00/0

3/01

00/0

4/01

00/0

5/01

00/0

6/01

00/0

7/01

00/0

8/01

00/0

9/01

00/1

0/01

00/1

1/01

00/1

2/01

01/0

1/01

01/0

2/01

01/0

3/01

01/0

4/01

01/0

5/01

01/0

6/01

01/0

7/01

01/0

8/01

01/0

9/01

runoff_grass.sim run_forest.sim

[ 20/24 ]

Mean

Dail

y R

un

off

[cfs

]

Results:

• available water will be reduced by forest plantations by

amounts ranging from 10.6% to 21.5%

• this reduction occurs as a result of higher interception and

evapotranspiration rates afforded by the afforestation.

• subsurface flow (SSF) will be noticeably more affected then

surface (SF) and groundwater flow (GWF), because of trees are

usually planted on hillslopes and available soil water on the slopes

will be taken up by the trees instead of generating interflow.

• Since surface runoff is mainly generated on bare soil/rock areas

and on grassland during intense rainfalls the SF is less affected

than on afforested grasslands.


MO

DE

L R

ES

MO

DE

L R

ES

Wetland-specific changes of water availabilityin Weatherley (prel.)

[ 21/24 ]

Results:

• wetlands total runoff losses (∆-Total) vary between 13.6% and

21.3%

• the total runoff loss of plateau wetlands (Plt), which are small in

size, range from 26.6% to 47.8%. In these wetlands, water input is

limited to rainfall, since water of planted surroundings is taken up

in plant growth

• medium-sized Slope wetlands (Slp) runoff loss varies between

11.1% and 19.9% and is mainly caused by reduced surface and

subsurface inflows from upslope areas

• Valley bottom wetlands (Val) are less affected (3.9-8.7% water

loss), since those wetlands are mainly controlled by groundwater

dynamics and modeled as saturated areas connected to the

stream.


CO

NC

LU

SIO

NC

ON

CL

US

ION

• MMS/PRMS has been successfully applied to analyse

and quantify the temporal and spatial impact of large scale afforestation on different wetland types within a

basin.

• forest plantation will reduce available water as a

consequence of higher interception considerably.

• the net loss of water due to increased evapotranspiration

varies with type and size of wetlands as well as the type

of water flow component.

• Additionally the modeling indicates that the wetland water

loss is strongly influenced by the surrounding area

(planted species).[ 22/24 ]

The study has shown that


AC

KN

OW

LE

DG

ME

NT

SA

CK

NO

WL

ED

GM

EN

TS

� German Research Association (Germany)

� United States Geological Survey (USA)

� Dr. Lorentz (University of Natal, South Africa)

� Forschungszentrum Jülich (Germany)

� VW-Association (Germany)

� National Research Foundation (South Africa)

� Mondi Forests Ltd. (South Africa)

Thanks are given to

[ 23/24 ]


[ 24/24 ]

Thank you very much for your attention!

Jörg Helmschrot

FSU Jena

Department of Geoinformatics,

Geohydrology and Modelling

Loebdergraben 32

07737 Jena, Germany

Ph: +49 3641 948858

Fax: +49 3641 948852

Eml: [email protected]

W³: www.geogr.uni-jena.de/wetlands

CO

NT

AC

TC

ON

TA

CT

assessment of temporal and spatial effects of land use...

Documents