Tho VISCHEL, Geoff PEGRAM, Scott SINCLAIR

1
Remote Sensing and Hydrology in Southern Africa
Théo VISCHEL, Geoff PEGRAM, Scott SINCLAIR Cork,
3rd April 2008
2
SOIL MOISTURE FROM SATELLITES Daily maps over
Republic of South Africa
Soil moisture estimation
1. Satellites
SAHG products ? Temperature gradient IR
(MeteoSat)
TUWIEN products ? Scatterometer (ERS1-2, MetOP)
? ASAR (ENVISAT)
2. Local probes
3. Hydrological modeling
Comparison with satellite estimations
3
OVERVIEW

• Hydrological modeling
• 2. From conceptual catchments
• 3. to a real catchment
• 4. Results
• 5. Conclusion and discussion

4
1.
Hydrological Modeling
5
1. Hydrological modeling
TOPKAPI Model (Liu and Todini, 2002)
Modeled processes

• 1. Soil
• Infiltration
• all the water infiltrates until soil saturation
• ? preferential paths
• Transfer

• 2. Overland flows
• Production
• exfiltration or production on saturated areas
• Transfer

• 4. Channel flows
• Transfer

Infiltration
Saturation
Exfiltration
Transfer
6
1. Hydrological modeling
TOPKAPI Model (Liu and Todini, 2002)
Modeled processes

• 1. Soil
• Infiltration
• all the water infiltrates until soil saturation
• ? preferential paths
• Transfer

Distributed

• 2. Overland flows
• Production
• exfiltration or production on saturated areas
• Transfer

• 4. Channel flows
• Transfer

5. Evapotranspiration
7
1. Hydrological modeling
TOPKAPI Model (Liu and Todini, 2002)
Modeled processes

• 1. Soil
• Infiltration
• all the water infiltrates until soil saturation
• ? preferential paths
• Transfer

Distributed
Distributed
Physically based

• 2. Overland flows
• Production
• exfiltration or production on saturated areas
• Transfer

• 4. Channel flows
• Transfer

5. Evapotranspiration
8
1. Hydrological modeling
TOPKAPI Model (Liu and Todini, 2002)
Distributed
Distributed
Physically based

• Continuity of mass
• Generalized Darcy equation

• Continuity of mass
• Mannings equation

Non-linear ODE
9
2.
From conceptual catchments
10
2. From conceptual catchments
Mathematical and numerical aspects
11
2. From conceptual catchments
Mathematical and numerical aspects
0
1
2
3
4 cells
Stable Fast
1. Runge-Kutta-Fehlberg
Unstable Very fast
2. Quasi-analytical solution
? Implementation of the core of the model in
Python
12
2. From conceptual catchments
Cell connectivity automation
13
2. From conceptual catchments
A new evapotranspiration scheme
Non saturated case
Saturated case
Evapotranspiration
Evapotranspiration
Overland
Gross infiltration
Gross infiltration
Soil layer
Lateral transfers
Lateral transfers
No evaporation of the overland flows ? Soil never
saturated

• During ?t
• Infiltrationtransfers

• During ?t
• Infiltrationtransfers

• At t?t
• Evapotranspiration

• At t?t
• Evapotranspiration

Liu and Todini version
14
2. From conceptual catchments
A new evapotranspiration scheme
Saturated case
Evapotranspiration
New version
Overland
Re-infiltration of overland flow ? Overland flow
evaporation

• During ?t
• Infiltrationtransfers

• At t?t
• Evapotranspiration
• Re-infiltration

15
3.
to a real catchment
16
3. to a real catchment
17
3. to a real catchment
Liebenbergsvlei (4625 km2)
18
3. to a real catchment
Liebenbergsvlei (4625 km2)

• Semi-arid climate
• - Mean annual rainfall
• 600-700 mm
• - Mean annual ETP
• 1400-1500 mm

South Africa
19
3. to a real catchment
Liebenbergsvlei (4625 km2)

• Landscape
• - Hills to steep relief
• - Grassland/Cropland

South Africa
20
3. to a real catchment
Liebenbergsvlei (4625 km2)

• Landscape
• - Hills to steep relief
• - Grassland/Cropland

South Africa

• Soil

Soil texture from WR90 (1994)
Soil type from SIRI (1987)
21
3. to a real catchment
Data availability
Liebenbergsvlei (4625 km2)
20 km
22
3. to a real catchment
Data availability
Protocol
Parameter adjustment
Comparison between modeled and remotely sensed
soil moisture
23
3. to a real catchment
Estimation a priori of the model parameters
Soil texture from WR90 (1994)
Soil type from SIRI (1987)
DEM from DLSI (1996)
Land use/Land cover from GLCC (1997)
Resolution 1km²
Calibration of the model is required
24
3. to a real catchment
Parameter adjustment methodology
25
3. to a real catchment
Parameter adjustment methodology
CALIBRATION

• Over one season 8 months

Initial soil moisture

• Based on the subcatchment flows

• At a 6 hours time step

26
4.
First results
27
4. First results
CALIBRATION
Season 1
2

• Initial soil moisture 40
• fac_Ks 100.
• Ks 0.3-2 m/h

Lateral transfers in preferential paths
28
4. First results
CALIBRATION and VALIDATION
Season 1
Season 2
2
Satisfactory flow simulations ? Soil moisture
1
2
29
4. First results

• Comparison between
• Modeled soil moisture
• Remotely sensed soil moisture

• Product from TUWIEN

• Scatterometer (Radar) ERS-1 et ERS-2
• Active
• 5.3 GHz (C band)
• Resolution 50 km
• Global daily covering 40
• Repeat cycle at a point 7 3 days

30
4. First results

• Comparison between
• Modeled soil moisture
• Remotely sensed soil moisture

• Product from TUWIEN

1. Resampling onto a global grid
2. Dry and wet reference backscatter
3. Relative surface soil moisture
4. Profile soil moisture values using a
simple infiltration model
Soil Water Index - SWI 10 day intervals
31
4. First results

• Comparison between
• Modeled soil moisture
• Remotely sensed soil moisture

• Product from TUWIEN

• Soil Water Index - SWI

• TOPKAPI simulations

• Soil Water Index - SWI

32
4. First results
At catchment scale
Season 1
Season 2
R20.780
R20.922
33
4. First results
At catchment scale
Season 1
34
4. First results
At catchment scale
3 months of wetting Philip a touch of
Hysteresis?
Season 1
35
4. First results
At footprint scale
36
4. First results
At footprint scale
37
4. First results
At footprint scale
Season 2
Season 1
38
5.
CONCLUSION DISCUSSION
39
5. Conclusion discussion
Two different approaches of soil moisture
estimation Hydrological modeling and Remote
sensing
Good correspondence!
Very encouraging results for
Hydrological modeling
Remote sensing

• Use of remote sensed soil moisture
• - Validation
• - Initialization
• - Assimilation

• Use of hydrological models
• - Validation
• - Disaggregation

40
5. Conclusion discussion
Why such good results? Peters-Lidard, C.
(EGU, 2007)

• The SWI (Soil Water Index) is considered
• ?Relevance of lateral transfers

• Scatterometer better on low vegetated area
• ?Grassland/Cropland

• The remotely sensed SWI comes from a simple
  conceptual infiltration model
• ?Parameters suited to the studied region

41
5. Conclusion discussion
Next steps

• Explicit modeling of surface processes

42
5. Conclusion discussion
Next steps

• Explicit modeling of surface processes

Evapotranspiration
Evapotranspiration
Explicit Infiltration
Runoff
First 5 cm of soil
Gross infiltration
Gross Infiltration
Lateral transfers
Lateral transfers
Future version
Actual version

• Runoff/Infiltration partition
• ? Infiltration excess runoff

43
5. Conclusion discussion
Next steps

• Explicit modeling of surface processes

Evapotranspiration
Evapotranspiration
Explicit Infiltration
First 5 cm of soil
Gross infiltration
Gross Infiltration
Lateral transfers
Lateral transfers
Future version
Actual version

• Runoff/Infiltration partition
• ? Infiltration excess runoff

• 2. Raw satellite measurements
• From Soil Water Index
• to Surface Wetness Index