Assimilation of Remotely Sensed Evapotranspiration Data into a SoilVegetationAtmosphereTransfer Sche

1
Assimilation of Remotely Sensed
Evapotranspiration Data into a Soil-Vegetation-Atm
osphere-Transfer Scheme Initial Results

• Valentijn R.N. Pauwels, Claudia De Pus,
• Niko E.C. Verhoest and Francois P. De Troch
• Laboratory of Hydrology and Water Management,
• Ghent University, Ghent, Belgium
• Roeland Samson and Raoul Lemeur
• Laboratory of Plant Ecology,
• Ghent University, Ghent, Belgium

Estec, Noordwijk, The Netherlands
April 3, 2003
2
Objectives

• To assess the possibility to use CHRIS-PROBA data
  for the remote sensing of evapotranspiration
  rates
• To update the soil moisture state of a
  Soil-Vegetation-Atmosphere-Transfer Scheme
  through the assimilation of the remotely sensed
  evapotranspiration rates

3
Preliminary Results

• Assessment of the possibility to measure
  evaptranspiration rates under non-ideal
  conditions (sloping grassland)
• Assimilation of the in-situ observed latent heat
  fluxes and soil moisture values

4
In-Situ Observations
5
Comparison BREB-EC
Sensible Heat Flux
Latent Heat Flux
Av x 56.45 Av y 57.12 R 0.83 RMSE 46.27
Av x 15.85 Av y -22.77 R 0.35 RMSE 59.15
Eddy correlation-based (Wm-2)
Eddy correlation-based (Wm-2)
Bowen ratio-based (Wm-2)
Bowen ratio-based (Wm-2)
6
Diurnal Patterns Energy Balance Closure
Daily Total Net Radiation 0-4 MJm-2 4-8
MJm-2 8-12 MJm-2
Energy Balance Closure (Wm-2)
Time (GMT)
7
Model Simulations

• Hourly simulations at 1 m spatial resolution
• Fully process-based water and energy balance
  model
• Topographically driven lateral redistribution of
  soil water
• Assimilation of observed latent heat fluxes
  through extended Kalman filtering

8
Baseline Model Results
NET RADIATION
LATENT HEAT FLUX
Av x 65.77 Av y 69.43 R 0.99 RMSE 10.63
Av x 53.06 Av y 46.49 R 0.87 RMSE 35.07
Simulations (Wm-2)
SENSIBLE HEAT FLUX
GROUND HEAT FLUX
Av x 3.11 Av y 1.72 R 0.80 RMSE 9.61
Av x 19.57 Av y 26.58 R 0.72 RMSE 35.64
Observations (Wm-2)
9
Effect of Data Assimilation
UPPER LAYER
LOWER LAYER
BASELINE RUN
BASELINE RUN
Soil Moisture (-)
COMBINED ASSIMILATION
COMBINED ASSIMILATION
Time (Day in 2002)
10
Effect of Data Assimilation
NET RADIATION
LATENT HEAT FLUX
Av x 53.06 Av y 51.51 R 0.89 RMSE 31.65
Av x 65.77 Av y 70.21 R 0.99 RMSE 11.39
Simulations (Wm-2)
SENSIBLE HEAT FLUX
GROUND HEAT FLUX
Av x 3.11 Av y 1.63 R 0.81 RMSE 9.48
Av x 19.57 Av y 22.65 R 0.75 RMSE 31.65
Observations (Wm-2)
11
Preliminary Conclusions

• BREB and EC methods yield similar estimates of
  latent heat fluxes on the sloping grassland
• Assimilation at a low temporal resolution of
  in-situ observed latent heat fluxes leads to an
  overall improvement in model results