DSD-INT 2018 Can we combine satellite derived Soil Moisture with hydrological models - Schellekens

Presentation by Jaap Schellekens (Vandersat) at the wflow User Day 2018, during Delft Software Days - Edition 2018. Friday 09 November 2018, Delft.

1. Can we combine satellite derived Soil Moisture with
hydrological models?

2. VanderSat
●Founded in 2015
●10, 14 23 employees (scientists ( > 50 % has a phd in
EO) and entrepreneurs)
●Transition from startup to SME
●Located in Haarlem
●Commercial EO services
●Prime focus on Microwave data and Soil Moisture

●Mission:
Build the best satellite products to solve the
global water and food crisis and always
keep innovating.

3. Soil moisture and hydrology
1.Soil moisture
●Indicator for drought
●Indicator for floods
●Derive discharge rating from soil moisture
●Input to models
2.Surface water
●Identify water seasonality
●Post flood analysis
●Flood monitoring/risk

The state of the system (now) is the
most important factor in determining
the quality of modelling results ->
Soil moisture helps a LOT in doing
that!
Input to models

Monitoring
Also VOD and Temp

4. Why Soil moisture?
1.Represent a store (not a flux)
2.Indicator for floods and droughts
3.Check the relation between Q and soil
moisture and Q and precipitation

5. Satellite Soil Moisture
Case study of a catchment in Australia, Van der Schalie et al., 2015: RSE

6. 7
Data architecture
<12 hr for L1

(7d for L4)
Downscaling //
LPRM //
Corrections //
Flagging //
L1
L3
+ 2 hr
Viewer //
API connected //

< 14 hr

7. https://maps.vandersat.com/

8. Success parameters….
1.The soil moisture to be used should:
a.be of sufficient quality and preferably
come with defined errors
b.have sufficient spatial and temporal
resolution
c.operational product
2.The model should be able to properly
represent (top) soil moisture
a.Sometimes derived parameters
(discharge) can be used
3.The model should be able to handle the
new input data

9. VanderSat Soil Moisture VS Cosmic Ray

10. VanderSat Soil Moisture VS Cosmic Ray

11. VanderSat Soil Moisture VS Cosmic Ray

12. Calibration with Soil
Moisture
1.Demonstrate here in PCR-GLOBWB
2.Can be done globally

13. Assimilation with Soil
Moisture
1.Demonstrate here in PCR-GLOBWB
2.Performance close to a locally calibrated
model

14. Assimilation of Q derived from
satellite soil moisture in
wflow_sbm

15. Mozambique
1.Used layered wflow_sbm to
match different zones
2.Large uncertainties in:
a.Precipi
b.Soil
c.but also the soil
moisture
3.When it matches we have
more certainty

17. Conclusions
1.Satellite soil moisture is available at NRT globally
(between 2 and 7 times per week)
2.New layered wflow_sbm can be configured to match the
satellite observed soil moisture
3.For models that do not have a surface soil moisture store
we can use a Derived Root Zone Soil Moisture or
discharge derived from soil moisture/wetness
4.Improvement is large for poorly or uncalibrated models
and smaller for well calibrated models
5.If a model is well calibrated but for the wrong reasons
adding soil moisture will show this -> recalibration
required

18. [email protected]
www.vandersat.com