Day 4

Detailed paper information

Back to list

Paper title Integrating satellite SAR time series and hydrological modelling to inform of soil moisture at the field scale
Authors
  1. Belen Marti University of Surrey Speaker
  2. Irene Seco University of Surrey
  3. Clement Atzberger BOKU - University of Natural Resources and Life Sciences Vienna
Form of presentation Poster
Topics
  • A7. Hydrology and Water Cycle
    • A7.01 Inland Water Storage and Runoff: Modeling, In Situ Data and Remote Sensing
Abstract text This study developed a method to derive field-specific SM information (as opposed to the large-footprint existing products) in near-real time by leveraging synergies of hydrological models and Earth observation (EO) data, both from SAR and optical sensors. The two components are further complemented by EO near-real time information on meteorological fields for drivers of precipitation and evapotranspiration. While the strength of the soil hydrological models consists of a physically based description of the rain infiltration and percolation processes, the satellite-based data permits to derive vegetation canopy properties at the field scale, and to obtain forcing variables such as precipitation and potential evapotranspiration to feed the models.
For several fields in the COSMOS UK soil moisture monitoring network, we retrieved time series of Sentinel-2 NDVI and Sentinel-1 backscattering values. We used the Hydrus-1D modelling tool for the simulation of surface and in-depth SM at the study fields at daily time steps during periods of low vegetation (NDVI < 0.25). The model’s upper boundary conditions were given by the time series of satellite-based estimates of precipitation and evapotranspiration. The lower boundary condition was set as free drainage, assuming that the water table is deeper than the root zone and the soil is well drained.
For C-band SAR and for the Sentinel-1 range of incidence angles, the literature reports an approximate linear relationship between the average VV backscattering coefficient of uniform, bare-soil crop fields and the surface soil moisture. For individual fields during the bare soil stage, it can be assumed that the main cause of Sentinel-1 VV backscattering changes is surface moisture, more rapidly variable than the surface roughness. Then, the fields’ soil hydraulic conductivity and other infiltration descriptors were obtained by optimizing the temporal trends of the modelled surface moisture to the temporal trends observed in the Sentinel-1 VV backscattering during low vegetation periods.
The soil moisture simulated in this way was compared to the moisture measured at the COSMOS fields at different depths. Our method achieved an excellent accuracy, only drifting away from the measured values at the end of the cropping cycle, after harvesting.
This work was carried out in collaboration with Mantle Labs Ltd. And received funding from the UK Research and Innovation SPace Research & Innovation Network for Technology (SPRINT) programme. By deriving valuable soil moisture information at the field level, Mantle Labs intends to offer enhanced drought related insurance products which can be made available to smallholder farmers. This index insurance will protect farmers against crop loss occurring due to extreme weather events.