Surface Soil Moisture from Multi-Mission Satellite Radar Altimetry
Berry, Philippa A. M.1; Dowson, Monica2; Carter, John3
1EAPRS Lab, UNITED KINGDOM; 2Arnesbury Ltd, UNITED KINGDOM; 3Queensland Climate Change Centre of Excellence (QCCCE), AUSTRALIA
Soil surface moisture is one of the key variables for climate systems; remote sensing currently provides a range of estimates from both passive and active sensors. However, modelling is required to relate the measured parameter to soil moisture, and this introduces uncertainties into the estimates. Satellite radar altimetry offers a unique perspective on this issue. Because these instruments are nadir-pointing, measurements from different passes can be combined, and the dense spatial sampling of the unique ERS1 Geodetic Mission can be exploited to yield detailed models of surface brightness, by reconciling and regressing the measurements to give a model as close to zero soil moisture as possible. This technique can be utilised over desert and semi-arid terrain globally. After cross-calibrating the brightness measurements from different missions over land, the differences between this model and multi-mission backscatter measurements can then yield time series of surface soil moisture. Because the complexity of the surface response (high spatial variability in surface roughness and composition) is vested in the model, the link between sigma0 difference and soil moisture estimate is greatly simplified. This paper reviews the model reconciliation and estimates performance for desert regions, using data from ERS1, ERS2 and EnviSat, demonstrating that data reconciliation can be achieved to better than 0.3dB if the extensive paleo-hydrological features found in many desert regions are excluded. Soil surface moisture time series are derived and examined for multiple deserts. Validation over desert regions is presented, and the altimeter derived values are shown to be in excellent agreement with ground truth. The impact of different along-track sampling rates and range window widths on the altimeter derived soil moisture estimates are examined, and the current capability and future potential of this technique is evaluated with reference to Sentinel3.