Exploring Approaches to Downscaling and Bias Correction of Smos and Smap Soil Moisture in Great Britain Using Reconstructed Climatology and the Cosmos-Uk Network

Satellite-based soil moisture monitoring has matured, notably as a result of the Soil Moisture Ocean Salinity (SMOS) and Soil Moisture Active Passive (SMAP) missions. Whilst spatial resolutions fall short of those ideal for field-scale monitoring, there are several 'value-added' RS soil moisture products available at the regional (e.g. SMOS: 25km, SMAP: 36km) to meso-scale (e.g. SMAP: 9km) resolution. Challenges arise as sensors only measure surface moisture (0-5cm), but frequent repasses, global coverage and low latency offer other advantages. Evaluation against ground-based measurements in the USA and elsewhere is encouraging, but little evaluation has been undertaken in Great Britain. In an initial phase we evaluated SMAP data against 28 COSMOS-UK sites for an approximately 18 month period (~2015-Dec 2016). The cosmic ray sensors integrate soil moisture over an area of ~12ha, and while not matching the spatial scale or soil depth of satellite measurements, do at least avoid some of the field-scale heterogeneity issues associated with point-based measurement. COSMOS-UK sites were installed from 2013 onwards, so records are short and a consistent temporal analysis was not possible. Time series were correlated and evaluated using the Pearson and Spearman's correlation coefficients, unbiased root-mean square error and mean bias. Reasonable time series correspondence was noted for many sites, but the majority showed a notable dry bias. In a second phase, we evaluate a wider variety of SMOS and SMAP products at various scales against ~40 COSMOS-UK sites for a longer period (2015 to present). Additionally, we use a new reconstructed historic (1960-2015) gridded monthly dataset of average monthly soil moisture (1km x 1km) to downscale satellite derived soil moisture and implement a bias-correction scheme based on this modelled climatology. We repeat the correlation analysis to evaluate performance of the approach and explore site-based differences in performance.