Comparison of Upscaled Soil Moisture Measurements from a SMAP Validation Network using Six Techniques

English Abstract:

This research examines six upscaling techniques: thiessen diagrams (TD), kriging (K), soil weighted average (SWA), arithmetic average (AA), inverse distance weighted (IDW) and temporal stability (TS) in order to quantify variation between techniques when utilized for the validation of a passive microwave footprint. In situ soil moisture data for this study were collected south of Saskatoon, at an extensive soil moisture network operated by the University of Guelph and Environment Canada. This 50x50 km network has been qualified for the calibration/validation of the Soil Moisture Active Passive Mission (SMAP) which is to be launched in 2015. The data used in this research were collected during the 2013 and 2014 growing season (May to October) from 25 and 31 sites for 2013 and 2014, respectively. SMAP's instrumentation operates at an L-band frequency which has a soil penetration depth of ~5 cm. Therefore, in situ measurements were taken from a horizontally oriented Steven's Hydra Probe at a 5 cm depth. The analysis was only conducted on days with 100% functionality from all stations, resulting in 26 and 22 days for the 2013 and 2014 growing season, respectively. The TS analysis was completed using a complete eight month time period with the 2013 and 2014 data. The results indicated good agreement among all six upscaling techniques through both 2013 and 2014 time periods, however, as network coefficient of variation (CV) increased, so did the disagreement between the techniques. The increase in network CV was the result of heterogeneous precipitation events. Additional analysis will be conducted to test the sensitivity of each upscaling technique to missing station data and to determine if there is a limit on the network CV beyond which network data should not be used for calibration/validation purposes.