Location: Hydrology and Remote Sensing Laboratory
Project Number: 1245-13610-028-75
Start Date: May 19, 2014
End Date: May 31, 2015
The project will address the soil moisture point-to-footprint upscaling problem by applying a triple collocation strategy to estimate and subsequently correct for the impact of random sampling error on comparisons between (satellite-based) SMAP soil moisture retrievals and point-scale ground observations. Triple collocation is a statistical technique commonly applied in the geosciences whereby the root-mean-square uncertainty of a single geophysical product is estimated through cross-comparisons against estimates of the same product acquired via two other independent means. Here the primary focus will be validating SMAP L2/3 soil moisture products using: 1) ground-based soil moisture observations obtained from a sparse ground-based network, and 2) gridded surface soil moisture products obtained independently from a land surface model. The analysis will be used to estimate the impact of spatial sampling uncertainty associated with using a single sparse ground-based observation to validate a time series of coarse-scale SMAP soil moisture retrievals. As a result, it will effectively broaden SMAP validation activities to fully leverage substantial investment in sparse ground-based instrument networks.