Hyperresolution Land Surface Modeling in the Context of SMAP Cal–Val

Abstract

his study explores the performance of Environment Canada?s Surface Prediction System (SPS) in comparison to in situ observations from the Brightwater Creek soil moisture observation network with respect to soil moisture and soil temperature. To do so, SPS is run at hyperresolution (100 m) over a small domain in southern Saskatchewan (Canada) during the summer of 2014. It is shown that with initial conditions and surface condition forcings based on observations, SPS can simulate soil moisture and soil temperature evolution over time with high accuracy (mean bias of 0.01 m3 m?3 and ?0.52°C, respectively). However, the modeled spatial variability is generally much weaker than observed. This is likely related to the model?s use of uniform soil texture, the lack of small-scale orography, as well as a predefined crop growth cycle in SPS. Nonetheless, the spatial averages of simulated soil conditions over the domain are very similar to those observed, suggesting that both are representative of large-scale conditions. Thus, in the context of the National Aeronautics and Space Administration?s (NASA) Soil Moisture Active Passive (SMAP) project, this study shows that both simulated and in situ observations can be upscaled to allow future comparison with upcoming satellite data.