资源环境科技发展态势分析平台

To date, the direct use of remote-sensing soil moisture data sets for examining surface/atmosphere coupling strengths has been hampered by the presence of significant random errors and data gaps in these products. This study investigates the potential for obtaining an improved observation-based lower bound of summertime soil moisture-air temperature coupling strength via the assimilation of long-term, remote-sensing soil moisture data sets into a simple, prognostic model driven by observed rainfall. In particular, we utilize simultaneous scatterometer- and radiometer-based soil moisture products obtained from the European Space Agency Climate Change Initiative soil moisture product and a triple collocation analysis approach to estimate the variance of modeling and observation errors (required as input by a data assimilation system). Results show that assimilating remotely sensed soil moisture leads to larger coupling strength estimates as measured by the absolute anomaly correlation with independent temperature observation data than those obtained from modeled or remotely sensed soil moisture alone. According to an analyses of soil moisture error impacts on coupling strength estimates, this increase in coupling strength is likely attributable to an improvement in the precision of the soil moisture product used to estimate it. Based on this, we conclude that data assimilation provides an improved lower bound on the magnitude of true coupling strength between soil moisture and screen-level air temperature.