Abstract:
An artificial neural network (ANN) is presented for the purpose of estimating passive microwave (PMW) emission from snow-covered land in North America. The NASA Catchment Land Surface Model (Catchment) is used to define snowpack properties. The Catchment-based ANN is then trained with PMW measurements acquired by the Advanced Microwave Scanning Radiometer (AMSR-E) or the Special Sensor Microwave/Imager (SSM/I). The intended use of the ANN is for eventual application as a predicted measurement operator in an ensemble-based data assimilation (DA) framework to be presented in a follow-on study. A comparison of ANN output against AMSR-E and SSM/I measurements not used during training activities as well as a comparison against independent PMW measurements collected during airborne surveys demonstrates the predictive skill of the ANN. When averaged over the study domain for the available PMW measurement collection period, computed statistics (relative to PMW measurements not used during training) for multiple frequencies and polarizations yielded a near-zero bias, a root mean squared error less than 10K, and an anomaly correlation coefficient of approximately 0.7. The ANN demonstrates skill at reproducing brightness temperatures during the ablation phase when the snowpack is ripe and relatively wet. The ANN demonstrates even greater skill during the accumulation phase when the snowpack is relatively dry. Overall, the results suggest the ANN should serve as an effective predicted measurement operator that is computationally efficient at the continental scale.