Measuring Snowfall From Satellite Microwave Observations

Abstract:

Snowfall is one of the important components in the global hydrological cycle. However, snowfall retrieval from satellite observations is very difficult (compared to rainfall retrieval) because of several reasons including weak radiative signal, surface contamination, cloud liquid water masking, etc. Several satellite sensors currently in operation are potentially capable of detecting and estimating snowfall, for example, cloud radar onboard CloudSat and high frequency microwave radiometers (SSMIS, MHS, GMI) on NOAA, MetOP, S-NPP, DMSP and GPM satellites. In this paper, we report our research results on how to best use these sensors to extract snowfall information. Specifically, we (1) studied the global snowfall frequency and rate distributions based on multiple years of CloudSat observations; (2) investigated the optimal channel selection for snowfall retrieval using collocated satellite SSMIS (19-183 GHz) and ground radar (NMQ) data; (3) developed a retrieval algorithm in which radar data (CloudSat and/or NMQ) are used as truth to train high-frequency passive microwave data and use high-frequency passive microwave data to broaden spatial and temporal coverage. Finally, retrievals based on observations from several satellites are compared with each other and compared with surface observations.