Improving Snowfall Rate Retrieval by Incorporating the Effect of Supercooled Cloud Liquid Water

A snowfall rate (SFR) product has been produced operationally at NOAA/NESDIS since 2012. The product uses measurements from Advanced Microwave Sounding Unit (AMSU) and Microwave Humidity Sounder (MHS) pair aboard four NOAA and EUMETSAT polar-orbiting satellites. Based on the operational product, an SFR algorithm was also developed for Advance Technology Microwave Sounder (ATMS) aboard Suomi-NPP. These algorithms mostly rely on the scattering signal from ice particles in the atmosphere to detect the presence of snowfall and retrieve snowfall rate. High frequency passive microwave measurements, which are used in the SFR algorithms, respond to the scattering effect with decreased values compared to when no precipitation-sized ice particles are present in cloud. However, snowfall often occurs when microwave observations are greater than when there is no precipitation. This is caused by the emission effect of cloud liquid water which can exist in snowing clouds. In fact, previous studies have shown that most snowfall is accompanied with supercooled cloud liquid water. Emission has the effect of elevating brightness temperatures (Tb) at the microwave frequencies. Precipitating snow and cloud liquid water exert competing effect on microwave measurements and result in complex response at different channels depending on the strength of each signal. The SFR algorithms employ an inversion method that involves forward modeling of Tbs at several frequencies. The effects of both ice and water vapor are taken into account in Tb simulation. However, liquid water is not included in the radiative transfer model (RTM) used as the forward model. This causes underestimation of snowfall rate when abundant liquid water is present in clouds such as in the case of convective snowfall. To improve the accuracy of the SFR algorithms, the RTM is modified by adding the effect of cloud liquid water. This presentation will show the result of the improved algorithms including cases where liquid water contaminates snowfall retrievals using the previous algorithms.