Abstract
Reliable precipitation forcing is essential for calculating the water
balance, seasonal snowpack, glacier mass balance, streamflow, and other
hydrological variables. However, satellite precipitation is often the
only forcing available to run hydrological models in data-scarce
regions, compromising hydrological calculations when unreliable. The
IMERG product estimates precipitation quasi-globally from a combination
of passive microwave and infrared satellites, which are intercalibrated
based on GPM’s DPR and GMI instruments. Current GPM-DPR radar algorithms
have satisfactorily estimated rainfall, but a limited consideration of
PSD, attenuation correction, and ground clutter have degraded snowfall
estimation, especially in mountain regions. This study aims to improve
satellite radar snowfall estimates for this situation. Nearly two years
(between 2019 and 2022) of aloft precipitation concentration, surface
hydrometeor size, number and fall velocity, and surface precipitation
rate from a high elevation site in the Canadian Rockies and collocated
GPM-DPR reflectivities were used to develop a new snowfall estimation
algorithm. Snowfall estimates using the new algorithm and measured
GPM-DPR reflectivities were compared to other GPM-DPR-based products,
including CORRA, which is employed to intercalibrate IMERG. Snowfall
rates estimated with measured Ka reflectivities, and from CORRA were
compared to MRR-2 observations, and had correlation, bias, and RMSE of
0.58 and 0.07, 0.43 and -0.38 mm h-1, and 0.83 and 0.85 mm h-1,
respectively. Predictions using measured Ka reflectivity suggest that
enhanced satellite radar snowfall estimates can be achieved using a
simple measured reflectivity algorithm. These improved snowfall
estimates can be adopted to intercalibrate IMERG in cold mountain
regions, thereby improving regional precipitation estimates.