Radar Snowfall Relationships during IMPACTS

This study investigates radar snowfall relationships, known as SWER(Ze), utilizing the Precipitation Imaging Package (PIP) and collocated in-situ and remote sensing measurements collected during Investigation of Microphysics and Precipitation for Atlantic Coast Threatening Snowstorms (IMPACTS) field campaign. The measurement site, Storrs, Connecticut, receives wintery precipitation from frontal systems and nor’easters and its proximity to the Atlantic Ocean often results in mixed-phase precipitation. The SWER(Ze) relationships have been previously derived based on snow density and habit classifications and dendrites and plates were the two major habits in these studies. This is still the case for a number of events but graupel, needle and wet snow categories contributed to the habit classification significantly in a number of other events making this dataset more diverse than the previous sites in South Korea and Marquette, Michigan. A number of mixed-phase events had contributions from all 6 habit classes including rain, showing the complicated nature of SWER(Ze) relationship. The multiple habits exists at the same temperature or wet-bulb temperature such that it is hard to distinguish the SWER(Ze) relationship based on environmental variables. From an operational point of view, weather radars use SWER(Ze) relationships based on their measured variables. This study simulates the four major radar variables (reflectivity, differential reflectivity, differential phase, and cross-correlation) utilizing newly developed PIP software where the particle aspect ratios and orientation have been greatly improved by the use of new shape-fitting algorithm. The PIP simulated radar variables are then used to determine the hydrometeor identification (HID) following Thomson et al. (2014) algorithm. The SWER(Ze) relationships have then been derived for each HID and evaluated on an event-by-event basis. This novel and innovative approach has potential to improve the quantitative precipitation estimation as long as the PIP-simulated and radar-measured variables matches well.