An Ocean Snowfall Detection Algorithm for ATMS

Abstract:

We will present a snowfall detection algorithm over ocean, sea-ice, and coast for the Advanced Technology Microwave Sounder (ATMS) onboard both S-NPP and NOAA-20 satellites. The algorithm was trained from collocated observations from ATMS and CloudSat Cloud Profiling Radar (CPR) from 2012 to 2017. Some Global Forecast System (GFS) model forecasts (e.g., relative humidity, temperature, and total precipitable water) have also been explored in the development process. Results show that the Heidke Skill Score (HSS) values are close to 0.60 over all three surface types, and the Probability of Detection (POD) values are close to 0.70. These statistics meet or exceed the JPSS Requirements for ocean Snowfall Rate (SFR). It is noted that the ocean (ice free) detection algorithm primarily depends on ATMS brightness temperatures (TBs). The performance only marginally degrades without GFS variables as predictors with POD and HSS at 0.66 and 0.57, respectively. However, GFS variables are critically important to snowfall detection over sea-ice and coastal regions. Without GFS variables, the HSS values over both sea-ice and coastal regions decrease sharply to about 0.30. Comparisons with the snowfall observations from OceanRain (shipborne in-situ data) and Global Precipitation Measurement (GPM) Mission Ku-band Precipitation Radar (KuPR) also indicate strong snowfall detection skills. This product will benefit coastal communities by providing information on snowstorms offshore before they transition to land. In particular, it can provide situational awareness to forecasters in their support of such activities like US Coast Guard search and rescue and other aviation activities along the Arctic coast of Alaska.