Machine Learning Enhanced Snowfall Rate Retrievals from Passive Microwave Sensors

Snowfall accounts for a large fraction of winter precipitation in mid and high latitudes. It is also an important indicator of climate change. The NOAA NESDIS operational Snowfall Rate (SFR) product is generated at near real time from a suite of passive microwave (PWM) sensors onboard several polar orbiting satellites operated by NOAA and its partner agencies. It estimates liquid equivalent snowfall rate over global land and has been in operation since 2012. The SFR algorithm consists of two main components: snowfall detection (SD) and snowfall rate estimation (SFR). Machine learning (ML) algorithms have been developed to improve both components. For SD, we first built a global precipitation dataset that consists of collocated ground weather reports, satellite measurements and NWP model analysis. And then developed a new ML algorithm based on eXtreme Gradient Boosting (XGB) model validated globally. Compare to the previously developed deep neural network (DNN) SD model, the XGB model significantly improved the SD performance under cold conditions, i.e., 2 meter air temperature below -15°C, and also improved the overall SD performance. For snowfall rate estimation, we developed ML algorithms to improve the Ice Water Path (IWP) initialization and the snowfall rate bias correction. Over the ocean, a preliminary machine learning based SD and snowfall rate estimation algorithm was developed using collocated CPR near surface snowfall rate, satellite measurements, and NWP model analysis. The ML SD and SFR algorithms have been successfully applied to cross scanning instrument such as ATMS and MHS sensors. We will present the development of the ML SD and SFR algorithms from the conical scanning GMI sensor and its global validation results. Case study compared with Stage IV radar and gauge combined precipitation rate, CPR near surface snowfall rate, NOHRSC hourly snowfall analysis and ERA5 hourly snowfall rate reanalysis will also be provided.