Russ Schumacher and Greg Herman demonstrate new forecasting product

During June and July 2017, the NOAA Weather Prediction Center and Hydrometeorology Testbed are hosting FFaIR: the Flash Flood and Intense Rainfall experiment. This program brings together researchers, forecasters, numerical weather model developers, and others to evaluate new tools for improved prediction of heavy rainfall and flash flooding. One of the products being demonstrated and evaluated this year was developed by CSU atmospheric science graduate student Greg Herman, who is advised by Professor Russ Schumacher.

The product uses machine learning algorithms to process historical observations of heavy precipitation, along with output of weather-prediction models and information about the past performance of those models, to generate probabilities of an extreme rain event occurring in regions all across the U.S. These probabilistic forecasts are being formally evaluated by the FFaIR participants, with the goal of eventually becoming a product used in forecast operations at the WPC.

This project is supported by the NOAA Joint Technology Transfer Initiative.

Link to experimental extreme precipitation forecasts

Photo: FFaIR experiment participants evaluate the experimental CSU heavy precipitation forecast product during the daily forecast activities.