CSU machine learning model helps forecasters improve confidence in storm prediction

Over the last several years, Russ Schumacher, professor in the Department of Atmospheric Science and Colorado State Climatologist, has led a team developing a sophisticated machine learning model for advancing skillful prediction of hazardous weather across the continental United States. First trained on historical records of excessive rainfall, the model is now smart enough to make accurate predictions of events like tornadoes and hail four to eight days in advance – the crucial sweet spot for forecasters to get information out to the public so they can prepare. The model is called CSU-MLP, or Colorado State University-Machine Learning Probabilities. 

Dropsonde Profile

Improving the use of Dropsondes in NOAA Operations with HWRF (Sippel FY17)

This project implements an algorithm to estimate drift of the dropsondes in the data assimilation system of the operational Hurricane Weather Research and Forecasting model. With the help of JTTI funds this dropsonde drift algorithm was operational tested and implemented in the National Weather Service (NWS) operations.  NWS is now able to assimilate more dropsonde observations…

RiVorS deployment (Photo by Matthew Mahalik/OU CIMMS)

Implications of Inconsistent Visuals

On end user uncertainty, risk perception, and behavioral intentions (Grundstein FY18) Within the last decade, operational meteorologists have raised concerns that the availability of weather information from a variety of sources may contribute to a perception that weather risk messages are inconsistent and result in negative consequences among end users. The challenge, however, is that…

Global Extratropical Storm and Tide Operational Forecast System (Global ESTOFS)

Improving Water Level Predictions

Title: Advancing ADCIRC U.S. Atlantic and Gulf Coast Grids and Capabilities to Facilitate Coupling to the National Water Model in the Extratropical Surge and Tide Operational Forecast System (ESTOFS) Operational Forecasting (Westerink FY18) The primary objective of this project is to improve predictions of water levels caused by extratropical tide and storm surge within the…

excessive rainfall noaa

Using Machine Learning to Improve Forecasts of Excessive Rainfall

Forecasters at NOAA’s Weather Prediction Center (WPC) are responsible for producing Excessive Rainfall Outlooks, which brings awareness to the potential for flood-inducing rains up to three days in advance. However, the amount of rain that qualifies as “excessive” varies from region to region, and pinpointing the specific areas likely to receive heavy rain is challenging,…