PSL contributes to new version of the NWS Global Ensemble Forecast System

A person walking with an umbrella on a rainy city sreet. Photo courtesy Osman Rana on Unsplash.com

On September 23, the National Weather Service (NWS) upgraded its Global Ensemble Forecast System (GEFS) to version 12. GEFSv12 provides improved ensemble forecast data to support NWS products for medium range (3-14 days) to monthly forecasts. NOAA and CIRES scientists at the Physical Sciences Laboratory (PSL) worked closely with NWS Environmental Modeling Center (EMC) colleagues to develop the GEFSv12 system. The PSL team adapted and tested ways of estimating the forecast uncertainty due to model imperfections, so that the resulting GEFSv12 ensembles now provide probabilistic forecasts that are much more reliable. As an example, in all the circumstances when 80% chance of rain is forecast, rain should happen 80% of the time.

PSL scientists also generated a 20-year reanalysis with a data assimilation system that very closely resembles the real-time operational analysis system, and worked with EMC colleagues on the production of multi-decadal reforecasts. These retrospective forecasts use the same GEFSv12 code base as the operational predictions. The resulting reforecasts are invaluable for making statistical adjustments to the real-time forecasts so that the product quality is further improved to provide high-resolution terrain detail. PSL now leads multiple efforts to use these reforecasts to provide improved products, such as sub-seasonal precipitation and hurricane forecasts, fire-weather forecasts, and more. PSL and EMC also worked together to make the data available inside and outside of NOAA, including storage on Amazon Web Services through NOAA’s Big-Data Project. PSL continues its collaborations with EMC to develop GEFS version 13 and supporting data sets. It is anticipated that GEFSv13 will be a coupled atmosphere-ocean-sea ice-land prediction system.

Medium-range to sub-seasonal products in coming years will be dramatically improved through use of data from the new Global Ensemble Forecast System that PSL scientists helped develop.

Learn More