CHIRPS-GEFS forecasts are now using the new version of NOAA’s Global Ensemble Forecast System (GEFS). A new historical CHIRPS-GEFS archive is under production. Data from October 1st, 2020 to present are currently available. 

Introduction to CHIRPS-GEFS

CHIRPS-compatible GEFS rainfall forecasts for anticipating flood and drought hazards

CHIRPS-GEFS is a bias-corrected and downscaled version of NCEP Global Ensemble Forecast System precipitation forecasts made to be spatially compatible with various CHIRPS products. The ESRL/PSD version 12 Reforecast Project version-12 runs an instance of the Global Ensemble Forecast System (GEFS) model 16 days into the future (  These data consist of 5 ensemble members, and the mean of these members is used as the target forecast for this product. Daily rainfall forecasts are accumulated to create 5-/10-/15-day totals.  The rank-based quantile of these totals is then quantile-matched to the empirical distribution of CHIRPS rainfall for the corresponding period. The result of the quantile-matching scheme is that the average and variance of the CHIRPS data is approximately retained in the resulting CHIRPS-GEFS values. The CHIRPS-GEFS forecast data product is a valuable resource for CHIRPS users in particular, as it provides 5-day to 15-day GEFS forecast precipitation totals and anomalies that are compatible with the historical CHIRPS. This feature allows for the timely assessment of how the latest forecast could alter the current agro-climatological situation.


Daily 5-day, 10-day, 15-day Forecasts




Dekadal Forecasts

First (5-8 day lead)

  • Total & Anomaly         Data

Last (0-day lead) 

  • Total & Anomaly         Data

Pentadal Forecasts

First (10-day lead)

  • Total & Anomaly          Data

Second (5-day lead) 

  • Total & Anomaly          Data

Last (0-day lead)

  • Total & Anomaly          Data



  • CHIRPS-GEFS is available online through the CHC EWX viewer. Users can view the entire data set and visually compare it to other rainfall products, as well as view it alongside other types of data, such as surface temperature and evapotranspiration. Users can also view historical time-series plots over a variety of regions, such as administrative zones, crop zones, and hydrologic catchment basins. Time-series statistics and raw data can be downloaded from the viewer.
  • To download GeoTIFF files for various daily, pentad, and dekad variations, we have provided the table above.  It contains links to data, and some select graphics, for the many variations of CHIRPS-GEFS. The top row of the table shows the forecast rainfall totals estimated for the next 5, 10, and 15 days, updated every day.  The second row is dekadal data. These correspond to the calendar dates 1-10, 11-20, and 21-end of each month. Given that the GEFS forecast is for the next 16 days, the “First” corresponds to the first chance we get to estimate a complete dekad (i.e. when the last day of the dekad is the 16th day of the forecast period). The “Last” is the forecast released on day 1 of the forecast interval, and since it has the shortest lead time, it represents the best estimate of the rainfall totals, because the model is initialized with the best representation of the actual conditions. The third row of the table links to pentadal data, this is the 5-day analog to the dekads in which each month contains six pentads; the first five contain exactly five days, while the sixth pentad of the month can range from three-to-six days depending on the month. Again, given the 16-day forecast period, there are three opportunities to estimate each pentad.  All pentadal estimates are made on the first day of each pentad (i.e. the 1st, 6th, 11th, 16th, 21st, or 26th of each month). For pentadal estimates, the forecast is used to create forecast estimates for the next three pentads, which may capture 13-16 days, depending on the time of year. For example, if we are evaluating the third pentad of a month (the 11th through 15th), the “First” estimate would be made on the 1st of the month, the “Second” estimate would be made on the 6th of the month, and the “Last” estimate would be made on the 11th of the month. Similarly, if it is the 11th of the month, the “Last” estimate would cover the 11th through the 15th of the month, the “Second” estimate would cover the 16th through 20th of the month, and the “First” estimate would cover the 21st through 25th. 
  • An external source for accessing CHIRPS-GEFS is SERVIR’s ClimateSERV portal. This tool is useful for creating a historical time series of user-specified or selected polygons. Click the about ClimateSERV button for more information about that tool’s capabilities.



  • CHIRPS-GEFS is routinely used in producing CHC Early Estimates, a data and map resource for agro-climatological monitoring. CHC Early Estimates show rainfall accumulations for recent time periods based on the CHIRPS final and preliminary data. The  “+ Forecast” version includes the CHIRPS-GEFS 15-day forecast. 
  • Monthly GeoGLAM Crop Monitor for Early Warning bulletins. CHIRPS-GEFS are used for evaluating season-to-date (plus forecast) rainfall conditions for areas of concern. This information is often featured in the Regional Outlooks for areas of concern. 
  • CHIRPS-GEFS and potential use for flood early warning: Blending CHIRPS Data and GEFS Forecasts for an Enhanced Rainfall Forecast Product
  • Examples of CHIRPS-GEFS in early warning for the March-to-May 2019 Eastern Horn drought, including generation of combined forecast seasonal rainfall scenarios.
  • A video presentation of Dr. Chris Funk showcasing a number of products for southern Africa is available on the CHC YouTube channel here.
  • Examples of CHIRPS-GEFS incorporated in operational monitoring of the 2018/19 growing season over Southern Africa can be found here. 
  • Application in crop early warning reports in Tanzania by the Tanzania Ministry of Agriculture.  The CHIRPS-GEFS is regularly featured in monthly bulletins on the right-hand side of this page.