CHC Early Estimates: Recent Rainfall and Seasonal Rainfall
The CHC Early Estimates provide a tool for monitoring rainfall conditions using combinations of CHIRPS, preliminary CHIRPS, and the CHIRPS-GEFS 15-day forecast. The Recent Rainfall products provide near-real-time estimates of rainfall accumulations for the most recent pentad (5 to 6 days) to the most recent 18 pentads (3 months). The Seasonal Rainfall products provide near-real-time estimates of rainfall accumulations for region-specific agricultural seasons ranging from 3 to 6 months.
The CHC Early Estimates have proven to be a valuable resource for agroclimatic monitoring with the CHIRPS data stream. Early Estimates have been operational since 2018. Starting in October 2025, the product input data stream changed from CHIRPS v2 to CHIRPS v3.
Indicators of rainfall performance
The CHC Early Estimates products include rainfall totals (Total), difference from the 1981–2020 average (Anomaly), percent of average (% Average), standardized anomalies (SPI), historical rank for the 1st to 3rd wettest or driest (Rank), and the Seasonal Precipitation Performance Probability (SPPP). SPPP is described in more detail on this page.
When mapping the latter four products—% Average, SPI, Rank, and SPPP—we apply a “dry mask,” in which areas that average less than 1 mm per pentad during the given monitoring period are masked; the data files themselves are unmasked. Additionally, 1 mm/pentad (max of 10 mm) is added to the numerator and denominator when calculating the percent of average in order to avoid extreme values in dry areas.
CHC Early Estimates use available data and a forecast
To-date rainfall values are based on preliminary CHIRPS data for the most recent pentads, and CHIRPS Final data, when available. The outlook (+Forecast) combines these to-date accumulations with a rainfall forecast for the next three pentads (15–16 days) using CHIRPS-GEFS. CHC Early Estimates have the same latency as CHIRPS Preliminary data, and are updated on the 2nd, 7th, 12th, 17th, 22nd, and 27th of each month, with the Seasonal Rainfall only being updated during the seasons of interest. As the Early Estimates include both preliminary and forecast data, the information is subject to change. All products are back-processed when CHIRPS Final is made available in order to ensure the most accurate estimates.
In October 2025, the CHC transitioned the Early Estimates from CHIRPS2 to CHIRPS3 products (e.g., CHIRPS3 Final, Prelim, and CHIRPS3-GEFS). The latest year of Recent Rainfall graphics and products are available from the data portal, and only the most recent season of the Seasonal Rainfall graphics are available. Early Estimates products using CHIRPS2 are accessible via the data portal and links at the bottom of this page. Users are encouraged to transition to CHIRPS3 products.
Data access: CHC Early Estimates Recent Rainfall
The CHC Early Estimates Recent Rainfall page provides easy access to regional maps (pngs). The ”Rainfall Accumulations” links below lead to all maps and data archives for those rainfall values. The “Outlook” section contains links to the "+Forecast" outlooks, which include the 3-pentad CHIRPS-GEFS forecast. The usage of preliminary or forecast data sources, and corresponding dates, are documented on the maps. GeoTIFFs of these products are available for download for the full CHIRPS extent (60N°–60°S, 180°W–180°E, 0.05° resolution). The geotiffs are not subset by region.
Data access: CHC Early Estimates Seasonal Rainfall
The CHC Early Estimates Seasonal Rainfall page provides easy access to regional maps (pngs). Archived maps (pngs) and data (geotiffs) can be retrieved using the links below. The +Forecast outlook maps and data shown below can be accessed through the top links, within the “png” and “tif” directories (look for “with_forecast”)
Middle East & North Africa Nov – Jul
CHC Early Estimates with CHIRPS2
The Early Estimates were originally produced using CHIRPS2 and transitioned to CHIRPS3 in October 2025. The CHIRPS2 Early Estimates will continue to be produced and be available through the links below until July 1, 2026.
Users should note that the increased variance in CHIRPS3 compared to CHIRPS2, and other features implemented to improve data performance, results in a larger spread of rainfall values as well as the estimation of higher rainfall amounts in CHIRPS3. In the CHC Early Estimates, the occurrence of more extreme values in CHIRPS3 compared to CHIRPS2 particularly influences percent of average rainfall values. Users may also observe more spatially variable rainfall patterns in the CHIRPS3-based product.
About the SPPP: Seasonal Precipitation Performance Probability
In addition to the five products available in the Recent Rainfall set, the Seasonal Rainfall Early Estimates also includes the Seasonal Precipitation Performance Probability (SPPP). The SPPP provides a historical rainfall-based outlook ofevaluates the probability that the current season’s precipitation total will fall within one of the three tercile categories. These merge current season rainfall to-date with season-completion rainfall values from every year of the historical CHIRPS archive. A “+Forecast” version, which used the 3-pentad CHIRPS-GEFS forecast, is also provided. The SPPP maps show probabilities of below-normal (< 33rd percentile), normal (33rd–66th percentile), and above-normal (> 66th percentile) seasonal rainfall totals, based on the assumption that season-completion rainfall totals observed during every historical year are equally likely to reoccur for the current season..
Observed season totals for all previous years are used to generate the tercile boundaries for the season of interest. The current season's potential totals are calculated by taking to-date rainfall and running out the remainder of the season using all available previous years (n) of data (1981 to last year) to generate n potential season totals. Probabilities of each tercile are then calculated by counting the number of those scenarios that fall in each historical tercile and dividing by n. The SPPP maps the percent probability of the most likely tercile; i.e. if 29 of the 42 simulated season totals fall in the bottom tercile of historical totals, then the SPPP would be 69% probability that the current season will be ‘Below Normal’.
Please send comments and inquiries to Will Turner and Laura Harrison (will.turner@geog.ucsb.edu ; harrison@geog.ucsb.edu)