Air temperature, global radiation, potential evapotranspiration and precipitation

C3S AgERA5 data stream replaces the in-house product merging ECMWF ERA5 and ENS used in previous ASAP versions.

AgERA5 air temperature and global radiation are used to their full global extent. AgERA5 precipitation is used outside the 60°N / -60°S latitudinal band covered by CHIRPS precipitation (see below).

CHIRPS Precipitation

CHIRPS v3 from the Climate Hazard Center of the University of Santa Barbara replaces CHIRPS v2 used in previous ASAP versions in the ASAP Warning Explorer and downstream services (e.g. admin unit spatial aggregation for national and subnational crop yield forecasting). CHIRPS v3 spans a larger latitudinal band (60°N to 60°S) compared to v2 (50°N to 50°S).

CHIRPS v2 is temporarily maintained in the computation of WSI to guarantee continuity between WSI historical archive and near real-time production. Reprocessing of the full archive with CHIRPS v3 is planned for Q1 2027.

Note: seasonal forecasts and forecast warning are not affected by such changes.

AgERA5 meteorological data are produced with a 7-day latency. Consequently, dekadal observational ASAP warnings are made available with a 5-day latency, compared with the previous ASAP versions, for which the latency was typically 2 to 3 days. That is, warnings for dekad D are published on the morning of day 5 of dekad D+1. As in the previous ASAP version, the meteorological processing pipeline produces preliminary products that are later updated to final, consolidated products. The mechanisms and timing of the preliminary and final products in the previous and current versions of ASAP are summarized in Table 1.

FAO GAUL 2024 edition replaces the FAO GAUL 2015 used in previous ASAP versions. Level 1 is based on GAUL1 globally, except for Australia, Canada, Libya, and Russia where GAUL2 is used due to the large size of GAUL1 units. Level 2 is only available in selected countries where Level 1 does not correspond to Integrated Food Security Phase (IPC) acute food insecurity reporting units, using GAUL2 by default, except for Ghana (IPC units) and Burundi, Haiti, and Tajikistan (FEWS NET livelihood zones). In previous ASAP versions, Level 1 and Level 2 corresponded globally to GAUL1 and GAUL2 respectively, with Level 2 applied across all countries covered by the system. Country-specific exceptions were in place for selected countries and no longer applied.

Cropland (rangeland) time series warning charts now display breakpoints at dekads where a consistent share of the unit crop (rangeland) area ends its season. Figure 1 shows an example of breakpoint determination for the Hirat ASAP unit in Afghanistan.

Figure 1. Time series of cropland warnings for a level 1 ASAP unit (Hirat, Afghanistan).

This is particularly helpful for units with long growing seasons or year-round agricultural activity, where a single ASAP season would otherwise hide the coexistence of overlapping but distinct cropping cycles within the same administrative unit.

Such conditions arise, for example, in the Sub-Saharan African transition zone where monomodal and bimodal rainfall regimes coexist within a single unit, or in Europe and central Asia where winter and summer crops occupy different portions of the same season (as identified in the average long-term phenology). A breakpoint is introduced when at least 25% of the total crop (rangeland) area ends the season within a two-month window; the resulting sub-seasons must each be at least nine dekads long (approximately three months).

Figure 2 shows for the same unit of Figure 1, the temporal evolution of percentage of pixels ending or starting the season per dekad and the breakpoint derivation.

Figure 2. Breakpoint determination for a level 1 ASAP unit (Hirat, Afghanistan). EOS and SOS curves show the percentage of pixels ending or starting the season per dekad; dashed lines are their 6-dekad backward running means. Vertical lines mark phenological transitions and breakpoints: identified candidates (top panel) and the final promoted breakpoint (bottom panel, yellow line). The original long ASAP season (dekad 5–29) is split into two (dekad 5–18 and dekad 19–29).

A unit is analyzed if at least 60 cropland (rangeland) pixels are present and these pixels represent more than 0.1% of the total unit area. For rangeland units, an additional condition applies: the mean cattle equivalent density, calculated using FAO GLW data, must exceed 0.44 livestock units per pixel. These reduced requirements allow a larger number of units to be analyzed compared with the previous thresholds.

ASAP documentation is currently being updated to reflect the latest developments and functionalities of the system. The updated documentation is expected to be made available in Q2 2027.