Input data set: Irrigation techniques — country level (Group III)

Country-level companion products for the irrigation techniques dataset (see the Irrigation techniques (Group III) entry for the gridded files).

The country CSVs give, for every country and year from 2015 to 2100, the irrigated area per irrigation technique in m2 (area_m2) and that technique's share of the country's total irrigated area in that year (share_of_country_irrigated_area). They were aggregated from the published gridded irrigation techniques inputs over the cells that carry a country code in the published country index.

Two reference files accompany them: the country index itself (irrigation-techniques-country-index_05deg.nc, a 0.5 degree NetCDF holding integer country_index 0-256, where 0 means the cell is not assigned to any country, plus valid_data_mask marking the cells that carry irrigation-techniques data) and a lookup CSV (irrigation-techniques-country-lookup.csv) mapping country_id to country name, ISO3 code where defined, color code, status, continent and region.

Together these three file types allow to map every grid cell to the same country assignment that was used to derive the country products, and to check a cell-to-country derivation against the published sums.

Time step: annual, 2015-2100 (86 years)

Spatial resolution: national (one row per country and year, 256 countries plus their years without irrigation), aggregated from the 0.5 degree gridded irrigation techniques inputs via the country index

Format: CSV (header row, comma separated) for the country and lookup files; NetCDF4 (0.5 degree regular lon/lat grid) for the country index

Files are organized in the same soc-scenario subfolders as the gridded files; the two reference files sit at the top of the dataset folder.

Aggregated from the published ISIMIP3b gridded irrigation techniques inputs, which were derived according to the methodology described in Yao et al. (2025, https://www.researchsquare.com/article/rs-6121279/v1) code repository: https://github.com/YiYao1995/Irrigation-technique-shares-projection.git

Country geometry: world-administrative-boundaries (https://public.opendatasoft.com/explore/dataset/world-administrative-boundaries/export/) as recorded in the country index NetCDF's boundary_source attribute.

Country assignment is the one of the original irrigation techniques input, not the unified ISIMIP country masks. Cells are assigned by center-point placement into the polygons of the world-administrative-boundaries dataset, followed by a single neighbor-fill pass (unassigned cells carrying valid data take an adjacent cell's id, without chaining). The unified ISIMIP regional masks differ from this index and cannot be used to reproduce the country sums.

Country sums are about 0.02-0.04 % below the gridded global totals. A small number of irrigated cells (26 for the ssp126soc-noadapt IMAGE input in 2015, at small-island states such as Malta, Barbados or Hawaii, plus a few ocean-centered coastal cells; Antarctica is not part of the boundary dataset) never received a country code. They carry data in the gridded files, where their irrigated area is entirely flood irrigation (drip and sprinkler are zero there for 2015-2100), but no country to sum into. To reproduce the gridded product from the country CSVs exactly, map cells with the published index and apply flood = 100 % to cells with country_index == 0.

country_index > 0 is not a test for "cell has data". Country polygons include inland waters whose cells carry no land-use data (for example in the Caspian Sea, the Aral Sea, the Great Lakes and Lake Victoria — 195 cells in the ssp126 IMAGE input). These cells have a country code but no irrigation data. A cell has data if country_index > 0 and the gridded input is not missing there; only such cells contributed to the country sums.

17 of the 256 polygons are disputed or administered territories without an ISO3 code (for example Jammu-Kashmir, Arunachal Pradesh, Aksai Chin). They have their own country_id and their own rows in the country CSVs. Join on country_name (unique across all 256 entries) or color_code; joining on iso3 alone silently drops them.

ISIMIP participants can download these files from the DKRZ server using the path /work/bb0820/ISIMIP/ISIMIP3b/InputData/socioeconomic/irrigation-techniques