The spatial distribution of irrigation dam benefits is poorly understood at the global scale due to a scarcity of spatial information on irrigation dam command areas. Several studies aimed at mapping irrigated lands globally, but the spatially explicit attribution of irrigated lands to dams has rarely been undertaken. First approaches attributing changes in agricultural production to dams were based on aggregated areal units, such as administrative districts (Duflo and Pande, 2007), or watershed boundaries (Strobl and Strobl, 2011). These approaches represent only indirect approximations of command areas, and may be improved by considering spatially explicit dam- and location-specific parameters (e.g. reservoir storage capacity or topography). Such a refined dataset is required for better understanding the spatial distribution and properties of irrigation dam command areas.
We approximated irrigation dam command areas for 1,370 dams with irrigation function which were commissioned across 71 countries since 1985. We approximated a) the extent and b) the location of irrigation dam command areas at 500m spatial resolution using global-scale assumptions motivated by existing literature. We first estimated command area extent [ha] based on reservoir storage capacity [m³], while accounting for country-level variations in the ratio of land irrigated with surface water per unit of total national reservoir storage capacity [ha/m³]. We then spatially allocated the estimated command area extent for each dam, accounting for parameters representing: irrigated cropland abundance (P1), topography relative to the dam (P2), watershed boundaries (P3), reservoir size (P4), national borders (P5), and distance to the dam (P6). To understand the sensitivity of the allocation towards the assumptions underlying the selected parameters, we tested 24 different allocation schemes with varying parameter settings. An overview of the datasets used for the command area extent estimation and the spatial allocation procedure, as well as an illustration of an exemplary allocation is included in the download. For a detailed description of the methods used to produce these data, please see Rufin et al. (2018).