Heat and heavy-rain days are district-mean values from ERA5-Land (~11km) and CHIRPS (~5.5km), 1990–2023. A heat day is ≥35°C daily max; a heatwave is 3+ consecutive heat days within the same calendar year; a heavy rain day is ≥20mm. At this resolution, large heterogeneous districts get an area-averaged value that can smooth over local extremes, and small/urban districts may be represented by very few pixels.
Flood frequency is the % of a district's area classified as seasonal or permanent surface water that year, from JRC Global Surface Water (30m, reduced at 300m), 1990–2021.
Prevalence is from Zimbabwe DHS HIV test results (AR recode), weighted by DHS sample weights, for the 2005–06, 2010–11 and 2015 survey rounds — the only rounds with HIV testing currently loaded. Each tested person is assigned to a district via their DHS cluster's GPS coordinates, which DHS deliberately displaces by ~2km (urban) to ~5km (rural, 10km for 1% of rural clusters) to protect anonymity — district assignment near boundaries is therefore approximate, not exact.
FAO GAUL 2015 level2 (62 districts), the same source used elsewhere in the CASCADE Zimbabwe flood model work, for consistency across outputs.
Each district's heat days, heavy rain days, flood frequency and HIV prevalence are averaged across all available years/rounds, min-max normalised to 0–1, then averaged into one composite score. It ranks relative exposure across districts; it is not an absolute risk measure.