Dataset of key derived spatial outputs for the article "From rainfall Seasons to Moisture-Season Regimes: a Daily Water-Balance Framework for Crop-Neutral Agricultural Water Mapping in Kenya"
Description
This minimum processed-data deposit contains key GeoTIFF outputs supporting the reported comparison of seasonal regime classifications and reliability/probability outcomes under a fixed 100 mm storage representation and SoilGrids-derived water-holding-capacity representation. Included spatial products comprise regime classifications, regime reliability surfaces, probabilities of one season and two-or-more seasons, scenario class-change surfaces, and the common-valid and no-data/water/unclassified masks required for interpretation. Raw external source datasets and large daily processing intermediates are excluded.
Files
Steps to reproduce
Daily rainfall, temperature and atmospheric forcing data for 1991–2020 were assembled for Kenya. CHIRPS daily rainfall was used as the rainfall input and as the 0.05° master spatial grid. Daily minimum and maximum temperatures were obtained from CHIRTS-ERA5. The completed processing workflow used NASA POWER solar radiation, dewpoint temperature and 2-m wind speed as drivers for FAO-56 Penman-Monteith reference evapotranspiration. Soil-water storage was represented using two alternatives: fixed reference storage of 100 mm and a SoilGrids-derived water-holding-capacity surface. All daily and static layers were clipped to Kenya and aligned to the common CHIRPS grid. Daily reference evapotranspiration was then calculated using the FAO-56 Penman-Monteith equation. A daily soil-water balance was implemented independently for the two storage representations, with soil-water storage constrained by the respective maximum storage capacity. Annual moisture seasons were identified using the study parameterization: rainfall onset required at least 25 mm accumulated over four days, with at least 1 mm rainfall on the starting day; a 10-day false-start check was applied; seasons had to persist for at least 45 days; moisture onset required storage of at least 30% of maximum storage; cessation was recorded when storage remained below 20% of maximum storage for 10 consecutive days; and two seasons were separated by a minimum 20-day dry break. Annual season outcomes were aggregated across 1991–2020 to derive, for each storage scenario, the probability of one moisture season, the probability of two or more moisture seasons, and the reliability of the dominant seasonal regime. Regime classes were assigned using the reliability rules applied in the analysis: reliable regimes occurred in at least 60% of years, strongly reliable regimes in at least 80% of years, and transitional regimes in 40–59% of years. Final GeoTIFF outputs were generated for regime classes, regime reliability, probabilities of one season and two-or-more seasons, scenario class changes and supporting masks. Direct spatial comparisons between the fixed-storage and SoilGrids WHC scenarios should be conducted only within `Figure_common_valid_mask_reference_and_soilgrids.tif`. The minimum repository deposit retains the selected derived GeoTIFF products, companion XML metadata files, key result tables, complete R code, metadata records and checksums, while excluding raw third-party inputs and large daily intermediate datasets.