CHARACTERIZING DROUGHT AT MALANVILLE USING A MULTISCALE STANDARDIZED WATER-BALANCE INDEX (NIGER BASIN, BENIN)

The rainfall regime at Malanville (northern Benin, Niger River basin) shows strong intra- and inter-annual variability typical of the Sudano-Sahelian domain, yet drought diagnostics based on rainfall alone do not account for evaporative demand. This study builds and consolidates a standardized water-balance index (SWBI) that integrates potential evapotranspiration (PET) to characterize drought episodes at Malanville over 1995-2024.
In the absence of a temperature record at Malanville itself, PET is estimated with the Hargreaves-Samani method from the 1991-2020 climatological normals of the nearby Kandi station (~100 km away), applied as a fixed seasonal cycle. The monthly water balance (P − PET) is standardized by calendar month at four time scales (1, 3, 6, 12 months), and dry episodes are extracted with run theory (threshold −1). Robustness is tested with seven independent indices: the Rainfall Anomaly Index, a gamma-fitted SPI, the decile method, the ETCCDI daily climate indices, the De Martonne aridity index, the Bagnouls-Gaussen xerothermic index, and the length of the wet period.
Cumulative annual PET (1,949.4 mm) far exceeds mean annual rainfall (823.9 mm), yielding an annual aridity index of 0.42. At the 3-month scale (77% data coverage), 15 dry episodes are identified, the most severe being May-July 2018 (peak −2.10). Four statistically independent methods agree perfectly in flagging 1995, 2013, 2021 and 2023 as robust deficit years, and 2003, 2019 and 2024 as surplus years. Gamma fitting (SPI-Gamma), better suited to the skewness of rainfall series, systematically increases the estimated severity of the driest years relative to a normal approximation. Indices targeting the intra-annual distribution of rainfall (dry months, wet months), nearly independent of the annual total (r < 0.2), reveal distinct information: 2024, classified wet on annual total alone, has as many dry months as 2023, a confirmed deficit year.
The convergence of seven methodologically independent indices around the same critical years (1995, 2013, 2021, 2023) gives substantial robustness to the drought diagnosis at Malanville, despite the limitations inherent in using a regional temperature climatology as a proxy. The multi-index approach proposed here, combining water balance, classical rainfall indices and standardized daily climate indices, offers a transferable framework for other Sahelian stations facing comparable data constraints.