Assessment of annual hydrological effects of mining and scenario sensitivity in a coastal sandy basin of southeastern Madagascar

Abstract

This version supersedes the previous version. The Introduction has been reorganised around two research questions, the hypothesis now precedes the objectives, and the documented 2008 operational threshold is distinguished from an instantaneous hydrological break. The Methods now provide the HRU definitions, spatial supports, area-weighting equation, annual BACI regression, Newey-West standard errors, finite-sample correction and Holm adjustment. The Results now report absolute annual water-balance magnitudes, the bounded 2008-2020 temporal-window check and the fully specified low, mid and high scenario responses. The limitations of spatial attribution and the absence of continuous local discharge validation are stated directly. The hydrological effects of coastal mines are difficult to separate from regional basin trajectories, particularly in tropical systems with sandy soils, wetlands and strong interannual variability. This study re-evaluates mine–climate interactions in the Taolagnaro basin of southeastern Madagascar using Soil and Water Assessment Tool (SWAT+) simulations for 1993–2022 and a mining-versus-reference before–after–control–impact (BACI) design centred on a 2008 threshold. Annual hydrological response unit (HRU) water balances are aggregated by area weighting to compare the Mandena mining footprint with an internal basin reference zone. No annual hydrological variable shows a post-2008 BACI contrast that clearly exceeds interannual variability. Percolation yields the most pronounced mean negative contrast but remains too dispersed to constitute an isolated mining response. Post-2008 trajectories of the mining and reference zones evolve largely together, indicating a dominant regional dynamic. A low/mid/high scenario chain nevertheless shows that SWAT+ strongly redistributes the water balance when a disturbance is imposed on the mining HRUs. The study therefore distinguishes annual BACI detection from simulated mechanistic sensitivity, and cautions against over-attributing post-2008 hydrological change to mining at the annual scale in this coastal setting. This is an independent author version released under CC BY 4.0.