Extreme rainfall spatiotemporal trend based on modified Mann-Kendall test under multiple hypothesis testing for Rajang Basin in Sarawak, Malaysia

Understanding the spatiotemporal dynamics of extreme rainfall is critical for flood risk management and water resource planning in tropical river basins. This study assesses extreme rainfall trends across the Rajang Basin in Sarawak, Malaysia, using CHIRPS v3 daily rainfall data from 1981 to 2024, at 0.25° resolution across 95 grids. Missing observations were reconstructed using predictive mean matching (PMM) with spatial neighbors to ensure robust imputation. Extreme rainfall indices, including Rx1day, Rx3day, Rx5day, Rx7day, R10, R20, R25, R95pTOT, and R99pTOT, were derived and analyzed using the modified Mann-Kendall test with Walker-style robustness. To control false discoveries, False Discovery Rate (FDR) and Benjamini-Yekutieli (BY) corrections were applied, and field significance was evaluated via Monte Carlo simulation. Spatial autocorrelation was examined using Moran’s I. Results show largely non-significant trends across indices, with seasonal variability characterized by higher extremes during the Northeast monsoon (NEM) and reduced intensity during the Southwest monsoon (SWM). Localized positive and negative trends occur but remain within expectations under randomness. Spatial patterns indicate moderate clustering, highlighting basin-scale heterogeneity. Overall, despite pronounced monsoonal seasonality, the absence of significant long-term temporal trends in rainfall extremes suggests that observed downstream flood risks may be influenced by a complex interplay of local factors, warranting further investigation into non-climatic drivers such as land-use changes. This framework demonstrates robust detection of subtle spatially coherent changes in tropical rainfall dynamics. These findings provide evidence supporting cautious interpretation of apparent local changes and emphasize the importance of integrating statistical rigor with spatial diagnostics in hydroclimatic studies across data-scarce tropical regions.