Comparative performance of nonparametric  methods for detecting rainfall trends in West Garo Hills, Meghalaya, India

Rainfall variability in Meghalaya’s West Garo Hills significantly impacts agriculture, water resources, and ecological stability, necessitating robust trend analysis for effective climate adaptation. This study examines long-term trends in annual and seasonal rainfall using advanced nonparametric techniques. Rainfall data (1984–2023) were analyzed using Mann-Kendall and its modified variants (MMKY, PWMK, TFPWMK, SMK, CSMK), supported by normality and randomness tests. Statistical diagnostics confirmed non-normality, non-randomness, and autocorrelation in the dataset.The MK test revealed a significant negative trend (Z = -2.697, p = 0.007; slope = -0.0738 mm/year), corroborated by TFPWMK (Z = -2.782, p = 0.0054) and CSMK (Z = -4.0468, p = 5.19e-05). Seasonal MK detected strong seasonal effects (Z = -5.9655, p = 2.44e-09), while autocorrelation-adjusted methods confirmed trend persistence.The novelty lies in a comparative scoring framework that identifies CSMK as the most robust method, offering high statistical power, interpretability, and resilience to autocorrelation and seasonality. These findings support the adoption of climate-resilient agriculture, water conservation, and effective policy planning in Northeast India.