Trend Analysis and Forecasting of Climate in the Ladakh region of Western Himalayas using the Mann-Kendall test and Machine Learning models

The cold-arid, high-altitude region of Ladakh is among the most climate-sensitive environments in the Western Himalayas, yet long-term assessments of its climatic trajectory remain limited. This study provides a comprehensive analysis of rainfall and temperature variability using IMD gridded data (1980–2024), combining the Mann–Kendall test, Sen’s slope estimator, and ensemble machine learning models (Random Forest and XGBoost) to detect past trends and forecast climate conditions for 2025–2054. Results reveal a significant and persistent decline in precipitation across all months and seasons, with an annual decrease of –47.13 mm/year. Winter and summer exhibit the sharpest reductions, highlighting weakening western disturbances that dominate Ladakh’s hydrometeorology. Maximum and minimum temperatures show robust warming, with Tmin rising more rapidly (+0.0175 °C/year) than Tmax (+0.0184 °C/year), indicating pronounced night-time warming and implications for permafrost and glacier stability. Machine-learning-based forecasts project continued aridification, with rainfall declining by 6–12% and winter Tmin increasing by +0.9 to +1.2 °C by 2054. XGBoost outperformed RF across all performance metrics, producing more stable and reliable predictions. The combined evidence points to warmer winters, reduced snow accumulation, altered meltwater timing, and heightened water stress in Ladakh’s fragile mountain environment. These findings underscore the urgent need for adaptive water-resource strategies, integration of advanced forecasting tools into regional climate services, and enhanced monitoring of cryosphere–climate interactions in the Western Himalayas.

Keywords: Ladakh; Climate variability; Mann–Kendall test; Sen’s slope; Rainfall trends; Temperature trends; Machine learning forecasting; Random Forest; XGBoost; High-altitude Himalaya.