Investigating long term rainfall trends (1901-2021) and forecast over the mountainous state of India using machine learning approach


Kumar P., Sarda R., Thakur S., Janmaijaya M., GÖNENÇGİL B., Yadav A.

Singapore Journal of Tropical Geography, 2026 (SSCI, Scopus) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1111/sjtg.70054
  • Dergi Adı: Singapore Journal of Tropical Geography
  • Derginin Tarandığı İndeksler: Social Sciences Citation Index (SSCI), Scopus, Environment Index, Geobase, Political Science Complete, Public Affairs Index
  • Anahtar Kelimeler: homogeneity test, Mann-Kendall test, rainfall forecast, rainfall variability, trend analysis
  • İstanbul Üniversitesi Adresli: Evet

Özet

This paper investigates the rainfall variability and trends (annual and seasonal) over Himachal Pradesh state from 1901 to 2021. The annual rainfall fluctuation ranges from 20 per cent to 36 per cent, indicating that rainfall in the mountainous regions has remained relatively stable over 121 years. While monsoon and annual rainfall exhibit a low degree of rainfall variability, there is greater fluctuation evident during the post-monsoon (99 per cent to 185 per cent) and winter seasons (37 per cent to 88 per cent). Homogeneity tests (Pettitt, Lanzante and Buishand's range) were employed to identify the pattern of any sudden changes in the rainfall data. The time series rainfall data reveals 1967 as a clear change or alteration point (p < 0.05). The pre-monsoon and winter season rainfall record post-1967 as characterized by increased rainfall trends and variability across the region. The Modified Mann-Kendall (M-MK) test has been found to be more significant than Trend Free Pre Whitening Mann-Kendall (TFPW-MK) and Mann-Kendall (MK) tests. The Artificial Neural Network-Multilayer Perception (ANN-MLP) technique was employed to forecast the next 30 years of rainfall (2022–2051). This investigation may help in formulating mitigation strategies to combat the impact of climate change in the mountainous region.