3D Pattern Characterization of Rainfall Trends and Change Point Detection in an Indian River Basin, Using Variable-Size Cluster Analysis

Abstract

Floods and droughts are significantly impacted by rainfall, a vital component of the hydrological cycle. This study evaluates long-term rainfall trends using variable-size cluster analysis (VSCA) to examine trends and change points over eight synoptic stations of the Damodar River Basin from 1922 to 2021. The Mann–Kendall (MK) test with Sen's slope estimator reveals monotonic trends and magnitudes, and VSCA analyzes rainfall patterns and change points. Changing climate statistics were summarized using a modified Pettitt–Mann–Whitney test version. Rainfall patterns that changed over time were shown graphically using 3D representations for 100 years of data with a minimum cluster size of 10. VSCA analysis showed a declining trend in rainfall beginning about 1990, with notable variations in 1970–1980 for Bardhaman, Dhanbad, Giridih, and Hazaribag. On the other hand, Koderma and Purulia had rising patterns starting in 1970 and lasting roughly from 1960 to 1980. Most of the time, West-Medinipur showed both declining and no-trend conditions. The MK test and Sen's slope technique revealed a significant negative trend in rainfall, with magnitudes of −1.28, −1.03, −1.67, −0.61, −2.54, and −1.92 mm/year for Bardhaman, Dhanbad, Giridih, Hazaribag, Ramgarh, and West-Medinipur, respectively. Purulia and Koderma displayed rising trends with magnitudes of 0.84. This research enhances our understanding and provides valuable insights for managing water resources.