Comparative Analysis of Single and Multiple Change Points Detection for Streamflow Variations

Abrupt changes in streamflow patterns significantly affect hydrological systems, making their detection critical for effective water resource management. This study uses the annual maximum streamflow (AMS) data to analyze and identify changes at the Kajang Station in the Langat Basin, Selangor, Malaysia. The objective of research is to determine the precise years of abrupt changes in streamflow and examine the underlying factors causing them. The problem lies in the increasing frequency and intensity of streamflow changes that could be related to factors such as changes in land use and climate variation, which require detailed investigation. This study conducts six complementary statistical tests to identify change points in the streamflow data. Six complementary statistical tests were conducted to identify change points. The Pettitt test (statistic: 276, p-value: 0.001), Buishand range test (statistic: 1.5881, p-value: 0.034), and standard normal homogeneity test (statistic: 11.349, p-value: 0.009) consistently identified 2003 as a significant single change point. For multiple change points, the sequential Mann-Kendall test indicated shifts in 2002 and 2007. The multiple structural change method and classification and regression trees revealed significant change points in 1985, 2003, and 2009. These changes are likely due to the 1982 massive flood event and subsequent changes in land use and river encroachments. The findings underscore the importance of monitoring and managing river systems, especially given the rapidly occurring environmental changes. It is vital to understand these change points to develop more resilient strategies for water resource management.