<![CDATA[topography, and ocean-land interactions, which shape weather patterns and rainfall intensity variability. This study analyzes rainfall patterns in Sulawesi Island from 1981 to 2015 using the Empirical Orthogonal Function (EOF) method and composite analysis with machine learning. The results show that the EOF method successfully identifies three primary modes of rainfall variability. EOF Mode 1 captures negative anomalies, while EOF Mode 2 and EOF Mode 3 capture both positive and negative rainfall anomalies. EOF Mode 1 is the dominant component, explaining nearly 70% of the total variance. EOF Modes 2 and 3 capture additional variations on a smaller scale, and collectively, these three modes explain 88.53% of the total rainfall variability. Meanwhile, composite analysis reveals that global factors such as ENSO and the Indian Ocean Dipole (IOD) also influence rainfall variability, impacting drought periods and extreme rainfall events. During El Niño and positive IOD phases, rainfall deficits occur, potentially leading to prolonged droughts. Conversely, during La Niña and negative IOD phases, Sulawesi experiences a significant rainfall surplus, increasing the risk of hydrometeorological disasters such as floods and landslides.]]>
Copyrights © 2025
