Seasonal patterns strongly influence traditional agriculture in Indonesia; therefore, climate change is likely to have a significant impact on crop production. This study explores the dependency relationships between meteorological conditions and agricultural yields in Indonesia by integrating meteorological data with agricultural yield data. The datasets are collected from a wide range (2010 - 2024) at the district level. The meteorological data are obtained from 100 meteorological stations across Indonesia. The proposed approach employs the Peter-Clark (PC) algorithm to generate causal graphs and an Intelligent Kernel K-Means (IKKM) method to classify regions based on similarities in meteorological conditions and agricultural yields. IKKM is effective for mapping regions according to shared climatic and yield characteristics. This study examines five major agricultural commodities (cocoa, coffee, oil palm, cayenne, and paddy). The IKKM method successfully groups each dataset into three clusters, achieving an average Silhouette score of 0.35. The resulting causal graphs reveal dependency relationships between meteorological variables and crop yields. The dependent relationships indicate that rising temperatures are likely influence the declining yields of cocoa, oil palm, and paddy. Panel regression results indicate statistically significant (p-value Formula: see text) negative effects of temperature on agricultural land use and crop productivity. Minimum temperature significantly reduces cocoa land area, while minimum, maximum, and average temperatures negatively affect oil palm productivity. Average temperature also has a significant adverse impact on paddy productivity. These findings serve as a warning that climate change may directly and indirectly affect agricultural industries, with potentially severe impacts on key regions that contribute substantially to annual crop production.
Handhayani et al. (Mon,) studied this question.