Silvina Rosita Yulianti

Department of Statistics, Universitas Sebelas Maret, Jl. Ir. Sutami 36A, Surakarta 57126, Indonesia

Sri Sulistijowati Handajani

Department of Statistics, Universitas Sebelas Maret, Jl. Ir. Sutami 36A, Surakarta 57126, Indonesia

DOI: https://doi.org/10.19184/mims.v26i1.60017

ABSTRACT 

The climate in Indonesia is becoming more unpredictable and has caused serious damage to many important sectors, especially agriculture. Floods, droughts, and plant pests have affected agricultural land and created major risks for the future of agriculture in Indonesia. This study aims to examine agricultural damage risk by developing a composite agricultural damage index and analyzing its changes over time. A composite index is used because it can provide a broader view of damage trends than looking at each hazard separately. This study used secondary data on the ratio of damaged agricultural land from Statistics Indonesia (BPS) and the Sustainable Agricultural Insurance Scheme Survey Report (Bappenas-JICA) for 2003-2017. The index was constructed using equal weights and principal component analysis (PCA) weights. Trend significance was tested using the Mann-Kendall test, Sen’s slope estimator, and ordinary least squares regression, while change points were examined using the Pettitt test. The equal-weights index (EWI) was used as the main measure, while the PCA-weighted index (PWI) was used for comparison. The results show that the EWI has no statistically significant monotonic trend, while the PWI shows only a weak declining tendency at the 10% level. The Pettitt test also detects no significant change point in either index. The main pattern is strong year-to-year variability rather than a clear long-term trend. The 2015 drought appears as a severe single hazard, but not as a peak in the composite indices, showing that composite indices may hide the effect of one dominant factor.

Keywords: Agriculture, climate, composite index, PCA, trend.

MSC2020: 62M10