SPI-BASED DROUGHT PREDICTION: COMPARING ARIMA AND MARKOV CHAIN MODELS

Nur Azizah Affandy (Departement Civil Engineering, Faculty of Sains and Technology, Universitas Islam Lamongan, Indonesia)
Muhammad Feri Erfinansah (Department of Civil Engineering, Faculty of Sains and Technology, Universitas Islam Lamongan, Indonesia)
Rifky Aisyatul Faroh (Department of Electrical Engineering, Faculty of Sains and Technology, Universitas Islam Lamongan, Indonesia)
Nur Nafi’iyah (Department of Informatics Engineering, Faculty of Sains and Technology, Universitas Islam Lamongan, Indonesia)
Entin Hidayah (Department of Civil Engineering, Faculty of Engineering, Universitas Jember, Indonesia)
Dedy Dwi Prastyo (Department of Statistics, Faculty of Science and Data Analytics, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia)
FX Suryadi (IHE Delft, Institute for Water Education Delft, The Netherlands)

AL-ULUM: JURNAL SAINS DAN TEKNOLOGI
Vol 12 No 2 (2026)

Drought poses severe risks to agricultural sustainability in climate-vulnerable regions. This study compares AutoRegressive Integrated Moving Average (ARIMA) and Fuzzy Time Series–Markov Chain (FTS-MC) models for multi-timescale Standardized Precipitation Index (SPI) forecasting in the Slahung sub-watershed, East Java. Historical monthly rainfall data were transformed into SPI series to model seasonal behaviors and uncertainty transitions. Evaluated via Mean Squared Error (MSE), ARIMA demonstrated superior numerical accuracy in capturing temporal and seasonal trends. Conversely, FTS-MC provided complementary probabilistic insights into drought state transitions under high climatic variability. Integrating these deterministic and probabilistic approaches enhances drought-prediction robustness, offering water resource managers a comprehensive framework for climate risk mitigation.