Predicting Agricultural Drought Vulnerability: A Machine Learning Approach Using Sentinel-2 NDDI Time Series in Rainfed Agroecosystems, East Java (2020–2025)
Keywords:
Agricultural Drought Prediction, NDDI, Machine Learning, Sentinel-2, Rainfed Agriculture, Early Warning SystemAbstract
Rainfed agricultural systems in tropical monsoon regions face increasing drought risks under climate variability. This study develops a predictive framework for agricultural drought vulnerability using Sentinel-2 derived Normalized Difference Drought Index (NDDI) time series (2020–2025) integrated with machine learning algorithms in Jatirogo Subdistrict, East Java. Unlike previous mapping-focused studies, this research applies Random Forest (RF) and Support Vector Machine (SVM) models to forecast drought severity classes one month ahead based on historical NDDI patterns, land use, and soil parameters. Results demonstrate that the RF model achieves superior predictive performance (accuracy = 87.3%, Kappa = 0.81) compared to SVM (accuracy = 79.6%, Kappa = 0.72 ), with the most important predictors being NDDI values from the preceding two months (importance score = 0.34) and land use type (importance score = 0.28). Forecasts for the 2025 dry season accurately predicted severe drought conditions in Kebonharjo and Sugihan villages with 89% spatial agreement. The study identifies a critical threshold: when August NDDI exceeds 0.45, the probability of severe drought in September–October reaches 0.82. The proposed predictive framework provides operational lead time for adaptive planting schedules, water resource allocation, and early warning systems, offering a replicable methodology for drought -prone rainfed agricultural regions across Southeast Asia .
References
Aldyan, R. A. (2023). The impact of climate change on water resources and food security in Indonesia.
Journal of Law, Environmental and Justice, 1(1), 50 –63.
Auliyani, D., Wahyuningrum, N., Supangat, A. B., & Basuki, T. M. (2022). Prediction of drought vulnerability and its mitigation measures in the Northern Coast of Java, Indonesia. IOP Conference Series: Earth and Environmental Science, 1016(1), 012003.
Badan Meteorologi, Klimatologi, dan Geofisika (BMKG). (2023). Analisis iklim dan kekeringan di Jawa Timur. Jakarta: BMKG.
Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5 –32.
Cahyono, B. E., Rahagian, R., & Nugroho, A. T. (2023). Analisis produktivitas padi berdasarkan indeks kekeringan (NDWI dan NDDI) lahan sawah menggunakan data citra sentinel-2A di Kecamatan Ambulu. Indonesian Journal of Applied Physics, 13(1), 88.
Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine Learning, 20(3), 273–297.
Davoodi, E., Soleimani-Motlagh, M., & Chamanpira, R. (2025). Prediction of temporal and spatial variations of the NDDI drought index in the Khorramabad watershed using remote sensing. Environmental Resources Research, 13(2), 275 –292.
Dikshit, A., Pradhan, B., Huete, A., & Park, H.-J. (2022). Spatial based drought assessment: Where are we heading? A review on the current status and future. Science of The Total Environment, 844, 157239.
Dinas Pertanian Kabupaten Tuban. (2022). Data pertanian Kabupaten Tuban. Tuban: Dinas Pertanian. Firdaus, R. A., Hermawan, E., & Kamilah, N. (2024). Implementasi metode Normalize Difference Drought
Index (NDDI) terhadap pemantauan produktivitas pertanian tanaman padi (Studi kasus: Kecamatan
Jonggol tahun 2019-2022). Infotech Journal, 10(1), 147–160.
Food and Agriculture Organization (FAO). (2023). The impact of disasters on agriculture and food security.
Rome: FAO.
Funk, C., Harrison, L., Shukla, S., Pomposi, C., Galu, G., & Korecha, D. (2023). Drought early warning in Africa: A review of current and emerging approaches. Bulletin of the American Meteorological Society, 104(2), E511-E531.
Gholinia, A., & Abbaszadeh, P. (2024). Agricultural drought monitoring: A comparative review of conventional and satellite-based indices. Atmosphere, 15(9), 1129.
Gu, Y., Hunt, E., Wardlow, B., Basara, J. B., Brown, J. F., & Verdin, J. P. (2023). Evaluation of MODIS NDVI and NDWI for vegetation drought monitoring using Oklahoma Mesonet soil moisture data. Geophysical Research Letters, 35(22).
Irsyad, F., Sari, N., Putri, A. E., & Filipović, V. (2025). Application of the Normalized Difference Drought Index (NDDI) for monitoring agricultural drought in tropical environments. Land, 14(12), 2431.
Mechler, R., Singh, C., Ebi, K., Djalante, R., Thomas, A., James, R., ... & Schinko, T. (2022). Loss and
damage from climate change: A new frontier for research and policy. Climate Policy, 22(5), 551 -566. Mishra, A. K., & Singh, V. P. (2020). A review of drought concepts. Journal of Hydrology, 391(1 -2), 202–
216.
Mokhtar, A., Jalali, M., He, H., Al-Ansari, N., Elbeltagi, A., Alsafadi, K., ... & Abuarab, M. (2022). Machine learning-based drought prediction in semi-arid regions. Agricultural Water Management, 271, 107809.
Nguyen, M. H., Dao, D. T., Le, M. S., & Le, T. H. (2024). A modification of normalized difference drought index to enhance drought assessment using remotely sensed imagery. Environmental Monitoring and Assessment, 196(10), 883.
Nugroho, A. T., Cahyono, B. E., & Rahagian, R. (2022). Validasi indeks kekeringan NDDI untuk pemantauan kekeringan lahan sawah di Jawa Timur. Jurnal Ilmu Tanah dan Lingkungan, 24(2), 45 – 58.
Park, H., Kim, K., & Lee, D. K. (2022). Random forest-based agricultural drought prediction using meteorological and remote sensing data. Remote Sensing, 14(15), 3658.
Rahmi, N., Mubarak, H., Faridah, S. N., & Sapsal, M. T. (2025). Drought level analysis of paddy fields
using the NDDI method based on Sentinel-2A imagery in South Polombangkeng District, Indonesia. Salaga Journal, 61–67.
Ruqoyah, R., Ruhiat, Y., & Saefullah, A. (2023). Analisis klasifikasi tipe iklim dari data curah hujan
menggunakan metode Schmidt-Ferguson (Studi kasus: Kabupaten Tangerang). Jurnal Teori dan Aplikasi Fisika, 11(1).
Siswanto, S., Wardani, K. K., Purbantoro, B., Rustanto, A., Zulkarnain, F., Anggraheni, E., ... & Dimyati,
M. (2022). Satellite-based meteorological drought indicator to support food security in Java Island. PLOS ONE, 17(6), e0260982.
West, H., Quinn, N., & Horswell, M. (2021). Remote sensing for drought monitoring and assessment: A
review of progress and operational applications. Remote Sensing, 13(8), 1521.
Wilhite, D. A., Sivakumar, M. V. K., & Pulwarty, R. (2021). Managing drought risk in a changing climate.
Climate Risk Management, 33, 100342.
Submitted to: International Conference on Climate Resilience and Agricultural Adaptation (ICCRAA) 2026.
