Analisis Spasial Indeks Bahaya Kebakaran Hutan dan Lahan pada Gambut dan Non-Gambut di Kabupaten Ogan Komering Ilir, Sumatra Selatan
Edra Elricadewi, Dr. Like Indrawati, S.Si., M.Sc.
2026 | Tugas Akhir | D4 SISTEM INFORMASI GEOGRAFIS
Kabupaten Ogan Komering Ilir memiliki wilayah gambut yang luas dan berpotensi besar mengalami kebakaran hutan dan lahan. Perbedaan kondisi lingkungan dan aktivitas manusia dapat menyebabkan variasi tingkat bahaya kebakaran secara spasial. Penelitian ini bertujuan memetakan indeks bahaya kebakaran hutan dan lahan, menganalisis distribusinya pada wilayah gambut dan non-gambut, serta mengevaluasi hasil pemodelan menggunakan data kejadian kebakaran independen. Metode Multi-Criteria Decision Analysis (MCDA) digunakan melalui pendekatan weighted overlay. Parameter bahaya mencakup karakteristik penutupan lahan, jenis tanah, curah hujan, jarak terhadap jalan, dan jarak terhadap permukiman. Pembobotan dilakukan menggunakan Frequency Ratio (FR), Relative Frequency (RF), dan Prediction Rate (PR) berdasarkan pola hotspot konsisten VIIRS NOAA-20 periode 2021 sampai 2025. Hasil menunjukkan bahwa curah hujan memiliki bobot terbesar sebesar 41,73%, diikuti jarak terhadap permukiman 19,02%, penutupan lahan 18,38%, jenis tanah 11,13%, dan jarak terhadap jalan 9,75%. Model A didominasi kelas bahaya sedang sebesar 83,47%, sedangkan pada Model B wilayah gambut didominasi kelas sedang sebesar 75,05 persen dan non-gambut sebesar 96,37 persen. Evaluasi dengan Prediction Rate Curve menghasilkan AUC sebesar 0,5764 pada Model A dan 0,6484 pada Model B, yang masing-masing termasuk kategori poor dan average.
Ogan Komering Ilir Regency has extensive peatland areas and is prone to forest and land fires. Differences in environmental conditions and human activities may result in spatial variations in fire hazard levels. This study aims to map the forest and land fire hazard index, analyze its distribution across peat and non-peat areas, and evaluate the model results using independent fire occurrence data. The method employed was Multi-Criteria Decision Analysis (MCDA) using a weighted overlay approach. The parameters consisted of soil type, rainfall, land cover, distance to roads, and distance to settlements. Parameter weighting was performed using the Frequency Ratio (FR), Relative Frequency (RF), and Prediction Rate (PR) methods based on consistent VIIRS NOAA-20 hotspot patterns from 2021 to 2025 to determine the relative weights of each parameter. The results show that rainfall had the highest weight at 41.73%, followed by distance to settlements at 19.02%, land cover at 18.38%, soil type at 11.13%, and distance to roads at 9.75%. Model A was dominated by the moderate hazard class at 83.47%, while in Model B, the peatland area was dominated by the moderate hazard class at 75.05% and the non-peatland area at 96.37%. The prediction rate curve evaluation resulted in AUC values of 0.5764 for Model A and 0.6484 for Model B, which were classified as poor and average, respectively.
Kata Kunci : kebakaran hutan dan lahan, indeks bahaya, frequency ratio, gambut, weighted overlay