Ekstraksi otomatis data spasial wilayah terbakar untuk persiapan rehabilitasi hutan dan lahan pasca kebakaran di Kawasan Suaka Alam Riam Kanan Provinsi Kalimantan Selatan

Abstract

The Natural Reserved Area Riam Kanan is experiencing forest and land fires that occur during each dry season. This will result in the potential of land to be rehabilitated every year. The purpose of this study is to automatically extract the extensive data and the severity of forest and land fire incidents occurring in KSA Riam Kanan. Future benefits are to identify and determine priorities for forest and land rehabilitation activities after the fire incident. The method used is Normalized Burn Thermal Ratio (NBRT) transformation on Landsat Image 8. Method accuracy tested using NASA FIRMS MODIS Thermal Anomalies / Fire Locations data, and land closure data of multispectral classification on Landsat Image 8. Based on the result of research, it is found that, from August 12, 2015 to October 15, 2015, the total area of burned land in KSA Riam Kanan is 28,133,73 hectares. With the classification of the severity of fire incident Low severity burn 27,134.5 hectares, Moderate to low severity burn 967.59 hectares, Moderate to high severity burn 27.32 hectares, and High severity burn 4.32 hectares. While the category of burned land is 888.07 hectares of forest, Cultivated Agricultural Crop 4,065.44 hectares, and Shrub 23,062.22 hectares. The results of the accuracy test show that the accuracy of NBRT classification to extract spatial data of fire is 96.77%, and the accuracy of the mapping of the NBRT method is 99.58%. Wrong area detection error occurs in cloud and open field objects.