Resumen
This research examined the use of fire index algorithms to detect and recognize the burnt area in West Kalimantan by applying the pre-fire and post-fire image comparison technique. The main data used were derived from remotely sensed data MODIS acquired from Januari to April 2014. The examined algorithms utilized the near-infrared (NIR) and short-infrared (SWIR) wavelength spectrums. in the case of forest and land fires, occured the value of NIR decreases as the amount of chlorophyll decrease, while the pixel values and the inceasing value of SWIR will increase due to the rising temperature. The research objective was to the capability of the algorithms in detecting burnt forest and land areas in several selected areas in West Kalimantan, using few indices generated from MODIS data. The examined indices were NDFI (Normalized Difference Fire Index) and MNDFI (Modified Normalized Difference Fire Index), which utilize the reflectance values of band 2 (NIR) and band 7 (SWIR) from MODIS. The study results show that both the NDFI and MNDFI were applicable in detecting burnt area having good performance with the Normalize Distance (D) values larger than 1. Based on D-Value and accuracy assessment, MNDFI algorithm gave better index than the NDFI in detecting both forest and land areas.