Performance of dNDVI, dNBR and dMIRBI spectral indices in burnt areas detection: Case study of Moyowosi game reserve, Kigoma, Tanzania
2025
KONGO, LAMECK SIMON
Context and background Fire is a major problem facing protected areas. This is due to several human activities such as agricultural activities taking place nearby those areas. Traditional method or remote sensing techniques can be employed in detecting fire severity in those areas. Detecting burnt areas by traditional method is cost fully and time consuming thus remote sensing approach has been employed in detecting those areas. This study has assessed different spectral indices on their performance on detecting burnt areas in wetland environment. Indices selected were difference normalized burnt ratio (dNBR), differenced normalized vegetation index (dNDVI) and mid-infrared burn index (MIRBI) derived from Landsat 8 images. Goal and Objectives: Assessing the performance of dNDVI. dNBR and MIRBI spectral indices in detecting burnt areas in wetland environment. Methodology: Spectral indices were extracted from Landsat 8 OLI images (2018 and 2019) and active fire data (VIIRS) were used to validate the performance on extracted indices. Results: From the obtained spectral indices, values were classified into four categories (unburnt, low severity, moderate severity and high severity) which were used for discriminating levels of burn severity. The areas which were highly affected by fire were selected and their corresponding active fire data were clipped. A number of counts of active fire data falling within classified highly bunt area were recorded. A total of 1278 counts about 84.6% under MIRBI, 1272 counts about 84.2% under dNBR and 973 counts about 64.4% under dNDVI. Thus highlighted the superior performance of the MIRBI in assessing burned areas in wetland environment.
Afficher plus [+] Moins [-]Informations bibliographiques
Cette notice bibliographique a été fournie par Institut Agronomique et Vétérinaire Hassan II
Découvrez la collection de ce fournisseur de données dans AGRIS