| Popis: |
Tropical wetlands are ecosystems of great ecological importance, but their loss has accelerated in recent decades. Even though its hydrological, geological, and chemical characteristics make the Pantanos de Centla Biosphere Reserve (PCBR) the most significant wetland in Mesoamerica, it is the worst affected by forest fires area each year of Tabasco, Mexico. Satellite remote sensing is proposed as a cost-effective alternative to locate and delimit fire-affected surfaces in these difficult-to-access environments more precisely. The objective of the present study was to determine the optimal Landsat 8 spectral index for the detection of burned areas within the PCBR. It was possible to identify and extract 41 disturbed areas in six years. These areas were submitted to calculate six spectral burn indices, widely tested in detecting fire scars in forest areas, BAI, MIRBI, NBR, NBR2, NBRT1, NBRT3, including the best-known vegetation index NDVI. The capacity of each index to discriminate burned areas was estimated by comparing them with each other by using a separability index. The comparisons of the annual coefficients of variation allowed the evaluation of separability dispersion, indicating how homogeneous each index was. MIRBI was identified as the index with the highest potential for discrimination of burned areas in the PCBR, followed by far by NBR2. It is expected that these results can serve, among other things, to characterize, evaluate and prioritize monitoring areas to contribute by updating and improving the fire management plan in the PCBR. They are also expected to help subsequent studies on fire dynamics associated with human activities that cause and impact fire, such as burn severity, biomass loss, and post-fire vegetation regeneration. |
Full Text from ScienceDirect