Índice de Umidade Perpendicular (PMI), NDVI e KERNEL na temporada de incêndios no Parque Nacional da Chapada dos Guimarães/MT

Abstract

In recent years, Brazil has seen an increase in the occurrence of fires, especially in the Cerrado, a biome highly susceptible to this type of event. The use of geotechnologies, such as remote sensing data and Geographic Information Systems (GIS), allows us to understand the phenological stage of vegetation, before and after fires, and has proven to be fundamental for public policies to prevent forest fires. The objective of this work was to understand the variability of the Perpendicular Moisture Index (PMI), NDVI and Kernel, in order to develop an annual metric of the duration of the fire season in the Chapada dos Guimarães National Park (PNCG) from 2003 to 2023. For this, fire outbreak data provided by the BDQueimadas/INPE and FIRMS/NASA databases were used, combined with the PMI anomaly (bands 4 and 5 of the MOD9/MODIS product) and images from the Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) Landsat 8 sensors, between August and September 2019, to calculate the Normalized Difference Vegetation Index (NDVI) and the fire maps from the Kernel estimator present in QGis 3.22.6, in order to evaluate the distribution and pattern of fires in the proposed area. The spatial patterns of the PMI between the months of July, August and September show lower values in the months of August and September of 2015 and 2019, compared to the same period in 2017. In the September 2019 compositions, the values were lower compared to 2015 and 2017. However, in relation to the fire outbreaks that occurred over the 20 years, these did not coincide with the most affected areas, as shown by the Kernel density. This reflects human interference, resulting from anthropogenic actions in the most impacted region of the park. The NDVI varied between 0.229 and 0.0050, reflecting the vegetation response to the fires that occurred up to September 2019. Regarding the spatialization of the fires in the fire dynamics, a concentration of hot spots with very high density was observed in the Chapada dos Guimarães National Park, originating from the Buffer Zone (ZA) of the PNCG, mainly in the years 2015, 2017 and 2019. The PMI proved to be effective in predicting fires, based on vegetation moisture maps, corroborating the NDVI results. The Kernel density estimate allowed identifying the areas with the highest frequency of forest fires in the park. It is concluded that the PMI was useful for understanding the spatial distribution of vegetation moisture and identifying areas prone to forest fires, as well as the NDVI for evaluating areas after fires. However, to improve the quality of the analyses, it is essential to consider human actions in the region. Based on the results obtained, it was found that there was a need to intensify public actions in conservation units, with strategies to control access to areas most prone to forest fires, aiming to mitigate the impacts of human actions in the region.