Assessing the impacts of catastrophic 2020 wildfires in the Brazilian Pantanal using MODIS data and Google Earth Engine: A case study in the world’s largest sanctuary for Jaguars

• Published in: Earth science informatics Vol. 16; no. 4; pp. 3257 - 3267
• Main Authors: Parra, Larissa M. P., Santos, Fabrícia C., Negri, Rogério G., Colnago, Marilaine, Bressane, Adriano, Dias, Maurício A., Casaca, Wallace
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-12-2023; Springer Nature B.V
• Subjects: Atmospheric correction; Change detection; Computation; Earth and Environmental Science; Earth Sciences; Earth System Sciences; Flora; Image filters; Information Systems Applications (incl.Internet); Land use; Missing data; MODIS; Ontology; Parks & recreation areas; Reflectance; Simulation and Modeling; Space Exploration and Astronautics; Space Sciences (including Extraterrestrial Physics; Temporal resolution; Wildfires; Pantanal biome; Remote sensing; MODIS; Fire-affected areas
• DOI: 10.1007/s12145-023-01080-x
• EISSN: 1865-0481
• ISSN: 1865-0473

The Encontro das Águas State Park (EASP), renowned as the world’s largest refuge for Jaguars (
Panthera onca
), is located within the Brazilian portion of the Pantanal biome, and it covers a vast area of approximately 1,080 square kilometers. This ecologically rich region suffered significant devastation from extensive fires in 2020. Given that the ongoing monitoring of wildfires is a crucial task for the preservation of fauna and flora in legally protected environments such as the Pantanal biome, this paper investigates the catastrophic 2020 fire incidents in the EASP reserve through a fully automated methodology capable of detecting and characterizing fire-devastated areas. By taking updated and accurate data from the Google Earth Engine platform, our approach integrates a comprehensive collection of MODIS sensor images, spectral indices, and filtering processes to generate a spatial map of fire-affected areas in a given period of analysis. Specifically, given a surface reflectance and atmospheric corrected MODIS (MOD09Q/A1) image series, the NBR index is computed from each image and then processed through Savitzky-Golay filtering to remove noisy and missing data. Next, the
Δ
NBR index is calculated for each consecutive pair of images so as to produce a frequency map of burned areas. In order to quantify and analyze the recent changes due to these successive wildfires that took place in this Pantanal portion, we focused on the devastating fire events that occurred in the EASP park from July to September 2020. The fire mappings were assessed and statistically validated using the kappa coefficient and significance tests computed through reference samples collected from official databases and visual inspection. The findings revealed that, tragically, 84% of the study area experienced at least one instance of fire during the three-month investigation period. The high temporal resolution of MODIS sensors proves to be extremely valuable in promptly and effectively detecting changes in land use.