Dados do Trabalho

Título

MARIANA'S DAM COLLAPSE AND ITS IMPACTS ON ENVIRONMENTAL RESOURCES CHARACTERIZED BY REMOTE SENSING DATA

Resumo

Brazil's most significant environmental tragedy was the rupture of the Fundão Dam in the municipality of Mariana, Minas Gerais State, in 2015. Around 34 million cubic meters of toxic waste and mud were released into the Doce River, causing socioeconomic and environmental damage. Therefore, our research focused on plant and soil issues to characterize the impact of the Mariana mining disaster by remote sensing time series analyses. For that, we established a polygon over the disaster area dividing it into four lots to discuss the impacts step by step, following the path of the river and the most impacted areas. For that, a buffer covered the entire length where the problem occurred. Subsequently, we obtained Landsat 5 and 8 imagery in the dry seasons of 1989, 2003, 2010 and 2014 (before the disaster) and 2015, 2016, 2018, and 2020 (after the disaster). We converted the images to the Normalized Difference Vegetation Index (NDVI) to know possible changes in the land cover. In addition, we ran the GEOS3 system to detect exposed soil (SYSI) and bare soil frequency (BSF) over these periods. We identified higher NDVI values in the previous years of dam construction and mining company establishment. This index gradually decreased with mining activity expansion until 2015, when the dam collapsed. The NDVI abruptly fell from then on because the sediments covered the river course and soil near streams, and the images presented bare soil features. After 2018, these areas have gradually increased due to the re-establishment of vegetation. As also noted by BSF, SISY revealed a notable increase in soil sediments, particularly close to the dam. The work shows the importance of remote sensing in monitoring the diagnosis and environmental impacts.

Palavras-chave

reomte sensing, soil helath, spectroscopy, mining, carbon

Instituição financiadora

CNPQ 2022-2108, FAPESP