Mapeamento e avaliação de qualidade em pastagens na bacia hidrográfica do rio Tijuco (Triângulo Mineiro)

Abstract

Faced of a prominent agricultural production on the world stage, some practices of appropriation of natural resources induce negative impacts on the Brazilian environment, including degradation in pastures. The country, assuming a voluntary commitment to reduce GHG emissions in agriculture, created in 2010 the ABC Program (Low Carbon Emissions Agriculture), which since 2011 has provided credit to rural producers so that they can adopt practices as recovery of degraded pastures. However, few studies to verify the efficiency of this program have been carried out so far, once that mapping pasture, especially, the quality of it in the Cerrado, requires the development of computational approaches and accurate data. From this perspective, the current study aims to improve the mapping of pastures and quality pastures techniques in the Rio Tijuco basin, located in the Triângulo Mineiro, and to evaluate the efficiency of the ABC Program in that region. Therefore, in the moment of pasture mapping, image segmentation was performed with Sentinel 2/MSI images from 2020, later associated with an object based classification using Google Earth Engine and Geographic Information System tools. In the step of quality evaluation in pastures, we used data from the Sentinel 2 Satellite and MSI sensor, and calculated the average of the Normalized Difference Vegetation Index (NDVI), from a set of 600 images, along two years (2019-2020), and compared them with field data as well. The total pasture area found was of 5.533,47 km² corresponding to 38.39% of the total area studied. The municipalities of Prata and Ituiutaba were found to be very relevant for the region’s livestock farming. Moreover, it was possible to identify 3 levels of degradation ("Moderate/Strong", "Slight" and "None"). NDVI values less than 0.32 indicate "Moderate/Strong" degraded pastures; those corresponding to the interval between 0.32 and 0.42 can be considered as having a “Slight” degradation level and NDVI values greater than 0.42 represent not degraded pastures. By using the intervals for mapping the quality of pastures in the Tijuco river basin, we identified a very good accuracy, corresponding to the Kappa Index of 0.77. Regarding the quality assessment of pastures on properties that captured investments from the ABC Program in the study area, the results indicated a gradual increase in NDVI values over the years observed, although the study has not been conclusive, since climatic factors can interfere with the results.