Land-use dynamics in a Brazilian agricultural frontier region, 1985-2017

摘要

Increasing demands for essential human needs (e.g. water, food, and energy) put pressure on natural resources, intensifies land use competition, and create challenges regarding environmental conservation. Understanding such dynamics is instrumental for the implementation of public policies in a more rational and planned manner, ensuring a better balance between uses and the maintenance of natural habitats, as well as fulfilling the demands of a rapidly growing society. Within this scope, in this study we applied statistical sampling techniques to remote sensing data (Landsat images from 1985 to 2017) in order to retrieve information on the dynamics of land use in one of the most prominent agricultural frontier regions in the world. Specifically, we chose the Brazilian state of Goias, located in the Cerrado biome, which is a biodiversity hotspot for its natural value and high endangerment level. Our results yielded estimates showing, in a first moment (1986 - 1996), the suppression of native vegetation concomitant with the expansion of pasture areas. From 1997 onward, this dynamic coexisted with the advance of annual crops and sugarcane over pasture areas (which serve as important land reserves). Our study corroborates the potential of using image-based statistical sampling to understand, in an efficient, agile, unbiased, and accurate manner, the dynamics and history of land uses in any region of interest. Likewise, the synergy between such low cost approach and other datasets (e.g. agricultural surveys) can increase the precision of land-cover and land-use assessments, complementing missing information and refining the identification of occupation patterns and trends. In particular, area estimates, as those generated in this study, can serve to quickly assess the outcomes of public policies focused on reducing deforestation rates and improving the use efficiency of areas already converted.