Análise da dinâmica do uso e cobertura da terra nas áreas desflorestadas do estado do Pará por meio da plataforma Google Earth Engine

Abstract

The Amazon Forest has suffered intense changes in the use and coverage of the Earth in recent years, mainly due to the incentives of territorial occupation policies and macroeconomic issues involving logging, livestock and agricultural crops. Among the States that make up the Brazilian Legal Amazon, the State of Pará contributed the most to the deforestation of the Amazon Forest with 148.399 km² of deforested areas until the year 2018. Thus, the objective of this work was to evaluate the dynamics of the use and land cover of the deforested areas of the State of Pará between 1988 and 2018 through the Google Earth Engine (GEE) platform. For the execution of this work, scripts were developed in the JavaScript programming language with application of the spectral indexes (vegetation and water) and the Random Forest algorithm in the GEE Code Editor. The Landsat-5/TM satellite images of the year 1988, 1998 and 2008 and the Landsat-8/OLI-TIRS satellite image of the year 2018 were used for mapping the land use and land cover. In this work we have mapped the classes: Agriculture, Forest, Hydrography, Others, Pasture and Secondary Vegetation. The validation of the mappings occurred with the analysis of the contingency matrix in order to verify the efficacy and the errors (of inclusion and omission), so the accuracy can be obtained through the coefficients of agreement (Global Accuracy, Kappa and Tau Index) and fieldwork. Between 1988 and 2018 there was deforestation of approximately 139.970,65 km², of which 48.3% became pasture, 28,2% were secondary vegetation, 22,6% were other, and 0,9% were agriculture. However, the agriculture class was the one that obtained the highest percentage growth in the last thirty years, with an increase of 674,1%, while pasture and secondary vegetation obtained gains of 165,3% and 192,4%, respectively. With this we can conclude that pasture is advancing on recent deforestation and yielding area for agriculture to a lesser extent. In relation to the accuracy of the data, the mappings obtained good results with precision from 89,8% to 94,8%, with ratings considered to be very good according to the results of the Kappa and Tau indices. It is worth noting that GEE is an important tool in the monitoring and control of illegal deforestation because it allows data acquisition in a fast and safe way, contributing with data for the development of public environmental policies in order to guarantee the maintenance of ecosystems, Amazonian Forest that remain, until the present day, conditioned by a scenario of extreme conflict of interests, involving actors, articulations and local to global conjunctures.