Modelagem de mudança de cobertura e uso do solo no município de Londrina

Abstract

Geotecnologies have become an important tool to understand the environmental dynamics and to simulate tendencies from future scenarios. The analysis of projections and of usage and coverage of the soil allows the manager to verify if the current actions are sustainable or need to be altered to prevent future problems, like the decision making assistance. The objective of this study was to model the dynamics of soil usage and coverage through the probabilistic model “Pesos de Evidência”, applied to the remote sensing images, for the north region of Londrina city, between the period of 2005 and 2010. A simulation for the soil coverage scenery for the period between 2015 and 2020 was done, from the change rates obtained in the modeling process. The simulation for the year of 2015 was done to validate the results, comparing them with a current image, and apply then to the 2020 ones. Landsat 5 satellite, sensor TM, images from 2005 and 2010 and images from the satellite Landsat 8, sensor OLI, for the year 2015 were used, as well the tools of the software Spring and Dynamic-EGO. The modeling applied requires as input data thematic maps of soil cover usage, which were generated from the classification of the images. The classes of soil usage and coverage determined in this paper were: agricultural area, urban area, vegetation and water. The method of “Pesos e Evidência” is based on the Bayes approach and has as premise to define the influence degree of a set of factors to model the landscape changes. With this method, it can be quantitatively evaluated, how favorable or unfavorable each variable is, to a certain changing of soil usage and coverage. The following factors were analyzed: distance to hydrograph, distance to roads, distance to the urban class, altimetry. The simulation showed a good degree of concordance with the real image, of 2015, because the similarity seen was of 68%. In face of this, the simulation of the future scenary of 2020 was done to observe how the changings of soil usage and coverage would affect the study region. The 2020 prognosticate landscape shows a tendency of the 7% growth of the urban class around the periphery area of the city, the vegetation class to expand 5% in the south region of the landscape, and the agricultural class to diminish in 12%.