Determinación de cambios de la cobertura arbórea usando imágenes satelitales Landsat 7ETM+A através de redes neuronales artificiales en la parroquia Achupallas, cantón Alausí, provincia de Chimborazo

Abstract

The current research proposed determining the tree coverage changes using Landsat 7ETM+ satellite images through artificial neural networks for Achupallas Parish, Alausí County, Chimborazo Province. The artificial neural networks (ARNs) are precise tools for estimating the forest changes occurring in previous and future periods. The supervised classification of satellite images were carried out determining six categories which are: paramo, forest, plantation, water bodies, crops, pastures. The exchange rate and net change produced in 1991, 2001 and 2011 were also determined as well as the creation of a prediction map for 2030 using the RNAs-based Land Change Modeler Module (LCM). For 1991 the forest covered 21.87% of the land, in 2011 they reached a 17.25%, and there is a projection of a 15.61% by 2030. In 1991 the plantations covered 0.06 hectares of land, in 2011 they reached 701 hectares and the projection 456 hectares by 2030. The paramo in 1991 covered a 65.46% of the land, but it changed to 60.65% in 2011, and a projection of 57.79% by 2030. The crops and pastures are the potential highest categories for the development, these are projected from 10% and 4.1% in 2001 to 20.38 and 4.96% in 2030 respectively. The tree coverage is represented by native forests and multiple-tendency plantations. While forests lose large areas, plantations experiment a slight growth, and the paramos invaded by crops and pastures is the category which experiments a highest loss. It is recommended to diffuse these results to adopt actions for protecting the natural ecosystems. Key words: Tree coverage, Satellite images, artificial neural networks, Pastures.