Análisis multitemporal de imágenes sentinel-2 para la cobertura vegetal de la zona andina de la provincia de el oro, ecuador

Abstract

The Andean area of the province of El Oro (ZAEO), belongs to the southern region of Ecuador, which is characterized by great variability and biological diversity. The objective of this study was to analyze the changes in vegetation cover and land use in the ZAEO. Two Sentinel-2 (MSI) level 1C images were used, for the period 2017-2020. A supervised classification of maximum likelihood was carried out, in an altitudinal range from 900 to 3900 masl, using 4 of the 6 level 1 classes proposed by the Ministry of the Environment of Ecuador, for each class 16 training areas (ROI) were elaborated. A general precision of 0.855 was obtained for 2017 and 0.944 for 2020. 120 pixels were used for the elaboration of the confusion matrix, where the Kappa index for 2017 was 0.6870, while for 2020 it was 0.9058. Finally, the rate of change between 2017 and 2020, there were increases in the classes: forests (BO), anthropic zone (ZA) and herbaceous and shrub vegetation (VAH), with values of 43.90%, 9.52% and 4.22%, respectively, while agricultural land (TA) recorded losses of 9.43%.