Characterization of anthropic intervention levels in paramo ecosystems by multispectral image analysis

Abstract

Vegetation plays an important role as an indicator of the wetland ecosystem state. In that regard, remote sensing provides the possibility to determine the ecosystem state, allowing reducing the time consumption of wetland assessment. Nonetheless, satellite images from high mountain regions are not commonly used because those images present a high level of cloudiness. Besides, the spatial resolution does not allow an adequate classification of the vegetation types, generating several difficulties to perform a ground analysis on high detailed-scale. In this work, it is presented an assessment of high-Andean wetlands using multispectral images acquired from an unmanned aerial vehicle. The study region is located at El Ocho sector, the buffer zone of the Los Nevados National Natural Park, in wasteland ecosystem with an altitude of 3500 m above sea level, and 20 ha of area. The methodology comprises three stages: a) The wetland was divided into four sectors according their anthropic affectation level; b) the vegetation indexes were obtained from aerial images captured using a RPAS (visible and multispectral), and then, the green index was calculated in the wetland and compared against NDVI in testing patches; Finally, c) it was performed the correlation between the green index and the environmental damage. As a result, the computed indexes showed a high correlation with the ecosystem intervention levels.