Mapping Land Cover Types for Highland Andean Ecosystems in Peru Using Google Earth Engine

Highland Andean ecosystems sustain high levels of floral and faunal biodiversity in areas with diverse topography provide varied ecosystem services, including the supply water to cities downstream agricultural valleys. Google (™) has developed a product specifically designed for mapping purposes (Earth Engine), which enables users harness computing power cloud-based solution near-real time land cover change monitoring. We explore feasibility using this platform types topographically complex terrain highly mixed vegetation (Nor Yauyos Cochas Landscape Reserve located central Andes Peru) classification machine learning (ML) algorithms combination different sets remote sensing data. The were trained 3601 sampling pixels (a) normalized spectral bands between visible near infrared spectrum Landsat 8 OLI sensor 2018 period, (b) indices vegetation, soil, water, snow, burned bare ground (c) topographic-derived (elevation, slope aspect). Six ML tested, CART, random forest, gradient tree boosting, minimum distance, naïve Bayes support vector machine. results reveal that produce accurate classifications when are used conjunction topographic indices, resulting better discrimination among classes similar signatures such as pajonal (tussock grass-dominated cover) short grasses or rocky groups, moraines, forested areas. model highest explanatory was obtained from forest algorithm (Kappa = 0.81). Our study presents first approach its kind Cordilleran we show GEE is particularly useful large-scale monitoring mountainous subject rapid changes conversions, replicability scalability other characteristics.