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

Remote sensing with large-scale satellite images for precision studies in grasslands has spatial and spectral resolution limitations. Against this, using spectral signs and vegetation indices obtained with microsensors transported by unmanned aerial vehicles (UAV) constitutes a more accurate alternative for biomass estimation. In the fieldwork, images were acquired with microsensors, and fixed transects of 100 m were used where vegetation samples were collected. The photographs acquired with the UAV were processed in Pix 4D, Arc Gis, and algorithms elaborated in R programming language. The biomass estimation was carried out with Multiple Linear Regression, Vector Support Machine, and Random (Forest Random) models. The Random model showed a Kappa coefficient of 0.94 in the training set and 0.901 in the test set (R2 = 0.482). The Random Forest model predicted 3 g/pixel of MV for Puna grass in the rainy season and 2 g/pixel for the dry season; the predicted biomass for the Tola bush was 15 g/pixel of MV for both seasons of the year. The estimation of biomass/hectare for the tolar plant community with its tola shrub and Puna grass components was 6,535.88 kg/ha for the rainy season and 6,588.81 kg/ha for the dry season. The difference between the biomass estimated in the field and the biomass estimated with Random Forest was 5.48% for the rainy season and 9.63% for the dry season.

La teledetección y los sistemas de información geográfica son herramientas que en la última década se utilizan con énfasis en la gestión de recursos naturales; sin embargo, éstas han presentado deficiencias para estudios de ganadería de precisión debido a la calidad de las resoluciones espacial, espectral y temporal de las imágenes, frente a esta limitación aparece como alternativa los microsensores en sistemas aéreos no tripulados (UAS) que permiten obtener ortofotografías con mejores resoluciones. Considerando estas ventajas se desarrolló un estudio para determinar la mejor altura de vuelo de las UAV en la detección e identificación de las comunidades vegetales tolar y bofedal de puna seca. Para el estudio se recopilaron fotografías RGB y NDVI con sensores ZENMUSE X3 DJI RGB-NDVI en UAS con alturas de vuelo de 25, 50, 75 y 100 m. En el campo se contabilizaron plantas de tola y cojines de DIMU en cuadrantes de 10 m x 10 m (100 m²). La preparación de ortofotografías se realizó en el software Pix 4D y para analizar la información se elaboró un algoritmo con capacidad de identificar un elemento segmentado (planta de tola y/o cojín de DIMU) utilizando el lenguaje de programación Python. El estudio determinó que el rango de NDVI para la identificación de tolares de Parastrephia lepidophilla es de 0,20 a 0,45 y para bofedales de Distichia muscoides es de 0,68 a 0,95; finalmente usando ortofotografías RGB y NDVI se determinó que la mejor altura de vuelo para identificar las especies segmentadas Tola y DIMU es de 25 m seguido de 50 m.

While the grazing of livestock has occurred for millennia in the Andes, current sustainability debates center on concerns with co-managing climate change and pastoralism. These discussions have special resonance in places protected by the state for biodiversity, scenery, and sustainable and traditional land uses, such as those found in protected areas and biosphere reserves. For this article, we integrate data from a social-ecological research project on the land use systems that affect high-elevation ecosystems in Peru’s Huascarán National Park, with special emphasis on the wetlands. We used land cover and land use data and insights from interactions with pastoralists to show that (1) wet meadows dominate the lower reaches of the park, while peatlands predominate above 4000 m elevation; (2) wet meadows are most useful for traditional grazing systems, while the peatlands are especially susceptible to trampling by livestock; and (3) there is limited ecological space at the highest elevations for the successful future upward relocation of either land use or potential habitats for species identified as of concern. We explore the implications of these findings for the adaptive strategies of biophysical and social processes in terms of livelihoods and biodiversity in and around a protected area. We conclude that there are many additional opportunities to be explored to inform the management of ecosystem services and provide improvements for the adaptive capacity of communities and park managers.

This study evaluated the use of visible and near-infrared (Vis-NIR) spectroscopy combined with machine learning (ML) algorithms to predict soil fertility-related properties in two contrasting agroecological regions of Peru: the Highlands and the Rainforest. A total of 297 soil samples were analyzed using portable spectroradiometers covering a spectral range of 350–2500 nm, applying transformations such as Savitzky–Golay smoothing, first derivative, and band depth. Predictive models were developed using PLSR, Random Forest, Support Vector Machines, and neural networks. Results show variable predictive performance across soil properties and ecosystems. Organic matter in Highland soils and calcium in Rainforest soils achieved the strongest test-set accuracy (R² > 0.70), while pH and texture fractions showed moderate performance (R² = 0.42–0.67), and mobile nutrients including phosphorus, potassium, and sodium showed limited predictive accuracy due to their weak spectral expression. Spectral predictions were further integrated into a structured nutrient balance framework to assess agronomic reliability. Nitrogen fertilizer recommendations showed the strongest agreement between observed and predicted values across both ecosystems, whereas K₂O and CaO recommendations in Highland soils were substantially underestimated, demonstrating that property-level statistical performance does not guarantee agronomic reliability. These findings confirm that Vis-NIR spectroscopy combined with ML represents a fast, cost-effective, and sustainable alternative to conventional soil analysis, especially in rural areas with limited laboratory infrastructure. Expanding regional calibration datasets and exploring mid-infrared FTIR spectroscopy as a complementary technology are identified as priority directions for improving predictions of agronomically critical nutrients.