Estudio Avanzado Del Efecto De Borde A Través De Indicadores Microclimáticos En Areas Protegidas Del Piedemonte Llanero, Villavicencio - Meta

Abstract

It was considered for this study that, in fragmented forests of the Colombian Llanos foothills, edge effects occur that generate abiotic alterations and, in turn, modify microclimatic conditions, influencing the ecological dynamics of forest fragments and overall forest health. This study aimed to determine the edge effect in forest fragments located within the Cuenca Alta del Caño Vanguardia and Cerro Vanguardia National Protective Forest Reserves, using microclimatic variables in order to generate technical inputs to guide conservation and management strategies for forest fragments. Multispectral images from the Landsat 8 (OLI/TIRS) satellite, provided by NASA, were used to characterize the fragments, establishing three transects per fragment across a total of eight study units. Edge effect detection was performed using segmented regression, which allowed the identification of inflection points in microclimatic gradients. This analysis was complemented by the Kruskal–Wallis test, used for statistical comparison between edge, transition, and core zones. Additionally, the correspondence between satellite-derived and in situ data was evaluated using error metrics such as RMSLE, BIAS, and the Pearson correlation coefficient, in order to validate the accuracy of satellite data in environmental monitoring based on field data collection. The results revealed the presence of microclimatic gradients associated with the edge effect in forest fragments, detected through both in situ measurements and satellite imagery. It was determined that NDMI showed significant spatial differences that allow the identification of edge effects, whereas in situ air temperature did not present statistically significant variations between transects. Regarding validation, a limited correspondence was observed between satellite data and field measurements, with high variability in errors, the presence of biases, and weak correlations. This indicates that, although remote sensing allows the identification of spatial patterns, it presents limitations in accurately estimating microclimatic conditions at the local scale; therefore, its application should be complemented with in situ measurements for an adequate interpretation and management of fragmented ecosystems.