Uso de imágenes satelitales en la estimación del volumen de madera de plantaciones de Pinus radiata D.Don. parroquia Palmira cantón Guamote, provincia de Chimborazo

Abstract

This research aimed to estimate the volume of wood using a database obtained in the field and spectral data from satellite images from the Sentinel-2 satellite of the Pinus radiata D. Don plantation in the town of Palmira, Guamote city, province of Chimborazo. The non-parametric Random Forest technique was used to estimate wood volume. The most important predictive variables were the Digital Terrain Model (DTM), Sentinel-2 band 10, the Digital Surface Model (DSM) and the Structure Insensitive Pigmentation Index (SIPI). The Random Forest model presented a coefficient of determination (𝑅2) of 0.80 and a value of the root mean square error (RMSE) of 23.86 𝑚3ℎ𝑎−1. These indicators were able to estimate the total volume of the plantation (𝑚3ℎ𝑎−1) with 84.13% accuracy. Spectral information from a medium resolution sensor in combination with field data is a viable alternative to estimate the volume of wood in forest plantations.