Resumen

HERNANDEZ-RAMOS, Jonathan et al. Inventory and mapping of forest variables through remote sensors in Quintana Roo state, México. Madera bosques [online]. 2020, vol.26, n.1, e2611884.  Epub 30-Jun-2020. ISSN 2448-7597.  https://doi.org/10.21829/myb.2020.2611884.

Remote sensors in combination with information derived from forest inventories estimate variables of interest with precision and low cost. The objective was to estimate the basal area (AB), timber volume (VTA) and aboveground biomass (B) in different forest ecosystems using Landsat ETM information and National Forest and Soil Inventory (INFyS) in Quintana Roo, Mexico. A correlation matrix was generated between INFyS data and spectral information, and later, a multiple linear regression model. With the selected equations, spatial distribution maps of AB (m2 ha-1), VTA (m3 ha-1) and B (Mg ha-1) were generated. The total inventory was estimated using three approaches: i) Reason Estimators (ERaz), ii) Regression Estimators (EReg), and iii) Estimators of Random Simple Sampling. The first two approaches correspond to the alternative inventory using remote sensors and the third corresponds to the traditional inventory. The correlation coefficient was greater than the normalized difference index with 0.35, 0.39 and 0.39 for AB, VTA and B. The regression models had adjusted determination coefficients of 0.28, 0.32 and 0.32 to estimate AB, VTA and B, respectively. The three estimators are statistically different and show that the EReg is the most conservative and with precision in AB, VTA and B of 2.73%, 2.92% and 2.71%, respectively, in addition to confidence intervals of smaller amplitude than the MSA and ERaz. By updating the inventory using remote sensors, the process of evaluating forest resources and their planning is improved.

Palabras llave : Aerial biomass; forest structure; Landsat; regression models; remote sensing.

        · resumen en Español     · texto en Español     · Español ( pdf )