Estimación de atributos forestales mediante teledetección en bosques mixtos de Durango, México

Abstract

Remote sensing images with fi eld data (forest inventory) provided more accurate information in assessing and monitoring forest variables compared to the conventional one. The objective of this study was to compare two methods of resource assessment in the mixed forests of Durango, Mexico: (i) conventional inventory (CI) using the classic Random Simple Sampling (MSA, abbreviation in Spanish) estimator and (ii) combination of inventory information and Landsat ETM spectral data with ratio and regression estimators to assess biomass, volume and basal area. Likewise, with information on forest variables (AB: basal area, volume and B: biomass) and the spectral bands and vegetation index, regression models were adjusted; In addition, the total inventory of forest variables was estimated using two methods: (i) alternative inventory through remote perception using ratio and regression estimators; and (ii) with the CI, using the MSA. Maps were generated explaining the spatial variability of each one of the variables of interest. The results indicated that the adjusted regression models showed excellent statistical bases to estimate B, V and AB, which allowed the construction of maps that described the spatial variation of the parameters in the entire forest area of interest. All the regression models were highly signifi cant at 95% reliability with the hypothesis test of the parameters and the adjusted determination coeffi cients of 0.58, 0.66 and 0.59 for the AB m2 ha-1, V m3 ha-1 and B Mg ha-1, respectively. A conservative inventory allows sustainable planning and exploitation of forest resources in a managed forest.