A non-destructive method for estimating above-ground forest biomass in threatened woodlands

Current techniques for calculation of biomass in agroforestry require felling of many trees. Such methods are not well suited to the natural environment, especially if the environment is subject to anthropic degradation and if the wood supply to local populations is at stake. The method we describe here was used in a socioeconomic and ecological study of the biomass of a thuriferous juniper woodland (Juniperus thurifera L.) in the High Central Atlas mountains (Morocco). This computerized method reconstructs the different component volumes of a tree from two orthogonal-view photographs. Then, using the volume and the density of each component, it estimates the biomass of the tree. Regression curves were established between dendrometric parameters for 102 trees (tree height x crown projection area) and their estimated biomass by this computer method. A second-order polynomial equation gave the best regressions with a high coefficient of determination (R-2 = 0.96). To validate the method, the biomass of seven trees (cut in a previous study) was compared with those obtained from our regression equations, while in a second study, the biomass of three trees, estimated from photographs using the computer program, was compared with values obtained after felling and weighing the same trees. The results show that the method is reliable with a mean error percentage varying between 2.5 and 7.5 per tree. For young trees, the accuracy of the biomass could be improved, for example, by using a tree morphology classification. (C) 2000 Elsevier Science B.V. All rights reserved.