Optimizing the National Forest Inventory of Mexico
Elena Diez Pastor, University of Delaware, ediezp@udel.edu (Presenter)
Van Huong Le, University of Delaware, vanle@udel.edu
Rodrigo Vargas, University of Delaware, rvargas@udel.edu
The optimization of environmental sampling is now more accessible thanks to the development of computational techniques applied to this purpose. It is relevant to the extent that it simplifies the process and facilitates the survival of projects of this nature, saving valuable resources. The National Forest Inventory of Mexico monitors and characterizes the country's forest resources, despite experiencing a decrease in the funding necessary to carry out this activity. However, it is essential to help redesign their sampling strategy and reduce their monitoring effort, saving costs but recovering enough information on the measured variables. For this specific objective, we have applied a data-driven method for identifying optimized samples from spatial information of environmental data: the conditioned Latin Hypercube Sampling. We obtain a representative sample of the univariate and the multivariate probability distribution function. We prove that it is possible to significantly reduce the sample size, still representing the statistical and the spatial behavior of the variables of interest, and saving a relevant amount of resources. Moreover, these results allow the viability of a project that helps to quantify how the ecosystem services that exist in Mexico contribute to alleviate climate change.
Associated Project(s):
Poster Location ID: 21
Presentation Type: Poster
Session: Poster Session 1
Session Date: Wednesday (9/27) 1:15 PM