Simpler is more effective than complex for a GIS-based simulation of deforestation for a conservation project in Madagascar

Keywords: land use change, deforestation, REDD++
Session Type: Virtual Paper Abstract
Day: Monday
Session Start / End Time: 2/28/2022 02:00 PM (Eastern Time (US & Canada)) - 2/28/2022 03:20 PM (Eastern Time (US & Canada))
Room: Virtual 56

Abstract

Land change simulation models play a fundamental role in quantifying the effectiveness of conservation projects for Reducing Emissions due to Deforestation and forest Degradation (REDD). A simulation model must attain validation accuracy measured by the Figure of Merit to qualify for REDD according to the Verified Carbon Standard (VCS). My research uses a proposed conservation project in Madagascar as a case study to compare five simulations of deforestation for a REDD project. The first simulation allocates deforestation on the highest carbon density, the second on the lowest carbon density, the third near previous deforestation, the fourth according to a multivariate model of seven drivers, and the fifth uniformly. The multivariate model has a Figure of Merit of 7%-10% while the simpler proximity model has a Figure of Merit of 10%-12%, thus both models qualify according to the VCS criterion. The rate at which both models disturb carbon is nearly identical to the rate for the uniform model.