How To Assess The Future Tree-Cover Potential For Reforestation Planning In Semi-Arid Regions? An Attempt Using The Vegetation Model ORCHIDEE

Abstract:

More and more reforestation projects are undertaken at local to continental scales to fight desertification, to address development challenges, and to improve local living conditions in tropical semi-arid regions. These regions are very sensitive to climatic changes and the potential for maintaining tree-covers will be altered in the next decades. Therefore, reforestation planning needs predicting the future "climatic tree-cover potential": the optimum tree-fraction sustainable in future climatic states.

Global circulation models projections provide possible future climatologies for the 21st century. These can be used at the global scale to force a land-surface model, which in turn simulates the vegetation development under these conditions. The tree cover leading to an optimum development may then be identified. We propose here to run a state-of-the-art model and to assess the span and the relevance of the answers that can be obtained for reforestation planning.

The ORCHIDEE vegetation model is chosen here to allow a multi-criteria evaluation of the optimum cover, as it returns surface climate state variables as well as vegetation functioning and biomass products. It is forced with global climate data (WFDEI and CRU) for the 20th century and models projections (CMIP5 outputs) for the 21st century. At the grid-cell resolution of the forcing climate data, tree-covers ranging from 0 to 100% are successively prescribed. A set of indicators is then derived from the model outputs, meant for modulating reforestation strategies according to the regional priorities (e.g. maximize the biomass production or decrease the surface air temperature). The choice of indicators and the relevance of the final answers provided will be collectively assessed by the climate scientists and reforestation project management experts from the KINOME social enterprise (http://en.kinome.fr). Such feedback will point towards the model most urging needs for improvement.