20 - Reconstructing and Modeling Past Changes in Terrestrial Primary Productivity

Abstract

Reconstruction of net primary productivity (NPP) from pollen data by statistical methods is a useful tool for assuming a relationship between NPP and taxonomic composition. This approach, practicable in regions with a lot of data, has some limitations too. Several parallel approaches have attempted to estimate the global terrestrial carbon storage of the Last Glacial Maximum: the large differences between these approaches prove the difficulty of the task. To mitigate this lack of data, this chapter proposes an approach based on model simulations, controlled by data. Waiting for more sophisticated earth system models, BIOME3 is used asynchronously coupled with a climate model to simulate final vegetation distributions and NPP. Evidently, there is yet no reliable way to reconstruct past global patterns of NPP. Data-based reconstruction is possible in regions where the data network is exceptionally dense. Biogeochemical models can be used to convert empirical regional climate and biome reconstructions to NPP and carbon storage, taking into account external factors including atmospheric carbon dioxide content. Global analysis, however, requires a complete “forward modeling” approach, based on climate model simulations coupled to an equilibrium biogeography/biogeochemical model. Nevertheless, past data can be used to test the climate and vegetation models by experimentally assigning to them conditions drastically different from the present one. If the models are reasonably able to simulate these paleoenvironments, they prove to be robust and one can have some confidence in them for predicting the future.