Bioclim Deliverable D4/5: global climatic characteristics, including vegetation and seasonal cycles over Europe, for snapshots over the next 200,000 years

The aim of the BIOCLIM project is to develop and present techniques that can be used to develop self-consistent patterns of possible future climate changes over the next million years (climate scenarios), and to demonstrate how these climate scenarios can be used in assessments of the long-term safety of nuclear waste repository sites. Within the project, two strategies are implemented to predict climate change. The first is the hierarchical strategy, in which a hierarchy of climate models is used to investigate the evolution of climate over the period of interest. These models vary from very simple 2-D and threshold models, which simulate interactions between only a few aspects of the earth system, through general circulation models (GCMs) and vegetation models, which simulate in great detail the dynamics and physics of the atmosphere, ocean, and biosphere, to regional models, which focus in particular on the European region and the specific areas of interest. The second strategy is the integrated strategy, in which intermediate complexity climate models are developed, and used to consecutively simulate the development of the earth system over many millennia. Although these models are relatively simple compared to a GCM, they are more advanced than 2D models, and do include physical descriptions of the biosphere, cryo-sphere, atmosphere and ocean. This deliverable, D4/5, focuses on the hierarchical strategy, and in particular the GCM and vegetation model simulation of possible future climates. Deliverable D3 documented the first step in this strategy. The Louvain-la-Neuve 2-D climate model (LLN-2D) was used to estimate (among other variables) annual mean temperatures and ice volume in the Northern Hemisphere over the next 1 million years. It was driven by the calculated evolution of orbital parameters, and plausible scenarios of CO2 concentration. From the results, 3 future time periods within the next 200,000 years were identified as being extreme, that is either significantly warmer or cooler than the present. The next stage in the hierarchical strategy was to use a GCM and biosphere model, to simulate in more detail these extreme time periods. This deliverable starts with a description of the GCM and biosphere models used, and continues with an assessment of the models' performance in simulating the present day climate. It then gives a summary of the boundary conditions that constrain the model for the future climate simulations, as indicated by the LLN-2D model, and describes how these boundary conditions are implemented in the models. The majority of the deliverable is a description of the GCM results for the future climate simulations, and the resulting vegetation distributions. This is followed by a section focusing on the areas of specific interest. Finally, there is a critical discussion, which includes a comparison of some of the results with previous work, and an assessment of the uncertainty in the model results. There are many figures in this deliverable, all of which are included at the back. However, not all of the potentially interesting data from the models is discussed or presented. Therefore, data from all the simulations are available on the business collaborator at http://cobweb.businesscollaborator.com/bc/bc.cgi