Deliverable D7: Continuous climate evolution scenarios over western Europe (1000 km) scale. Work Package 2, Simulation of the future evolution of the biosphere system using the hierarchical strategy. Modelling Sequential Biosphere Systems under Climate Change for Radioactive Waste Disposal (BIOCLIM)

The overall aim of BIOCLIM is to assess the possible long term impacts due to climate change on the safety of radioactive waste repositories in deep formations. This aim is addressed through the following specific objectives: • Development of practical and innovative strategies for representing sequential climatic changes to the geosphere-biosphere system for existing sites over central Europe, addressing the timescale of one million years, which is relevant to the geological disposal of radioactive waste. • Exploration and evaluation of the potential effects of climate change on the nature of the biosphere systems used to assess the environmental impact. • Dissemination of information on the new methodologies and the results obtained from the project among the international waste management community for use in performance assessments of potential or planned radioactive waste repositories. A key point of the project is therefore to develop strategies for representing sequential long-term climatic changes by addressing time scales of relevance to geological disposal of solid radioactive wastes. The integrated strategy, which first step is described in this deliverable (D7), consists of building an integrated, dynamic climate model, to represent all the known important mechanisms for long term climatic variations. The time-dependent results will then be interpreted in terms of regional climate using rulebased and statistical downscaling approaches. Therefore, the continuous simulation of the climate evolution of the next 200 000 years selected for study is a major objective of the BIOCLIM project. This requires models that account for the simultaneous evolution of the atmosphere, biosphere, land-ice and the ocean. To be able to perform several 200 000-yearlong transient climate simulations, the models have to include all these components, but also need to be simple enough to run fast. Therefore, climate models of intermediate complexity have been chosen to complete this part of the BIOCLIM project. In the present deliverable, we report on the results of two such models, MoBidiC (Louvain-la-Neuve) and CLIMBER-GREMLINS (LSCE). The overall objective of the work presented here is the simulation of the climate of the next 200 000 years for three different CO2 scenarios [Ref.1]. However, both models used for this work have been either modified for the project (MoBidiC) or developed within the project (CLIMBERGREMLINS). Therefore their performance, and the modifications and developments needed to be documented, especially as far as their ability to reproduce past and different climates is concerned. Therefore, a large section of the present deliverable is devoted to the evaluation of the models through past climate simulations. The deliverable is structured as follows: first, a brief description of the models is given. In the second section, results from the models for past climate situations are presented. The third section deals with the future climate simulations devised for the BIOCLIM project: for each CO2 scenario, the results of the two models are compared. It is emphasized that the model results, especially those for CLIMBER-GREMLINS, should be regarded as illustrations of possibilities rather than absolute predictions of climate evolution. The novel approach to long-term climate change adopted in BIOCLIM is based on research tools under continuing development, notably, the CLIMBER-GREMLINS model.