Reporting period October 2002 -Dec. 2003 Work package 1 Activities Work package 2 Activities Work package 3 Activities Work package 4 Activities Work package 5 Activities

Work package 1 Activities :

The aim of WP1 was to summarise what was known about the causes of climate change in order to set the scene for following work. The methods previously used to represent climate change in biosphere assessments and problems with such approaches were also set out (Deliverable 1). The original technical work plan was completed as planned.

Present day biosphere system descriptions were provided of the European areas that formed the focus for the climate change work (eastern France, central England, central/southern Spain, plus Czech Republic and Germany). A large effort was also put into collating and documenting relevant palaeological information for the same regions so that there was a common basis for the rule-based downscaling work that was to be developed later in the project (Deliverable 2). The original work plan was extended and completed.

Work package 2 Activities :

The general purpose of WP2 was to develop a hierarchical strategy for representing sequential climate changes to the geosphere-biosphere system for selected time slices on time scales relevant to the geological disposal of solid radioactive waste, i.e. typically up to one million years. Different models were used in the framework of this work package: an earth model of intermediate complexity (EMIC) (LLN-2D-NH model) to provide the long-term evolution of the climate for the Northern Hemisphere; a general circulation model (GCM) (IPSL_CM4_D) to provide a more detailed global view of the climate; and a regional climate model (RCM) (MAR model) to provide an even more detailed view at the regional scale. The three main phases of the work are described below.

First, the climate of the next million years was simulated using the UCL model LLN-2D-NH according to different scenarios for natural and anthropogenic greenhouse gas forcing conditions. Then, six time slices were selected to be further studied in later tasks with the more detailed GCM models (LMDz and IPSL_CM4_D). Discussion amongst all the project participants about the most appropriate time slices and the editing process of the deliverable delayed the completion date. Nevertheless, the work was completed and reported in Year 1 (Deliverable D3).

Second, the GCMs were used to provide a more detailed picture of the climate and ocean dynamics over Europe. Output was also used for input information in conjunction with the ORCHIDEE vegetation model to derive changing vegetation patterns accompanying the climate changes. The work was delayed because of the delay in appointing a key post-doctoral climate modeller at CEA/LSCE to help with the BIOCLIM model development. All the planned work was eventually completed as planned (Deliverable 4+5).

Third, a regional climate model (MAR) was used to produce estimates of local climate conditions at the selected sites/areas. Input to MAR was generated by the IPSL_CM4_D model. As both GCM and MAR models had to be adapted to the needs of the project, new test experiments had to be carried out. All climate simulations were completed as planned (Deliverable 6A).

In WP2, an overview of statistical downscaling methods, their underlying assumptions and advantages/disadvantages were reported in Deliverable 6B. Specific issues relating to downscaling applications within the BIOCLIM context (i.e., application to the IPSL_CM4_D snapshot simulations) were identified, for example, the issue of stationarity, i.e., the extent it is possible to apply relationships between the present-day large-scale and regional/local climate to the specific situations of the BIOCLIM snapshot simulations, particularly those with no Greenland Ice Sheet or more extensive ice sheets than present. The predictor and predictand data sets that would be required to implement these methods within the BIOCLIM hierarchical strategy were also outlined. Implementation of statistical downscaling techniques in WP2 was delayed in order to give priority to the application of the rule-based downscaling method developed in WP3 to WP2 EMIC output. This task was not originally planned, but allowed more comprehensive comparison and evaluation of the BIOCLIM scenarios and downscaling methods to be undertaken and this was reported in D10-12.

Work package 3 Activities :

The goal of WP3 was to provide long term climatic scenarios for future changes, with simplified, but still physically based models, the so-called "Earth Models of Intermediate Complexity" (EMICs). The two models were still in development in climate research laboratories at the start of the project, and using them for the BIOCLIM project was a challenge. The two models and associated tasks in WP3 were:

• MoBidiC - Full coupling between the atmosphere/ocean part of the model and the ice sheets was performed. Sensitivity experiments over a few millennia and simulations of interglacial climate, glacial climate and deglaciation processes were studied. The model was tested by comparing simulations for the last glacial-interglacial cycle with known data. Some fine-tuning was required. Then three climate simulations (designated A4, B3 and B4), each for a 200 ka, were undertaken and the results analysed.
• CLIMBER-GREMLINS - Coupling between the ocean/atmosphere model and the high resolution ice-sheet model was carried out – this took a number of months. Once achieved, test runs were performed and then the BIOCLIM simulations were undertaken. The same three simulations (A4, B3 and B4) were completed and the results compared with those obtained from MoBidiC.
  The extensive work required to modify or develop the two coupled models resulted in a time delay before the BIOCLIM climate simulations could be undertaken but the latter were completed as planned. Results obtained from the three climate scenarios using MoBidiC and CLIMBER_GREMLINS were documented in Deliverable D7. In addition, the report includes information on potential vegetation changes with changing climate. This obviated the need for Deliverable 9.

Downscaling – Two different downscaling methodologies were developed and applied in WP3 – rule-based method and a statistical/physical method.

Work for the rule-based downscaling approach involved an eight-step methodology.

1. Development of regional climatic sequences and indices for the last climatic cycle including Køppen/Trewartha climate classes.
2. Identification of relationships between climate states and Køppen/Trewartha climate classes.
3. Identification of appropriate analogue stations to describe each climate class.
4. Manipulation of analogue station data into appropriate formats for presentation of results and input to performance assessments.
5. Selection of appropriate MoBidiC simulations and variables for the identification of downscaling rules/thresholds.
6. Identification of objective rules/thresholds for defining climate classes in MoBidiC and LLN-2D-NH output.
7. Evaluation of the rule-based downscaling methodology for the last climatic cycle.
8. Application of the rule-based downscaling methodology to MoBidiC and LLN-2D-NH output for future time periods.

Some of the required input information came from the palaeo-climatological data collated in WP1; other data were derived from analogue climate stations. The methodology was applied to the output from MoBidiC and results reported in Deliverable 8A.

A second approach to downscaling was developed using both statistical and physical geographical parameters. The method was reported in Deliverable 8B. The objective of deliverable 8B was to devise a statistical downscaling method to obtain regional precipitation and temperature for the European study regions of interest. The new methodology was partly based on "physical" characteristics of the regions and partly on statistical (empirical) methods. It was recognized that a simple connection of coarse model results to regional results cannot be achieved on a purely empirical basis because there are not enough data available (neither from a climate model, nor from the real world). Therefore "physically based" relationships between large and regional scales were used to supplement the statistical approach. This physical/statistical method was applied to the results of three simulations performed with CLIMBER_GREMLINS covering 200 ka After Present.

Thus, BIOCLIM employed dynamical, rule-based and statistical downscaling methods as originally planned. A comparison of the outcome from the different downscaling methodologies was included as part of Deliverable 10-12.

Work package 4 Activities :

WP4 was concerned with application of the understanding of long-term climate change derived from Work Packages 1, 2 and 3 in a performance assessment context. The more detailed aims of Work Package 4 (WP4) were defined as being to:

• Provide a context for the synthesis of data from other work packages that will be useful in PAs;
• Show how those data are used to inform the development of biosphere system descriptions appropriate to PA;
• Investigate the potential radiological significance of transitions between different climatic states;
• Develop biosphere descriptions appropriate to selected climate states and transitions, and recommend how to derive conceptual models for radiological impact assessment;
• Develop recommendations on how the effects of climate change on the biosphere can be taken into account in PAs.

In order to provide an adequate description of biosphere systems for use in long term performance assessments, WP4 used the knowledge and understanding derived from the preceding work packages to identify and describe the potential effects of environmental evolution. In particular, the significance of dynamic environmental change due to transitions between climate states was investigated in relation to radionuclide migration in the biosphere. The biosphere systems representative of future conditions at the European regions of concern were described. The methods adopted were based on, and augmented, the Reference Biosphere Methodology developed within IAEA BIOMASS programme (BIOMASS, 2003).

Technical work on WP4 originally was not scheduled to commence until Month 25 (October 2002). However, it was recognised at Meeting 1 that the amount of work required to complete WP4 was substantial and that there was nothing to preclude an initial phase of the work being undertaken in parallel with the other Work Packages. In view of this, it was decided that development of the methodology to be used in WP4 would start early. Initial thinking relating to the development of such a methodology was included in a discussion document presented at Meeting 3. This discussion document was reviewed briefly at the meeting and extensive comments on the proposed outline methodology were subsequently received from various BIOCLIM participants. A synthesis of these comments was produced, together with an updated and substantially expanded discussion document. Both the synthesis of comments and the expanded discussion document were reviewed in detail at Meeting 4 (in January 2002) and a small working party was formed to develop the methodological approach to be adopted in WP4. Proposals from this working party were discussed at subsequent technical meetings. The methodology was further developed at two WP4 meetings and it was decided to combine the three deliverables into one document (subsequently designated as D10-12). This document includes detailed descriptions of the climatic implications of the BIOCLIM scenarios at the global, regional and local scales for the European regions of interest as well as a comparison of the various BIOCLIM downscaling methods.

Work package 5 Activities :

Work Package 5 activities were associated with the organisation of the final seminar. Extensive discussions were held between the ANDRA Project Co-ordinator and the EU representative and BIOCLIM participants to get ideas for the seminar format. At the suggestion of the EU representative, the idea of holding a joint seminar with the complementary EC BioMoSA project (that was also due to be completed at the same time as BIOCLIM) was pursued and eventually agreed upon. A joint final seminar was held at the EC offices in Luxembourg on 27 and 28 November 2003.

- " Reporting period October 2000 - April 2001 "
- " Reporting period April 2001 - September 2001 "
- " Reporting period October 2001 - March 2002 "
- " Reporting period April 2002 - September 2002 "
- " Reporting period October 2002 - December 2003 "