The spread of the temperature projections for the next century indicates a considerable uncertainty range for climate sensitivity of different climate models. Improving the knowledge about feedback processes (from e.g. natural and human-induced vegetation changes), regional climate change or the influence of external natural forcing factors (e.g. solar forcing) will narrow down the uncertainties. This will provide society and decision makers with a more solid basis to assess the urgency and necessary measures to mitigate future climate change.

RA2 specifically focuses on:

• Reconstruction of past land-cover, using pollen data and comparison to model simulations of natural vegetation and human-induced deforestation (model-data comparison)
• Studies of climate-vegetation/land-cover feedbacks in the past
• Comparison of climate-model simulations with proxy-based reconstructions of climate and circulation patterns
• Modelling of past natural climate change using up-to-date estimates of past natural external forcings (with focus on solar forcing)

Models and methods

• The REgional VEgetation Abundance from Large Sites (REVEALS) model for pollen-based quantitative reconstructions of past vegetation cover (natural and human-induced land-cover changes)
• Proxy-based reconstruction of circulation and hydroclimatic patterns for Fennoscandia (e.g. using stable isotope and tree ring data)
• Cosmogenic radionuclide-based solar activity reconstructions
• Application of global and regional climate models developed within RA1
• Application of extended climate models that use forward modelling of paleoproxies (e.g. water isotopes) and models that include the proposed relevant processes for studying the solar influence on climate

Cooperation within MERGE RA2 provides real-world data from the past that is crucial for testing the climate and dynamic vegetation models developed within RA1. Advanced statistical methods, including spatial statistics (RA4), are applied to interpret paleoclimatic information and to fill the gaps in the point datasets of pollen-based vegetation reconstructions to obtain spatially explicit land-cover descriptions that can be used in climate models.