This thesis deals with climate variability in the Nordic Seas during the Holocene derived from numerical model simulations, and compared to proxy-based reconstructions. The main focus of the research presented in this thesis aims at describing and understanding climate variability involving decaying ice sheets, in particular the effect of the Greenland ice sheet and its corresponding melt water during times of warmer than present-day climates. Another aspect of melting ice sheets is the effect of rising sea levels and the corresponding flooding of former land areas, such as the Siberian shelf in the Arctic Ocean. These two influences on the variability of the Nordic Seas climate have been studied in this thesis during the Holocene period, using numerical simulations performed with the LOVECLIM global climate model of intermediate complexity. The Holocene started about 11.700 yrs before present (BP, 1950 A.D.) and followed the preceding cold period, the last glacial. During the transition from a glacial to an interglacial state the climate system reorganises and large land-based ice sheets start to melt and release huge amounts of freshwater to the surrounding oceans. Initiated by the impact of increased orbital-based summer insolation on the Northern Hemisphere, the ice sheets decay and a thermal maximum is reached in the early Holocene. This thermal maximum is found in numerous proxy-based reconstructions of the Northern Hemisphere and it has been previously shown by Renssen et al. (2009) that the decaying ice sheets affected the spatial and temporal structure of the thermal maximum. Including such a factor improves the comparison to proxy-based reconstructions around the North Atlantic. However, the Greenland ice sheet (GIS) has been mostly neglected in this context, leading to the following question, addressed in the second chapter. What is the impact of GIS melting on the Holocene Ther- mal Maximum? Our model results show that the effect of melt water on the surface ocean is a stratification and subsequent cooling (up to 2◦C), as well as an increase of sea-ice cover. The cooling is stronger in the western Nordic Seas compared to the eastern, thus altering the spatial structure of the warmest temperatures and the timing of the Holocene thermal maximum in the Nordic Seas. The simulation results agree better with proxy-based reconstructions from this area and confirm that the Greenland ice sheet is important on a regional scale during its time of melting in the early Holocene. Related to effects of melt water from ice sheets is the rise of global sea level. However, there are regional differences and lower sea level stands have been reconstructed for the Siberian shelf during the early Holocene. An extrapolation of regional sea level stands to the whole Arctic indicates that large shelf areas must have been land in the early Holocene. Thus, the question arises: What impact does the flooding of an Arctic shelf introduce in the Nordic Seas? The Arctic shelf areas are zones of sea-ice production and supply the Arctic ocean and it’s main gateway, the Fram Strait, with sea ice. This major route of sea-ice export through Fram Strait and along the East Greenland current as far as the tip of Greenland posses a strong influence on the climate of the Nordic Seas. These cold and sea-ice covered waters on the western side are opposed by warm and salty Atlantic waters on the eastern side of the Nordic Seas. Our model results show that the flooding of the shelf increases the sea-ice production (15%), but decreases the Fram Strait sea-ice export (-15%). Contrary to our hypothesis, changes to the Nordic Seas’ climate are not caused by increased sea-ice export, but by local sea-ice production and forcing from the atmospheric winds. The conversion of land to ocean impacts the atmospheric circulation and enforces the Greenland winter High, yielding stronger winds. In response, the production and transport of sea-ice is increased, resulting in lower surface temperatures (up to 4◦C) during the whole year in the Nordic Seas. Related to all feedbacks the transport of heat to the higher latitudes is reduced, thus showing the importance of this in reality dynamic feedback to an orbitally-induced warming and rising of sea level. In direct consequence of results from our previous questions, we found that melting of the Greenland ice sheet has impacted the Atlantic Meridional Overturning Circulation (AMOC). The AMOC transports heat and salt from the tropical Atlantic to the Northern high latitudes and thereby contributes to a warmer climate over Europe compared to similar latitudes such as for example in North America. The recent efforts to measure the AMOC’s strength allow better insights into the complex ocean circulation and proves to be valuable information for assessing future changes. Expanding the history of measurements is a much-needed effort to increase confidence into future projections. Employing proxy relations to reconstruct past ocean circulation properties is a valuable tool that allows giving qualitative estimates of changes and indicates long-term trends. However, given the nature of these reconstructions, trends differ by proxy and location. Therefore, using numerical model simulations to compare to proxy-based reconstructions, allows us to address the following question in Chapter 4. What was the strength of the AMOC and its subcompo- nents during the Holocene? The climate and the ocean circulation in the North Atlantic region changed over the course of the Holocene, partly because of disintegrating ice sheets and partly because of an orbitally-induced summer insolation trend. In the early Holocene, the Laurentide ice sheet in North America, cooled large parts of the Northern Hemisphere by melt fluxes into the North Atlantic and by the cooling effect of the remnant ice sheet itself. In the Nordic Seas, this impact was accompanied by a rather small, but significant, amount of Greenland ice sheet melting that reduced local ocean surface temperatures even further. We compared transient simulations of the Holocene AMOC strength with proxy-based reconstructions. The modelled ocean circulation in the North Atlantic evolved from a melt water-impacted reduced state in the early Holocene to an evenly strong present-day state by 7,000 yrs BP. This evolution is partly confirmed by proxy-based reconstructions employing carbon isotopes, however showing a weaker response during the melting phase in the early Holocene compared to the model, whereas a comparison to proxy-based reconstructions focusing on the deep ocean currents is more problematic and only a few records agree with model results. The overall comparison shows that different trends in subcomponents of the AMOC can agree with each other and represent a long-term evolution of the AMOC similar to our model simulations. These results thus suggest that the evolution of the AMOC was mostly stable over the Holocene, with reduced values only in the early part. Our results on the early Holocene are relevant in the context of presentday warming and its impact on the contemporary Greenland ice sheet. However, the early Holocene was not the only time period in the past, when the Greenland ice sheet was melting and in particular the Last Interglacial is often referred to as a period of extreme Greenland ice sheet melting. Given the fact that there is evidence of Greenland ice sheet melting in the past and a large likelihood in future projections, we address the fourth and final question in Chapter 5. What is the impact of GIS melting beyond the Holocene and what can we learn for the future? The climate system responds to, for example, a radiative forcing in every component of the climate system with certain positive or negative responses, called feedbacks. Thus, the sensitivity of the climate system to a forcing is determined by the response of the individual components. When comparing past and future warm climates, it is important to account for the forcing that makes the climate warm. In the case of the considered past climates, the orbitally-induced insolation is redistributed to cause for example warmer summers as well as cooler winters in the Northern Hemisphere. The response of the climate system, however is than seasonally dependent and not as straight forward as one might anticipate. In the case of the future climate, with the projected increases in greenhouse gas concentrations, the forcing is more globally uniform compared to the Holocene and the Last Interglacial. When we consider the impact of Greenland ice sheet melting in these past and future climates in our climate model, we find different responses in the melt-water-reduced AMOC strength for the past climates compared to the future projections. This indicates that the origin of this difference is likely to be related to the nature of the radiative forcing and melt water-related feedbacks in the different components of the climate system. The most likely candidate is the dynamic response of the sea ice to Greenland ice sheet melting in a set of differently warm climates. However, this highlights that a simplified extrapolation from past climates towards the future might be misleading, as this dynamic feedback is likely to be different in the future compared to the past. This thesis provides a detailed introduction into the field of climate as well as individual introductions to the specific questions asked before and a final chapter summarising the major findings and future research suggestions.