The natural causes of climate variation (including abrupt change)
The combination of model predictions (Milankovitch, 1930-1940; Berger, 1970-1980) with measurements of diverse paleoceanographic proxies (Emiliani, 1950-1960; Shackleton, 1970-1990; and others) stated that the ultimate cause of climatic change is related with the orbital evolution of the planet. This context brings about a scenario of very slow climate evolution in which no variation could be observed within the current human life. However, the field data on marine sediments showed that transitions from glacial to interglacial periods occurred much faster than the slow orbital driving processes. This rapid climate transitions constitute a first indication that internal planetary processes may act on their own reinforcing and speeding climate transitions.

Further studies of paleoclimatic records during the past few decades have led to the discovery that our climate can exhibit abrupt changes with great speed, achieving sizable changes in a few centuries or even decades. Collaborations between several members of the GRACCIE team have contributed to understanding the impacts of these abrupt climate changes on continental and marine zones, particularly for the Iberian Peninsula. In fact, members of our team published the first ever literature report proving a close relationship between changes in atmospheric temperature increases in Greenland and changes in temperature increases of Mediterranean surface waters (Cacho et al., 1999). Said article established that these changes could comprise temperature variations of up to 6ºC within a few hundred years. Furthermore, pollen levels in marine sediments have indicated that the vegetation of the Iberian Peninsula was capable of transforming from predominantly Mediterranean arborea to steppe vegetation in a few decades (Sanchez-Goñi et al., 2002).

Further analyses of sediment samples from the Western Mediterranean by GRACCIE members showed that this type of abrupt climate variability also occurred in the last interglacial period, during which surface water oscillated by up to 10ºC within a few hundred years (Martrat et al., 2004; Figure 1). These results indicated that the frequency of abrupt transitions was lower during the last interglacial, but their intensity was significantly greater in relation to those which took place during the glacial age (Martrat et al., 2004).

Examination of the marine climate variability over a broader temporal perspective, e.g. the last 420 000 years, from sediment cores collected in the Iberian Margin showed that abrupt variability was a common robust feature over the past four climate cycles (Martrat et al., 2007). However, none of the climate cycles studied was an exact reproduction of another. In fact, SST variability increased while the Pleistocene progressed to the present. Furthermore, combined examination of surface and deep water proxies allowed to observing increases in deep sea floor ventilation with powerful arrival of north Atlantic deep water at the beginning of every climate cycle (Figs 2 and 3). Conversely, harsh drops in SST were preceded by steep decreases in deep water flows with changes from northern deep water to southern Antarctic Bottom Water predominance. Thus, abrupt changes occurred simultaneously with reorganization of the deep-water masses in the northern Atlantic Ocean and arrival of Antarctic Bottom Waters at latitudes such as those of the Iberian Peninsula. Due to the non-linear behavior of the ocean-atmosphere-sea ice system apparent gradual, external triggers or slow changes in the deep ocean are preceding rapid climate oscillations in the Mediterranean region at surface level (Grimalt and Martrat, 2008).

As concerns the present interglacial (Holocene), these types of oscillations could have a great impact on the future climate, perhaps even greater than that induced by human activities.

The role of the ocean in causing these abrupt changes is important from a future perspective considering the reported data on freshening of the North Sea and the Arctic Ocean, and consequent reduction in the production of deep Atlantic water (Hansen et al., 2001). Further observations indicate that this flow may have dropped up to 30% in the last 60 years (Bryden et al., 2005), thereby suggesting that the thermohaline system is entering a destabilization phase that could have devastating consequences for the short and mid-terms (i.e., one generation). Climate models suggest that in addition to the future atmospheric volume of CO2, the velocity with which it is amassed may have serious consequences for the Atlantic thermohaline system (Stocker and Schmittner, 1997). Hence we have an obligation to understand the processes responsible for the initiation, transmission and amplification of these rapid oscillations, and to assess their potential impact on the fate of our fragile planet. The results of Martrat et al. (2004) highlight the need for studying abrupt climate changes of earlier interglacial periods in order to predict future scenarios, thus the Holocene must be considered a high priority for paleoclimatic analysis.

Regarding the Holocene, a growing number of studies, some of which have been carried out by GRACCIE partners, have demonstrated that some type of abrupt climate variability did indeed occur, although of relatively minor intensity than in the glacial (Cacho et al., 2001; Marchal et al., 2002; Pla and Catalan 2005; Frigola et al., 2007). However, there is ever increasing evidence that some of these oscillations took place during historic periods (the segment of the Little Ice Age from the 18^th to the 19^th centuries, or the Medieval Warm Period), for which important documents exist on the impact of said oscillations on human cultural and economic development (deMenocal et al., 2000; Bond et al., 1999). Consequently, we can ascertain the vulnerability of our systems to this type of variability even when it is of a low intensity.