HB-1-4

Paleoceanographic records indicate substantial Holocene climate variations in the North Atlantic realm at timescales ranging from centuries to millennia. It has been suggested that the Holocene climate fluctuations at these timescales, specifically those reconstructed for the North Atlantic region, are linked to variations in the strength of the Atlantic meridional overturning circulation (AMOC) and the associated heat transport. Such type of variability does not only affect the climate system on a regional or even larger scale, but may also have a profound influence on the marine ecosystem. Understanding the mechanisms underlying millennial-scale oscillations in North Atlantic climate is therefore not only important from a paleoclimatic perspective, but it is also essential to understand the origin and dynamics of these climate variations in order to inform about potential interference with anthropogenic climate change. A potential role in this Holocene low-frequency climate variability has been suggested for the Labrador Sea. In long-term integrations with a coupled climate model of intermediate complexity (ECBilt-CLIO), these authors found centennial-to-millennial-scale climate variability in the North Atlantic realm associated with variations in AMOC strength. These climatic oscillations were attributed to an underlying bistability in Labrador Sea convection and state transitions induced by noise. On a multicentennial timescale these stochastic modetransitions can be phase-locked to small periodic external forcings. Reconstructions of Labrador Sea density stratification and winter convection during the Holocene revealed millennial-scale oscillations after ca. 8 ka BP. This finding corroborates the model-based notion of a bistable regime operating in the Labrador Sea. So far, low-frequency oscillations of this type have not been found in more comprehensive climate models. However, Yoshimori et al. reported two possible AMOC states (a strong and a weak mode with ca. 17 and 10 Sv, respectively) in multicentennial pre-industrial simulations using the low-resolution (T31 atmosphere, 3° ocean) version of the comprehensive global climate model CCSM3 (Community Climate System Model version 3). The stability and bifurcation properties have not been studied systematically. Supporting the finding of Yoshimori et al. we have recently found an abrupt and persistent mode transition from the strong to the weak AMOC state in a transient simulation of Holocene climate with the same climate model (not published). The existence of two AMOC modes holds the potential for noise-induced transitions and hence low-frequency variability similar as in ECBilt-CLIO. The overarching goal of this project is to improve our understanding of Holocene centennial-tomillennial-scale variations in North Atlantic climate. More specifically, we aim to identify boundary conditions that favor low-frequency AMOC variability as well as early-warning signals for North Atlantic climate transitions.

The PhD student will analyze tempo-spatial patterns associated with the mode transitions in ECBilt-CLIO and CCSM3. The student will search for early-warning signals for state transitions through statistical time series analysis; for instance, the phenomenon of ‘critical slowing down’ would result in changing statistical properties (e.g. increased autocorrelation and variance) in time series associated with AMOC variability. To identify the influence of the background climate conditions on deep water masses and the existence, behavior and timescale of North Atlantic mode transitions, sensitivity experiments with the low-resolution CCSM3 will be carried out using greenhouse gas concentrations, orbital forcing and surface freshwater fluxes as control parameters. Again, statistical analysis of time series will provide hints concerning the proximity to critical thresholds through ‘critical slowing down’. In-depth comparison of model output with Holocene proxy records will be carried out in order to assess the potential of the simulated low-frequency variability in the real climate system. To this end, the PhD student will extend proxy data bases by compilation of published data and new records produced in other projects of the Research Training Group (esp. HB-3, HB-8, HB-9, CA-2, CA-8, CA-9, CA-14)