CT 1

Core Theme 1:  Quantifying and modeling THC variability using palaeoclimate observations and simulations

Long-term observational time series and multi century climate model experiments suggest that broad-scale changes in surface temperature, precipitation, and storm/drought patterns are associated with low-frequency variations in the THC, such as the Atlantic Multidecadal Oscillation (AMO). These interdecadal to centennial changes exhibit similar time characteristics as the presently ongoing global warming. Therefore the THC and its response to anthropogenic global warming in the future represent a major uncertainty in predicting future climate change and its consequences. On the other hand, the persistent natural thermohaline circulation cycles imply that there is predictability of the THC. CT1 aims to generate the underpinning understanding of interdecadal to centennial THC variability and the underlying processes by combining high-resolution palaeo observations from marine cores and millennium-scale model simulations. Refining our understanding of the low-frequency variations, their scale and temporal characteristics, and subsequent impacts must rely heavily on observational records covering the centennial to millennium-scale, and therefore on geological archives. Substantial progress has been made in the assessment of past climate change using geological archives capable of high temporal resolution.  New proxies and state-of-the-art techniques will now allow the history of THC to be assessed at similarly high resolution.

For the first time, computational resources are available to carry out millennium-scale simulations with comprehensive atmosphere-ocean general circulation models (AOGCMs), such as those used in the IPCC AR4. Progress in model performance and quality allows analysing climate variability in unprecedented detail and to better understand the underlying causes and feedbacks.

The CT 1 is led by: Johann Jungclaus (MPG-M); co-lead: H. Kleiven (UiB)
Participants: UiB, UPMC-LOCEAN, CNRS, MET-O, MPG-M, NERSC

WP 1.1   Analyses of millennium-scale simulations with coupled atmosphere-ocean general circulation models

As an analysis of direct observations of oceanic transports and water mass properties is limited by the availability and quality of the data sets, the workpackage 1.1 will use model simulations to investigate inter-decadal to centennial variations of the THC and to evaluate the robustness of processes in the different models. To date, several activities are underway to apply state-of-the-art climate models to simulate the climate of the last millennium. Four of the THOR partners are directly involved in planning, performing and managing these experiments.

These simulations forced by reconstructions of external (solar, volcanic) and anthropogenic (land-use change, CO2, trace gases etc.) and their respective unforced control integration form a unique database for the investigation of THC variability on inter-decadal to centennial time scales. The simulations will be compared with the observations from WP 1.2, CT 3, and other sources. Statistical analyses will be performed to derive the main patterns of oceanic and atmospheric variability associated to the THC. The multi-model approach allows estimating how robust mechanisms identified in one model are and which processes are of key importance in all simulations. The existing long-term simulations will be complemented by sensitivity studies focusing on selected processes.

Task 1.1.1 Assessment of the Millennium runs and role of the external forcing

The THC in forced runs will be analysed in comparison with long control integrations to investigate the THC response to external forcing changes (e.g., solar and volcanic) as well as the THC response to GHG and aerosol forcing in the industrial era. The partners will also investigate how the models reproduce the properties determined from palaeo-observations in Task 1.2 in direct comparison and in a statistical sense: low-frequency variability in the properties and intensity of the exchanges across the GSR, the ventilation rates, the thermocline variability in the tropical Atlantic, the cross-equatorial fluxes, and large-scale North Atlantic SST patterns as well as the role of Southern Ocean upwelling. The signatures of the THC on decadal-to centennial scales in the Millennium experiments that can be compared to those derived from instrumental observations will be established. This includes links between major climate indices, lead/lag relationships, and space-time patterns of ocean – atmosphere variability.

In direct comparison with the observational and monitoring program of CT3 and other sources, the ability of the models to simulate key aspects of the THC and related (small-scale) processes will be assessed: mass, heat and salt transports across the GSR, overflow properties, transports variability, ventilation in Labrador and Nordic Seas. Additionally, recent observations of changes in oceanic heat and fresh water distributions will be used as tests of the model THC processes making use of a larger suite of simulations of the period since 1850 that are available from the Millennium and IPCC AR4 models.

Finally, the partners will design and apply a common framework for model testing to establish and quantify, in a coordinated way, to what extent the Millennium runs are consistent with available observations.

Task 1.1.2 THC variability on decadal to centennial timescale

In order to make reliable forecasts of the THC it is essential to advance understanding of the detailed processes that govern THC variability. In this Task, the dynamics of THC adjustments and decadal water mass changes will be analysed, including ventilation and water mass transformation processes and their variations. CT1 will investigate how overflow across the Greenland Scotland Ridge and deep-water formation in the subpolar gyre (Labrador and Irminger Seas) are transferred into THC variations. The role of the hydrological processes will be established and the core theme will determine whether and on what time scale salinity has an active or passive role in the THC variability. Particular emphasis will be given to the Arctic fresh water budget as contribution to the International Polar Year (IPY). Furthermore, the influence on the THC of atmospheric forcing in different ocean basins (in particular the Southern ocean) and of variations in inter basin exchanges of fresh water will be investigated.

Model results will be compared to recent observations and to the sensibility studies on THC variations to idealized fresh water perturbations carried out in Task 2.2. The analysis of the existing long-term integrations will be supported by selected coordinated sensitivity experiments that are designed to answer the question: what drives overflow variations on interdecadal time scales? The project partners will run dedicated experiments with idealized forcings/set-up and sensitivity experiments with prescribed modification of overflow properties.

Task 1.1.3 Ocean – atmosphere feedback and climatic impact of the THC changes

Studies with different AOGCMs have given different answers to the question if the observed and simulated low-frequency THC variations are the expression of a coupled ocean-atmosphere mode or of ocean dynamics that are either internally driven or forced by atmospheric noise that may have a strong impact on the oceanic internal mixing.. In this Task, CT1 will detect and attribute the atmospheric response to the THC changes in the Millennium and associated control simulations, assessing their climate impact in the multi model framework and establishing the implications for predictability (CT 4). CT1 will also establish whether there are coupled atmospheric/oceanic modes and an active atmospheric response to the THC changes. CT1 will separate between cause and effect, and establish lead/lag relationships between the oceanic fields and the atmosphere. If an active coupling is found, partners will perform sensitivity studies with the atmospheric component of the climate models to understand the nature of the atmospheric response and the dynamics involved. Guided by the results of these analyses, idealized forcing sets will be designed. The latter will be done by performing partial coupled integrations, i.e., integrations with fully coupled AOGCMs but where selected oceanic and/or atmospheric fields are prescribed in the region of interest (the North Atlantic Ocean).

Modules used in WP 1.1

WP 1.1 Lead: Johann Jungclaus (MPG-M)
Participants: UPMC-LOCEAN, MET-O, MPG-M, NERSC

WP 1.2: THC and related climate variables during the last Millennium from Palaeo observations

WP 1.2 aims to provide the extended high-resolution palaeoclimate reconstructions of ocean state variables necessary to resolve key issues concerning the mechanisms for THC variability on multi-decadal to centennial time scales. The scale and frequency of past variations in the transport and properties of the upper and lower branches of the THC will be reconstructed at key locations in the Nordic Seas and North Atlantic - tracking deep overturning circulation, inflow variability and the tropical response.  Once defined, the past THC variability will be compared to the climate variations (e.g. SST) co-registered from palaeoclimate proxy data at these locations elucidating the relationship between THC and climate in the time domain. Through comparison with the model simulations in WP1.1 these reconstructions will provide the empirical constraints required to gain a quantified, process-based understanding of THC variability and its impact on the climate of the last millennium. In addition, improved error estimates for each proxy will be provided by focused study of the period of overlap between proxy and observational records.  Palaeoclimate observations can also be used to test the long-term persistence of large scale features observed from the instrumental record, e.g. the Atlantic Multidecadal Oscillation (AMO) and will (within the limitations of the palaeoclimate methods) be used to diagnose how consistent properties identified in CT3 are within a millennial scale temporal framework.

Study Areas of WP 1.2

Task 1.2.1.  Characterize changes in the deep and intermediate return flow of THC - Determine how much it changed, which components, and why

CT1 will assess the variations in total NADW export over the last millennium using sedimentary Pa/Th within the deep Western Boundary Current from cores west of the mid Atlantic ridge (area VI). These records provide the baseline for assessing how accurately the scale and timing of THC variability is simulated in WP 1.1.

Proxy archives from North Atlantic drift sites record changes in the past circulation and properties of North Atlantic Deep Water (NADW) along it’s flow paths. Using newly recovered cores CT1 will reconstruct the properties and intensity of both the integrated overflows (area IV, Erik drift,) and the Eastern Nordic seas overflows (area V, Gardar Drift). Thus the CT1 will differentiate changes in overall THC intensity  (integrated overflows) from potential climate related shifts in the volume flux between overflows.  The integrated Nordic Seas overflows will be assessed using proxies for bottom water vigor (sedimentary and magnetic grain sizes). In addition, sediment Pa/Th will assess changes in integrated transports within both the western (Greenland-Iceland) and eastern (Iceland-Scotland) overflow pathways over the past millennium. Thus, it is possible to provide observational constraints for identifying how variability in specific branches of the overflows translate into THC variability (modelled in WP 1.1).

The physical and chemical deep water property variability associated with changes in overflow intensity (area IV and V) will be determined using benthic foraminiferal isotopes and Mg/Ca in order elucidate the specific changes in water mass formation processes related to THC variability.  These observational constraints will ground truth the model analyses in Task 1.1.2

Task 1.2.2. Characterize the upper limb of THC—Variations in the inflows to the Nordic Seas

Determine the variability in the properties and intensity of inflowing Atlantic Water to the Nordic Seas on decadal-centennial timescales over the last millennium. Preliminary results (Nyland and co-workers, in prep.) show a clear connection between the AMO index and reconstructed variations in the position of the Arctic front from foraminiferal census data (percent left coiling N. Pachyderma) back to 1850. Here CT1 will extend these reference records of AMO changes. In particular a full characterization of AMO related surface water changes will be developed for comparison with the high-resolution overflow records—to determine potential links between the extended AMO series and the reconstructed inflow-outflow variability (Task 1.2.1) from the Nordic Seas. Sediment grainsize work on cores from areas I and II will be used to constrain the vigour of Atlantic water inflow over the past millennium.  In addition, elements of the ongoing VAMOC project (area III) will be extended to provide quantitative reconstruction of upper ocean transport into the Nordic Seas using a palaeo-geostrophic approach--reconstructing the lateral and vertical extent of Atlantic Water along main inflow areas, and estimate slope changes of deep thermocline waters. 

Task 1.2.3. Characterize climate and thermocline evolution over the last millennium

CT1 will generate surface climate proxy records in its sites capturing deep transport properties allowing a direct assessment of the climate-THC link—to determine if upper ocean climate changes when the THC does.    These products will be added to the high resolution SST reconstructions produced through the PACLIVA project to build a spatially resolved data base for assessing the scale and pattern of climate variability observed in millennium model simulations (WP 1.1).

In idealized model experiments there is a tendency that changes in the strength of the THC are manifested by changes in the depth and the heat content of the tropical Atlantic thermocline. Palaeoclimate data can thus be a test bed for identifying the consistency and robustness of such model behavior.  CT1 will reconstruct the variability of the vertical temperature and density gradients in the main thermocline from the western tropical Atlantic (area VII) and estimate seasonal thermocline temperature changes during the last millennium using cores collected through the RETRO project (an ESF Eurocores program).  Thermocline variability is an important diagnostic for past THC variability: models predict large changes in response to variable THC and proxy records of thermocline temperatures tend to be more representative as local influences are significantly less important than they are in surface water climate records.

Lead: Helga Kleiven (UiB)
Participants: UiB, CNRS