3rd Annual Meeting

Bergen, 9-11 November 2011

CT Reports

CT 1 presentation of Johann Jungclaus: download

CT 2 presentation of Steffen M. Olsen: download

CT 3 presentation of Svein Oesterhus: download

CT 4 presentation of Wilco Hazeleger: download

CT 5 presentation of Johannes Karstensen: download

Science talks on observed North Atlantic/Arctic Ocean climate variability and its predictability

Hazeleger, Wilco et al.: Multimodel Decadal Predictability of the Subpolar Gyre (download PPT and PDF).
[Royal Netherlands Meteorological Institute, The Netherlands]
Multimodel decadal predictions made within the THOR project are presented. The THOR project focusses on the AMOC. The ocean analyses show that the AMOC may have increased slightly up to the 1990s after which a reduction took place associated with a reduction of Labrador Sea Water formation. However, the AMOC is not directly observed, hence the focus will shift to observed ocean phenomana. These include the Atlantic Multidecadal Variability, the interhemispheric dipole, Labrador Sea Water formation and the Great Salinity Anomaly.
It is shown that the interhemispheric dipole and the Atlantic Multidecadal Variability is predictable up to 9 years ahead. The upper ocean heat content is even better predictable. It appears to be hard to predict the Labrador Sea Water formation and the propagation of salinity anomalies in the subpolar gyre.
Finally, the predictability is partly originating from the external forcing by changing greenhouse gas concentrations and aerosols.

Langehaug, Helene et al.:
Mechanisms for decadal scale variability in the North Atlantic Ocean circulation in the Bergen Climate Model (download)
Full list of authors: H. R. Langehaug, I. Medhaug, T. Eldevik, O. H. Otterå, T. Furevik, and M. Bentsen
[University of Bergen, Norway]
From a 600-year pre-industrial control simulation with the Bergen Climate Model, we have identified potential mechanisms for decadal variability of the Atlantic Meridional Overturning Circulation (AMOC) and Subpolar Gyre strength. The three dominant patterns of North Atlantic atmospheric variability – the North Atlantic Oscillation (NAO), the East Atlantic Pattern and the Scandinavian Pattern (SP), are reflected in the ocean circulation. The Labrador Sea convection, giving the upper North Atlantic Deep Water, is driven by decadal scale variability in the heat flux related to the NAO. The lower North Atlantic Deep Water responds to fluctuations in the SP, where a negative phase is associated with northerly winds and an increased southward flow across the Greenland-Scotland Ridge. The deep water constitutes the lower limb of AMOC, where the variability is directly linked to convection in the Labrador Sea. The variability in the deep water, together with EAP, can partly explain the variations in the strength of the gyre circulation.

McCarthy, G. et al.: Interannual variability in the Atlantic meridional overturning circulation at 26.5°N (download)
[National Oceanography Centre, Southampton, United Kingdom]
Full list of authors: Gerard McCarthy, Stuart A. Cunningham, Eleanor Frajka-Williams, Harry Bryden
From March 2004, as part of the Rapid programme, we have been monitoring the Atlantic meridional overturning circulation at 26.5°N with a continent-to-continent array of moored instruments. Here we examine the first six years of observations to quantify inter-annual variability and trends in the components of the overturning circulation (Ekman transport from satellite scatterometer measurements, Florida Straits transport from electromagnetic cable measurements, and upper mid-ocean geostrophic recirculation from the Rapid array).
From 2004 to 2008 the MOC variability has a clear seasonal cycle, arising largely from the upper-mid ocean, which in turn is driven by seasonality in wind stress curl variations at the eastern boundary. In 2009 we see a very different characteristic to upper mid-ocean variability with a persistently high southward transport of about -20 Sv during 2009. Combined with a declining Gulf Stream transport the MOC declined to about 12 Sv near the end of 2009. The winter of 2009/2010 (DJF) was characterized by extreme negative NAO state (probably the most negative state in the instrumental period). The Atlantic wide disruption to the usual wind patterns has forced an extreme southward Ekman transport of more than -10 Sv. This projects directly on to the MOC and during December 2009 the MOC actually goes to zero, and is persistently low into the first quarter of 2010.
Perhaps more climatically significant is the reduction in annual average transport in 2009 of lower North Atlantic Deep Water from a long term mean of 7.4 Sv to 4.7 Sv. This reduced southward deep flow is compensated by an increase in southward thermocline circulation in the upper 1000m, and overall results in a reduction in the MOC to 14.3 Sv in 2009 from the 2004-2010 mean of 18.0 Sv. Meridional heat flux is dominated by the MOC and so this reduction in the mean MOC in 2009 is accompanied by a reduction in the meridional heat flux.
By the time of this talk we will have observations extending through the winter of 2010/11 and in this talk I hope to focus on the forcing and consequences of the recent changes in the interannual variability of the MOC at 26.5°N.

Dunstone, Nick: Impact of observations on MOC predictions (download)
[MET O UK, United Kingdom]
We present the experimental design and first results from the multi-model idealised predictability experiments currently underway in WP4.2. These experiments attempt to evaluate the suitability of the current observing system for initialising decadal predictions of the MOC, to assess the skill of decadal MOC predictions in the idealised model world (where we know the answers!) and finally to identify any shortcomings in the assimilation techniques used. We also briefly summarise the results of a study into the use of optimal perturbations to create forecast ensembles.

***Skagseth, Øystein: Observed North Atlantic/Arctic Ocean climate variability of predictive potential
[Bjerknes Centre for Climate Research, Norway]

Larsen, Karin et al.: Atlantic water hydrography in the Faroe area – sources and variability (download)
[Faroe Marine Research Institute, Faroe Islands]
Full list of authors: Karin Margretha H. Larsen*, Hjálmar Hátún, Bogi Hansen and Regin Kristiansen
Corresponding author: Karin Larsen
The inflow of Atlantic water across the Greenland-Scotland Ridge and into the Nordic Seas controls both physical and biological conditions in the northeastern Atlantic through its transport of heat, salt and other properties. The two main branches of this flow pass through the Iceland-Faroe Gap and the Faroe-Shetland Channel, respectively. Regular monitoring along four standard sections crossing these flows provides time series of the Atlantic water temperature and salinity variability since the late 1980s. The presented analysis of these series shows a persistent increasing trend in both temperatures and salinities, modulated by smaller subdecadal oscillations. Using supplementary data sources, we support the previously established link between the large-scale circulation in the North Atlantic and the Atlantic inflow properties. The salinity is also impacted by large changes in the Bay of Biscay source waters, while up-stream air-sea heat fluxes modulate the temperatures. We furthermore discuss the relation between changes in transports and associated residence times, and the modifying strength of air-sea interaction and mixing.

Lohmann, Katja: Atlantic Multidecadal Variability and Iceland-Scotland overflow strength in the millennium simulations (download)
[Max Planck-Meteorology, Hamburg, Germany]
A new sediment core proxy representing the strength of the Nordic Seas
overflow between Iceland and Scotland shows low-frequency variability in phase with the Atlantic Multidecadal Oscillation (AMO). A possible explanation for the in-phase relationship would be: strong overflow -> strong meridional overturning circulation (MOC) -> warm North Atlantic. In this study we use simulations of the last millennium, driven by external forcing reconstructions, with three different coupled climate models (MPI-M ESM, IPSLCM, BCM) to investigate possible mechanisms underlying the phase relation between overflow strength and AMO. Comparing the low-frequency variability of the simulated overflow strength with that of the proxy reveals some similarity during the last 200 years, indicating that during this period the external forcing played an important role in driving the overflow variability. Considering the simulated phase relation between overflow strength and AMO shows an out-of-phase relationship for MPI-M ESM (all eight ensemble members), and an in-phase relationship for IPSLCM and BCM (although less clear in the latter case).
We performed lag regression patterns of the low-frequency overflow and AMO indices from MPI-M ESM for various oceanic and atmospheric variables. An AMO state representing a warm North Atlantic goes along with a strengthening of the low pressure in the Icelandic region and a related strengthening of the subpolar westerlies. Maximum values are found over the western subpolar region. The strengthened westerlies (in combination with warm sea surface temperatures) lead to an increase in the evaporation, which in turn leads to an increase in the sea surface salinity. The subpolar density is dominated by salinity and the increase in the subpolar density leads to enhanced subpolar convection. The latter mixes up warmer water to the surface, leading to an increase in the sea surface temperature in the western subpolar region, and also increases (through enhanced deep water formation) the strength of the MOC. The latter leads to a warming in the eastern subpolar region. Concerning the surface heat flux reveals a damping over the entire northern North Atlantic, indicating that the sea surface temperature anomalies are created solely by the ocean dynamics.
In contrast to the AMO, a strong overflow goes along with a weakening of the low pressure in the Icelandic region and the subpolar westerlies. Minimum values are found in the eastern subpolar region. The wind stress pattern will influence the sea surface height (barotropic part of pressure) and/or the density structure (baroclinic part of pressure) south of the ridge, affecting the pressure gradient across the ridge at the depth of the overflow. Details still need to be investigated. The density structure north of the ridge shows a deepening of the isopycnals which lags the overflow strength, explainable by the removal of dense water from the Nordic Seas due to a strong overflow.
Preliminary analysis suggests basically the same mechanism for the AMO also in IPSLCM and BCM (triggered by a strengthening of the low pressure in the Icelandic region). The difference in the phase relation between overflow strength and AMO seems mainly to originate from the fact that in IPSLCM and BCM a strong overflow goes along with a strengthening of the low pressure in the Icelandic region, in contrast to MPI-M ESM. This difference might also explain the difference in the phase relationship of the two overflow branches (between Iceland and Scotland as well as through Denmark Strait), being in-phase for MPI-M ESM but out-of-phase in IPSLCM and BCM.
In MPI-M ESM, the relation between low-frequency overflow strength and AMO variability seems to be rather indirect with both being caused by a similar atmospheric circulation pattern. Also in IPSLCM and BCM, a direct link between overflow strength and AMO (strong overflow ? strong MOC ? warm North Atlantic) seems unlikely, as the AMO tends to lead the overflow as indicated by the cross-correlation-function. Furthermore, a study of the corresponding control integrations revealed no significant influence of the strength of the overflow between Iceland and Scotland on the MOC. For IPSLCM and BCM, however, it might be that the MOC has an effect on the density structure in the Nordic Seas which influences the strength of the overflow. Further investigation is needed to clarify this issue.

Hawkins, Ed et al.: Arctic Predictability in HadCM3 and the APPOSITE project
(download) Full list of authors: Ed Hawkins, Sarah Keeley, Dan Hodson, Nick Dunstone, Rowan Sutton
We explore the 'perfect model' predictability of the HadCM3 GCM in a large set of ensemble predictability experiments, and find that sea-ice extent and volume are potentially predictable for up to 5 years. The mechanisms of predictability will be discussed. Additionally, we outline the design of the APPOSITE project, which aims to (a) quantify predictability of Arctic climate in a wide range of GCMs, (b) identify the key physical mechanisms providing predictability, and (c) inform required developments for operational seasonal-to-interannual Arctic predictions.

Gascard, Jean Claude: Is the Ocean involved in and/or responsible for the recently observed climate shift in the Arctic? (download)
[ LOCEAN, UPMC, France]
During the past 10 years we observed drastic changes in the atmosphere (stratosphere and troposphere) involving the northern polar vortex, the Arctic Oscillation, the dipole anomaly, the storm tracks (polar lows) and cold air outbreaks (CAO) strongly correlated with sudden stratospheric warming (SSW) . We also observed drastic changes in Arctic sea-ice extent, drift, thickness and age. At the end of the summer (September) the extent of Arctic sea-ice is now approaching 4 millions km2 and the average thickness is twice less than it was 20 years ago (less than two compare with more than three). Combining extent and thickness reduction by half each would mean the sea-ice volume (or total mass) has decreased by 75% at the end of each summer. Sea-ice is drifting twice faster than before as illustrated by the 500 days transpolar drift of Tara in 2006-2008 compared with 3 years drift for the Fram more than hundred years ago. The Arctic multiyear sea-ice is vanishing very rapidly and is now replaced by first year sea-ice very much like what is happening every year around Antarctica. By comparison the Arctic Ocean looks much more robust and stable at first glance and its variability does not seem to have changed drastically. The strong vertical stratification due to the presence of the shallow halocline and the deep thermocline has not been eroded or significantly affected by climate change occurring in the Arctic atmosphere and sea-ice. But how long that would last ? It is quite clear that the upper ocean is responding very strongly to the sea-ice ocean albedo positive feedback by melting more sea-ice during summer and by delaying the freeze up until the heat stored in the upper ocean has been released in the atmosphere. So indeed the ocean is involved in the transformation of the lower troposphere triggering interaction with the stratosphere in addition to be responsible for melting more sea-ice and for the global warming polar amplification due to ice-ocean albedo positive feedback.

Spall, Mike: Precipitation and the shutdown of deep convection in marginal seas  (download)
[Woods Hole Oceanographic Institution, USA]
The influences of precipitation on water mass transformation and the strength of the meridional overturning circulation in marginal seas are studied using theoretical and idealized numerical models. Nondimensional equations are developed for the temperature and salinity anomalies of deep convective water masses, making explicit their dependence on both geometric parameters such as basin area, sill depth, and latitude, as well as on the strength of atmospheric forcing. The theory also predicts the magnitude of precipitation required to shut down deep convection and switch the circulation into the haline mode. High resolution numerical model calculations compare well with the theory. However, the model also shows that for precipitation levels that exceed this critical threshold, the circulation remains in a thermally direct mode even in the absence of deep convection.

Myers, Paul:
Freshwater Pathways from the Arctic to the sub-polar North Atlantic from modelling and data
[Institute, Country]
Significant fluxes of low salinity water pass from the Pacific Ocean into the Arctic Ocean, and after mixing with other freshwater sources in the Arctic, are transported into the North Atlantic. These pathways and the associated fluxes can have a significant impact on physical, chemical and biological ocean properties and potentially can impact aspects of the large scale circulation and climate. Specific details on how Pacific Water travels from Bering Strait to the Canadian Arctic Archipelago are uncertain. As well, the fate of the freshwater (of Pacific and other origin) that passes through the Canadian Arctic Archipelago is uncertain. In particular how does this freshwater leave the boundary current system and where is it taken up into the Atlantic Ocean. We examine these questions used a several eddy-permitting regional configurations of the NEMO coupled ocean/sea-ice numerical model. As well as examining hydrographic properties and fluxes, we use the lagrangian float tool Ariane to examine the freshwater pathways and their variability.

Glessmer, Mirjam et al.: Nordic Seas freshwater anomalies tracked back to Atlantic inflow? (download)
[UiB, Norway]
Full list of authors: M. S. Glessmer, T. Eldevik, J. E. Ø. Nilsen
Increased freshwater input into the Nordic Seas could ultimately lead to changes in the Atlantic Meridional Overturning Circulation. Freshwater anomalies in the Nordic Seas are often implicitly assumed as connected to the Arctic inflow, which is modulated by melting of Arctic sea ice as well as Greenland ice cover and hence strongly influenced by global warming. We investigate Nordic Seas’ freshwater anomalies in a hydrographic dataset collected within the NISE project and show that those anomalies are mainly located in the Norwegian Sea. This location suggests a connection with the Atlantic rather than the Arctic inflow. The 'routing' of freshwater deduced from the observations is compared with that of a multi-century Bergen Climate Model simulation. The implications for the role of Arctic freshwater anomalies in future climate change and predictions are briefly discussed.

Menary, Matthew et al.: Anthropogenic controls on the AMOC over the 20th Century? (download)
Speaker: Matt Palmer
[Met O, Country]
Full list of authors: Matthew Menary, Chris Roberts, Paul Halloran, Laura Jackson, Matt Palmer, and Richard Wood
Simulations using the HadGEM2 Earth System model show a 20% increase in the AMOC over the 20th Century, and a subsequent weakening into the 21st century. This is in stark contrast to previous modelling studies, which suggest that the AMOC has been stable or weakening over the last century. In our simulations the AMOC strengthening occurs primarily in response to atmospheric circulation changes, which arise due to the particular combination of anthropogenic and volcanic aerosols, solar activity, and land-use changes. A past AMOC strengthening is supported by recent multimodel reanalyses, by direct comparison with atmospheric reanalyses; by an ocean model forced with these same atmospheric reanalyses; and within the error bars of re-evaluated long term observational estimates of the AMOC. The rapid weakening of the AMOC in the early decades of the 21st Century may exacerbate the effects on European climate previously predicted by climate models.

Mecking, Jenny et al.:
Origins of Atlantic Decadal Variability (download)
[Leibniz Institute of Marine Sciences (IFM-GEOMAR), Kiel]
Full list of authors: Jennifer Mecking(1,2), Noel Keenlyside (1), Richard Greatbatch (2), Jin Ba (1)
1Geophysical Institute, University of Bergen
2Leibniz Institute of Marine Sciences (IFM-GEOMAR)
**Abstract missing**

Frankignoul, Claude and Gastineau, Guillaume:
The response of the atmospheric circulation to the variability of the Atlantic meridional overturning circulation (download)
[Université Pierre et Marie Curie, France]
In control simulations with six global climate models, a significant influence of the natural variability of the Atlantic meridional overturning circulation (AMOC) was found during the northern hemisphere cold season when the AMOC leads the atmosphere by a few years. Although the AMO-like sea surface temperature (SST) patterns vary slightly between models, in all cases an intensification of the AMOC is followed by a negative phase of the North Atlantic Oscillation (NAO). The atmospheric response seems to be primarily due to changes in the SST and the heat loss along the North Atlantic Current and the subpolar gyre, which shift southward the maximum lower-tropospheric baroclinicity and alter the North Atlantic storm track. However, in CCSM3, the AMOC-driven SST changes along the Gulf Stream and the North Atlantic Current shift the maximum heat release and the lower-tropospheric baroclinicity northward, so that a positive NAO phase follows an AMOC intensification.

Keenlyside, Noel:
Atmospheric response to North Atlantic Decadal variability
(download) [University of Bergen, Norway]
Full list of authors: Noel Keenlyside, Nour-Eddine Omrani, Sergey Gulev, Mojib Latif
Atlantic multi-decadal variability has major socio-economic impacts, yet it remains a poorly understood phenomenon. This presentation addresses ocean-atmosphere interaction in the extra-tropical Atlantic on multi-decadal time scales. First, we show using observations that the ocean drives SST and turbulent surface heat fluxes over the Gulf Stream extension on time scales longer than 10 years. Second, using atmospheric model experiments we show that the response to multi-decadal variations in SST involves the stratosphere, and can only be captured when stratosphere-troposphere coupling is resolved.

Wouters, Bert: North Atlantic interdecadal Meridional Overturning Circulation variability in the EC-EARTH coupled model (download)
[Royal Netherlands Meteorological Institute, The Netherlands]
Interdecal variability in the North-Altantic meridional overturning circulation (MOC) has been studied in a pre-industrial control run newly developed EC-EARTH model. An oscillation of approximately 2 Sv is observed, which manifests itself as a monopole causing the overturning to simultaneously strengthen (/weaken) and deepen (/shallow) as a whole with a periodicity of 50--60 yr. Eight years before the MOC reaches its maximum, density in the Labrador-Irminger Sea region reaches a maximum, triggering deep water formation. This density change is caused by a counterclockwise advection of temperature and salinity anomalies at lower latitudes, which we relate to the North-South excursions of subpolar-subtropical gyre boundary and variations in strength and position of the subpolar gyre and the North Atlantic Current. A weak link to the North Atlantic Oscillation is found, although this atmospheric forcing does not appear to be a dominant factor in setting the MOC oscillation in our model.

Latif, Mojib and Park, Wonsun: Internal and External Variability: Kiel Climate Model versus Data (download)
[Leibniz Institute of Marine Sciences (IFM-GEOMAR),Kiel]
We integrated the Kiel Climate Model (KCM) in a number of experiments. The multi-millennial control run depicts a wide range of variability, from seasonal to centennial. The talk will focus on the Atlantic Meridional Overturning Circulation (AMOC). The statistical investigation of KCM’s internal AMOC variability obtained from a multi-millennial control run yields three distinct modes: a multi-decadal mode with a period of about 60 years, a quasi-centennial mode with a period of about 100 years and a multi-centennial mode with a period of about 300-400 years. Most variance is explained by the multi-centennial mode, and the least by the quasi-centennial mode. The dynamics of the three modes will be discussed first and comparisons with data shown.
In the second part, KCM’s response to idealized (periodic) external (solar) forcing is studied. The AMOC response is rather complex and nonlinear. It involves strong changes in the frequency structure of the variability. While the control run depicts multi-timescale behaviour, the AMOC variability in the experiment with 100 year forcing period is channelled into a relatively narrow band centred near the forcing period. It is the quasi-centennial AMOC mode which is excited, although it is heavily damped in the control run. Furthermore, the quasi-centennial mode retains its period which does not correspond exactly to the forcing period. Surprisingly, the quasi-centennial mode is also most strongly excited when the forcing period is set to 60 years, the period of the multi-decadal mode which is rather prominent in the control run. It is largely the spatial structure of the forcing rather than its period that determines which of the three internal AMOC modes is excited. Interestingly, the instrumental data yield some evidence for such a quasi-centennial mode.

Marini, Camille and Frankignoul, Claude: The Atlantic Multidecadal Oscillation in a simulation of the last millennium with the IPSLCM4 climate model (download)
[Université Pierre et Marie Curie, France]
The links between the Atlantic Multidecadal Oscillation (AMO) and the Atlantic Meridional Overturning Circulation (AMOC) are investigated in a simulation of the last millennium climate with the IPSLCM4 model and compared with a control simulation of the same model. This forced simulation represents
the volcanic activity, but the solar forcing is underestimated. A dynamical filter based on linear inverse modelling (LIM) is used to isolate the radiative effect of volcanic eruptions and to remove it from the AMO. We also used the LIM-based filter to remove the influence of El Nino Southern Oscillation (ENSO) from the AMO. Whereas this leads to a slight increase in AMO-AMOC correlations in the control simulation, these correlations are reduced in the forced simulation.

Drijfhout, Sybren:
Transitory cooling after a thermohaline circulation collapse (download)
[Royal Netherlands Meteorological Institute, De Bilt, The Netherlands]
An ensemble of coupled climate model runs is subjected to a forced collapse of the thermohaline circulation while greenhouse gases follow the SRES A1b scenario.
It is investigated whether a THC collapse outweighs increasing CO2 forcing, and if so, for how long, and what processes are responsible for reversing the trends in temperature change and climate feedbacks. Also, the effect of the THC collapse proper is analyzed from the differential signal between two SRES A1b ensembles, with and without hosing.
It is shown that, even when CO2 increases, the fast response to a THC collapse is dominated by a reduced greenhouse effect. The atmosphere gets drier. This leads to hemisphere-wide cooling, contradicting the bipolar seesaw paradigm.
However, the slow response to the THC collapse reduces the cooling, instead of amplifying it, also for the differential signal with constant CO2.
A downward radiation anomaly arises at the top of the atmosphere, heating the earth's surface.
The radiation anomaly is caused by ocean heat uptake: atmospheric cooling is forced by reduced ocean heat release through increase of sea-ice cover.
In the differential signal this radiative damping removes the global mean temperature anomaly in 300 years.
When CO2 concentrations follow the SRES A1b scenario the radiative damping is reinforced; recovery takes 50 years.
The recovery-pattern is very inhomogeneous. First, because the fast response has a marked spatial pattern, secondly because Ocean circulation feedbacks counteract the radiative damping in a few regions.
Because CO2 increase also causes ocean heat uptake, it is hypothesized that abrupt climate change after a THC collapse can only occur when atmospheric CO2 decreases by enhanced ocean carbon uptake.

Behrens, Erik et al.: The impact of eddy processes on the freshwater distribution and AMOC in Greenland melting scenarios (download)
Full list of authors: Erik Behrens, Arne Biastoch, Claus Böning
[Leibniz Institute of Marine Sciences (IFM-GEOMAR),Kiel, Germany]
To analyse the effect of eddies in Greenland melting scenarios we use a hierarchy of ocean-only model configurations differing in horizontal resolution from non-eddying “coarse” (1/2°) global up to regional high-resolution eddy-resolving (1/20°) configuration. The high-resolution configuration is embedded in an eddy permitting global 1/4° configuration via a two-way nesting scheme. A hindcast simulation (1948-2007) is shown to capture the key features of the boundary current and mesoscale eddy activity in the Labrador and Nordic Seas which are important for a realistic representation of the boundary-interior exchange, convection processes and deep water formation. The host of model configurations was used to simulate a Greenland melting scenario (with additional runoff 100mSv, equally distributed around Greenland).
The results show that the eddy based exchange of Greenland melt water/freshwater between boundary current (WGC) and interior Labrador Sea has important consequences for changes in the convection intensity, with implications for the response of the subpolar gyre and the AMOC.

Mjell, Tor. Lien, et al.: Variability in ISOW Vigor Over the Last two Millennia and its Relationship to Climate (download)
Full list of authors T. L. Mjell (1,2), U. S. Ninnemann (1,2), H. F. Kleiven (1,2), Y. Rosenthal (3), I. R. Hall (4)
[(1) Department of Earth Science, University of Bergen, Norway]
[(2) Bjerknes Centre for Climate Research, University of Bergen, Norway]
[(3) Institute of Marine and Coastal Science and Department of Geology, Rutgers The State University of New Jersey, USA]
[(4) School of Earth, Ocean and Planetary Sciences, Cardiff University, Wales, UK]
Low frequency variability in the Atlantic Meridional Overturning Circulation (AMOC) constitutes a key uncertainty in predictions of future climate, ocean, and atmospheric CO2 changes. Although AMOC variability is commonly invoked to explain low frequency climate variability observed on millennial to multidecadal timescales during the Holocene, there is very little observational or paleoclimatic evidence available to test these hypotheses. In short, empirical constraints on the nature and magnitude of ocean variability (under a range of boundary conditions) are sorely needed before we can hope to evaluate and understand the role of the ocean in either past or future climate changes. Here we use well dated (210Pb and AMS 14C), high sedimentation rate, multi and gravity cores taken on the Gardar Sediment Drift (60°19N, 23°58W, 2081 m water depth) to reconstruct decadal to centennial variability in the properties and vigor of the eastern branch of the Nordic Seas overflows over the past two millennia. The Gardar drift accumulates on the eastern flank of the Reykjanes Ridge due to the supply of sediments provided by the overlying Iceland Scotland Overflow Water (ISOW)—an important contributor of North Atlantic Deep Water(NADW). We reconstruct changes in the vigor of ISOW from size variations in the sediment proxy mean sortable silt, while changes in local hydrography are reconstructed using the δ18O of the planktonic foraminifer species N. pachyderma (d), G. bulloides, and G. inflata, in addition to Mg/Ca paleotemperature reconstructions from N. pachyderma (d).
Our records provide a sub-decadally sampled history of ISOW variability spanning the last ~2000 years. Our results reveal that AMOC variability is tightly coupled to low frequency variations in basin-wide climate (Atlantic Multidecadal Oscillation-AMO). In particular, that the eastern branch of the Nordic Seas overflows (ISOW) has varied closely in phase with AMO over the past ~350 years on both inter-decadal and centennial timescales, with increased (decreased) ISOW vigor during warm (cold) AMO phases. The similarities suggest that key components of AMOC are tightly coupled to basin wide temperature perturbations.
Furthermore, our finding of a correlation between high medieval temperatures in Western Europe (800-850 AD) and a vigorous ISOW, as well as a correlation between low summer temperatures in Iceland and a markedly more sluggish ISOW (990-1050 AD), indicates that this coupling is representative for the past ∼2000 years.

Corti, Susanna: On the impact of initial conditions relative to external forcing on the skill of decadal predictions: preliminary results with the ECMWF coupled model (download)
[European Centre for Medium-Range Weather Forecasts, UK]
In this study, results of decadal ensemble hindcasts over the period 1960-2005 are presented. The experiments were performed at ECMWF using the IFS/NEMO coupled model initialised with analysis data. The ocean conditions have been produced with NEMOVAR, a multivariate 3D-var data assimilation method. The atmosphere and land surface initialization was from the ERA-40 and ERA-Interim reanalysis.
The decadal prediction experiments show a positive forecast quality that can be statistically significant over several areas. When the linear climate trend is subtracted, some regions (common to all the experiments carried out) of more pronounced predictability have been identified.
To address the question about the relative importance of external forcing and initial conditions on the skill of decadal forecasts, we performed a sensitivity experiment for the starting dates of Nov1965 and Nov1995. Preliminary results of this experiment indicate that over time scales longer than 5 years predictability arises mainly from the forcing. On a global domain, the correct initialisation has a strong impact only for the first year of integration. However, when the North Atlantic region is considered, the system seems to be sensitive to initial conditions much longer. For the (two) specific starting dates considered the signal of initial conditions is detectable up to the 5th year of forecast.

Park, Wonsun and Latif, Mojib:
Atlantic Meridional Overturning Circulation response to idealized external forcing (download)
[Leibniz Institute of Marine Sciences (IFM-GEOMAR),Kiel, Germany]
The response of the Atlantic Meridional Overturning Circulation (AMOC) to idealized external (solar) forcing is studied in terms of the internal (unforced) AMOC modes with the Kiel Climate Model (KCM), a coupled atmosphere-ocean-sea ice general circulation model. The statistical investigation of KCM’s internal AMOC variability obtained from a multi-millennial control run yields three distinct modes: a multi-decadal mode with a period of about 60 years, a quasi-centennial mode with a period of about 100 years and a multi-centennial mode with a period of about 300–400 years. Most variance is explained by the multi-centennial mode, and the least by the quasi-centennial mode. The solar constant varies sinusoidally with two different periods (100 and 60 years) in forced runs with KCM. The AMOC response to the external forcing is rather complex and nonlinear. It involves strong changes in the frequency structure of the variability. While the control run depicts multi-timescale behavior, the AMOC variability in the experiment with 100 year forcing period is channeled into a relatively narrow band centered near the forcing period. It is the quasi-centennial AMOC mode with a period of just under 100 years which is excited, although it is heavily damped in the control run. Thus, the quasi-centennial mode retains its period which does not correspond exactly to the forcing period. Surprisingly, the quasi-centennial mode is also most strongly excited when the forcing period is set to 60 years, the period of the multi-decadal mode which is rather prominent in the control run. It is largely the spatial structure of the forcing rather than its period that determines which of the three internal AMOC modes is excited. The results suggest that we need to understand the full modal structure of the internal AMOC variability in order to understand the circulation’s response to external forcing.

Swingedouw, Didier et al.: Decadal fingerprints of fresh water discharge around Greenland in a multi-models ensemble  (download)
Full list of authors: Swingedouw D., Rodehacke C., Behrens E., Menary M., Olsen S., Gao Y., Mikolajewicz U., Mignot J., Biastoch A.
In order to investigate the impact of a large Greenland ice sheet (GIS) melting on the ocean and climate, we use six models: one ocean-only model and five state-of-the art AOGCMs. We utilize an idealized framework where we put 0.1 Sv around Greenland for the historical era 1965-2004. We found similar fingerprints after 4 decades of hosing in the different models and a general weakening of Atlantic Meridional Overturning Circulation (AMOC) and of the gyres. The fresh water spreads along the main currents. A large part of the negative salinity anomalies can be found along the Canary Current after 4 decades, which we call the fresh water leakage. This leakage is actually an escape path for the fresh water anomalies from the North Atlantic. As a consequence, we show that the AMOC weakening is smaller if the leakage is large. We relate this leakage with the gyre asymmetry among the models. When compared with observations, we notice that this asymmetry is in better agreement with models showing a limited leakage and the largest AMOC weakening, indicating a potential larger than previously thought AMOC weakening in response to GIS melting.

Herbaut, Christophe, Iovino Dorothea and Houssais, Marie-Noelle: Impact of ice-ocean model resolution in the Nordic Seas on the simulated exchanges over the Greenland-Scotland sills (download)
[Université Pierre et Marie Curie, France]
In order to evaluate the impact of the resolution on the intensity and properties of the exchanges through the Greenland-Scotland passage, a realistic ice-ocean model of the Arctic/Atlantic Ocean sector is used with an horizontal resolution of 1/4° and forced by the ERA-Interim reanalysis during the period 1989-2009.The results of this moderate resolution configuration are compared with the outputs of a second model configuration which includes a local mesh refinement for the ice-ocean model over the Nordic Seas (horizontal resolution of 1/16°, eddy-resolving).
As expected the level of turbulent kinetic energy is increased in the nested model with eddies concentrated along topographic features. The mean flow is also more energetic and better constrained by the topography, leading to increased fresh water export from the Arctic through Fram Strait. Larger amounts of fresh water are therefore brought along the EGC and WGC to the Greenland Sea and Labrador Sea, acting to reduce convection there. Moreover, the Nordic Seas boundary currents, especially the Atlantic Current, are very stable leading to underestimation of eddy fluxes and basin-slope exchanges and perhaps damping of the seasonal cycle of the convection. The Denmark Strait overflow does not change much between the different simulations while the FSC overflow decreases. Reasons for this are being investigated. Another impact of finer resolution of the topography is a better representation of the entrainment in the DSO plume south of the strait. Changes are seen from a large diffuse overflow to a more confined overflow hanging over the upper slope in the high resolution model.

Jonsson, Steingrímur and Héðinn Valdimarsson: Variability and forcing of the flow of water masses on the north Icelandic shelf (download)
Authors: Steingrímur Jónsson (1,2) and Héðinn Valdimarsson (1)
[(1) Marine Research Institute, Skúlagata 4, 101 Reykjavík, Iceland]
[(2) University of Akureyri, Borgir v/Norðurslóð, 600 Akureyri, Iceland]
The north Icelandic shelf is an area that shows highly variable hydrographic properties. In Denmark Strait, between Greenland and Iceland, the warm saline Atlantic Water of the Irminger Current meets the cold and relatively fresh Polar Water of the East Greenland Current originating from the Arctic Ocean. A mixture of these two water masses then flows as the North Icelandic Irminger Current along the shelf north of Iceland and it can vary from being almost pure Atlantic Water to consisting nearly entirely of Polar Water. To determine the flow along the shelf, the Marine Research Institute in Iceland has been monitoring the flow with current meters on a section north of Iceland since 1994. Measurements with a vessel mounted ADCP have been done on several occasions along the section during this time. Together with the current measurements, CTD measurements have been made on standard sections in the area. All these measurements are used to study the structure of the flow, its variability and forcing. In the period 1994-2010 the flow consisted on the average of 68% of Atlantic Water with its transport being 0.88 Sv and the associated heat transport was estimated to be 24 TW. Since 1994 there is a positive trend in both the transport of Atlantic water and the heat transport. The increase in the heat transport is mainly due to an increase in the temperature of the Atlantic water. Both the transport time series and the water mass composition of the current are closely correlated with the NCEP wind over the area. The transport responds very quickly to changes in the wind whereas the water mass composition takes longer time to adjust. Thus the wind is a primary force driving the variability of the two parameters. However the direct wind forcing does not explain the mean flow of Atlantic water to the north Icelandic shelf.

Berx, Barbara et al.: Variability of volume, heat and salt transport in the Faroe Shetland Channel on seasonal and inter-annual time scales
Full list of authors: B. Berx (1), B. Hansen (2), S. Østerhus (3), and T. Sherwin (4)
Contact author: Barbara Berx
[(1) Marine Scotland Science, Marine Laboratory, Aberdeen, United Kingdom]
[(2) Havstovan, Faroe Marine Research Institute, Tórshavn, Føroyar]
[(3) Bjerknes Centre for Climate Research, Bergen, Norway]
[(4) Scottish Association for Marine Science, Oban, United Kingdom]
The inflow of Atlantic Water into the Nordic Seas through the Faroe Shetland Channel (FSC) is one of three branches of Atlantic Water crossing the Greenland-Scotland Ridge (GSR). Comparable transports of occur through the FSC and between Iceland and Faroes, with a minor contribution West of Iceland making up the third branch. Volume transport is monitored by an array of Acoustic Doppler Current Profilers deployed across the channel. Established in 1994, the observations upto 2010 will be presented. Using a new method, which takes account of the two water masses contributing to Atlantic Water (AW) circulation in the FSC (North-Atlantic Water and Modified North-Atlantic Water), mean volume transport of AW is estimated to be 4.4 Sv (± 1.4 Sv; 1 Sv = 106 m3 s-1). Combined with ship-based observations of temperature and salinity, mean heat and salt transport of AW are estimated to be 172 TW (± 59 TW; 1 TW = 1012 W) and 159 kt s-1 (± 52 kt s-1; 1 kt s-1 = 106 kg s-1), respectively.
The seasonal and inter-annual variability in these observational time series are presented, and the question whether volume, heat and salt transport are influenced by changing temperature and salinity will be investigated.

Paka, Vadim: Preparation for the final microstructure measurements in the Denmark Strait overflow (download)
[Atlantic Branch of Shirshov Institute of Oceanology, Russia]
The task for microstructure measurements in the Denmark Strait overflow plume is to examine the nature and processes controlling the entrainment of ambient water into the plume. During the first field experiment in 2009 the necessary instrument – a free falling vertical microstructure profiler BAKLAN-3500 specifically developed to measure fine structure and dissipation-scale turbulence in deep open ocean - was successfully tested and introduced into the MSM 12/1 Rosette survey. Some important features of the microstructure above and within the overflow plume at different distances from the Denmark Strait Sill were discovered. However the obtained data are still not sufficient for proper solving the task. Some disadvantages were found which could be suppressed, so, the continuation of the microstructure experiment is necessary which should be preceded by improvement of the instrument and measurements techniques.
To better understanding positive and negative features of our instrument, the comparison of BAKLAN-3500 with the popular free-falling microstructure profiler TurboMAP [www.jfe-alec.co.jp/html/turbulence.htm] was made. The main negative feature there was inadequate accuracy of dissipation estimates for some reasons, which had been studied and eliminated.
Next measures were undertaken to get better results from the final microstucture experiment:
-reconstruction of the sensors’ guarding cage to suppress the vibration noise;
-reconstruction of the buoyancy module to decrease the mass and dimensions and improve the shape of the streamline body: two formerly used 13” glass spheres were replaced by the two smaller (10”) ones;
-design of the Rosette-mounted drift compensator to avoid the cut of free-falling path length of the profiler;
-installation of standard calibrated PNS06 shear probes in addition to our own ones;
-improving the fast-response thermometer; further thermal microstructure measurements will be made by either original flat-spiral cupper thermometer with time constant of order of 10 ms or/and by standard FP7 micro-bead thermistor.

Hansen, Bogi et al.: Mixing of the Faroe Bank Channel overflow by convective events (download)
Full list of authors: Bogi Hansen, Karin Margretha H. Larsen, Svein Østerhus, Detlef Quadfasel
[Faroe Marine Research Institute, Box 3051, FO-110 Torshavn, Faroe Islands]
[UNI Bjerknes and Geophysical Institute, University of Bergen, Norway]
[University of Hamburg, KlimaCampus, Institute of Oceanography, Hamburg, Germany]
From May 2008 to May 2009, an Acoustic Doppler Current Profiler (ADCP) was deployed in a frame on the bottom of the Faroe Bank Channel (FBC). The frame was located on the northeastern slope of the channel in the overflow layer and recorded velocity profiles every 20 minutes. In the comprehensive data set acquired, a number of events can be identified with extreme vertical velocities in the overflow layer. These events have a pronounced asymmetric character and may have vertical velocities exceeding 10 cm/s upwards for several hours, implying that they are not caused solely by internal waves. The overflow layer in the FBC has a high degree of vertical homogeneity and it has been hypothesized that this is associated with the helical circulation system that has been shown in several studies. This circulation system will over the Faroe slope tend to bring lighter water below denser water and induce instabilities. The presentation discusses, to what extent the events observed can be interpreted as convective mixing events induced by this.