Lesley Allison*, The RAPIT Team
Met Office, NOC, BAS, LSE, Imperial College and the Universities of Reading, Oxford and Durham
The Risk Assessment, Probability and Impacts Team are working to estimate the risk of the collapse of the overturning circulation in the North Atlantic. Part of the project is concerned with understanding the mechanisms for the rapid slow down of the overturning and the impacts these slowdowns have on the wider climate system while part is concerned with the estimation of the risk using coupled climate GCM’s. During the year we have been examining a number of GCM’s to see if we can discern a common mechanism for rapid, unforced changes in the strength of the overturning circulation. The long control run of HADCM3, combined with a small perturbed parameter ensemble (THCQUMP), allows us to examine the multidecadal variability in the overturning. Computational problems have prevented us running our large ensemble experiments yet. However these are largely solved now and we are designing and starting to run experiments with Famous and HADCM3. The initial experiments will use hierarchical statistical methods to relate a large ensemble of Famous runs to a smaller HADCM3 ensemble. The emulators produced in this experiment will then be used to inform our designs for future HADCM3 ensembles.
Download full presentation
Lesley Allison1,*, Ed Hawkins1, Tim Woollings1
NCAS-Climate, University of Reading
The RAPID-WATCH Risk Assessment, Probability and Impacts Team (RAPIT) project aims to enhance our understanding of the uncertainties surrounding the risk of a future rapid change in the Atlantic Meridional Overturning Circulation (AMOC). As a step towards this goal, we study the largest rapid (decadal) natural fluctuations in AMOC strength within coupled general circulation model (GCM) control integrations. The eventual aim is to generate a time-varying fingerprint of precursors to, and climate impacts of, these rapid events, which is robust across different climate models. Understanding the mechanisms involved in the rapid fluctuations may help to explain the difference in AMOC stability between GCMs and lower-complexity models. To compare events across different climate models, which commonly have rather different AMOC characteristics in terms of mean state and variability, a consistent definition of a large rapid event must first be established. We then compare the mechanisms involved in these rapid events with the mechanisms associated with more general AMOC variability. The initial focus is on rapid events within a control integration of GFDL CM2.1; the analysis is then extended to several other climate models, and the latest results will be presented.
Christopher P. Atkinson1*, Harry L. Bryden1, Stuart A. Cunningham1, Brian A. King1
National Oceanography Centre, University of Southampton, European Way, Southampton, UK, SO14 3ZH
During January and February 2010, a full hydrographic section was completed in the sub-tropical Atlantic along 24ºN. This was the sixth occupation of this section over the past 5 decades. Calculation of water mass transports in depth classes for the 2010 section reveals an apparent return in strength of the Atlantic thermohaline circulation to that of the early 1990s of 18 Sv, in contrast to suggestions that the circulation may be weakening with time. This variability however falls within the apparent seasonal range of overturning circulation as recently established by the RAPID-WATCH array. In the deep ocean, a reduction in lower North Atlantic deep-water transports seen in the mid to late 1990s has persisted into 2010. The lower deep waters also appear to have continuously freshened since the 1950s. The significance of these trends in water mass properties is assessed through comparison with RAPID-WATCH mooring data and the associated changes in deep-water transport are discussed.
Magdalena Balmaseda 1, Adam Blaker2, Keith Haines3,*, Leon Hermanson3, Joel Hirschi2, Alan Iwi4, Franco Molteni1, Jon Robson3, Bablu Sinha2, Doug Smith5, Vladimir Stepanov3, Rowan Sutton3
1) ECMWF2) National Oceanography Centre3) University of Reading4) STFC5) Met Office Hadley Centre
The overarching goal of VALOR is to assess the value of the RAPID observations for predictions of the Atlantic meridional overturning circulation (AMOC) and its impacts on climate. In addition, the project seeks to make recommendations on the design of a potential AMOC prediction system. Most work has initially focussed on the development of assimilation methodologies for the RAPID observations, observing system experiments and creating syntheses to provide initial conditions for ocean only and coupled ocean-atmosphere hindcasts. Model representations of the different MOC components observed in the data have been studied and diagnostic tools have been developed for this purpose. Covariance relationships relevant to the AMOC have been examined in both the NEMO ocean model and the HadCM3 climate model in order to assess the appropriate influence of the Rapid array observations. High frequency MOC variability in models associated with internal gravity waves have also been found and new studies made. The groundwork has been laid for comparable high resolution NEMO model runs with and without data assimilation to be made over the coming 6 months. Work on assimilating the Rapid array observations into a lower 1 degree resolution NEMO has demonstrated some of the constraints involved and been optimized to give significant improvements in the modeled MOC. Both the Met Office and the University of Reading are working on developing the existing decadal prediction system based on HadCM3, DePreSys, for the use with RAPID observations. Both observational analysis methods and hindcast ensemble creation is being considered. Initial results show that a simple initial condition ensemble has less skill in hurricane prediction than a perturbed physics ensemble. Current results will be reviewed.
Download full presentation [Restricted to RAPID]
Adam T. Blaker1,*, Joel J-M. Hirschi1, Bablu Sinha1
National Oceanography Centre Southampton
Analysis of a short NEMO 1/4 degree integration with high frequency output reveals a field of fast propagating inertia-gravity waves which propagate equatorward due to beta dispersion. Propagation speed decreases with latitude. The inertia-gravity waves lead to zonally coherent bands of velocity anomalies. At 40°N propagation speed is 3-5 degrees/day. This decreases to 1-2 degrees/day at 20°N. Vertical velocities associated with these waves are large, typically 35-40 m/day. The passage of these inertia-gravity waves strongly influences the Atlantic MOC (Meridional Overturning Circulation) on short timescales (i.e. inertial period). The waves lead to deep-reaching MOC cells propagating equatorward. The meridional extent of these cells ranges from around 1° at mid-latitudes to about 3-5° close to the equator and the maximum values typically occur at depths of around 2000m. The MOC variability associated with the passage of inertia-gravity waves is large. At 26.5°N the MOC variability about the mean of 25 Sv can exceed 60 Sv in less than one day.
Stuart Cunningham, Sheldon Bacon
National Oceanography Centre
Brief overview of the Overturning in the Subpolar North Atlanic Program: OSNAP
Thomas L. Delworth1,*, Rong Zhang1, Shaoqing Zhang1, Tony Rosati1, Keith Dixon1, Ryn Msadek2
1) Geophysical Fluid Dynamics Laboratory/NOAA2) Princeton University
We review recent and ongoing research at GFDL that seeks to better understand the role of the Atlantic and the AMOC in the global climate system, including climatic impacts and potential predictability of the AMOC. We use a variety of modeling tools in concert with analyses of the observed record, incorporating both the instrumental and paleoclimate records. We show that Atlantic temperature changes, likely related to AMOC changes, have a major influence on monsoonal rainfall from India to Africa, North American rainfall, Atlantic atmospheric circulation of relevance for hurricanes, and even can project onto hemispheric temperature. Further, by combining the observed record and modeling studies we show that a fingerprint of the AMOC can be established using subsurface temperature records and satellite observations. These can be used to make inferences about past AMOC activity. We also summarize ongoing efforts to explore decadal predictability of the AMOC, and to develop a decadal prediction system.
Shane Elipot1,*, Chris Hughes1,*, Miguel Angel Morales Maqueda1,*, Ric Williams2,*
1) National Oceanography Centre, Liverpool2) University of Liverpool
The objective of the Rapid-WAVE and Rapid-WATCH programmes of the National Oceanography Centre, Liverpool, (formerly Proudman Oceanographic Laboratory) in collaboration with the Canadian Bedford Institute of Oceanography, is to monitor the lower limb of the North Atlantic Meridional overturning at 42 N. Specifically, our Rapid-WATCH programme is based on the theory of using boundary pressure only to estimate meridional transport. It is simply based on the dominance of geostrophic balance for large-scale flows, and results from a zonal integral of the zonal momentum equation. This theory has been validated so far by numerical ocean models. The Rapid Scotian array, deployed in 2008 across the continental slope on the western boundary off the coast of Nova Scotia, between approximately 1000 m and 4000 m, consists in five short and one tall deep moorings, all fitted with bottom pressure recorders (BPR), upward-looking Acoustic Doppler Current Profilers and CTD instruments. The "step method" consists in estimating the pressure variability down the slope by combining the estimates of horizontal pressure gradient from velocity data and the estimates of hydrostatic pressure variability from density data. This method is shown to work well, as it replicates the pressure differences along the slope obtain from the bottom pressure recorders alone, at time scales shorter than the seasonal time scales where the BPR data are most reliable. By subsequently integrating the pressure variability from current and density data alone down the western slope, a first time series estimate of the western contribution to the transport anomaly below 1000 m is obtained.
Eleanor Frajka-Williams1,*, Stuart Cunningham1, Joel Hirschi1, Harry Bryden2, Williams Johns3, Chrisopher Meinen4, Molly Baringer4
1) National Oceanography Centre Southampton, Ocean Observing and Climate Division, European Way, Southampton, SO14 3ZH, UK.2) National Oceanography Centre Southampton, School of Ocean and Earth Science, European Way, Southampton, SO14 3ZH, UK.3) Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker Causeway, Miami FL,USA.4) NOAA, Atlantic Oceanographic and Meteorological Laboratory, 4301 Rickenbacker Causeway, Miami FL,USA.
In the North Atlantic the meridional overturning streamfunction reveals two vertical cells atop one another. The upper cell consists of a northward-flowing, upper branch (shallower than ~1 km) and a southward-flowing lower branch (from ~1 km to ~3 km). Beneath the upper cell is an abyssal cell comprising the northward flow of Antarctic Bottom Waters (AABW), which progressively rise to turn and join the deep southward flow before reaching 40°N. Together these two cells comprise the Atlantic Meridional Overturning Circulation. Since 2004 the RAPID-MOC/MOCHA programme has been monitoring the daily strength and vertical structure of the basin-wide, full-water-column AMOC at 26.5°N. These estimates to date have used climatological values of the net northward flux of AABW deeper than 4800 m rather than direct measurements: approximately 2 Sv at 26.5°N. In May 2009 as part of the RAPID-MOC/MOCHA mid-ocean array we deployed two additional moorings to directly estimate the transport and variability of the lower AABW cell. These moorings measured density on either side of the western basin to a depth of 5600 m. Here we describe the vertical structure and variability of the AABW cell as inferred from these new data, and compare these to the structure of AABW transport inferred from six hydrographic cruises. Finally, we describe the impact of the abyssal cell for estimates AMOC strength.
Keith Haines1,, Vladimir Stepanov1, Leon Hermanson1,2, Greg Smith1,3, Adam Blaker4
1) University of Reading2) Met Office3) Environment Canada4) NOC Southampton
In ocean and climate models such as NEMO and HadCM3 the upper layer poleward transports near the western boundary at 26N are not separated cleanly by a florida strait barrier and tend to spill over into the upper levels of the deep basin further east. In consequence the first deep western boundary grid points of these models have to support a strong vertical shear at upper levels to accomodate this "Florida strait spillover". In consequence their upper layer density near the western wall cannot be consistent with observations from the Rapid mooring array eg. at wb2 the first full depth observation profile. The different model components of the MOC and their mass transport compensations as defined recently in Bryden et al (2009) will be considered. Also this has important implications for any attempt to assimilate Rapid array density data near the western boundary of these models in an attempt to constrain MOC transports. Examples will be shown from the NEMO and HadCM3 models and possible alternative methods of assimilating Rapid array data will be discussed.
Ed Hawkins1, Robin Smith1, Lesley Allison1,*, Jonathan Gregory1,2, Jose Rodriguez2, Richard Wood2
1) NCAS-Climate, University of Reading2) Met Office Hadley Centre
Previous studies have found that the Atlantic thermohaline circulation (THC) in intermediate complexity climate models (EMICs) can exhibit hysteresis when additional freshwater is applied at the surface (hosing) - i.e. more than one equilibrium THC state exists for some values of hosing; the state actually adopted by the system is dependent on the history of the simulation. The presence of such behaviour in the real climate system would be significant. However THC hysteresis has not previously been found in fully coupled GCMs. We make a specific search for hysteresis in the FAMOUS GCM (a low resolution version of HadCM3). We have attempted to map out the hysteresis curve, and find that both a stable THC 'on' and a THC 'off' state can be maintained with the same hosing, indicating the presence of hysteresis. We also explore the sensitivity of this finding to the hosing regions considered, and consider whether such behaviour is predictable.
Ed Hawkins1, Jon Robson1,*, Rowan Sutton1,*, Doug Smith2, Noel Keenlyside3
1) NCAS-Climate, University of Reading2) Met Office3) IFM-GEOMAR, Kiel
We explore the potential for making statistical decadal predictions of sea surface temperatures (SSTs) in a perfect model analysis, with a focus on the Atlantic basin. Various statistical methods (Lagged correlations, Linear inverse modelling and Constructed analogue) are found to have significant skill in predicting the internal variability of Atlantic SSTs for up to a decade ahead in control integrations of two different global climate models (GCMs), namely HadCM3 and HadGEM1. Which is the most successful statistical method depends on the region, GCM data used and prediction lead time. Importantly, the regions of greatest prediction skill can be very different to the regions identified as being potentially predictable from variance explained arguments, suggesting that significant local decadal variability is not a prerequsite for skillful decadal predictions, and that the statistical methods are capturing some of the dynamics of low-frequency SST evolution. In particular, using data from the HadGEM1 GCM, significant skill at lead times of 6-10 years is found in the tropical North Atlantic where there is relatively little decadal variability compared to interannual variability. Also, the estimated skill from the statistical methods is comparable to that of two operational decadal prediction systems, one using HadCM3 and another using ECHAM5/MPI-OM. Finally, we explore whether adding sub-surface temperature data improves these decadal statistical predictions, and find that, again, it depends on the region, prediction lead time and GCM data used. Overall, we argue that the estimated prediction skill motivates the further development of statistical decadal predictions of SSTs as a benchmark for current and future GCM-based decadal climate predictions.
Leon Hermanson1, Rowan Sutton1,2
1) University of Reading2) NCAS-Climate
The RAPID array consists of several moorings at about 26°N in the Atlantic that measure temperature and salinity from which the meridional overturning circulation (MOC) and its components at this latitude can be determined. The main aim of the VALue Of the RAPID array (VALOR) project is to assess the value of the RAPID array observations for predictions of the Atlantic MOC and its impact on climate. In addition, the project will explore a range of issues concerning the design of a potentional MOC prediction system. One of the models used in this work is the Hadley Centre HadCM3 model. In this work we have investigated the potential issues that may arise when comparing RAPID array observations and its derived transports with model variables. This could be for comparing an ocean assimilation or an ocean forecast with observations. The main issue is the low resolution of the HadCM3 ocean (1.25°x1.25°), which changes the bathymetry. As an example, most of Florida does not exist, so the Florida Straits, which is a crucial part of the observations, is missing. Moving the model array further north to where there is a Florida is considered, but no convincing reason to do so is found. Furthermore, initial assessments of correlations between the observed transport and basin-wide temperature or salinity show promise for an alternative method of assimilating the RAPID array observations into the model.
Dan Hodson1, Rowan Sutton1
Atlantic Sea Surface Temperatures (SSTs) exhibited distinct decadal variability during the 20th Century. Modelling studies suggest that this variability may be at least partly driven by variations in the Atlantic Meridional Overturning Circulation (AMOC). Decadal ocean variability may be an important source of climate predictability on decadal timescales via the modulation of atmosphere-forcing SST patterns. It is important therefore to understand the mechanisms involved in generating such decadal variability within the AMOC, and the accuracy and limitations of climate models in representing such processes. One source of decadal variability in the AMOC is variability in the sinking of dense waters in the subpolar north Atlantic. Adjustment of the AMOC requires that such high latitude density changes are communicated throughout the entire Atlantic Basin. Evidence from coupled climate models and idealised ocean models suggests that this adjustment is communicated by the propagation of density signals along the western Atlantic Ocean boundary. The timescale for complete adjustment may therefore depend on the speed of propagation of these boundary density signals. Previous studies have suggested that the propagation speed of these signals may be significantly influenced by model resolution, particularly in the high latitude regions where the density signals are generated. In this study we examine both the propagation of such boundary density signals, and the overall AMOC adjustment, in both a high-resolution coupled climate model (1/3 degree x 1/3 degree ocean) (HiGEM) and a standard resolution climate model (HadGEM). We assess the impact of model resolution on the large scale decadal variability of the AMOC, and on the nature of the associated ocean-atmosphere interactions.
Chris W. Hughes1, Shane Elipot1, Miguel Angel Morales Maqueda1
National Oceanography Centre (Liverpool)
We now have bottom pressure measurements from September 2004 to September 2009, at depths from 1100 m to 4100 m down the continental slope near to Halifax. We discuss the character of these measurements, highlighting the different kinds of variability to be seen. The dominant signal is a variability which is uniform across the slope, and explains more than 90% of the variance, consistent with expectations for a large spatial-scale barotropic variability. This has a power spectrum approximately proportional to frequency to the power -2. Differences between pressure measurements show much less variability, and a quite distinct spectrum. At frequencies higher than inertial, there is significant power in pressure differences, which is generally poorly correlated down the slope, consistent with the expected small length scales of internal waves (and including peaks of energy corresponding to the semidiurnal tide and its harmonics). Between the inertial period and about 5 day period, there is very little energy in the pressure differences. Energy then increases, with broad peaks near about 8 days and 25 days. Energy in these modes is very coherent down the slope to a depth of 3650 m, corresponding to a second pressure mode in which pressure variability is a linear function of depth. Deeper than 3650 m, variability in pressure differences at these frequencies increases dramatically, as the continental slope flattens out and the corresponding suppression of eddy energy ceases to operate. We therefore have a clear signal of changes in the vertical shear of zonally-integrated meridional transport, which are coherent over the depth range 1100-3650 m, exist at periods between 5 and 40 days, and persist over five years.
Laura Jackson1, Michael Vellinga1
Met Office Hadley Centre
Large, multidecadal changes in the meridional overturning circulation (MOC) are examined in a coupled GCM (HadCM3) with constant greenhouse gas concentrations. These changes are strongly related to salinity anomalies in the sinking regions and hence may help us understand possible mechanisms behind a rapid shutdown of the MOC. The drivers of the large multidecadal changes and coupled feedbacks associated with the changes are investigated. These involve the propagation of anomalies into the region as well as remote and local feedbacks. A perturbed physics ensemble based on HadCM3 (where physics of the atmosphere, land and sea-ice is perturbed) is also examined. The ensemble members have different MOC strengths and variability due to the different surface forcings caused by the different physics perturbations. The drivers and feedbacks of MOC changes in the ensemble are also analysed, and are compared to those in the control HadCM3 simulation, with the aim of understanding the sensitivity of the processes involved.
Bill Johns1, Tong Lee2, Fabrice Hernandez3, Nicolas Perry3, Eric Chassignet4, Erik Van Sebille1
1) RSMAS, University of Miami, Miami FL2) NASA Jet Propulsion Laboratory, Pasadena, CA3) Mercator Océan, Ramonville St Agne, France4) Center for Ocean-Atmospheric Prediction Studies, Florida State University, Tallahassee, FL
Estimates of MOC strength, MOC vertical streamfunction, and meridional heat transport derived from the RAPID-MOCHA array for 2004-2007 are compared to results from a number of data assimilation model reanalyses. The models include the ECCO and ECCO-2 ocean state estimate products and the high-resolution GODAE/HyCom and GODAE/Mercator products. We focus on how well these models are able to capture the characteristics of the observed MOC variability at 26ºN, in terms of overall strength, temporal variability, and seasonal cycle, and on the relationship between heat transport and MOC variability in the models as compared to observations. Substantial differences are found in the seasonal variation of maximum MOC at 26ºN between the RAPID data and two ECCO ocean state estimation products that have not assimilated the RAPID data. However, a comparison of the correlation and regression between MOC and meridional heat transport show reasonable agreement between the RAPID data and the ECCO products. The Mercator GLORYS1V1 reanalysis - also not assimilating RAPID data - reproduces the observed MOC seasonal cycle at 26ºN quite well, but with slightly lower amplitude, and has an overall correlation of 0.71 with the observed MOC variability from 2004-2007. There is a tendency in most of the models for the southward NADW return flow to occur at shallower average depth than in observations, with relatively more upper and middle NADW and less lower NADW.
Brian King1,*, Gerard McCarthy1, Harry Bryden1
National Oceanography Centre, Southampton
A transatlantic hydrographic section along 24°S was taken in March-April 2009 aboard RRS James Cook. Geostrophic estimates of the meridional circulation with an initial reference level at 1500 dbar indicate that there is a net northward transport of 21.5 Sv above 1300 dbar with a compensating southward flow below 1300 dbar. Estimates of northward heat transport and freshwater transport are made; in particular we estimate the mov freshwater transport, a key indicator of the stability of the Atlantic overturning circulation according to presentations by Dijkstra and Drijfhout at past Rapid meetings. We also compare results with the 1983 transatlantic section along 24°S to assess the robustness of the new estimates.
Valerie N. Livina1,*, Frank Kwasniok2,, Gerrit Lohmann3,, Jan Kantelhardt4,, Yuri Sapronov5,, Tim M. Lenton1
1) University of East Anglia, UK2) University of Exeter, UK3) Alfred Wegener Institute for Polar and Marine research, Germany4) Institute of Physics, Martin-Luther-Universitat-Halle-Wittenberg, Germany5) Voronezh State University, Russia
We apply a new method of potential analysis to study several paleo and historic climatic records. The method comprises (i) derivation of the number of distinct global states of a system from time series, (ii)derivation of the potential coefficients using an Unscented Kalman Filter (UKF), yielding indications of possible bifurcations and transitions of the system. The method is tested on artificial data and then applied to climatic records spanning progressively shorter time periods from 5.3 Myr (Raymo benthic stack) to recent times. In particular, we study Atlantic Multidecadal Oscillation index and detect a potential change in the last 25 years. The method can be applied to a wide range of geophysical systems where time series of sufficient length and temporal resolution are available and transitions or bifurcations are surmised. [1] Livina V., F. Kwasniok, and T. Lenton, Potential analysis reveals changing number of climate states during the last 60 kyr, Clim. Past Discuss., 5, 2223-2237, 2009.
Kevin Marsh1,*, Robin McCandliss2, Julie Collins2, Mark Hebden2
1) British Atmospheric Data Centre2) British Oceanographic Data Centre
The RAPID Data Centre (RDC) provides data support to all projects in the RAPID-WATCH programme. It is made up of the British Oceanographic Data Centre (BODC) and the British Atmospheric Data Centre (BADC), the NERC Designated Data Centres for Marine and Atmospheric sciences, respectively. The BODC is responsible for the archival of observational data, and the BADC holds numerical model output. The RDC provides a long term archive for RAPID and RAPID-WATCH datasets. It also gives comprehensive advice on all aspects of data management, and promotes data sharing and collaboration between the various RAPID-WATCH projects. The RDC supports the RAPID-WATCH community by providing third-party datasets, on-line tools for data analysis, and a dedicated user helpdesk.
David Marshall1,*, Helen Johnson1, Xiaoming Zhai1
University of Oxford
We present results from an analytical reduced-gravity model for the propagation and coherence of meridional circulation anomalies western and eastern boundaries, and their interaction with long Rossby waves in the ocean interior. Our key findings are: (i) For wave periods exceeding a few months, Kelvin waves play no role. Instead, propagation occurs through short and long Rossby waves at the western and eastern boundaries respectively. These Rossby waves propagate zonally, as predicted by classical Rossby wave theory, and cyclonically along the basin boundaries, in order to satisfy the no-normal flow boundary condition. (ii) The along-boundary propagation speed is inversely proportional to width of the appropriate frictional boundary layer; this result appears to hold across a wide range of parameter regimes, with either linear friction or lateral viscosity and a no-slip boundary condition. For parameters typical of contemporary ocean climate models, the propagation speed is coincidentally close to the Kelvin wave speed. (iii) Meridional transport anomalies are constant to leading order as they propagate equatorward along a western boundary, but decay with latitude as they propagate poleward along an eastern boundary, the latter being due to the radiation of long Rossby waves. Precisely the opposite result is obtained for pressure (or sea surface elevation) anomalies. (iv) In the limit of small dissipation, virtually all of the wave energy is dissipated at the western boundary, whereas virtually no energy is dissipated at the eastern boundary. The independence of energy dissipation on dissipation coefficients occurs because the area over which the short Rossby waves extend into the basin interior increases as the inverse of the dissipation coefficient, this precisely compensating for the decrease in the local energy sink. (v) A corollary of the latter result is that the western boundary in our model is a "graveyard" for energy propagating westward in long Rossby waves. (vi) Energy dissipation at the western boundary implies a convergence of eddy bolus transports which requires a small but non-negligible mean western boundary current transport (analogous to a rip current at a beach). (vii) The reduction of eddy energy at the western boundary is consistent with that seen in altimetric data at midlatitudes. This suggests that local eddies should not dominate meridional overturning circulation variability on interannual to decadal time scales, as monitored through end-point hydrographic observations. The relevance of these results to adjustment of the Atlantic Meridional Overturning Circulation and potentiaL implications for diapycnal mixing will be discussed.
Alex Megann1, Adrian New1, Adam Blaker1, Bablu Sinha1
The Coupled Hadley-Isopycnic Model Experiment (CHIME) has been constructed to be as similar as possible to the Hadley Centre’s HadCM3, but with a different ocean component, comprising constant density layers in the ocean interior, as opposed to fixed levels. Here we present key differences between 200-year control simulations of the two models, and early results of climate change experiments. In the control simulations we focus our attention on the respective abilities of the two models to maintain the structure of water masses which are important to the climate system, and use the insights gained to throw light onto the differences in the sea-surface temperatures and the rate at which the upper layers of the ocean warm or cool. The climate change experiments reveal insights into likely uncertainties in climate predictions arising from structurally different ocean components. Results will also be presented relating the variability of the Atlantic Meridional Overturning Circulation (AMOC) in CHIME to variability in surface air temperature over the UK and Western Europe.
Christopher S. Meinen1,*, Silvia L. Garzoli1, Molly O. Baringer1, Stuart A. Cunningham2, Darren Rayner2, William E. Johns3
1) NOAA/Atlantic Oceanographic and Meteorological Laboratory2) National Oceanography Centre, Southampton3) Rosenstiel School of Marine and Atmospheric Science, University of Miami
At 26.5°N the western boundary contributions to the Atlantic Meridional Overturning Circulation (MOC) are carried by the southward Deep Western Boundary Current (DWBC) at depth and by the northward Florida and Antilles Currents in the upper layer. Variability of the MOC in this region is enmeshed with other flow features such as the well-known recirculation cell in the deep-water layer east of the DWBC. Changes in the MOC are correlated in numerical climate models with socially important quantities such as precipitation over North America and northern hemisphere surface air temperatures. Four to five years of data from overlapping lines of pressure-equipped inverted echo sounders and current meter and dynamic height moorings from the Western Boundary Time Series and RAPID-MOC/MOCHA programs are now available to study variations in the Antilles Current and the DWBC east of the Bahamas. These data are used to present a preliminary analysis of the observed variability of these currents near the western boundary and to highlight connections to the basin-wide MOC fluctuations.
Matthew Menary1,*, Wonsun Park2, Michael Vellinga1, Mojib Latif2
1) Met Office Hadley Centre for Climate Prediction and Research2) Leibniz Institute of Marine Sciences
Centennial variability of the Atlantic Meridional Overturning Circulation (MOC) is analysed in multimillennial control simulations with the 3rd Hadley Centre coupled climate Model (HadCM3) and the Kiel Climate Model (KCM). This encompasses approximately forty cycles of our models' centennial MOC oscillation, which has significant power at time-scales of around 120 years in HadCM3 and in KCM. Some support for variability at these centennial timescales comes from palaeo-reconstructions of the last 4500 years that have indicated a similar centennial periodicity in Sea Surface Temperatures (SSTs) North of Iceland which themselves have been linked to ITCZ variability. Long period variability of the MOC may well be an important modulator of anthropogenically induced climatic changes, and its role in past abrupt climate change make understanding the centennial variability of vital importance. The length of our new simulations allows the detail of the centennial mechanism to be investigated as part of a systematic analysis of the similarities and differences in the mechanism between the two models. We show that, despite the length of time in which oceanic signals of MOC variability can remain apparent (for instance the slow northward propagation of Sea Surface Salinity (SSS) signals detected between equatorial regions and the subpolar gyre), the role of the atmosphere in the centennial mechanism is important. Atmospheric influences such as changes in the degree of precipitation within, and position of, the Inter-Tropical Convergence Zone (ITCZ) play a crucial role in switching the mode of oscillation from a positive to negative phase.
Aazani Mujahid1, Harry Bryden1
University of Southampton, School of Ocean and Earth Science
A unique feature of the Rapid array is the combination of full-depth moorings with instruments measuring temperature, salinity and pressure time series at many depths with co-located bottom pressure measurements so that dynamic pressure can be estimated from surface to bottom. After low-pass filtering to remove tidal pressure fluctuations of order 1 dbar, bottom pressure measurements show a zonally uniform rise (and fall) of bottom pressure of 0.015 dbar on a 5 to 10 day time scale, suggesting that the Atlantic basin is filling and draining on a short time scale. After removing the zonally uniform bottom pressure fluctuations, bottom pressure variations at 4000 m depth against the western boundary compensate instantaneously for baroclinic fluctuations in the strength and structure of the deep western boundary current so there is no basin-scale mass imbalance resulting from variations in the deep western boundary current. After removing the mass compensating bottom pressure, residual bottom pressure fluctuations at the western boundary just east of the Bahamas balance variations in Gulf Stream transport. Again the compensation appears to be especially confined close to the western boundary. Sea surface height fluctuations from satellite altimeter measurements are correlated with time series of dynamic height at 100m depth on four full-depth moorings. Variations in sea surface height and in dynamic height both decrease markedly as the western boundary at the Bahamas is approached. Fluctuations in sea surface height (and dynamic height at 100 m) are not significantly correlated with bottom pressure. The vertical structure of dynamic pressure exhibits a mixture of barotropic, first and second baroclinic modes. The strength of the baroclinic modes decreases the amount of correlation between surface dynamic height (and sea surface height) and bottom pressure below significance levels.
Holger Pohlmann1, Magdalena Balmaseda2, Noel Keenlyside3, Daniela Matei4, Wolfgang Müller4, Philippe Rogel 5, Doug Smith1
1) Met Office Hadley Centre, Exeter, UK2) ECMWF, Reading, UK3) IFM-GEOMAR, Kiel, Germany4) Max Planck Institute for Meteorology, Hamburg, Germany5) CERFACS, Toulouse, France
In the first part of this study we compare the Atlantic Meridional Overturning Circulation (AMOC) variability between 10 different analysis systems over the second half of the 20th century from the EU-ENSEMBLES and the German-North Atlantic project. When the AMOCs at 45°N from these different assimilation experiments are standardised they show similar long-term variability, i.e. an increase from the 1960s to the 1990s and a decline thereafter. Importantly, this signal matches observed variations in the North Atlantic Oscillation, Labrador Sea convection, and strength of the sub-polar gyre. Assessing the skill of the AMOC in historical hindcasts is hampered by the lack of observations. However, Pohlmann et al. [2009] found that decadal hindcasts of the AMOC are skilful when assessed against the model analyses from which they are initialised. We also find that the multi-model ensemble predictions of the AMOC at 45°N are skilful.
Darren Rayner1,*, Stuart Cunningham1
As part of the RAPID-MOC 26˚N mooring array we use self-logging current meters to directly record the currents in the western boundary sub-array. These measurements are used to determine the contribution to the Atlantic Meridional Overturning Circulation (MOC) from the Western Boundary Wedge inshore of our prime density mooring. An opportunity arose to test current meters from different manufacturers at the location where they are required for the project and a short test deployment mooring was added to the array. The mooring “WBCM” was deployed for 7 months between April and November 2009. A deep rated Aanderaa RCM11, a Sontek Argonaut MD, a Nortek Aquadopp, and an InterOcean S4 were deployed within 12m vertically of each other, at a depth 4270m. The instruments were deployed 500m above the seabed in a region of low concentration of backscatterers. To add to this we were loaned a 6000m rated Doppler Volume Sampler (DVS) from Teledyne RDI, to allow us to evaluate it against our existing instruments. Here we present the findings of this trial and highlight the differences in relative performance of each instrument. The currents measured during the deployment were highly variable and peaked at 40-45 cm/s.
The RAPID-MOC mooring array at 26˚N is now into its seventh year of deployment and continues to evolve in response to changes in the perceived importance of different moorings for monitoring the Atlantic Meridional Overturning Circulation (MOC). With this continuing evolution of the array it is important to track instrument usage and the design changes – both for the array as a whole and also for individual moorings – to allow calculations of the MOC to be repeatable at any time in the future. Data return and changes to the loss rate of equipment can also be quantified and compared year on year to check that design changes intended to improve either of these are in fact working. Here we present the latest array design, as serviced by the Autumn 2009 eastern boundary cruise aboard the RRS Discovery, and the Spring 2010 western boundary cruise aboard the RV Oceanus. In order to obtain data from the key moorings in a timelier manner – and to help secure data in the event of mooring loss – we are in the process of procuring a telemetry system to relay the data from the moorings to NOCS. This procurement is currently out for tender through the Research Councils Shared Service Centre, with interested companies due to submit their proposals by 5th July 2010, and on this poster we include details of the initial specification and intended timeline for deployment of the selected system. Data from the project in the form of the MOC component timeseries data and vertical transport profiles are available from the project website at http://www.noc.soton.ac.uk/rapidmoc/.
Peter Rhines1,*, Charlie Eriksen1, Nick Beaird1, Eleanor Frajka-Williams2
1) University of Washington2) NOC, Southampton England.
Between 2003 and 2009 gliders were deployed for the first time to explore high-latitude deep convection, upper-ocean biology, and circulation and mixing in dense overflow plumes near the polar front. We ran Seaglider transects of hydrography, bio-optics, oxygen, depth-averaged velocity and fine-scale vertical velocity in a NOAA-sponsored program in the Labrador Sea and most recently collected 17,800 slant-vertical, full-depth profiles near the polar front on the Iceland-Faroe Ridge and Faroe-Bank Channel. (2006-9, NSF sponsored). These key sites are the ‘headwaters of the AMOC’ (to quote Bob Dickson) where the vertical structure of North Atlantic Deep Water formation depends critically on water-mass transports and transformation at horizontal scales of 10s of km and smaller. This period saw extraordinary warming of the northward flowing warm-water branch of the AMOC (Holliday et al., GRL 2008, Hakkinen & Rhines JGR 2009 et seq.,). This and the altimetrically inferred weakening of the Atlantic subpolar gyre since 1995 coincide with decadal events in the century-long record of Atlantic Meridional Variability; during this time the deep waters began to respond with θ/S shifts, more so than with variability of density or transport. These field studies have shown: (i) the effect of low-salinity Arctic waters spreading over the Labrador Sea, in determining deep convection sites, hence Labrador Sea Water production rates, and in controlling the dominant spring phytoplankton bloom in the region; more generally, the geographic distribution of vertically-integrated buoyancy and its control over convection sites (Hatun et al., JPO. 2007; Frajka-Williams et al. DSR, 2009,2010) (ii) the confluence of the two dominant dense overflows (Faroe-Bank Channel and Iceland-Faroe Ridge) in the eastern subpolar Atlantic and their θ/S/transport evolution in that region (Beaird et al., EOS Trans. AGU 91(26), 2010); (iii) a rich geography of mixing and internal waves in the upper km of the Labrador Sea and deep overflow region at the Iceland-Scotland Ridge seen in vertical velocity reconstructions from the Seaglider slant trajectory. These three observational products are relevant to critical regions in circulation- and climate models, both in air/sea driven convective water-mass transformation, evolution of dense, descending overflows and deep-ocean mixing across θ/S space. Targeted Seaglider sections have begun to combine with satellite altimetry and ARGO float data in high-latitude sites important to the AMOC: boundary currents, overflows and passages connecting Atlantic, Nordic and Arctic seas. Sustained observations combining these techniques are a promising approach to observing future AMOC variability.
Jon I Robson1,*, Rowan T Sutton1, Katja Lohmann2, Doug M Smith3
1) NCAS-Climate, University of Reading2) Max Planck Institute for Meteorology3) Met Office
The variability in the North Atlantic subpolar gyre is primarily driven by the variability in the North Atlantic Oscillation (NAO) through changes in buoyancy forcing and changes in the wind stress applied to the surface of the ocean. In the mid-1990s the subpolar gyre of the North Atlantic underwent a rapid warming, with sea surface temperatures increasing by around 1C in just 2 years. This rapid warming followed a prolonged positive phase of the NAO, but also coincided with an unusually negative NAO in the winter of 95/96. By careful comparison between ocean analyses and idealised ocean model experiments we show that the rapid warming was very likely a delayed response to the prolonged positive phase of the NAO, and not simply an instantaneous response to negative NAO of 95/96. Furthermore, we infer that the warming was likely associated with a surge, and subsequent decline, in the northward heat transport of the Atlantic Ocean. Finally, using initialized predictions made with the UK Met Office Decadal Prediction System, we show that the rapid warming of the subpolar gyre could have been predicted in advance.
Jon I. Robson1,*, Rowan T. Sutton1, Doug M. Smith2
1) NCAS-Climate, University of Reading2) Met Office
In the mid 1990s the North Atlantic subpolar gyre underwent a large and rapid warming that followed a prolonged cooling through the 1980s and early 1990s. Much of the change in the North Atlantic subpolar gyre was driven by changes in the North Atlantic Oscillation (NAO). However, there is some evidence that a lagged strengthening of the Atlantic Meridional Overturning circulation (AMOC) in response to the persistent positive NAO forcing also played a role in warming the subpolar gyre. The UK Met Office has recently developed a Decadal Prediction System (DePreSys; Smith et al, 2007), which assimilates anomalies of temperature and salinity onto a model climatology to avoid climate drift. Hindcast predictions made using DePreSys suggest that the warming of the North Atlantic subpolar gyre could have been predicted in advance, but the skill of predictions is sensitive to the way in which hindcasts are initialised. Through controlled experiments we have shown that predictions of the AMOC and related quantities are sensitive to the model climatology used in the assimilation, and also to density errors that arise from the non-linear equation of state. Our results have important implications for the development of anomaly based decadal prediction systems.
Vassil Roussenov1,*, Richard Williams1, Doug Smith2, Susan Lozier3
1) University of Liverpool, Liverpool, UK2) Met Ofiice, Exeter, UK3) Duke University, Durham, NC, USA
Historical estimates of the overturning have been confined to the analyses along 26 N, which have been extended by the RAPID programme. To complement these historical analyses, a model assessment is performed using historical density data over the North Atlantic basin. Two independent hydrogrphic data sets (WHOI-HYDROBASE and MetOffice) have been used to assess the changes in the temperature, salinity and density over the last 50 years. Both data sets reveal similar gyre-specific patterns: from 1950 to 2000 the subtropical gyre warmed and became more saline, while the subpolar ocean cooled and became less saline with a strong degree of density compensation. Nonetheless, there is an overall decrease in the density of the basin for this fifty-year period. The implied changes in sea surface height reveal a spatially-varying pattern and rates reaching typically 2 mm/year over the last 50 years, which is broadly comparable with inferences from tide gauges. The data are used to initialise the MIT General Circulation model, allowed to dynamically adjust for few years, then enabling the overturning to be diagnosed. This procedure reveals a slight strengthening of the overturning at high latitudes and weakening in subtropics from 1950-1970 to 1980-2000. The uncertainties of these model estimates have been assessed by performing Baysian-type perturbation experiments using the standard errors for the historical data, providing confidence limits for the overturning changes. This model assessment then suggests that the changes in overturning over these two twenty year periods is less then 2 Sv. This analysis of the variability in the property changes and overturning is then extended for other 10 year periods from 1950 to 2006.
Patrick Scholz1,2 *, Lohmann Gerrit1,2
1) Alfred Wegener Institute, Bremerhaven, Germany2) MARUM, University of Bremen, Germany
The climate in the Atlantic region is essentially influenced by the Atlantic meridional overturning circulation (AMOC) which carries warm waters into northern latitudes and returns cold deep water southward across the equator. An important aspect in driving the AMOC is the deep-water mass formation at northern latitudes, but climate scenarios for the future indicate that deep-water formation rate in the North Atlantic could weaken during the 21st century due to global warming. Geological records already indicate that the ocean circulation had almost ceased several times in the geological past due to abrupt changes in the climate. We aim to determine the processes that are responsible for the fluctuations in the deep-water mass formation rates, on interannual to decadal timescales, by using a coupled finite-element sea-ice ocean model. This model has a special focus on the deep-water mass formation areas in the Atlantic (eg., Greenland Sea and Labrador Sea) as well as on areas in the Southern Ocean (eg.,Weddell Sea and Ross Sea). Furthermore, we test the importance of the equatorial and coastal upwelling regions, which also play a major role in driving the large-scale ocean circulation.
Uwe Send1,*, Matthias Lankhorst1., Torsten Kanzow2
1) Scripps Institution of Oceanography2) IfM-Geomar
The Atlantic Meridional Overturning Circulation has been observed continuously since early 2000 by the “MOVE” (Meridional Overturning Variability Experiment) ocean observatory located near 16 N. It uses a scaled-down array which assumes that it is sufficient to observe only the southward (NADW) branch of the MOC, and that most of that flow passes west of the Mid-Atlantic Ridge. Data are available until summer 2009, i.e. for nearly 10 years. The southward volume transport of the overturning circulation is measured in three components: a continental slope contribution observed directly with current meters, an internal geostrophic part obtained from end-point moorings measuring dynamic height, and an external part observed with bottom pressure measurements. The internal plus boundary part contains a significant trend towards reduced circulation over the observing period, with an estimated magnitude of 0.3Sv/yr. Such a trend is expected by climate models in an anthropogenic greenhouse scenario, but may also occur naturally as part of low-frequency variability. The observed results are insensitive to a variety of assumptions, and consistent with knowledge about water messes. In particular the vertical distribution of the trend agrees with water mass signatures, showing a maximum at the level of the Labrador Sea Water CFC maximum. There are also suggestions of changes in the water mass thicknesses/volumes. The observations are intended to provide additional constraints for circulation and climate models, and the variables of choice for assimilation or validation will be discussed.
Amrita Shravat1,*, Simon Tett1, Toby Sherwin2
1) University of Edinburgh2) Scottish Marine Institute (SAMS)
In order to assess what changes have been observed in the AMOC and, more generally, the Atlantic Ocean since 1960 we are using the recently available ocean re-analysis data. Several datasets exist from about 1960 to date. However, rather than naively using them we first evaluate them by comparison with high quality in situ data. We first use the decadal time-series (1995-2005) from Faroe-Bank region to compare ocean re-analyses including the ORA-S3. The metric designed for this comparison is based on the transport of the Atlantic inflow to the Nordic Seas (Holliday et al., 08). The metric is based on velocity transport because most ocean re-analyses do not assimilate velocity. This provides an independent data source. This way of examining the re-analyses with in-situ oceanographic data, determines if the re-analyses generate robust estimates of multi-annual changes and variability in the AMOC.
David Smeed1, Stuart Cunningham1, Paul Wright1*, Lucas Merckelbach1, Julie Collins2
1) National Oceanography Centre, Southampton2) British Oceanographic Data Centre
Since April 2004 the RAPID-MOC mooring array across the tropical North Atlantic Ocean has been monitoring the geostrophic mid-ocean component of the meridional overturning circulation (MOC). The long-term aim of the project is the development of an operational, cost-efficient observation system. Currently the bulk of the mooring array consists of 22 moorings instrumented with SeaBird SBE37 MicroCAT conductivity, depth and temperature (CTD) instruments and a variety of current meters. The moorings are not currently equipped with telemetry and require annual servicing. The eastern boundary of the Atlantic has been found to be important with regards to the seasonal cycle of the MOC. It is also the part of the array that has suffered the highest mooring losses, possibly due to fishing activity. Gliders have the potential to replace the shallower moorings on the eastern boundary where the currents are generally weaker. The principal benefits are that the data are received in near real-time and that any data losses are immediately apparent. A Webb Research Slocum Deep Electric glider was deployed from Gran Canaria from May 2009 to July 2009. The moorings EBH4 and EBH5 are co-located in 1000m of water between the Canary Islands and Morocco. Once the glider reached the mooring site she remained on station acting as a full depth virtual mooring for two months, within a horizontal watch circle of less than four kilometres. During this period the glider sent near real-time data back every eight or nine hours. The moorings were both successfully recovered by the RRS Discovery in November 2009. The sub-sampled data and profiles produced by the glider compared favourably with those from the moorings. The temporal resolution of the glider data was slightly poorer than that of the moored data, although sufficient for removal of tidal effects and calculating the time varying density profiles. The vertical resolution was more detailed with the additional benefit that the freshening and seasonal heating of the upper surface layer above the top MicroCAT was observed.
Vladimir Stepanov*, Keith Haines, Leon Hermanson
University of Reading
Experiments assimilating RAPID-array data (using both 5 day averaged data and 5 day averaged anomalies) into a 1° global NEMO model have been performed. Different methods of assimilating the eastern and western boundary data have been used. Assimilation gives higher values of the MOC transport at 26.5N (by 2-3 Sv), correcting the model’s low MOC bias however the MOC streamfunction undergoes some undesirable changes including intensification of the negative circulation cell near bottom and weakness of the MOC transport at upper levels to the south of the assimilation section near 20N. This is related to assimilation of the deeper density observations. The above experiments used standard covariance functions defined as in Carton et al. (2000) when zonal and meridional length scales depend only on the latitude, to distribute the influence of the boundary data. A new set of experiments focus on tailoring the covariance functions for boundary assimilation. Results are compared with covariance calculated from a climate model HadCM3 and different options for improved assimilation are discussed.
Zoltan Szuts1,*, Jochem Marotzke1
Max-Planck-Institut for Meteorology, Hamburg, Germany
The RAPID array monitors the vertical density structure across the North Atlantic at 26.5N in order to calculate the basin or interior component of the meridional overturning circulation (MOC). By decomposing density fluctuations into vertical modes whose dynamics are well understood, we test whether the theoretical assumptions of vertical modes can explain the observed fluctuations. In addition, this method extracts the first baroclinic mode for comparison with satellite altimetry. We consider two data sets: moored density or geopotential anomalies (GPA) from the RAPID temperature and salinity moorings, and sea surface height (SSH) from satellite altimetry. Because of the project’s sampling strategy of using moored CTDs to measure density profiles, we developed a new decomposition technique to determine modal amplitudes from these observations. Hydrographic records are 2-day low-pass filtered to remove tides, and are further 7-day lowpass filtered when compared with SSH. GPA from the original RAPID processing is called “standard,” while that which is synthesized using the modal decomposition is called “reconstructed”. All 5 full-depth RAPID moorings are considered from the base of the western continental slope (WB2) to the base of the eastern continental slope (EB1). Considered by itself, the modal decomposition accurately recovers most of the GPA variance in the vertical (r2 > 0.9), with slightly less effectiveness at the western boundary (r2 = 0.8). Using only the first baroclinic mode, the reconstructed GPA and its gradient agree closely with the standard GPA and with SSH in the central basin, which demonstrates that almost all of the variance at these locations is contained in the first baroclinic mode. Under these conditions, the first baroclinic mode and SSH contain the same information and both can recover the large fluctuations caused by mid-ocean eddies. In contrast, at the boundaries the signals are much weaker overall, and a significant amount of the variance is contained outside of the first baroclinic mode. Even though the first baroclinic mode extracts the near-boundary transport signal most correlated with SSH, the reconstructed GPA does a poor job of recovering total transport fluctuations adjacent to the boundaries. Given that the interior basin component of the MOC is calculated from boundary-to-boundary density differences, our findings provide an explanation for why satellite altimetry is not able to recover basinwide transports. Furthermore, the failure of simple vertical-mode dynamics at the boundaries suggests that boundary interactions are necessary for explaining the eddy signal contained in the observed MOC and for a clearer understanding of how wave perturbations affect the RAPID observations
Matthew Thomas1,*, Agatha De Boer1, Helen Johnson2, David Stevens1
1) University of East Anglia2) University of Oxford
In the 1940's Sverdrup derived a simple expression for the ocean upper layer transport as a function of the windstress by assuming a linear vorticity balance in the ocean and a level of no motion. This so called Sverdrup Balance has become a cornerstone of oceanography and has formed the basis of many oceanic theories, including theories of how the Meridional Overturning Circulation will adapt to changes in its forcing. It can also be used to derive the upper layer meridional transport remotely from the wind field at any geographical location where the balance holds to first order. Until recently it has not been possible to verify the accuracy of Sverdrup Balance due to the practical difficulties of obtaining observational data with sufficient spatial and temporal resolution. In this study we use the 1° ECCO reanalysis data to quantify the extent to which Sverdrup Balance holds and to determine the dynamical mechanisms responsible for any discrepancies. Averaging the data between 1992 and 2007, we found that, away from western boundaries between 40S and 40N, the deviation from the zonally averaged Sverdrup Balance is between 15% and 40%, but a pointwise balance is far reduced with deviations in many areas exceeding 100%. We argue that these discrepancies are largely due to non-linearity rather than the assumption of a level of no motion. Despite the presence of significant deep transport, past concerns that this flow will render Sverdrup Balance invalid are demonstrated to be unimportant in many areas due to the presence of a mid-water level of no motion. However, a lack of any level of no motion is in places still important, particularly at higher latitudes.
Craig Wallace1,*
National Oceanography Centre, Empress Dock, European Way, Southampton, UK.
Users of climate information, for planning and adaptation purposes, have highlighted precipitation as a key variable on which additional information is sought. Within the scope of the RAPID and RAPID-WATCH programmes data can be utilised to further our understanding of the potential changes to UK precipitation in the low-probability likelihood of a complete collapse of the AMOC this century. Regular hosing style GCM experiments indicate a broad drying pattern for the UK associated, in part, with the reduction of North Atlantic sea surface temperatures. To the end user, information on GCM-scale grid boxes is useful, however, of more use is local-scale information, either at RCM resolution or finer, where there may be deviations to the large scale signal. Here, a statistical methodology is presented to obtain station-level precipitation changes under a collapsed-AMOC scenario for a dense station network within the UK.
Neil Wells1,*, Vladimir Ivchenko1., Joel Hirschi2., Brian King2., Simon Josey2
1) University of Southampton 2) National Oceanography Centre, NSRD
The main goal of MONACO is to understand the links between the meridional overturning circulation (MOC) and the meridional heat transport (MHT) from the RAPID-WATCH observing system, and the subannual to interannual variability of oceanic heat content (OHC) inferred from Argo floats and sea surface temperatures (SSTs) in the North Atlantic inferred from Argo floats. The proposed work addresses question 1 (Qu.1) listed in the RAPID-WATCH announcement of opportunity. We have calculated the MHT variability in the North Atlantic for the 1999 to 2009 period, and the MOC observations are available from April 2004 to April 2008. The largest OHC signal is the seasonal cycle. Its amplitude and phase is similar to the seasonal heat uptake and release through the air-sea heat fluxes. However, North Atlantic MHT changes and air-sea fluxes can differ by more than 1PW (1015W) for periods extending over several months, suggesting that MHT fluctuations may leave a sizable imprint on the OHC. Lag correlations between deseasoned OHC and MOC variability suggest that the MOC is leading North Atlantic OHC changes by about 8 months. Similarly, we find that MOC leads the development of a tripolar SST pattern in the North Atlantic by 6 months. However, since these correlations are based on the short April 2004 and April 2008 period, and it is not clear yet whether the observed MOC – OHC, and MOC - SST links are robust.
Paul D. Williams1,*, Eric Guilyardi1,2, Gurvan Madec2, Silvio Gualdi3, Enrico Scoccimarro3
1) University of Reading, UK2) Universite Paris VI, France3) Istituto Nazionale di Geofisica e Vulcanologia, Italy
On geological time scales, the mean salinity of the world ocean is reported to have exceeded 50psu, because of imbalances between the input and extraction of salt. Such high salinities could have profound consequences for the thermohaline circulation (THC). For example, some authors have suggested that salinization through evaporation and brine rejection, which consume energy and thereby restrict the mean strength of the THC, would no longer be needed for deep water to be formed. Other authors have suggested that impacts could arise because haline forcing, rather than thermal forcing, is the dominant driver of the variability of the THC. Williams et al. (2010) study the sensitivity of the THC to mean ocean salinity, in a coupled atmosphere-ocean general circulation model. We analyse a sensitivity experiment in which the global-mean salinity is approximately doubled from its present observed value, by adding 35psu everywhere in the ocean. We find that the simulated equilibrium THC is surprisingly insensitive to the global-mean salinity increase: the mean strength is reduced by only 20% and the variability is not significantly altered. Our results appear to dispute the claims that higher salinities for the world ocean have profound consequences for the THC.
Tim Woollings1, Jonathan Gregory1
Dept. of Meteorology, University of Reading
There is considerable uncertainty in climate model projections for the North Atlantic storm track and jet stream. Here we show that the varied response of the Meridional Overturning Circulation in different climate models is an important factor contributing to this uncertainty. We demonstrate a clearly significant relationship in that climate models with particularly large reductions in overturning predict a strengthening and extension of the Atlantic jet stream into Europe. This can be explained by an increase in storm track activity arising from increased meridional sea surface temperature gradients, a dynamical signal which is consistent with the behaviour seen in idealised simulations of a weakened MOC. The spread in the MOC response explains about half of the spread in the wind response over Europe, further motivating attempts to narrow the uncertainty in future MOC change.
Hiro Yamazaki1, Kuniko Yamazaki1, Myles Allen1, Tolu Aina1, Milo Thurston1, Nicolas Bellouin2, Peter Hanappe3
1) University of Oxford, UK2) Met Office, UK3) Sony Computer Science Laboratory Paris, France
We have developed a 32-bit precision, Internet-wide distribution package of FAMOUS, a faster, coarser resolution variant of HadCM3, and tested it on a ClimatePrediction.net beta test site. A number of stability issues are identified and rectified in the Fortran source code level as well as tuning the parameters for atmospheric filtering. The current model behaves well under present-day and future conditions with elevated level of CO2 concentration. After an initial transient experiment for the last millennium from 800 to 2000 AD, perturbed physics spin-up experiment was performed under constant forcing condition with an extra diagnostics specifically designed for the RAPID WATCH, vertical cross-section of northward currents along 26.5 degrees north in Atlantic. We also report the initial result of the transient experiment towards 2200AD as well as the development status of HadCM3.
Igor Yashayaev1,*, John Loder1, Sheldon Bacon2, Femke de Jong3, Stephen Dye4, Shane Elipot5, Jürgen Fischer6, Penny Holliday2, Ed Horne1, Chris Hughes5, Dagmar Kieke7, Miguel Maqueda5, Monika Rhein7, Detlef Quadfasel8, Artem Sarafanov9, Hendrik Van Aken3
1) Bedford Institute of Oceanography (BIO), Fisheries and Oceans Canada (DFO), Dartmouth, Canada2) National Oceanography Centre, Southampton, UK3) Royal Netherlands Institute for Sea Research, Texel, Netherlands4) Centre for Environment, Fisheries and Aquaculture Science, Lowestoft, UK5) Proudman Oceanographic Laboratory, National Oceanography Centre, Liverpool, UK6) Leibniz Institute of Marine Sciences, Kiel, Germany7) University of Bremen, Bremen, Germany8) University of Hamburg, Hamburg, Germany9) Institute of Oceanology, Moscow, Russian Federation
The dense water overflows across the Greenland-Scotland Ridge via the Denmark Strait and Faroe-Shetland Channel form the Denmark Strait Overflow Water (DSOW) and Northeast Atlantic Deep Water (NEADW), respectively. Collectively with the convectively-formed Labrador Sea Water (LSW), these water masses form the deep limb of the Atlantic Meridional Overturning Circulation (AMOC) and are important components of the Atlantic and hence global climate systems. Their equatorward flow along the North Atlantic’s western margin is known as the Deep Western Boundary Current (DWBC). Drifter, hydrographic, moored, tracer and satellite data provide information on the fate and integrity of this flow as it moves to lower latitudes, but some important questions remain regarding its variability. Recent variability in the properties of the intermediate and deep water masses between the Iceland Basin and the Scotian Rise off Halifax will be described using observational (hydrographic, moored, profiling float and altimeter) data from several programs. We will show that the variability of intermediate-depth water in the subpolar North Atlantic, and in the associated exiting branch of DWBC, is strongly influenced by the strength and duration of winter convection in the Labrador Sea on the western side, and the advection of warmer more saline intermediate waters from the lower latitudes on the eastern side. In the 2010 annual BIO survey of the Labrador Sea, four variations of LSW produced in different years were identified. While gradually transforming in time, these waters have been preserved in different ranges of density and depth because of gradual weakening of winter convection since 2008, and are still distinguishable by their unique signatures in temperature, salinity and chemical tracers. The fate of each individual LSW class can now be followed by combining profiles from Argo floats and hydrographic data from several institutes. The seawater property anomalies captured by the DWBC are expected to propagate toward the Equator. However, it is only recently that we have been able to actually follow these signals advecting downstream from one moored/hydrographic array to another. We will show how an international array of hydrographic and tracer sections supported by moored, drifter, profiling float and satellite measurements has resolved the downstream propagation of some interesting events formed in the subpolar or Arctic seas. In particular, we are able to document strong fresh and cold anomalies in DSOW, first observed in moored and hydrographic data in the Irminger Sea in 1999, 2004 and 2009, and then with a year delay in the abyssal Labrador Sea. It appears that the 2004 cold fresh anomaly in Denmark Strait overflow water reached the Scotian Rise as a weak signal in 2007-08, suggesting a 3-4 year transit time with considerable water mass modification. We anticipate that the fresh and cold DSOW anomaly that was most recently recorded in the Irminger Sea (spring-fall of 2009) and in the Labrador Sea (spring of 2010), will be arriving on the Halifax Section, maintained by DFO BIO and UK RAPID, sometime between 2012 and 2014
Laure Zanna1,*, Eli Tziperman2, Patrick Heimbach3, Andrew Moore4
1) University of Oxford, AOPP2) Harvard University, EPS3) MIT, EAPS4) UCSC, Ocean Sciences
The variability and the limits on predictability of the Atlantic meridional overturning circulation (MOC) and sea surface temperature(SST) due to surface forcing are examined in an ocean general circulation model. Significant transient amplifications of MOC and SST anomalies occur due to the non-normal dynamics on timescales of 19 and 15 years respectively. The mechanisms responsible for the growth of anomalies is analyzed. Moreover, the growth of SST and MOC anomalies is slower and weaker when only the upper ocean is excited compared to an excitation at depth, leading to the conclusion that predictability experiments perturbing only the atmospheric initial state (equivalent to perturbing the upper ocean only) may overestimate the ocean predictability time.
Xiaoming Zhai1,*, Helen Johnson1, David Marshall1
We investigate heat content changes in the Atlantic in response to thermohaline forcing using MITgcm forced by switching northern relaxation boundary conditions. When the meridional overturning circulation is weakened at the northern boundary, the thermocline at lower latitudes deepens, causing warming of the upper ocean. Following Johnson and Marshall (2002), analytical solutions are sought for ocean heat content changes in response to thermohaline forcing at northern latitudes.
