Magdalena Balmaseda1, Gregory Smith2, Keith Haines2, David Anderson1, Tim Palmer1, Arthur Vidard3
1) ECMWF2) ESSC, Univ. Reading3) INRIA
The Atlantic meridional overturning circulation is composed of a warm near-surface branch flowing northward as part of the Gulf Stream and a return flow of cold waters at depth. Numerous modelling studies have highlighted the importance of this circulation as a regulating mechanism of the climate system, and its variability plays an important role in future climate change scenarios. However, knowledge of this variability is sorely lacking. Here we provide a continuous historical reconstruction of the meridional overturning circulation over the period 1959 - 2006 by combining an ocean model with a complete set of in situ and satellite observations from the historical database record. We find that the model agreement with previous estimates of MOC intensity, based on 5 snapshot ship cruises, is good once observations are synthesised. Results show recent changes in the vertical structure of the circulation with a shallower recirculation cell within the thermocline and a decrease of the deep ocean circulation. A significant downward trend of about 20% in the last 47 years is found, with an equivalent reduction in the northward heat transport. This trend is smaller than previously thought, although the changes in vertical structure are in good agreement with previous results.
Rory Bingham, Chris Hughes1, Vassil Roussenov2, Ric Williams2
1) Proudman Oceanographic Laboratory2) University of Liverpool
The North Atlantic Meridional Overturning Circulation (MOC) is associated with deep water formation at high latitudes, and climatically-important ocean-atmosphere heat fluxes. Results from highly idealised model studies suggest that the North Atlantic responds rapidly to changes in the deep water formation rate, implying that even on short timescales variations in the MOC in the North Atlantic are meridionally coherent. In this paper, we investigate the coherence of MOC changes in two eddy-permitting models with realistic forcing. We find that for periods greater than a year MOC variability to the north of 40N is dominated by a local overturning cell that varies on a decadal timescale. To the south of 40N MOC variability is dominated by a mode that has much more power at interannual timescales. This mode has its maximum amplitude at low to mid latitudes but does also extent to higher latitudes. The ocean dynamics giving rise to these modes of variability and the implications for monitoring the MOC are discussed.
Adam Blaker, Bablu Sinha
National Oceanography Centre, Southampton, UK
The FORTE (Fast Ocean Rapid Troposphere Experiment) model comprises of a T42 resolution IGCM3 atmosphere coupled to a 2 x 2 degree MOMA ocean. Under the UK THCMIP, FORTE will complete a 200 year control integration. After 60 years of integration two 140 year perturbation integrations will begin: a 1% transient co2 forcing run, which reaches 4x CO2 after 140 years; and a fresh water hosing run, in which a 0.1 Sv flux of fresh water will be evenly distributed over the North Atlantic between 50 and 70 degrees north. A further two 140 year partially coupled integrations will examine the relative importance of the heat and fresh water fluxes into the ocean during the perturbation runs. An update on the status of these integrations and analysis of the relationship between the meridional overturning circulation and the steric height gradient in FORTE is presented.
David Brayshaw1, Brian Hoskins2, Mike Blackburn1
1) Walker Institute for Climate Research, Department of Meteorlogy, University of Reading2) Department of Meteorlogy, University of Reading
The North Atlantic storm track dominates the weather and climate of north western Europe and is strongly influenced by the surface properties of the Atlantic ocean. In the context of recent climate simulations that suggest a weakening of the oceanic thermohaline circulation, it is crucial for confident predictions of climate change in the European region to understand how changing sea surface temperature patterns would affect the storm track. This paper adopts a structured and comprehensive approach to this issue by using the HadAM3 general circulation model to perform a series of simulations with both idealised and semi-realistic boundary conditions. This hierarchy enables the behaviour of the more complex simulations to be interpreted in terms of simpler configurations. In this manner, the effects of the sea surface temperature anomalies associated with the Gulf Stream and the North Atlantic Drift are investigated along with the non-linear interactions between them and the "background" flow created by features such as the Rocky Mountains. The storm track is found to be highly sensitive to changes in the midlatitude SST gradients on the poleward side of the subtropical jet, with the Gulf Stream enhancing and the North Atlantic Drift suppressing storm activity both locally and downstream over Europe. Interactions with the flow patterns created by the North American coastline and the Rocky Mountains are also shown to further intensify and localise the storm track. The physical mechanisms responsible for generating the signals are discussed and implications for climate scenarios featuring a weakening of the THC are drawn.
David Brayshaw1, Tim Woollings1, Michael Vellinga2
1) Dept. of Meteorology, University of Reading2) Hadley Centre, Met Office
Much of the European impact of a shutdown in the North Atlantic's meridional overturning circulation will be mediated though changes in atmosphere, particularly in the storm tracks and the North Atlantic Oscillation (NAO). Here we use a variety of techniques to analyse storm data from a simulation of the coupled model HadCM3 which is forced to maintain an equilibrium state without an MOC. When compared with a control run, the simulation shows dramatic differences, exhibiting a stronger Atlantic storm track extending deeper into Northern Europe. The differences are interpreted by comparison with more idealised model simulations. There is also an eastward shift of the NAO pattern and a slight tendency for the positive phase of the NAO to be realised more often.
Peter Challenor
National Oceanography Centre, Southampton
The historical observations of the strength of the overturning circulation at 26°N presented in the Bryden at al (2005) paper have been the subject of much discussion. A lot of this has centred on how much information can be extracted from only five numbers. Is there a real trend in the data or are we only seeing the effect of aliasing of the short term variability? Although there has been a lot of discussion there has been little statistical analysis of these data. In this paper we will present a number of statistical tests of whether we are really seeing a signal. We will show that it is possible to answer this question with only five data points and discussion the implications of our results.
María Paz Chidichimo1, Torsten Kanzow2, Jochem Marotzke1, Stuart A. Cunningham2
1) Max Planck Institute for Meteorology, Bundesstrasse 53, 20146 Hamburg, Germany.2) National Oceanography Centre, Empress Dock, Southampton, SO143ZH, UK.
The aim of this work is to study the contribution of eastern boundary density to the MOC, based on the RAPID 26.5°N mooring data between April 2004 and October 2006. It is generally expected that the density variability near the eastern boundary of the North Atlantic is smaller than near the western boundary. However, neither the amplitude nor the frequency distribution of eastern boundary density contributions to MOC variability have been studied systematically. Here we show that they are important. The RAPID array makes use of moored time series measurements of density at the western and eastern boundaries of the North Atlantic to estimate the basin scale, absolute zonally integrated mid-ocean geostrophic transport. To highlight the eastern boundary variability, calculations are made assuming that density at the western boundary is constant and only the eastern boundary density varies over time. At the eastern boundary there are two methods of sampling density profiles: using either of two 5000 m long (full water column) moorings located at the base of the African continental slope, or with an array of small moorings distributed between the African shelf and the base of the continental slope. Using the tall moorings, the largest variability is found in the thermocline layer (surface to 800 db, standard deviation of 1.1 Sv) and in lower North Atlantic Deep Water (below 3000 db, standard deviation of 1.34 Sv). The dominating vertical velocity mode exhibits large amplitudes in these layers with a zero crossing at ~ 1770 db. Moreover, flows in these two layers compensate each other. The eastern boundary contribution to the basin-scale velocity inferred from the inshore (small moorings) and offshore (tall moorings) data sets are investigated for potential redundancy. Preliminary results suggest that there are considerable differences between the two data sets in terms of amplitude, vertical structure and frequency distribution of the resulting mid ocean geostrophic transport fluctuations. The spectrum based on the inshore data shows much more variance in the high frequency limit. Thus, near boundary processes appear to play an important role. The analysis of the dominating vertical velocity modes associated with the inshore data set shows that variability is more concentrated in the upper ocean, and is very small in the deep ocean. Contribution of eastern boundary density variability to MOC variability is 1.9 Sv if the offshore data are used, and 1.5 Sv if the inshore data are used. The different sources of the eastern boundary transport variability as well as the associated processes will be addressed.
Stuart A. Cunningham1, Kanzow Torsten O.1, Rayner Darren1, Baringer Molly O.3, Johns William E.2, Marotzke Jochem Y.4, Hirschi Joël J-M1, Beal Lisa M.2, Meinen Christopher S.3, Harry L. Bryden1, Hannah R. Longworth1, Elizabeth M. Grant1
1) National Oceanography Centre, Southampton2) University of Miami, Rosenstiel School of Marine and Atmospheric Sciences3) NOAA-Atlantic Oceanographic and Meteorological Laboratory4) Max Planck Institute for Meteorology
Objective 1 of the RAPID programme is “to establish a pre-operational prototype system to continuously observe the strength and structure of the Atlantic meridional overturning circulation (MOC)”. The MOC is commonly defined as the zonally integrated meridional flow, as a function of latitude and depth. While parts of the MOC are wind-driven, the basin-scale Atlantic MOC is largely buoyancy-forced. Hence, observing the Atlantic MOC is the fundamental observational requirement of a programme aiming to assess the role of the Atlantic thermohaline circulation (THC) in rapid climate change. The RAPID-MOC array measures: the Gulf Stream transport through Florida Strait by cable and repeat direct velocity measurements; Ekman transports by satellite scatterometer and from climatologies; shallow and deep western boundary currents by direct velocity measurements and the basin wide interior baroclinic circulation from moorings measuring vertical density at the boundaries and; barotropic fluctuations are measured by an array of bottom pressure recorders. The array became operational in late March 2004. Here we describe results for a 368-day period from 29 March 2004 to 31 March 2005. We present the first direct observational evidence that the zonally integrated meridional flow tends to conserve mass, with the fluctuations of the different transport components largely compensating at periods longer than 10 days. We take this as experimental confirmation for the design of the array. This enables us to describe, for the first time, the intra-annual transport and variability of the MOC. We show evidence that the near surface Ekman transport is compensated by barotropic fluctuations and that density induced transports contribute as much to the overall MOC variability as the Ekman component. Mid-ocean geostrophic transports are estimated from the first year’s time series using daily eastern and western boundary dynamic height profiles with the reference level velocity chosen so that the mid-ocean geostrophic transport balances the Florida Straits plus Ekman transports on a daily basis. The annual mean MOC and standard deviation is 19.1±5.6 Sv. The Florida Straits, Ekman and mid-ocean upper ocean transports are not significantly correlated and contribute about equally to the temporal variability in the MOC. There are around 15 independent values in the year-long MOC timeseries, so if future years’ timeseries exhibit similar variability, we can define inter-annual variability in the overturning with a resolution of about 1.5 Sv.
Ruth Curry
Woods Hole Oceanographic Institution
Over the last fifty years, the subpolar basins and Nordic Seas exhibited a full cycle of changes in heat and freshwater content organized, to first-order, around the structure of the NAO/AO forcing. Although the freshening there between the 1960s and 1990s was perceived to be extraordinary, it was comparable in timing and amplitude to the natural variability of freshwater storage diagnosed from control runs of coupled climate models. The observed 50-year fluctuation of freshwater content can largely be accounted for as a combination of local storm track / precipitation anomalies, variability in moisture flux convergence onto the Arctic Ocean and its watersheds, and episodic accumulation and release of sea ice and freshwater from the Arctic governed by the regional atmospheric pressure patterns. Subpolar heat and freshwater content anomalies fluctuated in synchrony and were generally density-compensating. By contrast, the Nordic Seas heat and freshwater storage records diverged in the late 1980s resulting in reductions of density in the waters beneath the sill depths of the overflows. The vertical redistribution of temperature and freshwater anomalies, however, inhibited significant alterations of the density contrast that sustains the Atlantic MOC strength. Little, if any, significant change in the AMOC is evident in the modern oceanic instrumental record.
Lars Czeschel1, David P. Marshall1, Helen L. Johnson2
1) AOPP, University of Oxford2) Department of Meteorology, University of Reading
The MIT adjoint model in an 1 degree global configuration is used to study the variability of the meridional overturning circulation (MOC). The adjoint model allows to calculate the sensitivity of the MOC against several model inputs (e.g. initial conditions and atmospheric forcing) in a single model run. These adjoint sensitivities are used to identify how hydrographic and dynamic anomalies along the western boundary and the interior, and thus the MOC, are sensitive to local and remote forcing. We concentrate here on the sensitivities against winter-time buoyancy fluxes in the subpolar gyre. Our study suggests that the buoyancy forced variability of the mid-latitude MOC at a particular time is the sum of at least the last ~40 years of winter-time buoyancy forcing in the subpolar gyre. The variability is characterized by a damped basin mode. Positive heat loss anomalies, for example during NAO+ phases, are leading to an increased MOC for the next ~12 years but to a decreased MOC for the following ~10 years and so on.
Elizabeth J. Farmer, Julian E. Andrews, Mark R. Chapman
School of Environmental Sciences, University of East Anglia, Earlham Road, Norwich, Norfolk, NR4 7TJ
The temperature dependent nature of the ratio of magnesium to calcium (Mg/Ca) uptake in foraminiferal calcite has lead to the development and increasing use of this technique for reconstructing past seawater temperatures for both surface (planktonic) and deep (benthic) water environments. A high resolution record of Mg/Ca ratios from the planktonic foraminifera G. bulloides has been produced for the Holocene sequence of the North Atlantic core MD99-2251. The record presented here extends to Termination I at an approximate 70 year resolution, with a higher resolution (~20 year) section covering the 8.2 kyr cooling event. This event has been the focus of much recent debate, appearing as the most pronounced climatic event of the Holocene in a number of palaeoenvironmental records, and is thought to be related to catastrophic outburst events from proglacial meltwater lakes Agassiz and Ojibway. The Mg/Ca data show a muted Holocene response to the climatic shifts previously illustrated by foraminiferal assemblage counts and isotope data from the same core. Reconstructed temperatures using G. bulloides Mg/Ca ratios indicate Holocene variability of 1-3 ºC, suggesting that the species may be complicated by ecological issues which affect its capacity to record a single climatological variable, such as seasonal or monthly sea surface temperatures. This record complements a suite of other proxy data from the same core (Ellison et al. 2006) and potentially highlights some interesting issues relating to the recording of rapid climate change with respect to the Mg/Ca method and the choice of species analysed.
Philip Goodwin1, Ric Williams1, Andy Ridgwell2, Mick Follows3
1) University of Liverpool, UK2) University of Bristol, UK3) MIT, USA
Whilst there is widespread concern about the immediate climate response of raised atmospheric carbon dioxide (CO2) levels, there is also a long term response lasting many thousands of years. Current fossil fuel emissions are raising the partial pressure of atmospheric CO2 (PCO2), and cause an uptake of CO2 by the ocean. Once emissions cease and air-sea equilibrium has been reached the radiative forcing caused by carbon emission is varied by how much ocean uptake is achieved, itself a product of the atmosphere and ocean buffered carbon inventory. We analyse the atmosphere and ocean buffered carbon inventory for the present, past and future and find that at present the radiative forcing change with carbon emission is currently 1.5 to 1.7 W m-2 1000GtC-1 at air-sea equilibrium. This present day level of radiative forcing per unit carbon emission is higher than for the majority of the past 400 million years and for future scenarios. We conclude that the present day is a bad time to perturb the carbon system, since an unusually large radiative forcing is produced for each unit of carbon emitted.
Jeremy P. Grist, Simon A. Josey, Bablu Sinha, Adam T. Blaker
The relationship between Denmark Strait (DS) dense water transport and Nordic Seas heat flux is examined within the control runs of two IPCC Class Models: HadGEM1 and HadCM3, and also with prescribed perturbation experiments in the FORTE coupled climate model. Experiments with FORTE allow the role of heat flux anomalies to be isolated from other processes. The signature of winter surface cooling is vertical mixing leading to a negative heat anomaly between 600 and 800m. The anomaly propagates southward, descending to a depth of 2000m south of the DS sill. In all the models, DS overflow increases approximately linearly with Nordic Seas heat loss, the strongest decrease in DS temperature occurring within 8-12 months of the start of a winter of strong heat loss. The robustness of the results suggests that Nordic Sea heat flux variability is a key factor in determining DS overflow characteristics that must be considered in addition to other processes, in particular the impact of wind driven circulation anomalies.
Jeremy P. Grist, Simon A. Josey, Bablu Sinha
The ocean response to air-sea flux variability in the Greenland Sea is investigated using the 1000 year control run of the coupled ocean-atmosphere model, HadCM3. Evaluation of the density flux reveals that net heat flux anomalies have a greater impact on surface density changes than anomalies in both net evaporation and ice melt / formation. Averaged over the Greenland Sea, the annual mean density flux due to heat loss is 1.8 x 10^(-6) kg m^(-2) s^(-1), over an order of magnitude greater than the net evaporation and the ice melt / formation terms, which are –0.1 and –0.2 x 10^(-6) kg m^(-2) s^(-1) respectively. Extreme winter heat loss events reach 250 Wm^(-2) and are associated with reduced ice cover and anomalously strong northerly airflow over the Greenland Sea. They result in enhanced convection and modify the properties of deep water flowing south through the Denmark Strait. The deep water transport increases by about 30% when the strongest and weakest heat loss events are compared. We also find significant correlations between deep western basin temperatures at 60, 55 and 49N and the Greenland Sea heat flux anomalies which peak at lags of up to 4 years with the time delay increasing towards more southerly latitudes. Our results suggest that extreme Greenland Sea heat flux events are key to understanding recent observations of significant interannual variability in Denmark Strait transport characteristics.
Rachel Hadfield1, Neil Wells2, Simon Josey2
1) BMT Cordah Limited2) National Oceanography Centre
Since 1999, Argo floats have collected almost 50,000 temperature profiles in the North Atlantic Ocean. The accuracy with which this dataset can be used to estimate the upper ocean temperature and heat storage in the North Atlantic has been investigated. A hydrographic section across 36°N was used to assess uncertainty in the temperature field. The RMS difference in the Argo based temperature field relative to the section measurements is about 0.6 °C. In comparison, the difference of the section with respect to the World Ocean Atlas (WOA) is 0.8 °C. For the upper 100 m, the improvement with Argo is more dramatic, the RMS difference being 0.56 C, compared to 1.13 C with the WOA. The Ocean Circulation and Climate Advanced Model (OCCAM) was used to determine the Argo sampling error in mixed layer heat storage estimates. Using OCCAM subsampled to typical Argo sampling density, it is found that outside of the western boundary, the mixed layer monthly heat storage in the subtropical North Atlantic has a sampling error of 10-20 Wm-2 when averaged over a 10°x10° area. This error reduces to less than 10 Wm-2 when seasonal heat storage is considered. Further results will be presented that demonstrate closure of the heat budget to within 10 Wm-2 in the central and eastern subtropical ocean.
Ed Hawkins, Rowan Sutton
Walker Institute, University of Reading
We describe the use of 3d EOFs in characterising multi-decadal variability of the THC in HadCM3. We find that the leading two modes are well correlated with an index of the MOC on decadal timescales, with the leading mode alone accounting for 54% of the decadal variance. This approach retains only a few degrees of freedom but provides a better representation of the complex three-dimensional thermohaline circulation dynamics than a 2d streamfunction. Episodes of coherent oscillations in the sub-space of the leading EOFs are identified; these episodes are of great interest for the predictability of the THC, and could indicate the existence of different regimes of natural variability. The mechanism identified for the multi-decadal variability is an internal ocean mode, dominated by changes in convection in the Nordic Seas, which lead the changes in the MOC by a few years. Variations in salinity transports from the Arctic and from the North Atlantic are the main feedbacks which control the oscillation. This mode has a weak feedback onto the atmosphere and hence a surface climatic influence. Interestingly, some of these climate impacts lead the changes in the overturning. This is at odds with the traditional view that any climatic effects tend to follow a change in the MOC. There are also similarities between these modelled climate signatures and observed multi-decadal climate variability. Further analysis has shown that the largest, and most rapid, MOC changes are preceded by significant changes in sea surface properties in the Nordic Seas. If this predictability was present in the real world then it would indicate a need for more observations in the Nordic Seas to help monitor the state of the overturning circulation. We are continuing to explore the analysis of THC predictability using our EOF basis with the aim of estimating empirical singular vectors for the THC and will report our latest findings.
Ross Hendry, Kumiko Azetsu-Scott, John Loder, John Smith, Igor Yashayaev
Fisheries and Oceans Canada, Bedford Institute of Oceanography, Dartmouth, NS, Canada
Since 1990 annual repeat interdisciplinary measurements in spring or early summer have been made along the WOCE AR7W Labrador Sea section between southern Labrador and southern Greenland. This continues a historical connection between the Bedford Institute of Oceanography and the Labrador Sea stretching back to 1965–1967 when CSS Hudson and CCGS Labrador mapped hydrographic properties from Davis Strait through the Labrador and Irminger Seas. The physics goals of the modern AR7W surveys include monitoring interannual variations in the intensity of winter convection and associated property changes in intermediate-depth Labrador Sea Water and monitoring interannual changes in North Atlantic Deep Water properties in the deep boundary currents linked to the large-scale North Atlantic thermohaline circulation. The Halifax Section on the Scotian Shelf starts off Halifax, Nova Scotia and progresses southeastward to approximately the 2500 m isobath on the continental slope. Regular occupations of this line began in 1950 and continued irregularly until the late 1970s. Sampling resumed in the early-to-mid 1990s and is ongoing as part of the Atlantic Zone Monitoring Program (AZMP) formally established in 1998. The primary AZMP mandate is to collect and analyze biological, chemical, and physical data to characterize and understand the causes of oceanic variability in the Canadian NW Atlantic shelf region at seasonal, interannual, and decadal time scales. Labrador Sea Water is exported to the south into the Slope Water between the eastern Canadian continental slope and the Gulf Stream. In this sense, the Slope Water is a westward extension of the subpolar gyre. Changes in the properties of the Slope Water can cause corresponding changes in the waters over the shelf and influence the regional ecosystems. The recently-established Atlantic Zone Offshore Monitoring Program combines our Labrador Sea monitoring work and a new initiative to extend the Halifax Line offshore to monitor interannual variability in the circulation and water properties over the deeper continental slope. This coordinated effort seeks to better understand the links between the Labrador Sea and the Slope Water in the intermediate density range associated with Labrador Sea Water formation. It also offers the prospect of monitoring spatially coordinated interannual changes in the Deep Western Boundary Current in the North West North Atlantic.
Chris W. Hughes1, Rory J. Bingham1, Miguel Angel Morales Maqueda1, Richard G. Williams2
We set out the theory behind the idea of using western boundary pressure measurements to monitor the MOC, and demonstrate that the method works well in the context of the 0.25 degree OCCAM model. Although pressure measurements are in principle all that is needed, in practice it is currently impossible to maintain an accurate multiyear timeseries of bottom pressure from direct measurements. We show how, instead, a combination of boundary measurements of current and density can be used to calculate the needed pressure field, and illustrate the principle with reference to data recently collected from the WAVE experiment.
Vladimir Ivchenko1, Sergey Danilov1, Dmitry Sidorenko1, Jens Schroeter1, Manfred Wenzel1, Dmitry Aleynik2
1) Alfred Wegener Institute for Polar and Marine Research2) University of Plymouth
Anomalies of dynamic height derived from an analysis of Argo profiling buoys data are analysed to assess the relative roles of contributions from temperature and salinity over the North Atlantic for the period of 1999-2004. They are compared with dynamic topography anomalies based on TOPEX/Poseidon and Jason altimetry. It is shown that the halosteric contribution to the anomalies of dynamic height is comparable in magnitude to the thermosteric one over the period analyzed. Taking both salinity and temperature into account improves the agreement between zonally averaged trends in the satellite dynamic topography and dynamic height increasing the correlation between them to 0.73 from 0.63 when only temperature variability is taken into account. The implication of this result is that the salinity contribution cannot be neglected in the North Atlantic and that one cannot rely on estimating the thermosteric part by anomalies in the sea surface dynamic topography derived from the satellite altimetry.
Bill Johns1, Lisa Beal1, Molly Baringer2, Jonathan Molina1, Stuart Cunningham3, Torsten Kanzow3
1) Division of Meteorology and Physical Oceanography, Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL2) Physical Oceanography Division, NOAA Atlantic Oceanographic and Meteorological, Miami, FL3) Ocean Observing and Climate Research Group, National Oceanography Centre, Southampton, UK
The measurements presented here are part of a UK-US collaborative effort to measure the strength and variability of the meridional overturning circulation (MOC) in the mid-latitude North Atlantic. Transport variability of the northward flowing Antilles Current within the upper 1000 m, and of the southward-flowing Deep Western Boundary Current (DWBC) below, are important components of MOC variability. These currents are measured using an array of five current meters out to 47 km offshore, the historic mean position of the DWBC core, and from 47 to 500 km offshore using dynamic height moorings, which measure the spatially-integrated geostrophic flow. The Antilles Current is a thermocline-intensified flow with a mean transport of 6 Sv and a variability of -15 to +25 Sv. Its variability is dominated by westward propagating mesoscale eddies with time scales of a few months. The DWBC, including offshore recirculations, has a mean transport of -26.5 Sv, which is almost equally divided between Upper and Lower North Atlantic Deep Water (NADW). For the first time a persistent and intense upper core was found in the DWBC close to the coast, above the Bahamas Escarpment, near 1200 m depth. The mean structure from the array compares well with the mean of three LADCP sections collected in 2004/2005, which show a DWBC with an offshore extent of between 80 and 100 km and anything from two to four recirculation cells, or eddies, out to 500 km offshore. The transport variability of the DWBC is between -60 and +3 Sv and is dominated by barotropic fluctuations (probably barotropic Rossby Waves) with time scales of a few days to a few weeks and zonal scales of about 2000 km. Upon removal of these barotropic fluctuations, slower baroclinic variability is revealed, highlighting a period between October and December 2004 when the transport of Lower NADW was considerably reduced (including a temporary stoppage) due to a downward shift in isopycnals.
Lukas Jonkers1, Maarten Prins2, Gert Jan Weltje3, Andrey Ganopolski4, Simon Troelstra2, Noortje Dijkstra2, Geert-Jan Brummer1
1) Royal Netherlands Institute for Sea Research. Dept. Marine Chemistry and Geology. PO Box 59, 1790 AB Den Burg, Texel, The Netherlands2) Vrije Universiteit Amsterdam, Dept. Paleoclimatology & Geomorphology. De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands3) Delft University of Technology, Department of Applied Earth Sciences, PO Box 5028, 2600 GA Delft, The Netherlands4) Potsdam Institute for Climate Impact Research, Telegrafenberg A 31, 14473 Potsdam, Germany
We present grain-size based inferences of variations in ice-rafting and ocean-circulation during the last glacial based on a core from the North-Atlantic. An inverse modelling technique was used to quantify the relative contributions of ice-rafted and current-sorted sediments, of which the proportion and mean grain-size were used to infer freshwater forcing and Atlantic Meridional Overturning Circulation (AMOC) strength, respectively. The record shows that the AMOC slowed down in response to glacial freshwater forcing events, suggesting a strong link between freshwater forcing and AMOC strength. In addition our record indicates consistently that the ocean response lagged the applied freshwater forcing. We have critically evaluated the conversion of the static proxy data to ocean-circulation dynamics, facilitating comparison of the inferred AMOC response with model simulations. The inferences show remarkable similarities with model simulations and may suggest non-linearity in the AMOC response to freshwater forcing. Oxygen isotope curves of N. pachyderma s. for three ice-rafting events reveal a similar sequence. They suggest an important role for sea-ice, both as a carrier of ice-rafted sediment as well as a source of freshwater. Delayed response of the AMOC to freshwater forcing is a robust feature of many climate models and our data provide the first empirical evidence that indeed such delay was a real feature of the glacial AMOC. Such information on the response of the AMOC further constrains the effects of glacial melting during modern climate warming.
Torsten Kanzow1, Stuart Cunningham1, Joel Hirschi1, Darren Rayner1, Jochem Marotzke2, Christopher Meinen3, William Johns4, Molly Baringer3, Harry Bryden1, Lisa Beal4
1) NOC, Southampton, UK2) MPI, Hamburg, Germany3) AOML/NOAA, Miami, USA4) RSMAS, Miami, USA
The Atlantic meridional overturning circulation (MOC) carries roughly one Petawatt of heat northward through the subtropical North Atlantic. For the first time a continuous one-year-long time series of the meridional overturning volume transports across 25°N in the Atlantic has been computed by combining five components in the framework of the RAPID-MOC experiment. The Gulf Stream and the zonally integrated Ekman transport are estimated from cable measurements across Florida Strait and space-borne scatterometry, respectively. The three components that combine to give the mid-ocean meridional transports (integrated across the six thousand kilometre wide section below the surface Ekman layer) are derived from eastern to western boundary differences in density and bottom pressure, complemented by direct current measurements. The present study represents a first attempt to highlight the different processes that contribute to the temporal evolution of the MOC. We find that density near the western boundary contributes significantly more to the MOC variability than eastern boundary density, however, the latter contribution is not negligible. Densities measured at both sides of the Mid-Atlantic Ridge have a significant effect on the mean MOC transport profile, however appear to contribute an order of magnitude less to the variability. We investigate the vertical structure of the different MOC contributions as well as their frequency distribution. Our results suggest that pronounced oscillations at periods shorter than 10 days are may be linked global atmospheric pressure forcing whereas variability found at periods near 20 days point to phenomena dynamically trapped at western boundary. Indications for possibly wind-driven seasonal variability has also been observed near the western boundary both in the Straits of Florida and east of the Bahamas. We conclude that understanding the dynamics and spectral distribution of MOC fluctuations is essential for the detection of long-term changes.
Klaus Keller, David McInerney
Department of Geosciences, Penn State
Model simulations as well as the geological record suggest that the Earth System can change abruptly and with little warning time between different basins of attraction in response to small and smooth forcing. One example of such a climate threshold response is a possible shutdown of the North Atlantic meridional overturning circulation (MOC) due to anthropogenic climate forcing. An MOC shutdown has often been characterized as a low-probability high-impact event, but these projections are deeply uncertain. Here we analyze three main questions: (i) What is the ability of proposed MOC observation systems to deliver an actionable early warning sign of a future MOC threshold response? (ii) How could MOC observation systems be improved to better inform the decision-making process? (iii) What are economically efficient risk management strategies in the face of deep uncertainty and potential learning about climate thresholds? We show that the currently implemented MOC observation system may well succeed in reliably detecting statistically significant MOC trends on a time scale of decades. However, predicting future changes (not just the detection of past changes) is required for many decision-making frameworks. Determining whether an observed MOC trend differs in a statistically significant way from an unforced scenario (the detection problem) can impose lower requirements on an observation system than the determination whether the MOC will shut down in the future (the prediction problem). As a result, observation systems designed for early detection of MOC changes might fail at the task of early and confident prediction. We demonstrate how the analysis of additional tracer observations (e.g., oxygen) and the optimization of the spatial design of the observation system can improved the signal-to-noise ratio and hence the detection and prediction capabilities. We use an integrated assessment model of climate change to analyze how potential MOC observation systems may inform the design of climate risk management strategies. Considerably reducing the odds for a future MOC collapse would require a sizable decarbonization of the economy within a few decades in this simple framework. Reducing the current uncertainty about the MOC response can have a large economic value, as it would allow for the design of more efficient risk management strategies. The economic value of information associated with a confident and early prediction of an MOC threshold response could exceed the costs of typically implemented ocean observation systems by orders of magnitude. Identifying a feasible observation and prediction system that would enable a confident and early MOC prediction across the range of relevant uncertainties is, so far, an open challenge.
Thomas Kleinen, Tim Osborn, Keith Briffa
Climatic Research Unit, School of Environmental Sciences, University of East Anglia, Norwich, UK
We are investigating the climate of the Little Ice Age. While cooling as a simple, direct response to the reduced insolation due to solar minima, as well as high concentrations of volcanic aerosols, may explain part of the changes in climate during the Little Ice Age, it is possible that dynamic changes in the climate system also play an important role. Dynamical changes in the atmosphere or ocean may themselves be part of the response to solar and volcanic forcing, but this is currently unclear because many climate models simulate only a relative weak dynamical response to changes in external forcings. In order to explore the potential causes of the Little Ice Age climate anomaly, we are conducting sensitivity experiments with HadCM3 to investigate the consequences of three candidate mechanisms for North Atlantic climate, and different combinations of these three. The candidate mechanisms are: (i) a reduction in North Atlantic MOC, (ii) changes in North Atlantic atmospheric circulation to negative NAO states, and (iii) reduced radiative forcing. We will present results from experiments involving these mechanisms. We will present the simulated consequences of a 25% reduction in North Atlantic MOC for climate in the North Atlantic area. In addition we will present the methodology of pattern nudging and its implementation in HadCM3. We are using this method to nudge the HadCM3 atmospheric circulation towards a state with weaker westerly winds across the North Atlantic (i.e., a negative NAO state) during winter, without suppressing the synoptic-scale variability such as storm tracks. We will present results from nudged HadCM3 simulations for circum North Atlantic climate. Finally we will show the consequences of reduced radiative forcing, as well as of forcing combinations.
Daniel Lea1, Keith Haines1, Greg Smith1, Matt Martin2
1) ESSC, Univ. Reading2) UK Met Office
We modify the UK Met Office FOAM (NEMO) OI ocean data assimilation system to assimilate profile data (including Argo) on density levels. The model background field is converted to spiciness on a set of specified density levels and the depths of those density levels are calculated. The profile observations are also converted to spiciness and density level depth for the same set of specified density levels. We perform two analyses, first of spiciness on density levels with the data being spread over large spatial scales, and second of density level depth with smaller spatial scales. The order of the analyses is not important. The resulting increment fields are converted back to temperature and salinity on model z-levels. The results of this are compared to the standard FOAM assimilation on z-levels over a 1 year period assimilating the same observations. The advantage of spreading the information along density levels isillustrated by comparing the results to withheld data.
Richard Levine, Grant Bigg
Department of Geography, University of Sheffield
The impact of the disintegration of European Ice Sheets is less well known than for the North American Ice Sheets, while the meltwater and iceberg release due to European events is likely to be upstream of the deep convection regions of the North Atlantic. In order to simulate these events we have coupled a dynamic and thermodynamic iceberg model to a global atmosphere-ocean model for the first time. We are able to produce steady states in the model that are consistent with a steady climate at the LGM. We use this state as a base to additionally impose Heinrich event-scale iceberg fluxes, sub-glacial lake floods, and North Sea ice break up events that are thought to have occurred from the European Ice Sheets. These events are found to have a very different impact than similar events from the Laurentide Ice Sheet. Also the difference between the Heinrich simulations and freshwater flood simulations from the same sources highlights the importance of freshwater addition through icebergs.
Jean Lynch-Stieglitz
School of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta, GA 30332-0340
Geostrophy has provided a powerful constraint for the reconstruction of ocean circulation from modern oceanographic measurements. We use oxygen isotope measurements on benthic foraminifera to reconstruct density distributions in the past, and can use this information to make inferences about past circulation states. In this talk I will discuss some of the work we have done reconstructing the geostrophic shear associated with the meridional overturning circulation in the South Atlantic during the Last Glacial Maximum which suggest a dramatically different circulation at that time. I will also show some reconstructions of the horizontal density gradients in the Florida Straits, which can then be used to calculate the vertical shear in the volume transport. This information is combined with other paleoceanographic data as well as our knowledge of ocean and atmospheric dynamics to gain insight into past changes in the circulation of the North Atlantic Ocean. This data supports the idea of a dramatic weakening in the AMOC during the Younger Dryas and Heinrich Event 1, times of enhanced melt water flux to the North Atlantic. Because the density structure changes simultaneously with the changes in ocean circulation, detailed radiocarbon dating of the abrupt shifts in the Florida margin record can yield insights into the relationship between changes in ocean circulation and abrupt climate change.
Alison MacLeod1, Adrian Palmer1, John Lowe1, Jim Rose1, Jon Merritt2
1) Centre for Quaternary Research, Department of Geography, Royal Holloway University of London2) British Geological Survey, Murchison House, West Mains Road, Edinburgh, EH9 3LA
Development of a high resolution chronological framework spanning the LGIT in Britain addresses a key research objective of RAPID: determining the rate and timing of climatic changes during the last glacial - interglacial transition and consequent environmental responses. Varved sediments provide the potential for dating discrete events with annual resolution. Construction of a robust chronology, however, relies on an ability to securely differentiate between sequences which are simply rhythmically laminated and those which preserve verifiable annual records of sedimentation. At the present time, analysis of glaciolacustrine varved sequences is most commonly based on core description, X-radiography and micromorphology. These methods provide us with considerable detail of the structure and sedimentological properties of the sediments at macro- and micro-scales. However, there are problems associated with each technique: (1) visual description and X-radiography is rapid but does not always provide the resolution to enable the classification of sediments as ‘varves’ – especially in sequences where the varves are sub millimeter in thickness. (2) Thin-section micromorphology provides more precise sedimentological detail on the internal structure and the nature of the contacts between each couplet; however, it is a relatively time consuming, expensive and semi-destructive technique. Here we present the potential for an approach which employs well-established techniques, such as colour/grey scale analysis, X-radiography and magnetic susceptibility, in combination with down-core XRF analysis as a tool for the identification, delimitation and classification of varved sediments. Preliminary results presented here indicate that variations in elemental ratios recorded in the varves provide a coherent signal, which can be validated against both X-radiography and thin-section analysis. Elements such as Zirconium, Strontium and Rubidium exhibit cyclic variations related to changes in grain size which may be related to summer and winter deposition. This shows the potential for this relatively rapid, non-destructive core scanning technique to provide more rapid and comprehensive assessment of varved sediments in the palaeo record. Allied to this, identification of key marker layers within the sequences allows intra- and inter-site verification and correlation of results; while sedimentological descriptions obtained independently using thin section analysis provide a basis for validating the results.
Robin McCandliss1, Julie Collins1, Mark Hebden1, Zoe Aston1, Anne De Rudder2, Kevin Marsh2
1) British Oceanographic Data Centre2) British Atmospheric Data Centre
The RAPID Data Centre (RDC) provides data support to all projects in the RAPID Directed Mode 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 purpose of the RDC is to provide a long term archive for RAPID datasets, to give comprehensive advice on all aspects of data management, and to promote data sharing and collaboration between projects. The RDC also supports the RAPID community through the provision of resources such as the recovery and provision of third-party datasets, on-line tools for data analysis, mailing lists, and a user helpdesk.
Ben McDonald, Gregory C Smith, Keith Haines
Environmental Systems Science Centre, University of Reading
Part of the motivation for the work is the development of an assimilation system which uses a short-term forecast model to reanalyse MOC. MOC in the global model over a period of four years from 2002 to 2005 is presented for an operational run, a reanalysis, a control run and an additional reanalysis without assimilation of Argo observations. The effects of assimilating Argo observations on the representation of MOC in the system are discussed in particular. Principally, the effect is to increase MOC magnitude and produce a more coherent phase pattern in latitude and time, especially, a more pronounced peak in overturning around 36N. MOC at 24N in the model is compared with the recent Rapid array estimates.
Lorna M. McLean, Brian A. King
The scales over which ocean properties vary play an important role in both ocean data assimilation and the delayed mode quality control for Argo floats. At present Argo use one set of scales for all parts of the global ocean with adjustments made for coastal regions and the poles. The default scales used by the Argo group as of 2005 are 4° longitude and 2° latitude. These scales have been adequate for use in the Argo quality control method, however, extensive research into the variability of these scales with region has not been carried out. In this study one year of salinity data from the ENACT data set is used to develop a method of estimating the correlation scales. For all pairs of observations in a region, the difference in salinity is found and divided into 50 km bins according to the distance between the observations. The rms difference in salinity is then calculated for each bin. A curve is fitted to the data increasing exponentially from the near field to a far field limit based on a function devised by Bohme and Send [Bohme, L. and U. Send, 2005 Deep Sea Research II, 52, 651-664]. By varying the scale parameter, and finding the best fit to the data (the lowest rms error), a best estimate for the correlation scale is found. Three sections of the Pacific are chosen to test the method. It is found that determining the scale from the spatial separation alone does not provide a sufficiently accurate fit to the observational data. As the distance between the observations reduces to zero the difference in salinity between observations does not tend to zero as the equation suggests. The temporal separation is investigated but found to have a small effect relative to the spatial separation. An offset is therefore incorporated into the equation. The offset in the near field is chosen to be half of the far field limit and this gives a reasonable fit to the data for all 3 sections. Estimates of zonal and meridional scales are found to be 400 to 500 km and 100 to 300 km respectively.
Doug McNeall
Climate models are expensive to run, we wish to predict output at untested inputs. Progress has been made recently with statistical models to the full climate model - emulators. These techniques offer large efficiency gains of over running ensembles with a full climate model. Efforts have focused on univariate (1D) climate model output - MOC strength, or global mean air temperature for example. Here we present a method to statistically model multivariate climate model output, for example fields of surface air temperature or ocean overturning. While a field may appear to be of several thousand dimensions (datapoints), the number of effective degrees of freedom may be far fewer. We use output of an ensemble of GENIE as a training data set, and principal component analysis (PCA) to express the output field as a linear combination of a few orthogonal variables. We use a Gaussian Process emulator to estimate the climate model output at untried inputs in this reduced-dimension space, and then transform back into the full model space. In this way, we reduce the dimension of model output from a few thousand, to a few. Validation of the technique is performed using leave-one-out cross validation and full holdout methods. We examine and discuss the implications of spatial variability of emulator error.
Doug McNeall, Peter Challenor
If we are to take action to mitigate against abrupt climate changes we need to be able to specify the risk, and how that risk can be modified. We present a general method for the calculation of such risks and apply it to the risk of the collapse (or severe reduction) of the Atlantic overturning circulation given a proposed emission path. We investigate this problem using an intermediate complexity climate model - GENIE - to give us a good test bed for developing methods that can later be applied to GCMs. We apply methods developed by the statistical community to analyse computer experiments, centred around the concept of an emulator. An emulator is a statistical approximation to the dynamical model. Where we have run the model the fit is exact. At untried inputs, the emulator gives an estimate of the model output plus a measure of uncertainty. We build such an emulator for the GENIE model. This involves running the model in a Latin Hypercube design, extending to reasonable limits for its input parameters. Later we simulate from the uncertainty distributions of these input parameters. We spin GENIE up for 4000 years to remove (as far as possible) the dependence on initial conditions, and train the emulator using the model output, giving us an emulator for the strength of the MOC at the present day. The addition of two extra parameters – the rate that the ice sheets on Greenland melt and the rate at which carbon is removed from the atmosphere – enables us to run the model to 2100 under each of the six illustrative SRES scenarios (A1B, AIFI, A2, B1, B2 and A1T) and build an emulator for GENIE at 2100. We propagate the uncertainty on the inputs through to the uncertainty on the outputs. Using a Monte Carlo technique, we can now calculate the probabilities of the overturning circulation being less than some specified value, 5Sv say. However we need not simply look at the strength of the overturning at 2100, we can examine the shape of the trajectory to get to that value. Does the overturning reduce at a constant rate or are there sudden reductions in its strength? In this talk we update our earlier work on the probability of a severe reduction in the MOC and examine the trajectories to get there.
Alex Megann, Adrian New, Adam Blaker, Bablu Sinha
Ocean Modelling and Forecasting, NOC
The CHIME (Coupled Hadley-Isopycnic Model Experiment) model comprises the Hadley Centre's atmosphere and ice models, coupled to the Hybrid-Coordinate Ocean Model (HYCOM), which uses an isopycnic vertical coordinate over most of the ocean interior. CHIME is designed to be identical to the Hadley Centre's HadCM3 model except for its ocean component, allowing us to investigate the sensitivity of the climate to the choice of vertical coordinate in the ocean model. CHIME is being run as part of the RAPID THCMIP climate model intercomparison project. A 200-year control integration is now well under way, and a 1% per year increasing CO2 run and a run with freshwater hosing in the North Atlantic are planned. Although the meridional overturning in the control runs of CHIME and HadCM3 is similar, CHIME shows several differences compared with HadCM3. In particular, CHIME does not show the 4 degree cold bias seen in the North Pacific in HadCM3, instead having a warm bias in this region. General comparisons will be made between CHIME and HadCM3, and in addition more detailed results will be presented on the relationship between the Atlantic overturning circulation and the steric height gradient.
Katie Miller1, Mark Chapman1, Julian Andrews1, Nalân Koç2
1) School of Environmental Sciences, University of East Anglia (UEA), Norwich, NR4 7TJ ,UK.2) Norwegian Polar Institute N-9296 Tromsø, Norway
While it is widely accepted that interglacial climates are more stable than glacial climates, palaeoceanographic studies of the subpolar North Atlantic have indicated significant climate variability during the current interglacial period. A number of periods of lower temperatures and increased iceberg discharge have been identified, most notably the ‘8.2 kyr event’, throughout the Holocene. These variations are thought significant due to their potential association with changes in the Meridional Overturning Circulation. This study concentrates on the 8.2 kyr climate event as recorded in marine sediment core MD99-2251 from the Gardar Drift in the subpolar North Atlantic. Previous analyses of the foraminiferal assemblages, deep flow speed and oxygen isotopes through this time interval have indicated a pulsed two step meltwater discharge from the proglacial Lakes Agassiz and Ojibway as having forced a reduction in the flow rate of Iceland Scotland Overflow Water and a freshening and cooling of the surface ocean (Ellison et al., 2006). This study seeks to identify whether this sequence of events can be identified in the diatom and sedimentary record. We have generated high resolution (1-2cm) multi-proxy data using quantitative diatom assemblages as an indicator of sea surface conditions, variations in the diatom genus Chaetoceros as an indicator of palaeoproductivity and analyses of terrigenous material in the sediments as an indication of drift ice flux over the coring site. A peak in the dominance of subarctic flora is associated with the 8.2 kyr event, with the onset of cooling dated to ~8.300 years ago based on a calibrated 14C time scale. Variations in the amount of ice rafted material also reveal distinct multidecadal and centennial signals but no strong enhancement in terrigenous input associated with the 8.2 kyr cooling.
Aazani Mujahid, Torsten Kanzow, Guillaume Charria, Harry L. Bryden
National Oceanography Centre, Southampton.
A recent numerical study from Hirschi et. al. (2006) suggests that on short timescales there is a link between baroclinic components and westward moving transport anomalies (as long baroclinic Rossby waves) which could contribute several Sverdrups to the MOC variability. With recent interest focused on the MOC (Schiermeier, 2006) and findings by Bryden et. al. (2005) that the MOC at 25°N may have slowed by around 30% - all possible contributions to long term trends must be accounted for including propagating features. When using satellite altimeters to track features in longitude/time plots of sea surface height (SSH) anomalies, we do observe westward propagation. Do we observe propagating features in real-time mooring array datasets as well? The mooring array systems of RAPID-MOC (Monitoring the Atlantic Meridional Overturning Circulation at 26.5N) and MOCHA (Meridional Overturning Circulation and Heat-flux Array) have been continuously monitoring the Atlantic Meridional Circulation (MOC) in the subtropical North Atlantic since spring 2004. With the opportunistic use of the western boundary moorings from 2004-2005 along with historical data (1992 – 1993), we analyze the vertical structure of the meridional flow in the complex Antilles Current (AC)/ Deep Western Boundary Current (DWBC) regime, quantifying the temporal and spatial characteristics of the variability in the regime. What is the cause of this variability? Could the variability from the moorings be correlated to variability in SSH? First results show the vertical structure as a decoupled regime of upper northward flowing AC from the southward DWBC, with sub-annual variability. We have observed coherent propagating flow events across the different mooring sites possibly related to meandering, bifurcation (Pickart, 1994), or recirculation (Lee et. al., 1996). Spectral analysis methods reveal the time-varying spectral content of the dataset, in this case using FFT with Welch windowing shows periodicity of over 100 days. Variations in the inter-annual meridional flow are derived using empirical orthogonal functions (EOFs) to determine the leading modes and associated principal components (PCs). For example, 97% of the variance in meridional velocities can be explained by the first 3 EOF modes which are associated with the barotropic mode (70%), followed by the first two leading baroclinic modes (21% and 6% respectively).
Eleanor A O'Rourke1, Christopher W Hughes1, Richard G Williams2
1) Proudman Oceanographic Laboratory2) Department of Earth & Ocean Sciences, University of Liverpool
A coherent signal in the sub-surface pressure is seen along the western boundary of the North Atlantic. Johnson & Marshall (2002) and Kawase (1987) showed that, using idealised models with vertical sidewalls, a change in forcing in the North Atlantic results in baroclinic Kelvin waves propagating along this same pathway reaching the equator in a matter of months. Such waves could therefore prove to be a rapid indicator of change to buoyancy forcing in the North Atlantic. However, in the real ocean, with complex topographic and stratification profiles, the full spectrum of coastal trapped waves, from baroclinic Kelvin waves through to barotropic shelf waves, are found. In this work, the structure of coastal trapped waves was examined based on free modes, firstly in a barotropic model with realistic topography, and then in a baroclinic model also incorporating realistic stratification profiles. The barotropic results show dramatic differences in the nodal structure of the free modes linked to detailed changes in the topography along the western boundary. Early baroclinic output has suggested that comparable results may be found due to similarities in the velocity and pressure fields. Therefore, it can be suggested that the propagating coastal trapped wave modes are not an adequate description of the coherent signal observed.
Matthew Palmer1, Keith Haines2, Simon Tett1, Tara Ansell1
1) Met Office Hadley Centre for Climate Change2) NERC Environmental Systems Science Centre, University of Reading
An analysis of historical ocean temperature data relative to fixed isotherms is presented. The method is designed to filter out high frequency “noise” and the influence of circulation changes on estimates of ocean warming. The new approach reveals a more globally-uniform spatial pattern of warming and is more closely related to changes in air-sea heat exchange than previous fixed depth studies. The method also removes some sources of error in the observing system and provides an improved baseline for evaluation of climate models.
Clare F. Postlethwaite, Graham R. Tattersall, Miguel A. Morales Maqueda, Jason T. Holt, Andrew J. Willmott
Proudman Oceanographic Laboratory, 6 Brownlow Street, Liverpool, L3 5DA, UK.
The Arctic shelf seas are key regions of dense water formation contributing to the global thermohaline circulation. Reports of decreased Arctic summer sea-ice extent highlight the importance of ensuring all relevant Arctic processes are sufficiently represented in climate prediction models. We investigate whether the interaction between tides and sea-ice significantly alters model results of dense water formation. Tidal currents can cause divergence of sea-ice, bringing about the opening of leads within the ice pack. Although these areas of open water are small and short lived, they can lead to large heat fluxes from the ocean to atmosphere and increased salt fluxes to the surface ocean as new ice forms. We present results from a dynamic/thermodynamic sea ice model coupled to a baroclinic coastal ocean model for the Barents and Kara Seas. Comparing results from the model run both with and without tidal forcing indicates differences in the ice distribution for the two scenarios. Results show that including tidal forcing in a coupled coastal ocean/ice model can increase the salt flux to the ocean and affect the dense water formation in the region.
Sean D.F. Pyne-O'Donnell1, Simon P.E. Blockley2, Alison MacLeod1, J. John Lowe1, Mike J.C. Walker3
1) Centre for Quaternary Research, Geography Department, Royal Holloway University of London2) Research Laboratory for Archaeology and the History of Art, Oxford University3) Department of Archaeology and Anthropology, University of Wales, Lampeter
A key research objective for RAPID is to establish rates of change in North Atlantic THC (Thermohaline Circulation) during the LGIT (Late Glacial-Interglacial Transition: 16-8 cal. yr BP). The LGIT was a period of pronounced climatic perturbation thought to reflect ‘forcing’ by THC changes. Constructing an accurate radiocarbon timescale for the inferred changes during this period, however, is problematic due to a number of complications such as calibration and marine reservoir offsets. In order to achieve the more precise decadal resolution required for RAPID objectives we are employing (volcanic ash isochrones) to correlate continental NW European and Greenland ice-core sequences with shallow shelf and deep marine records. By using precise ash isochrones, the duration of events and degree of synchroneity between THC changes and their forcing effects in the terrestrial and ice realms can be assessed. A number of well-dated and widespread distal ashes (‘microtephras’) are known to have been deposited during the LGIT in terrestrial and ice-core sequences throughout the North Atlantic region, with additional discoveries regularly being added to the tephrostratigraphy of this period. In collaboration with other institutions, ten North Atlantic marine sequences have been analysed to-date and found to contain a number of discrete ash isochrones which correlate with those in terrestrial and ice-core sequences. Three sequences from the NW Atlantic contain tephras which geochemically correlate with the mid-GS-1 Vedde Ash (ca. 12,000 GRIP yr BP) and possibly the Penifiler (ca. GI-1c) and early GI-1 Borrobol tephras (ca. 14,400 cal. yr BP). Two adjacent sequences from just south of Greenland jointly contain the early Holocene Saksunarvatn Ash (ca. 10,200 GRIP yr BP) in addition to the Vedde Ash. Both these sequences also encompass the Last Termination and thus possibly also contain the Penifiler and Borrobol tephras. Two sequences from the NE Atlantic (Hebridean Sea) contain the Vedde Ash, along with the similarly aged I-THOL-2 and early Holocene I-THOL-1. In most cases the isochrones contain shard maxima within discrete distributions which are similar in nature to those seen in terrestrial lake sequences. Analysis of peak maxima between rhyolitic and basaltic shards at different size ranges is expected to provide valuable information about the mode of delivery and deposition of these tephras. For example, the suggestion that the Vedde Ash was ice-rafted in the NW Atlantic sequences can be used to assess current flow directions during the Younger Dryas. Proxy palaeoclimate reconstructions exist or are in development for a number of these sequences based on chironomids and forams etc. The next stage is to use the ash isochrones to test the degree of synchroneity of THC changes inferred from the proxy records. Attention will be focused most on those records that provide quantified estimates of rates and magnitudes of climate change.
Darren Rayner1, Stuart A. Cunningham1, Torsten O. Kanzow1, Harry L. Bryden1, Joel J.-M. Hirschi1, Jochem Y. Marotzke2, William E. Johns3, Lisa M. Beal3, Molly O. Baringer4, Christopher S. Meinen4
1) National Oceanography Centre, Southampton2) Max Planck Institute for Meteorology, Hamburg3) University of Miami, Rosenstiel School of Marine and Atmospheric Sciences4) NOAA-Atlantic Oceanographic and Meteorological Laboratory
Objective 1 of the RAPID programme is to establish a pre-operational prototype system to continuously observe the strength and structure of the basin wide Atlantic circulation. The meridional circulation is a balance of the northward flux of the Florida Current plus Ekman transport by the southward thermocline and North Atlantic Deep Water flux. The zonal integral of this circulation defines the meridional overturning circulation (MOC). While parts of the MOC are wind-driven, the basin-scale Atlantic MOC is largely buoyancy-forced. Hence, observing the Atlantic MOC is the fundamental observational requirement of a programme aiming to assess the role of the Atlantic thermohaline circulation (THC) in rapid climate change. The northward flow of warm water in the Florida Current is measured using telephone cables between the US and Bahamas; and the compensating southward circulation of cool thermocline and cold deep water are observed using the RAPID-MOC transatlantic array of moorings. The moorings measure top-to-bottom density profiles and bottom pressure near America and Africa and on either side of the mid-Atlantic Ridge. The transatlantic mooring array was first deployed in spring 2004, and has now been operating continuously for 3 years. The array consists of 3 sub-arrays; the Eastern Boundary; the Mid Atlantic Ridge; and the Western Boundary. This poster presents the current status of the array including changes implemented in the Autumn 2006 Eastern Boundary service cruise and the Spring 2007 Western Boundary service cruise. The Western Boundary sub-array is a joint operation between the National Oceanography Centre (Southampton), the Rosenstiel School of Marine and Atmospheric Science (Miami); and the Atlantic Oceanographic and Meteorological Laboratory (Miami).
Vassil M. Roussenov1, Richard G. Williams1, Chris W. Hughes2, Rory J. Bingham2
1) University of Liverpool, Department of Earth and Ocean Sciences2) Proudman Oceanographic Laboratory
Changes in high latitude forcing induce changes in overturning, which are communicated over the basin through wave propagation and advection along the western boundary. This communication process is explored here using an isopycnal model (MICOM) with horizontal resolution of 0.23 degrees on a Mercator grid. The model is integrated over realistic topography and driven by wind and buoyancy forcing. A transient model tracer is released in Labrador Sea in order to monitor the deep advection after the climatological mean state has been reached. Twin experiments are conducted during the last 10 years of integration with either unchanged or altered wind and buoyancy forcing. High latitude forcing perturbations lead to a pressure signal rapidly propagating along the western boundary on a timescale of several months. This rapid communication is induced by boundary waves, a mixture of Kelvin and topographic Rossby waves, and is followed by an intermediate response involving changes in local circulation and layer thickness, excited by the waves. Eventually, there is a slower far-field, advective response, as marked by the arrival of the Labrador Sea tracer with a timescale of typically several years. The frequency response of the wave signals alters with latitude with shorter periods confined to northern latitudes. This wave communication process is considered to be associated with corresponding changes in the meridional overturning circulation (MOC). Different types of forcing functions (annual-mean constant, repeating seasonal variations and realistic forcing including inter-annual variability) have been applied in order to identify the magnitudes of the internal and imposed variability of the overturning. For each case, the variations of the sea-surface height and the bottom pressure along the western boundary have been diagnosed and compared with changes in the MOC in order to identify the potential link between the western boundary signals and the overturning.
Gregory Smith, Keith Haines, Dan Lea, Jon Blower, Alastair Gemmell, Ben McDonald
ESSC, Univ. Reading
Our focus is on improving the realism of ocean data assimilation schemes and using assimilation to investigate ocean climate signals. The main aim is to produce a high-resolution reconstruction of the global ocean over the last 50 years, assimilating temperature and salinity observations. The relative abundance of collocated temperature and salinity observations provided by Argo are used to develop an assimilation scheme whereby temperature and salinity profiles are assimilated on isotherms and isopycnals. This allows us to exploit the larger spatial and temporal decorrelations of these quantities, compared with assimilation on geopotential surfaces, allowing flow dependent assimilation and recovery of water mass information. Two configurations of the NEMO ice-ocean model will be employed in this study. A 1 degree resolution version with a tropical enhancement to 1/3 degree will be used mainly for testing and sensitivity studies, while the ¼ degree DRAKKAR version will be used to perform the high-resolution reanalysis. Results from a series of experiments over the Argo period, assimilating on depth, temperature and density surfaces will be presented. Preliminary results from a first 50 year reanalysis using this model-data synthesis method will also be shown.
Jennifer D. Stanford1, Eelco J. Rohling1, Sally E. Hunter1, Andrew P. Roberts1, Sune O. Rasmussen2, Edouard Bard3, Jerry McManus4, Richard G. Fairbanks5
1) National Oceanography Centre, Southampton, University of Southampton, Southampton, SO14 3ZH, United Kingdom2) Ice and Climate Research, Niels Bohr Institute, University of Copenhagen, Denmark3) Collège de France, CEREGE, UMR-6635, Le Trocadéro, Europole de l'Arbois BP80, 13545 Aix-en-Provence cedex 4, France4) Woods Hole Oceanographic Institution, 121 Clark Lab, MS #23, Woods Hole, MA 02543, USA5) Lamont Doherty Earth Observatory and Department of Earth and Environmental Science, Columbia University, Palisades, NY 10964, USA
The temporal relationship between meltwater pulse 1a (mwp-1a) and the climate history of the last deglaciation remains a subject of debate. By combining the Greenland Ice Core Project δ18O ice core record on the new Greenland ice core chronology 2005 timescale with the U/Th-dated Barbados coral record, we conclusively derive that mwp-1a did not coincide with the sharp Bølling warming but instead with the abrupt cooling of the Older Dryas. To evaluate whether there is a relationship between meltwater injections, North Atlantic Deep Water (NADW) formation, and climate change, we present a high-resolution record of NADW flow intensity from Eirik Drift through the last deglaciation. It indicates only a relatively minor 200-year weakening of NADW flow, coincident with mwp-1a. Our compilation of records also indicates that during Heinrich event 1 and the Younger Dryas there were no discernible sea level rises, and yet these periods were characterized by intense NADW slowdowns/shutdowns. Clearly, deepwater formation and climate are not simply controlled by the magnitude or rate of meltwater addition. Instead, our results emphasize that the location of meltwater pulses may be more important, with NADW formation being particularly sensitive to surface freshening in the Arctic/Nordic Seas.
Remi Tailleux
Walker Institute for Climate System Research, University of Reading
There is currently a controversy among oceanographers as to whether the Atlantic meridional overturning should be regarded as a classical heat engine, or as a mechanically-driven heat transport heat engine powered by the winds and tides. We show in this paper that the mechanically-driven hypothesis violates the second law, because it implicitly assumes that the mechanical forcing can produce irreversible changes in the mean stratification with an efficiency of about 20 percent, whereas the second law dictates that it can only produce irreversible changes after dissipating into heat, for which the efficiency is less than 0.1 percent in general. It follows that the oceans must be a heat engine, and that the existence of diabatic thermally indirect cells in the oceans is extremely unlikely. This result conflicts with the energetics of current ocean general circulation models, in which the work of expansion/contraction is in general negative. This seems to suggest that current OGCMs may underestimate the effect of the buoyancy forcing, and overestimate their dependencies upon the mechanical forcing. Another implication is that non Boussinesq effects are likely to be two or three orders of magnitude larger than previously thought.
Ian Walkington1, Miguel Angel Morales Maqueda2, Andrew Willmott2
1) Department of Engineering, University of Liverpool2) Proudman Oceanographic Laboratory, NERC
Latent heat polynyas are regions of low ice cover in areas characterised by high ice concentration, and are formed and maintained by the dynamic action of wind and ocean currents. Polynyas may contribute to the formation of Arctic intermediate and deep waters due to the large ice production rate and associated brine rejection found within them. The impact of these small scale features may therefore have regional and global significance through these water masses. Polynyas are not well represented in large scale sea ice models due to two main factors. Firstly, the coarse resolution of these models means that polynyas cannot be properly captured by them. Secondly, the nature of large scale sea ice models means that they are valid only for length scales larger than the typical length scale involved in the physics of polynya formation. It is therefore not possible at present to accurately represent polynyas with large scale sea ice models. The only option is to parameterise the effect of polynyas and include these into the models. A polynya model has been developed utilising shock techniques commonly used in gas dynamics. The polynya edge, where the frazil ice is consolidated to the ice pack, is viewed as a jump, or shock, in the ice cover. Within the model, ice mass, momentum and energy are all conserved across the polynya edge. The evolution of the polynya and an estimate of the brine rejected may then be calculated. This model may then be used as a parameterisation within large scale sea ice models so that the effect of polynyas may be better understood on the larger scales.
Craig J. Wallace
All NERC thematic programmes undertake knowledge transfer activities to ensure key scientific findings are delivered to appropriate members of the public and private sector. Although many results from RAPID are still accumulating, initial results are now available and are actively being distributed to our stakeholders. In this presentation I will outline: RAPID's strategy for ensuring effective knowledge transfer; progress since last year, and; plans for the future - with consideration of RAPID's finale in 2008.
Paul D Williams1, Eric Guilyardi12, Rowan Sutton1, Jonathan Gregory13, Gurvan Madec2
1) Walker Institute & NCAS-Climate, Department of Meteorology, University of Reading, UK2) Laboratoire d'Oceanographie et de Climat: Experimentation et Approche Numerique (LOCEAN/IPSL), CNRS/Universite Paris VI, Paris, France3) Hadley Centre for Climate Prediction and Research, Exeter, UK
An intensification of the hydrological cycle is a likely consequence of global warming. But changes in the hydrological cycle could affect sea-surface temperature by modifying diffusive ocean heat transports. We investigate this mechanism by studying a coupled general circulation model sensitivity experiment in which the hydrological cycle is artificially amplified. We find that the amplified hydrological cycle depresses sea-surface temperature by enhancing ocean heat uptake in low latitudes. We estimate that a 10% increase in the hydrological cycle will contribute a basin-scale sea-surface temperature decrease of around 0.1 K away from high latitudes, with larger decreases locally. We conclude that an intensified hydrological cycle is likely to contribute a weak negative feedback to anthropogenic climate change.
Tim Woollings, Brian Hoskins, Mike Blackburn
Dept of Meteorology, University of Reading
The North Atlantic Oscillation (NAO) is the dominant pattern of natural climate variability affecting Europe. The strong trend from the negative to the positive NAO phase over recent decades is associated with about half of the observed changes in European winter climate. However, it is not known to what extent this trend is due to random internal variability or to external forcing. Here we present a new interpretation of the NAO as a signal of variations in the occurrence of blocking-like Rossby wave-breaking events on the tropopause. A few dynamical precursors to wave-breaking are identified, and these suggest mechanisms by which low-frequency variability in the oceans and the stratosphere could modulate the occurrence of wave-breaking, and so affect the NAO. A preliminary analysis is undertaken to investigate if any of these mechanisms are responsible for the recent dramatic trend.
Gerard van der Schrier1, Sybren Drijfhout1, Wilco Hazeleger1, Ludovic Noulin2
1) Royal Netherlands Meteorological Institute2) ecole d'Ingenieurs en Modelisation Mathematique et Mechanique
This study examines a recent hypothesis concerning abrupt climate change over the North Atlantic sector in a General Circulation Model of intermediate complexity. Central to this hypothesis is that a substantial strengthening of the Atlantic subtropical jet will lead to a reorganization of the mid-latitude atmospheric circulation in which the Atlantic subtropical and eddy-driven jets coincide. The subtropical jet over the North Atlantic sector is increased by calculating optimal model tendency perturbations, using a recently developed technique. Strengthening of the subtropical jet is achieved by model tendency perturbations which 1) sharpen the meridional gradient in streamfunction or 2) force a strongly and persistently negative North Atlantic Oscillation pattern in the upper troposphere. Both approaches lead to similar forcing patterns. It is found that when the subtropical jet is sufficiently strong, the eddy-driven jet is indeed drawn to the northern rim of the subtropical jet. This reduces atmospheric meridional heat transport over the North Atlantic. The southern sea ice-edge extends further south and an overall cooling of the circum-North Atlantic region is observed.
Gerard van der Schrier, Nanne Weber
Royal Netherlands Meteorological institute
We present gridded Sea-Surface Temperatures (SSTs) for the Atlantic basin (45S-60N) for the period AD 1790-1825, based on early-instrumental SST data. The original measurements were compiled by Major James Rennell and made by numerous British naval vessels on behalf of the British Admiralty. We describe the digitization of this dataset and the reconstruction of spatially coherent, averaged conditions for the boreal warm (November-March) and cold (May-September) season using a reduced space Optimal Interpolation technique, in which the data is projected on a limited number of empirical orthogonal functions. This approach is validated on modern data which are sampled in a similar way as the early-instrumental data. The reconstruction for the November-March period shows a large area with anomalously high temperatures from the point where the Gulf Stream separates from the coast until ca. 15W. A tongue of anomalous cool water is found at the eastern side of the North Atlantic basin, along the coast of Europe and northern Africa. In the north eastern South Atlantic, anomalously high temperatures are found, while temperatures in the south western South Atlantic are anomalously cool. For the March-September season, anomalous temperatures in the South Atlantic are similar, but stronger, compared to those in the boreal cold season. In the central North Atlantic, anomalously high temperatures are found.