2nd Annual Meeting
Copenhagen, 17-18 November 2010
- Agenda of the meeting, with list of interventions and confirmed participants: download
- Minutes of the meeting: introduction, scientific talks and project management: download
Science talks on variability of overflows
1. Kjetil Våge: Upstream pathways of Denmark Strait Overflow Water (download) (UiB Bergen NO)
The dense Nordic seas' overflow waters constitute the lower limb of the Atlantic meridional overturning circulation (MOC), and, as such, are a crucial component of the Earth's climate system. Warm subtropical-origin waters flow northward across the Greenland–-Scotland Ridge where they are subject to intense air-sea interaction. After releasing heat to the atmosphere, the resulting dense water returns southward by flowing through the gaps in the ridge and descending the continental slope as overflow plumes. These overflows represent the headwaters of the MOC; the largest of these is the Denmark Strait Overflow Water plume which passes southward between Greenland and Iceland and contributes to the Deep Western Boundary Current. It is commonly thought that the primary source of the overflow water is the East Greenland Current (EGC), a southward-flowing current along the continental slope of Greenland. However, this view has been called into question with the discovery of the Northwest Icelandic Jet (NIJ) -- a current transporting overflow water towards the Denmark Strait along the continental slope of Iceland. Results from two shipboard hydrographic/velocity surveys in October 2008 and August 2009 highlight the importance of the NIJ. During the period of the surveys, the NIJ was found to supply both the densest overflow water and a significant fraction of the total overflow transport (30% - 50%). The NIJ was traced as far upstream as northeast of Iceland, confirming it as a distinct pathway from the EGC.
2. Rolf Käse: Model pathways of dense water across the Iceland-Scotland-Ridge in NAO+/- composites (download) (R. Käse, G. Voet, D. Quadfasel, C. Rodehacke) (UHAM DE)
Recent observations show that the deep flow from the Nordic Seas across the ridge directly east of Iceland has a more persistent character than thought which was previously known as sporadic and dominated by eddies. To elucidate the importance of the role of seasonal forcing by surface fluxes, a regional model (ROMS) was driven by composite climatological NAO+ or NAO- cycles. Tracer release experiments reveal that the origin of the overflowing water in both cases is located north and east of Iceland and the channel is reached faster for NAO- situations with stronger transport. A splitting of the flow into barotropic and baroclinic parts shows that the annual cycle is the same for the latter, while the amplitude of the former is larger. Maximum overflow occurs during summer seasons.
3. Hendrik van Aken: Long time series of Labrador Sea Water near its origin, and in the Irminger Sea (download) (H. M. van Aken and M. Femke de Jong ) (NIOZ NL)
The hydrographic reference data sets for the Irminger and the Labrador Sea mainly contain section data from the WOCE AR7/A1 section, surveyed since 1990. In order to extend this data set, hydrographic data since 1950 were collected from the WODB and ICES data centres for the central parts of the Irminger and Labrador gyres. While the resulting time series of temperature and salinity mainly reflect a large but slow inter-decadal variability, the oxygen content and stability of the water column reveal the near decadal occurrence of deep (~1500 m) reaching convective ventilation events in the Labrador Sea. These deep hydrographic signals are advected to the Irminger Sea, where they occur with a delay of 1 to 1½ year.
4. Amrita Shravat: How well can the ocean-reanalyses represent the formation of Deep Western Boundary Current? (download) (SAMS UK)
A volume transport based metric has been used to compare ocean re-analyses with observations across the Greenland-Scotland Ridge and in the Labrador Sea. The transport obtained from the re-analyses ORA-S3 and ECCO-JPL for the Atlantic water inflows and the Labrador Sea outflow is unrealistically small. On the other hand, DePreSys produces a rather large inflow across the Greenland-Scotland ridge (more than 15 Sv) and a substantially weak outflow in the Labrador Sea (less than 10 Sv). The relatively higher resolution re-analyses NCEP (1/3 degrees in latitudinal direction) and SODA (1/4 degrees) in particular, seem to produce the observed inflow across the ridge, while capturing the inter-annual variability. They do not produce the observed outflow from the Labrador Sea. The analysis at 26 N suggests that the re-analyses, except for SODA and DePreSys do reproduce the expected northward and southward transport. We are therefore not confident that the hindcasts produced from re-analyses will be fit to determine the strength of the AMOC.
MOC Variability in general and over the last millennium
5. Marie-Noelle Houssais: The influence of the Arctic Ocean fresh water content on the fresh water outflow to the Atlantic MOC (download) (C. Herbaut, M.-.N. Houssais, D.Iovino)
In a hindcast simulation of the Arctic-North Atlantic region, the Atlantic meridional overturning circulation (MOC) is shown to be sensitive to the surface fresh water outflow through the Fram Strait with enhanced fresh water export leading to reduced MOC intensity at about 60ºN. The interannual variability of this export impacts on the surface salinity in the Labrador Sea, with a strong response in the deep water formation area. It is also correlated with the variability of the mean annual fresh water content in the Beaufort Gyre. The latter appears to be the result of distinct seasonal contributions which have been analysed in details. While sea ice growth/melt anomalies appear to play a marginal role in the fresh water content evolution, Ekman pumping dominates the response in all but the summer seasons and is shown to be related to a dipolar atmospheric seal level pressure pattern having a strong signature in the northern Pacific.
6. Christian Rodehacke: First glance on simulated fresh water releases around Greenland: Common features and differences between models (download) (MPI-M DE)
In a warming world enhanced surface melting and a increased ocean ice sheet ocean interaction release more fresh water from Greenland into the surrounding coastal ocean. The ocean's reaction has been simulated in standalone and coupled climate models for a prescribed additional fresh water hosing of 0.1 Sv (100,000 m3/s) along the coast Greenland. A preliminary analysis highlights the spreading paths of the additional fresh water: In most models the fresh water anomaly flows predominately into the Labrador Sea and afterwards into the subpolar Atlantic. There it splits into two branches. One takes a route towards the equator the other into the Greenland-Island-Norwegian (GIN) Sea. A common feature is a distinct warming of the SST after four decades in the northern GIN Sea. A quantitative analysis shows for most models a decrease of the overflow water mass densities after several decades. Also the by deep reaching convection in the winter enclosed volume decreases for both the Labrador and the GIN Seas. However the amplitude of the latter signal is in some models as large as its variability. As a consequence of the reduced overflow densities and ventilated volumes decreases the meridional overturning circulation (MOC) at 36N by up to 10Sv. The reduction ranges typically from 0.5 to 5Sv, whereby the reduction seems to be independent of its absolute strength in the undisturbed case. Further analysis are required to weight the different model results and to obtain a more consistent picture of the uncertainties.
7. Katja Lohmann: Sensitivity of the (simulated) THC to changes in the northern North Atlantic (download) (K. Lohmann, J. Jungclaus, M. Botzet, H. Haak, S. Lorenz, D. Matei) (MPI-M DE)
We analyze a 1000 year long control experiment as well as a 1000 year long sensitivity experiment with suppressed variability of subpolar deep water formation (performed with a relatively coarse, but highly efficient version, of the Max-Planck-Institute Earth System Model) to investigate the response of the low-frequency THC variability to suppressed variability of subpolar deep water formation. Regression of an index for Labrador Sea Water thickness onto the THC at 1000m depth (depth of maximum THC strength) in the control experiment suggests the largest influence of subpolar deep water formation variability onto the THC between 40oN and 50oN. The strongest reduction in the THC variability in the sensitivity experiment is also found at these latitudes. Preliminary results concerning variability of the overflows from the Nordic Seas suggest a similar response of the THC variability as seen for the subpolar deep water formation. It is noteworthy that the largest THC variability in the control experiment is found at the same latitudes where we see the largest influence from subpolar deep water formation / overflows. Long control integrations from the other models in CT1 will be used to check the robustness of above results.
8. Guillaume Gastineau: Atmospheric response to the natural variability of the AMOC (download) (G. Gastineau, C. Frankignoul and J. Mignot) (UPMC-LOCEAN FR)
The influence of the natural variability of the Atlantic meridional overturning circulation (AMOC) on the atmosphere is studied in multi-centennial simulations, using Maximum Covariance Analysis. In all models, a significant but weak Sea-Level-Pressure response that resembles a negative phase of the North Atlantic Oscillation (NAO) is found during the Northern Hemisphere cold-season, when the ocean leads the atmosphere by 4 to 8 years. The signal amplitude is typically 0.5 hPa and explains about 1% of the interannual NAO. The atmospheric response occurs while the sea surface temperature (SST) increases in the Atlantic subpolar gyre and sea ice retreats in the Labrador Sea.
The atmospheric response seems to be due to an enhanced ocean heat loss along the Gulf Stream and the North Atlantic Current, in the subpolar gyre, which occurs a few year after the AMOC intensification. Stronger heating decreases the baroclinicity of the lower troposphere and the eddy activity in the North Atlantic storm track. This in turn drives an equivalent barotropic perturbation resembling the NAO.
9. Helga Kleiven: Climate and North Atlantic Deep Water variability since 600 AD (download) (H. Kleiven, Y. Rosenthal & U. Ninnemann) (UiB, Bjerknes Centre NO)
Few paleoclimate archives provide evidence for natural climate variability on multi-decadal to millennial timescales. Understanding the origin and expression of these low frequency modes of natural climate variability is crucial for constraining their potential role in current and future climate changes. Multi-decadal climate oscillations are often postulated to result from changes in the ocean's meridional overturning circulation (MOC). Testing this hypothesis for historically recorded climate changes such as the Little Ice Age (LIA) requires decadally resolved constraints on the state of ocean circulation spanning these events. Here we use a high sedimentation rate multicore (GS06-144 03 MC A, 57°29'N, 48°37'W, 3432 m) from the Eirik drift south of Greenland to infer the timing, amplitude, and nature of climate and deep water changes since 600 AD. The Eirik Drift accumulates rapidly as a result of the influx of sediments eroded from the Denmark Strait and eastern Greenland margin suspended in Denmark Strait Overflow Water (DSOW). DSOW combines with North West Atlantic Deep Water (NWADW) to form the Western Boundary Undercurrent (WBUC). Thus, this location is well situated to monitor changes in the past circulation and properties of nearly mature NADW along its western boundary flow path.
We reconstruct the bottom water physical and chemical properties of the lower branch of the MOC (the deep overflowing branches from the Nordic Seas) using the oxygen and carbon isotopes of benthic foraminifera, whereas changes in the vigor of near bottom flow are inferred from size variations in the sediment mean sortable silt. The changes in surface hydrography co-registered in the planktonic foraminiferal isotopic records are used to assess possible linkages between MOC and climate. The planktonic foraminiferal oxygen isotopic and Mg/Ca temperature reconstructions show centennial-scale variations characterized by several distinct cooling events, the most pronounced being an abrupt cooling around 1400 AD in both surface and deep water records closely aligned with the timing of the LIA. Superimposed on the centennial variations are higher frequency oscillations, which are similar in duration and amplitude to those of the Atlantic Multidecadal Oscillation (AMO). The results suggest a close coupling between surface climate and the properties of proto North Atlantic Deep Water, primarily on centennial timescales. Coeval changes in both kinetic and chemical deep water proxies provides strong support that both the intensity and properties of NWADW vary on multi-decadal timescales over the last 600 AD
10. Ulysses Ninnemann: Variability in ISOW vigor over the last millennium and its relationship to climate (download) (U. Ninnemann, T. Mjell, I. Hall and H. Kleiven) (UiB, Bjerknes Centre NO)
As part of the EU-FP7 program THOR we are reconstructing variability in the deep Nordic Seas overflows over the last millennium. Here we use well dated (210Pb and AMS 14C), high sedimentation rate, multi and gravity cores taken on the Gardar Sediment Drift (60°19N, 23°58W, 2081 m) to reconstruct decadal to centennial variability in the properties and vigor of the eastern branch of the Nordic Seas overflows over the past millennium. The Gardar drift accumulates on the eastern flank of the Reykjanes ridge due the supply of sediments provided by the overlying Iceland Scotland Overflow Water (ISOW). We reconstruct the bottom water physical and chemical properties of ISOW using the oxygen and carbon isotopes of benthic foraminifera (C. wuellerstorfi), while changes in the vigor of near bottom flow are inferred from size variations in the sediment mean sortable silt ( ). The data produced to date provide a sub-decadally sampled history of ISOW variability spanning from ~600-2000 AD. In addition, changes in surface hydrography are reconstructed using ´18O of the planktonic foraminifera N. pachyderma (d) and G. inflata. Our results show multi-decadal and centennial variability in ISOW vigor throughout the past millennium that is in phase with reconstructed AMO on both inter-decadal and centennial timescales—with increased (decreased) ISOW vigor during positive (negative) AMO phases. This strong coherence suggests that low frequency variability in key components of AMOC are tightly coupled to basin wide temperature perturbations. However, our reconstructions of local near surface hydrography are out of phase with both the ISOW vigor and basin scale temperature pattern (AMO). We discuss the nature of the AMO-ISOW coupling as well as the possible influence of subpolar gyre dynamics in affecting an antiphase between local hydrography and AMO.
11. Juliette Mignot : Oceanic response to volcanic eruptions over the last Millennium (download) (J. Mignot, M. Khodri, J. Servonnat and C. Frankignoul) (UPMC-LOCEAN FR)
The oceanic response to volcanic eruptions over the last millennium is investagted over a fully coupled ocean-atmosphere general circulation model forced by reconstructions of the total solar irradiance modulation, modulations of the stratospheric aerosols optical depth due to volcanic eruptions and tropospheric greenhouse gases. Anomalies following volcanic eruptions typically remain persistent for more than 10 years in surface and longer in subsurface. The oceanic barotropic and zonally averaged meridional circulations are both rapidly modified in response to anomalous wind stress resulting from the tropical cooling. At longer timescales, the response seems to depend more on the succession and magnitude of the volcanic events.
Decadal variability and predictions
12. Nuno Nunes: Interannual and decadal variability of the Labrador Sea outflow (download) (J. Fischer, M. Visbeck, R. Zantopp, N. Nunes) (IFM GEOMAR DE)
A decade of weak convection in the Labrador Sea associated with decreasing water mass transformation, in combination with advective and eddy fluxes into the convection area, caused significant warming of the deep waters in both the central Labrador Sea and boundary current system along the Labrador shelf break. The connection to the export of Deep Water was studied based on moored current meter stations between 1997 and 2009 at the exit of the Labrador Sea, near the shelf break at 53N. More than 100 year-long current meter records spanning the full water column have been analyzed with respect to high frequency variability, decaying from the surface to the bottom layer, and for the annual mean flow, showing intra- to interannual variability but no detectable decadal trend in the strength of the deep and near-bottom flow out of the Labrador Sea.
13. Camille Marini: On the links between the Atlantic Multidecadal Oscillation and the AMOC Streamfunction (download) (C. Marini, C. Frankignou) (UPMC-LOCEAN FR)
The Atlantic Meridional Overturning Circulation (AMOC) is a key component of the global climate system. Observations remain far too sparse to directly study its variability, since it is a large scale circulation, involving deep currents and having a long time scale. A candidate to assess the decadal and multidecadal variability of the AMOC may be the Atlantic Multidecadal Oscillation (AMO) (ie, Latif et al. 2004, Knight et al. 2005), which can be defined as the time series of linearly detrended Sea Surface Temperature (SST) anomalies, averaged over the North Atlantic basin and filtered with a cutoff period of 10 yr. Nevertheless, the AMO is also infuenced by other climate modes, such as the El Niño Southern Oscillation, or externally forced variations.
In order to find the AMO part related to the AMOC variability, we use a dynamically based filter to remove the ENSO signal and the global trend. Assuming that SST anomalies are well approximated by a linear multivariate Markov process, the Linear Inverse Modeling approach is used to decompose the SST field into a global trend- and ENSO-related part and a global trend and ENSO-unrelated part (Penland and Matrosova 2006, Compo and Sardeshmukh 2010). Using the latter, a "filtered" AMO is computed and is likely to be a better proxy of the AMOC variability than the raw AMO.
We perform this analysis using HadISST reanalysis and output of global climate models (IPSLCM4 and CCSMv3), in which the AMOC is provided. In these two models, correlations between the AMOC and the "_ltered" AMO are slightly improved.
This is a work in progress and we aim at applying this approach to other reanalysis datasets and other climate models.
14. Odd Helge Otteraa: External forcing as a metronome for Atlantic multidecadal variability (download) (O. Otteraa, M. Bentsen, H. Drange, L. Suo) (UiB Bergen NO)
Instrumental records, paleo data and climate modelling show that multidecadal variability is a dominant feature of North Atlantic surface temperature variations. It has been suggested that these variations are internally driven, related to multidecadal fluctuations in the Atlantic thermohaline circulation. However, the governing mechanisms are still uncertain. Identifying the underlying physical mechanisms and exploring the spatial and temporal characteristics of natural variability modes are essential for understanding the observed variability and for addressing the potential for climate predictions.
We have examined the multidecadal variability in Bergen Climate Model by analyzing model simulations covering the last 600 years including the most important anthropogenic and external forcing agents during this period. The simulated NH temperatures and Atlantic SSTs over the last 600 years show many similarities with instrumental records and multi-proxy reconstructions on multidecadal time scales, highlighting the important role of external solar and volcanic forcing agents for such variability. In particular, our results suggest a causal link between tropical SSTs and the applied external forcing. Furthermore, in the idealised case of no changes in the radiative forcing of the Earth (i.e. pre-industrial control simulation), we find that basinwide multidecadal Atlantic SST variations are primarily controlled by variations in the Atlantic meridional overturning criculation (AMOC). However, when the external forcings are included we find a lagged out-of-phase relationship, with positive Atlantic SST anomalies in periods with weaker AMOC and vice versa. Consequently, there is not a one-to-one relationship between the AMOC and the basinwide Atlantic SSTs, in contrast to what is commonly assumed. We also find that interdecadal variations in the North Atlantic Oscillation (NAO), through its dominant buoyancy forcing in the Labrador Sea, influence the phasing of the multidecadal variability of AMOC. In particular, our results indicate that large volcanic eruptions play an important role in forcing the AMOC. They influence the upper ocean conditions at high northern latitudes through their direct radiative cooling, but in our simulations also through their clear tendency for inducing a positive phase of the NAO. The demonstrated role of variations in the external forcing give new insight into the causes and timing of the Atlantic multidecadal variations and, consequently, for the potential for developing decadal prediction schemes for the Atlantic climate.
15. Trond Dokken: Multidecadal variability in properties of inflowing waters to the Nordic Seas and possible connection to sub polar gyre (download) (UiB, Bjerknes Centre NO)
Here we present paleoclimate proxy records that extend the instrumental time series back to 1750 years AD. The records suggest coherent multidecadal variability in the strength of the North Atlantic subpolar gyre and the cyclonic circulation in the Nordic Seas during the past 250 years. The results are based on the carbon isotopes (´13C) of the spring dwelling planktic foraminifera species Globogerina bulloides from a sediment core in the Nordic Seas recording property changes in the Atlantic water masses. To complement these data the census counts of the Arctic frontal species Turburotalita quinqueloba from three records obtained in transect north of the Faroe Islands are used to monitor Ekman driven variability in the Arctic front position in the Nordic Seas. We show that these dynamical changes have been tightly coupled to the strength of the North Atlantic Oscillation (NAO) (Hurrell, 1995) on multidecadal timescales. Using proxy data validated against instrumental data sets, we infer that variations in the subpolar gyre are important for the properties of the North Atlantic inflow to the Nordic Seas on multidecadal timescales and in turn for the circulation pattern in the Nordic Seas, whether there is any active overturning circulation or a more enhanced horizontal (i.e. cyclonic) circulation in the Nordic Seas. (Upcoming publication: Reconstructed multidecadal variability of the North Atlantic
subpolar gyre and the Nordic Seas circulation pattern coupled to the North Atlantic Oscillation by B. Friestad, T. Dokken, Ø. Lie, T. Furevik)
16. Jin Ba: North Atlantic Multi-decadal Variability Simulated in CGCMs (download) (J. Ba, N. Keenlyside, W. Park, E. Hawkins) (IFM GEOMAR DE)
Several climate models, e.g. CMIP3, have in common that the Meridional Overturning Circulation (MOC) varies on multi-decadal timescale. Understanding the causes of this Atlantic Multi-decadal Variability (AMV) is a key to address its role regarding the subpolar heat transport. On the other hand, models show also a large spread in the role of the North Atlantic Oscillation (NAO) in air-sea interaction processes.
In the Kiel Climate Model (KCM), MOC shows a typical 60-years period and leads corresponding sea surface temperature (SST) variations. Because three different convection regions, Greenland-Iceland- Norwegian (GIN) Sea, Irminger and Labrador Sea, play different roles on MOC variability, there is no unique role of the NAO. Mixed layer depth variability in the Irminger Sea leads MOC by 15 years, with salinity being the larger contributor to density. Correspondingly, Subpolar Gyre (SPG) strength is also leading MOC by 15 years. We propose that dimensional temperature and salinity joint EOF-analysis is a useful tool for AMV studies and demonstrate that a strong relationship between MOC and SPG exists in the dominant principal components when applied to KCM.
17. Matthew Menary: Centennial variability in GCMs (download) (Met Office UK)
A mechanism contributing to centennial variability of the Atlantic Meridional Overturning Circulation (AMOC) is tested robustly with multi-millennial control simulations of several coupled general circulation models (CGCMs). Significant AMOC variability on time scales of around 100 years is simulated in these models. The centennial mechanism links changes in the strength of the AMOC with oceanic salinities and surface temperatures, and atmospheric phenomena such as the Intertropical Convergence Zone (ITCZ). A comparison with a high resolution paleo-proxy for Sea Surface Temperatures (SSTs) north of Iceland over the last 4000 years, also linked to the ITCZ, suggests that this mechanism may also be detectable in the real world.
18. Ed Hawkins: THC hysteresis in a coupled GCM (download) (UREAD UK)
Abstract: We demonstrate the existence of THC hysteresis to freshwater hosing in a coupled GCM (FAMOUS). The existence of this bistable THC regime can be diagnosed by the sign of the freshwater transport by the overturning circulation at the southern boundary of the Atlantic. This provides the potential to determine the stability of the THC through present day observations.
19. Bert Wouters: Preliminary validation of the EC-EARTH decadal experiments (download) (KNMI NL)
KNMI's contribution to THOR CT 4 is focused on the decadal predictability of
the overturning circulation and its associated climate variables, for which the EC-Earth v2.2 model is used. The model configuration used consists of the ECMWF IFSc31 model at T159/L62 resolution, the NEMO2 ocean model at 1° resolution and the LIM2 sea-ice model. Analysis of a ~500 year run with pre-industrial conditions shows an overturning circulation dominated by decadal variability, indicating the potential predictability. To assess the actual skill in predictability of the model, a hindcast experiment is carried out. Each 5 years, starting 1 Nov. 1960, a ten year model run consisting of 5 members is launched. A full initialization method is used, where the ocean is initialized using NEMOVAR re-analysis data and sea-ice is obtained from a forced ocean-only model run. In this presentation, we present the first results of this decadal prediction experiments. As
expected, the model shows a drift in the first years due to the initialization shock from the full initial state, but stabilizes afterwards with excellent agreement between observed and predicted surface temparure at lead times up to 9 years. When we remove the trend, several regions remain where the model proves to be skillful, e.g., the North Atlantic and ENSO region in the Pacific.
20. Susanna Corti: Preliminary results from ECMWF decadal experiments (download) (ECMWF UK)
As part of the THOR WP4.1 activities to explore the potential of predictability of the THC on decadal timescales, in this study we use the ECMWF new coupled forecast system (i.e. IFS/NEMO) in multi-decadal integrations.
The atmospheric model is the IFS cycle 36r1 with a horizontal truncation of TL159 and 91 vertical levels. The ocean model is NEMO V3.0 in the ORCA1 configuration. The sea-ice is randomly been sampled from historical data. The system includes the interannual evolution of global mean annual greenhouse trace gases and anthropogenic aerosols, as well as interannual variations of total solar irradiance. Volcanic eruptions are not included.
The model is initialized from an observed state. As initial conditions for the atmosphere the ERA Interim reanalysis has been used. The ocean conditions have been produced with NEMOVAR, a multivariate 3D-var data assimilation method.
The re-forecasts were started once every five years over the period 1960 to 2005, i.e. in 1960, 1965, and so on. Each 3-member ensemble simulation starts at 00 GMT on the 1st of November of each year and run for 120 months. For the period 1980-2005 7 ensemble members have been run.
A set of three different experiments have been run. The baseline experiment does not have any relaxation or flux correction during the forecast. The model is initialised from the observed state (full initialisation). We will refer to this experiment as “the control experiment”. In the second experiment observed anomalies with respect to the observed mean state have been added to the model climate. We will refer to this experiment as “anomaly initialisation”. The third experiment introduces artificial sources and sinks of momentum and heat in the coupling between the atmosphere and the ocean in order to avoid a model drift. This experiment will be referred to as “flux correction”.
We carried out a very preliminary and general diagnostics for the ocean and a more complete diagnostics for the atmosphere which includes an evaluation of the skill of the model in reproducing the observed coupled teleconnection patterns and the leading modes of the interannual extratropical variability in the atmosphere. An assessment of the extent to which atmospheric variables are skilfully predicted in the forecast range from one to ten years has been carried out as well.
We found a predictable signal (in terms of anomaly correlation coefficient score) over the North Atlantic and (partially) over continental Europe over the forecast range of 2-5 and 6-9 years. The flux corrected integrations exhibit an higher predictability and better climatology in the Tropical Pacific and in all the other regions “teleconnected” to this area, like for example the North America. More in general, flux correction helps in reproducing more realistic ocean-atmosphere couple modes of variability. More diagnostics is necessary to assess the state of the ocean and the predictability of the overturning circulation.
21. Jürgen Kröger: Impact of Different Ocean Reanalyses on Decadal Climate Prediction (download)
(J. Kröger, W. Müller, J. von Storch) (MPI-M DE)
We study the impact of several ocean state estimates on decadal predictability in one particular forecast system. Decadal or near-term climate prediction needs to take into account both external forcing and internal climate variability. Much of the memory of variations on decadal and longer time-scales in the Earth system is thought to lie in surface and subsurface layers of the ocean. Hence, in a forecast system accurate initialization of the ocean is crucial when aiming at better skill with respect to pure climate projection efforts. Our forecast system is a recent version of the Earth system model from the Max Planck Institute for Meteorology (MPI-M) in Hamburg. Initial conditions for our forecast system stem from three different ocean state estimates (GECCO, SODA, ECMWF-ORA-S3).
Our forecast procedure follows two steps. First, anomalies of the observational estimates are assimilated into our coupled model. Second, the assimilation runs are then used to initialize every year 10-year-long hindcasts/forecasts from 1960 to 2001. In order to assess the suitability of the particular ocean state estimates, prediction skill is computed for various climate parameters such as North Atlantic (NA) sea surface temperature (SST) and
upper-levels ocean heat content (OHC). In addition, we compare variations in the Atlantic meridional overturning circulation (MOC) within the model experiments, that is, between assimilation run and respective hindcasts of each reanalysis product ('potential prediction skill'). The hindcasts of NA SST up to ten years ahead reveal comparable prediction skill in all experiments. In terms of MOC predictability, the SODA-initialized runs perform worse than the other two initializations, lacking skill after 1 to 2 years. Based on MOC prediction alone, it is not possible to establish a clear preference for either GECCO or ORA-S3. When we compare hindcasts of NA OHC, on the other hand, ECMWF-ORA-S3-initialization proves to be the best for our decadal forecast system, providing superior skill up to ten years ahead.
22. Nick Dunstone: Impact of initialisation of the Sub-Polar Gyre region on multi-annual predictions of the THC and the tropical Atlantic atmosphere (download) (Met Office UK)
Using idealised model experiments we show that predictions of 5-year mean surface air temperature, precipitation and wind shear in the Atlantic main hurricane development region (MDR) has significant skill for lead times of up to four years. Furthermore, tropical storm frequency and the position of the inter-tropical convergence zone are also skilfully predicted on these timescales. We perform additional experiments that withhold data in different parts of the ocean and identify the north Atlantic sub-polar gyre as the key region for driving the skill in the MDR in our model.