Estimating trends of Atlantic meridional overturning circulation from long-term hydrographic data and model simulations
- 664 Downloads
The ocean meridional overturning circulation (MOC) plays a central role for the climate in the Atlantic realm. Since scenarios for future climate change indicate a significant reduction of the MOC under global warming, an assessment of variations and trends of the real MOC is important. Using observations at ocean weather ship (OWS) stations and along oceanic sections, we examine the hydrographic information that can be used to determine MOC trends via its signature in water mass properties obtained from model simulations with the climate model ECHAM5/MPI-OM. We show that temperature trends at mid-latitudes provide useful indirect measure of large-scale changes of deep circulation: A mid-depth warming is related to MOC weakening and a cooling to MOC strengthening. Based on our model experiments, we argue that a continuation of measurements at key OWS sites may contribute to a timely detection of a possible future MOC slowdown and to separate the signal from interannual-to-multidecadal MOC variability. The simulations suggest that the subsurface hydrographic information related to MOC has a lower variability than the MOC trend measured directly. Based on our model and the available long-term hydrographic data, we estimate non-significant MOC trends for the last 80 years. For the twenty-first century, however, the model simulations predict a significant MOC decline and accompanied mid-depth warming trend.
KeywordsOcean circulation Ocean weather ship stations Atlantic meridional circulation Temperature trends
A major unknown in assessing global change is the future fate of the world ocean. Climate model experiments suggest that the Atlantic meridional overturning circulation (MOC) and its associated heat transport may weaken under global warming conditions in the twenty-first century (IPCC 2007). The MOC consists of deep convection from surface cooling at high latitudes, equatorward flow of the resulting deep water, and upwelling of deep water at lower latitudes. Paleoclimate proxy records suggest that massive and abrupt climate changes have occurred in the Northern Hemisphere, especially during and just after the last glacial, with MOC changes as the most plausible mechanism (e.g., Broecker and Denton 1989; Alley et al. 2003; McManus et al. 2004). By analogy with paleoclimatic reconstructions, it is possible that such a MOC perturbation would have an impact on future climate, e.g., by influencing the global cycle of uptake, release, and storage of carbon.
Analyzing hydrographic records between the 1950s and 1990s, it has been suggested that North Atlantic Deep Water (NADW) have steadily become warmer (Levitus et al. 2000; Arbic and Owens 2001) and less saline (Dickson et al. 2002; Curry et al. 2003; Curry and Mauritzen 2005), which could have affected the MOC. However, it is neither straightforward to trace MOC weakening under global warming conditions nor to detect if it is already in progress.
Observation-based estimates of MOC strength are given by Roemmich and Wunsch (1985) and Ganachaud and Wunsch (2000). However, the existing hydrographic measurements are too sparse to obtain a time evolution in the mass budgets. In order to systematically measure the MOC strength, moored instrument arrays have been deployed at transatlantic sections (Hirschi et al. 2003, Baehr et al. 2004). MOC monitoring at 26.5° N was also recently reported in Cunningham et al. (2007) and Kanzow et al. (2007). Of particular relevance to the focus on temperature and salinity responses to MOC changes, Cunningham and Alderson (2007) diagnose the vertical structure of changes in temperature and salinity at 24.5° N and discuss how these changes may be linked to MOC slowdown reported by Bryden et al. (2005). Bryden et al. (2005) used ship-based hydrographic transects along 24.5° N from 1957, 1981, 1992, 1998, and 2004. Missing data and natural climate variations on time scales from days to centuries make MOC estimates extremely difficult (Von Storch and Haak 2008).
Mooring array programs have only recently been established, and their observational record is too short to draw a conclusion about MOC trends since the ocean circulation is also masked by strong variability on subannual to interdecadal timescales (e.g., Cunningham et al. 2007; Dima and Lohmann 2007). As a different approach, Levermann et al. (2005) suggest that the observed changes in sea level can be effectively used for monitoring the strength of MOC. Others propose to use the surface signature in temperature and precipitation (Lohmann 2003; Latif et al. 2004; Hu et al. 2004) for estimating changes in MOC. Disadvantages of these methods may be that other phenomena project on the signature (e.g., north–south gradient or “interbasin” balances) and the derived indices are partly different from MOC trends.
Here, we propose a different approach to use long-term hydrographic data. The World Ocean Circulation Experiment (WOCE) was a component of the World Climate Research Program and remains the most ambitious oceanographic experiment undertaken to date. The field phase of the project lasted from 1990 to 1998. On the completion of WOCE, a global study of the Earth’s climate variability and predictability (CLIVAR) is being actively pursued. In the present paper, we trace circulation trends based on long-term quality-checked observational temperature and salinity data in combination with climate scenario simulations. The climate model experiments are used to derive characteristic temperature and salinity patterns of MOC weakening along hydrographic sections (WOCE-CLIVAR) and station data at weather ships (OWS) in the North Atlantic. The ocean weather station idea originated in the early days of radio communications and transoceanic aviation. Oceanographic observations were recommended for weather ships almost from the start ∼1910 (e.g., temperature and salinity profiles).
As we find out, the information of the whole water column at these stations is necessary to detect ocean circulation trends and to separate them from natural climate variability. These characteristic patterns are used to estimate the past 100 years of MOC trends, which is important to bring recent data and predicted changes into a long-term context. The use of the long-term observations provided by the OWS bring the more recent data related to MOC into the long-term context.
The simulations were carried out for the period 1860–2100 in the framework of the fourth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) published in 2007. The model was initialized from the end of a pre-industrial control experiment and then forced with observed greenhouse gas and aerosol concentrations for the period 1860–2000 (20C Experiment). Thereafter, greenhouse gas and aerosol concentrations followed different scenarios (A1B, A2, and B1 Experiments). A summary of the forcing functions for the scenarios A2, B1, and A1B can be found in the IPCC report (IPCC 2007). Experiment 20C has been continued with fixed emissions for the twenty-first century at values from year 2000 AD.
2.3 Hydrographic sections and Ocean Weather Ship stations
For the sections in the Atlantic Ocean (Fig. 1), repeated measurements across major current systems and interbasin channels have been made in WOCE and CLIVAR in order to document interannual and longer period changes in circulation. In our analysis, section A17 has been extended to the northern North Atlantic in relation to the WOCE-CLIVAR definition in order to include more information in the meridional direction. The detailed structure of the section is, however, of minor importance for our results. Other data sets, e.g., satellite altimetry and ARGO profiling floats, are not considered in our analysis.
The hydrographic database (Curry 2002) includes long-term OWS data in combination with more recent section and station measurements. For our analysis, we have used the quality-checked data (Curry 2002; Curry et al. 2003). For OWS E, we applied an area average over (50–46° W; 33–37° N) covering the period 1910–1993; for OWS S, an area average over (66–62° W; 30–34° N) covering the period 1914–1989. In order to examine the regressions and ocean circulation (re-)constructions, we interpolate the station data to the 40 vertical model levels.
3.1 Model climate and scenario integrations
The time evolution of the Atlantic maximum meridional overturning stream function is shown in Fig. 4b (annual mean values). The seasonal MOC strength has a standard deviation of appr. 3 Sv in our simulation, but in the following, we concentrate on the annual mean values. The meridional overturning for A1B shows a weakening of 5 Sv from year 2000 to 2100 due to warming and freshening of high-latitude surface water. We find, though the large-scale MOC structure seen in Fig. 2c is not affected, considerable changes in the deep water formation regions and their relative role (not shown). It is interesting to note that the run 20C shows a very moderate MOC reduction for the twenty-first century, but an in-depth discussion is not within the scope of this paper. For our further analyses, we concentrate on the SRES A1B scenario (IPCC 2007), which is a scenario with a moderate climate response (Fig. 4a) and with an overturning response similar to that of A2 and B1 in the twenty-first century (cf. Fig. 4b).
3.2 Regression analyses
The temperature regression at OWS S has three minima at 500, 1,500, and 2,500 m, indicating cold conditions when there is a strong MOC (Fig. 5a). The pattern is close to the baroclinic first mode as described by Cunningham and Alderson (2007). The changes in the 300–800 m are related to the thermoclince, the second peak between 900 to 1750 m is related to intermediate water, 1,750 to 2,500 m is related to upper North Atlantic Deep Water, and below that is lower Atlantic Deep Water. A decrease in MOC is related to temperature increase (Fig. 5a), which is dominated by isopycnal heave. The temperature trend analysis (Figs. 5b, 6b) for the period 2000–2100, when the MOC weakens, shows a warming in the upper 2,000 m. The temperature variability is relatively small at mid-depth (700–2,200 m), which could be used for an optimized signal-to-noise analysis (not shown). In the upper 1,000 m, temperature variability is highly dominated by interannual-to-decadal variability (not shown).
The section A17 temperature trend analysis (Fig. 7b) shows a general warming, as expected from the increasing atmosphere greenhouse gases. The warming trend at intermediate depths, which is accompanied by a cooling in the downward branch of the North Atlantic Deep Water, is associated with Atlantic MOC weakening during this period. The A17 salinity trend analysis (Fig. 7d) is determined by enhanced evaporation in the subtropics in conjunction with high-latitude freshening and projects not directly onto the MOC trend.
3.3 Constructing the overturning trends
The analysis reveals that the temperatures at OWS E and OWS S, as well as temperatures at section A17, provide useful indirect measure of large-scale changes of deep circulation, with a standard deviation even smaller than the simulated MOC. The reason for the smaller standard deviation could be that the MOC is affected by the atmospheric noise (e.g. the North Atlantic Oscillation), whereas the subsurface quantities see already a reduced noise integrated by oceanic processes (e.g., through vertical mixing, advection). The atmospheric forcing can contribute to decadal MOC variations via Labrador Sea Water (Haak 2004) and mask longer-term MOC trends (Delworth and Dixon 2000; Vellinga and Wood 2004).
In our analysis, we extend the analysis to the last century but rely of course on the model’s ability to simulate the hydrographic changes with MOC. We admit that the modeled and observed temperature structure is not identical. We find, however, that the structure of temperature changes with MOC is a robust feature in our simulations and that the principal temperature structure (mid-depth warming for MOC reduction) are similar when looking at different model simulations (Rühlemann et al. 2004; Drijfhout and Hazeleger 2006) and even model intercomparison using the same forcing function (Stouffer et al. 2006).
Our integrated temperature indicator of MOC trend (based on the whole water column) has an error of less than 1 Sv within one century (Fig. 11), although the regression error at each depth can be higher. From Fig. 12, we derive the conclusion that there is no significant MOC trend over the last century. However, if one looks at shorter periods like 1960–2000, one may come to the conclusion of a moderate MOC decline, which could be part of decadal variability. This would be consistent with Cunningham and Alderson (2007), indicating mid-depth warming for that period (their Fig. 2a) caused by an upward motion of the isopycnals. The hydrographic change in A05 (Fig. 8) is also consistent with an increase in salt in the west and reduced salinity in the east for the last decades (Cunningham and Alderson 2007). The sections A1, A2, and A3 turn out to be less useful for constructing MOC trends and could be more related to with the changes in thermocline thickness implied by the temporal changes in Ekman pumping (Leadbetter et al. 2007).
For the time when hydrographic data is available, our method should be combined with direct calculations of the MOC strength or direct estimates of mass transports based on velocity measurements and hydrographic fields (e.g., Hirschi et al. 2003; Bryden et al 2005; Baehr et al. 2004, 2007a, b; Cunningham et al. 2007), as well as indirect methods (Latif et al. 2004; Hu et al. 2004; Levermann et al. 2005). Global inverse calculations from hydrographic observations or other volume transport estimates typically yield a standard error of 2–6 Sv in the MOC strength (Roemmich and Wunsch 1985; Ganachaud and Wunsch 2000; Ganachaud 2003; Smethie and Fine 2001; Bryden et al. 2005). Furthermore, detecting MOC changes needs to account for both variability and observational uncertainty.
One should mention two principal problems for MOC construction. The noise is relatively high for intra-annual variations; therefore, the use of annual mean values of stations and arrays is preferred. It could be that the much more noisy ‘real’ observed station and cross-section data would cause a detection of MOC slowdown to occur much later than indicated by the idealized model case. The variance of the hydrography at the OWS sites is lower in the model as compared to the measured data (not shown). We speculated that this could be due to not continuous measurements for the real observations and/or model deficits. In our case, it appears, however, that the MOI values based on modeled and measured hydrography at OWS E and OWS S have almost the same variance (Fig. 12). The usual problem of calculating MOC from observations (hydrographic and velocity measurements) is transferred in our approach to the models’ ability to simulate the temperature and salinity responses to MOC changes. It would be interesting to check the robustness in other IPCC runs showing also a MOC decline for the next century (IPCC 2007).
The MOC trends for the twenty-first century are only reconstructed correctly in 10% of the cases. From a statistical point of view, one may argue that this is a pure mathematical artifact. However, we underline that the temperature response due to MOC trend has a clear physical reason: Less advection of cold high-latitude water in the North Atlantic causes a mid-depth warming in the tropical and subtropical North Atlantic. Such mechanism has been shown to be valid for MOC reductions in the geological past induced by deglacial meltwater and moderate Holocene sea surface salinity variations (Rühlemann et al. 2004).
The AR4 simulations did not take into account any additional freshwater input from the melting Greenland ice sheet. Mass balance calculations (Huybrechts et al. 2004) estimate an input of about 6 × 10−3 Sv in the year 2100. The most recent estimates tend to be larger (e.g., Rignot and Kanagaratnam 2006), suggesting that freshwater input by 2100 could be considerably larger. Sensitivity experiments (Jungclaus et al. 2006b) showed an additional reduction of the MOC but even for an upper estimate of fresh water forcing; the MOC evolution in the twenty-first century was close to the baseline scenario.
Another difficulty of observational and model-based methods is that the ocean shows transient behavior and small-scale signals. In the MOC reconstructions, which are based on hydrographic fields, one has to implicitly assume that the noisy signals are filtered out when considering monthly means or seasonal averages (when considering observations) or in the case of models that the net effect of eddies are well captured in current non-eddy resolving climate models.
Paleoceanographic proxy records indicate a strong relationship between MOC strength and mid-depth tropical Atlantic temperatures during the last deglaciation and Holocene millennial variability (Rühlemann et al. 2004). In our analysis, we see that a similar relationship helps to detect recent and future changes of the MOC. The temperature indicators suggest no MOC trend for the past 100 years using the existing long-term observations from the OWS. Our use of a spatial (albeit extremely coarse) network of hydrographic observations opens the prospect of MOC analysis beyond single latitude transects.
As implied from our work, improving the ocean observation system is critical for detecting the beginning of any MOC weakening. The almost-century-long data sets from the ocean weather ships put recent direct MOC measurements and possible future MOC changes into a long-term perspective. Therefore, the continuation of these measurements would constitute one important element of a MOC observing system. Furthermore, other quantities like the ocean heat transport can be analyzed in the same way as MOC trends. Drijfhout and Hazeleger (2006) found out that the ocean heat transport depends not only on MOC but also on vertical temperature contrasts, which may compensate for a weaker MOC.
As a logical next step, the assimilation of ocean observing systems provides a dynamical interpolation and synthesis of hydrographic data sets. Unfortunately, the ocean observing systems based on satellite altimetry, ARGO profiling floats, XBT, and CLIVAR hydrography, do not cover many decades, which makes an estimate of MOC trends with respect to the last century and their relation to decadal to interdecadal MOC variability difficult. Alternatively, one may use the information of well-dated, high-resolution proxy data. Lohmann et al. (2004) used high-resolution data from Cariaco Basin (64.67° W, 10.5° N; Black et al. 1999), covering the last 800 years in order to trace variations in MOC in terms of the interhemispheric sea surface temperature signature, which is consistent with the temperature reconstruction for the multiproxy network utilized by Mann et al. (1998). The proxy records show large interdecadal variations with no pronounced trend. A more detailed analysis of the mechanisms related to variations in MOC (Dima and Lohmann 2007) emphasizes a coupled atmosphere–ocean–sea ice mode, including an atmospheric wave-number-one pattern and sea ice transport on multidecadal time scales. It is conceivable that more high-resolution proxy data in sensitive regions may become available in order to trace past MOC variations and trends. In the future, a detailed analysis of the underlying physics influencing the effectiveness of such factors on MOC trends will be tested also in comparison with other coupled climate models.
We thank R. Curry for making the observational data available to us. We are thankful for constructive suggestions by three anonymous referees. The model simulations were done at the German Climate Computing Center (DKRZ) in Hamburg.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Baehr J, Keller K, Marotzke J (2007a) Detecting potential changes in the meridional overturning circulation at 26° N in the Atlantic. Clim Change, http://www.dx.doi.org/10.1007/s10584-006-9153-z
- Curry RG (2002) HydroBase2: a global database of hydrographic profiles and tools for climatological analysis. http://www.whoi.edu/science/PO/hydrobase
- Haak H (2004) Simulation of low-frequency climate variability in the North Atlantic Ocean and the Arctic. Reports on Earth System Science 1, Max Planck Institute for Meteorology, Hamburg, Germany, pp 115Google Scholar
- IPCC (2007) Intergovernmental panel on climate change. The physical basis of climate change (http://ipcc-wg1.ucar.edu/wg1/wg1-report.html)
- Latif M, Roeckner E, Botzet M, Esch M, Haak H, Hagemann S, Jungclaus J, Legutke S, Marsland S, Mikolajewicz U, Mitchell J (2004) Reconstructing, monitoring, and predicting multidecadal-scale changes in the North Atlantic thermohaline circulation with sea surface temperature. J Climate 17:1605–1613CrossRefGoogle Scholar
- Lohmann G (2003) Atmospheric and oceanic freshwater transport during weak Atlantic overturning circulation. Tellus 55A:438–449Google Scholar
- Roeckner E, Bäuml G, Bonaventura L, Brokopf R, Esch M, Giorgetta M, Hagemann S, Kirchner I, Kornblueh L, Manzini E, Rhodin A, Schlese U, Schulzweida U, Tompkins A (2003) The atmospheric general circulation model ECHAM 5. PART I: model description. MPI-Report 349, Max Planck Institute for Meteorology, Hamburg, Germany pp 127Google Scholar
- Rühlemann C, Mulitza S, Lohmann G, Paul A, Prange M, Wefer G (2004) Intermediate depth warming in the tropical Atlantic related to weakened thermohaline circulation: combining paleoclimate data and modeling results for the last deglaciation. Paleoceanography 19, PA1025, doi: 10.1029/2003PA000948
- Stouffer R, Yin J, Gregory JM, Dixon KW, Spelman MJ, Hurlin W, Weaver AJ, Eby M, Flato GM, Hasumi H, Hu A, Jungclaus JH, Kamenkovich IV, Levermann A, Montoya M, Murakami S, Nawrath S, Oka A, Peltier WR, Robitaille DY, Sokolov A, Vettoretti G, Weber N (2006) Investigating the causes of the response of the thermohaline circulation to past and future climate changes. J Climate 19:1365–1387CrossRefGoogle Scholar