The Southern Ocean carbon cycle response to historical climate change

This page is part of the topic Antarctic climate and environment change in the instrumental period

Introduction

The Southern Ocean, with its energetic interactions between the atmosphere, ocean and sea ice, plays a critical role in ventilating the global oceans and regulating the climate system though the uptake and storage of heat, freshwater and atmospheric CO₂ (Rintoul et al., 2001; Sarmiento et al., 2004a). The Southern Ocean is dominated by the eastward flowing ACC. The surface of the ACC is characterized by a northward Ekman flow creating a divergent driven deep upwelling south of the ACC and convergent flow north of the ACC. In the southern part, the upwelling of mid-depth (2-2.5 km) water to the surface provides a unique connection between the deep ocean and the atmosphere while in the northern part the downwelling provides a strong connection to water masses that resurface at lower latitudes e.g. Sarmiento et al. (2004a). These connections make the Southern Ocean extremely important in controlling the storage of carbon in the ocean and a key driver in setting atmospheric CO₂ levels (Caldeira and Duffy, 2000).

CO₂ fluxes in the Southern Ocean

4.55 The annual cycle of epCO₂ (pCO₂^ocean-pCO₂^atm; i.e. negative/positive = atmospheric CO₂ sink) in the Southern Ocean for the regions 40°-50°S (black line) and 50°-62°S (dashed line; Takahashi et al., 2009). The Sub-Antarctic zone (40°-50°S) acts a permanent CO₂ sink but at higher latitudes the ocean acts as an atmospheric sink during summer and a source during winter (Metzl et al., 2006).

The Southern Ocean carbon cycle response uptake can be described as a combination of seasonal and non-seasonal variability. In the Southern Ocean seasonal variability is the dominant mode of variability, and the mechanisms that drive the air-sea CO₂ fluxes are well known but their magnitudes remained poorly constrained (Metzl et al., 2006 and references therein). The seasonal cycle of CO₂ uptake is a complex interplay between the biological and physical pumps and can be described both in terms of its magnitude and phase. A recent climatological synthesis of more than 3 million measurements of surface pCO₂ measurements provides important insights into Southern Ocean behaviour at the seasonal timescales (Takahashi et al., 2009). During the austral summer biological production reduces surface ocean pCO₂ through photosynthetic activity and then exports part of this organic matter to the deep ocean. This reduction in surface pCO₂ is offset by the changes in the physical pump that reduce the capacity of the surface water to store CO₂ through upper ocean warming, which lowers solubility, thereby increasing the surface ocean pCO₂. The net result of this competition between the biological and physical pumps is that the Southern Ocean acts a sink of atmospheric CO₂ in the summer in the sub-Antarctic zone (SAZ; nominally 40°S – 50°S) and south of the Polar Front (PF; ~50°S; Figure 4.55). During the austral winter, a picture of two distinct zones separated by the PF is evident. South of the PF relatively little biological activity occurs in winter (deep mixing and light limitation); therefore surface pCO₂ values are set by the competition within the physical pump between deep winter mixing bringing up CO₂ water from the carbon rich deep ocean, leading to increased pCO₂ surface levels, and a cooling that increases the ability of the surface waters to store CO₂. As the deep winter mixing dominates in this region a net out-gassing of CO₂ to the atmosphere occurs. North of the PF during the austral winter the same cooling and mixing processes that occur further south exist, but with the addition of biological activity during the period, albeit at a reduced rate, reducing the pCO₂ in surface waters. The result of the combined responses of the biological and physical pumps means this region acts as a weak sink of atmospheric CO₂ during the austral winter.

When the observed winter and summer fluxes are integrated, the annual mean uptake is small south of 50°S (about -0.08 PgC/yr); conversely the SAZ (40°S-50°S) behaves as a strong sink (-0.74 PgC/yr) (Takahashi et al, 2009; see comments below). The response of the SAZ is consistent with other studies that suggest the SAZ is also a strong CO₂ sink approaching -1 PgC/yr (McNeil et al., 2007; Metzl et al., 1999) and an important region of mode and intermediate water formation and transformation. In comparison to the total uptake of 2 GtC/yr for the global ocean, the Southern Ocean south of 40°S takes up more than 40% of the total uptake. Note in these calculations we have used the gas transfer coefficent of Wanninkhof (1992) with the dataset of Takahashi et al., (2008).

4.56 Annual mean uptake of air-sea CO₂ fluxes as calculated from OPA/PISCES 1990-1999 (Lenton et al., 2006) and that from the new climatology of Takahashi et al. (2009). The sub-Antarctic region (40-50°S) represents a strong sink (blue colors), whereas south of 50°S, large regions act as a CO₂ source for the atmosphere (red).

Significant progress has been made in recent years in simulating the annual mean uptake of CO₂ by the Southern Ocean as precursor to understanding interannual to decadal variability. Figure 4.56 shows that the spatial pattern of the annual mean uptake is well represented in comparison with that simulated from the current class of ocean biogeochemical models (e.g. OPA/PISCES model; Aumont and Bopp, 2006)). Although different models do contain different representations of the magnitude of the seasonal cycle, the phase between each model and observations shows good agreement. This suggests that seasonal processes that drive this variability are well captured, and that as a result, we are in a better position to explore changes at interannual and longer timescales.

Historical Change - Observed Response

The Southern Ocean has undergone significant changes in response to climate changes; such as a net increase in heat and freshwater fluxes, and a poleward movement and intensification of winds e.g. Thompson and Solomon (2002). The major driver of these changes has been the SAM that is associated with changes in the Antarctic Vortex, primarily in response to depletion of stratospheric ozone and increasing atmospheric greenhouse gas concentrations (Arblaster and Meehl, 2006). While the SAM is a significant driver of variability it cannot explain all the variability present, as it has both linear and non-linear interactions with other climatic modes such as ENSO and Indian Ocean Dipole (IOD) that drives diverse responses in different regions. These physical changes impact directly on the physical pump and to a lesser extent on the biological pumps, therefore on the concentration of CO₂ in surface waters and the magnitude of both uptake and export of CO₂ from the atmosphere to the deep ocean.

In the Southern Ocean the interannual to decadal changes in biological production, ocean dynamics and thermodynamics that drive oceanic pCO₂ and air-sea CO₂ exchanges remain poorly understood and very undersampled. Decadal and interannual variations have been observed at high latitudes, but at only very few locations and during the austral summer (Jabaud-Jan et al., 2004; Brévière et al., 2006, Borges et al., 2008). Although these analyses provide important information on the response of the ocean to climate variability in the Southern Hemisphere, there is no clear detection of the decadal trends of oceanic CO₂ and associated air-sea CO₂ fluxes, unlike the situation in the north Atlantic or the north and equatorial Pacific where there are long-term time series of such data (e.g. Bates, 2001; Feely et al., 2002). Takahashi et al. (2009) have recently constructed a pCO₂ data synthesis, from which a significant increase of oceanic pCO₂ during winter has been calculated, about +2.1±0.6 µatm/yr, which is close to or faster than the growth rate in the atmosphere (1.7 ppm/yr) over the period 1986-2007.

4.57 Annual mean trends of temperature normalized fCO₂ in 4 regions of the South-Western Indian Ocean (based on summer and winter observations in 1991-2007). The open bars indicate the growth rates estimated for summer and black bars for winter. Standard errors associated to each trend are also indicated. The dashed line indicates the atmospheric CO₂ annual growth rate (figure reproduced from Metzl, 2009).

Repeat underway measurements of surface pCO₂ have been made regularly in the Southern Indian Ocean since the 1990s (e.g. Metzl et al.,1999). Although these measurements are quite often confined to regions where ships travel to resupply Antarctic and sub-Antarctic bases, they represent a valuable timeseries for exploring the evolution of the surface ocean. A recent study by Metzl (2009; Figure 3.68) in the South Western Indian Ocean calculated surface trends of pCO₂ between 1991-2007 and showed that oceanic pCO₂ increased at all latitudes south of 20°S (1.5 to 2.4 µatm/yr depending the location and season). More specifically, at latitudes of less than 40°S, they determined that oceanic pCO₂ increased faster than in the atmosphere since 1991, suggesting the strength of the oceanic sink decreased. In addition, when pCO₂ data are normalized to temperature, removing the effect of solubility on CO₂, this analysis showed that the system is increasing much faster in the winter than in the summer (Figure 4.57). These results suggest that the increase may be due to changes in ocean dynamics, given that the largest response occurs in the austral winter, when the winds are strongest. In the recent period (since the 1980s) the increase of pCO₂ appeared to be faster compared to the trends based on historical observations from 1969-2002 (Inoue and Ishii, 2005), suggesting that the Southern Ocean CO₂ sink has continued to evolve in response to climate change.

4.58 Sea-air CO₂ flux anomalies in the Southern Ocean (PgC/y, >45°S) based on atmospheric CO₂ data and inversed transport model (on top) and a global biogeochemical ocean model (bottom). Compared to experiments that do not take into account the climate variability (in red), both approaches suggest a stabilization or reduction of the ocean CO₂ sink since the 1980s (Le Quéré et al., 2007).

As oceanic pCO₂ in recent years has been observed to be increasing close to, or faster than in the atmosphere, the signature of these changes in atmospheric CO₂ data should be detectable, as has been observed in the Equatorial Pacific during ENSO events (e.g. Peylin et al., 2005). At latitudes south of 40°S the ocean has a very large surface and it is expected that continental carbon source/sink variability has a low imprint in atmospheric CO₂ records (compared to the tropics and north hemisphere). This is clearly seen in the CO₂ record at La Nouvelle Amsterdam Island (in the South-Indian Ocean), for example, where the seasonality of atmospheric CO₂ is very low. A recent study by Le Quéré et al. (2007) using a combination of atmospheric observations and inverse methods reported that in the period 1981-2004, the strength of the Southern Ocean CO₂ sink (south of 45°S) was reduced (Figure 4.58). Although this result remains controversial (e.g. Law et al., 2008), it does suggest that the observed increase in oceanic pCO₂ acts to reduce the strength of the air-sea CO₂ gradient ( pCO₂) and this in turn translates to reduction in the strength of the Southern Ocean CO₂ sink. This result is significant as it was expected in response to strengthening air-sea CO₂ gradient that the Southern Ocean CO₂ uptake would increase. These results are also consistent with a number of recent modeling studies, that have also suggested a decrease in uptake in the last decades and which attributed the upper ocean pCO₂ values to increases in the wind speed increasing the ventilation of carbon rich deep waters e.g. Lenton and Matear (2007).

Historical Changes – Simulated View

4.59 Annual-averaged Southern Ocean uptake of: a) total carbon; b) natural carbon; c) anthropogenic carbon. The different experiments use the following colour coding, total experiment (black line), 1948 (red line), wind stress (tau, green line), heat flux (blue line) and freshwater flux (cyan line).

To explore how changes in Southern Ocean air-sea CO₂ fluxes have responsed to historical climate change between 1948-2003. Matear and Lenton (2008) used a biogeochemical ocean model driven with observed changes (NCEP R1; Kalnay et al., 1996). They explored how the total carbon cycle as well as the natural and anthropogenic carbon uptake responded to the observed increases in windstress, heat and freshwater fluxes (Figure 4.59). Their results show a complex picture: when only either heat or freshwater fluxes increase, the total uptake is slightly greater than the total CO₂ flux response (with all fields increasing). In contrast, when only the wind speed increased, the total uptake was less than the total response. In addition, the anthropogenic response was always much smaller than the natural carbon cycle response and hence the natural response dominated the total response. The natural carbon sink dominates the total response over the last 50 years for two reasons: (i) because wind stress changes are larger than the corresponding changes in heat and freshwater flux, therefore changes in solubility are dominated by changes in ocean dynamics; and ii) the CO₂ gradient between the atmosphere and the ocean due to anthropogenic emissions is strong enough over this period to counter the increased CO₂ in surface waters caused by winds bringing up water rich in dissolved inorganic carbon from the deep ocean. Although there is some question of the validity of the changes in the pre-satellite era (1948-1979; Marshall, 2003), the largest changes occur in the later period 1979-2003, as seen in Figure 4.58.

As reanalysis products used in Matear and Lenton (2008) are assimilated products, all climate modes/variability are represented. Other studies using different ocean models and experiments to explore the response of the Southern Ocean air-sea CO₂ flux to the SAM alone e.g. Lenton and Matear (2007), Lovenduski et al. (2007) are very consistent with the view determined using all the superposition of all the climatic modes. This is not surprising given that these studies suggest that more than 40% of the total variance in CO₂ flux is be explained by the SAM in the recent period Lenton and Matear (2007).

Over the period only a small increase in primary production and export production is evident, suggesting only a weak link between atmospheric forcing and export production, despite the increased upwelling of deep waters in response to increased winds (i.e. supplying also macro and micro nutrients). The largest response was on the northern boundary of the High Nutrient Low Chlorophyll (HNLC) area of the Southern Ocean. Over the rest of the Southern Ocean, very little response was seen, demonstrating only a weak link between atmospheric forcing and production.

The increased ventilation of the Southern Ocean from simulations does not only alter the concentration of upper ocean CO₂ and the air-sea CO₂ fluxes, it also alters the carbonate chemistry of the upper ocean. These changes in carbonate chemistry affect the ability of the ocean to take up atmospheric CO₂, through changes in the Revelle factor (Revelle and Suess, 1957) and through the lowering of seawater pH. This pH reduction, or ocean acidification, reduces the ability of organisms that use calcite to build shells, potentially adversely impacting the marine ecosytem (Feely et al., 2004). In response to ocean acidification, a key carbon parameter is the aragonite saturation state (Ω_A), which influences the rate of calcification of marine organisms (Riebesell et al., 2000; Langdon and Atkinson, 2005). Simulations show that the observed increases and variability in heat, freshwater fluxes and in particular wind stress in the last 50 years, has moved the saturation horizon closer to the surface (Orr et al, 2005), potentially already impacting on the ecosystems in the Southern Ocean, although we do not yet have the observational evidence to support this hypothesis.

Changes in CO₂ inventories

4.60 Anthropogenic carbon concentrations (colour scale, in umol/kg) derived using a back-calculation method in the Indian Ocean sector of the Southern Ocean between Antarctica (left) and Africa (right) (redrawn from Lo Monaco et al., 2005).

In the previous sections, we focused on the changes of the air-sea CO₂ fluxes as observed and simulated over the last 50 years. When discussing the capacity of the ocean to reduce the impact of climate changes, the change in anthropogenic CO₂ in the water column since the pre-industrial era must be evaluated. This is important not only to estimate the global ocean’s capacity to absorb anthropogenic CO₂ emissions, but also to detect the changes in carbonate saturation levels and the potential increase of ocean acidity, especially in the Southern Ocean (Feely et al., 2004; Orr et al., 2005). It has been estimated that in recent decades 30% of the total uptake since the preindustrial period has been taken up by Southern Ocean mode waters (Sabine et al., 2004). The anthropogenic CO₂ in the ocean (C_ant) cannot be directly measured, but under several assumptions, it can be derived from in-situ observations. This was first suggested by Brewer (1978) and Chen and Millero (1979), and in the last ten years several data-based methods have been investigated at regional and global scales (see a review in Wallace 2001; Sabine et al., 2004; Waugh et al., 2006; Lo Monaco et al., 2005). Comparisons of data-based methods (Lo Monaco et al., 2005) clearly show that all methods converge to estimate large inventories associated with mode and intermediate waters (Figure 4.60). Meanwhile in Southern Ocean uptake south of 50°S uncertainties in C_ant still exist, but these differences appear to reflect techniques that have been recognized to underestimate anthropogenic carbon in deep and bottom waters along the Antarctic coast (Lo Monaco, personal communication).

Concluding Remarks

The Southern Ocean plays a critical role in the uptake of atmospheric CO₂, accounting for more that 40% of the annual mean CO₂ uptake. Modelling and observational studies show that the Southern Ocean has undergone significant changes in the last 50 years; these views appear to be converging towards a coherent view. The largest change has been a reduction in the total CO₂ uptake in recent decades in response to the observed changes in climatic forcings, particularly changes in wind speed. The increased wind speed drives strong changes in the physical carbon pump, specifically through ocean dynamics, rather than through changes in solubility or in the Revelle factor; this view is further reinforced by only a weak response in the biological pump. In the future, in response to climate change, both the biological and physical pumps are expected to be impacted (see Section 5.8 for details).