Australia: The Land Where Time Began

A biography of the Australian continent 

Southern Ocean Mixed Layer Depths – Assessment in CMIP5 Models: Historical Bias and Response to Forcing

In this paper Sallée et al. report the results of their assessment of the development of the deep water Southern Ocean winter mixed layer in the climate models that participated in the 5th Coupled Model Intercomparison Project (CMIP5). A key to the ventilation of the ocean interior are the deep winter convection regions, and property changes of these regions have been related to changes of climate in many studies. When compared to observations the simulation of these models is consistently too shallow, too light and shifted towards the equator. The shallow bias is associated mostly with an excess of annual-mean freshwater input at the surface of the ocean that over stratifies the surface layer thereby preventing the development of deep water in winter. Contrasting with this, future changes that have been modelled are mostly associated with reduced loss of heat during winter that leads to winter mixed layers that are even shallower. In the Pacific Basin the mixed layers shallow most strongly under future scenarios, and this is associated with a reduction of the volume of ventilated water in the interior. Sallée et al. found that there was a strong state dependency on change of depth of the mixed layer in the future, with models with larger historical mixed layer depths. It was expected that, given the biased shallow in most models, most CMIP5 climate models might underestimate winter mixed-layer shallowing, with implications that are important for the sequestration of heat, as well as gases such as carbon dioxide, and therefore for climate.

At the ocean surface the mixed layer is the gateway for all exchanges between the atmosphere and the ocean. In the Southern Ocean, where intense winds and buoyancy flux extremes lead to the formation of the thickest mixed layers on Earth (de Boyer et al., 2004) this gateway function is especially important. These deep mixed layers provide a conduit a conduit for the sequestration of heat and gases (including carbon dioxide), from the atmosphere into the interior of the ocean (Sabine et al., 2004; Ito et al., 2010; Sallée et al., 2012). Therefore, the assessment of how well the mixed layer of the Southern Ocean is represented in climate models as it can affect the accuracy of future projections.

In the Southern Ocean the deepest mixed layers form in winter directly north of the Antarctic Circumpolar Current (ACC) (e.g., McCartney, 1977; Sallée et al., 2006; Dong et al., 2008). Mode and intermediate waters acquire their physical and biogeochemical properties in the circumpolar band of thick mixed layer prior to being subducted into the ocean interior. The thermocline of the Southern Hemisphere subtropical gyres are then ventilated by these waters (Sallée et al., 2010a). It has long been recognised that mode and intermediate waters are key water masses in the determination of the global distribution and budgets of heat, carbon and nutrients (e.g., Sarmiento et al., 2004; Sabine et al., 2004; Ito at al., 2010; Sallée et al., 2012), as a result of their large thickness and their surface formation. Estimates suggest, in particular, that more than 40 % of the total oceanic anthropogenic carbon has entered the ocean south of 40oS. There are also indications from palaeoclimatic records that in the Southern Ocean a breakdown in stratification contributed to the rise of atmospheric carbon dioxide at the end of the Last Glacial Maximum (LGM) (Toggweiler & Russel, 2008; Anderson et al., 2009). This emphasises the importance of representing accurately the mixed layer of the Southern Ocean so that past present and future climate can be modelled accurately.

A wide variation in the ability to represent deep water mixed layer in the Southern Ocean is exhibited by climate models of the 3rd Coupled Model Intercomparison Project (CMIP3) (Downes et al., 2009, 2010). Various improvements to the parameterisation of mixed layer dynamics have been suggested since then, and some of these have been implemented in the models that contribute to Coupled Model Intercomparison Project Phase 5 (CMIP5).

The presence of surface waters is possibly the most significant characteristic of the mixed layer of the ocean, which is contrasts with the atmospheric boundary layer, which results in both wave breaking and Langmuir circulation at the surface (Noh & Min, 2004). There are a number of turbulent kinetic closure schemes that have been developed with the aim of parameterising this complex physics, which is associated with convection and restratification of the mixed layer. In some models new generation turbulence closure schemes have been implemented (e.g., IPSL group, J. L. Dufresne et al.), climate change projections using IPSL CM5 Earth System Model: from CMIP3 to CMIP5, submitted to Climate Dynamics, 2012, with representation of double diffusion processes (Merryfield et al., 1999), Langmeyer cells (Axell, 2002) and surface wave breaking (Mellor & Blumberg, 2004; Burchard &  2008). Also, the restratification effects of the finite-amplitude, sub-mesoscale mixed layer eddies have been included in some models (e.g., CCSM4 group, Danabasoglu et al., 2012) using the mixed layer parameterisation of Fox-Kemper et al. (2008) as was implemented by Fox-Kemper et al. (2011). It might also be expected that other model developments would improve the representation of the mixed layer. It has been shown (Lee et al., 2011) that increased resolution of a model can improve the representation of the ocean advection of buoyancy and the stratification of the Southern Ocean, which translates into a mixed layer representation that is much more realistic.

 Sallée et al. suggest it might be expected that if improvements have been made to the representation of fluxes of heat, freshwater and momentum at the air surface interface, through improvements to the atmospheric models or to atmospheric-ocean coupling, to benefit the representation of the ocean mixed layer. Surface winds in the Southern Hemisphere also have a strong impact on the mixed layer in the Southern Ocean (e.g. Sallée et al., 2010b) and it has been shown that they are influenced by the representation of the recovery of ozone over the first half of the 21st century (Son et al., 2008, 2010). All CMIP5 models differ from CMIP3 in that they include a representation of changes in the stratospheric ozone (see information about ozone forcing in Bracegirdle et al., 2013); Sallée et al. note that all models use the same ozone forcing, and therefore, in different models the effect may be of different intensity.

In this paper Sallée et al. assess the present-day skill and projected changes that are simulated by the CMIP5 models, with a focus on the mixed layer of the Southern Ocean. No study to date has consistently analysed the causes of the present-day Southern Ocean mixed layer bias in climate models, in spite of the result of 10 years of Argo profiles in the Southern Ocean that are now available allow a robust understanding of the structure, characteristics and construction of the real Southern Ocean winter mixed layer (Dong et al., 2008; Sallée et al., 2010b). Sallée et al. present here such a consistent analysis. Sallée et al. attempt to identify the most important forcings that lead to biased information, though a detailed study of the influence of particular parametrisation schemes is beyond the scope of this paper. For projections in the future Sallée et al. provide a summary of multi-model projections and examine if there is a state dependence in the model response. The analysis is focused on the assessment of the deep mixed-layer band that develops in winter and leads to the formation of mode and intermediate water. The implications of representation of mixed layer for modelled mode and intermediate water masses are added to the end of this paper, and are discussed further in the framework of the Southern Ocean overturning circulation in a companion paper (Sallée et al., 2013).

Southern Ocean Mixed Layer Representation

In the Southern Ocean there is a strong seasonal cycle in the Mixed Layer Depth (MLD) that exceeds more than 400 m in some locations to the north of the Antarctic Circumpolar Current (ACC) (Sallée et al., 2010b). The water column is destabilised by the winter cooling and the MLD is increased to the extent that the maximum MLD are found in the late austral winter (September) before the shallow summer mixed layer is rapidly re-established by warming in spring and early summer. In this paper, as introduced above, Sallée et al. focused mostly on the deep winter mixed layer convection that develops on the equatorwards, northern, edge of the ACC. Sallée et al. found it useful for documenting the ability of models to represent summer mixed layer and the amplitude of the seasonal cycle before they tackled the analysis of winter Mixed Layer depth representation. The characteristics of water subducted in winter are set all year round; while the winter mixed layer depth is crucial for the ventilation of the Southern Ocean (McCartney, 1977; Hanawa & Talley, 2001; Sallée et al., 2010a). Also, the depth in summer is critical for the chemistry of the ocean surface and biological activity (Lovenduski & Gruber, 2005; Sallée et al., 2010a), which are processes that are implemented in Earth System Models that participated in CMIP5.

In the Southern Ocean the structure of the mixed layer is characterised by a circumpolar band of deep mixed layers that reach 60-90 m in summer (February), in the latitude band 50oS-60oS. Mixed layers are shallower, at around 50 m, outside of this band. In winter the deep circumpolar band is strongly destabilised to reach depths of up to 400-700 m. In winter the band is narrower and concentrated only on the equatorwards edge of the ACC. Mode and intermediate waters are formed at this location (McCartney, 1977; Hanawa & Talley, 2001; Sallée et al., 2006, 2008a). The models tend to be biased shallow, on average, compared with observations, in summer and in winter, in the band of deepest mixed layers. In winter as well as summer the multi-model average of bias is significant. The multi-model average of bias reaches 50-70 m in summer, while it reaches 100-200 m in winter. Sallée et al. also found that models simulate a band of deep mixed layer that is too wide, which extends too far towards the equator, as was revealed by deep bias on the northern edge of the deepest mixed layer band (average deep bias of 100-200 m).

Compared with observations the summer and winter mixed layer depth biases translate into a misrepresentation of the mixed layer seasonal cycle. According to Sallée et al. the MLD seasonal cycle amplitude is too small by as much as 200 m in regions of deep mixed layer convection in the Eastern Indian Ocean, the mid-Pacific and eastern Pacific basins. In subtropical regions immediately to the north of the maximum mixed layer depth sector, in contrast, the seasonal cycle amplitude is too large when compared with observations by 100-200 m on average. There are important implications for the formation of mode and intermediate water of these significant biases. (Sallée et al., 2010a) have shown the importance of seasonal cycle and regional structure of the deepest mixed layer depth of the Southern Ocean in the subduction of water masses. In winter mixed layer depth and seasonal cycle amplitude the significant deep bias in the subtropical regions, western Indian Ocean and western Pacific Ocean, suggest that the amount of subtropical mode and intermediate water subducted in climate models is too large.

According to Sallée et al. they have shown, consistent with this, in a companion paper that the density of mode water is biased light as the result of unrealistically too large formation of subtropical mode water in the western Indian and western Pacific sectors, and the formation of subantarctic mode water in the eastern Indian and Pacific sectors is too weak (Sallée et al., 2013).

Understanding the bias in winter mixed layer depth

As described above, the multimodal average of MLD bias can be used to understand the general shortcomings of the ensemble of models, though it hides a range of structures that are very distinct across the models. In this section the spread of MLD patterns in each model is detailed, and the forcings are analysed to understand better what primarily leads to the distinct MLD representations across the models. The focus was on the circumpolar band of very deep mixed layer (MLDmax) that develops on the northern edge of the Antarctic Circumpolar Current (ACC) in winter, where the mode and intermediate waters form (McCartney, 1977; Hanawa & Talley, 2001; Sallée et al., 2006).

Conclusions and discussion

The representation of the winter mixed layer in the Southern Ocean has been evaluated in 21 climate models that participated in the CMIP5 exercise. In the climate models that were analysed the region MLDmax, where mode and intermediate waters form and are preconditioned, is shallower, lighter and more towards the equator than it has been observed. There are important implications for characteristics of mode and intermediate waters and the rate at which they enter the interior of the ocean. For the dissolution and sequestration of carbon dioxide in the interior of the ocean this is of primary importance (Séférian et al., 2012; Sallée et al., 2012).

This paper is the first to have unravelled the primary drivers of the bias that need to be looked at if the representation of deep MLD in the Southern Ocean in climate models, while MLD bias in the Southern Ocean in climate models has previously been documented (e.g. Downes et al., 2009). Sallée et al. identified that fluxes of freshwater increase artificially the stratification of MLDmax, and this biases the depth and density of the surface layer by preventing convection in the deep mixed layer. According to Sallée et al. they are not arguing for the thermal stratification or winter buoyancy flux not having an impact on MLDmax, but they identified the annual mean freshwater flux as the primary source of error. One of the largest shortcomings of knowledge in the Southern Ocean has long been identified as observational uncertainty and technical difficulties in obtaining good estimates of annual mean buoyancy flux in the Southern Ocean (e.g. Liu et al., 2011). The analyses of Sallée, however, offers a fresh perspective for modelling teams to adjust their Southern Ocean fluxes to best represent the amount of haline stratification at the base of the mixed layer that has been well observed.

CMIP5 models simulate shallowing, a lightening and a meridional shift of MLDmax under increased radiative forcing scenarios. The meridional shift is towards the equator in the Pacific sector and towards the pole in the Indian sector, and is associated with shift in ACC position (Meijers et al., 2012). The shallowing is linked strongly to increased fluxes of heat in winter and occurs mostly in the Pacific region. Sallée et al. found a strong state of dependency between historical and future change in MLD: those models that have the strongest historical bias in MLD indicate there will be little change in the future, whereas those that have a present-day MLD that is closer to observations indicate significant shallowing of the MLD under future forcing scenarios. Importantly, this suggests that changes in the future in MLD might be larger than in indicated by most models, given that most models are biased shallow. Mixed-layer properties are linked tightly to the volume and properties of ventilated layers in the interior of the ocean, in historical runs, as well as for future changes. The state dependency in mixed layer could therefore potentially indicate that most models simulate a reduction that is too weak in the volume of the ventilated layer. This would have large implications for sequestration of heat, freshwater and gases such as oxygen and carbon, and could indicate that this potential climate change feedback may be underestimated by the current generation of models.

Sources & Further reading

Sallée, J. B., et al. (2013). "Assessment of Southern Ocean mixed-layer depths in CMIP5 models: Historical bias and forcing response." Journal of Geophysical Research: Oceans 118(4): 1845-1862.


Author: M. H. Monroe
Last Updated 08/10/2017
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading