CGD's Oceanography Section

ANALYSIS OF THE COUPLED CLIMATE SYSTEM and the role of the ocean continue to be a priority activity. In particular, the ocean's Thermohaline Circulation (THC) is important in the oceanic meridional heat transport and uptake of anthropogenic CO₂. Paleoclimate observations indicate that it has undergone abrupt changes in the past and some modeling studies have suggested that it may be susceptible to abrupt changes in response to global warming. Building on previous years work from the IPCC AR4 studies, [Bryan #1, and Bryan #2] have examined the response of the THC in CCSM3 under different greenhouse gas stabilization scenarios, with a particular focus on the behavior when stabilization is approached non-monotonically. These studies show that the THC recovers to a slightly higher strength for the same greenhouse gas concentrations when the ultimate stabilization level is temporarily exceeded. Further, the conventional measure of THC strength, the maximum of the streamfunction for the zonally integrated transport, was shown to be a rather poor indicator of the THC behavior. While the maximum strength recovers following greenhouse gas stabilization, the overturning cell continues to shallow, and the deep part of the circulation weakens long after stabilization.

Simulated changes in the Arctic climate system are being investigated using CCSM and other coupled climate model simulations. The projected changes in the Arctic Ocean freshwater budgets from a number of models have been analyzed (Holland et al., 2007, in press). The model simulations clearly indicate acceleration in the Arctic freshwater cycle with increased net river input and net precipitation balanced by an increase in freshwater transport to the north Atlantic. The role that extratropical cyclone activity plays in the changing Arctic precipitation has been further examined in CCSM integrations (Finnis et al., 2007, in press). Results suggest that high latitude cyclone activity remains similar in the warming climate but produces more precipitation due to increasing available moisture.

An assessment of climate model simulated sea ice conditions continues. Our comparison between late-20th century IPCC-AR4 model simulated Arctic sea ice trends and observations (Stroeve et al., 2007) shows that the models are conservative in the rate of ice loss over the late 20th-early 21st century. Work is underway to understand these results and the large inter-model scatter in sea ice conditions. In particular, the changing sea ice mass budgets and their relationship to changing Arctic heat budgets are being assessed.

Multi-decadal variability of the Atlantic Meridional Overturning Circulation (AMOC) is investigated diagnostically in the NCAR CCSM3 present-day simulations, using the highest (T85x1) resolution version. This variability has a 21-year period and is present in many other ocean fields in the North Atlantic. In the AMOC, the oscillation amplitude is about 4.5 Sv, corresponding to 20% of the mean maximum MOC transport. The northward heat transport (NHT) variability has an amplitude of about 0.12 PW, representing 10% of the mean maximum NHT. In sea surface temperature (SST) and salinity (SSS), the peak-to-peak changes can be as large as 6-7 Celsius and 3 psu, respectively. Comparisons with observations show that neither the pattern nor the magnitude of the SST variability is realistic. These mid-latitude SST and SSS anomalies are created by the fluctuations of the subtropical-subpolar gyre boundary driven by wind stress curl (WSC) anomalies. The Labrador Sea region is identified as the deep water formation site associated with the MOC oscillations. These multi-decadal oscillations are primarily driven by atmospheric variability associated with the NAO which also shows variability at the 21-year period.

COUPLED EXPERIMENTATION into the climate response to changes in ocean physics has led to a number of unexpected and exciting results. First, lateral ocean viscosity was reduced as much as the model numerics would allow in order to resolve Tropical Instability Waves (TIW), so that these waves could pump heat laterally into the cold tongue of the equatorial Pacific. Jochum et al. (2008) document the reduction of the regional cold SST bias as well as several other positive impacts of low viscosity. There was a dramatic improvement is the sea-ice distribution in the Labrador Sea due to stronger currents advecting warmer water into the region, melting ice and initiating a positive ice-albedo feedback, improved the current structure and temperatures in the Kuroshio extension, with positive impacts on the atmospheric circulation over the North Pacific, and a very different water mass exchange between the North Atlantic and the Arctic. A particularly surprising result was a decrease the transport of the Antarctic Circumpolar Current through Drake Passage.

In the same spirit enhanced tidal mixing in the vicinity of the Indonesian Throughflow affects SST only a little, but improves regional convection and precipitation a great deal (Jochum and Potemra, 2007). Replacing the hitherto used constant vertical background diffusivity with its observed latitudinal structure led to a dramatic improvement of subtropical coastal upwelling and tropical Indian Ocean precipitation. These improvements have not been anticipated, and are the result of non-obvious atmospheric feedback processes. Therefore, future CCSM developments in all components should be assessed in the coupled context.

OCEAN OBSERVATIONS are an essential element of modeling studies and model development. Together with a group of PMEL scientists, mooring and satellite based observations have been combined to measure the heat flux maintained by atropical instabilty waves (Jochum et al., 2007). It is found that the new CCSM performs surprisingly well in that regard; most importantly though, it shows that theoretical ideas developed at NCAR lead to new observations and provide directions for the wider community.

Sea surface salinity is much less well observed than temperature, and efforts on behalf of the CLIVAR Salinity Working Group and the NASA AQUARIUS Satellite Science team are aimed at narrowing this gap.

The PHYSICS OF THE OCEAN as a forced, rather than coupled, system is also a primary research goal. An analysis of global upper-ocean observations collected by ARGO profiling floats supports model results which have drawn attention to the role of ocean boundary layer mixing in generating interannual variations in ocean spice in select subtropical regions characterized by large salinity inversions (Yeager and Large 2007). The observational confirmation of a large seasonal cycle in near surface bulk vertical Turner Angle (an indicator of the degree of density compensation of temperature and salinity profiles in late winter) suggests that the spice injection phenomenon, first identified in model hindcasts (Yeager and Large 2004), may indeed play a significant role in generating interannual-decadal variability in the ocean's pycnocline.

An attempt to quantify the role of mesoscale eddies in ocean transport is underway, following a number of years of work in developing and tuning high-resolution (0.1 deg.) global ocean circulation models. In one study, the age spectrum, or transit time distribution (TTD) of the global ocean is being computed. This entity describes the probability density function of ages in a fluid subject to turbulent mixing in addition to mean advection. The statistics of the TTD will be used to constrain parameter choices in the eddy mixing parameterization used in coarser resolution versions of the CCSM ocean component model, and in estimating ocean uptake of trace gases such as CO2. In a second study, eddy tracer diffusivities are being computed by inverting the eddy flux-gradient relationship for a suite of approximately orthogonal passive tracers. These can in turn be used to evaluate the assumptions and limitations underlying extant mixing parameterizations, and guide future development of more accurate and comprehensive parameterizations.

OCEAN BIOGEOCHEMISTRY is an important component of the coupled carbon cycle modeling effort with CCSM. Extensive experiments have performed with the BEC ecosystem model (Moore, Doney, and Lindsay 2004) in the fully coupled CCSM3 model in its low resolution configuration (~3 degree for the ocean). Feedbacks between biogeochemistry and climate are being analyzed and compared to previous studies and observational constraints. Recent improvements to the parameterization of the iron cycle have been incorporated and initial experiments with the modified model have been performed in an ocean only configuration at the higher ~1 degree resolution of CCSM.

The sensitivity of coral reefs to future bleaching events has been shown using data from both observations and an ensemble of CCSM3 Twentieth Century simulations. Two important discoveries are: First, exposure of coral reefs to increased SSTs has not been uniform over the last 50 years. SSTs greater than 29°C have warmed less than SSTs < 29°C. This is consistent with, but does not prove, the notion that various negative feedbacks may act to slow the rate of warming in some tropical regions. The numbers of bleaching events that have occurred in regions where SSTs are already greater than 29°C appear to be less than in other regions. Second, coral reef sensitivity also varies geographically. The bleaching events that have occurred in these regions appear to have occurred at lower thresholds than the commonly used 1-2°C above the maxima.

Improving CCSM3 BIASES is a priority of the CCSM Advisory Board, the CGD Advisory Committee, and CCSM scientists. In response a two year effort has led to realistic representation of ENSO in CCSM (Neale et al. 2007). The major achievement has been to remove the strong two-year period in ENSO variability and radically improve the ENSO correlation with SST over the whole Pacific Ocean. Noteworthy here is that scientists from across CGD took part in this effort. Moreover, as with the observations, they were able to break new grounds by showing that the delayed oscillator model is a poor paradigm for ENSO. Rather ENSO should be treated as a series of events. This is an old debate, but the new results end a stalemate in the discussion by showing that the delayed oscillator regime is possible and does exist, but leads to an unrealistic ENSO if present exclusively. The current focus is on the amplitude of ENSO variability and in at least one case a very reasonable amplitude has been produced all across the low frequency band (1 - 10 years). Soon a new effort will be launched to address the seasonal cycle of winds and SST across the equatorial Pacific, which have compared poorly with observations ever since CSM1 days.

The largest local cold SST errors, with magnitudes approaching 10 degC, occur east of the Grand Banks off Newfoundland and result from a poor simulation of the path of the North Atlantic Current (NAC). The broader scale climate impacts of this error were investigated in [Bryan #3], by nudging the NAC toward a more realistic path using a conservative and adiabatic adjustment to the pressure field. Downstream, in the Nordic Seas, the subsurface ocean responds favorably to this adjustment, as the vertical profiles of potential temperature and salinity converge towards the observations. Atmospheric stationary wave patterns show a modest improvement, with a slight weakening of the excessively deep Icelandic low.

Continual effort on OCEAN MODEL DEVELOPMENT is needed to maintain a state of the art ocean model for coupled climate studies, as well as ocean physics research. Apart from the very practical steps described above in coupled analysis, the more long term and fundamental research is done in collaboration with universities. For example Jochum is lead-PI in 2 proposals that investigate the benefits of high resolution regional coupled modelling in the tropical Atlantic and Indian oceans. It is found that on ocean mesoscales the rectifying effect of TIW induced wind field disturbances is rather small (Seo et al. 2007a), but that the key to a realistic ITCZ structure is to properly represent African Easterly Waves (Seo et al. 2007b). For the Indian Ocean, it appears that baroclinic instabilities of the current that carries the ITF water can lead to large scale atmospheric anomalies, possibly affecting southern Indian Ocean tropical storms (Zhou et al., 2007ab).

Major ocean model developments are proceeding under the auspices of the CLIVAR Climate Process Teams (CPTs) on both gravity current entrainment, and eddy mixed layer interaction. The former has imbedded a parameterized Mediterranean overflow scheme in the ocean component of the CCSM3, and finally succeeded in obtaining stable solutions for overflows from the Nordic seas.

The latter, has produced promising results in two areas. First, is a new near-surface eddy flux parameterization that includes the effects of diabatic mesoscale fluxes within the surface layer? Experiments show significant improvements compared to CCSM3, and a comparison with available observations and eddy-resolving model solutions are favorable. The improvements include the elimination of strong, near-surface, eddy-induced circulations, a better heat transport profile in the upper-ocean, reduced abyssal cooling and diminished trends in the potential temperature drifts.

Second, is a new prescription for the surface intensification and abyssal reduction of the tracer diffusivities? This structure is simply achieved using a stratification dependent vertical profile. Resulting solutions with a variety of diffusivity choices compare more favorably with the available observations than those of CCSM3. Positive aspects include an improved representation of the vertical structure and transport of the eddy-induced velocity in the upper-ocean North Pacific, a reduced warm bias in the upper ocean, including the equatorial Pacific, improved southward heat transport in the low- to mid-latitude Southern Hemisphere, and a modest enhancement of abyssal stratification in the Southern Ocean.

While the fidelity of high-resolution ocean simulations has increased considerably over the last several years, there remain significant biases in aspects of the simulated circulation and substantial uncertainties in the robustness of the results with respect to parameterization choices. A new configuration of the 0.1deg global ocean model was developed and a series of sensitivity studies carried out to test various closure assumptions. Innovations over previous generations of the high-resolution model include a tripole grid that provides a more uniform discretization of the sphere than the dipole grid used previously, and partial bottom cells that provide a more accurate representation of topography. These changes have resulted in considerable improvements in the circulation in the North Atlantic basin over previous versions of the model.

In order to maintain the leading edge in SEA-ICE MODELING, work continues on sea ice model development. Improvements to the sea ice meltpond parameterization have been implemented and plans have been developed for improvements to the snow aging and albedo parameterizations. The assessment of sea ice model conditions in the latest fully coupled runs is ongoing.

The ADMINISTRATION OF CCSM has been a very significant burden again during 2007. Positions held include the chair and a member of the CCSM Science Steering Committee, co-chair of the CCSM Polar Climate Working Group and co-chair of the Ocean Working Group. Also, there were liaison responsibilities for the Biogeochemistry, Polar Climate and Ocean working groups.

---

References:

Moore, K., S. C. Doney, and K. Lindsay, 2004: Upper ocean ecosystems dynamics and iron cycling in a global 3D model. Global Biogeochemical Cycles, 18, GB4028.

[1] Nakashiki, N., D.-H. Kim, F. O. Bryan, Y. Yoshida, D. Tsumune, K. Maruyama, H. Kitabata 2006: Recovery of the thermohaline circulation and sea ice area under CO2 stabilization and overshoot scenarios. Ocean Modelling, 15, 200-217.

[2] Bryan, F.O., N. Nakashiki, Y. Yoshida, and K. Maruyama (2007) Response of the thermohaline circulation during different pathways toward greenhouse gas stabilization. In: Ocean Circulation: Mechanisms and Impacts. Geophysical Monograph Series, Volume 173, 351-364. A. Schmittner, J. Chiang and S. Hemming (Eds). AGU, Washington, D.C.

[3] Weese, S.R. and F.O. Bryan, 2006: Climate impacts of systematic errors in the simulation of the path of the North Atlantic Current. Geophys. Res. Lett., 33, L19708, doi:10.1029/2006GL027669.

Yeager, S. G. and W. G. Large, 2004: Late-Winter Generation of Spiciness on Subducted Isopycnals. J. Phys. Oceanogr., 34, 1528-1547.

Yeager, S. G. and W. G. Large, 2007: Observational Evidence of Winter Spice Injection. J. Phys. Oceanogr., in press.

Danabasoglu, G., and J. Marshall, 2007: Effects of vertical variations of thickness diffusivity in an ocean general circulation model. Ocean Modelling, 18, 122-141, doi:10.1016/j.ocemod.2007.03.006.

Wu, W., G. Danabasoglu, and W. G. Large, 2007: On the effects of parameterized Mediterranean overflow on North Atlantic Ocean circulation and climate. Ocean Modelling, 19, 31-52, doi:10.1016/j.ocemod.2007.06.003.

Danabasoglu, G., R. Ferrari, and J. C. McWilliams, 2007: Sensitivity of an ocean general circulation model to a parameterization of near-surface eddy fluxes. J. Climate, in press.

Danabasoglu, G., 2007: On multi-decadal variability of the Atlantic meridional overturning circulation in the Community Climate System Model version 3 (CCSM3). J. Climate, submitted.

Kleypas, J. A., J. M. Lough, and G. Danabasoglu, 2007: The potential role of the ocean thermostat in determining future coral reef survival and distribution. Geophys. Res. Lett., submitted.