Coordinated Ocean-ice Reference Experiments (COREs)

https://doi.org/10.1016/j.ocemod.2008.08.007Get rights and content

Abstract

Coordinated Ocean-ice Reference Experiments (COREs) are presented as a tool to explore the behaviour of global ocean-ice models under forcing from a common atmospheric dataset. We highlight issues arising when designing coupled global ocean and sea ice experiments, such as difficulties formulating a consistent forcing methodology and experimental protocol. Particular focus is given to the hydrological forcing, the details of which are key to realizing simulations with stable meridional overturning circulations.

The atmospheric forcing from [Large, W., Yeager, S., 2004. Diurnal to decadal global forcing for ocean and sea-ice models: the data sets and flux climatologies. NCAR Technical Note: NCAR/TN-460+STR. CGD Division of the National Center for Atmospheric Research] was developed for coupled-ocean and sea ice models. We found it to be suitable for our purposes, even though its evaluation originally focussed more on the ocean than on the sea-ice. Simulations with this atmospheric forcing are presented from seven global ocean-ice models using the CORE-I design (repeating annual cycle of atmospheric forcing for 500 years). These simulations test the hypothesis that global ocean-ice models run under the same atmospheric state produce qualitatively similar simulations. The validity of this hypothesis is shown to depend on the chosen diagnostic. The CORE simulations provide feedback to the fidelity of the atmospheric forcing and model configuration, with identification of biases promoting avenues for forcing dataset and/or model development.

Introduction

Simulations with global coupled ocean-ice models can be used to assist in understanding climate dynamics, and as a step towards the development of more complete earth system models. Unfortunately, there is little consensus in the modelling community regarding the design of global ocean-ice experiments, especially those run for centennial and longer time scales. In particular, there is no widely agreed method to force the models. Furthermore, some relatively small differences in forcing methods can lead to large deviations in circulation behaviour and sensitivities. Such difficulties create practical barriers to comparing simulations from different modelling groups.

A central purpose of this paper is to present Coordinated Ocean-ice Reference Experiments (COREs). COREs provide a common reference point for research groups developing and analyzing global ocean-ice models. They do so by establishing a standard practice for the design of a baseline set of experiments that is useful for model development and ocean-ice research. By standard practice, we envision an experimental protocol that satisfies the following goals:

  • Provides model simulations that can be tested directly against a broad suite of ocean and sea ice observations;

  • Is not specific to a particular model or model framework, facilitating cooperation between groups and model communities;

  • Is not so complex or computationally expensive so as to make it too onerous for smaller groups to implement;

  • Can be incorporated into a more comprehensive model development or research program; e.g., by providing spun-up initial conditions for fully coupled climate simulations or control experiments in sensitivity studies;

  • Facilitates sharing of expertise and reduces redundant efforts in forcing data set design.

Prior to the availability of atmospheric reanalysis products, a de facto standard practice existed in the ocean modelling community: wind stress was prescribed by the only widely available global dataset (Hellerman and Rosenstein, 1983), and surface temperature and salinity were damped toward observed conditions (see Section 3.1). With the emergence of more comprehensive and realistic atmospheric reanalysis and remote sensing products, the choices have expanded but also become more complex. Our proposal for COREs does not provide the definitive resolution of these forcing issues, but can provoke discussion and debate leading to improved scientific convergence onto a common experimental protocol.

We distinguish the research focus of COREs from that of model intercomparison projects. In an intercomparison project, simulations follow a strict protocol and output is generated for analyses by a broad community. Projects, such as the Atmospheric Model Intercomparison Project (AMIP) (Gates, 1993), help document model similarities and differences, and can be of great use for various research and development purposes. Prior to deciding whether an analogous global ocean-ice model intercomparison project (i.e., an OMIP) would be a useful exercise, it is important for the research community to converge to a baseline experimental design. We believe that COREs provide a useful step toward this convergence.

Given the broad selection of models participating in this study, the simulations presented here can provide some feedback to the fidelity of the atmospheric forcing. That is, places where each model produces a similar behaviour that is biased relative to observations may signal a problem with the atmospheric dataset, thus suggesting areas requiring reexamination. The common bias could, in contrast, indicate a common problem amongst the full suite of models that may highlight problems in the model fundamentals and/or configurations. Analogously, in the case where a single model produces a widely varying behaviour, this outlier model may result from problems in the model’s fundamentals and/or configuration.

This paper contains three main parts. The first part summarizes the state of the art in global ocean-ice coupled modelling, and it starts in Section 2, which highlights some uses of ocean-ice models and argues for the relevance of a reference experimental design. Section 3 reviews methods used to force the ocean-ice models, with emphasis on limitations of these methods. Section 4 then presents our proposal for COREs. The second part of the paper is given in Sections 5 Globally averaged ocean temperature and salinity, 6 Horizontally averaged temperature and salinity, 7 SST and SSS bias maps, 8 Annual cycle at Ocean Weather Ship Echo, 9 Sea ice concentrations, 10 Tropical Pacific, 11 Mixed layer depths, 12 Zonal average potential temperature and salinity, 13 Volume transport through Drake Passage, 14 Poleward heat transport, 15 Meridional overturning streamfunction, 16 Surface freshwater forcing and MOC behaviour, where we consider a selection of diagnostics from seven global ocean-ice simulations run with the CORE-I (repeating annual cycle) forcing. For each diagnostic, we provide rationalizations for why the diagnostic is useful to examine in global simulations; present guidance towards observational datasets that can be used for model-observational comparisons; display model results; and offer hypotheses that could explain model differences and which could be followed-up with more focused studies. We do not provide a complete mechanistic understanding of model differences for the exhibited diagnostics. Doing so is nontrivial from many perspectives, and would require a new study no less lengthy than the present. Section 17 closes the main paper with discussion and conclusions. The third part of this paper is comprised of appendices that detail aspects of the models used in this study; the experimental protocol; the methods use to force the models; formulational aspects of certain diagnostics; and acronyms used in the manuscript.

Section snippets

Uses of ocean-ice models

To study the earth’s climate, and possible climatic changes due to anthropogenic forcing, various research teams have successfully built realistic global climate or earth system models with interactive ocean, sea ice, land, atmosphere, biogeochemical, and ecosystem components (referred to as climate models in the following). These models are generally built incrementally, with components considered initially in isolation, then sub-groups of components are coupled, and finally the full set of

Boundary fluxes for ocean-only and ocean-ice models

A coupled ocean-ice model requires momentum, heat, and hydrological exchanges with the atmosphere to drive the simulated ocean and ice fields. These exchanges take the form of stress from atmospheric winds, of radiative and turbulent fluxes of heat, and of precipitation, continental runoff and evaporation. Notably, evaporation has an associated turbulent latent heat flux which links the thermal and hydrological fluxes. When decoupling the ocean and sea ice models from the atmosphere, one must

A proposal for COREs

The previous section highlights some issues that arise when decoupling the ocean and sea ice components from the rest of the climate system, in particular from an interactive atmosphere. Quite simply, it is ambiguous how one specifies interactions with unrepresented components, and these ambiguities can introduce nontrivial and often unphysical sensitivities. It is thus important to recognize the limitations of ocean-ice models, as no methodology for specifying interactions with missing

Globally averaged ocean temperature and salinity

Amongst the most basic of model diagnostics is the globally averaged ocean temperature and salinity. Assuming no interior sources and sinks, the globally averaged ocean temperature and globally averaged ocean salinity are affected by surface fluxes, and by exchange of heat and salt with the sea ice. Sections D.4 Budget for volume averaged temperature, D.5 Budget for volume averaged salinity in the Appendix detail the various processes contributing to the globally averaged temperature and

Horizontally averaged temperature and salinity

More details about the temperature and salinity spin-up in the CORE-I simulations are provided in Fig. 5, Fig. 6. These figures show time series for the anomalous annual mean temperature and salinity as a function of depth, where the anomalies were created by taking the difference between the annual mean from the model and the annual mean from Conkright et al., 2002, Steele et al., 2001. The near-surface and thermocline conditions show a rapid adjustment during the first 50–100 years, with

SST and SSS bias maps

Global maps of SST and SSS from the simulations are compared in Fig. 7, Fig. 8 to those from Conkright et al. (2002) for the World Ocean outside the Arctic, with Steele et al. (2001) used for the Arctic. Despite the strong negative feedback provided by the prescribed air-temperature, the models develop some regions with nontrivial difference patterns, and these are often found in multiple models. These results suggest that common modelling problems and/or deficiencies in the forcing may be

Annual cycle at Ocean Weather Ship Echo

Analysis of the long-term adjustment behaviours seen in Sections 5 Globally averaged ocean temperature and salinity, 6 Horizontally averaged temperature and salinity, 7 SST and SSS bias maps is complemented by an inspection of the annual cycle of near-surface thermal properties. In general, the temporal rate of change of ocean heat storage is balanced by the surface heat flux and the horizontal divergence of heat advection and diffusion in the ocean. As suggested by theory (Gill and Niiler, 1973

Sea ice concentrations

High-latitude processes, including the distribution and strength of convective areas and the seasonal cycle of polar sea ice cover, are among the most challenging aspects of the climate system to accurately simulate. In particular, these aspects are very sensitive to the choice of atmospheric boundary conditions and model configurations, as emphasized by Proshutinsky et al. (2001) for AOMIP. A detailed examination of the parameter sensitivities encountered in the host of CORE experiments

Tropical Pacific

Realistic ocean simulations in the Tropical Pacific are important for coupled ocean atmosphere simulations of El Niño Southern Oscillations (ENSO) (Latif et al., 1998). Integrity of the simulation is dependent especially on the wind stress and the ability to maintain a tight thermocline, with the latter dependent on vertical mixing (Meehl et al., 2001) as well as horizontal friction.

For comparison with observed hydrography and currents at the equator, we employ the isopycnal analysis of Johnson

Mixed layer depths

The surface mixed layer is deepened by wind-driven mixing processes, Ekman-induced subduction, and convective overturn of gravitationally unstable water columns. It is the latter process that is particularly important in the formation of Subantarctic Mode Water (SAMW), North Atlantic Deep Water (NADW), and High Salinity Shelf Water around Antarctica. The mixed layer depth (MLD) attained in late winter is thus a crucial model diagnostic, as it reflects the depth of rapid overturn of surface

Zonal average potential temperature and salinity

One of the most widely assessed benchmarks of ocean model performance is the distribution of global ocean potential temperature and salinity (θ-S) in the latitude and depth plane. There are several reasons that this diagnostic is popular. Foremost, the way heat and salt are distributed with latitude and depth is directly set by the global thermohaline and wind-driven circulation, reflecting the rate and properties of large-scale water-mass ventilation. Water-mass overturning rates are

Volume transport through Drake Passage

Vertically integrated volume transport of seawater through selected regions of the ocean provide modellers with an important benchmark to evaluate the integrity of simulated water-masses and ocean currents, as well as the boundary forcing impacting the transport.

Poleward heat transport

Poleward heat transport by the climate system is a response in the atmosphere and ocean to differential solar heating, with more warming in the tropics than the poles (Peixoto and Oort, 1992). In contrast to the atmosphere, ocean heat transport is greatly modified by meridional land-sea boundaries. It is critical for climate models to have a partitioning of poleward transport that is well distributed between the atmosphere and ocean in order for the model to maintain a relatively stable climate

Meridional overturning streamfunction

The meridional overturning streamfunction diagnoses the transport of volume or mass, and it is commonly used to summarise various features of the large scale circulation, particularly the effects from thermohaline forcing. In Section D.1 in the Appendix, we detail methods for how this streamfunction is computed in the models. Here, we present results from the CORE-I simulations.

Surface freshwater forcing and MOC behaviour

The long-term behaviour of global ocean-ice models depends critically on surface freshwater forcing, particularly in the high-latitude areas important for setting properties of the deep and bottom water masses. As discussed in Section 3.3, salinity restoring is generally used to damp drifts in water-mass properties, and in particular for the purpose of maintaining a stable MOC. During the development of CORE-I, some groups examined sensitivities to salinity/water restoring. Given that four of

Discussion and concluding remarks

Simulations with global ocean-ice models are more difficult to run in isolation than coupled atmosphere-land simulations. There are two main reasons for this distinction. First, the atmospheric fields needed to force the ocean-ice system are more numerous and less well known than the relatively well observed surface temperatures used to derive atmospheric fluxes over the ocean and specified sea-ice. Second, assuming that the atmosphere rapidly responds to a slowly varying SST, as in the

Acknowledgements

This paper was prompted by discussions occurring within the CLIVAR Working Group for Ocean Model Development (WGOMD) since the panel’s first meeting in 2000. The WGOMD sponsored a Pilot-OMIP starting in 2001, with this project involving just a few models. The Pilot-OMIP was itself based on an earlier comparison of German ocean-ice models (Fritzsch et al., 2000). Both projects used the forcing dataset compiled by Röske (2006). The culmination of the Pilot-OMIP occurred at a CLIVAR workshop on

References (148)

  • Barnier, B., 1998. Forcing the ocean. In: Chassignet, E.P., Verron, J. (Eds.), Ocean Modeling and Parameterization,...
  • B. Barnier et al.

    Impact of partial steps and momentum advection schemes in a global ocean circulation model at eddy permitting resolution

    Ocean Dyn.

    (2006)
  • A. Beljaars

    The parameterization of surface fluxes in large-scale models under free convection

    Q. J. Roy. Meteor. Soc.

    (1994)
  • M. Bentsen et al.

    Simulated variability of the Atlantic meridional overturning circulation

    Clim. Dynam.

    (2004)
  • R. Bleck et al.

    Salinity-driven thermocline transients in a wind and thermohaline forced isopycnic coordinate model of the North Atlantic

    J. Phys. Oceanogr.

    (1992)
  • C.W. Böning et al.

    Annual cycle of poleward heat transport in the oceans: results from high-resolution modeling of the Equatorial North Atlantic

    J. Phys. Oceanogr.

    (1994)
  • Briegleb, B., Bitz, C., Hunke, E., Lipscomb, W., Holland, M., Schramm, J., Moritz, R., 2004. Scientific description of...
  • Briegleb, B., Hunke, E., Bitz, C.,, Lipscomb, W., 2002. Description of the Community Climate System Model Version 2 Sea...
  • Brodeau, L., Barnier, B., Treguier, A., Penduff, T., Gulev, S., 2008. An era40-based atmospheric forcing for global...
  • F. Bryan

    High latitude salinity effects and interhemispheric thermohaline circulations

    Nature

    (1986)
  • D. Cavalieri et al.

    30-year satellite record reveals conrasting Arctic and Antarctic decadal variability

    Geophys. Res. Lett.

    (2003)
  • E. Chassignet et al.

    Generalized vertical coordinates for eddy-resolving global and coastal forecasting

    Oceanography

    (2006)
  • E. Chassignet et al.

    North Atlantic simulation with the HYbrid Coordinate Ocean model (HYCOM): impact of the vertical coordinate choice, reference density, and thermobaricity

    J. Phys. Oceanogr.

    (2003)
  • A. Colin de Verdière et al.

    Baroclinic instability: a wavemaker for oceanic interdecadal variability

    J. Phys. Oceanogr.

    (1999)
  • W. Collins et al.

    The community climate system model: CCSM3

    J. Clim.

    (2006)
  • Comiso, J., 1999 (updated 2005). Bootstrap sea ice concentrations for NIMBUS-7 SMMR and DMSP SSM/I, June to September...
  • Conkright, M., Antonov, J., Baranova, O., Boyer, T., Garcia, H., Gelfeld, F., Johnson, D., Locarnini, R., Murphy, P.,...
  • M. Cox et al.

    A numerical model of the ventilated thermocline

    J. Phys. Oceanogr.

    (1984)
  • S. Cunningham et al.

    Transport and variability of the Antarctic Circumpolar Current in Drake Passage

    J. Geophys. Res.

    (2003)
  • G. Danabasoglu et al.

    Sensitivity of an ocean general circulation model to a parameterization of near-surface eddy fluxes

    J. Clim.

    (2008)
  • G. Danabasoglu et al.

    Diurnal coupling in the tropical oceans of CCSM3

    J. Clim.

    (2006)
  • G. Danabasoglu et al.

    The role of mesoscale tracer transports in the global ocean circulation

    Science

    (1994)
  • C. De Boyer Montegut et al.

    Mixed layer depth over the global ocean: an examination of profile data and a profile based climatology

    J. Geophys. Res.

    (2004)
  • T.L. Delworth et al.

    GFDL’s CM2 global coupled climate models – part 1: formulation and simulation characteristics

    J. Clim.

    (2006)
  • S.C. Doney et al.

    Antarctic bottom water formation and deep water chlorofluorocarbon distributions in a global ocean climate model

    J. Phys. Oceanogr.

    (2002)
  • J.K. Dukowicz et al.

    Implicit free-surface method for the Bryan–Cox–Semtner ocean model

    J. Geophys. Res.

    (1994)
  • M.H. England

    The age of water and ventilation timescales in a global ocean model

    J. Phys. Oceanogr.

    (1995)
  • M.H. England et al.

    Using chemical tracers to assess ocean models

    Rev. Geophys.

    (2001)
  • Fritzsch, B., Gerdes, R., Hiller, W., Latif, M., Legutke, S., Maier-Reimer, E., Olbers, D., Röske, F., 2000. The Ocean...
  • A. Ganachaud

    Large-scale mass transports, water mass formation, and diffusivities estimated from World Ocean Circulation Experiment (WOCE) hydrographic data

    J. Geophys. Res.

    (2003)
  • A. Ganachaud et al.

    Improved estimates of global ocean circulation, heat transport and mixing from hydrographic data

    Nature

    (2000)
  • P. Gaspar et al.

    A simple eddy kinetic energy model for simulations of the oceanic vertical mixing: tests at station Papa and long-term upper ocean study site

    J. Geophys. Res.

    (1990)
  • W. Gates

    AMIP: the atmosphere model intercomparison project

    Bull. Am. Meteorol. Soc.

    (1993)
  • P. Gent et al.

    What sets the mean transport through Drake Passage?

    J. Geophys. Res.

    (2001)
  • P.R. Gent et al.

    Isopycnal mixing in ocean circulation models

    J. Phys. Oceanogr.

    (1990)
  • P.R. Gent et al.

    Parameterizing eddy-induced tracer transports in ocean circulation models

    J. Phys. Oceanogr.

    (1995)
  • R. Gerdes et al.

    Reaction of the oceanic circulation to increased melt water flux from greenland – a test case for ocean general circulation models

    Clivar Exchanges

    (2005)
  • R. Gerdes et al.

    Simulated history of convection in the Greenland and Labrador Seas 1948–2001

  • R. Gerdes et al.

    Comparison of Arctic sea ice thickness variability in IPCC Climate of the 20th Century experiments and in ocean–sea ice hindcasts

    J. Geophys. Res.

    (2007)
  • A.E. Gill et al.

    The theory of the seasonal variability in the ocean

    Deep-Sea Res.

    (1973)
  • Cited by (560)

    View all citing articles on Scopus
    View full text