CCSM OCEAN MODEL WORKING GROUP MEETING REPORT

January 14-15, 2004

NCAR, Damon Room

 

The CCSM Ocean Model Working Group (OMWG) met in the Damon Room of the NCAR Mesa Laboratory on January 14 and 15, 2004. Sadly, Co-Chair Rick Smith was unable to attend because he was recovering from a major car accident. Our best wishes to Rick.

 

The major topics of the meeting were the IPCC version of the ocean model, the papers and experiments planned for the Journal of Climate Special Issue, high-resolution modeling, and model developments. Also, there were reports from related meetings of the ocean mixing CPTs in December and CCSM/GFDL ocean collaboration meeting on January 13, 2004.

 

A special thanks is due to Peter Gent, Antonietta Capotondi, and Gokhan Danabasoglu for presenting three short science talks on model results. The respective titles were "Ocean Heat Uptake in CCSM2," "Pacific Subtropical/Tropical Cells and Decadal Variability," and "Meridional Overturning Circulation in CCSM2."

 

IPCC INTEGRATIONS AND THE OCEAN COMPONENT

 

To view the presentation on the configuration of the ocean component for IPCC integrations, please click on PDF or PPT (PowerPoint). The ocean model results can be found at www.cgd.ucar.edu/oce/ccsm_data.html.

 

The current CCSM controls as of Friday, Jan. 23 06:01:00 MST 2004 are:

 

b30.004 0479-01-01-00000 - T42_gx1v3 present day control

b30.018 0410-09-01-00000 - T42_gx1v3 1% CO₂ (branch b30.004 0400-01-01)

bxx.xx5 0000-00-00-00000 - T42_gx1v3 1% CO₂ branch at doubling

bxx.xx1 0000-00-00-00000 - T42_gx1v3 1870 control

bxx.xx2 0000-00-00-00000 - T42_gx1v3 1870-2000 historical

b30.009 0284-01-01-00000 - T85_gx1v3 present day control

b30.014 0169-12-13-00000 - T85_gx1v3 1% CO₂ (branch b30.009 0070-01-01)

bxx.xx6 0000-00-00-00000 - T85_gx1v3 1% CO₂ branch at doubling

b30.013 0238-01-01-00000 - T85_gx1v3 1870 control

b30.017 0262-01-01-00000 - T85_gx1v3 1870 control (branch b30.013 0230-01-01)

bxx.xx3 0000-00-00-00000 - T85_gx1v3 1870-2000 historical

b30.015 0062-01-01-00000 - 2x2.5_gx1v3 present day control

bxx.xx4 0000-00-00-00000 - T31_gx3v5 present day control

bxx.xx7 0000-00-00-00000 - T31_gx3v5 1% CO₂ branch from present day run

bxx.xx8 0000-00-00-00000 - T31_gx3v5 1% CO₂ branch at doubling

bxx.xx9 0000-00-00-00000 - T31_gx3v5 1870 control

bxx.x10 0000-00-00-00000 - Special runs for transport fields

 

See www.cgd.ucar.edu/ccr/ipcc/index.html for the planned timelines for these runs, and www.cgd.ucar.edu/ccr/ipcc/status.html for their status, updated every 5 minutes.

 

PROPOSED CONTRIBUTIONS TO THE JOURNAL OF CLIMATE SPECIAL ISSUE

 

Preliminary plans for OMWG contributions to the Special Issue were prepared. The emphasis is on the ocean solutions from fully coupled model integrations. These plans try to align the Special Issue papers with the "key topics" given in the January 9 email to Working Group Co-Chairs entitled "Special Journal of Climate Issue - Information Needed - JAN 20 2004." The list that must be included was given as:

 

1. Overview of CCSM3

2. Description of the climate states produced by each component in the coupled system, e.g., a description of the polar climate from CCSM3

3. Analysis of the climate sensitivity of the model

4. Analysis of the response of the model to, e.g., paleoclimate scenarios and pre-industrial conditions

5. Discussion of simulation of major modes of variability that affect climate-change attribution studies, e.g., simulation of ENSO

 

We also made a preliminary identification of the "control" runs that would be needed for analyses to write these papers. These are referenced as:

 

2000_T42 -- present day control at T42, b30.004 (D')

2000_T85 -- present day control at T85, b30.009 (D)

1870_T42 --1870 control at T42

1870_T85 --1870 control at T85, b30.017

C20_T42 -- 20th century 1870 to 2000 at T42

C20_T85 -- 20th century 1870 to 2000 at T85

1%_T42 -- 1% per year increase in CO₂ at T42, b30.018

1%_T85 -- 1% per year increase in CO₂ at T85, b30.014

 

The points of contact, as stated below, are interim. Currently, there are no commitments to whether they will be authors of the proposed papers, but I would ask that they assess if there is sufficient interest for the paper to proceed, and if so, to recommend a lead (not necessarily themselves).

 

1) Contribution to the "Overview of CCSM3" if invited (Key Topic 1)

 

Science Theme: The ocean component of CCSM3, a general description with reference to other papers for specifics, and coupling to other components, plus anything else required for balancing with contributions from other components (e.g., selected highlight(s)). A suggestion is to include at least a summary of the major differences between T42 and T85, as seen by all the components.

 

Control Integrations: 2000_T42, 2000_T85, ???

 

Numerical Experiments: None at this time.

 

Point of Contact (Interim): W. Large (wily@ncar.ucar.edu)

 

2) Attribution of Ocean Biases in CCSM3 (Key Topic 2)

 

Science Theme: A description of the climate state of the ocean, but, in order to relieve boredom, from the point of view of documenting the biases, and attributing these as much as possible between the ocean model, the atmosphere model, the sea ice model, and coupled behavior. Biases will not only include SST and SSS, but also deep T and S and Ideal Age, as well as the mean and seasonal cycle of equatorial T and S and velocity.

 

Control Integrations: 2000_T42, 2000_T85, C20_T42, C20_T85

1870_T42, 1870_T85

 

Numerical Experiments:  Regional Restoring in Areas of Large Biases

  LOA (Long Ocean Alone cycles over historical record)

  COI (Coupled Ocean-Ice of LOA)

 

Point of Contact (Interim): W. Large (wily@ncar.ucar.edu)

 

3) Diurnal Coupling (Key Topic 5)

 

Science Theme: The response of the climate system to diurnal coupling, in particular ENSO variability in both the ocean and the atmosphere, since we know that this is the primary response. The paper is to demonstrate (if preliminary investigations continue to hold) that it is sufficient to capture the daily SST rectification with a simple ocean diurnal cycle (SW heating), while retaining daily ocean-atmosphere coupling.

 

Control Integrations: 2000_T85, 2000_T42

 

Numerical Experiments: b30.010 (E), companion to 2000_T85, no DC

 b30.012 (E'), companion to 2000_T42, no DC

1 hour coupling at T85

3 hour coupling at T85

 

Point of Contact (Interim): G. Danabasoglu (gokhan@ncar.ucar.edu)

 

4) Invasion of the Ocean by CFCs and Heat (Key Topics 3 and 4)

 

Science Theme: How does the ocean distribute unbalanced fluxes (active heat and freshwater, as well as passive transient tracers like CFCs) in the vertical and horizontal? The CFCs are the critical validation because of the wealth of ocean observations. A very relevant sub-theme will be verifying, or not, the utility of Slab Ocean Models (SOMs) for climate sensitivity determinations and sea ice modeling.

 

Control Integrations: 2000_T42, 2000_T85, C20_T42, C20_T85

 1870_T42, 1870_T85, 1%_T85, 1%_T42

 

Numerical Experiments: At least 1 C20_T85 and 1 C20_T42 integration need to have CFCs active in the ocean, starting in the 1930s when CFCs first appeared. These are purely passive with no real impact on the T85 coupled model performance, and the codes have been implemented and tested.

LOA

COI

 

Point of Contact (Interim): P. Gent (gent@ncar.ucar.edu)

 

5) Modes of Ocean Climate Variability (Key Topics 4 and 5)

 

May be separate parts, a single paper or a combination, with some parts perhaps in regular journals.

Part I Meridional Overturning Circulation

Part II, Subtropical-Tropical Cells

Part III, High Latitudes (North Atlantic/Arctic, Southern Ocean)

Part IV, Tropical Oceans

 

Science Theme: How does the ocean variability relate to the major climate indices of the atmosphere (NAO, AO, AAO, ENSO, NP)? How well are these relationships represented in the model in the few instances where some validation is possible over the historical record? Do these relationships change during the industrial era post 1870?

 

Control Integrations: 2000_T42, 2000_T85, C20_T42, C20_T85

 1870_T42, 1870_T85, 1%_T85, 1%_T42

 

Numerical Experiments:  LOA

  COI

 

Points of Contact (Interim): Part I, G. Danabasoglu (gokhan@ncar.ucar.edu)

    Part II, I. Wainer (wainer@usp.br)

    Part III, M. Holland (mholland@ncar.ucar.edu)

    Part IV, A. Capotondi (Antonietta.Capotondi@noaa.gov)

 

HIGH-RESOLUTION OCEAN MODELING AND THE EARTH SIMULATOR

 

A large number of sensitivity experiments with global and North Atlantic basin OGCMs have been completed by the OMWG and allied projects pursuing improvements in the representation of the climatological means state of the simulated flow. At resolutions near 10 km, western boundary current dynamics remain quite sensitive to subgrid-scale closure choices, with somewhat different manifestations in the Gulf Stream, Kuroshio, and Agulhas regions, for example. Of particular interest are experiments using anisotropic formulations of GM mixing and viscosity, and experiments with different representations of topography (partial vs full cell and different digital terrain databases) in both global and North Atlantic models. The conclusion from the North Atlantic experiments is that for the current formulation of the model, there is a fairly narrow domain of parameter space within which a realistic Gulf Stream separation and North Atlantic Current path can be achieved. Thus far, we have not identified the location of that target domain for the global 0.1 degree configuration, though the guidance of the North Atlantic model experiments is narrowing down the possibilities. The plan for the next six months is to begin configuring a tripole version of the global 0.1 model more suitable for eventual coupling to the rest of the CCSM system than is the present Hudson Bay dipole grid. For further details, see Frank Bryan's presentation, Matthew Hecht's presentation, and Norikazu Nakashiki's presentation.

OCEAN MODEL DEVELOPMENTS

 

There will be NO change in the ocean component until the IPCC version is released on or about May 18, 2004.

 

T31x3

An integral part of the May 18 release will be a low-resolution CCSM3 with a T31 atmosphere (and land), and nominal 3 degree (x3) ocean (and ice). The x3 ocean grid has been reconfigured since CCSM2, as reported at the 2003 Breckenridge workshop. The companion T31 model should be available soon, whereupon coupled testing can proceed.

 

POP2

The transition to POP2 will begin in summer 2004, after Phil Jones (LANL) completes updating of the CCSM version of POP1.4 (by end of February, 2004). The transition is not to change the ocean climate, and some testing will be needed to confirm this. It will also bring two new features with it, a tripole grid option and partial bottom cells. Ocean biogeochemistry modules will also need to be transitioned.

 

Upper Ocean Model

There is widespread support for an ocean model whose equilibration time is comparable to that of the atmosphere (~20 years). It is hoped that this can be accomplished in a simple and straightforward manner by strongly restoring to observed temperature and salinity at all depths below about 400m. The problem will be scoped out, and assuming it remains simple to implement, it will be developed in the POP2 framework following the summer transition to POP2.

 

Embedded High-Resolution Models

Development of high-resolution coupled atmosphere (WRF) and ocean (ROMS) models, embedded within the coupled CCSM at ocean eastern boundaries, has begun as an NCAR Opportunity Fund Project. The present focus is on transferring CAM physics to WRF.

 

CCSM/GFDL COLLABORATIONS

 

Forcing

Comparison of ocean/ice solutions revealed that the CCSM models had not been forced correctly. The problem was identified and the CCSM ocean/ice system will be rerun. However, a general tendency to keep too little sea ice in the Arctic is expected to persist. Joint investigations of the forcing will continue to determine the best way to alter the forcing (e.g., reduced radiation, colder air temperature), so that the ice volume becomes closer to the observed.

 

Ice Model

There would be some merit in setting ice albedos to their fully coupled values. But in CCSM these have been reduced significantly during the coupled model tuning exercise, so the forcing changes required to compensate may be too great. An ice-alone run at NCAR will test the sensitivity of the ice volume to surface roughness.

 

Equatorial Circulation

How close are the CCSM and GFDL model solutions when forced identically? CCSM will send results to GFDL, interpolated to a 1-degree zonal and 0.5-degree meridional grid between 15 degrees north and south latitudes. Here the comparison is assumed to be independent of Arctic forcing and can proceed immediately.

 

After Forcing is Fixed

How similar are the ocean solutions? Is the MOC still a factor of 2 different? Compare ocean-ice fluxes. Set up parallel investigations of the effects of different GM tapering, starting from a simple constant and working toward the fully (IPCC) ocean schemes of both CCSM and GFDL. Compare the rates of SSS drift. How different are the weak restoring fluxes in the two models? Is use of freshwater flux by GFDL and virtual salt flux by CCSM a source of significant difference?

 

OCEAN MIXING CPTS

 

NCAR is participating on the two CLIVAR Climate Process Teams (CPTs) on ocean mixing, which provides an avenue for OMWG participation as well. CPT scientists from the university community will be invited and encouraged to participate in CCSM.

 

Gravity Current Entrainment (link in preparation)

 

The NCAR/CCSM plans for contributing to this CPT over the next year or so are:

 

1) Implement a Price/Baringer/Yang parameterization of the Mediterranean Outflow into the North Atlantic.

2) Generalize this parameterization to other outflows from marginal seas, such as the Red Sea, Black Sea, and Persian Gulf.

3) Generalize the parameterization to overflows in more open ocean regions, such as Denmark Strait, Faroe Channel, Weddell Sea, and Ross Sea.

4) Assess the impacts of these overflows on local ocean circulation, such as the North Atlantic, and Antarctic MOC.

 

Eddy Mixed-Layer Interactions (CPT-EMILIE) (www.cpt-emilie.org).

 

The NCAR/CCSM activities associated with CPT-EMILIE will be largely in collaboration with GFDL (see above) and are to begin with:

 

1) Explicit calculation of the bolus velocity, U*, in the GM parameterization of eddy mixing

2) Implementation of a scheme where U* in the boundary layer has no shear and has a transport equal to the depth integrated transport below

3) Demonstration of the effect of additional horizontal diffusion in the boundary layer

4) Implementation of a GM coefficient dependence of the local Rossby radius

 

 

PARTICIPANTS