The goal of the Climate System Model (CSM) project is to develop a comprehensive modeling system for climate, including both physical and biogeochemical aspects, and to apply appropriate configurations of the model to important scientific problems. The CSM project is committed to using comprehensive coupled models without flux adjustments. A major success of the project over the last year has been the completion of a 300-year simulation with the initial version of the CSM, in which there are no significant drifts in the surface temperatures.
The CSM project is committed to widening the involvement of the U.S. climate science community in the development of the CSM, in its application to scientific questions, and in the analysis of major simulations undertaken by the project. The project also is committed to making the modeling system and the results of major experiments freely available to the wider scientific community.
The initial development of the CSM was completed in 1996 by a team of scientists centered in the Climate and Global Dynamics (CGD) Division and lead by Byron Boville (Climate Modeling Section, CMS) and Peter Gent (Oceanography Section, OS), with participants from the Atmospheric Chemistry Division (ACD) and the Scientific Computing Division (SCD). The first version of the model (CSM1.0) was made freely available on the World Wide Web in June, 1996. In January, 1997, the oversight of the CSM project passed to a Scientific Steering Committee (SSC) consisting of Maurice Blackmon (chair), Boville, Robert Dickinson (University of Arizona), Jeffrey Kiehl (CMS), Gent, David Randall (Colorado State University), Jagadish Shukla (Center for Ocean-Land-Atmosphere Studies, COLA), Susan Solomon (NOAA Aeronomy Laboratory), and Warren Washington (Climate Change Research Section, CCR). To foster wider participation in the CSM project, the SSC has established eight working groups and named co-chairs for each group:
The second annual CSM workshop was held in Breckenridge, Colorado, in June, 1997. Results of the 300-year CSM control simulation and of a transient carbon dioxide (CO2) increase simulation were presented at the workshop. All of the CSM working groups met at least once prior to the workshop and met again at the workshop.
The primary foci of the CSM project over the last year have been: the completion and analysis of a long (300-year) control simulation for present day solar forcing and atmospheric trace gas concentrations; an initial experiment on transient CO2 increase; the planning and development of forcing scenarios for simulations of the last century (1870-1990); the development of improvements to the component models; and the refinement of the coupled modeling system.
The CSM is a flexible modeling system that includes atmospheric and oceanic general circulation models (GCMs), a land surface biophysics and basic soil hydrology model, and a sea-ice dynamics and thermodynamics model, each communicating with a driver program (flux coupler) using message passing. (This figure (9K) shows the presently available configurations of the CSM.) A high-resolution tropical Pacific model is available in place of the lower-resolution, full-depth global ocean GCM. The component models are driven primarily by fluxes at the Earth's surface. Those interfacial fluxes that directly depend on the state of more than one component model, e.g., turbulent fluxes of latent and sensible heat, are computed within the flux coupler. The flux coupler is also responsible for interpolating and averaging between the differing grids of the component models while conserving local and integral properties and for coordination of the time integration.
The modeling system can be configured in a variety of ways, depending on the scientific objectives for a particular simulation. For example, the atmospheric GCM, Community Climate Model version 3 (CCM3), and the land surface model version 1 (LSM1) can be run coupled to:
Similarly, the ocean GCM and sea-ice model can be run coupled to CCM3 and LSM1, or the atmosphere and land models can be replaced by inputting datasets from previous simulations or from observations/analyses.
Except for the coupler, each component of the modeling system has at least two options, a full model and a simple program to read existing datasets. In principle, replacing any of the full models with an alternate version is straightforward. However, many practical implementation problems are involved, since the software for the modeling system is not mature, and most alternate models have not been designed with a similar coupling strategy in mind.
To perform relatively drift-free coupled simulations, compatible initial states for the component models are required. Several sequential integrations of the component models, referred to as the spin-up procedure, are needed to ensure compatibility. (This figure (2K) illustrates the spin-up procedure for CSM runs, beginning from equilibrium ocean, atmosphere, and land solutions.) The flexible nature of the modeling system is exploited during the spin-up procedure.
First, the atmosphere and land models were integrated for 6.5 years using observed climatological monthly SSTs and sea-ice extents diagnosed from the SSTs. Second, the NCAR CSM Ocean Model (NCOM) was integrated to an equilibrium annual cycle solution using surface insolation from the International Satellite Cloud Climatology Program (ISCCP), reanalyzed atmospheric fields from the National Centers for Environmental Prediction (NCEP), precipitation estimates from the microwave sounding unit (MSU), and the salinity dataset from Levitus (1982). Sea-ice extent in this integration is also diagnosed from the climatological SSTs. The equilibrium simulation contains a weak (two-year time constant) restoring of surface salinity to Levitus. This restoring term is not included in the subsequent spin-up phase or in coupled simulations. Third, the ocean and sea-ice models were integrated for 60 years using daily atmospheric forcing from the first step and an equilibrium January 1 initial condition for the ocean from the second step. For the first 25 years of this simulation, the absorbed insolation at the surface from CCM3 was used in place of the downward insolation during this period. During the first 50 years of this phase, the ocean tracer equations were accelerated by a factor of 6 with respect to the momentum equations, and the deep ocean was accelerated by a factor of 10 with respect to the surface. The acceleration was turned off for the last 10 years of the spin-up.
The fully coupled model was integrated for 300 years. The global annual mean surface temperatures exhibit an adjustment of 0.7 K over the first 5 to 10 years of the simulation and are remarkably stable afterwards. (This figure (8K) shows 12-month running means of surface temperature globally averaged over: all surfaces; over land only; and over ocean and sea-ice. The mean of each series from years 11 to 300 is indicated by a horizontal line.) The initial adjustment is largely due to a 1.5 K decrease in the land temperatures, resulting from the fact that a generic initial condition was inadvertently used in LSM1 instead of the equilibrated state from the end of the CCM3/LSM1 simulation. There is also a rapid adjustment of the ocean temperatures in the first few months of the simulation, with the initial month being approximately 0.2 K warmer than any subsequent month. The coupled simulation has strong variability on multi-year timescales but no surface temperature trends after year 10. The trends in land and ocean/ice temperatures, determined by least squares fits for years 11 to 299, are 0.03 K per century and are small compared to the standard deviations of the annual means of 0.2 K and 0.07 K, respectively.
Over much of the globe, the annual mean simulated SSTs are similar to the observed SSTs. (This figure (13K) shows SSTs from the coupled simulation averaged over years 91-100. The contour interval is 2°C.) Over much of the globe, the SST errors are less than 1 K. (This figure (11K) shows the difference between the simulated SST and climatological temperatures for 1950-1979. The contour interval is 1°C, and magnitudes less than 1°C are unshaded.) The marine stratus regions off the western coasts of North and South America and off Africa are too warm by 2 to 3 K, resulting from a bias in cloud simulation in CCM3. In higher northern latitudes, a shift in the Gulf Stream is apparent with a warm bias off Labrador, while the SSTs are too cold near Norway and in the North Pacific. These biases are accompanied by shifts in the ice distribution. The high latitude southern ocean is slightly too warm, although the largest difference with climatology is associated with deviations of the flow over large ridges in bottom topography.
The area covered by sea-ice in the northern hemisphere increases for the first 20 years of the coupled simulation, then stabilizes with about 15% too much ice area compared to observations. (This figure (8K) shows the total area of sea-ice in a) the Northern Hemisphere and b) the Southern Hemisphere. Solid curves are the 12-month running means from the model, and the shading indicates the annual range of the model monthly mean values. Horizontal lines give observational estimates of the climatological maximum and minimum areas.) The excess is somewhat larger in winter than in summer. There is an 80-year period from years 110-190 with increased winter ice areas, then the area returns to its earlier, somewhat overly large value. This increased winter ice is quite thin and does not have a clear signal in the total ice volume. The maximum ice areas in the Southern Hemisphere drop to the observed level almost immediately in the coupled simulation, giving an annual cycle of ice area that matches observations and remains stable throughout. The ice retreats back to the Antarctic coast in summer, while extensive regions of relatively thin, non-compact ice are found in winter, in agreement with observations.
A coupled simulation in which the atmospheric CO2 concentration increased by 1% per year was performed in collaboration with scientists from the Japanese Central Research Institute of Electric Power Industry (CRIEPI). This simulation used initial conditions from year 15 of the 300-year control run. The CO2 concentration was held fixed at 355 parts per million by volume (ppmv) for 10 years, while instantaneous data was output every 6 hours. CO2 was then increased at 1% per year for 115 years, at which time the concentration had increased by a factor of slightly more than three. Output was again obtained every 6 hours for 10 years beginning at the time of CO2 doubling. The 6-hour output is being used by CRIEPI scientists as boundary and forcing data for regional model simulations. The globally averaged temperature increases by 1.25 K at the time of CO2 doubling and 2 K at the time of CO2 tripling, consistent with a 2 K equilibrium temperature increase for CCM3 coupled to a slab ocean. (This figure (13K) shows the 12-month running means of globally averaged surface temperature for the control simulation (red) and the increasing CO2 experiment (blue).)
The CSM Working Group on Atmospheric Biogeochemistry and Anthropogenic Climate Change (ABACC) chaired by Kiehl (CMS) and Solomon (NOAA/Aeronomy Laboratory) has identified two major projects as foci for the CSM. The first project, Simulation of the Climate of the 20th Century, will carry out three simulations of the time period 1870 to 1990. The first simulation will employ time varying spatially prescribed changes in trace gas concentrations of CO2, O3, CH4, N2O, and chlorofluorocarbons (CFCs) from 1870 to 1990. The CH4, N2O, and CFCs are advected in the model. Oxidation of CH4 to water (H2O) in the stratosphere is also accounted for in this simulation. This simulation, which uses "greenhouse gases" only, will provide information on the model response in the absence of any negative anthropogenic forcing agents. The second simulation for the 1870 to 1990 time period will include time varying, three-dimensional prescribed sulfate aerosol distributions obtained from the version of CCM3 that includes a sulfur chemistry cycle. This simulation will include both the positive greenhouse effects of the first simulation and the negative forcing due to sulfate aerosols. Both the direct and indirect radiative effects of sulfate aerosols will be calculated by the CSM. The final simulation will include two sources of natural variability: solar and volcanic in nature. The second foci of the ABACC will be to carry out scenario-based simulations of climate change of the 21st century. The changes in chemical composition for these simulations will rely on collaborations with Guy Brasseur and other members of ACD.