IDENTIFICATIONS AND APPLICATIONS OF COUPLED CLIMATE MODELS Jin-Yi Yu Department of Earth System Science, University of California Irvine, USA. Keywords: Climate system, climate model, atmosphere, ocean, land surface, chemistry, coupling interface, information exchange Contents 1. Introduction 2. Elements of Coupled Climate Models 3. Component Models 4. Coupling Interfaces 5. Computational Aspects of Coupled Climate Models 6. Important Issues in Climate Model Coupling 7. Hierarchy of Coupled Climate Models 8. Applications of Coupled Climate Models Acknowledgements Related Chapters Glossary Bibliography Biographical Sketch Summary Climate system models are at the core of environmental system models. Identification and application of coupled climate models constitute an important research area of environmental system modeling. They also represent the ultimate challenge to the performance of numerical models for the atmosphere, ocean, land surface, and chemistry. Environmental system models are also characterized by strong interactions among sub-systems. They share concerns and problems similar to those faced by coupled climate models. Therefore, the technologies developed for coupled climate systems pave the way for the future success of environmental system models. 1. Introduction The Earth climate system consists of several major components, such as atmosphere, ocean, land, and chemistry. Many climate phenomena that have profound impacts on the environment and human society are produced by strong interactions among two or more of those components. The ozone hole depletion involves interactions between chemical processes and the atmospheric Search Print this chapter Cite this chapterPage 1 of 15IDENTIFICATIONS AND APPLICATIONS OF COUPLED CLIMATE MODELS10/18/2006http://greenplanet.eolss.net/EolssLogn/searchdt_categorysorted.asp?cmd=getdoc&maxSi...circulation; the El Nio and La Nia events are results of mutual forcing between the atmosphere and ocean; and the Amazon deforestation involves not only land surface but also atmospheric processes. In the most challenging problem of global warming, all four components of the Earth climate system are involved. Coupled climate models, which integrate together models of the individual climate system components and allow them to interact with each other, are powerful tools to study and to predict those climate phenomena. There is a hierarchy of coupled models with varying complexity being developed in the climate research community. Each type of coupled model has its own unique strength. Selection of coupled model type primarily depends on the nature of specific climate problem being addressed and the availability of computational resources. 2. Elements of Coupled Climate Models There are at least two basic elements in a coupled climate model: component model and coupling interface. For more complex coupled climate models that are computational demanding and generate large amounts of outputs, two additional elements are needed: high-performance computational tools and a model data information system. A component model is a numerical representation of one climate component in the coupled model. Component models are responsible for simulating the state of that climate component and its responses to forcing from other components. Interactions within the climate system are represented in coupled models by exchanging simulated state parameters (e.g., atmospheric circulations, ocean temperatures, ozone concentrations, land surface types, etc.) among component models. The back and forth passing of simulated parameters between any two component models constitutes a coupling interface. Coupling interfaces are responsible for coordinating and processing information flow among component models. Complex coupled climate models solve large sets of partial differential equations to provide comprehensive information on climate states. Furthermore, those complex models are usually developed for studying climate phenomena that have long characteristic time scales. Long integrations of these models are usually needed. It is desirable that those long model integrations can be completed in a reasonably short "wall clock" time. High-performance computational tools are necessary for coupled climate modeling. Processing and analyzing the huge amounts of output from long-term simulations of coupled climate models can be time-consuming and difficult. An efficient data information system is needed to simplify the use of simulation results See Data Information System and Coupling Interfaces. Page 2 of 15IDENTIFICATIONS AND APPLICATIONS OF COUPLED CLIMATE MODELS10/18/2006http://greenplanet.eolss.net/EolssLogn/searchdt_categorysorted.asp?cmd=getdoc&maxSi...3. Component Models Various complexities of component models have been developed by the climate research community for the simulations of the atmosphere, ocean, land surface, and chemistry components of the climate system. In general, those models can be categorized into three types: Statistical models, simplified physics-based models, and sophisticated, general circulations models. 3.1 Statistical Type Statistical models use observational data to empirically determine the state parameters of climate components. They are usually simple and computationally efficient, but cannot provide comprehensive information on climate states. Statistical models are usually used in coupled climate models for generating force to drive another component models. The empirical relationships derived from the present climate situation that underlies these models are not necessary valid when significant changes occur in the climate system. Therefore, statistical component models are not appropriate for coupled climate models that aim at climate change applications. 3.2 Simplified Physics-based Type Simplified physics-based models solve simplified governing equations. The simplification eliminates equation terms that are of secondary importance to generate specific features in climate components. Many factors can affect how the governing equations can be simplified, such as time scales of the climate feature to be simulated or its spatial scales, geographic location, or dynamic characteristics. Simplified physics-based models are computationally more efficient than complex general circulation models and physically more realistic than statistical models. Shallow-water equation models are one example of simplified