JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 103, NO. D3, PAGES 3399-3417, FEBRUARY 20, 1998 A global interactive chemistry and climate model: Formulation and testing Chien Wang, Ronald G. Prinn, and Andrei Sokolov Joint Program on the Science and Policy of Global Change, Massachusetts Institute of Technology, Cambridge Abstract. In order to elucidate interactions between climate change and biogeochemical processes and to provide a tool for comprehensive analysis of sensitivity, uncertainty, and proposed climate change mitigation policies, we have developed a zonally averaged two- dimensional model including coupled biogeochemical and climate submodels, as a part of an integrated global system nxxlel. When driven with calculated or estimated trace gas emissions from both anthropogenic and natural sources, it is designed to simulate centennial-scale evolution of many radiatively and chemically imix•rtant tracers in the atn•sphere. Predicted concentratiom of chemical species in the chemistry submodel are used interactively to calculate radiative forcing in the climate submodel, which, in turn, provides winds, teng•eratures, and other variables to the chemistry submodel. Model predictions of the surface trends of several key species are close to observations over the past 10-20 years. Predicted vertical distributions of climate-relevant species, as well as seasonal variations, are also in good agreement with observatiom. Runs of the model imply that if the current increasing trends of anthropogenic emissions of climate-relevant gases are continued over the next centm•, the chemical corr•sition of the atmosphere would be quite different in the year 2100 than that currently observed. The differences involve not only higher concentrations of major long-lived trace gases such as CO2, N20, and CH4 but also about 20% lower concentrations of the major tropospheric oxidizer (OH free radical), and almost double the current concentrations of the short-lived air pollutants CO and N O•. 1. Introduction Concerns about future climate change involve not only the impact of increases of long-lived trace gases in the atmosphere on the climate (e.g., surface temperature, precipitation, and cloud coverage) but also the influence of climate change on atmospheric chemical processes and the possible effects of products of these processes, such as aerosols, on climate. The lifetimes of chemical species in the atmosphere are affected by many processes, including atmospheric transport, chemical reactions, wet and dry deposition, exchange with the ocean or vegetation, and emissions from both human-related and natural processes. One needs to include treatment of each of these processes to address accurately the climatic effects of atmospheric chemical species and the influence of climate change on atmospheric chemistry. To address feedbacks between chemistry and climate, it is clear that a coupled numerical model including explicit descriptions of both climate dynamics and atmospheric chemistry is needed. In recent years, models addressing both climate and chemistry with various levels of coupling between these two components have been developed [e.g., Hauglustaine et al., 1994; Taylor and Penner, 1994; Roelofts and Lelieveld, 1995; Moxim et al., 1996; Levy et al., 1996]. However, to quantify the magnitude and uncertainty of the effects on Copyright 1998 by the American Geophysical Union. Paper number 97ID03465. 0148-0227/98/97ID-03465509.00 climate of long-lived trace gases (and vice versa), a large number of long (>100 years) transient integrations using fully coupled models with special attention to surface fluxes need to be carried out. A model would in particular have to be computationally efficient so that many runs can be made in order to understand the relative importance of various climate- chemistry feedbacks, to determine sensitivity of the results to the many critical assumptions in submodels, and (through comparison with observations) to make improvements. Constructing such a coupled model is both a theoretical and computational challenge. Toward the goal of obtaining a better quantitation of climate-chemistry interactions and their uncertainty, we have developed a coupled global-scale modeling system including a global climate model and an atmospheric chemistry model. This system is designed to predict as functions of time, latitude and altitude the zonally averaged concentrations of the major chemically and radiatively important trace species in the atmosphere. The chemistry and climate submodels in this system are fully interactive. Specifically, the transport of chemical species is driven by dynamical variables predicted by the climate model, and the calculations of gaseous and aqueous phase reactions are based on the temperatures, radiative fluxes, and precipitation rates computed in the climate model. Predicted mixing ratios of trace species are then used, in turn, to calculate the radiative forcing in the climate model. This paper describes the coupled model in detail, with a special focus on the chemistry component that has not yet been published. Results from a 124-year "reference" run and 33993400 WANG ET AL.' A GLOBAL INTERACTIVE CHEMISTRY AND CLIMA• MODEL tests of the model using observational data are then presented. The last section provides a critique of these results and some conclusions. Further analyses focusing on the dynamics of the interactions between atmospheric chemistry and climate change are discussed by C. Wang and R. Pfinn (Interactions among emissions, atmospheric chemistry, and climate change: Implications for future trends, submitted to Journal of Geophysical Research, 1997). where ' = Sadvection + Seddy + Seonvection + Semission + Sreaetion + Sdeposition (1) (2) 2. Model Descriptions The climate submodel used in the coupled model is the Massachusetts Institute of Technology (MIT) two- dimensional (2-D) land-ocean-resolving (LO) statistical- dynamical model [Sokolov and Stone, 1995; 1997]. It is a modified version of a model developed at the Goddard Institute for Space Studies (GISS) [Yao and Stone, 1987; Stone and Yao, 1987; 1990]. The original version of the 2-D model was developed from the 3-D GISS general circulation model (GCM) of Hansen et al. [1983]. As a result, the model's numerics and parameterizations of physical processes, such as radiation and convection, closely