Testing the impact of clouds on the radiation budgets of19 atmospheric general circulation modelsGerald L. PotterProgram for Climate Model Diagnosis and Intercomparison, Lawrence Livermore National Laboratory, Livermore,California, USARobert D. CessMarine Sciences Research Center, State University of New York at Stony Brook, Stony Brook, New York, USAReceived 28 July 2003; revised 4 November 2003; accepted 13 November 2003; published 27 January 2004.[1] We compare cloud-radiative forcing (CRF) at the top-of-the atmosphere from19 atmospheric general circulation models, employing simulations with prescribed sea-surface temperatures, to observations from the Earth Radiation Budget Experiment(ERBE). With respect to 60°Nto60°S means, a surprising result is that many of the19 models produce unusually large biases in Net CRF that are all of the same sign(negative), meaning that many of the models significantly overestimate cloud radiativecooling. The primary focus of this study, however, is to demonstrate a diagnostic procedure,using ERBE data, to test if a model might produce, for a given region, reasonable CRF as aconsequence of compensating errors caused either by unrealistic cloud vertical structure,cloud optical depth or cloud fraction. For this purpose we have chosen two regions, one inthe western tropical Pacific characterized by high clouds spanning the range from thincirrus to deep convective clouds, and the other in the southeastern Pacific characterized bytrade cumulus. For a subset of eight models, it is found that most typically produce morerealistic regionally-averaged CRF (and its longwave and shortwave components) for thesoutheastern region as opposed to the western region. However, when the diagnosticprocedure for investigating cloud vertical structure and cloud optical depth is imposed, thissomewhat better agreement in the southeastern region is found to be the result ofcompensating errors in either cloud vertical structure, cloud optical depth or cloud fraction.The comparison with ERBE data also shows large errors in clear-sky fluxes for many of themodels.INDEX TERMS: 1620 Global Change: Climate dynamics (3309); 1610 Global Change:Atmosphere (0315, 0325); 3359 Meteorology and Atmospheric Dynamics: Radiative processes; 3309Meteorology and Atmospheric Dynamics: Climatology (1620); KEYWORDS: radiation, climate, cloudsCitation: Potter, G. L., and R. D. Cess (2004), Testing the impact of clouds on the radiation budgets of 19 atmospheric generalcirculation models, J. Geophys. Res., 109, D02106, doi:10.1029/2003JD004018.1. Introduction[2] Cloud-climate interactions comprise one of the great-est uncertainties in attempting to model climate changeusing general circulation models (GCMs), and there is aneed to devise ways of testing such interactions withinmodels. As Webb et al. [2001] emphasize: ‘‘If we are tohave confidence in predictions from climate models, anecessary (although not sufficient) requirement is that theyshould be able to reproduce the observed present-daydistribution of clouds and their associated radiative fluxes.’’Since the availability of the Earth Radiation Budget Exper-iment (ERBE) data [Ramanathan et al., 1989; Harrison etal., 1990], there have been numerous comparisons ofmodels to the ERBE top-of-the-atmosphere (TOA) radiativefluxes. For example, a comparison of seasonal changes ofcloud-radiative forcing (CRF), as produced by 18 GCMs, toERBE satellite data showed substantial differences, an dmore importantly provided clues as to the deficiencies ofsome models [Cess et al., 1997]. In their study, Webb et al.[2001] added an additional constraint by combining ERBEdata with cloud-top pressures determined by the Interna-tional Satellite Cloud Climatology Project (ISCCP) [Rossowand Schiffer, 1991]. Such a combined test is important,because as we later emphasize, it is possible for a model toproduce, for a given region, reasonable TOA CRF as aconsequence of compensating errors caused either by unre-alistic cloud vertical structure, cloud optical depth or cloudfraction.[3] In the present study we adopt the procedure used byCess et al. [2001a , 2001b], who studied cloud verticalstructure anomalies over the tropi cal Pacifi c during thestrong 1997/1998 El Nin˜o. This involved interfacing short-wave (SW) and longwave (LW) CRF, as measured both byERBE and the Clouds and the Eart h’s Radiant EnergyJOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109, D02106, doi:10.1029/2003JD004018, 2004Copyright 2004 by the American Geophysical Union.0148-0227/04/2003JD004018$09.00D02106 1of9System (CERES), so as to infer changes in cloud verticalstructure. These changes were in turn shown to be consis-tent with measurements of the vertical distribution of cloudsbased on observations by the Stratospheric Aerosol and GasExperiment (SAGE) II [Cess et al., 2001b]. In the presentstudy we employ five years of ERBE data (1985 –1989) todetermine both CRF and cloud vertical structure for tworegions of the Pacific, a tropical western region character-ized by high clouds and a southeastern region characterizedby trade cumulus. For these two study regions, this data setis used to test the impact of clouds on the radiation budgetsof 19 atmospheric GCMs for the same time period as theERBE data. Although there have been numerous compar-isons of GCMs to ERBE data, this study additionallyaddresses the issue of how well the models depict cloudvertical structure and cloud optical depth for the 1985 –1989period.2. Earth Radiation Budget Experiment (ERBE)Data[4] The ERBE data consist of the monthly-mean CRFmeasurements on 2.5°  2.5° grids and for the 5-year ERBEperiod (1985–1989) [Harrison et al., 1990]. The SW andLW components of CRF [Ramanathan et al., 1989;Harrison et al., 1990] are defined asSWCRF ¼ RC Rwhere R denotes the TOA all-sky reflected SW and RCthatfor clear skies, whileLWCRF ¼ FC Fwhere F and FC, respective l y, denote the all-sky and clear-sky TOA emitted LW. Typically SW CRF is negative(cooling) and LW CRF positive (heating). The Net CRF issimply the sum of the componentsNet CRF ¼ SW CRF þ LW CRFand cancellation between SW CRF cooling and LW CRFheating (Net CRF = 0) can be expressed by the ratioN ¼ SW CRFðÞ= LW CRFðÞ¼1If N > 1, SW cooling dominates, whereas LW heatingdominates for N < 1.[5] Figure 1 shows the geographical distribution of theNet CRF for the DJF period and averaged over the fiveERBE years (1985 –1989). The western and southeasternstudy