• Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor, 4 , 11471167.

    • Search Google Scholar
    • Export Citation
  • Boer, G. J., and Coauthors, 1992: Some results from an intercomparison of the climates simulated by 14 atmospheric general circulation models. J. Geophys. Res, 97 , 1277112786.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., and S. Levis, 2006: Evaluating aspects of the Community Land and Atmosphere Models (CLM3 and CAM) using a dynamic global vegetation model. J. Climate, 19 , 22902301.

    • Search Google Scholar
    • Export Citation
  • Boville, B. A., P. J. Rasch, J. J. Hack, and J. R. McCaa, 2006: Representation of clouds and precipitation processes in the Community Atmosphere Model version 3 (CAM3). J. Climate, 19 , 21842198.

    • Search Google Scholar
    • Export Citation
  • Brutsaert, W., 1984: Evaporation into the Atmosphere. Kluwer Academic Publishers, 299 pp.

  • Cess, R. D., and Coauthors, 1990: Inter-comparison and interpretation of climate feedback processes in 19 atmospheric general circulation models. J. Geophys. Res, 95 , 1660116615.

    • Search Google Scholar
    • Export Citation
  • Chahine, M. T., 1992: The hydrologic cycle and its influence on climate. Nature, 359 , 373380.

  • Choudhury, B. J., N. E. DiGirolamo, J. Susskind, W. L. Darnell, S. K. Gupta, and G. Asrar, 1998: A biophysical process-based estimate of global land surface evaporation using satellite and ancillary data II. Regional and global patterns of seasonal and annual variations. J. Hydrol, 205 , 186204.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., 2001: Parameterization of generalized cloud overlap for radiative calculations in general circulation models. J. Atmos. Sci, 58 , 32243242.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., P. J. Rasch, B. E. Eaton, B. Khattatov, J-F. Lamarque, and C. S. Zender, 2001: Simulating aerosols using a chemical transport model with assimilation of satellite aerosol retrievals: Methodology for INDOEX. J. Geophys. Res, 106 , 73137336.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., J. K. Hackney, and D. P. Edwards, 2002a: A new parameterization for infrared emission and absorption by water vapor in the National Center for Atmospheric Research Community Atmosphere Model. J. Geophys. Res, 107 .8028, doi:10.1029/2001JD001365.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., P. J. Rasch, B. E. Eaton, D. W. Fillmore, J. T. Kiehl, T. C. Beck, and C. S. Zender, 2002b: Simulation of aerosol distributions and radiative forcing for INDOEX: Regional climate impacts. J. Geophys. Res, 107 .4664, doi:10.1029/2001JD001365.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., and Coauthors, 2006a: The Community Climate System Model version 3 (CCSM3). J. Climate, 19 , 21222143.

  • Collins, W. D., and Coauthors, 2006b: The formulation and atmospheric simulation of the Community Atmosphere Model version 3 (CAM3). J. Climate, 19 , 21442161.

    • Search Google Scholar
    • Export Citation
  • Dai, A., and K. E. Trenberth, 2002: Estimates of freshwater discharge from continents: Latitudinal and seasonal variations. J. Hydrometeor, 3 , 660687.

    • Search Google Scholar
    • Export Citation
  • da Silva, A., A. C. Young, and S. Levitus, 1994: Algorithims and Procedures. Vol. 1, Atlas of Surface Marine Data 1994, NOAA, 83 pp.

  • Davey, M. K., and Coauthors, 2002: Stoic: A study of coupled model climatology and variability in tropical ocean regions. Climate Dyn, 18 , 403420.

    • Search Google Scholar
    • Export Citation
  • DeWeaver, E., and C. M. Bitz, 2006: Atmospheric circulation and its effect on Arctic sea ice in CCSM3 at medium and high resolution. J. Climate, 19 , 24152436.

    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., K. W. Oleson, G. B. Bonan, F. M. Hoffman, P. E. Thornton, M. Vertenstein, Z-L. Yang, and X. B. Zeng, 2006: The Community Land Model and its climate statistics as a component of the Community Climate System Model. J. Climate, 19 , 23022324.

    • Search Google Scholar
    • Export Citation
  • Döll, P., and S. Siebert, 2002: Global modeling of irrigation water requirements. Water Resour. Res, 38 , 8.18.10.

  • Doney, S. C., W. G. Large, and F. O. Bryan, 1998: Surface ocean fluxes and water-mass transformation rates in the coupled NCAR Climate System Model. J. Climate, 11 , 14201441.

    • Search Google Scholar
    • Export Citation
  • Fekete, B. M., C. J. Vorosmarty, and W. Grabs, 2002: High-resolution fields of global runoff combining observed river discharge and simulated water balances. Global Biogeochem. Cycles, 16 .1042, doi:10.1029/1999GB001254.

    • Search Google Scholar
    • Export Citation
  • Goosse, H., and T. Fichefet, 1999: Importance of ice-ocean interactions for the global ocean circulation: A model study. J. Geophys. Res, 104 , 2333723355.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., B. A. Boville, J. T. Kiehl, and P. J. Rasch, 1994: Climate statistics from the National Center for Atmospheric Research Community Climate Model (CCM2). J. Geophys. Res, 99 , 2078520813.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., J. T. Kiehl, and J. W. Hurrell, 1998: The hydrologic and thermodynamic characteristics of the NCAR CCM3. J. Climate, 11 , 11791206.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., J. M. Caron, G. Danabasoglu, K. W. Oleson, C. Bitz, and J. E. Truesdale, 2006: CCSM–CAM3 climate simulation sensitivity to changes in horizontal resolution. J. Climate, 19 , 22672289.

    • Search Google Scholar
    • Export Citation
  • Holland, M. M., C. M. Bitz, M. Eby, and A. J. Weaver, 2001: The role of ice–ocean interactions in the variability of the North Atlantic thermohaline circulation. J. Climate, 14 , 656675.

    • Search Google Scholar
    • Export Citation
  • Holland, M. M., C. M. Bitz, E. C. Hunke, W. H. Lipscomb, and J. L. Schramm, 2006: Influence of the sea ice thickness distribution on polar climate in CCSM3. J. Climate, 19 , 23982414.

    • Search Google Scholar
    • Export Citation
  • Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, and D. Xiaosu, Eds. 2001: Climate Change 2001: The Scientific Basis. Cambridge University Press, 944 pp.

    • Search Google Scholar
    • Export Citation
  • Kiehl, J. T., 1994: Clouds and their effects on the climate system. Phys. Today, 47 , 3642.

  • Kiehl, J. T., and K. E. Trenberth, 1997: Earth's annual global mean energy budget. Bull. Amer. Meteor. Soc, 78 , 197208.

  • Kiehl, J. T., and P. R. Gent, 2004: The Community Climate System Model, version 2. J. Climate, 17 , 36663682.

  • Kiehl, J. T., J. J. Hack, G. B. Bonan, B. B. Boville, D. L. Williamson, and P. J. Rasch, 1998: The National Center for Atmospheric Research Community Climate Model: CCM3. J. Climate, 11 , 11311149.

    • Search Google Scholar
    • Export Citation
  • King, M. D., and Coauthors, 2003: Cloud and aerosol properties, precipitable water, and profiles of temperature and water vapor from MODIS. IEEE Trans. Geosci. Remote Sens, 41 , 442458.

    • Search Google Scholar
    • Export Citation
  • Klein, S., and C. Jakob, 1999: Validation and sensitivities of frontal clouds simulated by the ECMWF model. Mon. Wea. Rev, 127 , 25142531.

    • Search Google Scholar
    • Export Citation
  • Kwok, R., G. F. Cunningham, and S. S. Pang, 2004: Fram Strait sea ice outflow. J. Geophys. Res, 109 .C01009, doi:10.1029/2003JC001785.

  • Large, W. G., and S. G. Yeager, 2004: Diurnal to decadal global forcing for ocean and sea-ice models: The data sets and flux climatologies. Tech. Rep. NCAR/TN-460+STR, National Center for Atmospheric Research, Boulder, CO.

  • Large, W. G., and G. Danabasoglu, 2006: Attribution and impacts of upper-ocean biases in CCSM3. J. Climate, 19 , 23252346.

  • Lejenäs, H., R. A. Madden, and J. J. Hack, 1997: Global atmospheric angular momentum and Earth-atmosphere exchange of angular momentum simulated in a general circulation model. J. Geophys. Res, 102 , 19311941.

    • Search Google Scholar
    • Export Citation
  • Menon, S., J. Hansen, L. Nazarenko, and Y. Luo, 2002: Climate effects of black carbon aerosols in China and India. Science, 297 , 22502253.

    • Search Google Scholar
    • Export Citation
  • Oberhuber, J. M., 1988: An atlas based upon the “COADS” data set: The budgets of heat, buoyancy, and turbulent kinetic energy at the surface of the global ocean. Tech. Rep. 15, Max-Planck-Institut für Meteorologie, 199 pp.

  • Oleson, K. W., and Coauthors, 2004: Technical description of the Community Land Model (CLM). Tech. Rep. NCAR/TN-461+STR, National Center for Atmospheric Research, Boulder, CO, 174 pp.

  • Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • Ramanathan, V., P. J. Crutzen, J. T. Kiehl, and D. Rosenfeld, 2001: Aerosols, climate, and the hydrological cycle. Science, 294 , 21192124.

    • Search Google Scholar
    • Export Citation
  • Randel, D. L., T. H. V. Haar, M. A. Ringerud, G. Stephens, T. J. Greenwald, and C. L. Combs, 1996: A new global water vapor dataset. Bull. Amer. Meteor. Soc, 77 , 12331246.

    • Search Google Scholar
    • Export Citation
  • Rasch, P. J., W. D. Collins, and B. E. Eaton, 2001: Understanding the Indian Ocean Experiment (INDOEX) aerosol distributions with an aerosol assimilation. J. Geophys. Res, 106 , 73377356.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and E. N. Dueñas, 2004: The International Satellite Cloud Climatology Project (ISCCP) Web Site: An online resource for research. Bull. Amer. Meteor. Soc, 85 , 167172.

    • Search Google Scholar
    • Export Citation
  • Schlesinger, W. H., 1991: Biogeochemistry: An Analysis of Global Change. Academic Press, 433 pp.

  • Stephens, G. L., and P. J. Webster, 1981: Clouds and climate: Sensitivity of simple systems. J. Atmos. Sci, 38 , 235247.

  • Stowe, L. L., C. G. Wellemeyer, T. F. Eck, and H. Y. M. Yeh, and The Nimbus-7 Cloud Data Processing Team, 1988: Nimbus-7 global cloud climatology. Part I: Algorithms and validation. J. Climate, 1 , 445470.

    • Search Google Scholar
    • Export Citation
  • Stowe, L. L., H. Y. M. Yeh, T. F. Eck, C. G. Wellemeyer, and H. L. Kyle, and The Nimbus-7 Cloud Data Processing Team, 1989: Nimbus-7 global cloud climatology. Part II: First year results. J. Climate, 2 , 671709.

    • Search Google Scholar
    • Export Citation
  • Timmermann, R., A. Worby, H. Goosse, and T. Fichefet, 2004: Utilizing the ASPeCt sea ice thickness data set to evaluate a global coupled sea ice-ocean model. J. Geophys. Res, 109 .C07017, doi:10.1029/2003JC002242.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., and C. J. Guillemot, 1995: Evaluation of the global atmospheric moisture budget as seen from analyses. J. Climate, 8 , 22552272.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., D. P. Stepaniak, and J. M. Caron, 2000: The global monsoon as seen through the divergent atmospheric circulation. J. Climate, 13 , 39693993.

    • Search Google Scholar
    • Export Citation
  • Vaughan, D. G., J. L. Bamber, M. Giovinetto, J. Russell, and A. P. R. Cooper, 1999: Reassessment of net surface mass balance in Antarctica. J. Climate, 12 , 933946.

    • Search Google Scholar
    • Export Citation
  • Vinje, T., 2001: Fram Strait ice fluxes and atmospheric circulation: 1950–2000. J. Climate, 14 , 35083517.

  • Warren, S. G., C. J. Hahn, J. London, R. L. Chervin, M. Robert, and R. Jenne, 1988: Global distribution of total cloud cover and cloud type amounts over the ocean. Tech. Rep. NCAR Tech. Note TN-317+STR, National Center for Atmospheric Research, Boulder, CO. 42 pp. + 170 maps.

  • Weatherly, J. W., B. P. Briegleb, and W. G. Large, 1998: Sea ice and polar climate in the NCAR CSM. J. Climate, 11 , 14721486.

  • Webb, M., C. Senior, S. Bony, and J. J. Morcrette, 2001: Combining ERBE and ISCCP data to assess clouds in the Hadley Centre, ECMWF, and LMD atmospheric climate models. Climate Dyn, 17 , 905922.

    • Search Google Scholar
    • Export Citation
  • Wentz, F. J., and R. W. Spencer, 1998: SSM/I rain retrievals within a unified all-weather ocean algorithm. J. Atmos. Sci, 55 , 16131627.

    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., R. D. Cess, M. D. King, D. A. Randall, and E. F. Harrison, 1995: Mission to planet Earth: Role of clouds and radiation in climate. Bull. Amer. Meteor. Soc, 76 , 21252153.

    • Search Google Scholar
    • Export Citation
  • Williamson, D. L., 1988: The effect of vertical finite difference approximations on simulations with the NCAR Community Climate Model. J. Climate, 1 , 4058.

    • Search Google Scholar
    • Export Citation
  • Willmott, C. J., and K. Matsuura, cited. 2000: Terrestrial air temperature and precipitation: Monthly and annual climatologies. [Available online at http://climate.geog.udel.edu/~climate.].

  • Xie, P., and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc, 78 , 25392558.

    • Search Google Scholar
    • Export Citation
  • Xie, S., and M. Zhang, 2000: Impact of the convection triggering function on single-column model simulations. J. Geophys. Res, 105 , 1498314996.

    • Search Google Scholar
    • Export Citation
  • Yeager, S. G., C. Shields, W. G. Large, and J. J. Hack, 2006: The low-resolution CCSM3. J. Climate, 19 , 25452566.

  • Zhang, G. J., 2003: Convective quasi-equilibrium in the tropical western Pacific: Comparison with midlatitude continental environment. J. Geophys. Res, 108 .4592, doi:10.1029/2003JD003520.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The simulated global water budget. Storage terms are in mm and exchange rates are in mm day−1. Quantities in [] are derived from MODIS, while quantities in () are a renormalized version of the global water cycle described in Peixoto and Oort (1992).

  • View in gallery

    Zonally averaged annual, DJF, and JJA precipitation rate in mm day−1 for CAM3, CCSM3, and CMAP.

  • View in gallery

    Same as in Fig. 2, but for net surface water flux (EP) in W m−2 for CAM3 and CCSM3.

  • View in gallery

    Same as in Fig. 2, but for precipitable water in kg m−2 for CAM3, CCSM3, NVAP, and CCM3.

  • View in gallery

    Zonally averaged total cloud fraction for CAM3 compared with ISCCP, Nimbus-7, and MODIS.

  • View in gallery

    Zonally averaged, ocean-only cloud liquid water path (LWP) in g m−2 for CAM3, CCSM3, MODIS, SSMI, and NVAP.

  • View in gallery

    Same as in Fig. 2, but for cloud ice water path (IWP) in g m−2 for CAM3.

  • View in gallery

    Zonally averaged annual mean temperature, specific humidity, and relative humidity differences between CAM3 and ERA-40.

  • View in gallery

    Profiles of equivalent potential temperature and specific humidity for CAM3, ERA-40, and RAOBS at Yap Island (9.4°N, 138.1°E) for a July climatology.

  • View in gallery

    CAM3 zonally averaged annual mean cloud liquid water concentration (color contours), ice water concentration (black contours), with 273-K freezing level contour (red) for reference.

  • View in gallery

    Zonally averaged annual, DJF, and JJA (left) mean and (right) transient meridional moisture transport in g m kg−1 s−1 for CAM3.

  • View in gallery

    Annual mean precipitation in mm day−1 for (top) CAM3, (middle) CMAP, and (bottom) their difference.

  • View in gallery

    Annual, DJF, and JJA net surface water flux (EP) in W m−2 for CAM3.

  • View in gallery

    Annual mean precipitable water in kg m−2 for (top) CAM3, (middle) NVAP, and (bottom) their difference.

  • View in gallery

    Annual mean cloud IWP and cloud LWP in g m−2 for CAM3.

  • View in gallery

    Equatorial (10°S–10°N) precipitation anomalies for CAM3, CMAP, and CCSM3. The CAM3 and CMAP period includes 1979–2000, and the CCSM3 is an arbitrary representative 22-yr period from the control simulation.

  • View in gallery

    ENSO warm (July 1994) minus ENSO cold (June 1999) event precipitation anomalies for CAM3 and GPCP.

  • View in gallery

    Amplitude of seasonal precipitable water changes in mm for CAM3 and NVAP.

  • View in gallery

    Annual mean precipitation in mm day−1 for (top) CCSM3, (middle) CAM3, and (bottom) their difference.

  • View in gallery

    Net freshwater flux for the CCSM3, uncoupled ocean model, and observations.

  • View in gallery

    Global (solid lines) and Atlantic (dotted lines) mean northward ocean freshwater transport from the CCSM3 (thick) and the uncoupled model (thin).

  • View in gallery

    (top) Zonally averaged annual mean precipitation and evaporation for CCSM3 and observations and (bottom) zonally averaged annual mean land hydrology values from CCSM3.

  • View in gallery

    The simulated annual cycle of Fram Strait sea ice transport in Sv. The solid line is CCSM3. The monthly average values over a 9-yr time series from observational estimates (Kwok et al. 2004) are shown as diamonds with the standard deviation from this time series indicated by the error bars. Note that the observational estimates for June–September are identical on this figure because Kwok et al. (2004) only provide a single average for these months.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 314 176 19
PDF Downloads 135 75 11

Simulation of the Global Hydrological Cycle in the CCSM Community Atmosphere Model Version 3 (CAM3): Mean Features

View More View Less
  • 1 National Center for Atmospheric Research, Boulder, Colorado
© Get Permissions
Full access

Abstract

The seasonal and annual climatological behavior of selected components of the hydrological cycle are presented from coupled and uncoupled configurations of the atmospheric component of the Community Climate System Model (CCSM) Community Atmosphere Model version 3 (CAM3). The formulations of processes that play a role in the hydrological cycle are significantly more complex when compared with earlier versions of the atmospheric model. Major features of the simulated hydrological cycle are compared against available observational data, and the strengths and weaknesses are discussed in the context of specified sea surface temperature and fully coupled model simulations.

The magnitude of the CAM3 hydrological cycle is weaker than in earlier versions of the model, and is more consistent with observational estimates. Major features of the exchange of water with the surface, and the vertically integrated storage of water in the atmosphere, are generally well captured on seasonal and longer time scales. The water cycle response to ENSO events is also very realistic. The simulation, however, continues to exhibit a number of long-standing biases, such as a tendency to produce double ITCZ-like structures in the deep Tropics, and to overestimate precipitation rates poleward of the extratropical storm tracks. The lower-tropospheric dry bias, associated with the parameterized treatment of convection, also remains a simulation deficiency. Several of these biases are exacerbated when the atmosphere is coupled to fully interactive surface models, although the larger-scale behavior of the hydrological cycle remains nearly identical to simulations with prescribed distributions of sea surface temperature and sea ice.

Corresponding author address: Dr. James J. Hack, NCAR, P.O. Box 3000, Boulder, CO 80307. Email: jhack@ncar.ucar.edu

Abstract

The seasonal and annual climatological behavior of selected components of the hydrological cycle are presented from coupled and uncoupled configurations of the atmospheric component of the Community Climate System Model (CCSM) Community Atmosphere Model version 3 (CAM3). The formulations of processes that play a role in the hydrological cycle are significantly more complex when compared with earlier versions of the atmospheric model. Major features of the simulated hydrological cycle are compared against available observational data, and the strengths and weaknesses are discussed in the context of specified sea surface temperature and fully coupled model simulations.

The magnitude of the CAM3 hydrological cycle is weaker than in earlier versions of the model, and is more consistent with observational estimates. Major features of the exchange of water with the surface, and the vertically integrated storage of water in the atmosphere, are generally well captured on seasonal and longer time scales. The water cycle response to ENSO events is also very realistic. The simulation, however, continues to exhibit a number of long-standing biases, such as a tendency to produce double ITCZ-like structures in the deep Tropics, and to overestimate precipitation rates poleward of the extratropical storm tracks. The lower-tropospheric dry bias, associated with the parameterized treatment of convection, also remains a simulation deficiency. Several of these biases are exacerbated when the atmosphere is coupled to fully interactive surface models, although the larger-scale behavior of the hydrological cycle remains nearly identical to simulations with prescribed distributions of sea surface temperature and sea ice.

Corresponding author address: Dr. James J. Hack, NCAR, P.O. Box 3000, Boulder, CO 80307. Email: jhack@ncar.ucar.edu

1. Introduction

The accurate treatment of earth's hydrological cycle (the circulation of water in the climate system) is central to the scientific explorations of climate dynamics and climate change. The global water cycle is an integral part of the global energy cycle and hence plays a fundamental role in determining large-scale circulation and precipitation patterns. The complex web of feedbacks linking hydrological processes with the energy cycle operate over a continuum of time and space scales.

Water can be found in all three phases in the climate system, and is a strongly radiatively active atmospheric constituent in all forms. Clouds, composed of water in the liquid and frozen phases, play a dominant role in regulating the energy budget of the planet, and their behavior remains a major source of uncertainty in our ability to project the effects of climate change (e.g., Stephens and Webster 1981; Cess et al. 1990; Houghton et al. 2001). They cool the earth by reflecting solar radiation back to space, while at the same time warming the planet by absorbing thermal radiation emitted from the surface and lower regions of the atmosphere. The processes responsible for phase transitions of water also contribute to the diabatic forcing of the earth's dynamical circulations, and are key to the overall energy budget, particularly for the thermally driven circulations in the Tropics (Chahine 1992).

The processes of the hydrological cycle operate on a wide range of time and space scales, and are very difficult to quantify observationally. The most reliable observations of the hydrologic cycle are limited to relatively long time and large spatial scales. As such, current observational data provide relatively weak constraints on the formulation of hydrological processes in global models. A major challenge to the design of global climate models is to realistically incorporate the physical processes involved in the hydrological cycle that operate on scales of motion distinctly separate from those of the resolvable large-scale circulation, but strongly influence the behavior of the atmosphere on all motion scales.

The Community Atmosphere Model version 3 (CAM3) is the latest in a succession of general circulation models that have been made widely available to the scientific community, originating with the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM). This model is the atmospheric component of the Community Climate System Model version 3 (CCSM3), which is a fully coupled modeling framework that can be used for a broad class of scientific problems. The CCSM3 represents the latest generation of this modeling framework, and is discussed in more detail by Collins et al. (2006a). CAM3 incorporates a significant number of changes to the dynamical formulation, the treatment of cloud and precipitation processes, radiation processes, and atmospheric aerosols, and is discussed more fully in Collins et al. (2006b). The representation of cloud and precipitation processes has been significantly revised, including separate treatments of liquid and ice condensate, large-scale advection, detrainment, and sedimentation of cloud condensate; and separate treatments of frozen and liquid precipitation (Boville et al. 2006). The parameterization of radiative transfer has been updated to include a generalized treatment of cloud overlap (Collins et al. 2001) and new treatments of longwave and shortwave interactions with water vapor (Collins et al. 2002a, 2004, manuscript submitted to J. Geophys. Res.). Finally, a prescribed climatological distribution of sulfate, soil dust, carbonaceous species, and sea salt based upon a three-dimensional assimilation (Collins 2001; Rasch et al. 2001) is used to calculate the direct effects of tropospheric aerosols on heating rates (Collins et al. 2002b). This latter change is noteworthy in the context of what follows given that the radiative effects of atmospheric aerosols have been shown to strongly influence the behavior of hydrological processes (Ramanathan et al. 2001; Menon et al. 2002).

CAM3 has been designed to provide simulations with comparable large-scale fidelity over a range of horizontal resolutions for several different dynamical approximations. These modifications require adjustments to parameters in the physics package associated with cloud and precipitation processes. Consequently, the detailed hydrological process behavior has some dependence on horizontal resolution and the formulation of the dynamical core. Some of these issues, along with the sensitivity of the simulated climate to model resolution are discussed in more detail in Hack et al. (2006), Yeager et al. (2006), and DeWeaver and Bitz (2006). The standard configuration of the CAM3 is based on an Eulerian spectral dynamical core, where the vertical discretization makes use of 26 levels (L26) treated using second-order finite differences (Williamson 1988). The vertical domain is essentially the same as in earlier models, but employs eight additional levels to better refine the upper troposphere and lower stratosphere. The discussion that follows will focus on the standard CAM3 configuration that uses a 26-level 85-wave triangular spectral truncation (T85L26). This truncation translates to a 1.4° transform grid (∼150 km near the equator) on which nonlinear and parameterized physics terms are evaluated. This reflects a fourfold increase in the number of horizontal degrees of freedom when compared to earlier models (Hack et al. 1994; Kiehl et al. 1998; Kiehl and Gent 2004). A 22-yr five-member ensemble using observed sea surface temperature (SST) and observed sea ice is used to characterize the mean features of the simulated hydrological cycle in the uncoupled configuration. These simulation characteristics are then contrasted with simulation properties obtained from the fully coupled CCSM3.

There are a large number of observational datasets related to the earth's hydrological cycle. For cloud amount we use the Nimbus-7 total cloud data derived from the Temperature Humidity Infrared Radiometer (THIR) and the Total Ozone Mapping Spectrometer (TOMS) measurements for the period April 1979–March 1985 (Stowe et al. 1988, 1989), and the International Satellite Cloud Climatology Project (ISCCP) D2 gridded cloud product data for total cloud cover (Rossow and Dueñas 2004). Several datasets are used to evaluate precipitation. The Global Precipitation Climatology Project (GPCP) version-2 Monthly Precipitation Analysis extends from 1979 to 2003 (Adler et al. 2003). It is a merged dataset consisting of satellite microwave and infrared data and surface rain gauge data. Another global blended dataset, the Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP), covers the period 1979–98 (Xie and Arkin 1997). Several datasets are also used to evaluate water storage in the atmosphere. The National Aeronautics and Space Administration (NASA) Water Vapor Project global water vapor dataset (NVAP) is a water vapor and liquid water path archive that extends from 1988 to 1999. The blended analysis includes satellite retrievals of water vapor from the Television Infrared Observational Satellite (TIROS) Operational Vertical Sounder (TOVS), SSM/I, and radiosonde measurements (Randel et al. 1996). The SSM/I column water vapor and cloud liquid water are derived from satellite microwave radiometers as described in Wentz and Spencer (1998). The Moderate Resolution Imaging Spectroradiometer (MODIS) total column water vapor product was obtained from near-IR and IR algorithms, while cloud liquid water path is derived, along with a number of physical and radiative cloud properties, using IR and visible algorithms (King et al. 2003). We will supplement these datasets with the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) to examine selected features of the hydrological cycle and thermodynamic state, and will compare some simulation results with the predecessor model, CCM3.

2. Mean-state simulation properties: Uncoupled configuration

a. Global mean properties

We begin our discussion of the simulated hydrological cycle by examining the global annual water budget in the CAM3 using a 22-yr five-member ensemble as described earlier. The land surface in these simulations is fully interactive. The simulated global annual precipitation rate of 2.86 mm day−1 is approximately 7% larger than the CMAP satellite estimate, but represents an 8% reduction over the CCM3, the previously documented atmospheric model in this series (Hack et al. 1998). The improvement in the magnitude of the hydrological cycle is largely attributable to changes in the surface energy budget associated with the introduction of specified atmospheric aerosols. Figure 1 shows the breakdown of the precipitation and evaporation exchanges of water in absolute terms between the atmosphere and land, ocean, and sea ice surfaces, along with runoff from the land and ice to the oceans (defined as the difference between precipitation and evaporation). The relative distribution of surface exchange is remarkably similar to renormalized observational estimates (e.g., Peixoto and Oort 1992). The observational estimates of precipitation, evaporation, and runoff are normalized by the CAM3 simulated annual global precipitation rate, and show that all terms in the surface exchanges generally agree to within a few percent of the relative partitioning contained in the original observational estimates.

Figure 1 also shows the time-averaged storage of water in the atmosphere in all three phases as simulated by CAM3, along with estimates of water vapor (i.e., precipitable water) and cloud liquid water retrieved from MODIS. We have used MODIS estimates in this comparison because these retrievals provide the most comprehensive global coverage of cloud liquid water, while recognizing that they may not be the most definitive retrievals of water in its liquid or vapor form. The total reservoir of water in its frozen, liquid, and vapor state is broken down by its distribution over land, ocean, and ice. From this perspective, CAM3 does a reasonable job of simulating the distribution of water with regard to surface type. Since these numbers are strongly weighted by the fractional areas of ocean, land, and sea ice, it is no surprise that the largest fraction for each phase resides over the oceans. Figure 1 suggests that CAM3 generally overestimates precipitable water, overestimates cloud liquid water over the oceans, and underestimates cloud liquid water over land and sea ice.

Another way to look at the distribution of water, and its exchange with the surface, is to quantify the average properties for each of the three underlying surfaces, as shown in Tables 1 and 2. Table 1 shows precipitation and evaporation data for the CAM3, evaporation data for the ERA-40 reanalysis, and precipitation data from the CMAP climatology. We have not included ERA-40 precipitation estimates because of known spinup problems in the analysis of precipitation (S. Uppala 2004, personal communication). The ERA-40 evaporation rates, however, do not suffer as much from hydrological spinup problems, and compare well with the da Silva et al. (1994) climatology (A. Beljaars 2004, personal communication). CMAP provides a useful quasi-independent measure of the magnitude of the hydrological cycle, and the distribution of precipitation across surface types. To some extent, the disagreements in these estimates help illustrate continuing uncertainty in quantifying the magnitude of earth's hydrological cycle, even on long time scales. Although the magnitude of the CAM3 hydrological cycle is larger than the CMAP estimate using global annual precipitation as the measure, it bears a much closer relationship to the reanalysis if we use global annual evaporation as the measure. In a relative sense precipitation rates are greater over land and sea ice in CAM3 when compared to oceanic rates using CMAP estimates as the reference observation.

The components of water storage in the atmosphere are generally difficult quantities to observe on global scales. Much of these data come from satellite retrievals, often blended with in situ and/or analysis data, and are most often limited to vertical integrals of precipitable water and cloud liquid water. Table 2 shows the simulated precipitable water, cloud liquid water, and cloud ice water by surface type for CAM3, along with comparable estimates from NVAP, MODIS, and ERA-40. We include the ERA-40 estimates of cloud condensate for reference since there are no global observational measurements of ice water loading. One can see that there are considerable differences between the various observational estimates, which are of similar magnitude to differences with CAM3. Generally speaking, CAM3 appears to agree reasonably well with NVAP and ERA-40 estimates of precipitable water, which are slightly higher than the MODIS retrievals. Cloud liquid water over the oceans is higher in CAM3 than in any of the estimates, but falls between the ERA-40 and MODIS values over both land and sea ice.

b. Zonal mean properties

The zonally averaged seasonal and annual distribution of precipitation for CAM3 is shown in Fig. 2 in comparison with precipitation estimates from CMAP. In most respects CAM3 exhibits similar biases to those seen in the CCM3 simulation. The amplitude of the tropical precipitation in the intertropical convergence zone (ITCZ) is generally well captured, although there is a slightly more exaggerated double ITCZ than in CCM3, most notably during December–February (DJF). The simulated location of the DJF ITCZ maximum, more than 10° north of the analyzed CMAP maximum, illustrates a major problem with the representation of tropical precipitation, which is the persistence of ITCZ-like precipitation in the Northern Hemisphere year-round. This is in contrast to the observational estimates, which show a clear seasonal migration of ITCZ precipitation across the equator. Subtropical precipitation minima are generally displaced too far poleward seasonally and annually, as are the secondary precipitation maxima in the extratropical storm tracks. The poleward shift of the CAM3 Southern Hemisphere storm track, when compared to CCM3, results in a modest positive precipitation bias when compared to the satellite retrievals.

The zonally averaged seasonal and annual evaporation rate (not shown) is significantly reduced in CAM3 when compared to CCM3, our benchmark for this discussion. There is a pronounced and systematic reduction in surface evaporation at all latitudes compared with CCM3, which is dominated by changes over ocean surfaces and primarily associated with the introduction of a climatologically specified distribution of atmospheric aerosol. The inclusion of aerosol effects produces a significant reduction of absorbed solar radiation at the surface, directly affecting surface evaporation rates. As was the case for the CCM3 atmospheric model, the most vigorous transfer of water to the atmosphere occurs in the subtropics with evaporation rates reaching maximum values near 15°N and 15°S. Consistent with observational analyses, the Southern Hemisphere oceans are the principal source of water powering the atmospheric hydrologic cycle in CAM3. CAM3 exhibits a realistic suppression of evaporation in the vicinity of ITCZ, and is in good agreement with corresponding oceanic estimates (e.g., see Oberhuber 1988; Kiehl and Trenberth 1997; Doney et al. 1998; Large and Yeager 2004). We also note a substantial reduction in the surface flux of water to the atmosphere poleward of 80°N, which introduces a large change in the water budget over the Arctic when compared to CCM3.

The zonally averaged seasonal and annual net surface exchange of water [i.e., evaporation minus precipitation (EP)], is shown for CAM3 and CCSM3 in Fig. 3, and is quantified in units of energy (where 1 mm day−1 ∼29.055 W m−2). CAM3 simulates a strong seasonal meridional oscillation in the source regions, but a relatively weak seasonal movement of the equatorial water sink. The weak meridional excursion of the net water sink in the deep Tropics arises from the unrealistically weak seasonal migration of ITCZ precipitation. The regions 10°–40°N and 10°–40°S are well-defined source regions of total water, where the deep Tropics and high-latitude extratropics represent the principal sinks. In most respects, the CAM3 net water budget bears remarkable similarity to CCM3, especially considering the relatively large local changes to the individual components of the water exchange. The largest changes in EP occur along the equator, and poleward of 60°N. The equatorial differences are largely a consequence of reduced precipitation rates, manifested in the form of a stronger double ITCZ, particularly over the Indian Ocean. Over the Arctic the water deficit is nearly twice as large as in CCM3, largely due to systematic reductions in the surface evaporation rate.

The zonally averaged precipitable water, or vertical integral of the specific humidity, is shown in Fig. 4, along with the NVAP climatology. CAM3 is systematically moister than the CCM3, and in better agreement with zonally averaged observational estimates such as NVAP. The largest bias is present year-round near 30°N, exceeding 4 kg m−2 (or 4 mm) over most of the region between 10° and 40°N. As we will show, the agreement in the zonal mean distribution of precipitable water is the consequence of a fortuitous cancellation of errors in the longitudinal distribution.

As mentioned earlier, clouds provide important forcings on the climate system through their modulation of the radiative heating field. The climatological distribution of cloud and cloud condensate is therefore worthy of some discussion. The annually and zonally averaged meridional distributions of total cloud amount are shown in Fig. 5 for CAM3, ISCCP, Nimbus-7, and MODIS. The CAM3 cloud field is markedly different from CCM3. Total cloud cover in the Tropics and poleward of the extratropical storm tracks is significantly reduced in CAM3. This is dominated by a sharp reduction in high cloud over most of the globe, and reductions in mid- and low-level cloud at high latitudes. The reductions in tropical high cloud are compensated by increases in midlevel cloud, where low-level cloud has systematically increased equatorward of 60°N and 60°S. The reduction in high cloud is more consistent with ISCCP estimates, while the increase in low-level cloud amount is more consistent with the surface-based observations of Warren et al. (1988). Despite some improvement in the distribution of simulated cloud amount, important biases in its meridional distribution remain in CAM3. One of the more obvious long-standing deficiencies is the location of the minima in subtropical cloud cover, near 20° latitude in the observational record, but closer to 30° in the CAM3 simulation. This difference has important consequences for the radiative budget of the subtropics where, for example, the tropical shortwave cloud forcing is too broad and therefore too strong for most of the subtropics. The radiative issues related to the CAM3 simulation of cloud and cloud optical properties is not the focus of this manuscript, but will be discussed elsewhere using the ISCCP simulator developed by Klein and Jakob (1999) and Webb et al. (2001).

As noted in Table 2, condensed water in the atmosphere is several orders of magnitude smaller than storage in the vapor state, and yet is of comparable climate importance in terms of modulating the global radiation balance (e.g., Wielicki et al. 1995; Kiehl 1994). Zonally and annually averaged distributions of liquid water path are shown in Fig. 6 for CAM3, and for several ocean-only satellite-derived products. CAM3 exhibits sharply defined maxima for cloud liquid water in the deep Tropics and at 60°N and 60°S. Simulated cloud water in the Tropics and subtropics agrees most closely with the SSM/I retrievals of Wentz and Spencer (1998), and represents a ∼30% overestimate of cloud water in the ITCZ for both the MODIS and NVAP retrievals. Cloud water in the extratropical storm tracks is approximately twice as large as in the ITCZ, and approximately twice as large as diagnosed by any of the available retrievals. The one exception is a new MODIS retrieval under development by members of the NASA Clouds and the Earth's Radiant Energy System (CERES) Science Team, which shows high-latitude cloud liquid water paths of comparable magnitude to the CAM3 simulation (P. Minnis 2004, personal communication). Unlike the CCM3, the seasonal behavior of the simulated zonal average of cloud liquid water does not show a strong seasonal oscillation at high latitudes. The June–August (JJA) simulation shows the strongest departure from the annual mean distribution, with slightly enhanced liquid water paths in the ITCZ and Northern Hemisphere high latitudes. Similar behavior is seen in the various observational estimates, although the relative biases discussed earlier remain.

A quantity for which little in the way of globally observed data are available is the ice water path. Zonally, annually, and seasonally averaged distributions of the ice water path as simulated by CAM3 are shown in Fig. 7. As is the case for cloud liquid water, cloud ice water exhibits large differences between the Tropics and extratropics. Southern Hemisphere extratropical ice water paths are nearly 3 times as large as in the ITCZ, exceeding 40 gm m−2. Unlike the liquid water distribution, ice water has a very strong seasonal cycle at high latitudes, with maximum extratropical values occurring in the respective winter hemispheres. There is also a much stronger seasonal shift in cloud ice at low latitudes with greater tropical ice water loading during boreal summer.

c. Vertical structure

Temperature and water vapor are the two state variables that jointly define the moist static stability of the atmosphere. The ability to properly simulate the vertical distribution of water vapor is strongly constrained by biases in the simulated temperature structure. Figure 8 shows the CAM3 annual zonal average differences of temperature, specific humidity, and relative humidity from the ERA-40 reanalysis climatology. Overall, CAM3 does a relatively good job of reproducing the analyzed thermal structure. Simulated temperatures are within 1–2 K of the analyzed field for most of the domain equatorward of 50°N and 50°S. The CAM3 temperature simulation represents a modest improvement over CCM3. Tropical tropopause errors are nearly halved when compared to CCM3, and high-latitude lower-tropospheric temperatures have been significantly warmed, in better agreement with observations. A sizable part of the warming change is associated with the increased horizontal resolution, most notably in high-latitude mid- to upper-tropospheric temperatures (see Hack et al. 2006). Improvements to the formulation of the CAM3 cloud processes explain the remaining improvements to the temperature simulation, particularly for the lower troposphere at high latitudes (Boville et al. 2006). Despite these improvements, the difficulty in properly simulating polar tropopause temperatures remains a long-standing documented problem for atmospheric general circulation models (Boer et al. 1992).

Global observational data on the vertical distribution of water in the atmosphere are still nonexistent, where analysis products provide the best available estimates. Atmospheric analyses continue to contain uncertainties in the moisture field (e.g., Trenberth and Guillemot 1995), since the water vapor distribution strongly depends upon the parameterized treatment of processes involved in the hydrological cycle. Nevertheless, comparison of the reanalysis products against locally available radiosonde observations suggest that the vertical structure of the biases shown in Fig. 8 are robust. These zonally averaged biases generally show a wetter than analyzed atmosphere throughout most of the domain. The major exception is the meridionally broad low-level dry bias between 600 and 900 mb in the Tropics, exceeding 1 gm kg−1 in the zonal annual mean. Generally, relative humidity errors are within 10% of the analyzed field, where the exceptions are in the vicinity of the lower-stratospheric cold biases and in the lower troposphere over Antarctica. The overall structure of the water vapor bias is similar to CCM3, but slightly larger in amplitude. Preliminary analyses of this error suggest that the vertical structure in the Tropics, such as the positive water vapor anomaly at 500 mb, is strongly determined by the form of parameterized moist convection. One in situ example of the large-scale water vapor bias is shown in Fig. 9, which illustrates climatological vertical profiles of θe and specific humidity over Yap Island in the tropical west Pacific during the month of July. These figures show how the ERA-40 reanalysis compares to radiosonde data, and that the lower-tropospheric dry bias and mid- to upper-tropospheric moist bias are robust features of the CAM3 simulation. The dry bias maximizes near 850 mb locally reaching 3 gm kg−1. Water vapor biases of this magnitude and structure have a significant impact on the moist static stability of the tropical atmosphere, as is seen in the θe profiles, and are likely to play an important role in the low-latitude dynamical response to diabatic heating.

The zonal average of the simulated vertical distribution of condensate is shown in Fig. 10. The color-shaded regions show liquid water concentration, and the contoured regions show ice water concentration. The location of the freezing level is also shown for reference. Most of the liquid water shown in Fig. 6 resides below 900 mb with concentrations ranging from 0.05 to 0.15 gm m−3 in the zonal annual mean. The mid- and high-latitude extratropics exhibit very strong vertical gradients, while the vertical distribution in the deep Tropics is considerably more diffuse. The cloud ice water distribution generally reaches its maximum concentration several kilometers above the freezing level, between 500 and 600 mb in the extratropical storm tracks and near 300 mb in the deep Tropics. Maximum ice water concentrations reach 0.007 and 0.003 gm m−3 in the zonal annual extratropical and deep tropical means, respectively. The high-latitude seasonal swing in ice water path is primarily determined by changes in ice water concentration in the lowest kilometer of the atmosphere in the Northern Hemisphere. The Southern Hemisphere seasonal cycle is largely determined by ice water loading changes throughout the troposphere. Also, much of the high-latitude cloud condensate consists of mixed phase clouds, whereas ice and liquid water regimes are more clearly separated in the Tropics.

Finally, we illustrate the vertical structure of the meridional transport of water vapor in Fig. 11. Mean meridional water transport is shown in the left panels, and transient meridional water transport is shown in the right panels. As shown in Fig. 3, the deep Tropics are a sink of moisture, the subtropics are a source of moisture, and regions poleward of 40° are sinks of moisture. As might be expected, the transport of water from the subtropics into the deep Tropics is generally confined to the lower 1500 m of the atmosphere and largely handled by the mean meridional circulation. This transport exhibits the strong seasonal asymmetries associated with the Hadley circulation (Trenberth et al. 2000). The majority of water vapor transport to higher latitudes occurs over a slightly deeper portion of the atmosphere, occurring largely in the form of transient eddy transport. Although much weaker than the equatorward transport by the Hadley circulation, a third of the total poleward transport occurs in the indirect or Ferrel circulation. At higher latitudes eddy transports toward the Pole dominate transport by the mean circulation (polar cell), which acts to move water vapor from the polar regions to lower latitudes.

d. Horizontal structure

In this section, we will examine the horizontal distribution of vertically integrated measures of water storage and surface water exchange in CAM3. We begin with the annually averaged precipitation field shown in Fig. 12. Although the CAM3 simulation captures many of the observed features in the global precipitation distribution, it continues to share many of the same biases exhibited by CCM3. Most of the available retrieval data agree that the most serious simulation errors occur in the form of excessive precipitation over the western Indian Ocean, the central subtropical Pacific, and in the vicinity of Central America. CAM3 also continues to underestimate the strength of the Atlantic ITCZ. Another long-standing simulation deficiency is a tendency for the simulated tropical precipitation maxima to remain in the Northern Hemisphere year-round, and a slightly greater tendency for reduced precipitation along the equator, particularly over the Indian Ocean. This is in sharp contrast with most observational datasets, which show a clear seasonal migration of ITCZ precipitation across the equator.

The precipitation anomalies in the western Indian Ocean are related in part to deficiencies in the Zhang–McFarlane closure assumptions, a topic which has been explored by several investigators (Xie and Zhang 2000; Zhang 2003). The most serious manifestation of these problems appears in the form of excessive precipitation rates over the Arabian Peninsula during the Northern Hemisphere summer months. Other factors contribute to this bias throughout the year, including the tendency to inaccurately shift precipitation to the northern Indian Ocean during the boreal winter, coupled with an overactive and displaced precipitation regime to the west and northwest of Madagascar extending from the Mozambique Channel into the Indian Ocean east of Tanzania. The boreal summer also exhibits an extremely overactive precipitation regime just north of the equator and 1000 km to the southwest of the Indian subcontinent. The excessive precipitation in the central Pacific subtropics is associated with two simulation challenges. CAM3 continues to have difficulty in properly positioning the South Pacific convergence zone (SPCZ), which is too strong in amplitude and too zonal in structure, and does not extend far enough into the southern extratropics. The SPCZ also has a tendency to extend too far east, another symptom of the tendency for the model to produce a double ITCZ. The northern Pacific bias is associated with a poor simulation of the very well defined precipitation pattern that extends from the South China Sea through the Philippine Sea and into the tropical equatorial Pacific during the months of July and August. This precipitation pattern is represented as a relatively diffuse extension of the Southeast Asian monsoon well into the central Pacific subtropics, and is a long-standing precipitation bias in the CCM and CAM models. Other notable features of the precipitation distribution is the inability to capture the seasonal cycle of precipitation off the west coast of Central America, and weaker than analyzed precipitation rates along the extratropical western boundary currents. Precipitation over the land surfaces generally tends to be excessive, especially over the Congo. Exceptions include large areas over the Amazon basin, and much of United States east of the Continental Divide. Finally, simulated precipitation rates in the high-latitude extratropical storm-track regions continue to be slightly higher than current satellite retrievals suggest.

The simulated evaporation field (not shown) illustrates the important role played by the ocean surface as a source of water vapor to the atmospheric general circulation. As suggested by the zonal means, both the northern and Southern Oceans contribute to the evaporation of water vapor year-round, but with important seasonal longitudinal migrations of evaporation centers. The CAM3 simulation exhibits a clear evaporation minimum in the ITCZ year-round, with extensive regions of high evaporation in the respective winter hemispheres. The boreal winter includes evaporation maxima along the western boundary currents (the Kuroshio and Gulf Stream), the Red Sea, the eastern Bay of Bengal and eastern equatorial Pacific, all of which exceed 10 mm day−1 (∼290 W m−2) in the 3-month seasonal mean. Other features include maxima in the western subtropical Pacific, the western equatorial Atlantic, and southern Indian Ocean with evaporation rates >6 mm day−1 (∼175 W m−2). As was the case with CCM3, broad regions of evaporation are also seen over much of South America and southern Africa exceeding 4 mm day−1 (∼120 W m−2) in the seasonal average. During the boreal summer, the well-defined evaporation centers transition to an extensive region of high evaporation across the Southern Ocean, with maxima in the southern Indian Ocean and tropical west Pacific. Evaporation maxima in the northern Pacific Ocean migrate eastward to the vicinity of the Hawaiian Islands with maximum evaporation rates of 6 mm day−1. The principal evaporation regime in the Atlantic migrates into the Southern Hemisphere, and continental evaporation moves into the Northern Hemisphere, most notably eastern North America, India, large portions of East and Southeast Asia, and sub-Saharan Africa.

Together, the evaporation and precipitation fields define the properties of freshwater exchange between the atmosphere and the earth's surface. The annually averaged horizontal distribution of EP is shown in Fig. 13 for CAM3. We note that a comparable global observational dataset does not exist. The principal tropical precipitation features are clearly visible. Local water deficits in the ITCZ generally exceed 4 mm day−1 in the annual mean. The eastern Pacific subtropics, central Atlantic subtropics, and southern Indian Ocean subtropics, are the principal sources of water for the atmosphere. CAM3 simulates a large seasonal cycle in EP over much of South America, central and southern Africa, India, and Southeast Asia, mostly a reflection of the seasonal migration of deep convection in response to solar insolation. Similar seasonal variability is seen over most of Europe extending into central Asia, and over much of North America. Most of Europe and a large portion of North America can be clearly characterized as water source regions during JJA, and water sink regions during DJF.

The horizontal distribution of the annually averaged precipitable water, and its difference from the NVAP climatology, is shown in Fig. 14. To a large extent, CAM3 does a very good job of capturing the structure and correct magnitude of precipitable water in the atmosphere. There are, however, important large-scale systematic biases, despite exceptionally good agreement in the zonal mean structure. The longitudinally compensating arrangement of these biases is responsible for the good agreement in the zonal mean, where some of the regional biases are well correlated with biases in the precipitation distribution. Precipitable water is generally overestimated throughout most of the Pacific basin, in the western Indian Ocean, Arabian Sea, and central Africa. In sharp contrast, the simulation exhibits a large spatially coherent dry region stretching from the Americas, across the equatorial Atlantic, northern Africa, and into southern and Southeast Asia. In general terms, the simulation is systematically dry over continental regions, most notably during the warm season. These water vapor biases are locally significant, particularly over Saharan Africa where they can exceed 10 mm, or often half of the observed precipitable water.

Figure 15 shows the annual global distribution of cloud liquid water and cloud ice for the CAM3 simulation. The extratropical storm tracks, features of the low-latitude tropical circulation, such as the subtropical subsidence regimes, and continental deserts, are all clearly visible in the cloud liquid water field. The liquid water path frequently exceeds 200 gm m−2 in the storm tracks, in contrast with many of the much lower satellite-retrieved cloud liquid water paths. Liquid water loading at low latitudes is generally in better agreement with satellite-derived values, although pathlengths in the subtropical subsidence regimes are considerably smaller, particularly in the Southern Hemisphere. This is a surprising feature of the simulation, given the greater than observed cloud radiative forcing of these regions in the simulation. The cloud ice shares many of the same regional characteristics as the cloud water. The tropical distribution is highly correlated with areas of deep convective activity, such as over the Congo, western Indian Ocean, tropical west Pacific, and the Amazon. As suggested by the zonal means in Fig. 7, significantly greater ice water loading is found at high latitudes in the storm-track regions, where ice water paths are frequently well in excess of 2 times the maximum ice water paths seen in the Tropics.

3. Low-frequency forced variability: Uncoupled configuration

The seasonal cycle and changes to the equatorial SST distribution associated with El Niño–Southern Oscillation (ENSO) are two examples of major modes of low-frequency variability in the climate system. These are essentially forced modes of variability in uncoupled integrations of CAM3, and provide a useful basis for evaluating the simulated local and far-field responses as compared to observations.

The CMAP and GPCP analyses of global precipitation provide an observational opportunity to quantitatively examine the CAM3 simulated precipitation response to ENSO. Figure 16 is a Hovmöller plot of precipitation anomalies averaged over the deep Tropics as estimated by CMAP and simulated by CAM3 for the period January 1979–December 2000. The CMAP product shows the evolution of strong positive and negative precipitation anomalies in response to the warm and cold phases of the observed ENSO cycle. Generally, CAM3 does extremely well at capturing both the structure and amplitude of the anomaly pattern in the central and eastern Pacific. The eastward extension of the warm phase anomalies are well reproduced. The most serious deficiency is in the simulation of the anomaly pattern west of the Maritime Continent into the Indian Ocean, which is much more weakly simulated when compared to observations.

A second way of examining the response of the hydrological cycle to ENSO is to explore the spatial pattern of the anomaly response associated with the time-averaged precipitation difference between a specific warm and cold event. This approach has the advantage of amplifying the response to the ENSO cycle. Figure 17 shows the monthly averaged precipitation difference between July 1994 (warm phase) and June 1999 (cold phase) as analyzed by GPCP and as simulated by CAM3. Both panels show a very large positive precipitation anomaly stretching across the central equatorial Pacific flanked by negative anomalies to the north, west, and south (in the SPCZ). The CAM3 simulation does a very good job of representing the structure and amplitude of the positive anomaly. The structure and magnitude of the negative anomaly response is not as well represented, particularly in the western and eastern equatorial Pacific. The western Pacific anomaly is too strong immediately north of the equator and too weak along the equator. This response pattern is consistent with the weak time-dependent response shown in the Fig. 16 Hovmöller diagram. Nevertheless, the CAM3 simulation does a remarkably good job of capturing the overall pattern and amplitude of the responses, including the far-field responses seen in the Atlantic and the western Indian Ocean. An important exception is the rainfall anomaly over the Amazon basin, which is poorly represented.

Finally, we examine the ability of the CAM to simulate the seasonal migration of water vapor between the Northern and Southern Hemispheres, which represents a regular major meridional redistribution of mass in the atmosphere, and has an impact on the earth's angular momentum budget (e.g., Lejenäs et al. 1997). As seen in the zonal means in Fig. 4, CAM3 correctly simulates a strong seasonal meridional migration of precipitable water. Figure 18 shows this seasonal redistribution of water vapor by subtracting the JJA distribution of precipitable water from the DJF distribution. Despite the local biases discussed earlier, the seasonal redistribution of water vapor is well represented in CAM3. The structure of the seasonal response is very similar to the observed climatology, even on relatively small spatial scales. There are some large-scale systematic biases, such as the slightly weaker seasonal cycle in the Northern Hemisphere, and slightly stronger seasonal cycle in the Southern Hemisphere, that lead to local differences in amplitude. But the properties of this mode of variability are generally well represented in the CAM3 simulation. These results demonstrate the ability of the CAM3 hydrological cycle to respond to externally imposed low-frequency forcing.

4. Mean-state simulation properties: Coupled configuration

In this section we provide an overview of the hydrological cycle as represented in the CCSM3 coupled simulations, which employ CAM3 as the atmospheric component. We will examine the principal differences in the hydrological cycle as simulated by the atmosphere, along with the more important features of the hydrological cycle as seen from the perspective of the land, ocean, and sea ice component models. This discussion will employ a standard CCSM control simulation in which the atmosphere and land models are represented on a T85L26 transform grid, and the ocean and sea ice models make use of a nominal 1° horizontal discretization (T85 × 1). This T85 × 1 configuration of the coupled model, which we will generically refer to as CCSM3, has been used to document the CCSM3 simulations for international climate-change assessments (see Collins et al. 2006a).

a. Atmosphere

In an overall sense, the CCSM3 atmospheric global water and energy cycle budget remains remarkably similar to the uncoupled CAM3 simulation. The top of the atmosphere energy budget remains within 0.2 W m−2, while the individual components of the surface energy balance remain well within 1 W m−2 in the global annual mean. The global cycling of water in the CCSM3 atmosphere is nearly identical to the characterization shown in Fig. 1 for the uncoupled model. The magnitude of the global hydrological cycle is reduced by approximately 1%, primarily due to a reduction in the magnitude of the water exchange over land and sea ice, but with comparable levels of runoff. Global annual storage of water vapor and condensate in the atmosphere also remains well within 1% of the uncoupled control simulation. On seasonal time scales, differences in measures of the global water cycle vary only slightly more than in their respective annual means.

Although global annual measures of the hydrological cycle are virtually identical, the detailed regional behavior of the simulated hydrological cycle in coupled mode includes some notable differences. These anomalies can be seen in the zonal mean quantities related to the exchange and storage of water in the atmosphere (see Figs. 2, 3 and 4). There is a remarkable shift in the surface exchange of water from the Northern to Southern Hemisphere Tropics in the coupled model. Northern Hemisphere tropical precipitation rates are reduced by 1 mm day−1 in the zonal annual mean, and are enhanced by more than twice this rate near 10°S. This meridional shift produces a significant and unrealistic change to the freshwater budget over the tropical oceans, most notably during the boreal winter (see section 4b). Although precipitation anomalies appear in both the Atlantic and Pacific basins, the zonal mean anomaly is dominated by changes over the Pacific. This takes the form of an unrealistic enhancement of a southern and more vigorous branch of ITCZ convection extending across the Pacific basin from the warm pool to the Ecuador coast (see Fig. 19). The change to the precipitation distribution is symptomatic of the so-called double-ITZC problem that plagues many coupled models (e.g., see Davey et al. 2002). Despite the overestimated precipitation rates in the southern tropical Pacific and southeastern tropical Atlantic, several other features in the precipitation distribution are significantly improved in the coupled configuration. These include precipitation over Central America and the Caribbean, along the western midlatitude boundary currents, over the Arabian Peninsula, and over the northern Indian Ocean. Precipitation reductions in the north-central subtropical Pacific also represent modest improvements when compared to the uncoupled simulation.

Changes in the precipitation distribution are associated with similar shifts in the storage of water in the atmosphere. Precipitable water moves from the Northern to Southern Hemisphere showing a double-peaked tropical distribution in the zonal mean, which is dominated by anomalies that maximize in DJF. Large positive anomalies exceeding 10 mm appear in the south-central tropical Pacific and southeastern tropical Atlantic. Negative anomalies of similar magnitude are located over much of the tropical and subtropical Northern Hemisphere with maxima centered over the Arabian Peninsula and Central America. Generally speaking, the coupled simulation of the precipitable water distribution is further degraded when compared to observational estimates. The cloud condensate distribution reflects the changes to the distribution of precipitation and precipitable water. Cloud water and cloud ice follow the convective source regions, which have migrated to the Southern Ocean. These changes to the horizontal distribution of water strongly impact the energy budget at both the top of the atmosphere and the surface. Large local anomalies are seen in both clear-sky and all-sky radiative fluxes at the surface and at the top of the atmosphere, exceeding 40 W m−2 for all-sky fluxes. Since precipitation and radiative processes are integrally involved in driving the tropical circulation, significant changes to the low-level wind field are also seen in the central Pacific and eastern Atlantic. These changes give rise to differences in the meridional distribution of surface latent and sensible heat transfers, further affecting the freshwater and surface heat budgets over the oceans.

An example of the CCSM3 tropical variability of precipitation is shown in the right Hovmöller panel in Fig. 16. This shows much weaker tropical variability than in the uncoupled model. The response is related to the strength of CCSM3 simulated ENSO events (as opposed to specified ENSO events in the uncoupled simulation), coupled with the altered dynamical structure of the deep Tropics and a greater tendency for most of the simulated precipitation to occur away from the equator.

b. Ocean

Ocean transport of freshwater prevents the development of significant local salinity trends in regions where the mean net freshwater flux is strongly positive or negative. The net freshwater flux (see Fig. 13) is a function of precipitation and evaporation, which have a geographic distribution determined by large-scale atmospheric circulations, and to some extent have a strong meridional structure. The ocean's principal role in the hydrological cycle is to transport the net positive freshwater flux resulting from precipitation in the deep Tropics and high latitudes toward the midlatitude evaporation zones. The ocean also moves freshwater poleward from ice-melting regions to ice-forming regions where the mean net freshwater flux is negative. Last, the ocean must redistribute the large, highly localized influx of freshwater arising from river runoff.

Net freshwater flux into the ocean as a function of latitude is shown in Fig. 20, which compares the fully coupled CCSM3 control (T85 × 1) to an uncoupled ocean model solution (× 1ocn), as well as to an estimate of climatological freshwater flux derived from observed atmospheric and ocean datasets spanning 1984–2000, as described in Large and Yeager (2004, hereafter referred to as LY04). The uncoupled ocean and observationally estimated LY04 curves track each other closely, since they use the same blended precipitation dataset and the same climatological gauge estimate of river runoff. Although there is significant uncertainty, these data provide a reference measure of real freshwater fluxes, with and without polar processes, which can be contrasted with the coupled model solution.

The uncoupled simulation and LY04 curves deviate at high latitudes because the ocean model includes ice formation and ice melt flux algorithms for which there are no corresponding observational datasets. Thus, at extreme polar latitudes the uncoupled simulation shows large negative freshwater fluxes where ice formation results in the rejection of salt to the ocean. The uncoupled solution indicates more positive freshwater input than LY04 near ice edge latitudes (65°S, 70°N) where melting takes place. As in the uncoupled ocean, the CCSM3 freshwater flux is negative at high polar latitudes, and shows a jump to positive values near the ice edge, which exceed the observed flux estimate. Compared to both observationally constrained benchmarks, the coupled model exhibits excessive freshwater flux to the ocean between 45° and 65° and an insufficient flux between 20° and 45°, in both hemispheres. The Southern Hemisphere excess (point A) arises primarily from excessive precipitation combined with an equatorward displacement of the maximum in ice melt flux. The positive anomaly in ice melt flux is related to overly extensive ice coverage in the Atlantic and Indian Ocean sectors of the Southern Ocean (Holland et al. 2006).

The CCSM3 midlatitude freshwater flux deficit near 30°S (point B) arises from a lack of precipitation compared to observations as well as a systematically larger evaporation flux over these latitudes. The Northern Hemisphere high latitudes are also characterized by excessive precipitation and an equatorward shift in ice melt. However, most of the excess freshwater flux in the vicinity of 60°N (point C) is due to much higher than observed river runoff fluxes in the Arctic (see section 4c). As in the Southern Hemisphere, there is less precipitation and more evaporation between 20° and 45°N in the CCSM3 simulation, resulting in a freshwater flux deficit (point D).

The freshwater flux to the coupled ocean is most unrealistic in the Tropics, where the double ITCZ creates a spurious peak in zonal mean precipitation near 10°S. A peak in the CCSM3 freshwater flux at the equator (point E) is related to colder SSTs, which generate less evaporation in the Pacific, combined with excessive precipitation in the western equatorial Pacific and Indian Ocean. The positive flux bias south of the equator in Fig. 20 is further exacerbated by excessive runoff from the Congo. River runoff anomalies are related to precipitation anomalies over continents, which are discussed in section 4c.

The biases in the CCSM3 zonal mean surface freshwater flux give rise to biases in the mean meridional transport of freshwater by the ocean. In Fig. 21, the northward freshwater transports of the coupled and uncoupled ocean models, computed from Eulerian mean advection, are compared, where the uncoupled model results are used in lieu of observations (eddy transports are not included). Both the coupled and uncoupled results show poleward freshwater transport at high latitudes associated with ice growth/melt processes. While the uncoupled model shows significant freshwater transport south across the equator (about one-third of which occurs in the Atlantic basin), the global zonal mean freshwater transport across the equator in the coupled simulation is very near zero. In CCSM3, the negative freshwater flux zones at the southern midlatitudes (10°–40°S) receive freshwater via ocean transport from southern high latitudes as well as from the southern Tropics, where freshwater input is too high. Since observational estimates of tropical precipitation are much more asymmetric about the equator than as simulated by CCSM3 (see Fig. 20), the uncoupled model transports freshwater southward across the equator in each ocean basin in order to balance the southern midlatitude evaporation zones. The CCSM3 ocean transports about as much freshwater northward across the equator in the Atlantic as the uncoupled simulation transports southward across the equator, and there is much less transport of freshwater southward across the equator in the Pacific.

The overly active hydrological cycle is manifested in the form of excessive midlatitude evaporation in the Southern Ocean, which is related to excessive precipitation in the southern deep Tropics. The ocean must therefore transport more freshwater poleward from the Tropics than is estimated from observations, much of it northward. Excessive high northern latitude river runoff (section 4c) results in too much freshwater transport southward to the extratropics, contributing to subtropical sea surface salinity (SSS) and SST in the south that are fresher and warmer than observed (Large and Danabasoglu 2006). This suggests that the excess tropical precipitation transported to the subtropics by the ocean renders the midlatitude upper ocean too fresh and stable, thus inhibiting deep mixing that would lower the SST. Anomalously high SST results in the enhanced evaporation rates that, after atmospheric transport back to the Tropics, recurs as excessive precipitation. Process studies indicate that, at least for the Atlantic basin, correcting the west coast ocean SST bias greatly reduces the excessive tropical precipitation seen in the CCSM3 simulation (Large and Danabasoglu 2006).

c. Land

The hydrological budget over land in CCSM3 is a balance between precipitation, evapotranspiration, runoff, and storage changes in soils or snow. As seen in Table 3, there is no appreciable change in storage during the time period analyzed here. Both annual mean precipitation and runoff compare favorably with observations, with precipitation being about 3% high and runoff about 4% low if glaciers are included in the model estimate, and about right when glaciers are excluded. We note that observations of runoff do not include contributions from most of Greenland, or any runoff associated with Antarctica.

Reliable estimates of global evapotranspiration are not available. However, Brutsaert (1984), based on a number of estimates, proposes that evapotranspiration is about 60%–65% of precipitation. Simulated evapotranspiration in the CCSM3 is 63% of precipitation. Evaporation from the ground is the largest component of the simulated evapotranspiration (59%) followed by canopy evaporation (28%) and transpiration (13%). However, observational estimates of the partitioning of global evapotranspiration suggest that transpiration should be the dominant component followed by ground evaporation and canopy evaporation. In particular, Choudhury et al. (1998), using a process-based biophysical model of evaporation validated against field observations, found that the partitioning was 52% (transpiration), 28% (ground evaporation), and 20% (canopy evaporation). Furthermore, since photosynthesis is coupled to transpiration through stomatal conductance, the underestimate of transpiration has implications for the accurate treatment of carbon assimilation. Global photosynthesis is about 57 PgC, which appears to be about 50% low (Table 3).

The dominant form of runoff in CCSM3 is surface runoff (52% of total runoff); followed by drainage from the soil column (41%); and runoff from glaciers, lakes, and wetlands (7%). This latter runoff term is calculated from the residual of the water balance for these surfaces. This term may also be nonzero for other surfaces as well because the snowpack is limited to a maximum snow water equivalent of 1000 kg m−2.

Zonal annual average values of the land surface hydrological cycle are shown in Fig. 22. As discussed earlier, the CCSM3 simulation overestimates precipitation north of 45°N, generally underestimates it in the northern Tropics, and overestimates it in southern portions of South America. The meridional distribution of evaporation over land areas generally tracks the precipitation with a maximum in the Tropics. Runoff biases also track precipitation biases suggesting that improvements in the simulated precipitation would lead to improvements in the simulation of runoff. At high latitudes, the primary active hydrological component is runoff. At other latitudes, ground evaporation generally dominates. An exception to this is in the deep Tropics (10°S–10°N) where canopy evaporation is equally important. Transpiration is the smallest component of evaporation at all latitudes.

Total runoff from the land model is routed to the ocean using a river transport model (Oleson et al. 2004). Thus, biases in runoff have the potential to affect sea surface salinity and regional ocean circulations, as discussed in section 4b. The annual river discharge into the global ocean is 1.33 Sv (1 Sv ≡ 106 m3 s−1). Excluding Antarctica, river discharge is 1.25 Sv, which is about 6% higher than the estimate of Dai and Trenberth (2002). The river transport scheme does not account for loss of water due to human withdrawal or impoundment of water, seepage into groundwater, or evaporation from the river channel. In particular, consumption of water for irrigation may account for some of the discrepancy. Döll and Siebert (2002) estimate net and gross global irrigation requirements as 0.035 and 0.078 Sv, respectively. The loss of freshwater from Antarctica is estimated to be 0.07 Sv by Vaughan et al. (1999), which is the same as what is simulated by CCSM3. However, this comparison is fortuitous because the majority of Antarctic runoff from CCSM3 comes from the capping of snow over glaciers. More detailed glacier models need to be incorporated into CCSM3 to properly describe glacial processes including iceberg calving and basal melting. There are also notable deficiencies in the modeled discharge at certain latitudes that arise from the land runoff fields. In particular, discharge from the Amazon and the Congo is 42% low and 109% high, respectively.

The deficiencies in partitioning of evapotranspiration described earlier are most evident in the hydrological budget of the Amazon basin (not shown) and figure strongly in the runoff biases. The partitioning of annual evapotranspiration in the model is 49% canopy evaporation, 30% ground evaporation, and 21% transpiration. As discussed in Dickinson et al. (2006), too much water is intercepted by the canopy and reevaporated in the wet season, thus resulting in limited water availability for plant roots particularly in the dry season. Photosynthesis exhibits a significant decline in the dry season, which would affect the ability of a dynamic vegetation model to correctly simulate the composition of vegetation in this region (Bonan and Levis 2006). The year-round warm bias in this region, most pronounced in the dry season, is further confirmation that the simulation is deficient.

While improvements in the precipitation field supplied by the atmosphere model would help to improve the land hydrologic simulation in places like the Amazon basin, the coupled simulation shows that there are aspects of the land hydrology that clearly require attention. Current research is focused on improving the sunlit/shaded treatment of photosynthesis, stomatal conductance, and transpiration, and the parameterization of canopy interception.

d. Sea ice

The seasonal formation and melting of sea ice represents an important component of freshwater transport in the climate system. As ice grows from seawater, it rejects salt back to the ocean, resulting in a relatively fresh ice cover with approximately 4-ppt salinity. If ice dynamics is excluded and equilibrium climate conditions are considered, the local ice growth is balanced by local ice melt and the net long-term mean sea ice freshwater flux to the ocean is zero. Even under thermodynamic-only conditions, however, the considerable seasonal cycle of the ice/ocean freshwater flux can modify the ocean buoyancy forcing and influence ocean mixing. When sea ice dynamics are considered, the transport of relatively fresh sea ice redistributes water in the system, influencing the global hydrological cycle. This has the potential to modify the large-scale ocean circulation in both the Southern (e.g., Goosse and Fichefet 1999) and the Northern (e.g., Holland et al. 2001) Hemispheres.

In the Southern Hemisphere there is net sea ice growth along the Antarctic continent which is then transported equatorward. As the sea ice has only 4-ppt salinity, this ice volume transport is nearly equivalent to a freshwater transport. The transport reaches a maximum of approximately 0.25 Sv at 65°S. Since there are only sparse observations of Southern Hemisphere ice thickness (e.g., Timmermann et al. 2004), no standard observations of Southern Hemisphere meridional ice transport exist. Estimates from (Weatherly et al. 1998) suggest a maximum value of between 0.05 and 0.1 Sv. All indications are that the simulated transport is too large. Although CCSM3 ice motion compares quite well to satellite-derived estimates, the ice is excessively thick, particularly in the Weddell Sea (Holland et al. 2006). This results in the high values of meridional ice transport. This excessive transport and melting along the ice edge, modifies the ocean sea surface salinity conditions, resulting in a fresh bias along the Antarctic sea ice edge in the southwestern Atlantic.

In the CCSM3 simulation, the long-term average freshwater storage in Antarctic sea ice equals 15 630 km3. This corresponds to an annual average area of 12 × 106 km2 with an average thickness of approximately 1.4 m and a salinity of 4 ppt. As discussed in Holland et al. (2006), the simulated area of Antarctic sea ice is large compared to observations, which have an annual average of approximately 9–10 × 106 km2.

In the Northern Hemisphere, there is net sea ice growth in the Arctic basin, resulting in a net loss of water from the Arctic Ocean. The Arctic ice is transported by winds and ocean currents and enters the North Atlantic through Fram Strait. This provides an important source of freshwater to the Greenland–Iceland–Norwegian Seas and has the potential to influence oceanic deep-water formation in this region (e.g., Holland et al. 2001). The annual mean flux of sea ice through Fram Strait in the CCSM3 control integration is 0.08 Sv. This agrees well with the observed estimates of 0.09 Sv given by Vinje (2001) and 0.07 Sv given by Kwok et al. (2004). The flux has a considerable annual cycle (Fig. 23), which also agrees well with observational estimates. It reaches a maximum value of almost 0.12 Sv in late winter when the ice thickness is at a maximum and the winds are at their strongest. As the ice volume flux depends on both the thickness and velocity of the sea ice, the good agreement with observations suggests that both of these properties are reasonably simulated. This does appear to be the case, as discussed further in Holland et al. (2006) and DeWeaver and Bitz (2006). In the long-term average, the Northern Hemisphere CCSM3 sea ice covers 10 × 106 km2 with a mean thickness of approximately 2 m. Accounting for the ice salinity, this represents a freshwater storage of 18 450 km3.

5. Summary

We have presented selected features of the simulated hydrological cycle for CAM3 for both coupled and uncoupled configurations of the model. CAM3 exhibits a weaker hydrological cycle when compared with predecessor models, which is closer in magnitude to current observational estimates. The relative distribution of surface water exchange and atmospheric water storage by surface type is in good agreement with observational estimates. Major observable features of the hydrological cycle, such as precipitation rate, and vertical integrals of atmospheric water storage, are generally well captured in zonal-mean seasonal and annual averages. For other zonally averaged measures, like liquid water loading at high latitudes, there are significant differences with the evolving observational estimates.

To some extent, zonal averages and vertical integrals mask some important long-standing biases in the simulation. Although the simulated thermal structure of the CAM3 atmosphere is improved, biases in the vertical distribution of water vapor remain. The most serious deficiency is a lower-tropospheric dry bias equatorward of 40°latitude, and an upper-tropospheric moist bias over the same latitude band. Even though the zonally averaged vertical integrals of water vapor compare well with observations, there are significant differences with respect to longitudinal structure. As with earlier models, the Atlantic portion of the domain is drier than observed, while the Pacific sector is moister than observed. Similarly, the horizontal distribution of precipitation continues to exhibit a very weak Atlantic ITCZ, and double ITCZ-like structures in the Pacific and Indian Oceans. An overactive precipitation regime in the western Indian Ocean, and enhanced extratropical storm-track precipitation, also continue to be undesirable features of the time-mean simulation.

CAM3 does a very good job at representing important low-frequency variability measures of the hydrological cycle. Despite local biases in the mean state, the detailed structure of the seasonal cycle of precipitable water is very well reproduced. Similarly, CAM3's response to ENSO is also well captured, including features of the far-field response. An important exception is the representation of precipitation and precipitation variability over Amazonia. CAM3 also continues to weakly simulate equatorial precipitation variability over the Maritime Continent and westward into the Indian Ocean. Although reliable measurements do not yet exist, CAM3 shows only a weak seasonal redistribution of cloud liquid water at high latitudes and a very strong seasonal cycle of cloud ice water.

Many of the CAM3 simulation features using specified SST and sea ice distributions carry over to the fully coupled configuration of CCSM3. However, two of the most important weaknesses of the CAM3 simulation, the tendency to form double-ITCZ structures in the deep Tropics and to inadequately simulate the seasonal meridional migration of tropical precipitation, are significantly amplified in the coupled configuration. This simulation challenge affects most atmospheric general circulation models to some degree, and remains one of the most serious simulation biases facing coupled climate modeling. As discussed in section 4b, this simulation bias is significant with regard to the ocean freshwater budget in the Tropics, playing a major role in the erroneous meridional transport of freshwater by the ocean circulation. Biases in precipitation over the continents, such as the positive bias over the Congo and the negative bias over the Amazon, also present challenges to ocean freshwater transport away from regions of river discharge. CCSM3 freshwater transport by sea ice processes has mixed results. Meridional ice transport in the Southern Hemisphere is approximately 2–3 times larger than observational estimates, largely due to an overly extensive sea ice distribution. The Northern Hemisphere simulation is substantially better, where ice transport through the Fram Strait very closely tracks observational estimates.

Overall CAM3 does a very good job of representing the major features of the global hydrological cycle as it is presently known. There are, however, some important deficiencies in the simulation that require a better physical understanding so that they can be addressed through better parameterized treatments of hydrological processes, such as the partitioning of evapotranspiration over land. Given its importance to the coupled simulation, identifying the reasons for the atmosphere's double-ITCZ bias is likely to be a very high priority simulation challenge that will need to be addressed in the next generation of the CCSM–CAM.

Acknowledgments

The authors acknowledge the substantial contributions to the CCSM project from the National Science Foundation (NSF), Department of Energy (DOE), the National Oceanic and Atmospheric Administration (NOAA), and the National Aeronautics and Space Administration (NASA). In particular, Hack, Truesdale, and Caron acknowledge support from the DOE Office of Science, DOE Climate Change Prediction Program, and the NCAR Water Cycle Across Scales Initiative. Support for Holland is acknowledged from NSF Grant OPP-0242290.

This study is based on numerical integrations performed by NCAR and the Central Research Institute of the Electric Power Industry (CRIEPI) with support and facilities provided by NSF, DOE, MEXT, and ESC/JAMSTEC.

REFERENCES

  • Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor, 4 , 11471167.

    • Search Google Scholar
    • Export Citation
  • Boer, G. J., and Coauthors, 1992: Some results from an intercomparison of the climates simulated by 14 atmospheric general circulation models. J. Geophys. Res, 97 , 1277112786.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., and S. Levis, 2006: Evaluating aspects of the Community Land and Atmosphere Models (CLM3 and CAM) using a dynamic global vegetation model. J. Climate, 19 , 22902301.

    • Search Google Scholar
    • Export Citation
  • Boville, B. A., P. J. Rasch, J. J. Hack, and J. R. McCaa, 2006: Representation of clouds and precipitation processes in the Community Atmosphere Model version 3 (CAM3). J. Climate, 19 , 21842198.

    • Search Google Scholar
    • Export Citation
  • Brutsaert, W., 1984: Evaporation into the Atmosphere. Kluwer Academic Publishers, 299 pp.

  • Cess, R. D., and Coauthors, 1990: Inter-comparison and interpretation of climate feedback processes in 19 atmospheric general circulation models. J. Geophys. Res, 95 , 1660116615.

    • Search Google Scholar
    • Export Citation
  • Chahine, M. T., 1992: The hydrologic cycle and its influence on climate. Nature, 359 , 373380.

  • Choudhury, B. J., N. E. DiGirolamo, J. Susskind, W. L. Darnell, S. K. Gupta, and G. Asrar, 1998: A biophysical process-based estimate of global land surface evaporation using satellite and ancillary data II. Regional and global patterns of seasonal and annual variations. J. Hydrol, 205 , 186204.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., 2001: Parameterization of generalized cloud overlap for radiative calculations in general circulation models. J. Atmos. Sci, 58 , 32243242.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., P. J. Rasch, B. E. Eaton, B. Khattatov, J-F. Lamarque, and C. S. Zender, 2001: Simulating aerosols using a chemical transport model with assimilation of satellite aerosol retrievals: Methodology for INDOEX. J. Geophys. Res, 106 , 73137336.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., J. K. Hackney, and D. P. Edwards, 2002a: A new parameterization for infrared emission and absorption by water vapor in the National Center for Atmospheric Research Community Atmosphere Model. J. Geophys. Res, 107 .8028, doi:10.1029/2001JD001365.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., P. J. Rasch, B. E. Eaton, D. W. Fillmore, J. T. Kiehl, T. C. Beck, and C. S. Zender, 2002b: Simulation of aerosol distributions and radiative forcing for INDOEX: Regional climate impacts. J. Geophys. Res, 107 .4664, doi:10.1029/2001JD001365.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., and Coauthors, 2006a: The Community Climate System Model version 3 (CCSM3). J. Climate, 19 , 21222143.

  • Collins, W. D., and Coauthors, 2006b: The formulation and atmospheric simulation of the Community Atmosphere Model version 3 (CAM3). J. Climate, 19 , 21442161.

    • Search Google Scholar
    • Export Citation
  • Dai, A., and K. E. Trenberth, 2002: Estimates of freshwater discharge from continents: Latitudinal and seasonal variations. J. Hydrometeor, 3 , 660687.

    • Search Google Scholar
    • Export Citation
  • da Silva, A., A. C. Young, and S. Levitus, 1994: Algorithims and Procedures. Vol. 1, Atlas of Surface Marine Data 1994, NOAA, 83 pp.

  • Davey, M. K., and Coauthors, 2002: Stoic: A study of coupled model climatology and variability in tropical ocean regions. Climate Dyn, 18 , 403420.

    • Search Google Scholar
    • Export Citation
  • DeWeaver, E., and C. M. Bitz, 2006: Atmospheric circulation and its effect on Arctic sea ice in CCSM3 at medium and high resolution. J. Climate, 19 , 24152436.

    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., K. W. Oleson, G. B. Bonan, F. M. Hoffman, P. E. Thornton, M. Vertenstein, Z-L. Yang, and X. B. Zeng, 2006: The Community Land Model and its climate statistics as a component of the Community Climate System Model. J. Climate, 19 , 23022324.

    • Search Google Scholar
    • Export Citation
  • Döll, P., and S. Siebert, 2002: Global modeling of irrigation water requirements. Water Resour. Res, 38 , 8.18.10.

  • Doney, S. C., W. G. Large, and F. O. Bryan, 1998: Surface ocean fluxes and water-mass transformation rates in the coupled NCAR Climate System Model. J. Climate, 11 , 14201441.

    • Search Google Scholar
    • Export Citation
  • Fekete, B. M., C. J. Vorosmarty, and W. Grabs, 2002: High-resolution fields of global runoff combining observed river discharge and simulated water balances. Global Biogeochem. Cycles, 16 .1042, doi:10.1029/1999GB001254.

    • Search Google Scholar
    • Export Citation
  • Goosse, H., and T. Fichefet, 1999: Importance of ice-ocean interactions for the global ocean circulation: A model study. J. Geophys. Res, 104 , 2333723355.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., B. A. Boville, J. T. Kiehl, and P. J. Rasch, 1994: Climate statistics from the National Center for Atmospheric Research Community Climate Model (CCM2). J. Geophys. Res, 99 , 2078520813.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., J. T. Kiehl, and J. W. Hurrell, 1998: The hydrologic and thermodynamic characteristics of the NCAR CCM3. J. Climate, 11 , 11791206.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., J. M. Caron, G. Danabasoglu, K. W. Oleson, C. Bitz, and J. E. Truesdale, 2006: CCSM–CAM3 climate simulation sensitivity to changes in horizontal resolution. J. Climate, 19 , 22672289.

    • Search Google Scholar
    • Export Citation
  • Holland, M. M., C. M. Bitz, M. Eby, and A. J. Weaver, 2001: The role of ice–ocean interactions in the variability of the North Atlantic thermohaline circulation. J. Climate, 14 , 656675.

    • Search Google Scholar
    • Export Citation
  • Holland, M. M., C. M. Bitz, E. C. Hunke, W. H. Lipscomb, and J. L. Schramm, 2006: Influence of the sea ice thickness distribution on polar climate in CCSM3. J. Climate, 19 , 23982414.

    • Search Google Scholar
    • Export Citation
  • Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, and D. Xiaosu, Eds. 2001: Climate Change 2001: The Scientific Basis. Cambridge University Press, 944 pp.

    • Search Google Scholar
    • Export Citation
  • Kiehl, J. T., 1994: Clouds and their effects on the climate system. Phys. Today, 47 , 3642.

  • Kiehl, J. T., and K. E. Trenberth, 1997: Earth's annual global mean energy budget. Bull. Amer. Meteor. Soc, 78 , 197208.

  • Kiehl, J. T., and P. R. Gent, 2004: The Community Climate System Model, version 2. J. Climate, 17 , 36663682.

  • Kiehl, J. T., J. J. Hack, G. B. Bonan, B. B. Boville, D. L. Williamson, and P. J. Rasch, 1998: The National Center for Atmospheric Research Community Climate Model: CCM3. J. Climate, 11 , 11311149.

    • Search Google Scholar
    • Export Citation
  • King, M. D., and Coauthors, 2003: Cloud and aerosol properties, precipitable water, and profiles of temperature and water vapor from MODIS. IEEE Trans. Geosci. Remote Sens, 41 , 442458.

    • Search Google Scholar
    • Export Citation
  • Klein, S., and C. Jakob, 1999: Validation and sensitivities of frontal clouds simulated by the ECMWF model. Mon. Wea. Rev, 127 , 25142531.

    • Search Google Scholar
    • Export Citation
  • Kwok, R., G. F. Cunningham, and S. S. Pang, 2004: Fram Strait sea ice outflow. J. Geophys. Res, 109 .C01009, doi:10.1029/2003JC001785.

  • Large, W. G., and S. G. Yeager, 2004: Diurnal to decadal global forcing for ocean and sea-ice models: The data sets and flux climatologies. Tech. Rep. NCAR/TN-460+STR, National Center for Atmospheric Research, Boulder, CO.

  • Large, W. G., and G. Danabasoglu, 2006: Attribution and impacts of upper-ocean biases in CCSM3. J. Climate, 19 , 23252346.

  • Lejenäs, H., R. A. Madden, and J. J. Hack, 1997: Global atmospheric angular momentum and Earth-atmosphere exchange of angular momentum simulated in a general circulation model. J. Geophys. Res, 102 , 19311941.

    • Search Google Scholar
    • Export Citation
  • Menon, S., J. Hansen, L. Nazarenko, and Y. Luo, 2002: Climate effects of black carbon aerosols in China and India. Science, 297 , 22502253.

    • Search Google Scholar
    • Export Citation
  • Oberhuber, J. M., 1988: An atlas based upon the “COADS” data set: The budgets of heat, buoyancy, and turbulent kinetic energy at the surface of the global ocean. Tech. Rep. 15, Max-Planck-Institut für Meteorologie, 199 pp.

  • Oleson, K. W., and Coauthors, 2004: Technical description of the Community Land Model (CLM). Tech. Rep. NCAR/TN-461+STR, National Center for Atmospheric Research, Boulder, CO, 174 pp.

  • Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • Ramanathan, V., P. J. Crutzen, J. T. Kiehl, and D. Rosenfeld, 2001: Aerosols, climate, and the hydrological cycle. Science, 294 , 21192124.

    • Search Google Scholar
    • Export Citation
  • Randel, D. L., T. H. V. Haar, M. A. Ringerud, G. Stephens, T. J. Greenwald, and C. L. Combs, 1996: A new global water vapor dataset. Bull. Amer. Meteor. Soc, 77 , 12331246.

    • Search Google Scholar
    • Export Citation
  • Rasch, P. J., W. D. Collins, and B. E. Eaton, 2001: Understanding the Indian Ocean Experiment (INDOEX) aerosol distributions with an aerosol assimilation. J. Geophys. Res, 106 , 73377356.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and E. N. Dueñas, 2004: The International Satellite Cloud Climatology Project (ISCCP) Web Site: An online resource for research. Bull. Amer. Meteor. Soc, 85 , 167172.

    • Search Google Scholar
    • Export Citation
  • Schlesinger, W. H., 1991: Biogeochemistry: An Analysis of Global Change. Academic Press, 433 pp.

  • Stephens, G. L., and P. J. Webster, 1981: Clouds and climate: Sensitivity of simple systems. J. Atmos. Sci, 38 , 235247.

  • Stowe, L. L., C. G. Wellemeyer, T. F. Eck, and H. Y. M. Yeh, and The Nimbus-7 Cloud Data Processing Team, 1988: Nimbus-7 global cloud climatology. Part I: Algorithms and validation. J. Climate, 1 , 445470.

    • Search Google Scholar
    • Export Citation
  • Stowe, L. L., H. Y. M. Yeh, T. F. Eck, C. G. Wellemeyer, and H. L. Kyle, and The Nimbus-7 Cloud Data Processing Team, 1989: Nimbus-7 global cloud climatology. Part II: First year results. J. Climate, 2 , 671709.

    • Search Google Scholar
    • Export Citation
  • Timmermann, R., A. Worby, H. Goosse, and T. Fichefet, 2004: Utilizing the ASPeCt sea ice thickness data set to evaluate a global coupled sea ice-ocean model. J. Geophys. Res, 109 .C07017, doi:10.1029/2003JC002242.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., and C. J. Guillemot, 1995: Evaluation of the global atmospheric moisture budget as seen from analyses. J. Climate, 8 , 22552272.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., D. P. Stepaniak, and J. M. Caron, 2000: The global monsoon as seen through the divergent atmospheric circulation. J. Climate, 13 , 39693993.

    • Search Google Scholar
    • Export Citation
  • Vaughan, D. G., J. L. Bamber, M. Giovinetto, J. Russell, and A. P. R. Cooper, 1999: Reassessment of net surface mass balance in Antarctica. J. Climate, 12 , 933946.

    • Search Google Scholar
    • Export Citation
  • Vinje, T., 2001: Fram Strait ice fluxes and atmospheric circulation: 1950–2000. J. Climate, 14 , 35083517.

  • Warren, S. G., C. J. Hahn, J. London, R. L. Chervin, M. Robert, and R. Jenne, 1988: Global distribution of total cloud cover and cloud type amounts over the ocean. Tech. Rep. NCAR Tech. Note TN-317+STR, National Center for Atmospheric Research, Boulder, CO. 42 pp. + 170 maps.

  • Weatherly, J. W., B. P. Briegleb, and W. G. Large, 1998: Sea ice and polar climate in the NCAR CSM. J. Climate, 11 , 14721486.

  • Webb, M., C. Senior, S. Bony, and J. J. Morcrette, 2001: Combining ERBE and ISCCP data to assess clouds in the Hadley Centre, ECMWF, and LMD atmospheric climate models. Climate Dyn, 17 , 905922.

    • Search Google Scholar
    • Export Citation
  • Wentz, F. J., and R. W. Spencer, 1998: SSM/I rain retrievals within a unified all-weather ocean algorithm. J. Atmos. Sci, 55 , 16131627.

    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., R. D. Cess, M. D. King, D. A. Randall, and E. F. Harrison, 1995: Mission to planet Earth: Role of clouds and radiation in climate. Bull. Amer. Meteor. Soc, 76 , 21252153.

    • Search Google Scholar
    • Export Citation
  • Williamson, D. L., 1988: The effect of vertical finite difference approximations on simulations with the NCAR Community Climate Model. J. Climate, 1 , 4058.

    • Search Google Scholar
    • Export Citation
  • Willmott, C. J., and K. Matsuura, cited. 2000: Terrestrial air temperature and precipitation: Monthly and annual climatologies. [Available online at http://climate.geog.udel.edu/~climate.].

  • Xie, P., and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc, 78 , 25392558.

    • Search Google Scholar
    • Export Citation
  • Xie, S., and M. Zhang, 2000: Impact of the convection triggering function on single-column model simulations. J. Geophys. Res, 105 , 1498314996.

    • Search Google Scholar
    • Export Citation
  • Yeager, S. G., C. Shields, W. G. Large, and J. J. Hack, 2006: The low-resolution CCSM3. J. Climate, 19 , 25452566.

  • Zhang, G. J., 2003: Convective quasi-equilibrium in the tropical western Pacific: Comparison with midlatitude continental environment. J. Geophys. Res, 108 .4592, doi:10.1029/2003JD003520.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

The simulated global water budget. Storage terms are in mm and exchange rates are in mm day−1. Quantities in [] are derived from MODIS, while quantities in () are a renormalized version of the global water cycle described in Peixoto and Oort (1992).

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 2.
Fig. 2.

Zonally averaged annual, DJF, and JJA precipitation rate in mm day−1 for CAM3, CCSM3, and CMAP.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 3.
Fig. 3.

Same as in Fig. 2, but for net surface water flux (EP) in W m−2 for CAM3 and CCSM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 4.
Fig. 4.

Same as in Fig. 2, but for precipitable water in kg m−2 for CAM3, CCSM3, NVAP, and CCM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 5.
Fig. 5.

Zonally averaged total cloud fraction for CAM3 compared with ISCCP, Nimbus-7, and MODIS.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 6.
Fig. 6.

Zonally averaged, ocean-only cloud liquid water path (LWP) in g m−2 for CAM3, CCSM3, MODIS, SSMI, and NVAP.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 7.
Fig. 7.

Same as in Fig. 2, but for cloud ice water path (IWP) in g m−2 for CAM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 8.
Fig. 8.

Zonally averaged annual mean temperature, specific humidity, and relative humidity differences between CAM3 and ERA-40.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 9.
Fig. 9.

Profiles of equivalent potential temperature and specific humidity for CAM3, ERA-40, and RAOBS at Yap Island (9.4°N, 138.1°E) for a July climatology.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 10.
Fig. 10.

CAM3 zonally averaged annual mean cloud liquid water concentration (color contours), ice water concentration (black contours), with 273-K freezing level contour (red) for reference.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 11.
Fig. 11.

Zonally averaged annual, DJF, and JJA (left) mean and (right) transient meridional moisture transport in g m kg−1 s−1 for CAM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 12.
Fig. 12.

Annual mean precipitation in mm day−1 for (top) CAM3, (middle) CMAP, and (bottom) their difference.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 13.
Fig. 13.

Annual, DJF, and JJA net surface water flux (EP) in W m−2 for CAM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 14.
Fig. 14.

Annual mean precipitable water in kg m−2 for (top) CAM3, (middle) NVAP, and (bottom) their difference.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 15.
Fig. 15.

Annual mean cloud IWP and cloud LWP in g m−2 for CAM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 16.
Fig. 16.

Equatorial (10°S–10°N) precipitation anomalies for CAM3, CMAP, and CCSM3. The CAM3 and CMAP period includes 1979–2000, and the CCSM3 is an arbitrary representative 22-yr period from the control simulation.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 17.
Fig. 17.

ENSO warm (July 1994) minus ENSO cold (June 1999) event precipitation anomalies for CAM3 and GPCP.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 18.
Fig. 18.

Amplitude of seasonal precipitable water changes in mm for CAM3 and NVAP.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 19.
Fig. 19.

Annual mean precipitation in mm day−1 for (top) CCSM3, (middle) CAM3, and (bottom) their difference.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 20.
Fig. 20.

Net freshwater flux for the CCSM3, uncoupled ocean model, and observations.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 21.
Fig. 21.

Global (solid lines) and Atlantic (dotted lines) mean northward ocean freshwater transport from the CCSM3 (thick) and the uncoupled model (thin).

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 22.
Fig. 22.

(top) Zonally averaged annual mean precipitation and evaporation for CCSM3 and observations and (bottom) zonally averaged annual mean land hydrology values from CCSM3.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Fig. 23.
Fig. 23.

The simulated annual cycle of Fram Strait sea ice transport in Sv. The solid line is CCSM3. The monthly average values over a 9-yr time series from observational estimates (Kwok et al. 2004) are shown as diamonds with the standard deviation from this time series indicated by the error bars. Note that the observational estimates for June–September are identical on this figure because Kwok et al. (2004) only provide a single average for these months.

Citation: Journal of Climate 19, 11; 10.1175/JCLI3755.1

Table 1.

Annual average precipitation and evaporation rates by surface type (mm day−1).

Table 1.
Table 2.

Annual average storage of vapor, cloud water, and cloud ice by surface type (mm).

Table 2.
Table 3.

Annual averages of global land precipitation (P), evapotranspiration (E), and runoff (R) (mm day−1). The components of E are transpiration (ET), evaporation of canopy intercepted water (Ec), and ground evaporation (Eg). The components of runoff are surface runoff (RS), drainage from the soil column (RD), and runoff from glaciers, wetlands, and lakes and snow-capped surfaces (RGWL) (mm day−1). Photosynthesis (PS) has units of PgC. Observations for P are from Willmott and Matsuura (2000), R from Fekete et al. (2002), and PS from Schlesinger (1991).

Table 3.
Save