• Arakawa, A., 2004: The cumulus parameterization problem: Past, present, and future. J. Climate, 17, 24932525, doi:10.1175/1520-0442(2004)017<2493:RATCPP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Arakawa, A., , and W. H. Schubert, 1974: Interaction of a cumulus cloud ensemble with the large-scale environment, part I. J. Atmos. Sci., 31, 674701, doi:10.1175/1520-0469(1974)031<0674:IOACCE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Arakawa, A., , and C.-M. Wu, 2013: A unified representation of deep moist convection in numerical modeling of the atmosphere. Part I. J. Atmos. Sci., 70, 19771992, doi:10.1175/JAS-D-12-0330.1.

    • Search Google Scholar
    • Export Citation
  • Arakawa, A., , J.-H. Jung, , and C.-M. Wu, 2011: Toward unification of the multiscale modeling of the atmosphere. Atmos. Chem. Phys., 11, 37313742, doi:10.5194/acp-11-3731-2011.

    • Search Google Scholar
    • Export Citation
  • Bretherton, C. S., , and S. Park, 2009: A new moist turbulence parameterization in the Community Atmosphere Model. J. Climate, 22, 34223448, doi:10.1175/2008JCLI2556.1.

    • Search Google Scholar
    • Export Citation
  • Chen, Y.-L., 1985: Diagnosis of the net cloud mass flux in GATE. J. Atmos. Sci., 42, 17571769, doi:10.1175/1520-0469(1985)042<1757:DOTNCM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Dai, A., 2006: Precipitation characteristics in eighteen coupled climate models. J. Climate, 19, 46054630, doi:10.1175/JCLI3884.1.

  • Del Genio, A. D., , J. Wu, , and Y. Chen, 2012: Characteristics of mesoscale organization in WRF simulations of convection during TWP-ICE. J. Climate, 25, 56665688, doi:10.1175/JCLI-D-11-00422.1.

    • Search Google Scholar
    • Export Citation
  • Donner, L. J., 1993: A cumulus parameterization including mass fluxes, vertical momentum dynamics, and mesoscale effects. J. Atmos. Sci., 50, 889906, doi:10.1175/1520-0469(1993)050<0889:ACPIMF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K., 1991: A scheme for representing cumulus convection in large-scale models. J. Atmos. Sci., 48, 23132329, doi:10.1175/1520-0469(1991)048<2313:ASFRCC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K., , and R. Pierrehumbert, 1996: Microphysical and dynamical control of tropospheric water vapor. Clouds, Chemistry and Climate, P. Crutzen and V. Ramanathan, Eds., NATO ASI Series, Vol. 35, Springer, 17–28, doi:10.1007/978-3-642-61051-6_2.

  • Fu, Q., , S. K. Krueger, , and K. N. Liou, 1995: Interactions of radiation and convection in simulated tropical cloud clusters. J. Atmos. Sci., 52, 13101328, doi:10.1175/1520-0469(1995)052<1310:IORACI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gamache, J. F., , and R. A. Houze, 1983: Water budget of a mesoscale convective system in the tropics. J. Atmos. Sci., 40, 18351850, doi:10.1175/1520-0469(1983)040<1835:WBOAMC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., , C.-P. Cheng, , C. A. Leary, , and J. F. Gamache, 1980: Diagnosis of cloud mass and heat fluxes from radar and synoptic data. J. Atmos. Sci., 37, 754773, doi:10.1175/1520-0469(1980)037<0754:DOCMAH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Iacono, M. J., , J. S. Delamere, , E. J. Mlawer, , M. W. Shephard, , S. A. Clough, , and W. D. Collins, 2008: Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res., 113, D13103, doi:10.1029/2008JD009944.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., 1980: Diagnosis of convective and mesoscale motions during phase III of GATE. J. Atmos. Sci., 37, 733753, doi:10.1175/1520-0469(1980)037<0733:DOCAMM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., , T. M. Rickenbach, , S. A. Rutledge, , P. E. Ciesielski, , and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12, 23972418, doi:10.1175/1520-0442(1999)012<2397:TCOTC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kuo, H. L., 1974: Further studies of the parameterization of the influence of cumulus convection on large-scale flow. J. Atmos. Sci., 31, 12321240, doi:10.1175/1520-0469(1974)031<1232:FSOTPO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Leary, C. A., , and R. A. Houze, 1980: The contribution of mesoscale motions to the mass and heat fluxes of an intense tropical convective system. J. Atmos. Sci., 37, 784796, doi:10.1175/1520-0469(1980)037<0784:TCOMMT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lindzen, R. S., , M.-D. Chou, , and A. Y. Hou, 2001: Does the Earth have an adaptive infrared iris? Bull. Amer. Meteor. Soc., 82, 417432, doi:10.1175/1520-0477(2001)082<0417:DTEHAA>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liu, Y.-C., , J. Fan, , G. J. Zhang, , K.-M. Xu, , and S. Ghan, 2015: Improving representation of convective transport for scale-aware parameterization: 2. Analysis of cloud-resolving model simulations. J. Geophys. Res. Atmos., 120, 35103532, doi:10.1002/2014JD022145.

    • Search Google Scholar
    • Export Citation
  • Loeb, N. G., , B. A. Wielicki, , D. R. Doelling, , G. L. Smith, , D. F. Keyes, , S. Kato, , N. Manalo-Smith, , and T. Wong, 2009: Toward optimal closure of the Earth’s top-of-atmosphere radiation budget. J. Climate, 22, 748766, doi:10.1175/2008JCLI2637.1.

    • Search Google Scholar
    • Export Citation
  • Lohmann, U., 2008: Global anthropogenic aerosol effects on convective clouds in ECHAM5-HAM. Atmos. Chem. Phys., 8, 21152131, doi:10.5194/acp-8-2115-2008.

    • Search Google Scholar
    • Export Citation
  • Lord, S. J., 1982: Interaction of a cumulus cloud ensemble with the large-scale environment. Part III: Semi-prognostic test of the Arakawa-Schubert cumulus parameterization. J. Atmos. Sci., 39, 88103, doi:10.1175/1520-0469(1982)039<0088:IOACCE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Luo, Y., , and S. K. Krueger, 2004: Evaluation of detrainment and microphysics parameterization in the NCEP GFS single-column model. Proc. 14th ARM Science Team Meeting, Albuquerque, NM, Atmospheric Radiation Measurement Program. [Available online at https://www.arm.gov/publications/proceedings/conf14/extended_abs/luo2-y.pdf.]

  • Mace, G. G., , and Q. Zhang, 2014: The CloudSat radar-lidar geometrical profile product (RL-GeoProf): Updates, improvements, and selected results. J. Geophys. Res. Atmos., 119, 94419462, doi:10.1002/2013JD021374.

    • Search Google Scholar
    • Export Citation
  • Mace, G. G., , M. Deng, , B. Soden, , and E. Zipser, 2006: Association of tropical cirrus in the 10–15-km layer with deep convective sources: An observational study combining millimeter radar data and satellite-derived trajectories. J. Atmos. Sci., 63, 480503, doi:10.1175/JAS3627.1.

    • Search Google Scholar
    • Export Citation
  • Morrison, H., , and A. Gettelman, 2008: A new two-moment bulk stratiform cloud microphysics scheme in the Community Atmosphere Model, version 3 (CAM3). Part I: Description and numerical tests. J. Climate, 21, 36423659, doi:10.1175/2008JCLI2105.1.

    • Search Google Scholar
    • Export Citation
  • Neale, R. B., , J. H. Richter, , and M. Jochum, 2008: The impact of convection on ENSO: From a delayed oscillator to a series of events. J. Climate, 21, 59045924, doi:10.1175/2008JCLI2244.1.

    • Search Google Scholar
    • Export Citation
  • Park, S., 2014: A unified convection scheme (UNICON). Part I: Formulation. J. Atmos. Sci., 71, 39023930, doi:10.1175/JAS-D-13-0233.1.

    • Search Google Scholar
    • Export Citation
  • Park, S., , and C. S. Bretherton, 2009: The University of Washington shallow convection and moist turbulence schemes and their impact on climate simulations with the Community Atmosphere Model. J. Climate, 22, 34493469, doi:10.1175/2008JCLI2557.1.

    • Search Google Scholar
    • Export Citation
  • Ramanathan, V., , and W. Collins, 1991: Thermodynamic regulation of ocean warming by cirrus clouds deduced from observations of the 1987 El Niño. Nature, 351, 2732, doi:10.1038/351027a0.

    • Search Google Scholar
    • Export Citation
  • Randall, D. A., 2013: Beyond deadlock. Geophys. Res. Lett., 40, 59705976, doi:10.1002/2013GL057998.

  • Randall, D. A., , Harshvardhan, , D. A. Dazlich, , and T. G. Corsetti, 1989: Interactions among radiation, convection, and large-scale dynamics in a general circulation model. J. Atmos. Sci., 46, 19431970, doi:10.1175/1520-0469(1989)046<1943:IARCAL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rennó, N. O., , K. A. Emanuel, , and P. H. Stone, 1994: Radiative-convective model with an explicit hydrologic cycle: 1. Formulation and sensitivity to model parameters. J. Geophys. Res., 99, 14 42914 441, doi:10.1029/94JD00020.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and R. A. Houze, 2003: Stratiform rain in the tropics as seen by the TRMM Precipitation Radar. J. Climate, 16, 17391756, doi:10.1175/1520-0442(2003)016<1739:SRITTA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Song, X., , and G. J. Zhang, 2011: Microphysics parameterization for convective clouds in a global climate model: Description and single-column model tests. J. Geophys. Res., 116, D02201, doi:10.1029/2010JD014833.

    • Search Google Scholar
    • Export Citation
  • Song, X., , G. J. Zhang, , and J.-L. F. Li, 2012: Evaluation of microphysics parameterization for convective clouds in the NCAR Community Atmosphere Model CAM5. J. Climate, 25, 85688590, doi:10.1175/JCLI-D-11-00563.1.

    • Search Google Scholar
    • Export Citation
  • Stephens, G. L., , S. van den Heever, , and L. Pakula, 2008: Radiative–convective feedbacks in idealized states of radiative–convective equilibrium. J. Atmos. Sci., 65, 38993916, doi:10.1175/2008JAS2524.1.

    • Search Google Scholar
    • Export Citation
  • Tiedtke, M., 1989: A comprehensive mass flux scheme for cumulus parameterization in large-scale models. Mon. Wea. Rev., 117, 17791800, doi:10.1175/1520-0493(1989)117<1779:ACMFSF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wall, C., , C. Liu, , and E. Zipser, 2013: A climatology of tropical congestus using CloudSat. J. Geophys. Res. Atmos., 118, 64786492, doi:10.1002/jgrd.50455.

    • Search Google Scholar
    • Export Citation
  • Wang, Z., , and K. Sassen, 2001: Cloud type and macrophysical property retrieval using multiple remote sensors. J. Appl. Meteor., 40, 16651682, doi:10.1175/1520-0450(2001)040<1665:CTAMPR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , S. Esbensen, , and J.-H. Chu, 1973: Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611627, doi:10.1175/1520-0469(1973)030<0611:DOBPOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yang, S., , and E. A. Smith, 2008: Convective–stratiform precipitation variability at seasonal scale from 8 yr of TRMM observations: Implications for multiple modes of diurnal variability. J. Climate, 21, 40874114, doi:10.1175/2008JCLI2096.1.

    • Search Google Scholar
    • Export Citation
  • Yasunaga, K., , H. Kida, , T. Satomura, , and N. Nishi, 2004: A numerical study on the detrainment of tracers by cumulus convection in TOGA COARE. J. Meteor. Soc. Japan, 82, 861878, doi:10.2151/jmsj.2004.861.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , W.-K. Tao, , S. W. Powell, , R. A. Houze, , P. Ciesielski, , N. Guy, , H. Pierce, , and T. Matsui, 2013: A comparison of the water budgets between clouds from AMMA and TWP-ICE. J. Atmos. Sci., 70, 487503, doi:10.1175/JAS-D-12-050.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, G., , and N. A. McFarlane, 1995: Sensitivity of climate simulations to the parameterization of cumulus convection in the Canadian Climate Centre general circulation model. Atmos.–Ocean, 33, 407446, doi:10.1080/07055900.1995.9649539.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., , U. Lohmann, , and P. Stier, 2005: A microphysical parameterization for convective clouds in the ECHAM5 climate model: Single-column model results evaluated at the Oklahoma Atmospheric Radiation Measurement Program site. J. Geophys. Res., 110, D15S07, doi:10.1029/2004JD005128.

    • Search Google Scholar
    • Export Citation
  • Zhang, Y., , and S. A. Klein, 2010: Mechanisms affecting the transition from shallow to deep convection over land: Inferences from observations of the diurnal cycle collected at the ARM Southern Great Plains site. J. Atmos. Sci., 67, 29432959, doi:10.1175/2010JAS3366.1.

    • Search Google Scholar
    • Export Citation
  • Zhao, M., 2014: An investigation of the connections among convection, clouds, and climate sensitivity in a global climate model. J. Climate, 27, 18451862, doi:10.1175/JCLI-D-13-00145.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a),(b),(c) Water (liquid + ice) detrained from the Zhang–McFarlane deep convective parameterization. The tendency is temporally averaged and vertically integrated. (d),(e),(f) Frequency of occurrence of ZM deep convection, averaged temporally. (top to bottom) The CTL and MPHY simulations followed by the difference.

  • View in gallery

    Detrainment from the ZM scheme, time averaged and integrated vertically for (a) ice (T −35°C), (b) mixed-phase (−35° T 0°C), and (c) warm (T 0°C) regions. Data are shown as a difference between CTL and MPHY simulations.

  • View in gallery

    Detrainment from the ZM scheme (g kg−1 day−1), averaged spatially and temporally for each region.

  • View in gallery

    Stratus cloud fraction averaged spatially and temporally for each region. Also plotted is the cloud fraction from the CloudSat 2B-GEOPROF-lidar, averaged over a 2-yr period (2009–10).

  • View in gallery

    Water budget terms for the four regions as described in section 2b for the CTL and MPHY simulations. Budget terms are integrated, averaged temporally and spatially, and normalized by the average total precipitation in each region.

  • View in gallery

    Spatially and temporally averaged profiles of sources (condensation/deposition) and sinks (autoconversion/accretion/collection) of cloud water for each region.

  • View in gallery

    Precipitation by cloud type, as a percentage of total precipitation in each region.

  • View in gallery

    Terms in the water budget for the oceanic regions, split by temperature. (top to bottom) The layers are ice (T −35°C), mixed phase (−35 T 0°C), and warm cloud (T 0°C). Given for each region/simulation are (left) ZM convection, (center) the large-scale stratus cloud, and (right) shallow convection. Blue numbers represent the net condensation/evaporation of cloud water (liquid + ice), green is the production of precipitation, and red is evaporation of precipitation. Numbers with black arrows represent the transport of cloud water vertically between cloud layers or horizontally between cloud types. Purple arrows (showing downward movement) represent the vertical flux of precipitation through cloud layers and at the surface.

  • View in gallery

    As in Fig. 8, but for the land regions.

  • View in gallery

    As in Fig. 8, but for SCAM simulations of the TWP-ICE case.

  • View in gallery

    (left) Shortwave and (right) longwave cloud forcing (W m−2) (a),(b) averaged over 10 yr of CERES EBAF 2.8 data; the difference between observations and the (c),(d) CTL and (e),(f) MPHY simulations; and (g),(h) the difference between the MPHY and CTL simulations.

  • View in gallery

    OLR (W m−2) (a) averaged over 10 yr of CERES EBAF 2.8 data; the difference between observations and the (b) CTL and (c) MPHY simulations; and (d) the difference between the MPHY and CTL simulations.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 63 63 11
PDF Downloads 46 46 8

Effects of Convective Microphysics Parameterization on Large-Scale Cloud Hydrological Cycle and Radiative Budget in Tropical and Midlatitude Convective Regions

View More View Less
  • 1 Scripps Institution of Oceanography, La Jolla, California
© Get Permissions
Full access

Abstract

A two-moment microphysics scheme for deep convection was previously implemented in the NCAR Community Atmosphere Model version 5 (CAM5) by Song et al. The new scheme improved hydrometeor profiles in deep convective clouds and increased deep convective detrainment, reducing the negative biases in low and midlevel cloud fraction and liquid water path compared to observations. Here, the authors examine in more detail the impacts of this improved microphysical representation on regional-scale water and radiation budgets. As a primary source of cloud water for stratiform clouds is detrainment from deep and shallow convection, the enhanced detrainment leads to larger stratiform cloud fractions, higher cloud water content, and more stratiform precipitation over the ocean, particularly in the subtropics where convective frequency is also increased. This leads to increased net cloud radiative forcing. Over land regions, cloud amounts are reduced as a result of lower relative humidity, leading to weaker cloud forcing and increased OLR. Comparing the water budgets to cloud-resolving model simulations shows improvement in the partitioning between convective and stratiform precipitation, though the deep convection is still too active in the GCM. The addition of convective microphysics leads to an overall improvement in the regional cloud water budgets.

Corresponding author address: Rachel L. Storer, Scripps Institution of Oceanography, 8622 Kennel Way, La Jolla, CA 92037. E-mail: rachel.l.storer@jpl.nasa.gov

Abstract

A two-moment microphysics scheme for deep convection was previously implemented in the NCAR Community Atmosphere Model version 5 (CAM5) by Song et al. The new scheme improved hydrometeor profiles in deep convective clouds and increased deep convective detrainment, reducing the negative biases in low and midlevel cloud fraction and liquid water path compared to observations. Here, the authors examine in more detail the impacts of this improved microphysical representation on regional-scale water and radiation budgets. As a primary source of cloud water for stratiform clouds is detrainment from deep and shallow convection, the enhanced detrainment leads to larger stratiform cloud fractions, higher cloud water content, and more stratiform precipitation over the ocean, particularly in the subtropics where convective frequency is also increased. This leads to increased net cloud radiative forcing. Over land regions, cloud amounts are reduced as a result of lower relative humidity, leading to weaker cloud forcing and increased OLR. Comparing the water budgets to cloud-resolving model simulations shows improvement in the partitioning between convective and stratiform precipitation, though the deep convection is still too active in the GCM. The addition of convective microphysics leads to an overall improvement in the regional cloud water budgets.

Corresponding author address: Rachel L. Storer, Scripps Institution of Oceanography, 8622 Kennel Way, La Jolla, CA 92037. E-mail: rachel.l.storer@jpl.nasa.gov

1. Introduction

Deep convection plays a fundamental role in the climate system. To correctly simulate its interaction with the environment in global climate models (GCMs), much of the past effort has been on representing convective effects on temperature and moisture fields because of their roles in atmospheric energetics and the hydrological cycle (Kuo 1974; Arakawa and Schubert 1974; Tiedtke 1989; Emanuel 1991; Donner 1993; Zhang and McFarlane 1995; and many more). More recently, attempts to develop scale-aware (or unified) convective parameterization schemes, again largely for temperature and moisture fields, have been explored to meet the need of increasing GCM resolutions (Arakawa 2004; Arakawa et al. 2011; Arakawa and Wu 2013; Randall 2013; Liu et al. 2015). On the other hand, not as much work has been done to link convection to stratiform clouds in large-scale models. Numerous observational studies have shown that deep convection is important to anvil cloud generation by detraining hydrometeors from convection and that these anvil clouds have a tremendous impact on Earth’s radiative energy budget climatologically (Randall et al. 1989; Ramanathan and Collins 1991; Mace et al. 2006; Del Genio et al. 2012). The radiative effect of the anvil clouds in turn can provide a negative feedback by stabilizing the troposphere and suppressing convective activity (Fu et al. 1995; Stephens et al. 2008).

The generation of anvil and cirrus clouds strongly depends on the microphysics of precipitation formation within convective updrafts (Emanuel and Pierrehumbert 1996). Based on satellite data analysis, Lindzen et al. (2001) propose an “iris hypothesis,” arguing that anvil clouds associated with convection can have a negative feedback on climate. They suggest that the area coverage of anvil clouds associated with tropical convection is less extensive when sea surface temperatures are higher, thus leading to increased outgoing longwave radiation (OLR) and a cooling effect. Central to this argument is the effect of convection on cirrus/anvil clouds. Rennó et al. (1994) tested several convection parameterization schemes in a radiative–convective equilibrium model and found that the equilibrium climate was very sensitive to precipitation efficiency in convection. They found that clouds with high precipitation efficiency produced a cold and dry climate and clouds with low precipitation efficiency produced a warm and moist climate. Zhao (2014) looked at the effects of changing convective precipitation efficiency (or detrainment efficiency) on climate simulations and noted a large sensitivity in cloud cover, particularly in low clouds. This had impacts on climate sensitivity as well. Therefore, a proper treatment of convective microphysical processes in global climate models is a crucial ingredient in producing reliable simulations of the present climate and future climate projection.

Deep convection also affects the atmospheric hydrological cycle. A large portion of global rainfall can be attributed to deep convection, particularly in the tropics and the midlatitude summers. Estimates from satellite data have determined that convective precipitation accounts for between 40% and 60% of the rainfall in the tropics (Schumacher and Houze 2003; Yang and Smith 2008); in addition, much of what falls as stratiform precipitation results from the anvil portion of mesoscale convective systems (Leary and Houze 1980). Climate models overestimate (underestimate) the contribution from convective (stratiform) rainfall to the total precipitation (Dai 2006). Dai (2006) suggests that this is because convection often triggers too early, producing light rainfall too often; however, it is likely that this error is partly due to not enough water being detrained from deep convection to form precipitating anvil clouds. Based on radar observations of a tropical mesoscale convective system, Gamache and Houze (1983) estimate that as much as 60% to 75% of the condensate in the stratiform portion of the convective system comes from detrainment of hydrometeors from the convective updrafts. Cloud-resolving model (CRM) simulations of tropical convective systems show a similar percentage (Zeng et al. 2013, hereafter Z13). Thus, correctly modeling convective/stratiform partitioning of precipitation will require understanding the budgets of cloud water and precipitation formation in both large-scale and convective clouds.

Recently, more attention has been paid to representing microphysical processes in convective parameterization (Zhang et al. 2005; Lohmann 2008; Song and Zhang 2011, hereafter SZ11). One such convective microphysics scheme (SZ11) was implemented in the Zhang–McFarlane (ZM) convective parameterization scheme (Zhang and McFarlane 1995) and tested in the NCAR Community Atmosphere Model version 5 (CAM5) by Song et al. (2012). They found that with the addition of the SZ11 microphysics, the deep convective clouds produced more liquid and ice, and the profiles of hydrometeors better matched observations. The increased condensate led to larger amounts of water detrained from deep convection to the large scale. Not only were the convective properties improved, but also, through the effects of the detrainment, the large-scale cloud properties showed reduced bias compared to observations, with increased low-cloud amounts particularly in the subtropics.

In this paper, we examine the changes to the large-scale hydrological cycle and radiative forcing of clouds that result from the implementation of the SZ11 convective microphysics in CAM5. Large-scale water budgets will be calculated over four regions in both the tropics and midlatitudes. Where possible, they will also be compared with previous cloud-resolving model simulations and observations. Section 2 will describe the methods used for the analysis. Section 3 will present the results. A summary and discussion will be provided in section 4.

2. Methods

a. Model description

The simulations examined here were carried out using CAM5. The basic physics parameterizations in CAM5 are as follows. The radiation is handled by the Rapid Radiative Transfer Model for GCMs (RRTMG; Iacono et al. 2008) and the moist turbulence scheme is based on Bretherton and Park (2009). A new cloud macrophysics scheme (Park 2014) is used to integrate several cloud processes. The macrophysics calculates cumulus cloud fractions based on the mass flux output from the deep and shallow convective schemes. The stratus fraction is calculated as a function of grid-mean relative humidity, taking into account some simple assumptions about cloud overlap. Either evaporation of cloud water or condensation of water vapor is then performed based on the grid-mean relative humidity to provide an in-stratus cloud water content for use with the large-scale microphysics. The microphysics for stratus clouds is a prognostic, two-moment scheme from Morrison and Gettelman (2008), which predicts mass mixing ratios and number concentrations of cloud droplets and ice crystals. The microphysics scheme includes several processes such as condensation/evaporation, freezing/melting, and autoconversion and collection of hydrometeors. Shallow convection in CAM5 is calculated after the deep convection scheme and uses a convective inhibition-based closure (Park and Bretherton 2009). Precipitation from the shallow convection scheme is simply any condensate that is over a threshold value in the updraft.

Deep convection in CAM5 is handled by the ZM parameterization (Zhang and McFarlane 1995), with modifications in the calculation of the convective available potential energy (CAPE) closure by Neale et al. (2008). In the standard version of the ZM scheme the conversion from cloud water to rainwater is handled by a single formula, from Lord (1982):
e1
where is the rain production rate, l is the cloud water content, is the updraft mass flux, and is a tunable parameter. The rainwater produced by the ZM scheme is available for evaporation in the downdraft and cloud environment before falling to the ground as precipitation. The remaining water in convective updrafts is then partitioned between liquid and ice using a simple linear function of temperature, and the mean cloud droplet and ice particle radius are prescribed. This condensate is then detrained from convection and made available for the large-scale stratus clouds.

The SZ11 microphysics scheme is a two-moment diagnostic parameterization that calculates mass mixing ratio and number concentration of four hydrometeor species (cloud liquid water, cloud ice, rain, and snow). Both cloud droplet activation and ice nucleation by aerosols are parameterized, thus allowing for aerosol indirect effects on deep convection to be modeled. The SZ11 scheme is based on the large-scale microphysics parameterization of Morrison and Gettelman (2008) but with some modifications to make it more suitable for deep convection. Microphysical processes parameterized include homogeneous and heterogeneous freezing of cloud water to form ice, autoconversion of cloud liquid water and ice, self-collection/aggregation of rain/snow, accretion between the various hydrometeors, the Bergeron–Findeisen process, and fallout of rain and snow. The microphysics acts within the updrafts of ZM deep convection, which are assumed to be saturated, and the calculated mixing ratio and number concentrations are applied to the detrained cloudy air. More details of the microphysics formulation can be found in SZ11 and Song et al. (2012).

The model simulations analyzed here are the same as those described by Song et al. (2012). Two simulations were conducted using CAM5; the first was a control simulation with the standard ZM configuration for CAM5 as described above (CTL), and the second was identical but with the addition of SZ11 convective microphysics (MPHY). The runs were forced with climatological sea surface temperatures and aerosol number concentrations that were diagnosed from the output of a CAM4 simulation with a chemistry model (Song et al. 2012). The simulations used the finite-volume dynamical core, with grid spacing of 1.9° latitude by 2.5° longitude and 30 vertical levels. The last 5 yr of the 6-yr simulations were analyzed.

b. Cloud water budgets

To examine the large-scale cloud water (liquid + ice) budget, the following equation is considered:
e2
where DPD represents the sum of liquid and ice detrained from the ZM deep convection scheme. This is the first of the cloud physics parameterizations that is run, and SHD is the detrainment from the shallow convection. The macrophysics (MAC) calculates the large-scale evaporation/condensation based on the grid-mean relative humidity and the cloud fractions. This results in a profile of in-stratus cloud condensate, which then is fed into the microphysics (MIC) for conversion into other forms of hydrometeors. The grid-scale advection of cloud water is small compared to the other terms and is neglected in the following analysis.
The microphysical source or sink of large-scale cloud water can be broken down further:
e3
where cond includes terms involving condensation and deposition, evap is evaporation and sublimation, and precprod is the precipitation production or the conversion from cloud water (liquid + ice) to precipitation (rain + snow). This term can be calculated as follows:
e4
The processes included in precipitation production by Morrison and Gettelman (2008) are the following: autoconversion of cloud liquid water (au); accretion of cloud liquid water by rain (accr); autoconversion of ice crystals into snow (aui); accretion of cloud ice by rain (acir); accretion of cloud water, cloud ice, and rain by snow (accs); and conversion of cloud water to snow through the Bergeron process (berg). The evaporation/sublimation of rain/snow directly influences the precipitation production and specific humidity but not cloud water content and therefore does not appear in Eq. (4).
The budget of cloud water l within the ZM deep convective parameterization is given by
e5
where is the updraft mass flux, is the detrainment of mass from the updraft, l is the convective cloud water content, is the net condensation, and is the conversion from cloud water to rainwater. In the CTL simulation, is calculated using Eq. (1), while in the MPHY run, the production of precipitation is a sum of processes similar to that in Eq. (4) and detailed by SZ11.

3. Results

a. Overall convective properties

This work is a direct follow on from that of Song et al. (2012), who described several important effects of implementing the SZ11 convective microphysics scheme into CAM5. Overall, the MPHY simulation showed increased liquid and ice water path in deep convective clouds, particularly over the tropical oceans, which is an improvement compared to observations (Song et al. 2012). The MPHY simulation also showed reduced error in global analyses of annual precipitation, cloud fraction, and cloud liquid water path. Song et al. (2012) also performed an analysis over a region in the western Pacific showing that convective mass flux increased when SZ11 microphysics were included, which in combination with the higher convective cloud water content led to an increase in the detrainment of ice and liquid water from deep convection. The authors proposed the existence of a positive feedback in this region, whereby the increased convective detrainment leads to stratus clouds that produce more precipitation and have more heating aloft as a result of more active microphysics freezing more condensate. The additional heating aloft can then enhance the atmospheric circulation and convective activity even further. It is this change in detrainment, as well as its effects on the large scale in different convective regimes, that we concentrate on here.

The increase in detrainment from the ZM scheme can be seen in Figs. 1a–c. Detrainment here is the sum of liquid and ice and is vertically integrated and time averaged. The control simulation with the default ZM parameterization (CTL) is shown, as well as the simulation including the SZ11 convective microphysics (MPHY) and the difference between the two model runs. The increase in detrainment in MPHY is consistent nearly globally, though with much spatial variation. The change is largest over the ocean, particularly in the subtropics. To explain some of the differences between ocean and land, we can examine similar plots of the frequency of occurrence of deep convection (Figs. 1d–f). In the CTL simulation, detrainment from oceanic deep convective clouds was quite small compared to over land, even in regions of similar frequency of occurrence of deep convection (Fig. 1d). This is likely due to the simplistic calculation of convective precipitation formation in the CTL run shown in Eq. (1). Over land the conversion factor is set to a smaller value (0.0059) than over the ocean (0.045) in order to represent the lower precipitation efficiency in the more polluted regions—this allows for more condensate available for detrainment over land. Including convective microphysics prevents this stark difference between land-based and oceanic convection and results in a detrainment map that more reasonably follows from the location of deep convective clouds. In Fig. 1f, it is apparent that the only significant changes in deep convective frequency occur in the subtropics—the increase in frequency of occurrence of convection here along with the more realistic representation of precipitation formation leads to the largest increases in detrainment in these regions. Concurrently, the additional moisture associated with the enhanced detrainment increases the relative humidity in this region, helping to precondition the atmosphere and leading to convection triggering more frequently.

Fig. 1.
Fig. 1.

(a),(b),(c) Water (liquid + ice) detrained from the Zhang–McFarlane deep convective parameterization. The tendency is temporally averaged and vertically integrated. (d),(e),(f) Frequency of occurrence of ZM deep convection, averaged temporally. (top to bottom) The CTL and MPHY simulations followed by the difference.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

The temperature at which detrainment is occurring is a worthwhile consideration, as it can have important radiative impacts. To ascertain at what levels the increased detrainment occurs in MPHY, we have integrated the detrainment separately over the ice (T −35°C), mixed-phase (−35° T 0°C), and warm cloud (T 0°C) layers of the atmosphere. Differences between the CTL and MPHY runs for each region are shown in Fig. 2. Changes in detrainment in the ice phase portion of clouds are small and largely confined to the ITCZ and the warm pool regions of the Pacific and Indian Oceans. As described by Song et al. (2012), the contribution to detrainment from ice is small compared to the change in convective ice water content because of the weaker convective mass flux in the upper regions of deep convective clouds. Therefore, the ice region does not significantly contribute to the total change in detrainment between CTL and MPHY. In the mixed phase, the magnitude of the detrainment increase in MPHY is widely variable, increasing particularly in the ITCZ and tropical warm pool, as well as the midlatitude storm-track regions. The detrained condensate in the mixed-phase region consists largely of liquid, as the convective liquid water content is higher than that of ice, and the increase from CTL to MPHY is largely in the form of liquid as well. In the warm cloud region, large areas of increased detrainment exist in the subtropical oceans. The increased detrainment in the MPHY simulation is largely due to changes within the warm and mixed-phase regions of deep convection. This follows with Song et al. (2012), who saw the greatest change in cloudiness in low- and midlevel clouds.

Fig. 2.
Fig. 2.

Detrainment from the ZM scheme, time averaged and integrated vertically for (a) ice (T −35°C), (b) mixed-phase (−35° T 0°C), and (c) warm (T 0°C) regions. Data are shown as a difference between CTL and MPHY simulations.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

b. Regional differences

We have chosen four regions to examine in more detail, with the goal of examining the regional differences of the effects of increased detrainment on the water budget in different convective regimes. There are two ocean regions: the western Pacific (10°–20°N, 130°–150°E), chosen as a representative convectively active region in the tropics, and a region in the subtropics (30°–10°S, 120°–100°W), chosen because of the large changes seen in this region from shallower convection in the ZM scheme. The two land regions examined are the U.S. southern Great Plains (SGP; 33°–41°N,103°–95°W), representing a midlatitude continental convective regime, and the Amazon (12°S–4°N, 72°–56°W), representing a tropical land convective regime. To isolate the influence of convection, only the most convectively active season was used for each land region [May–July (MJJ) for SGP and December–February (DJF) for the Amazon].

The average profile of ZM detrainment for both the CTL and MPHY run for each of the regions is shown in Fig. 3. In both simulations there are notable differences between the regions because of the environmental regimes and the resulting types of convection that develop. However, features in the vertical profiles of the MPHY run are in much better agreement with observations. In the western Pacific, convection follows a trimodal structure, with noticeable peaks corresponding to shallow, congestus, and deep convection. This is consistent with what is typically observed in tropical convection (Johnson et al. 1999). A similar structure can be noted in the Amazon, though the peak associated with congestus clouds is much more pronounced, and there is little shallow detrainment from the ZM scheme. A recent study of congestus clouds using satellite data (Wall et al. 2013) noted that these clouds were found most frequently over the Amazon. The congestus mode is noticeably lacking in the SGP detrainment profile. Previous observations over the SGP region during the warm season indicate two main cloudy regimes determined largely by lower-tropospheric moisture: one dominated by fair-weather shallow cumulus clouds and one dominated by deep convection (Zhang and Klein 2010). Radar observations over the same region show similar vertical structure (Wang and Sassen 2001). The subtropics region is dominated by shallow convective detrainment. In the CTL run, although tropical convection (both land-based and oceanic) shows relatively large congestus detrainment, shallow and deep convective detrainment is significantly weaker. The subtropical region shows very little shallow convection detrainment in the CTL run with a marked increase in MPHY. The SGP site is probably the closest between CTL and MPHY in terms of the vertical structure of detrainment.

Fig. 3.
Fig. 3.

Detrainment from the ZM scheme (g kg−1 day−1), averaged spatially and temporally for each region.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

The differences between the CTL and MPHY simulations in Fig. 3 are consistent with the global plots. Detrainment from the ZM parameterization increases in all regions with the addition of SZ11 microphysics, though the changes are most apparent in the ocean regions. In the CTL simulation, the stark distinction between convective precipitation production over land and ocean leads to oceanic convection producing too much precipitation, leaving not enough liquid for detrainment into the large scale. The addition of convective microphysics decreases the efficiency of convective precipitation formation over the ocean; however, over land the differences between CTL and MPHY are smaller as a result of the already low convective autoconversion rate. The largest differences exist in the subtropics, where the MPHY simulation now results in an improvement in low-cloud fraction (Song et al. 2012), which suggests that the CTL produced too little convective detrainment. In the four regions, the average detrainment of cloud water in MPHY has values ranging from 0.1 to 1 g kg−1 day−1. Previous attempts to diagnose detrainment using observations and cloud-resolving model simulations (Yanai et al. 1973; Houze et al. 1980; Johnson 1980; Chen 1985; Luo and Krueger 2004; Yasunaga et al. 2004) have found similar values, though it is difficult to find exact agreement because of the large number of assumptions that have to be made to calculate detrainment in both observations and models.

The large-scale cloud fraction resulting from the CAM5 physics can be compared with satellite observations. In Fig. 4, the regional stratus fraction for the four regions is plotted against cloud fraction from the CloudSat geometric profiling product (2B-GEOPROF)-lidar data (Mace and Zhang 2014; http://www.cloudsat.cira.colostate.edu), averaged over two years (2009–10, annually or seasonally where appropriate). For the most part, in all four regions the cloud fraction produced by CAM5 is too high. The shapes of the cloud fraction profiles match fairly well though, as described above, though deep convection in the Amazon is not deep enough in the model.

Fig. 4.
Fig. 4.

Stratus cloud fraction averaged spatially and temporally for each region. Also plotted is the cloud fraction from the CloudSat 2B-GEOPROF-lidar, averaged over a 2-yr period (2009–10).

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

Song et al. (2012) discussed the change in low-level clouds in the subtropics and attributed it to the increase in deep convective detrainment. This is confirmed in Fig. 4, where the large spike in stratus fraction near 800 mb in the MPHY simulation matches the enhanced detrainment from the ZM scheme in the warm cloud region seen in Fig. 3. This spike in stratus clouds, though too high in altitude, matches much more closely with satellite data. In the western Pacific, convective detrainment increases substantially throughout the whole column, leading to a more modest increase in stratus fraction at nearly all levels. The cloud fraction was already too high in the CTL simulation, however, and so this is not an improvement compared to observations. Over land regions there is a small increase in detrainment; however, the stratus fraction does not increase consistently. To understand these differences we look in more detail at the water budgets in the four regions.

Houze et al. (1980) made an early attempt to diagnose the terms in the water budgets of deep convection and the associated stratiform anvil clouds using observations from the Global Atmospheric Research Program Atlantic Tropical Experiment (GATE). The water budget includes condensation, evaporation, rainfall, and detrainment. They found that a large portion of the detrained condensate from deep convection must fall out as large-scale precipitation, as little evaporation of the condensate in the large scale was calculated. However, these calculations depended on simple assumed model parameters, and few studies exist that test these assumptions.

Here we examine similar large-scale budgets to Houze et al. (1980) using the terms in Eq. (2). The four terms are vertically integrated and averaged temporally and spatially for each region. The quantities are then normalized by the average total precipitation in each region so as to better compare the changes in precipitation efficiency between the two runs. These values are plotted in Fig. 5 as bar charts. The first term, detrainment from the ZM deep convective parameterization, increases in all regions, though as described above the increases are more substantial in the oceanic regions owing to the large underrepresentation of detrainment in these areas in the CTL simulation. Detrainment from the shallow convective scheme generally decreases over much of the globe, with just a few areas showing a slight increase (not shown). The differences seen in Fig. 5 are consistent with this—in the regions examined, only the Amazon shows a small increase in detrainment from shallow convection.

Fig. 5.
Fig. 5.

Water budget terms for the four regions as described in section 2b for the CTL and MPHY simulations. Budget terms are integrated, averaged temporally and spatially, and normalized by the average total precipitation in each region.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

Next is the macrophysics, which takes detrained water from deep and shallow convection, reconciles the various overlapping cloud fractions, and evaporates it in cloud-free regions or adds it to water in stratus clouds (depending on the grid-mean relative humidity) to provide a mean profile of cloud water to the large-scale microphysics parameterization. Macrophysics, on average, generally results in condensation above the freezing level and evaporation below, where the detrainment provides the largest amounts of excess liquid. When the macrophysics term is vertically integrated it results in a sink of cloud water over the ocean and a source over the land. Macrophysics decreases in magnitude for the MPHY run in all four simulations. Over the ocean regions, increases in convective detrainment lead to a more moist atmosphere and result in increased condensation and decreased evaporation from the macrophysics scheme. Thus, when vertically integrated the overall moisture sink is reduced in the MPHY simulation. Over land the relative humidity decreases despite an increase in detrainment (though more modest than the increase over ocean), leading to a reduction in condensation and a decrease in the source of cloud water from the macrophysics.

The last term in the budget represents the tendency in cloud water from the Morrison and Gettelman (2008) stratiform cloud microphysics scheme. This includes the formation of precipitation through interactions between various hydrometeors and also residual condensation and evaporation terms. Precipitation formation is the largest component of the tendency, and hence the microphysics scheme is generally a sink of cloud water globally. As seen in Fig. 5, the microphysics sink increases in magnitude in all regions, with the subtropics having the most dramatic change because of a large positive feedback between increased low-level detrainment, low-level cloudiness, and relative humidity.

To summarize the cloud water budgets in Fig. 5, the main source of water for large-scale clouds is detrainment and the main sink is the precipitation formation by the stratiform microphysics scheme, both of which increase in the MPHY simulation. This leads to a small total tendency as expected, indicating the budget calculation is accurate.

It is useful to examine separate microphysical processes of stratiform clouds to help understand the increased microphysical sink in the MPHY simulation. Plotted in Fig. 6 are the sums of the sources and sinks of cloud water for each region. The solid lines are the sum of those terms involved in transforming hydrometeors between water and vapor: condensation, evaporation, deposition, and sublimation. This term is positive through much of the column, as condensation and deposition within cloud largely outweigh evaporation and sublimation, and so overall this is a source of cloud water. The largest contribution to the source term comes from deposition in all four of the regions (not shown). The sink of cloud water from the microphysics parameterization is the production of precipitation, plotted using dotted lines. This is the sum of the microphysical processes of autoconversion, accretion, and collection. An examination of contribution from individual processes (not shown) finds that the most important processes in precipitation production over the ocean are autoconversion and accretion of cloud water and autoconversion of ice to form snow, with smaller contributions from the other collection terms. In the land regions, the autoconversion of ice is the predominant process for forming precipitation. It is not surprising that the ice phase processes are more dominant over land, as the presence of more aerosols would suppress the warm rain process, leading to a smaller contribution from the autoconversion and accretion of liquid water. The precipitation production is greater in the MPHY run than the CTL run; this is especially noticeable in the subtropics where the increase in detrainment is the largest. The increase is largely due to increases in autoconversion and accretion of cloud water—the ice phase processes are less sensitive to the additional convective detrainment.

Fig. 6.
Fig. 6.

Spatially and temporally averaged profiles of sources (condensation/deposition) and sinks (autoconversion/accretion/collection) of cloud water for each region.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

In terms of relative contribution to the total precipitation from deep and shallow convection and stratiform clouds (Fig. 7), both deep and shallow convection produce less precipitation in the MPHY run in the oceanic regions as well as the SGP. In these regions, the stratiform microphysics is overall more active, leading to more precipitation from the large-scale stratus clouds, particularly in the subtropics. However, the fraction of precipitation originating in stratiform clouds is still smaller than seen in previous observational studies (Schumacher and Houze 2003; Yang and Smith 2008). The changes from CTL to MPHY in the Amazon region are less significant in both convective and large-scale precipitation.

Fig. 7.
Fig. 7.

Precipitation by cloud type, as a percentage of total precipitation in each region.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

The cloud water budgets can be compared to those seen in a CRM, in which convective and stratiform microphysics are not separated and include the type of detail that the SZ11 convective microphysics emulates. A previous study (Z13) looked at cloud water budgets in their CRM simulations by dividing the cloud depth into ice cloud layer (T −35°C), mixed-phase cloud layer (−35° T 0°C) and warm cloud layer (T 0°C). To compare values to those found in their study, we calculate the water budget for the same three layers of the atmosphere. Results are shown in Figs. 8 and 9. There are several differences in the calculations in Figs. 8 and 9 compared to the CRM calculations made by Z13 (shown in their Fig. 11). In the CRM budget, only contributions from mesoscale convective systems are included, while in CAM there is no distinction between isolated and organized convection. Z13 splits up the water budget into components from convective, stratiform raining, and stratiform nonraining (or anvil) clouds. As CAM deals with grid-average quantities, there is no clear separation between raining and nonraining clouds, so the center columns of Figs. 8 and 9 contain information about all stratiform clouds. In the figures, we also include water budget for shallow convection, which is not considered in Z13.

Fig. 8.
Fig. 8.

Terms in the water budget for the oceanic regions, split by temperature. (top to bottom) The layers are ice (T −35°C), mixed phase (−35 T 0°C), and warm cloud (T 0°C). Given for each region/simulation are (left) ZM convection, (center) the large-scale stratus cloud, and (right) shallow convection. Blue numbers represent the net condensation/evaporation of cloud water (liquid + ice), green is the production of precipitation, and red is evaporation of precipitation. Numbers with black arrows represent the transport of cloud water vertically between cloud layers or horizontally between cloud types. Purple arrows (showing downward movement) represent the vertical flux of precipitation through cloud layers and at the surface.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

Fig. 9.
Fig. 9.

As in Fig. 8, but for the land regions.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

Rather than including only one number for the microphysical source of water, we have broken the calculation down into three terms. On the left side of each box in blue is printed the source of cloud water (cloud ice + liquid) from the net condensation/evaporation processes. The number in green in the center of the box represents the conversion from cloud water to precipitation (snow and rain). For the stratiform region this is the term calculated using Eq. (4). Precipitation production from ZM convection in the CTL simulation is calculated using Eq. (1) and in the MPHY simulation using a similar equation to Eq. (4). The negative values in red (on the right of each box) represent the evaporation of precipitation. In the left column of the MPHY simulation, this term is included in the precipitation production, so the number in green should be thought of as the net precipitation production by the convective microphysics. For shallow convection the only term included is the net production of cloud water (equivalent to the sum of the three microphysical terms in the other boxes). As in Z13, horizontal arrows represent fluxes between the different cloud regimes (i.e., detrainment) and vertical arrows represent vertical fluxes between the cloud layers. Here we have calculated the flux of cloud water (in black) separately from that of precipitation (in purple). As in Z13, all quantities are normalized by the total precipitation in the region. Aside from those variables mentioned previously, missing numbers in the budgets indicate quantities that represent less than 1% of the total precipitation amount.

The processes described in Fig. 5 can be seen in Figs. 8 and 9 more clearly, which now include rain and snow. Detrainment from deep and shallow convection is a large source of water in the stratiform clouds, though in the warm cloud layer much of this is evaporated through macrophysics. Additional condensation and deposition occur in the mixed-phase and ice layers, with deposition being the greatest source of cloud water. The precipitation production is a sink of cloud water, effectively the microphysical term in the Fig. 5 bar plots. The conversion from cloud water to precipitation nearly balances the sources of cloud water, consistent with the small overall cloud water tendency in Fig. 5. Whatever precipitation (rain + snow) does not evaporate generally falls out of the layer. In shallow convection, the net condensation is either detrained or falls out as precipitation. The budgets within each cloud type and layer balance to within in the ocean regions, with slightly larger imbalances over land.

Song et al. (2012) showed increased convective mass flux over the western Pacific with the addition of convective microphysics. The regional budgets in Figs. 8 and 9 are consistent with that finding; the vertical flux of cloud water is increased significantly in all four regions from the CTL to the MPHY simulation. As these values are normalized by total precipitation, this points to more water cycling through the physical system in the MPHY simulation for the same amount of resulting surface precipitation. The increased movement of water is also apparent in the increased detrainment of water from convective to stratiform region in the MPHY simulation, and again the increase is more apparent in the ocean regions. The increased detrainment results in an increase in precipitation production, and in all regions except the Amazon the fraction of precipitation falling from stratiform clouds increases. Again, this change is largest in the subtropics.

In comparing our results to that of Z13, it would be expected that the MPHY simulation is more similar to the CRM results, as the SZ11 microphysics includes multiple hydrometeors and microphysical processes closer in complexity to the microphysics utilized in Z13. The first notable difference is that in both the CTL and MPHY simulations the stratiform precipitation is much lower than that simulated by the CRM. Z13 calculates a stratiform precipitation fraction of 41% and 51% for the African Monsoon Multidisciplinary Analyses (AMMA) and Tropical Warm Pool International Cloud Experiment (TWP-ICE) regions respectively, which falls within the range of what has been seen in observations (Schumacher and Houze 2003; Yang and Smith 2008). The MPHY simulation is a slight improvement in this regard. Even with the improvements of the addition of convective microphysics, the ZM scheme is still overproducing convective precipitation.

In the Z13 CRM simulations, detrainment is fairly evenly split between the layers of the atmosphere with values of around 10%–20% of the total precipitation. Detrainment in CAM is too low for the amount of precipitation produced, following with the low stratiform precipitation fraction. There is not enough ice detrainment in any of the simulations, and the shallow cumulus detrainment is far too active compared to what is seen in CRM simulations. Again, the MPHY simulation looks closer than CTL to CRM simulations, but the portioning of liquid water between convective and stratiform clouds still needs improvement.

For a more direct comparison with the CRM simulations, we also completed two single-column model simulations of the TWP-ICE case simulated in Z13, using the single-column version of CAM5 (SCAM). As with the CAM5 simulations, the CTL simulation contains the default ZM configuration and the MPHY utilizes the SZ11 microphysics. Budgets for these simulations are shown in Fig. 10. The SCAM simulations are different from the CAM5 simulations in a few notable ways. In the SCAM simulations, the source of water from large-scale condensation is huge. Most of this water is converted to precipitation and then evaporates before reaching the ground. The SCAM simulations produce more ice, which is closer to what is seen in the CRM. In the SCAM simulations, the difference between CTL and MPHY does not play out in quite the same way. The detrainment is increased in MPHY, but the deep convection actually produces more precipitation, and so the large-scale rainfall decreases in MPHY.

Fig. 10.
Fig. 10.

As in Fig. 8, but for SCAM simulations of the TWP-ICE case.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

c. Radiative impacts

Differences in radiative forcing are expected with changes in cloud amount. Song et al. (2012) noted that cloud fraction of low, midlevel, and high clouds all improved in the MPHY simulation compared to observations, particularly for low clouds over low-latitude oceans. To quantify the resulting changes in the radiation balance, we compare radiative fields to observational data from the NASA Clouds and the Earth’s Radiant Energy System, Energy Balanced and Filled, edition 2.8 (CERES EBAF 2.8; Loeb et al. 2009; http://ceres.larc.nasa.gov) dataset, using a 10-yr (2000–10) annual (or seasonal, where appropriate) average.

Figure 11 shows shortwave and longwave cloud radiative forcing (CRF), from the CERES data and the model simulations. Corresponding global and regional biases are shown in Table 1. The largest biases in the CAM simulations exist in the shortwave CRF field. In the CTL simulation, a large area of cool bias occurs over the western Pacific, while over much of the subtropical oceans there are not enough low clouds [as pointed out by Song et al. (2012)] and therefore not enough shortwave cooling. The addition of convective microphysics increases both the low-cloud amount and the liquid water path over the subtropical oceans, contributing to an improvement in these quantities compared to observations (Song et al. 2012). However, this combination of increased low cloud and liquid water path leads to a worse representation of shortwave CRF in the MPHY simulation compared to observations. The magnitude of the global-mean bias in shortwave CRF increases from −7.36 to −14.57 W m−2 in the MPHY simulation, and the root-mean-square error (RMSE) increases from 20.7 W m−2 in the CTL simulation to 25.5 W m−2 in MPHY.

Fig. 11.
Fig. 11.

(left) Shortwave and (right) longwave cloud forcing (W m−2) (a),(b) averaged over 10 yr of CERES EBAF 2.8 data; the difference between observations and the (c),(d) CTL and (e),(f) MPHY simulations; and (g),(h) the difference between the MPHY and CTL simulations.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

Table 1.

Global- and regional-average (temporally and spatially) radiative fluxes averaged over 10 yr of CERES EBAF 2.8 data and the global-mean bias of the CTL and MPHY simulations. All quantities have units of W m−2.

Table 1.

Longwave cloud forcing in the CAM5 simulations is too large compared to observations, particularly over the Maritime Continent and other tropical land regions. The MPHY simulation is a small improvement, reducing the RMSE by 1.2 W m−2. The regional changes are larger, with the MPHY simulation showing the reduction in the warm bias over land regions. Over much of the globe the changes in longwave and shortwave forcing balance out to produce only small changes in the net CRF. This is not the case in the subtropical oceans, where the substantial increases in shortwave cooling dominate the trend in net CRF. Globally, the magnitude of the net CRF bias increases by 7.36 W m−2, and the RMSE increases by 7.5 W m−2.

OLR has a cool bias when compared to observations (Fig. 12; Table 1). This is driven largely by overactive deep convection in the western Pacific. OLR shows smaller changes between CTL and MPHY than other radiative fields and generally increases over land and decreases over the ocean. SGP shows the largest sensitivity, and the MPHY simulation has a warm bias of 10.84 W m−2 while the CTL simulation had a very small cool bias. This is likely due to a combination of decreased high-cloud amount and increased surface temperature in this region in the MPHY simulation. The global average OLR bias shows a small improvement of 0.46 W m−2.

Fig. 12.
Fig. 12.

OLR (W m−2) (a) averaged over 10 yr of CERES EBAF 2.8 data; the difference between observations and the (b) CTL and (c) MPHY simulations; and (d) the difference between the MPHY and CTL simulations.

Citation: Journal of Climate 28, 23; 10.1175/JCLI-D-15-0064.1

4. Summary and conclusions

We have analyzed two 5-yr GCM simulations using CAM5: a CTL run with the standard ZM convective parameterization and a MPHY run that incorporates the convective microphysics scheme of Song and Zhang (2011). These simulations were also examined by Song et al. (2012), who noted improved convective cloud water content. In the western Pacific, they noted a positive feedback where larger convective detrainment provided more water to the large scale, and increased microphysical activity in the large-scale clouds released more latent heat, leading to enhanced convective activity. We have looked at four regions (the western Pacific warm pool, the subtropics, the U.S. southern Great Plains, and the Amazon) and calculated the water budgets in order to understand the sensitivity to convective microphysics.

Detrainment increased in all four regions, though the change was more pronounced over the ocean, where the simple conversion from cloud water to precipitation in the CTL run led to too high convective precipitation efficiency and low mass of water detrained. Allowing for more physical precipitation formation within deep convection leads to more-realistic differences between land-based and oceanic convection, and the explicit activation of cloud droplets and ice from aerosols in the SZ11 scheme provides the possibility for simulating aerosol indirect effects in a more realistic manner. The increased convective detrainment in the MPHY simulation led to increased large-scale cloudiness, reducing the bias compared to observations (Song et al. 2012). The average values of detrainment for each region in the MPHY run were similar to values calculated in previous studies (Yanai et al. 1973; Houze et al. 1980; Johnson 1980; Chen 1985; Luo and Krueger 2004; Yasunaga et al. 2004), though the calculation of detrainment is dependent on assumptions of model parameters or of the partitioning between convective and stratiform clouds, making it difficult to make direct comparisons.

The water budgets showed that a large percent of the condensate in stratiform clouds comes from convective detrainment, particularly over the ocean, similar to what was seen by previous studies (Gamache and Houze 1983; Zeng et al. 2013). This led to an increase in the precipitation formed from the large-scale clouds. This is an improvement, although the partitioning between convective and stratiform rainfall is still too skewed. Previous observations (Schumacher and Houze 2003; Yang and Smith 2008) put the stratiform rain fraction at closer to half of the total rainfall, so even the improved MPHY simulation is still underestimating this number. In the SCAM simulations, both CTL and MPHY largely overdo the convective precipitation. Convective/stratiform partitioning is a difficult parameter to define and has been specified in different ways in observational data depending on such factors as precipitation intensity and radar reflectivity. This is different from a climate model such as CAM5 that has an arbitrary split between the two, and the amount of precipitation coming from each cloud type is clearly sensitive to many factors. Generally, however, there is too much convective precipitation in CAM5, and so reducing the convective precipitation efficiency is a positive result.

Over both of the ocean regions analyzed, there is evidence for the feedback discussed by Song et al. (2012)—that is, increased detrainment coupled with increased large-scale microphysical activity and enhanced convective mass flux (not shown). In the subtropics, however, this may not be attributable to enhanced vertical motion in the large scale (this is a region of primarily subsidence). Instead, enhanced relative humidity from the large increases in detrainment acts to precondition the atmosphere such that more convection will form. The increased frequency of convective occurrence adds even more to the detrainment—a positive feedback that results in the substantial differences seen in this region. In the western Pacific, convection occurs too frequently compared to observations, and so this increased cloudiness is not an improvement; however, in the subtropics the low-cloud cover in MPHY much better matches what is seen in satellite data. Over land, although there do exist smaller increases in detrainment, the relative humidity decreases in the MPHY simulation, leading to an overall reduction in cloudiness in these regions.

While cloud fields show general improvement with the addition of SZ11 convective microphysics, radiative fields in the MPHY simulation show degradation compared to observations. Globally, there are very small improvements in longwave CRF and in OLR; however, small global biases often result from offsetting larger errors on the regional scale. Additionally, the already substantial low bias in shortwave CRF is doubled in the MPHY simulation. This large radiative cooling must be reduced to increase confidence in simulations utilizing the convective microphysics.

In many ways the MPHY simulation is an improvement compared to observations. While it is not yet known how much this result can be generalized, it is promising that a microphysics parameterization that is more detailed and physically more realistic can improve on some known issues in climate simulations. It is important to include realistic effects of convective microphysics, as we have shown they have wide-ranging impacts on the rest of the hydrologic cycle.

Acknowledgments

This research was supported by the U.S. Department of Energy under Grant DE-SC0008880, the National Science Foundation Grants AGS1015964 and EaSM-1048995, and the National Oceanic and Atmospheric Administration Grant NA11OAR4321098. We acknowledge the NASA CloudSat team and the CERES team for making data publicly available. The authors thank the three anonymous reviewers who contributed useful suggestions and comments to improve this manuscript.

REFERENCES

  • Arakawa, A., 2004: The cumulus parameterization problem: Past, present, and future. J. Climate, 17, 24932525, doi:10.1175/1520-0442(2004)017<2493:RATCPP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Arakawa, A., , and W. H. Schubert, 1974: Interaction of a cumulus cloud ensemble with the large-scale environment, part I. J. Atmos. Sci., 31, 674701, doi:10.1175/1520-0469(1974)031<0674:IOACCE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Arakawa, A., , and C.-M. Wu, 2013: A unified representation of deep moist convection in numerical modeling of the atmosphere. Part I. J. Atmos. Sci., 70, 19771992, doi:10.1175/JAS-D-12-0330.1.

    • Search Google Scholar
    • Export Citation
  • Arakawa, A., , J.-H. Jung, , and C.-M. Wu, 2011: Toward unification of the multiscale modeling of the atmosphere. Atmos. Chem. Phys., 11, 37313742, doi:10.5194/acp-11-3731-2011.

    • Search Google Scholar
    • Export Citation
  • Bretherton, C. S., , and S. Park, 2009: A new moist turbulence parameterization in the Community Atmosphere Model. J. Climate, 22, 34223448, doi:10.1175/2008JCLI2556.1.

    • Search Google Scholar
    • Export Citation
  • Chen, Y.-L., 1985: Diagnosis of the net cloud mass flux in GATE. J. Atmos. Sci., 42, 17571769, doi:10.1175/1520-0469(1985)042<1757:DOTNCM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Dai, A., 2006: Precipitation characteristics in eighteen coupled climate models. J. Climate, 19, 46054630, doi:10.1175/JCLI3884.1.

  • Del Genio, A. D., , J. Wu, , and Y. Chen, 2012: Characteristics of mesoscale organization in WRF simulations of convection during TWP-ICE. J. Climate, 25, 56665688, doi:10.1175/JCLI-D-11-00422.1.

    • Search Google Scholar
    • Export Citation
  • Donner, L. J., 1993: A cumulus parameterization including mass fluxes, vertical momentum dynamics, and mesoscale effects. J. Atmos. Sci., 50, 889906, doi:10.1175/1520-0469(1993)050<0889:ACPIMF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K., 1991: A scheme for representing cumulus convection in large-scale models. J. Atmos. Sci., 48, 23132329, doi:10.1175/1520-0469(1991)048<2313:ASFRCC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K., , and R. Pierrehumbert, 1996: Microphysical and dynamical control of tropospheric water vapor. Clouds, Chemistry and Climate, P. Crutzen and V. Ramanathan, Eds., NATO ASI Series, Vol. 35, Springer, 17–28, doi:10.1007/978-3-642-61051-6_2.

  • Fu, Q., , S. K. Krueger, , and K. N. Liou, 1995: Interactions of radiation and convection in simulated tropical cloud clusters. J. Atmos. Sci., 52, 13101328, doi:10.1175/1520-0469(1995)052<1310:IORACI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gamache, J. F., , and R. A. Houze, 1983: Water budget of a mesoscale convective system in the tropics. J. Atmos. Sci., 40, 18351850, doi:10.1175/1520-0469(1983)040<1835:WBOAMC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., , C.-P. Cheng, , C. A. Leary, , and J. F. Gamache, 1980: Diagnosis of cloud mass and heat fluxes from radar and synoptic data. J. Atmos. Sci., 37, 754773, doi:10.1175/1520-0469(1980)037<0754:DOCMAH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Iacono, M. J., , J. S. Delamere, , E. J. Mlawer, , M. W. Shephard, , S. A. Clough, , and W. D. Collins, 2008: Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res., 113, D13103, doi:10.1029/2008JD009944.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., 1980: Diagnosis of convective and mesoscale motions during phase III of GATE. J. Atmos. Sci., 37, 733753, doi:10.1175/1520-0469(1980)037<0733:DOCAMM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., , T. M. Rickenbach, , S. A. Rutledge, , P. E. Ciesielski, , and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12, 23972418, doi:10.1175/1520-0442(1999)012<2397:TCOTC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kuo, H. L., 1974: Further studies of the parameterization of the influence of cumulus convection on large-scale flow. J. Atmos. Sci., 31, 12321240, doi:10.1175/1520-0469(1974)031<1232:FSOTPO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Leary, C. A., , and R. A. Houze, 1980: The contribution of mesoscale motions to the mass and heat fluxes of an intense tropical convective system. J. Atmos. Sci., 37, 784796, doi:10.1175/1520-0469(1980)037<0784:TCOMMT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lindzen, R. S., , M.-D. Chou, , and A. Y. Hou, 2001: Does the Earth have an adaptive infrared iris? Bull. Amer. Meteor. Soc., 82, 417432, doi:10.1175/1520-0477(2001)082<0417:DTEHAA>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liu, Y.-C., , J. Fan, , G. J. Zhang, , K.-M. Xu, , and S. Ghan, 2015: Improving representation of convective transport for scale-aware parameterization: 2. Analysis of cloud-resolving model simulations. J. Geophys. Res. Atmos., 120, 35103532, doi:10.1002/2014JD022145.

    • Search Google Scholar
    • Export Citation
  • Loeb, N. G., , B. A. Wielicki, , D. R. Doelling, , G. L. Smith, , D. F. Keyes, , S. Kato, , N. Manalo-Smith, , and T. Wong, 2009: Toward optimal closure of the Earth’s top-of-atmosphere radiation budget. J. Climate, 22, 748766, doi:10.1175/2008JCLI2637.1.

    • Search Google Scholar
    • Export Citation
  • Lohmann, U., 2008: Global anthropogenic aerosol effects on convective clouds in ECHAM5-HAM. Atmos. Chem. Phys., 8, 21152131, doi:10.5194/acp-8-2115-2008.

    • Search Google Scholar
    • Export Citation
  • Lord, S. J., 1982: Interaction of a cumulus cloud ensemble with the large-scale environment. Part III: Semi-prognostic test of the Arakawa-Schubert cumulus parameterization. J. Atmos. Sci., 39, 88103, doi:10.1175/1520-0469(1982)039<0088:IOACCE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Luo, Y., , and S. K. Krueger, 2004: Evaluation of detrainment and microphysics parameterization in the NCEP GFS single-column model. Proc. 14th ARM Science Team Meeting, Albuquerque, NM, Atmospheric Radiation Measurement Program. [Available online at https://www.arm.gov/publications/proceedings/conf14/extended_abs/luo2-y.pdf.]

  • Mace, G. G., , and Q. Zhang, 2014: The CloudSat radar-lidar geometrical profile product (RL-GeoProf): Updates, improvements, and selected results. J. Geophys. Res. Atmos., 119, 94419462, doi:10.1002/2013JD021374.

    • Search Google Scholar
    • Export Citation
  • Mace, G. G., , M. Deng, , B. Soden, , and E. Zipser, 2006: Association of tropical cirrus in the 10–15-km layer with deep convective sources: An observational study combining millimeter radar data and satellite-derived trajectories. J. Atmos. Sci., 63, 480503, doi:10.1175/JAS3627.1.

    • Search Google Scholar
    • Export Citation
  • Morrison, H., , and A. Gettelman, 2008: A new two-moment bulk stratiform cloud microphysics scheme in the Community Atmosphere Model, version 3 (CAM3). Part I: Description and numerical tests. J. Climate, 21, 36423659, doi:10.1175/2008JCLI2105.1.

    • Search Google Scholar
    • Export Citation
  • Neale, R. B., , J. H. Richter, , and M. Jochum, 2008: The impact of convection on ENSO: From a delayed oscillator to a series of events. J. Climate, 21, 59045924, doi:10.1175/2008JCLI2244.1.

    • Search Google Scholar
    • Export Citation
  • Park, S., 2014: A unified convection scheme (UNICON). Part I: Formulation. J. Atmos. Sci., 71, 39023930, doi:10.1175/JAS-D-13-0233.1.

    • Search Google Scholar
    • Export Citation
  • Park, S., , and C. S. Bretherton, 2009: The University of Washington shallow convection and moist turbulence schemes and their impact on climate simulations with the Community Atmosphere Model. J. Climate, 22, 34493469, doi:10.1175/2008JCLI2557.1.

    • Search Google Scholar
    • Export Citation
  • Ramanathan, V., , and W. Collins, 1991: Thermodynamic regulation of ocean warming by cirrus clouds deduced from observations of the 1987 El Niño. Nature, 351, 2732, doi:10.1038/351027a0.

    • Search Google Scholar
    • Export Citation
  • Randall, D. A., 2013: Beyond deadlock. Geophys. Res. Lett., 40, 59705976, doi:10.1002/2013GL057998.

  • Randall, D. A., , Harshvardhan, , D. A. Dazlich, , and T. G. Corsetti, 1989: Interactions among radiation, convection, and large-scale dynamics in a general circulation model. J. Atmos. Sci., 46, 19431970, doi:10.1175/1520-0469(1989)046<1943:IARCAL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rennó, N. O., , K. A. Emanuel, , and P. H. Stone, 1994: Radiative-convective model with an explicit hydrologic cycle: 1. Formulation and sensitivity to model parameters. J. Geophys. Res., 99, 14 42914 441, doi:10.1029/94JD00020.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and R. A. Houze, 2003: Stratiform rain in the tropics as seen by the TRMM Precipitation Radar. J. Climate, 16, 17391756, doi:10.1175/1520-0442(2003)016<1739:SRITTA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Song, X., , and G. J. Zhang, 2011: Microphysics parameterization for convective clouds in a global climate model: Description and single-column model tests. J. Geophys. Res., 116, D02201, doi:10.1029/2010JD014833.

    • Search Google Scholar
    • Export Citation
  • Song, X., , G. J. Zhang, , and J.-L. F. Li, 2012: Evaluation of microphysics parameterization for convective clouds in the NCAR Community Atmosphere Model CAM5. J. Climate, 25, 85688590, doi:10.1175/JCLI-D-11-00563.1.

    • Search Google Scholar
    • Export Citation
  • Stephens, G. L., , S. van den Heever, , and L. Pakula, 2008: Radiative–convective feedbacks in idealized states of radiative–convective equilibrium. J. Atmos. Sci., 65, 38993916, doi:10.1175/2008JAS2524.1.

    • Search Google Scholar
    • Export Citation
  • Tiedtke, M., 1989: A comprehensive mass flux scheme for cumulus parameterization in large-scale models. Mon. Wea. Rev., 117, 17791800, doi:10.1175/1520-0493(1989)117<1779:ACMFSF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wall, C., , C. Liu, , and E. Zipser, 2013: A climatology of tropical congestus using CloudSat. J. Geophys. Res. Atmos., 118, 64786492, doi:10.1002/jgrd.50455.

    • Search Google Scholar
    • Export Citation
  • Wang, Z., , and K. Sassen, 2001: Cloud type and macrophysical property retrieval using multiple remote sensors. J. Appl. Meteor., 40, 16651682, doi:10.1175/1520-0450(2001)040<1665:CTAMPR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , S. Esbensen, , and J.-H. Chu, 1973: Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611627, doi:10.1175/1520-0469(1973)030<0611:DOBPOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yang, S., , and E. A. Smith, 2008: Convective–stratiform precipitation variability at seasonal scale from 8 yr of TRMM observations: Implications for multiple modes of diurnal variability. J. Climate, 21, 40874114, doi:10.1175/2008JCLI2096.1.

    • Search Google Scholar
    • Export Citation
  • Yasunaga, K., , H. Kida, , T. Satomura, , and N. Nishi, 2004: A numerical study on the detrainment of tracers by cumulus convection in TOGA COARE. J. Meteor. Soc. Japan, 82, 861878, doi:10.2151/jmsj.2004.861.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , W.-K. Tao, , S. W. Powell, , R. A. Houze, , P. Ciesielski, , N. Guy, , H. Pierce, , and T. Matsui, 2013: A comparison of the water budgets between clouds from AMMA and TWP-ICE. J. Atmos. Sci., 70, 487503, doi:10.1175/JAS-D-12-050.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, G., , and N. A. McFarlane, 1995: Sensitivity of climate simulations to the parameterization of cumulus convection in the Canadian Climate Centre general circulation model. Atmos.–Ocean, 33, 407446, doi:10.1080/07055900.1995.9649539.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., , U. Lohmann, , and P. Stier, 2005: A microphysical parameterization for convective clouds in the ECHAM5 climate model: Single-column model results evaluated at the Oklahoma Atmospheric Radiation Measurement Program site. J. Geophys. Res., 110, D15S07, doi:10.1029/2004JD005128.

    • Search Google Scholar
    • Export Citation
  • Zhang, Y., , and S. A. Klein, 2010: Mechanisms affecting the transition from shallow to deep convection over land: Inferences from observations of the diurnal cycle collected at the ARM Southern Great Plains site. J. Atmos. Sci., 67, 29432959, doi:10.1175/2010JAS3366.1.

    • Search Google Scholar
    • Export Citation
  • Zhao, M., 2014: An investigation of the connections among convection, clouds, and climate sensitivity in a global climate model. J. Climate, 27, 18451862, doi:10.1175/JCLI-D-13-00145.1.

    • Search Google Scholar
    • Export Citation
Save