• Abel, S. J., , D. N. Walters, , and G. Allen, 2010: Evaluation of stratocumulus cloud prediction in the Met Office forecast model during VOCALS-REx. Atmos. Chem. Phys., 10, 10 54110 559, doi:10.5194/acp-10-10541-2010.

    • Search Google Scholar
    • Export Citation
  • Andrejczuk, M., , W. W. Grabowski, , A. Gadian, , and R. Burton, 2012: Limited-area modelling of stratocumulus over South-Eastern Pacific. Atmos. Chem. Phys., 12, 35113526, doi:10.5194/acp-12-3511-2012.

    • Search Google Scholar
    • Export Citation
  • Boo, K.-O., , B. B. B. Booth, , Y.-H. Byun, , J. Lee, , C. Cho, , S. Shim, , and K.-T. Kim, 2015: Influence of aerosols in multidecadal SST variability simulations over the North Pacific. J. Geophys. Res. Atmos., 120, 517531, doi:10.1002/2014JD021933.

    • Search Google Scholar
    • Export Citation
  • Booth, B. B. B., , N. J. Dunstone, , P. R. Halloran, , T. Andrews, , and N. Bellouin, 2012: Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability. Nature, 484, 228232, doi:10.1038/nature10946.

    • 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
  • Burleyson, C. D., , S. P. de Szoeke, , S. E. Yuter, , M. Wilbanks, , and W. A. Brewer, 2013: Ship-based observations of the diurnal cycle of Southeast Pacific marine stratocumulus clouds and precipitation. J. Atmos. Sci., 70, 38763894, doi:10.1175/JAS-D-13-01.1.

    • Search Google Scholar
    • Export Citation
  • Calisto, M., , D. Folini, , M. Wild, , and L. Bengtsson, 2014: Cloud radiative forcing intercomparison between fully coupled CMIP5 models and CERES satellite data. Ann. Geophys., 32, 793807, doi:10.5194/angeo-32-793-2014.

    • Search Google Scholar
    • Export Citation
  • Chen, G., , W.-C. Wang, , and J.-P. Chen, 2015: Aerosol–stratocumulus–radiation interactions over the southeast Pacific. J. Atmos. Sci., 72, 26122621, doi:10.1175/JAS-D-14-0319.1.

    • Search Google Scholar
    • Export Citation
  • Chen, J. P., , and S. T. Liu, 2004: Physically based two-moment bulkwater parametrization for warm-cloud microphysics. Quart. J. Roy. Meteor. Soc., 130, 5178, doi:10.1256/qj.03.41.

    • Search Google Scholar
    • Export Citation
  • Cheng, C.-T., , W.-C. Wang, , and J.-P. Chen, 2007: A modelling study of aerosol impacts on cloud microphysics and radiative properties. Quart. J. Roy. Meteor. Soc., 133, 283297, doi:10.1002/qj.25.

    • Search Google Scholar
    • Export Citation
  • Cheng, C.-T., , W.-C. Wang, , and J.-P. Chen, 2010: Simulation of the effects of increasing cloud condensation nuclei on mixed-phase clouds and precipitation of a front system. Atmos. Res., 96, 461476, doi:10.1016/j.atmosres.2010.02.005.

    • Search Google Scholar
    • Export Citation
  • Diehl, T., , A. Heil, , M. Chin, , X. Pan, , D. Streets, , M. Schultz, , and S. Kinne, 2012: Anthropogenic, biomass burning, and volcanic emissions of black carbon, organic carbon, and SO2 from 1980 to 2010 for hindcast model experiments. Atmos. Chem. Phys. Discuss., 12, 24 89524 954, doi:10.5194/acpd-12-24895-2012.

    • Search Google Scholar
    • Export Citation
  • Ekman, A. M. L., 2014: Do sophisticated parameterizations of aerosol-cloud interactions in CMIP5 models improve the representation of recent observed temperature trends? J. Geophys. Res. Atmos., 119, 817832, doi:10.1002/2013JD020511.

    • Search Google Scholar
    • Export Citation
  • European Centre for Medium-Range Weather Forecasts, 2012: ERA-Interim project, monthly means. National Center for Atmospheric Research Computational and Information Systems Laboratory Research Data Archive, accessed 6 March 2015, doi:10.5065/D68050NT.

  • Flato, G., and et al. , 2013: Evaluation of climate models. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 741–866.

  • Granier, C., and et al. , 2011: Evolution of anthropogenic and biomass burning emissions of air pollutants at global and regional scales during the 1980–2010 period. Climatic Change, 109, 163190, doi:10.1007/s10584-011-0154-1.

    • Search Google Scholar
    • Export Citation
  • Grenier, H., , and C. S. Bretherton, 2001: A moist PBL parameterization for large-scale models and its application to subtropical cloud-topped marine boundary layers. Mon. Wea. Rev., 129, 357377, doi:10.1175/1520-0493(2001)129<0357:AMPPFL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hazra, A., , V. Mandal, , and J.-P. Chen, 2013: Study of cloud microphysical properties over India during CAIPEEX using a mesoscale model with new cloud microphysical scheme—Part I. J. Atmos. Sol.-Terr. Phys., 93, 2944, doi:10.1016/j.jastp.2012.11.010.

    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., , Y. Noh, , and J. Dudhia, 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 23182341, doi:10.1175/MWR3199.1.

    • Search Google Scholar
    • Export Citation
  • Hubanks, P. A., , M. D. King, , S. A. Platnick, , and R. A. Pincus, 2008: MODIS Atmosphere L3 gridded product. MODIS Algorithm Theoretical Basis Doc. ATBD-MOD-30, 96 pp. [Available online at http://modis-atmos.gsfc.nasa.gov/_docs/L3_ATBD_2008_12_04.pdf.]

  • 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
  • Kato, S., , N. G. Loeb, , F. G. Rose, , D. R. Doelling, , D. A. Rutan, , T. E. Caldwell, , L. S. Yu, , and R. A. Weller, 2013: Surface irradiances consistent with CERES-derived top-of-atmosphere shortwave and longwave irradiances. J. Climate, 26, 27192740, doi:10.1175/JCLI-D-12-00436.1.

    • Search Google Scholar
    • Export Citation
  • Kubar, T. L., , D. E. Waliser, , J. L. Li, , and X. Jiang, 2012: On the annual cycle, variability, and correlations of oceanic low-topped clouds with large-scale circulation using Aqua MODIS and ERA-Interim. J. Climate, 25, 61526174, doi:10.1175/JCLI-D-11-00478.1.

    • Search Google Scholar
    • Export Citation
  • Lamarque, J. F., and et al. , 2010: Historical (1850–2000) gridded anthropogenic and biomass burning emissions of reactive gases and aerosols: Methodology and application. Atmos. Chem. Phys., 10, 70177039, doi:10.5194/acp-10-7017-2010.

    • Search Google Scholar
    • Export Citation
  • Large, W. G., , and G. Danabasoglu, 2006: Attribution and impacts of upper-ocean biases in CCSM3. J. Climate, 19, 23252346, doi:10.1175/JCLI3740.1.

    • Search Google Scholar
    • Export Citation
  • Min, Q., and et al. , 2012: Comparison of MODIS cloud microphysical properties with in-situ measurements over the Southeast Pacific. Atmos. Chem. Phys., 12, 11 26111 273, doi:10.5194/acp-12-11261-2012.

    • Search Google Scholar
    • Export Citation
  • Nam, C., , S. Bony, , J. L. Dufresne, , and H. Chepfer, 2012: The ‘too few, too bright’ tropical low-cloud problem in CMIP5 models. Geophys. Res. Lett., 39, L21801, doi:10.1029/2012GL053421.

    • Search Google Scholar
    • Export Citation
  • National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce, 2000: NCEP FNL Operational Model Global Tropospheric Analyses, continuing from July 1999 (updated daily). National Center for Atmospheric Research Computational and Information Systems Laboratory Research Data Archive, accessed 16 December 2013. [Available online at http://rda.ucar.edu/datasets/ds083.2/.]

  • Painemal, D., , and P. Zuidema, 2011: Assessment of MODIS cloud effective radius and optical thickness retrievals over the Southeast Pacific with VOCALS-REx in situ measurements. J. Geophys. Res., 116, D24206, doi:10.1029/2011JD016155.

    • Search Google Scholar
    • Export Citation
  • Park, S., , C. Deser, , and M. A. Alexander, 2005: Estimation of the surface heat flux response to sea surface temperature anomalies over the global oceans. J. Climate, 18, 45824599, doi:10.1175/JCLI3521.1.

    • Search Google Scholar
    • Export Citation
  • Pollard, R. T., , P. B. Rhines, , and R. O. R. Y. Thompson, 1972: The deepening of the wind—Mixed layer. Geophys. Fluid Dyn., 4, 381404, doi:10.1080/03091927208236105.

    • Search Google Scholar
    • Export Citation
  • Rahn, D. A., , and R. Garreaud, 2010a: Marine boundary layer over the subtropical southeast Pacific during VOCALS-REx—Part 1: Mean structure and diurnal cycle. Atmos. Chem. Phys., 10, 44914506, doi:10.5194/acp-10-4491-2010.

    • Search Google Scholar
    • Export Citation
  • Rahn, D. A., , and R. Garreaud, 2010b: Marine boundary layer over the subtropical southeast Pacific during VOCALS-REx—Part 2: Synoptic variability. Atmos. Chem. Phys., 10, 45074519, doi:10.5194/acp-10-4507-2010.

    • Search Google Scholar
    • Export Citation
  • Rapp, A. D., 2015: Cloud responses in the AMIP simulations of CMIP5 models in the southeastern Pacific marine subsidence region. Int. J. Climatol., 35, 29082921, doi:10.1002/joc.4181.

    • Search Google Scholar
    • Export Citation
  • Richter, I., 2015: Climate model biases in the eastern tropical oceans: Causes, impacts and ways forward. Wiley Interdiscip. Rev.: Climate Change, 6, 345358, doi:10.1002/wcc.338.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., , H. L. Wang, , and P. J. Rasch, 2012: The roles of cloud drop effective radius and LWP in determining rain properties in marine stratocumulus. Geophys. Res. Lett., 39, L13801, doi:10.1029/2012GL052028.

    • Search Google Scholar
    • Export Citation
  • van der Werf, G. R., , J. T. Randerson, , L. Giglio, , G. J. Collatz, , P. S. Kasibhatla, , and A. F. Arellano, 2006: Interannual variability in global biomass burning emissions from 1997 to 2004. Atmos. Chem. Phys., 6, 34233441, doi:10.5194/acp-6-3423-2006.

    • Search Google Scholar
    • Export Citation
  • Voldoire, A., , M. Claudon, , G. Caniaux, , H. Giordani, , and R. Roehrig, 2014: Are atmospheric biases responsible for the tropical Atlantic SST biases in the CNRM-CM5 coupled model? Climate Dyn., 43, 29632984, doi:10.1007/s00382-013-2036-x.

    • Search Google Scholar
    • Export Citation
  • Wang, C. Z., , L. P. Zhang, , S. K. Lee, , L. X. Wu, , and C. R. Mechoso, 2014: A global perspective on CMIP5 climate model biases. Nat. Climate Change, 4, 201205, doi:10.1038/nclimate2118.

    • Search Google Scholar
    • Export Citation
  • Wang, S., , L. W. O’Neill, , Q. Jiang, , S. P. de Szoeke, , X. Hong, , H. Jin, , W. T. Thompson, , and X. Zheng, 2011: A regional real-time forecast of marine boundary layers during VOCALS-REx. Atmos. Chem. Phys., 11, 421437, doi:10.5194/acp-11-421-2011.

    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., , B. R. Barkstrom, , E. F. Harrison, , R. B. Lee, , G. Louis Smith, , and J. E. Cooper, 1996: Clouds and the Earth’s Radiant Energy System (CERES): An Earth Observing System experiment. Bull. Amer. Meteor. Soc., 77, 853868, doi:10.1175/1520-0477(1996)077<0853:CATERE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wilcox, L. J., , E. J. Highwood, , and N. J. Dunstone, 2013: The influence of anthropogenic aerosol on multi-decadal variations of historical global climate. Environ. Res. Lett., 8, 024033, doi:10.1088/1748-9326/8/2/024033.

    • Search Google Scholar
    • Export Citation
  • Wood, R., and et al. , 2011: The VAMOS Ocean-Cloud-Atmosphere-Land Study Regional Experiment (VOCALS-REx): Goals, platforms, and field operations. Atmos. Chem. Phys., 11, 627654, doi:10.5194/acp-11-627-2011.

    • Search Google Scholar
    • Export Citation
  • Zeng, X. B., , and A. Beljaars, 2005: A prognostic scheme of sea surface skin temperature for modeling and data assimilation. Geophys. Res. Lett., 32, L14605, doi:10.1029/2005GL023030.

    • Search Google Scholar
    • Export Citation
  • Zheng, Y., , G. N. Kiladis, , T. Shinoda, , E. J. Metzger, , H. E. Hurlburt, , J. Lin, , and B. S. Giese, 2010: Upper-ocean processes under the stratus cloud deck in the southeast Pacific Ocean. J. Phys. Oceanogr., 40, 103120, doi:10.1175/2009JPO4213.1.

    • Search Google Scholar
    • Export Citation
  • Zheng, Y., , T. Shinoda, , J.-L. Lin, , and G. N. Kiladis, 2011: Sea surface temperature biases under the stratus cloud deck in the southeast Pacific Ocean in 19 IPCC AR4 coupled general circulation models. J. Climate, 24, 41394164, doi:10.1175/2011JCLI4172.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Setups of (a) the domain location and (b) the vertical levels. One domain without further nesting was used in this study. In (a), the two dashed rectangles indicate the VOCALS region (larger one) and the coastal region (smaller one); the black line indicates the flight track of C130-RF04 during the VOCALS campaign. In (b), the solid and dashed lines indicate the 63- and 37-level vertical setups, respectively.

  • View in gallery

    Cloud properties from (a)–(d) MODIS and simulations with the fixed SST [(e)–(h) PF and (i)–(l) CF]. The first 2 days of simulations were discarded as the model spinup. Cloud droplet number concentration in (c) is calculated with the formula in Painemal and Zuidema (2011). The simulated cloud fractions in (e) and (i) were calculated by assuming maximum/random overlapping. The simulated LWP in (f) and (j) only included cloud droplets to be consistent with the calculation of radiation transfer. Cloud droplet effective radius in (h) and (l) only considered cloud droplets around cloud top to be consistent with the MODIS observation.

  • View in gallery

    Cloud radiative forcing at the surface from (a),(b) CERES SYN1deg and (c),(d) simulations with the fixed SST. (e),(f) Different scales are used for the differences in SWCRF and LWCRF between PF and CF (PF minus CF).

  • View in gallery

    (a) The initial SST in PM and CM, the simulated final SST in (b) PM and (c) CM, and (d) the difference between the latter two (PM minus CM). The SST near the coast is reduced by as much as 2°C owing to the CRF increase caused by anthropogenic aerosols.

  • View in gallery

    Temporal evolution of mean SST for the coastal region (smaller dashed rectangle in Fig. 1) from PM (black) and CM (gray). The thin lines include the initial SST and the final SSTs of each forcing round, while the thick lines include the mean SSTs within each forcing round.

  • View in gallery

    Correlation between the surface cooling and the CRF increase caused by anthropogenic aerosols for the coastal region. The solid line represents the linear regression result (ΔSST = 0.0975ΔCRF + 0.0514) with the ordinary least squares method. The correlation coefficient is 0.594.

  • View in gallery

    Vertical profiles of (a) potential temperature, (b) water vapor mixing ratio, (c) cloud water mixing ratio, and (d) horizontal wind speed over the coastal region, averaged over the last four forcing rounds (28 days). The dashed lines indicate results from PF (black) and CF (gray), while the solid lines indicate results from PM (black) and CM (gray).

  • View in gallery

    Diurnal cycle of (a) cloud fraction and (b) LWP over the coastal region, averaged over the last four forcing rounds (28 days). The dashed lines indicate results from PF (black) and CF (gray), while the solid lines indicate results from PM (black) and CM (gray). The local daytime is around 1200–2300 UTC.

  • View in gallery

    Time series of the low-cloud fraction over the coastal region from PF during the first forcing round (thin) and the last 15 forcing rounds (thick). Open circles represent the GOES-10 observation.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 61 61 5
PDF Downloads 57 57 2

Aerosol–Stratocumulus–Radiation Interactions over the Southeast Pacific: Implications to the Underlying Air–Sea Coupling

View More View Less
  • 1 Atmospheric Sciences Research Center, University at Albany, State University of New York, Albany, New York
© Get Permissions
Full access

Abstract

Recently, Chen et al. used a combination of observations and WRF simulations to illustrate that the anthropogenic aerosol–cloud microphysics–radiation interactions over the southeast Pacific can potentially reduce the excessive shortwave radiation reaching the sea surface, a common bias identified in CMIP5 models. Here, with the aid of a mixed-layer ocean, the authors further study the implications of the shortwave radiation reduction to the underlying air–sea coupling, focusing on the SST sensitivity to the changes. Results show that responses of the air–sea coupling include two negative feedbacks (a large decrease in the latent heat flux and a small decrease in the sensible heat flux, both associated with the surface cooling) and a positive feedback (an increase in the cloud cover, caused by the increase in the relative humidity within the boundary layer, especially during the daytime). The 0.1°C (W m−2)−1 SST sensitivity is about half that documented in CMIP5 models. In addition, an effective daytime cloud fraction weighted with the solar diurnal cycle is proposed to facilitate diagnosing the intensity of cloud–radiation interactions in general circulation models.

Corresponding author address: Wei-Chyung Wang, Atmospheric Sciences Research Center, University at Albany, State University of New York, 251 Fuller Road, Albany, NY 12203. E-mail: wcwang@albany.edu

Abstract

Recently, Chen et al. used a combination of observations and WRF simulations to illustrate that the anthropogenic aerosol–cloud microphysics–radiation interactions over the southeast Pacific can potentially reduce the excessive shortwave radiation reaching the sea surface, a common bias identified in CMIP5 models. Here, with the aid of a mixed-layer ocean, the authors further study the implications of the shortwave radiation reduction to the underlying air–sea coupling, focusing on the SST sensitivity to the changes. Results show that responses of the air–sea coupling include two negative feedbacks (a large decrease in the latent heat flux and a small decrease in the sensible heat flux, both associated with the surface cooling) and a positive feedback (an increase in the cloud cover, caused by the increase in the relative humidity within the boundary layer, especially during the daytime). The 0.1°C (W m−2)−1 SST sensitivity is about half that documented in CMIP5 models. In addition, an effective daytime cloud fraction weighted with the solar diurnal cycle is proposed to facilitate diagnosing the intensity of cloud–radiation interactions in general circulation models.

Corresponding author address: Wei-Chyung Wang, Atmospheric Sciences Research Center, University at Albany, State University of New York, 251 Fuller Road, Albany, NY 12203. E-mail: wcwang@albany.edu

1. Introduction

As revealed by Calisto et al. (2014) and Flato et al. (2013), the CMIP5 models underestimate the solar cloud radiative forcing by stratocumulus clouds and have too much shortwave radiation reaching the surface. This is consistent with the simulated warmer SSTs over the regions topped by extensive marine stratocumulus clouds (Wang et al. 2014). The biases in cloud radiative forcing can be attributed to biases in cloud macroproperties (e.g., too few, too bright issue; Nam et al. 2012) or cloud microproperties or both. It has been noticed that the CMIP5 models that allow aerosol–cloud microphysics interactions can better simulate the observed surface temperatures (Ekman 2014; Wilcox et al. 2013) and the decadal temperature variations over the North Atlantic and North Pacific (Boo et al. 2015; Booth et al. 2012). Thus, it is possible that the SST biases are caused by deficiencies in representing these processes in these models.

Recently, we (Chen et al. 2015) showed that anthropogenic aerosols emitted from South America greatly increased the shortwave cloud radiative forcing of stratocumulus clouds over the southeast Pacific (SEP) and decreased the shortwave radiation reaching the sea surface by up to 30 W m−2 over the coastal region. Here, we further investigate the implications of aerosol–cloud–radiation interactions to the air–sea coupling and subsequently the SST sensitivity to changes in the shortwave radiation.

The surface heat flux feedback is an important issue related to the SST responses to the shortwave radiation changes. The net surface heat flux feedback to the SST anomaly is generally negative (i.e., damping effect on the SST anomaly) over the global ocean, but positive-feedback pathways are also identified over some regions associated with responses of surface wind speed (positive turbulent heat flux feedback), stratocumulus cloud fraction (positive shortwave flux feedback), and atmospheric water vapor (positive longwave flux feedback) (Park et al. 2005). The response of stratocumulus cloud fraction is the most concern over the SEP region. The stratocumulus has an inverse seasonal cycle to the SST with the maximum cloud cover leading the minimum SST, suggesting that the cooler SST induces more cloud cover, which further enhances the surface cooling by reducing solar radiation reaching the surface (Kubar et al. 2012).

In addition to biases in the surface heat fluxes, biases in the simulated cold water upwelling and ocean currents also contribute to the warm SST biases (e.g., Large and Danabasoglu 2006; Richter 2015). Over the SEP, the upwelling cold water along the coast is transported offshore by ocean currents, which plays a dominating role in maintaining the annual-mean SST (Zheng et al. 2010). Zheng et al. (2011) argued that, although biases of individual surface heat components are pronounced (especially those of the latent heat flux and shortwave and longwave radiation), biases of the net surface heat fluxes are negative for most models and therefore cannot explain the warm SST biases, and that the biases in heat transport by Ekman currents largely contribute to the warm SST biases over both the coastal region and the remote ocean. However, the results from the CNRM-CM5 model showed that a correction to the surface solar net heat flux could reduce the warm SST biases by more than 50% over the equatorial Atlantic Ocean (Voldoire et al. 2014). Therefore, investigating how much the shortwave reduction can potentially lessen the SST biases is still warranted.

This study was conducted by examining the equilibrium responses of the air–sea coupling when a simple mixed-layer ocean was coupled to the WRF Model [the same version used in Chen et al. (2015)]. Stratocumulus clouds during the 2008 Variability of American Monsoon Systems Ocean–Cloud–Atmosphere–Land Study (VOCALS) campaign (Wood et al. 2011) were simulated and compared with satellite observations. In addition, we also addressed a few issues closely related to the marine stratocumulus clouds over the SEP, notably the diurnal variation of cloud cover, which are sensitive to atmospheric conditions. Descriptions of the model configuration, experiment design, and observation datasets are given in section 2. In section 3, we present the aerosol–cloud–radiation interactions over the SEP, in which the effects of cloud droplet number/size on the radiation transfer were explicitly considered. The implications of the interactions to the air–sea coupling, including SST sensitivity, atmospheric feedback, and responses of cloud properties and radiative forcing, are examined in section 4. Conclusions and discussion are given in section 5.

2. Approach

a. Model configuration

This study used the WRF Model (version 3.3) coupled with a physics-based two-moment microphysical scheme (Chen et al. 2015; Cheng et al. 2007, 2010; Hazra et al. 2013) that resolves aerosol effects on cloud properties. A brief description is given here while interested readers are referred to these papers for details. This scheme predicts both mass and number mixing ratios for five hydrometeors (cloud droplet, rain droplet, cloud ice, graupel, and snow). It includes a simple aerosol module, which explicitly calculates activation/resuspension and transport by tracking aerosol mass in hydrometeors and represents the two-way interactions between aerosols and clouds to some degree. While the aerosol concentration is simulated in domains, it is held constant at lateral boundaries of the outermost domain throughout the simulation with prescribed initial values (see section 2b for aerosol initialization), allowing aerosols to flow in and out based on local winds. At the surface, aerosols are replenished with specific rates.

The RRTMG shortwave and longwave radiation schemes (Iacono et al. 2008) were used to simulate the radiation transfer in the model simulations. The cloud droplet effective radius re, which was a diagnostic variable from the prognostic cloud droplet mass and number mixing ratios with the formula in Chen and Liu (2004), was used to account for the effect of cloud droplet size. Simulations used the Yonsei University (YSU) PBL scheme (Hong et al. 2006), although sensitivity simulations were also conducted with the University of Washington (UW) PBL scheme (Bretherton and Park 2009) and the GBM PBL scheme (Grenier and Bretherton 2001) (using the WRF Model, version 3.6.1) as well. Configurations of other physical processes (surface layer, land surface, and cumulus convection) were as in Chen et al. (2015).

A simple mixed-layer ocean scheme was included to estimate the response of SST to aerosol–stratocumulus–radiation interactions. The time change rate of SST is calculated with the following equation:
e1
where cw, ρw, α, and ε are, respectively, the heat capacity, density, shortwave reflectivity and longwave emissivity of the seawater; σ is the Stefan–Boltzmann constant; l and T are, respectively, the depth and temperature (SST) of the mixed layer; SW and LW are, respectively, the shortwave and longwave fluxes reaching the surface; and SH and LH are, respectively, the outgoing sensible and latent heat fluxes from the surface. The last term, HD, calculated with the output of a simulation with the fixed SST, is the heat divergence in the mixed layer due to horizontal and vertical transport. Note that the mixed-layer depth l is only a parameter to facilitate the air–sea coupling and does not correspond to the depth of the bulk mixed layer (e.g., Pollard et al. 1972) or the skin surface layer (e.g., Zeng and Beljaars 2005). This scheme is simple and straightforward, highlighting the effects of the atmospheric radiative forcing on the surface energy balance, but does not consider the responses of ocean currents and the cold water upwelling. These processes are much more complicated and interact with the atmosphere at much longer time scales, which are beyond the scope of this study.

b. Experimental design

Figure 1 presents the model setup of the domain location and the vertical levels. Besides the SEP region, the domain in Fig. 1a includes large part of South America in order to better simulate atmospheric dynamics near the coast and aerosol transport from the continent to the ocean. There were 171 × 138 grids horizontally with the resolution of 27 km. No nested domain was involved, but the analyses concentrated on the VOCALS region (larger dashed rectangle) and the coastal region (smaller dashed rectangle, where clouds are greatly influenced by anthropogenic aerosols). There were 63 levels vertically (solid line in Fig. 1b) with the resolution of about 40 m in the cloud and subcloud layers. This relatively high vertical resolution is to lessen the underestimation of the boundary layer depth in simulating low cloud cover (Wang et al. 2011). The influence of a lower vertical resolution (37 levels; dashed line in Fig. 1b) versus the higher resolution was also discussed.

Fig. 1.
Fig. 1.

Setups of (a) the domain location and (b) the vertical levels. One domain without further nesting was used in this study. In (a), the two dashed rectangles indicate the VOCALS region (larger one) and the coastal region (smaller one); the black line indicates the flight track of C130-RF04 during the VOCALS campaign. In (b), the solid and dashed lines indicate the 63- and 37-level vertical setups, respectively.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

The simulation period was 20–26 October 2008, a subperiod of the VOCALS campaign. The meteorology forcing was from the NCEP Final (FNL) Operational Global Analysis data (1° × 1°; National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce 2000). The SEP region was weakly disturbed by a midlatitude synoptic system around 24–26 October, and the WRF Model was shown to have relatively better performance in capturing cloud properties and the boundary layer structure for this disturbed period than for other steady periods (Rahn and Garreaud 2010b). Unlike usual WRF forecasting or hindcasting studies, this study applied the 7-day forcing for not only 1 round, but recurrently for 16 rounds to force the model to reach an equilibrium state. The unit forcing round, 7 days, involves meteorological variability within a week length (the typical length of a synoptic system), but avoids the variability at longer time scales, which demands a longer time for the model to reach the equilibrium state. Data were stored every 3 h.

Two cases—polluted and clean (including and excluding effects of anthropogenic aerosols, respectively)—were simulated. Both cases determined the aerosol size distribution and initial concentration through fitting the mean aerosol spectrum below cloud base (<500 m) observed by C130-RF04 (http://data.eol.ucar.edu/codiac/dss/id=89.115) during the VOCALS campaign with a trimodal lognormal distribution as in Chen et al. (2015). All aerosols were assumed to be ammonium sulfate. Initially, aerosols were horizontally homogeneous in the domain, vertically well mixed below 850 hPa, and decreasing exponentially above with a scale height of 800 m. In the polluted case (P for short), the aerosol replenishment from the surface was assumed to be proportional to the anthropogenic SO2 emission during October 2008 in the Monitoring Atmospheric Composition and Climate (MACC)/Megacity–Zoom for the Environment (CityZen) European Union projects (MACCity dataset in http://eccad.sedoo.fr/; Diehl et al. 2012; Granier et al. 2011; Lamarque et al. 2010; van der Werf et al. 2006) with a constant ratio for all grids. The ratio was tuned and determined by using the observed cloud droplet size and its east–west gradient as references. The final value was about 0.08. The strongest replenishing grids were mostly located on the continent and corresponded to large industries and cities in South America (figure not shown). In the clean case (C for short), no aerosol was replenished from the surface, and only the lateral boundaries provided the sources because of constant aerosol concentrations. Hence, the resulted aerosol spatial distribution reflected effects by dynamic transport and cloud processing. This case provides a reference to estimate the effects of anthropogenic aerosols. Both cases did not consider natural aerosol emissions from the surface because we are concerned with the coastal region, where anthropogenic aerosols dominate the total aerosol loading.

Each case included two simulations. In one simulation, SST was fixed to the prescribed initial value (F for short); in the other simulation, SST was simulated with the coupled mixed-layer ocean scheme (M for short). Hereafter, the four simulations are labeled with PF, CF, PM, and CM, respectively. The HD values used in PM and CM were calculated with the output from CF. As our concern was mainly the SST equilibrium response rather than its transient evolution, the mixed-layer depth l was set to 5 m in PM and CM to shorten the required time of the model to reach the equilibrium state.

c. Observational datasets

Cloud properties from the Moderate Resolution Imaging Spectroradiometer (MODIS) level-3 (L3; 1° × 1°) daily product (collection 051 in http://giovanni.gsfc.nasa.gov/giovanni; Hubanks et al. 2008) during 20–26 October 2008 were used to tune the aerosol replenishment rate in the polluted case and evaluate the model uncertainties. The MODIS instruments are carried by two sun-synchronous polar-orbiting satellites, Terra and Aqua, which cross the equator at 1030 and 1300 local time, respectively. Data were averaged between the two satellites to reduce the effect of the cloud diurnal cycle. We took liquid cloud fraction, gridded LWP and cloud (top) re from the dataset, and calculated cloud droplet number concentration Nc with cloud optical depth and re using the formula suggested by Painemal and Zuidema (2011). Because this formula assumes vertically uniform Nc and does not consider cloud adiabaticity, the Nc data derived by Min et al. (2012), which included corrections of cloud adiabaticity, were also compared with the simulated Nc.

Cloud radiative forcing (CRF) at the surface estimated by the Clouds and the Earth’s Radiant Energy System (CERES; Wielicki et al. 1996) during 20–26 October 2008 was compared with the simulated CRF. To be consistent with the simulations and MODIS observations, we mainly used the SYN1deg daily product (1° × 1°, edition 3A; http://ceres.larc.nasa.gov/products.php?product=SYN1deg). The surface fluxes are produced using the Langley Fu–Liou radiative transfer model with inputs including 3-hourly geostationary satellite (GEO) radiances, MODIS and GEO cloud properties, atmospheric profiles provided by Global Modeling and Assimilated Office (GMAO), and MODIS aerosols. The monthly CRF at the surface estimated by energy balanced and filled (EBAF) surface (1° × 1°; edition 2.8; http://ceres.larc.nasa.gov/products.php?product=EBAF-Surface; Kato et al. 2013) during October 2008 was also used as the climatology reference.

The low cloud fraction retrieved from Geostationary Operational Environmental Satellite-10 (GOES-10) channel-4 (wavelength λ = 10.7 μm) radiances (Abel et al. 2010) was used to examine the simulated cloud diurnal cycle. This dataset includes the low-cloud-cover fraction on a 0.25° latitude–longitude grid with a temporal frequency of once per 15–30 min. The 7-day mean cloud fraction from this dataset is similar to that from the MODIS L3 dataset.

3. Aerosol–stratocumulus–radiation interactions over the southeast Pacific

This part differs from our previous study (Chen et al. 2015) mainly in the method of evaluating cloud radiative forcing. To resolve the effects of droplet size on radiation transfer, our previous study employed an offline method, in which the diagnosed re was passed to an offline radiation transfer model, while this study employed the online method, in which the diagnosed re was passed to the online RRTMG schemes. The online method allows more-realistic aerosol–cloud–radiation interactions. In all simulations, the first 2 days of the first forcing round were discarded as the model spinup unless noted otherwise; the data were smoothed in all contour plots with the 25-grid (5 × 5) running-mean method to reduce the noise.

Figure 2 presents the comparisons of the low-cloud fraction, the gridded LWP, Nc, and re between the MODIS observation and the PF and CF simulations. The simulated low-cloud fraction was calculated with the simulated 3D cloud fraction below 2500 m assuming maximum/random overlapping. The simulated LWP only included cloud droplets to be consistent with the calculation of radiation transfer. As the MODIS retrieved re only represents the droplet population in the uppermost portion of a cloud, re from simulations was calculated with cloud droplet mass and number mixing ratios at the (top down) second cloud layer as in Rosenfeld et al. (2012).

Fig. 2.
Fig. 2.

Cloud properties from (a)–(d) MODIS and simulations with the fixed SST [(e)–(h) PF and (i)–(l) CF]. The first 2 days of simulations were discarded as the model spinup. Cloud droplet number concentration in (c) is calculated with the formula in Painemal and Zuidema (2011). The simulated cloud fractions in (e) and (i) were calculated by assuming maximum/random overlapping. The simulated LWP in (f) and (j) only included cloud droplets to be consistent with the calculation of radiation transfer. Cloud droplet effective radius in (h) and (l) only considered cloud droplets around cloud top to be consistent with the MODIS observation.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

The simulated cloud fraction in both PF and CF has a distribution close to the MODIS observation, but is slightly smaller, especially to the east of 75°W (Figs. 2a,e,i). The cloud cover generally decreases from the coast to the remote ocean and has three maximum regions (17°–19°S, 20°–25°S, and 30°–32°S) near the coast and one minimum region at the lower-right corner of the VOCALS region. The simulated maximum cloud cover is shifted southward compared with the observed. Between the two simulations, no significant difference is shown.

The simulated gridded LWP shows northward shifted peaks and great overestimations over the whole region compared with the MODIS observation (Figs. 2b,f,j). The observed gridded LWP is about 50 g m−2 near the coast and increases westward to about 100 g m−2 over the remote ocean, while the simulated LWP is about 150 g m−2 near the coast and increases westward to more than 250 g m−2. As the MODIS observations already overestimate the gridded LWP (Min et al. 2012), the overestimation by the model is more than 100 g m−2 for the whole region. Besides, the cloud geometric thickness is overestimated as well. The difference between PF and CF is minor, even for the coastal region, indicating clouds hardly precipitate there in both simulations. This is further explained below.

The simulated Nc in PF has a distribution comparable with the MODIS observation but has obvious overestimations for both the coastal region and the remote ocean (Figs. 2c,g). The observed Nc decreases toward the southwest from about 220 cm−3 near the coast to less than 50 cm−3 over the remote ocean, while the simulated Nc in PF is about 300 cm−3 near the coast and about 80 cm−3 over the remote ocean. The Nc values derived by Min et al. (2012), which included corrections of cloud adiabaticity to the derivation formula, are slightly larger than those shown in Fig. 2c, but still smaller than the simulated by PF. The Nc overestimation by the model is caused by the strong aerosol emissions from the surface (mainly affecting the coastal region) and the lateral boundaries (mainly affecting the remote ocean). In tuning, the surface aerosol emission was constrained by the cloud droplet size, not the droplet number, so in fact the overestimation in Nc benefits the simulation of re near the coast, as shown below. In contrast, the CF simulation does not show obvious east–west gradient in Nc (Fig. 2k). It differs from PF mainly to the east of 80°W (especially near the coast) owing to the lack of anthropogenic aerosols from the continent, and is almost the same as PF to the west of 80°W, confirming that the aerosol replenishment from lateral boundaries was too strong in the model.

The simulated re in PF also has a distribution comparable with the MODIS observation (Figs. 2d,h). The cloud-top re has its minimum of about 6 μm at the coastal region around 20°S and increases toward the southwest to more than 16 μm over the remote ocean. To the east of 80°W, the east–west gradients of re in PF are similar to those shown by the MODIS observation; the overestimation in Nc by PF offsets the overestimation in LWP so that the model can produce droplet sizes close to the observed. To the west of 80°W, the simulated re clearly has smoother east–west gradients than the MODIS observed re; the area with re greater than 16 μm in the simulation is also smaller, which further confirms that the model overestimates Nc over the remote ocean. The CF simulation also shows increasing re from the coast to the remote ocean (Fig. 2l), but the gradient is much smoother near the coast, mainly caused by the gradient in LWP. It is noted that the cloud-top re represents a very small portion of cloud droplets in the column because the model overestimates the cloud vertical geometric thickness. Therefore, the re value of about12 μm near the coast in CF does not ensure efficient water removal by precipitation as expected. This leads to the similar LWP in CF and PF as shown above.

As the MODIS observation tends to underestimate the stratocumulus clouds, which usually peak in the early morning, the simulated cloud fraction along the 20°S was also compared with the GOES-10 observation, and the simulated cloud-mean LWP, Nc, and columnar re were compared with the C130 aircraft observations during the VOCALS field campaign (figure not shown). The results remain the same. Overall, the model slightly underestimates the cloud fraction but significantly overestimates the cloud-mean LWP; anthropogenic aerosols mainly affect cloud microproperties (Nc and re), but have little effect on cloud macroproperties (cloud cover and LWP), which are most likely dominated by the large-scale meteorology. This is consistent with our previous results (Chen et al. 2015).

Figure 3 presents the surface CRF from the CERES SYN1deg data and the PF and CF simulations. The shortwave and longwave cloud radiative forcings (SWCRF and LWCRF) in both simulations (only PF shown in the figure) show similar patterns to the SYN1deg data. For example, the SWCRF has peak values over three regions coinciding with the regions with the maximum cloud cover. However, the model clearly overestimates both SWCRF and LWCRF, mainly caused by the overestimation of cloud thickness (LWP). Nevertheless, uncertainties of the SYN1deg data were also noted: the monthly mean surface CRF of October 2008 calculated with the SYN1deg data is weaker than that from the EBAF-surface data (figure not shown). The difference between PF and CF is limited to the east of 80°W and the largest to the east of 75°W (Figs. 3e,f), which is consistent with the difference in cloud droplet number and size shown above. Meanwhile, the difference is significant in the shortwave band (as much as 20 W m−2) but ignorable in the longwave band (less than 1 W m−2). This confirms that anthropogenic aerosols from the continent induce significant increase in SWCRF via aerosol–cloud–radiation interactions but have little effect on LWCRF.

Fig. 3.
Fig. 3.

Cloud radiative forcing at the surface from (a),(b) CERES SYN1deg and (c),(d) simulations with the fixed SST. (e),(f) Different scales are used for the differences in SWCRF and LWCRF between PF and CF (PF minus CF).

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

4. Implications to the air–sea coupling

a. SST sensitivity

The HD values in Eq. (1) were calculated for each ocean grid with the simulated heat budget at the surface from the CF simulation. Assuming the LHS to be zero, all terms except HD in the RHS were set to the temporal means for the whole 16 forcing rounds, so the resulted HD had only one time slice and kept constant throughout the simulation in PM and CM. The HD data have positive values near the coast and negative values over the remote ocean (figure not shown), consistent with the climatological strong upwelling of cold water near the coast. As the model biases in cloud fraction and thickness are similar in the polluted and clean cases, their effects are considered to be absorbed into HD values, so the difference in SST between PM and CM can only be attributed to their differences in cloud droplet number and size.

As the HD data in the mixed-layer ocean kept constant, the simulated SST has a simple diurnal cycle (warming during the day and cooling during the night) due to the solar diurnal cycle. This tends to modify the surface heat budget, the boundary layer development, cloud properties, and CRF, which would further affect SST. Therefore, the simulated SST tends to drift from the initial value (shown below) that also corresponds to the prescribed SST used in PF and CF, although none of other model configurations is changed.

Figure 4 presents the initial SST, the simulated final SSTs in PM and CM, and the difference between the latter two. A remarkable feature in the initial SST is the warm water patch near 20°S, 70°W, which was also shown in Rahn and Garreaud (2010a). We checked the monthly SST in the ERA-Interim during 2004–10 (ds627.1; European Centre for Medium-Range Forecasts 2012) and found that the feature is seasonal and usually prevails from October to May. The simulated final SSTs in PM and CM are higher than the initial SST for most area but show similar patterns to the initial SST (i.e., increasing from the coast to the remote ocean). This warm water patch is still apparent in the final SST of CM but much weaker in that of PM. The difference between PM and CM mainly occurs near the coast, where PM has a significant cooling of as much as 2°C. This is consistent with the shortwave reduction at the surface in PF as shown in Fig. 3c.

Fig. 4.
Fig. 4.

(a) The initial SST in PM and CM, the simulated final SST in (b) PM and (c) CM, and (d) the difference between the latter two (PM minus CM). The SST near the coast is reduced by as much as 2°C owing to the CRF increase caused by anthropogenic aerosols.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

For more in-depth analyses, we concentrate on the coastal region (smaller dashed rectangle in Fig. 1a). The time series of the mean SST for this region is shown in Fig. 5. Both the final SST and the mean SST of each forcing round reach the equilibrium state at the eighth forcing round. Within each round, the mean SST is slightly lower than the final SST, but their difference is stable when in equilibrium. The equilibrium final SST is about 0.6°C higher in CM and 0.8°C lower in PM than the initial SST, so the average cooling caused by the shortwave reduction is about 1.4°C for the coastal region.

Fig. 5.
Fig. 5.

Temporal evolution of mean SST for the coastal region (smaller dashed rectangle in Fig. 1) from PM (black) and CM (gray). The thin lines include the initial SST and the final SSTs of each forcing round, while the thick lines include the mean SSTs within each forcing round.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

Figure 6 shows the correlation between the SST cooling and the CRF increase at the surface from all the grids in the coastal region. The SST cooling generally has a positive correlation with the CRF increase with a correlation coefficient of about 0.6. On average, the sensitivity of SST cooling to the CRF increase is 0.1°C (W m−2)−1—about half the strength [0.19°C (W m−2)−1] by the direct radiation cooling for this region. Note that larger variances are shown in the surface cooling for grids with larger CRF increase, implying that the feedback intensity of the air–sea coupling is sensitive to local meteorological conditions. It is also interesting to mention that the 0.1°C (W m−2)−1 sensitivity with the prescribed HD here is 50%–100% smaller than the values [0.15°–0.2°C (W m−2)−1] of the North Atlantic SST to the aerosol-caused surface SW reduction shown by the HadGEM2 results [Figs. 3c and 3d in Booth et al. (2012)], in which responses of the ocean currents were also included.

Fig. 6.
Fig. 6.

Correlation between the surface cooling and the CRF increase caused by anthropogenic aerosols for the coastal region. The solid line represents the linear regression result (ΔSST = 0.0975ΔCRF + 0.0514) with the ordinary least squares method. The correlation coefficient is 0.594.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

b. Atmospheric feedback

The atmospheric feedback to the aerosol–cloud–radiation interactions was examined in two aspects: the surface heat budget and the boundary layer structure. Our analyses concentrate on the coastal region using results from the last four forcing rounds of all four simulations.

Table 1 presents changes of the surface heat budget between the polluted and clean cases (polluted minus clean). When SST is fixed, aerosols mainly affect the downward shortwave radiation (SW↓), which is consistent with the stronger SWCRF in PF as shown in Fig. 3. SW↓ is decreased by 19.1 W m−2 owing to the increased cloud albedo while the downward longwave radiation (LW↓) is increased by 1 W m−2 owing to the slightly increased cloud fraction and thickness (Table 2); both sensible (SH↑) and latent heat (LH↑) fluxes are slightly perturbed by changes in the near-surface atmospheric conditions; the upward longwave (LW↑ = εσT4) is identical because of the same prescribed SST. When the mixed-layer ocean is coupled, different responses in the surface energy balance are simulated. The LH↑, SH↑, and LW↑ fluxes are all decreased, imposing negative feedback to the SST changes that directly respond to the decrease of SW↓. The feedback in LH↑ offsets large part of the SW reduction caused by the increase of cloud albedo. However, the most interesting finding is the further decrease in both SW↓ (from −19.1 to −24 W m−2) and LW↓ (from 1 to −3.3 W m−2) reaching the surface. The former is caused by the increase of cloud cover (especially the daytime cloud cover, shown in Table 2), while the latter is caused by the decrease of the cloud-base temperature (Fig. 7 and Tcbase in Table 2).

Table 1.

Changes of the surface heat budget (polluted minus clean; W m−2) as a result of anthropogenic aerosols for the coastal region (18°–22°S, 71°–75°W), averaged over the last four forcing rounds (28 days). The mean SSTs during this period are 17.6°, 17.6°, 16.8°, and 18.2°C for PF, CF, PM, and CM, respectively.

Table 1.
Table 2.

Cloud properties and radiative forcing at the surface for the coastal region (18°–22°S, 71°–75°W), averaged over the last four forcing rounds (28 days). CldFrac and DCldFrac are acronyms for the cloud fraction and the effective daytime cloud fraction, respectively. The definition of DCldFrac is given by Eq. (2). Based on the retrieval from the GOES-10 channel-4 radiances (Abel et al. 2010), the mean CldFrac over this region is 81.5% during 20–26 Oct 2008, while the mean DCldFrac is 74.7%. Cloud top and base are defined, respectively, by the uppermost and the lowermost layers with the single-layer LWP greater than 0.003 g m−2, which is the same as in RRTMG schemes.

Table 2.
Fig. 7.
Fig. 7.

Vertical profiles of (a) potential temperature, (b) water vapor mixing ratio, (c) cloud water mixing ratio, and (d) horizontal wind speed over the coastal region, averaged over the last four forcing rounds (28 days). The dashed lines indicate results from PF (black) and CF (gray), while the solid lines indicate results from PM (black) and CM (gray).

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

Figure 7 presents the mean vertical profiles of temperature, water vapor mixing ratio, cloud water mixing ratio, and horizontal wind speeds of four simulations. The differences between PF and CF are minor, while the differences between PM and CM are significant in most parameters. Because of less SH↑ and LH↑ outgoing from the surface, the boundary layer in PM tends to be cooler and have less water vapor than that in CM, but the relative humidity in PM is higher, especially in the subcloud layer. This leads to a lower cloud height and a larger cloud cover in PM (shown below). Recalling that LW↓ is decreased as shown above, it is inferred that the effect of the boundary layer cooling in PM overwhelms the effects of the cloud base descending and the cloud fraction increasing in modulating LW↓. The wind profiles are not affected much except that the easterly wind is slightly stronger in PM. This tends to enhance the aerosol transport from the continent to the stratocumulus over the ocean, which further affects the offshore aerosol–cloud interactions.

c. Responses of cloud properties and radiative forcing

Table 2 presents the mean cloud properties and radiative forcing at the surface for all four simulations, which provides more details on how anthropogenic aerosols affect aerosol–cloud–radiation interactions in the air–sea coupling process. As shown in section 3, when SST is fixed, aerosols mainly affect cloud microproperties (cloud droplet number and effective radius) but have little effect on cloud macroproperties (cloud fraction and LWP). Therefore, compared with CF, PF has stronger SWCRF owing to the smaller droplet size and similar LWCRF owing to the similar cloud cover, LWP, and top and base heights (temperatures). When the mixed-layer ocean is coupled, however, aerosols affect both cloud macro- and microproperties. First, the cloud fraction is increased by 3.7% (PM minus CM; versus the 1.1% of PF minus CF) by the higher aerosol concentrations. As the cloud-mean LWP (gridded LWP/cloud fraction) is larger in PM than in CM, this further increases the gridded LWP in PM and tends to increase both SWCRF and LWCRF. Second, as the whole boundary layer is cooler in PM (shown in Fig. 7), the cloud tops and bases in PM are about 0.7°C cooler than those in CM. Nevertheless, the clear-sky longwave radiation is also decreased owing to the lower temperature, so LWCRF is not significantly affected by changes of the cloud temperature. The responses of cloud microproperties (droplet number and effective radius) to aerosols are similar as in PF and CF. Overall, both SWCRF and LWCRF at the surface are apparently increased by the higher aerosol concentrations when the SST response is included.

As the stratocumulus shows significant diurnal variations (forming during the nighttime and dissipating during the daytime; e.g., Burleyson et al. 2013), we further examine responses of the diurnal cycle of cloud fraction and LWP to aerosols. Generally, the polluted case has larger cloud fraction than the clean case all the time (Fig. 8a; solid black versus solid gray, dashed black versus dashed gray). The difference between PF and CF is nearly uniform throughout the day, while the difference between PM and CM is clearly larger during the daytime than during the nighttime, indicating that the aerosol increase reduces the diurnal variation of cloud fraction. It is because the surface cooling caused by the SWCRF increase prevails during the daytime as a result of the solar diurnal cycle, which subsequently leads to the increase of the boundary layer relative humidity and cloud fraction and further increases the SWCRF in return.

Fig. 8.
Fig. 8.

Diurnal cycle of (a) cloud fraction and (b) LWP over the coastal region, averaged over the last four forcing rounds (28 days). The dashed lines indicate results from PF (black) and CF (gray), while the solid lines indicate results from PM (black) and CM (gray). The local daytime is around 1200–2300 UTC.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

To quantify this effect, we defined an effective daytime cloud fraction (DCldFrac) as
e2
where CldFrac is the instant cloud fraction, ϕ is the grid longitude, ϕsun is the noontime longitude, and nt is the number of data record for the grid. In this way, DCldFrac represents the mean daytime cloud fraction weighted with the solar diurnal cycle. When cloud fraction changes uniformly throughout the day, DCldFrac changes by the same amount as CldFrac; when the daytime cloud fraction changes more (less) than the nighttime cloud fraction, DCldFrac changes more (less) than CldFrac.

Shown in Table 2, the difference in DCldFrac between PM and CM is larger than that in CldFrac, indicating that the increase of daytime cloud fraction is larger than that of nighttime cloud fraction. This is consistent with the result shown in Fig. 8a. Meanwhile, the CldFrac and DCldFrac calculated with the retrieval from GOES-10 channel-4 radiances (Abel et al. 2010) are 81.5% and 74.7%, respectively. This implies that the model underestimates the daytime cloud fraction and overestimates the nighttime cloud fraction, although it predicts a daily mean cloud fraction close to the observed. The similar bias might also exist in general circulation models because the models were shown to have overestimated both cloud fraction and LWP of low clouds but underestimated SWCRF during January–May [Fig. 5 in Rapp (2015)]. Note that uncertainties also exist in the GOES-10 estimation of cloud fraction. The instruments cannot see clouds that are not optically thick and have difficulties when the cloud-top temperatures are close to the surface temperatures. In addition, the cloud-mean SWCRF is better defined as SWCRF/DCldFrac instead of SWCRF/CldFrac. The former definition gives 1.97, 1.75, 1.98, and 1.75 for PF, CF, PM, and CM, while the latter gives 1.63, 1.45, 1.65, and 1.43. Obviously the former definition better represents the similarity in cloud microproperties from simulations with the same aerosol condition and highlights the important role of the response of cloud fraction to aerosols.

The diurnal cycle of LWP is less affected by using fixed or simulated SST (Fig. 8b). During the nighttime, the polluted case has slightly larger LWP than the clean case, mainly caused by the slightly larger cloud fraction in the former. During the daytime, the difference is minor. However, recalling that the daytime cloud fraction in PM is larger than that in CM, PM therefore has smaller daytime cloud thickness (lower cloud-mean LWP) than CM. In this sense, the aerosol increase could increase the cloud fraction and decrease the cloud thickness (cloud-mean LWP) during the daytime when SST and the relevant air–sea coupling are allowed to respond to aerosol–cloud–radiation interactions.

5. Conclusions and discussion

This study investigated the responses of the air–sea coupling over the SEP to the surface SW reduction caused by increased cloud albedo that is associated with increasing anthropogenic aerosols. The coupling was analyzed by examining the equilibrium responses of the surface energy balance components—LH and SH fluxes and SW and LW radiation simulated using the WRF Model coupled with a simple mixed-layer ocean module. The SST sensitivity was then derived and discussed within the context of SST biases in CMIP5 models. Highlights of the findings are given here.

Anthropogenic aerosols emitted from South America substantially reduce the solar radiation reaching the sea surface over the coastal region through increasing cloud radiative forcing by as much as 20 W m−2. Both negative and positive feedbacks are involved when SST responds to the surface SW reduction. The negative feedback stems from decreased LH and SH fluxes from the surface as a result of surface cooling; the LH decrease offsets most of the SW reduction. The positive feedback comes from an increase in cloud fraction (which further decreases the SW reaching the surface) caused by increased relative humidity in the boundary layer. The increase of cloud fraction mainly occurs during the daytime, when clouds shield the boundary layer and the surface from the solar warming and help to form a cooler, moister (larger relative humidity), and more stable boundary layer. An effective daytime cloud fraction (DCldFrac) to quantify this effect is proposed, which can also serve as a good diagnostic parameter to compare the cloud–radiation interactions among climate models. In the end, the model calculates a 0.1°C (W m−2)−1 sensitivity of the SST cooling to the SW reduction.

There are several relevant issues worthy of comments. First, the recurrent forcing method well retained the cloud evolution with time. Figure 9 shows the time series of cloud fraction over the coastal region in PF. Generally, the simulated cloud fraction has significant diurnal variations in the first forcing round (PF1 for short; thin dashed line in Fig. 9) and the mean of the other 15 forcing rounds (PF15; thick dashed line) and well reproduces the observed synoptic disturbance (open circles) around 24–26 October (days 4–6 in the plot) in both periods.

Fig. 9.
Fig. 9.

Time series of the low-cloud fraction over the coastal region from PF during the first forcing round (thin) and the last 15 forcing rounds (thick). Open circles represent the GOES-10 observation.

Citation: Journal of the Atmospheric Sciences 73, 7; 10.1175/JAS-D-15-0277.1

Second, the effect of vertical resolution was examined because the 27-km horizontal resolution used in the present study is much larger than the 3-km resolution in Chen et al. (2015). Nevertheless, similar results are obtained: lower vertical resolution leads to a lower cloud fraction, a higher LWP, and a thinner boundary layer depth. Moreover, some clouds were shown to form at the first level above the surface when using low resolutions in both horizontal and vertical dimensions, which was not discussed in Chen et al. (2015). It suggests that higher vertical resolution is needed to simulate stratocumulus clouds when the horizontal resolution is relatively coarse (e.g., in general circulation models).

Finally, as many previous studies have already pointed out, the model biases of the underestimation of cloud fraction and the overestimation of cloud thickness (LWP) still persist. We believe that these biases are related to the PBL scheme. The boundary layer top predicted by the YSU scheme is closer to the cloud base instead of the cloud top (Andrejczuk et al. 2012), causing weaker mixing in the cloud layer. This not only suppresses the cloud-top evaporation and thus the elevation of the boundary layer top but also contributes to the excessive cloud water content. In addition, the cloud fraction and thickness are shown to be sensitive to PBL schemes: the UW and the GBM schemes simulate smaller cloud fraction and LWP than the YSU scheme. Further improvement of the PBL scheme seems to be the key to reduce these biases.

Acknowledgments

We thank Dr. Steven Abel and Lanxi Min, respectively, for providing the GOES-10 low-cloud fraction and the revised MODIS cloud droplet number concentration. We acknowledge the two anonymous reviewers whose comments and suggestions have clarified the presentation. The research is supported by a grant from the Office of Science (BER), U.S. DOE.

REFERENCES

  • Abel, S. J., , D. N. Walters, , and G. Allen, 2010: Evaluation of stratocumulus cloud prediction in the Met Office forecast model during VOCALS-REx. Atmos. Chem. Phys., 10, 10 54110 559, doi:10.5194/acp-10-10541-2010.

    • Search Google Scholar
    • Export Citation
  • Andrejczuk, M., , W. W. Grabowski, , A. Gadian, , and R. Burton, 2012: Limited-area modelling of stratocumulus over South-Eastern Pacific. Atmos. Chem. Phys., 12, 35113526, doi:10.5194/acp-12-3511-2012.

    • Search Google Scholar
    • Export Citation
  • Boo, K.-O., , B. B. B. Booth, , Y.-H. Byun, , J. Lee, , C. Cho, , S. Shim, , and K.-T. Kim, 2015: Influence of aerosols in multidecadal SST variability simulations over the North Pacific. J. Geophys. Res. Atmos., 120, 517531, doi:10.1002/2014JD021933.

    • Search Google Scholar
    • Export Citation
  • Booth, B. B. B., , N. J. Dunstone, , P. R. Halloran, , T. Andrews, , and N. Bellouin, 2012: Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability. Nature, 484, 228232, doi:10.1038/nature10946.

    • 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
  • Burleyson, C. D., , S. P. de Szoeke, , S. E. Yuter, , M. Wilbanks, , and W. A. Brewer, 2013: Ship-based observations of the diurnal cycle of Southeast Pacific marine stratocumulus clouds and precipitation. J. Atmos. Sci., 70, 38763894, doi:10.1175/JAS-D-13-01.1.

    • Search Google Scholar
    • Export Citation
  • Calisto, M., , D. Folini, , M. Wild, , and L. Bengtsson, 2014: Cloud radiative forcing intercomparison between fully coupled CMIP5 models and CERES satellite data. Ann. Geophys., 32, 793807, doi:10.5194/angeo-32-793-2014.

    • Search Google Scholar
    • Export Citation
  • Chen, G., , W.-C. Wang, , and J.-P. Chen, 2015: Aerosol–stratocumulus–radiation interactions over the southeast Pacific. J. Atmos. Sci., 72, 26122621, doi:10.1175/JAS-D-14-0319.1.

    • Search Google Scholar
    • Export Citation
  • Chen, J. P., , and S. T. Liu, 2004: Physically based two-moment bulkwater parametrization for warm-cloud microphysics. Quart. J. Roy. Meteor. Soc., 130, 5178, doi:10.1256/qj.03.41.

    • Search Google Scholar
    • Export Citation
  • Cheng, C.-T., , W.-C. Wang, , and J.-P. Chen, 2007: A modelling study of aerosol impacts on cloud microphysics and radiative properties. Quart. J. Roy. Meteor. Soc., 133, 283297, doi:10.1002/qj.25.

    • Search Google Scholar
    • Export Citation
  • Cheng, C.-T., , W.-C. Wang, , and J.-P. Chen, 2010: Simulation of the effects of increasing cloud condensation nuclei on mixed-phase clouds and precipitation of a front system. Atmos. Res., 96, 461476, doi:10.1016/j.atmosres.2010.02.005.

    • Search Google Scholar
    • Export Citation
  • Diehl, T., , A. Heil, , M. Chin, , X. Pan, , D. Streets, , M. Schultz, , and S. Kinne, 2012: Anthropogenic, biomass burning, and volcanic emissions of black carbon, organic carbon, and SO2 from 1980 to 2010 for hindcast model experiments. Atmos. Chem. Phys. Discuss., 12, 24 89524 954, doi:10.5194/acpd-12-24895-2012.

    • Search Google Scholar
    • Export Citation
  • Ekman, A. M. L., 2014: Do sophisticated parameterizations of aerosol-cloud interactions in CMIP5 models improve the representation of recent observed temperature trends? J. Geophys. Res. Atmos., 119, 817832, doi:10.1002/2013JD020511.

    • Search Google Scholar
    • Export Citation
  • European Centre for Medium-Range Weather Forecasts, 2012: ERA-Interim project, monthly means. National Center for Atmospheric Research Computational and Information Systems Laboratory Research Data Archive, accessed 6 March 2015, doi:10.5065/D68050NT.

  • Flato, G., and et al. , 2013: Evaluation of climate models. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 741–866.

  • Granier, C., and et al. , 2011: Evolution of anthropogenic and biomass burning emissions of air pollutants at global and regional scales during the 1980–2010 period. Climatic Change, 109, 163190, doi:10.1007/s10584-011-0154-1.

    • Search Google Scholar
    • Export Citation
  • Grenier, H., , and C. S. Bretherton, 2001: A moist PBL parameterization for large-scale models and its application to subtropical cloud-topped marine boundary layers. Mon. Wea. Rev., 129, 357377, doi:10.1175/1520-0493(2001)129<0357:AMPPFL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hazra, A., , V. Mandal, , and J.-P. Chen, 2013: Study of cloud microphysical properties over India during CAIPEEX using a mesoscale model with new cloud microphysical scheme—Part I. J. Atmos. Sol.-Terr. Phys., 93, 2944, doi:10.1016/j.jastp.2012.11.010.

    • Search Google Scholar
    • Export Citation
  • Hong, S.-Y., , Y. Noh, , and J. Dudhia, 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 23182341, doi:10.1175/MWR3199.1.

    • Search Google Scholar
    • Export Citation
  • Hubanks, P. A., , M. D. King, , S. A. Platnick, , and R. A. Pincus, 2008: MODIS Atmosphere L3 gridded product. MODIS Algorithm Theoretical Basis Doc. ATBD-MOD-30, 96 pp. [Available online at http://modis-atmos.gsfc.nasa.gov/_docs/L3_ATBD_2008_12_04.pdf.]

  • 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
  • Kato, S., , N. G. Loeb, , F. G. Rose, , D. R. Doelling, , D. A. Rutan, , T. E. Caldwell, , L. S. Yu, , and R. A. Weller, 2013: Surface irradiances consistent with CERES-derived top-of-atmosphere shortwave and longwave irradiances. J. Climate, 26, 27192740, doi:10.1175/JCLI-D-12-00436.1.

    • Search Google Scholar
    • Export Citation
  • Kubar, T. L., , D. E. Waliser, , J. L. Li, , and X. Jiang, 2012: On the annual cycle, variability, and correlations of oceanic low-topped clouds with large-scale circulation using Aqua MODIS and ERA-Interim. J. Climate, 25, 61526174, doi:10.1175/JCLI-D-11-00478.1.

    • Search Google Scholar
    • Export Citation
  • Lamarque, J. F., and et al. , 2010: Historical (1850–2000) gridded anthropogenic and biomass burning emissions of reactive gases and aerosols: Methodology and application. Atmos. Chem. Phys., 10, 70177039, doi:10.5194/acp-10-7017-2010.

    • Search Google Scholar
    • Export Citation
  • Large, W. G., , and G. Danabasoglu, 2006: Attribution and impacts of upper-ocean biases in CCSM3. J. Climate, 19, 23252346, doi:10.1175/JCLI3740.1.

    • Search Google Scholar
    • Export Citation
  • Min, Q., and et al. , 2012: Comparison of MODIS cloud microphysical properties with in-situ measurements over the Southeast Pacific. Atmos. Chem. Phys., 12, 11 26111 273, doi:10.5194/acp-12-11261-2012.

    • Search Google Scholar
    • Export Citation
  • Nam, C., , S. Bony, , J. L. Dufresne, , and H. Chepfer, 2012: The ‘too few, too bright’ tropical low-cloud problem in CMIP5 models. Geophys. Res. Lett., 39, L21801, doi:10.1029/2012GL053421.

    • Search Google Scholar
    • Export Citation
  • National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce, 2000: NCEP FNL Operational Model Global Tropospheric Analyses, continuing from July 1999 (updated daily). National Center for Atmospheric Research Computational and Information Systems Laboratory Research Data Archive, accessed 16 December 2013. [Available online at http://rda.ucar.edu/datasets/ds083.2/.]

  • Painemal, D., , and P. Zuidema, 2011: Assessment of MODIS cloud effective radius and optical thickness retrievals over the Southeast Pacific with VOCALS-REx in situ measurements. J. Geophys. Res., 116, D24206, doi:10.1029/2011JD016155.

    • Search Google Scholar
    • Export Citation
  • Park, S., , C. Deser, , and M. A. Alexander, 2005: Estimation of the surface heat flux response to sea surface temperature anomalies over the global oceans. J. Climate, 18, 45824599, doi:10.1175/JCLI3521.1.

    • Search Google Scholar
    • Export Citation
  • Pollard, R. T., , P. B. Rhines, , and R. O. R. Y. Thompson, 1972: The deepening of the wind—Mixed layer. Geophys. Fluid Dyn., 4, 381404, doi:10.1080/03091927208236105.

    • Search Google Scholar
    • Export Citation
  • Rahn, D. A., , and R. Garreaud, 2010a: Marine boundary layer over the subtropical southeast Pacific during VOCALS-REx—Part 1: Mean structure and diurnal cycle. Atmos. Chem. Phys., 10, 44914506, doi:10.5194/acp-10-4491-2010.

    • Search Google Scholar
    • Export Citation
  • Rahn, D. A., , and R. Garreaud, 2010b: Marine boundary layer over the subtropical southeast Pacific during VOCALS-REx—Part 2: Synoptic variability. Atmos. Chem. Phys., 10, 45074519, doi:10.5194/acp-10-4507-2010.

    • Search Google Scholar
    • Export Citation
  • Rapp, A. D., 2015: Cloud responses in the AMIP simulations of CMIP5 models in the southeastern Pacific marine subsidence region. Int. J. Climatol., 35, 29082921, doi:10.1002/joc.4181.

    • Search Google Scholar
    • Export Citation
  • Richter, I., 2015: Climate model biases in the eastern tropical oceans: Causes, impacts and ways forward. Wiley Interdiscip. Rev.: Climate Change, 6, 345358, doi:10.1002/wcc.338.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., , H. L. Wang, , and P. J. Rasch, 2012: The roles of cloud drop effective radius and LWP in determining rain properties in marine stratocumulus. Geophys. Res. Lett., 39, L13801, doi:10.1029/2012GL052028.

    • Search Google Scholar
    • Export Citation
  • van der Werf, G. R., , J. T. Randerson, , L. Giglio, , G. J. Collatz, , P. S. Kasibhatla, , and A. F. Arellano, 2006: Interannual variability in global biomass burning emissions from 1997 to 2004. Atmos. Chem. Phys., 6, 34233441, doi:10.5194/acp-6-3423-2006.

    • Search Google Scholar
    • Export Citation
  • Voldoire, A., , M. Claudon, , G. Caniaux, , H. Giordani, , and R. Roehrig, 2014: Are atmospheric biases responsible for the tropical Atlantic SST biases in the CNRM-CM5 coupled model? Climate Dyn., 43, 29632984, doi:10.1007/s00382-013-2036-x.

    • Search Google Scholar
    • Export Citation
  • Wang, C. Z., , L. P. Zhang, , S. K. Lee, , L. X. Wu, , and C. R. Mechoso, 2014: A global perspective on CMIP5 climate model biases. Nat. Climate Change, 4, 201205, doi:10.1038/nclimate2118.

    • Search Google Scholar
    • Export Citation
  • Wang, S., , L. W. O’Neill, , Q. Jiang, , S. P. de Szoeke, , X. Hong, , H. Jin, , W. T. Thompson, , and X. Zheng, 2011: A regional real-time forecast of marine boundary layers during VOCALS-REx. Atmos. Chem. Phys., 11, 421437, doi:10.5194/acp-11-421-2011.

    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., , B. R. Barkstrom, , E. F. Harrison, , R. B. Lee, , G. Louis Smith, , and J. E. Cooper, 1996: Clouds and the Earth’s Radiant Energy System (CERES): An Earth Observing System experiment. Bull. Amer. Meteor. Soc., 77, 853868, doi:10.1175/1520-0477(1996)077<0853:CATERE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wilcox, L. J., , E. J. Highwood, , and N. J. Dunstone, 2013: The influence of anthropogenic aerosol on multi-decadal variations of historical global climate. Environ. Res. Lett., 8, 024033, doi:10.1088/1748-9326/8/2/024033.

    • Search Google Scholar
    • Export Citation
  • Wood, R., and et al. , 2011: The VAMOS Ocean-Cloud-Atmosphere-Land Study Regional Experiment (VOCALS-REx): Goals, platforms, and field operations. Atmos. Chem. Phys., 11, 627654, doi:10.5194/acp-11-627-2011.

    • Search Google Scholar
    • Export Citation
  • Zeng, X. B., , and A. Beljaars, 2005: A prognostic scheme of sea surface skin temperature for modeling and data assimilation. Geophys. Res. Lett., 32, L14605, doi:10.1029/2005GL023030.

    • Search Google Scholar
    • Export Citation
  • Zheng, Y., , G. N. Kiladis, , T. Shinoda, , E. J. Metzger, , H. E. Hurlburt, , J. Lin, , and B. S. Giese, 2010: Upper-ocean processes under the stratus cloud deck in the southeast Pacific Ocean. J. Phys. Oceanogr., 40, 103120, doi:10.1175/2009JPO4213.1.

    • Search Google Scholar
    • Export Citation
  • Zheng, Y., , T. Shinoda, , J.-L. Lin, , and G. N. Kiladis, 2011: Sea surface temperature biases under the stratus cloud deck in the southeast Pacific Ocean in 19 IPCC AR4 coupled general circulation models. J. Climate, 24, 41394164, doi:10.1175/2011JCLI4172.1.

    • Search Google Scholar
    • Export Citation
Save