• Bellenger, H., , E. Guilyardi, , J. Leloup, , M. Lengaigne, , and J. Vialard, 2014: ENSO representation in climate models: From CMIP3 to CMIP5. Climate Dyn., 42, 1999–2018, doi:10.1007/s00382-013-1783-z.

    • Search Google Scholar
    • Export Citation
  • Boyle, J., , and S. A. Klein, 2010: Impact of horizontal resolution on climate model forecasts of tropical precipitation and diabatic heating for the TWP-ICE period. J. Geophys. Res., 115, D23113, doi:10.1029/2010JD014262.

    • Search Google Scholar
    • Export Citation
  • Brient, F., , and S. Bony, 2012: How may low-cloud radiative properties simulated in the current climate influence low-cloud feedbacks under global warming? Geophys. Res. Lett., 39, L20807, doi:10.1029/2012GL053265.

    • Search Google Scholar
    • Export Citation
  • Chen, L., , Y. Yu, , and D.-Z. Sun, 2013: Cloud and water vapor feedbacks to the El Niño warming: Are they still biased in CMIP5 models? J. Climate, 26, 29472961, doi:10.1175/JCLI-D-12-00575.1.

    • Search Google Scholar
    • Export Citation
  • Cole, J., , H. Barker, , N. Loeb, , and K. von Salzen, 2011: Assessing simulated clouds and radiative fluxes using properties of clouds whose tops are exposed to space. J. Climate, 24, 27152727, doi:10.1175/2011JCLI3652.1.

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

  • Dai, A., , and K. E. Trenberth, 2004: The diurnal cycle and its depiction in the Community Climate System Model. J. Climate, 17, 930951, doi:10.1175/1520-0442(2004)017<0930:TDCAID>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., , P. Braconnot, , F. F. Jin, , S. T. Kim, , M. Kolasinski, , T. Li, , and I. Musat, 2009a: Atmosphere feedbacks during ENSO in a coupled GCM with a modified atmospheric convection scheme. J. Climate, 22, 56985718, doi:10.1175/2009JCLI2815.1.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., , A. Wittenberg, , A. Fedorov, , M. Collins, , C. Wang, , A. Capotondi, , G. J. van Oldenborgh, , and T. Stockdale, 2009b: Understanding El Niño in ocean–atmosphere general circulation models: Progress and challenges. Bull. Amer. Meteor. Soc., 90, 325340, doi:10.1175/2008BAMS2387.1.

    • Search Google Scholar
    • Export Citation
  • Ham, Y.-G., , and J.-S. Kug, 2014: Effects of Pacific Intertropical Convergence Zone precipitation bias on ENSO phase transition. Environ. Res. Lett., 9, 064008, doi:10.1088/1748-9326/9/6/064008.

    • Search Google Scholar
    • Export Citation
  • Kim, D., , J.-S. Kug, , I.-S. Kang, , F.-F. Jin, , and A. T. Wittenberg, 2008: Tropical Pacific impacts of convective momentum transport in the SNU coupled GCM. Climate Dyn., 31, 213226, doi:10.1007/s00382-007-0348-4.

    • Search Google Scholar
    • Export Citation
  • Kim, D., , Y.-S. Jang, , D.-H. Kim, , Y.-H. Kim, , M. Watanabe, , F.-F. Jin, , and J.-S. Kug, 2011: El Niño–Southern Oscillation sensitivity to cumulus entrainment in a coupled general circulation model. J. Geophys. Res., 116, D22112, doi:10.1029/2011JD016526.

    • Search Google Scholar
    • Export Citation
  • Kim, S. T., , W. Cai, , F.-F. Jin, , and J.-Y. Yu, 2014: ENSO stability in coupled climate models and its association with mean state. Climate Dyn., 42, 3313–3321, doi:10.1007/s00382-013-1833-6.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., , and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 15871606, doi:10.1175/1520-0442(1993)006,1587:TSCOLS.2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Li, G., , and G. J. Zhang, 2008: Understanding biases in shortwave cloud radiative forcing in the National Center for Atmospheric Research Community Atmosphere Model (CAM3) during El Niño. J. Geophys. Res., 113, D02103, doi:10.1029/2007JD008963.

    • Search Google Scholar
    • Export Citation
  • Li, L. J., , B. Wang, , and G. J. Zhang, 2014: The role of nonconvective condensation processes in response of surface shortwave cloud radiative forcing to El Niño warming. J. Climate, 27, 67216736, doi:10.1175/JCLI-D-13-00632.1.

    • Search Google Scholar
    • Export Citation
  • Lin, J. L., , B. Mapes, , M. Zhang, , and M. Newman, 2004: Stratiform precipitation, vertical heating profiles, and the Madden–Julian oscillation. J. Atmos. Sci., 61, 296309, doi:10.1175/1520-0469(2004)061<0296:SPVHPA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , E. Guilyardi, , H. Weller, , and J. Slingo, 2009: The role of atmosphere feedbacks during ENSO in the CMIP3 models. Atmos. Sci. Lett., 10, 170176, doi:10.1002/asl.227.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , E. Guilyardi, , and H. Weller, 2011: The role of atmosphere feedbacks during ENSO in the CMIP3 models. Part II: Using AMIP runs to understand the heat flux feedback mechanisms. Climate Dyn., 37, 12711292, doi:10.1007/s00382-010-0895-y.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , E. Guilyardi, , and H. Weller, 2012: The role of atmosphere feedbacks during ENSO in the CMIP3 models. Part III: The shortwave feedback. J. Climate, 25, 42754293, doi:10.1175/JCLI-D-11-00178.1.

    • Search Google Scholar
    • Export Citation
  • Ma, H. Y., , S. Xie, , J. S. Boyle, , S. A. Klein, , and Y. Zhang, 2013: Metrics and diagnostics for precipitation-related processes in climate model short-range hindcasts. J. Climate, 26, 15161534, doi:10.1175/JCLI-D-12-00235.1.

    • Search Google Scholar
    • Export Citation
  • Mauritsen, T., and et al. , 2012: Tuning the climate of a global model. J. Adv. Model. Earth Syst., 4, M00A01, doi:10.1029/2012MS000154.

    • 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
  • 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
  • Philander, S., , D. Gu, , G. Lambert, , T. Li, , D. Halpern, , N.-C. Lau, , and R. Pacanowski, 1996: Why the ITCZ is mostly north of the equator. J. Climate, 9, 29582972, doi:10.1175/1520-0442(1996)009<2958:WTIIMN>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., , D. E. Parker, , E. B. Horton, , C. K. Folland, , L. V. Alexander, , D. P. Rowell, , E. C. Kent, , and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Robertson, F. R., , D. E. Fitzjarrald, , and C. D. Kummerow, 2003: Effects of uncertainty in TRMM precipitation radar path integrated attenuation on interannual variations of tropical oceanic rainfall. Geophys. Res. Lett., 30, 1180, doi:10.1029/2002GL016416.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., , and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 22612288, doi:10.1175/1520-0477(1999)080<2261:AIUCFI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and R. A. Houze Jr., 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
  • Schumacher, C., , R. A. Houze Jr., , and I. Kraucunas, 2004: The tropical dynamical response to latent heating estimates derived from the TRMM Precipitation Radar. J. Atmos. Sci., 61, 13411358, doi:10.1175/1520-0469(2004)061<1341:TTDRTL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Seo, K. H., , and W. Q. Wang, 2010: The Madden–Julian oscillation simulated in the NCEP Climate Forecast System model: The importance of stratiform heating. J. Climate, 23, 47704793, doi:10.1175/2010JCLI2983.1.

    • Search Google Scholar
    • Export Citation
  • Shen, X., , Y. Wang, , and X. Li, 2011: Effects of vertical wind shear and cloud radiative processes on responses of rainfall to the large-scale forcing during pre-summer heavy rainfall over southern China. Quart. J. Roy. Meteor. Soc., 137, 236249, doi:10.1002/qj.735.

    • Search Google Scholar
    • Export Citation
  • Shonk, J., , R. Hogan, , G. Mace, , and J. Edwards, 2010: Effect of improving representation of horizontal and vertical cloud structure on the Earth’s global radiation budget. Part I: Review and parametrization. Quart. J. Roy. Meteor. Soc., 136, 11911204, doi:10.1002/qj.647.

    • Search Google Scholar
    • Export Citation
  • Sun, D.-Z., , J. Fasullo, , T. Zhang, , and A. Roubicek, 2003: On the radiative and dynamical feedbacks over the equatorial Pacific cold tongue. J. Climate, 16, 24252432, doi:10.1175/2786.1.

    • Search Google Scholar
    • Export Citation
  • Sun, D.-Z., and et al. , 2006: Radiative and dynamical feedbacks over the equatorial cold tongue: Results from nine atmospheric GCMs. J. Climate, 19, 40594074, doi:10.1175/JCLI3835.1.

    • Search Google Scholar
    • Export Citation
  • Sun, D.-Z., , Y. Yu, , and T. Zhang, 2009: Tropical water vapor and cloud feedbacks in climate models: A further assessment using coupled simulations. J. Climate, 22, 12871304, doi:10.1175/2008JCLI2267.1.

    • Search Google Scholar
    • Export Citation
  • Sundqvist, H., 1978: A parameterization scheme for non-convective condensation including prediction of cloud water content. Quart. J. Roy. Meteor. Soc., 104, 677690, doi:10.1002/qj.49710444110.

    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., , R. J. Stouffer, , and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485498, doi:10.1175/BAMS-D-11-00094.1.

    • Search Google Scholar
    • Export Citation
  • Toniazzo, T., , M. Collins, , and J. Brown, 2008: The variation of ENSO characteristics associated with atmospheric parameter perturbations in a coupled model. Climate Dyn., 30, 643656, doi:10.1007/s00382-007-0313-2.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and et al. , 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 29613012, doi:10.1256/qj.04.176.

  • Vannière, B., , E. Guilyardi, , G. Madec, , F. J. Doblas-Reyes, , and S. Woolnough, 2013: Using seasonal hindcasts to understand the origin of the equatorial cold tongue bias in CGCMs and its impact on ENSO. Climate Dyn., 40, 963981, doi:10.1007/s00382-012-1429-6.

    • Search Google Scholar
    • Export Citation
  • Vannière, B., , E. Guilyardi, , T. Toniazzo, , G. Madec, , and S. Woolnough, 2014: A systematic approach to identify sources of tropical SST errors in coupled models using the adjustment of initialised experiments. Climate Dyn., 43, 22612282, doi:10.1007/s00382-014-2051-6.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , M. Chikira, , Y. Imada, , and M. Kimoto, 2011: Convective control of ENSO simulated in MIROC. J. Climate, 24, 543562, doi:10.1175/2010JCLI3878.1.

    • Search Google Scholar
    • Export Citation
  • Webb, M., , C. Senior, , S. Bony, , and J. J. Morcrette, 2001: Combining ERBE and ISCCP data to assess clouds in the Hadley Centre, ECMWF and LMD atmospheric climate models. Climate Dyn., 17, 905922, doi:10.1007/s003820100157.

    • Search Google Scholar
    • Export Citation
  • Wu, R., , and B. P. Kirtman, 2007: Regimes of seasonal air–sea interaction and implications for performance of forced simulations. Climate Dyn., 29, 393410, doi:10.1007/s00382-007-0246-9.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., 2005: The shape of continents, air-sea interaction, and the rising branch of the Hadley circulation. The Hadley Circulation: Present, Past and Future, H. F. Diaz and R. S. Bradley, Eds., Advances in Global Change Research, Vol. 21, Kluwer Academic Publishers, 121–152.

    • Search Google Scholar
    • Export Citation
  • Xu, K.-M., , and D. A. Randall, 1996: A semiempirical cloudiness parameterization for use in climate models. J. Atmos. Sci., 53, 30843102, doi:10.1175/1520-0469(1996)053<3084:ASCPFU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zebiak, S. E., , and M. A. Cane, 1987: A model El Niño–Southern Oscillation. Mon. Wea. Rev., 115, 22622278, doi:10.1175/1520-0493(1987)115<2262:AMENO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, M., and et al. , 2005: Comparing clouds and their seasonal variations in 10 atmospheric general circulation models with satellite measurements. J. Geophys. Res., 110, D15S02, doi:10.1029/2004JD005021.

    • Search Google Scholar
    • Export Citation
  • Zhang, T., , and D.-Z. Sun, 2006: Response of water vapor and clouds to El Niño warming in three National Center for Atmospheric Research atmospheric models. J. Geophys. Res., 111, D17103, doi:10.1029/2005JD006700.

    • Search Google Scholar
    • Export Citation
  • Zhang, T., , and D.-Z. Sun, 2014: ENSO asymmetry in CMIP5 models. J. Climate, 27, 40704093, doi:10.1175/JCLI-D-13-00454.1.

  • View in gallery

    SW flux feedback during ENSO in the tropical Pacific for (a) ISCCP, and (b)–(r) 17 CMIP5 coupled models, which is measured as the linear SW regression against the Niño-3 index (W m−2 K−1). Niño-3 index is defined as the Niño-3 SST anomaly. Red and black rectangles represent the Niño-4 and Niño-3 regions, respectively.

  • View in gallery

    Averaged values of (a) αsw, (b) αcldtot (% K−1), and (c) αlwp (g m−2 K−1) and (d) αω500 (10−2 Pa s−1 K−1) in the Niño-3 region (red columns) and the Niño-4 region (blue columns) for ISCCP and the 17 CMIP coupled models. Observations (marked with letter a), also shown as red (Niño-3) and blue (Niño-4) dashed lines, are from ISCCP data in (a)–(c) and ERA-40 in (d). The letters on the x axis are sequenced according to the αsw strength of their corresponding models and observations in the Niño-3 region.

  • View in gallery

    As in Fig. 1, but for total cloud fraction (color shading) and total LWP (contours). The solid and dashed contour lines, with an interval of 8 g m−2 K−1, represent the positive and negative LWP feedbacks, respectively. The values separated by a forward slash in the top-right corner of each panel are the correlation coefficients between SW feedback and total cloud fraction feedback and between SW feedback and total LWP feedback, respectively, over the equatorial Pacific (5°S–5°N, 160°E–90°W). Correlation coefficients exceeding 0.13 are significant at the 99% level.

  • View in gallery

    Vertical cross section of cloud fraction feedback (color shading), LWC feedback (black contours), and ice water content feedback (green contours) for 15 coupled models over the equatorial Pacific (5°S–5°N). The solid and dashed lines, with an interval of 4 × 10−3 g kg−1 K−1, represent the positive and negative feedbacks, respectively.

  • View in gallery

    As in Fig. 4, but for RH (% K−1; color shading) and vertical velocity (10−2 Pa s−1 K−1; black contours). The solid and dashed lines, with an interval of 0.5 × 10−2 Pa s−1 K−1, represent the positive (descent) and negative (ascent) feedbacks, respectively.

  • View in gallery

    Average monthly (a) SW feedback, (b) total cloud fraction feedback, (c) total LWP feedback, and (d) dynamical feedback for observation (black), ensemble mean of strong models (red; CNRM-CM5, FGOALS-g2, MIROC5, and CCSM4 producing strong negative SW feedbacks in Niño-3), and mean of weak models (blue; the other 13 models).

  • View in gallery

    As in Fig. 1, but for the 17 uncoupled models.

  • View in gallery

    As in Fig. 2, but for the 17 uncoupled models.

  • View in gallery

    Hovmöller plots of total LWP anomalies (g m−2; color shading) and cloud fraction anomalies (contours) from January 1979 to December 2008 for (a) ISCCP, (e) CNRM-CM5, (f) CSIRO Mk3.6.0, and (k) HadGEM2-ES over the equatorial Pacific (5°S–5°N). The solid and dashed lines represent larger than 20% and smaller than −20% cloud fraction anomalies, respectively. The red and blue dots indicate the El Niño (SSTA > 1.0) and La Niña (SSTA < −1.0) events, respectively.

  • View in gallery

    Scatterplots of Niño-3 averaged SW feedback as a function of (a) total rainfall feedback, (b) convective rainfall feedback, and (c) stratiform rainfall feedback. Black dots are for observation, red asterisks are for CMIP runs, blue asterisks are for AMIP runs, dark green plus signs are for multimodel mean of CMIP runs, and light green circles are for multimodel mean of AMIP runs. The values separated by a forward slash in the top-right corner of each panel are the correlation coefficients between SW feedback and rainfall feedback in the 17 CMIP and AMIP simulations, respectively. Significant correlation coefficient is 0.48 at the 95% level and 0.61 at the 99% level.

  • View in gallery

    As in Fig. 10, but for (a)–(c) stratiform rainfall feedback and (d)–(f) convective rainfall feedback vs total cloud fraction feedback, total LWP feedback, and dynamics feedback.

  • View in gallery

    Niño-3 averaged coefficients of linear regression against SST of (a) αsw, (b) αcldtot, (c) αlwp, and (d) αω500 in observation and 17 models for SSTA < 0 (yellow bars for AMIP and red bars for CMIP) and SSTA > 0 (green bars for AMIP and blue bars for CMIP) separately. The red and blue dashed lines represent the feedback in observation during La Niña and El Niño, respectively. The letters on the x axis are sequenced according to the αsw strength of their corresponding models and observation during El Niño.

  • View in gallery

    Scatterplots of (a) differences of SW feedback between CMIP runs and AMIP runs vs their differences of convective rainfall feedback and (b) their differences of stratiform rainfall feedback during El Niño; SW response vs stratiform rainfall response to La Niña in the (c) AMIP and (d) CMIP runs in the Niño-3 region. The value in parentheses on the top-right corner of (b) is the correlation coefficient between changes in SW responses and those in stratiform rainfall responses to El Niño for 17 CMIP5 models except for GISS-E2-R (marked with letter j).

  • View in gallery

    Vertical cross section of multiannual mean cloud fraction (color shading) and LWC (contours) for the ensemble mean of (a) models (CCSM4, FGOALS-g2, and MIROC5) with strong SW flux feedback, (b) models (except for CCSM4, FGOALS-g2, and MIROC5) with weak SW flux feedback, and (c) the differences (significant at 95%) between (b) and (a) from their historical runs. The contour lines have an interval of 5 × 10−3 g kg−1 in (a),(b) and 3 × 10−3 g kg−1 in (c), where the solid (dashed) lines represent positive (negative) values.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 42 42 7
PDF Downloads 17 17 2

The Role of Moist Processes in Shortwave Radiative Feedback during ENSO in the CMIP5 Models

View More View Less
  • 1 State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
  • | 2 Ministry of Education Key Laboratory for Earth System Modeling, Center for Earth System Science, Tsinghua University, and State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
  • | 3 Ministry of Education Key Laboratory for Earth System Modeling, Center for Earth System Science, Tsinghua University, Beijing, China, and Scripps Institution of Oceanography, La Jolla, California
© Get Permissions
Full access

Abstract

The weak negative shortwave (SW) radiative feedback αsw during El Niño–Southern Oscillation (ENSO) over the equatorial Pacific is a common problem in the models participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5). In this study, the causes for the αsw biases are analyzed using three-dimensional cloud fraction and liquid water path (LWP) provided by the 17 CMIP5 models and the relative roles of convective and stratiform rainfall feedbacks in αsw are explored. Results show that the underestimate of SW feedback is primarily associated with too negative cloud fraction and LWP feedbacks in the boundary layers, together with insufficient middle and/or high cloud and dynamics feedbacks, in both the CMIP and Atmospheric Model Intercomparsion Project (AMIP) runs, the latter being somewhat better. The underestimations of SW feedbacks are due to both weak negative SW responses to El Niño, especially in the CMIP runs, and strong positive SW responses to La Niña, consistent with their biases in cloud fraction, LWP, and dynamics responses to El Niño and La Niña. The convective rainfall feedback, which is largely reduced owing to the excessive cold tongue in the CMIP runs compared with their AMIP counterparts, contributes more to the difference of SW feedback (mainly under El Niño conditions) between the CMIP and AMIP runs, while the stratiform rainfall plays a more important role in SW feedback during La Niña.

Corresponding author address: Dr. Lijuan Li, LASG, Institute of Atmospheric Physics, Chinese Academy of Sciences, P.O. Box 9804, Beijing 100029, China. E-mail: ljli@mail.iap.ac.cn

Abstract

The weak negative shortwave (SW) radiative feedback αsw during El Niño–Southern Oscillation (ENSO) over the equatorial Pacific is a common problem in the models participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5). In this study, the causes for the αsw biases are analyzed using three-dimensional cloud fraction and liquid water path (LWP) provided by the 17 CMIP5 models and the relative roles of convective and stratiform rainfall feedbacks in αsw are explored. Results show that the underestimate of SW feedback is primarily associated with too negative cloud fraction and LWP feedbacks in the boundary layers, together with insufficient middle and/or high cloud and dynamics feedbacks, in both the CMIP and Atmospheric Model Intercomparsion Project (AMIP) runs, the latter being somewhat better. The underestimations of SW feedbacks are due to both weak negative SW responses to El Niño, especially in the CMIP runs, and strong positive SW responses to La Niña, consistent with their biases in cloud fraction, LWP, and dynamics responses to El Niño and La Niña. The convective rainfall feedback, which is largely reduced owing to the excessive cold tongue in the CMIP runs compared with their AMIP counterparts, contributes more to the difference of SW feedback (mainly under El Niño conditions) between the CMIP and AMIP runs, while the stratiform rainfall plays a more important role in SW feedback during La Niña.

Corresponding author address: Dr. Lijuan Li, LASG, Institute of Atmospheric Physics, Chinese Academy of Sciences, P.O. Box 9804, Beijing 100029, China. E-mail: ljli@mail.iap.ac.cn

1. Introduction

The shortwave (SW) radiative feedback, defined as the regression coefficient between the net SW radiative flux anomalies at the surface and the sea surface temperature (SST) anomalies in the Niño-3 region, denoted by αsw, in the equatorial Pacific is one of the dominant components of negative heat flux feedbacks that drive El Niño–Southern Oscillation (ENSO) evolution (Zebiak and Cane 1987; Guilyardi et al. 2009b; Lloyd et al. 2009; Bellenger et al. 2014). During El Niño, a sufficiently large increase in SST leads to an increase of convective cloud, which decreases the SW flux reaching the surface, and hence a negative SW feedback that reduces the El Niño warming. During La Niña, a negative SST anomaly increases the stability of the atmospheric boundary layer and the amount of marine stratiform clouds, which decrease the surface SW flux, thus producing a positive SW feedback that will contribute to further La Niña cooling (e.g., Philander et al. 1996; Xie 2005; Bellenger et al. 2014). On average, the SW feedback during ENSO events is negative. In addition, the nonlinearity of the SW feedback, being negative in El Niño and positive in La Niña, also plays an important role in the ENSO skewness (toward warm events), amplitude, and phase lock in models from phases 3 and 5 of the the Coupled Model Intercomparison Project (CMIP3 and CMIP5, respectively; Lloyd et al. 2012; Bellenger et al. 2014).

However, both the average strength and the nonlinearity of SW feedback are poorly reproduced by most of the CMIP models. The models tend to simulate smaller negative or even positive αsw, and underestimate the nonlinearity when compared to the observations and reanalyses, especially over the eastern equatorial Pacific (Sun et al. 2003, 2006; Zhang and Sun 2006; Lloyd et al. 2011; Bellenger et al. 2014; Kim et al. 2014). The biases of the SW feedback in the CMIP models are directly linked to the atmosphere, the ocean, and their coupling. In the atmospheric models, the αsw biases are closely associated with the cloud-related process schemes, that is, the convective parameterization and nonconvective condensation schemes, tuning parameters, and model resolution (Kim et al. 2008, 2011; Li and Zhang 2008; Neale et al. 2008; Toniazzo et al. 2008; Guilyardi et al. 2009a; Watanabe et al. 2011; Li et al. 2014). Both the excessive cold tongue or SST bias, which could be attributed to the atmospheric models via deficient surface winds (Vannière et al. 2013, 2014), and the amplified errors (e.g., weakened dynamical feedback) during the coupling process are also important factors (Sun et al. 2009; Lloyd et al. 2012; Chen et al. 2013).

To identify the sources of the αsw biases, Lloyd et al. (2012) introduced an idealized shortwave feedback decomposition method [see Eq. (3) in their study] under the assumption that shortwave cloud feedback depends on cloud cover only. By applying the chain rule, they further related the total cloud cover to atmospheric dynamics, which in turn is linked to local SST only. In nature, shortwave flux also depends on the liquid water path (LWP, an integral of liquid water content or mass fraction between two points in the atmosphere). Li et al. (2014) included the LWP feedback in the decomposition [their Eq. (2)] and suggested that the underestimations of low- and midcloud cover feedback and LWP feedback are the main sources of shortwave biases in the Grid-Point Atmospheric Model of the Institute of Atmospheric Physics (IAP) of the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), version 1 (GAMIL1), one of the CMIP3 models. In fact, the “too few, too bright” low-cloud problem, underestimating the low-cloud cover and overestimating its optical thickness, has been a common issue in global climate models (GCM) (e.g., Webb et al. 2001; Zhang et al. 2005; Nam et al. 2012). Hence, this study will extend the analysis of Li et al. (2014) to explore the causes of the αsw biases, including its nonlinearity in the CMIP5 models from the standpoint of the vertical distribution of cloud cover and LWP.

The problem of too little stratiform rainfall over the tropical Pacific is another issue in climate models (e.g., Dai 2006). Although the definition of stratiform rainfall in satellite observations is different from that of GCM simulations, the impacts of the stratiform process on the vertical heating profiles and general circulation have been emphasized in both observations/reanalyses and simulations (e.g., Schumacher et al. 2004; Lin et al. 2004; Boyle and Klein 2010; Ma et al. 2013). Furthermore, the enhanced stratiform processes contribute to the improved shortwave feedback during ENSO by affecting the profiles of cloud fraction and LWP in version 2 of GAMIL (Li et al. 2014). Thus, this paper also aims to investigate whether the same relationship or physical mechanism between stratiform processes and shortwave feedback in the GAMIL exists in other CMIP5 models.

The paper is organized as follows: Section 2 describes the CMIP5 models, the observational data, and the method used to estimate the feedback. The SW feedback analysis in the coupled and uncoupled models, and their relationship with the convective and stratiform rainfall feedback are examined in section 3. Section 4 gives a summary and discussion.

2. Models, data, and analysis method

a. Models and observational data

There are over 40 CMIP5 models. However, not every model has the output data that are needed for this study. Therefore, we chose the 17 CMIP5 models that performed the historical and AMIP runs and published three-dimensional (3D) cloud variables, and analyzed the first ensemble member (r1i1p1) of their historical runs (1951–2005) and AMIP runs (1979–2008) (Taylor et al. 2012; Table 1). The 3D variables [e.g., cloud fraction and liquid water content (LWC)] at the model vertical levels are first linearly interpolated onto common pressure levels according to their formulas describing the vertical coordinates. In the 17 models, the CNRM-CM5 did not publish the 3D cloud fraction and LWC and the HadGEM2-ES did not provide the surface elevation needed for calculating the model layers. Thus, only 15 models are included in the analysis of 3D cloud distributions.

Table 1.

List of the CMIP5 models used in this study.

Table 1.

For comparison with the model results, the SW flux, cloud cover, and LWP from the International Satellite Cloud Climatology Project (ISCCP; Rossow and Schiffer 1999) for the period of July 1983–December 2008 are used. The precipitation (including convective and stratiform components) datasets are from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager level 3 data (TRMM 3A12; Robertson et al. 2003), covering January 1998–December 2013. The relative humidity (RH) and vertical velocity are from the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40; Uppala et al. 2005). The SST is obtained from the Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST; Rayner et al. 2003). All of the abovementioned datasets (observations, reanalysis, and simulations) are bilinearly interpolated into the uniform grid of 1.875° × 1.875°.

b. Method

For an atmospheric variable F (e.g., shortwave flux), the feedback during ENSO is measured by the linear regression coefficient α:
e1
where FA is the anomaly of F after removing the annual cycle and 〈SSTA〉 is the SST anomaly averaged over the Niño-3 region (5°S–5°N, 150°–90°W). Because of the nonlinearity of the regression (i.e., α being nonconstant) with respect to El Niño and La Niña, the linear regression of FA against SSTA at each grid point is computed separately for SSTA > 0 and SSTA < 0 and then averaged over the Niño-3 region (Lloyd et al. 2012).
The shortwave flux feedback can be written as
e2
where dCLD/dSST and dLWP/dSST are cloud fraction feedback and LWP feedback, respectively (Li et al. 2014). Since cloud fraction is a function of RH in most climate models and since it responds to changes in the large-scale atmospheric dynamics (represented by 500-hPa vertical velocity) (e.g., Sundqvist 1978; Xu and Randall 1996; Lloyd et al. 2012), the cloud fraction feedback can be further decomposed into dynamic feedback and relative humidity feedback:
e3
where 500/dSST is the dynamics feedback and dRH/dSST is the RH feedback. Note that, Eq. (3) implies that 500-hPa vertical velocity and RH are independent of each other although this assumption, in fact, is not entirely true.

3. Results

a. Biases in CMIP runs

The SW radiative feedbacks during ENSO for ISCCP and the 17 CMIP5 coupled simulations are shown in Fig. 1. In ISCCP, the negative SW feedback mainly occurs in the equatorial Pacific east of 150°E with a maximum near the date line. The positive values in the tropical Pacific west of 150°E and the off-equatorial central and eastern Pacific should be interpreted as a response to Niño-3 SSTA. During El Niño, cooling occurs in the western Pacific Maritime Continent region. Thus, the remote positive SW response to El Niño warming has a damping effect on the local SSTA. The average negative feedback in the Niño-4 region is −12.9 W m−2 K−1 and is much larger in magnitude than that in the Niño-3 region (−5.5 W m−2 K−1) (Fig. 2a). For the models, 13 out of the 17 models (except for CCSM4, CNRM-CM5, FGOALS-g2, and MIROC5) have large biases in both spatial negative/positive pattern distribution and feedback strength. The positive feedback values in these models extend westward to the central equatorial Pacific, even across the entire equatorial Pacific (CSIRO Mk3.6.0), and the negative values further shift into the western tropical Pacific, where ISCCP shows positive values. The feedback strengths in these 13 models range from 6 (IPSL-CM5A-LR) to −1 W m−2 K−1 (GISS-E2-R) in Niño-3; they are much less negative, −5.5 W m−2 K−1 in ISCCP; and from 6.9 (CSIRO Mk3.6.0) to −15.1 W m−2 K−1 (NorESM1-M) in Niño-4, compared to −12.9 W m−2 K−1 in ISCCP (Fig. 2a). The other four models simulate a more reasonable negative SW feedback distribution and strength, with CNRM-CM5 producing a somewhat overly strong negative feedback.

Fig. 1.
Fig. 1.

SW flux feedback during ENSO in the tropical Pacific for (a) ISCCP, and (b)–(r) 17 CMIP5 coupled models, which is measured as the linear SW regression against the Niño-3 index (W m−2 K−1). Niño-3 index is defined as the Niño-3 SST anomaly. Red and black rectangles represent the Niño-4 and Niño-3 regions, respectively.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

Fig. 2.
Fig. 2.

Averaged values of (a) αsw, (b) αcldtot (% K−1), and (c) αlwp (g m−2 K−1) and (d) αω500 (10−2 Pa s−1 K−1) in the Niño-3 region (red columns) and the Niño-4 region (blue columns) for ISCCP and the 17 CMIP coupled models. Observations (marked with letter a), also shown as red (Niño-3) and blue (Niño-4) dashed lines, are from ISCCP data in (a)–(c) and ERA-40 in (d). The letters on the x axis are sequenced according to the αsw strength of their corresponding models and observations in the Niño-3 region.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

SW flux feedback is closely associated with cloud fraction and LWP feedbacks [Eq. (2)]. The spatial correlation coefficient in the Niño-3+4 region (5°S–5°N, 160°E–90°W) between αsw and the regression coefficient between total cloud fraction and SST anomalies αcldtot is −0.91 for ISCCP and ranges from −0.80 to −0.99 for CMIP5 models with the exception of GISS-E2-R, which has a much lower (in magnitude) correlation (−0.35) than the other models. Likewise, the spatial correlation coefficient between αsw and the total LWP feedback αlwp is −0.89 for ISCCP and range from −0.49 to −0.98 for the models (except two IPSL models, which have positive correlation), all exceeding the significance level of 0.01, that is, 0.135 (Fig. 3). The high spatial correlation coefficients imply that the biases in αsw result from those in αcldtot and αlwp. For instance, when compared to ISCCP, the negative αcldtot and/or αlwp over the eastern equatorial Pacific in some models (BCC_CSM1, CanESM2, CSIRO Mk3.6.0, FGOALS-s2, HadGEM2-ES, INM-CM4.0, IPSL-CM5A-LR, IPSL-CM5A-MR, MPI-ESM-LR, and MPI-ESM-MR) are responsible for their positive αsw, and the too strong positive αcldtot and αlwp over the entire equatorial Pacific in CNRM-CM5 are also the causes for its overestimated negative αsw. Quantitatively, in Niño-3 the relatively small αsw biases in some models (e.g., CCSM4, FGOALS-g2, and MIROC5 marked with letters d, g, and o, respectively, in Figs. 2b,c) are a result of error compensation between underestimated cloud fraction feedback and overestimated LWP feedback. In Niño-4, this type of error compensation, weak αcldtot and strong αlwp, is also true for FGOALS-g2, CCSM4, and NorESM1-M. The right signs of αsw (negative) and αcldtot (positive) in Niño-4 are simulated by more models than those in Niño-3, while the right simulations of αlwp in both Niño-3 and Niño-4 regions are more challenging.

Fig. 3.
Fig. 3.

As in Fig. 1, but for total cloud fraction (color shading) and total LWP (contours). The solid and dashed contour lines, with an interval of 8 g m−2 K−1, represent the positive and negative LWP feedbacks, respectively. The values separated by a forward slash in the top-right corner of each panel are the correlation coefficients between SW feedback and total cloud fraction feedback and between SW feedback and total LWP feedback, respectively, over the equatorial Pacific (5°S–5°N, 160°E–90°W). Correlation coefficients exceeding 0.13 are significant at the 99% level.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

A further look into the vertical profiles of cloud fraction and liquid water content indicates that when compared to the models with small αsw biases (e.g., CCSM4, FGOALS-g2, and MIROC5), there are obvious underestimations of cloud fraction and liquid water content feedbacks below 400 hPa in those models with large αsw biases (e.g., CanESM2, CSIRO Mk3.6.0, FGOALS-s2, GFDL CM3, GISS-E2-R, two IPSL models, and two MPI models) (Fig. 4). Especially, the negative values of cloud fraction and LWP feedbacks in the boundary layer, together with insufficient mid- or high-cloud feedbacks, are responsible for the negative αcldtot and αlwp in the eastern Pacific, consistent with the findings that the lowest clouds have an overly negative SW (e.g., Lloyd et al. 2012) and with the results from GAMIL1, which simulated too negative low-cloud and LWP feedback, and thus positive αsw in the eastern Pacific (Li et al. 2014). Furthermore, because of the diagnostic relationship between cloud fraction and RH in most climate models, the negative cloud fraction feedback corresponds well with the negative RH feedback [Eq. (3)], in particular near the surface and eastern Pacific (Fig. 5). The positive cloud fraction feedbacks in the CMIP5 models are not identical to the positive RH feedbacks, because different cloud schemes or RH thresholds are used in different models. The pattern correlation coefficients between cloud fraction and RH feedbacks over the Niño-3+4 region (160°E–90°W at 1000–100 hPa) are greater than 0.4 in most models, except for four models (i.e., 0.17 for CSIRO Mk3.6.0, 0.18 for MPI-ESM-LR, 0.3 for MPI-ESM-MR, and 0.19 for NorESM1-M), all exceeding the significance level of 0.01 (corresponding to a correlation coefficient of 0.081), and GISS-E2-R with a negative correlation (−0.28). The feedbacks of cloud fraction and RH above the boundary layer also coincide with the vertical circulation feedback, for example, the positive cloud fraction and RH feedbacks with anomalous ascent and the negative cloud fraction and RH feedbacks with anomalous descent. When quantifying the circulation feedback using the vertical velocity feedback at 500 hPa (αω500, also known as dynamical feedback), the models (except for GISS-E2-R) simulating stronger dynamical feedbacks—for example, CCSM4, CNRM-CM5, FGOALS-g2, and MIROC5—also simulate stronger positive αcldtot as well as stronger negative αsw relative to other models [Fig. 2; Eqs. (2) and (3)].

Fig. 4.
Fig. 4.

Vertical cross section of cloud fraction feedback (color shading), LWC feedback (black contours), and ice water content feedback (green contours) for 15 coupled models over the equatorial Pacific (5°S–5°N). The solid and dashed lines, with an interval of 4 × 10−3 g kg−1 K−1, represent the positive and negative feedbacks, respectively.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

Fig. 5.
Fig. 5.

As in Fig. 4, but for RH (% K−1; color shading) and vertical velocity (10−2 Pa s−1 K−1; black contours). The solid and dashed lines, with an interval of 0.5 × 10−2 Pa s−1 K−1, represent the positive (descent) and negative (ascent) feedbacks, respectively.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

To further understand the model biases, using the αsw strength in Niño-3 (Fig. 2a), we divide the 17 CMIP5 models into two groups—strong (CNRM-CM5, FGOALS-g2, MIROC5, and CCSM4) and weak (the other 13 models) groups—and show their seasonal variations in αsw, αlwp, andαcldtot, as well as αω500 (Fig. 6). The ISCCP shows a clear seasonal variation, a strong/weak negative SW feedback during the first/second half of the year. The strong group reproduces the observed seasonal variation with a somewhat underestimation (overestimation) of αsw in magnitude during the first (second) half of the year, whereas the weak group displays a constant positive αsw throughout the year (Fig. 6a). The strong group models show large seasonal variations and most of the weak group models (except for GFDL CM3 and NorESM1-M) have small seasonal variations. This behavior is similar to that seen in Bellenger et al. (2014) in spite of the different experiments (historical run versus preindustrial run) and integration lengths (55 versus 300–1000 yr). The seasonal variation of αsw is directly connected to that of αlwp and αcldtot as well as αω500 (Figs. 6b–d). In the observation, the strong SW feedback in spring corresponds to large magnitudes of αlwp, αcldtot, and αω500, and the weak SW feedback in summer and fall corresponds to small magnitudes of αlwp, αcldtot, and αω500. In the weak group, the nearly unchanging SW feedback in all seasons agrees with the changeless αlwp and αcldtot in spite of the weak αω500 seasonal cycle. In the strong group, the underestimation of αsw magnitude in spring results from smaller magnitudes of αcldtot and αω500, and the overestimation of αsw in summer and fall results from larger magnitudes of αlwp and αω500 when compared to the observation. In addition, in the four strong models the overestimation of annual mean αlwp is from all seasons and the underestimation of annual mean αcldtot is mainly from the spring.

Fig. 6.
Fig. 6.

Average monthly (a) SW feedback, (b) total cloud fraction feedback, (c) total LWP feedback, and (d) dynamical feedback for observation (black), ensemble mean of strong models (red; CNRM-CM5, FGOALS-g2, MIROC5, and CCSM4 producing strong negative SW feedbacks in Niño-3), and mean of weak models (blue; the other 13 models).

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

b. Biases in AMIP runs

The errors in SW feedback in coupled models can be due to either errors in shortwave simulation or errors in SST simulation, or both. To isolate their contributions to the SW feedback errors in coupled models, we analyze the uncoupled model results that are forced by the observed SST in this section. As noted in previous studies, the AMIP runs generally produce better αsw in both spatial pattern and strength compared to their CMIP counterparts in the equatorial central and eastern Pacific, where SST forcing dominates (Wu and Kirtman 2007) (Fig. 7). In the Niño-3 region, 8 out of the 17 atmospheric models fall within the 50% range of the observed αsw strength and five models fall within the 20% range, as opposed to only two coupled models (CCSM4 and MIROC5) in the two (20% and 50%) ranges (Fig. 8a). The ratio for coupled models (2 out of 17 models, ~12%) is comparable to 10% of CMIP5 coupled models within 25% of the observation in Bellenger et al. (2014) even though different experiments and time integration lengths are used in the two studies. In the Niño-4 region, all 17 atmospheric models fall within 50% of the observed values, of which 10 models fall within the 20% range of the observation versus 7 (50% range) and 3 (20% range) coupled models. Nonetheless, there are still obvious αsw biases (Guilyardi et al. 2009a; Lloyd et al. 2012). Here we focus on the Niño-3 region, where the observed negative αsw has a local damping effect on interannual SSTA. Of the 17 models, 11 underestimate the negative αsw, including three models (CSIRO Mk3.6.0, IPSL-CM5A-LR, and IPSL-CM5A-MR) that simulate the positive responses, and the underestimations occur in all seasons (figure not shown), as in the CMIP runs (Fig. 6a). Two models (GFDL CM3 and GISS-E2-R marked with letters i and j, respectively, in Fig. 8a) simulate the closest strength to the ISCCP. The other four models (CCSM4, CNRM-CM5, FGOALS-g2, and MIROC5) overestimate the αsw to various extents. The underestimations or overestimations of αsw are due to biases in αlwp and/or αcldtot as well as αω500 (Figs. 8b–d). It is noted that the seemingly correct αsw in GISS-E2-R in the Niño-3 region is a result of error compensation between underestimated αlwp and overestimated ice water path (IWP) feedback αiwp, which is also seen in its coupled simulation (green lines in Fig. 3j).

Fig. 7.
Fig. 7.

As in Fig. 1, but for the 17 uncoupled models.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

Fig. 8.
Fig. 8.

As in Fig. 2, but for the 17 uncoupled models.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

These biases of cloud fraction and LWP feedback using the linear regression are also clearly reflected in their temporal evolutions. For example, based on the strength of αcldtot and αlwp in the Niño-3 region, we choose three models: CNRM-CM5 and CSIRO-Mk3.6.0, presenting large positive or negative biases, and HadGEM2-ES having small biases, and show their temporal evolutions in Fig. 9. Compared with ISCCP, the negative αlwp of CSIRO Mk3.6.0 in the Niño-3 region agrees with the negative LWP anomalies in the eastern Pacific during El Niño warming years (e.g., 1982/83, 1987/88, and 1997/98) and the positive LWP anomalies between 150° and 120°W during La Niña (e.g., 1988/89 and 1999/2000). The overestimated positive αlwp by the CNRM-CM5 coincides with the overly positive (negative) LWP anomalies during El Niño (La Niña) events, and the correct αlwp in the HadGEM2-ES is consistent with the correct simulation of LWP anomalies. It is also true for cloud fraction. The abovementioned findings indicate that the results are independent of the methods used to represent the feedback/response, and that the biases in SW, LWP, and cloud fraction responses are rooted in AGCMs, especially in the cloud-related processes.

Fig. 9.
Fig. 9.

Hovmöller plots of total LWP anomalies (g m−2; color shading) and cloud fraction anomalies (contours) from January 1979 to December 2008 for (a) ISCCP, (e) CNRM-CM5, (f) CSIRO Mk3.6.0, and (k) HadGEM2-ES over the equatorial Pacific (5°S–5°N). The solid and dashed lines represent larger than 20% and smaller than −20% cloud fraction anomalies, respectively. The red and blue dots indicate the El Niño (SSTA > 1.0) and La Niña (SSTA < −1.0) events, respectively.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

c. Relationship between the SW feedback and rainfall feedback

The convective and nonconvective condensation processes are two of the most important cloud-related processes in the AGCM. Here the convective and stratiform rainfall is used as proxies to represent the moist processes. Many studies have suggested that both the climatological mean precipitation and precipitation variation with SSTA (or precipitation feedback) play an important role in ENSO simulations by affecting the lower-tropospheric circulation and surface heat flux (Toniazzo et al. 2008; Sun et al. 2009; Watanabe et al. 2011; Ham and Kug 2014). In the 17 CMIP runs, the SW feedback and total rainfall feedback αpr, convective rainfall feedback αprc, and stratiform rainfall feedback αprl in the Niño-3 region are significantly correlated among the models, with a negative linear correlation coefficient of −0.73, −0.59, and −0.67, respectively (Fig. 10). Models with positive SW feedbacks have small rainfall feedbacks and models with negative SW feedbacks have large rainfall feedbacks. These suggest that moist processes (both convective and nonconvective) play an important role in SW feedback. In the AMIP runs, however, the relationship between them is not as significant, especially between αsw and αprc. This is due to the rainfall responses to the same SST changes among the AMIP runs being closer than in the CMIP runs, while the spread of αsw in the AMIP runs is nearly as much as that in the CMIP runs. The total rainfall responses in the AMIP runs cluster around that from TRMM, which may be a result of model tuning (e.g., Mauritsen et al. 2012), whereas their convective (stratiform) (except for GISS-E2-R) components are heavily overestimated (underestimated). It is noted that the total rainfall feedback (1.35 mm day−1 K−1) from TRMM is very close to those in the Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP) and the Global Precipitation Climatology Project (GPCP) (1.3 and 1.2 mm day−1 K−1, respectively), and the stratiform rainfall feedback accounts for about 45% of the total rainfall feedback, which has about the same proportion as their climatological mean states (Schumacher and Houze 2003). The overestimations of αprc in the AMIP runs are erroneously reduced in CMIP, due to the excessive cold tongue in CMIP (Sun et al. 2009; Chen et al. 2013), while the underestimations of αprl do not change much between the coupled and uncoupled runs. The different biases in αprc and αprl between CMIP and AMIP experiments may indicate their different roles in SW feedback.

Fig. 10.
Fig. 10.

Scatterplots of Niño-3 averaged SW feedback as a function of (a) total rainfall feedback, (b) convective rainfall feedback, and (c) stratiform rainfall feedback. Black dots are for observation, red asterisks are for CMIP runs, blue asterisks are for AMIP runs, dark green plus signs are for multimodel mean of CMIP runs, and light green circles are for multimodel mean of AMIP runs. The values separated by a forward slash in the top-right corner of each panel are the correlation coefficients between SW feedback and rainfall feedback in the 17 CMIP and AMIP simulations, respectively. Significant correlation coefficient is 0.48 at the 95% level and 0.61 at the 99% level.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

The convective and stratiform processes influence the SW feedback by affecting (or interacting with) the feedbacks of cloud fraction, LWP, and dynamics (Fig. 11). The impacts of convective and stratiform processes on total cloud fraction feedback are similar (Figs. 11a,d), with the main differences lying in their correlations with LWP and dynamics feedback. In CMIP, the correlation coefficients between αprl and αlwp, αprl and αω500, αprc and αlwp, and αprc and αω500 are 0.68, −0.85, 0.79, and −0.63, respectively, statistically significant at 99%. In AMIP, the coefficients do not even exceed the significance level of 90%, except between αprl and αω500 (−0.55 significant at 90%). In particular, the poor correlation (−0.08) between αprc and αlwp is the main factor responsible for the low correlation between αprc and αsw (Fig. 10b). The strong relationship between αprl and αω500 in both CMIP and AMIP is in agreement with previous studies, which showed that the increase in stratiform rainfall strengthens the upper-level circulation (e.g., Schumacher et al. 2004; Ma et al. 2013), indicating another pathway of the stratiform process affecting SW feedback besides interacting with cloud fraction and LWP. By comparing the strong correlations among different CMIP runs, the weak correlations among AMIP runs may be related to the fact that all the AMIP runs are performed under the same SST forcing and all the feedbacks lack air–sea coupling in forced simulations.

Fig. 11.
Fig. 11.

As in Fig. 10, but for (a)–(c) stratiform rainfall feedback and (d)–(f) convective rainfall feedback vs total cloud fraction feedback, total LWP feedback, and dynamics feedback.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

d. Nonlinearity

To determine whether the errors in αsw, αcldtot, αlwp, and αω500 in the models mainly come from El Niño or La Niña, Fig. 12 shows the regression coefficients of SW, total cloud fraction, LWP, and 500-hPa vertical velocity in Niño-3 under SSTA > 0 and SSTA < 0 conditions in AMIP and CMIP. In general, the responses of the atmospheric variables to El Niño are much stronger than those to La Niña in the observations (or reanalyses) and most AMIP simulations (Li et al. 2014). Compared with the ISCCP, the underestimations of αsw in most models are from both the underestimation of the negative response (responses in CSIRO Mk3.6.0, IPSL-CM5A-LR, and IPSL-CM5A-MR are even positive) to El Niño and overestimation of the positive response to La Niña (Fig. 12a). The overestimations of αsw in CNRM-CM5, FGOALS-g2, and MIROC5 are from both the overestimated negative response to El Niño and the incorrect negative response to La Niña. As demonstrated in the previous section, the biases of αsw come from those of αlwp, αcldtot, and/or αω500 (Figs. 12b–d). In particular, the overestimated negative cloud fraction and LWP responses to La Niña in the models are directly related to their negative responses in the boundary layers (Fig. 4), indicating the excessive low cloud (including LWP) under La Niña. Li et al. (2014) showed that the excessive low cloud in GAMIL1 is due to enhanced atmospheric stability in the lower troposphere (as measured by the differences in potential temperature between 700 hPa and the surface; Klein and Hartmann 1993) under cold SST. Here we found no significant correlation between stability feedback and αsw, αcldtot, or αlwp in the CMIP5 models (figures not shown).

Fig. 12.
Fig. 12.

Niño-3 averaged coefficients of linear regression against SST of (a) αsw, (b) αcldtot, (c) αlwp, and (d) αω500 in observation and 17 models for SSTA < 0 (yellow bars for AMIP and red bars for CMIP) and SSTA > 0 (green bars for AMIP and blue bars for CMIP) separately. The red and blue dashed lines represent the feedback in observation during La Niña and El Niño, respectively. The letters on the x axis are sequenced according to the αsw strength of their corresponding models and observation during El Niño.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

In addition, compared to their AMIP counterparts, the increased αsw biases in CMIP are mostly due to changed responses to El Niño—for example, the reversed positive response in BCC_CSM1, CanESM2, FGOALS-s2, HadGEM2-ES, INM-CM4.0, MPI-ESM-LR, and MPI-ESM-MR; the reduced negative response in GFDL CM3, GISS-E2-R, and NorESM1-M; and the increased positive response in CSIRO Mk3.6.0, IPSL-CM5A-LR, and IPSL-CM5A-MR. These changes are closely related to convective precipitation changes. Figures 13a and 13b show the differences of SW feedback between the CMIP and AMIP runs versus those of convective and stratiform precipitation feedback, respectively, during El Niño for each of the 17 models. They are well correlated, with a correlation coefficient of −0.68. During El Niño, the changes in SW feedback from the AMIP runs to the CMIP runs are mainly attributed to changes in convective rainfall responses; that is, models (e.g., the two MPI models and CanESM2 marked with letters p, q, and c, respectively, in Fig. 13) with large decrease in convective rainfall feedback in CMIP when compared to that in AMIP produce the most deteriorated SW feedback, with stratiform rainfall responses playing an insignificant role (Figs. 13a,b). By contrast, during La Niña, changes in SW feedback from the AMIP runs to the CMIP runs are mainly attributed to changes in stratiform rainfall responses, with convective rainfall responses playing an insignificant role (figure not shown). Furthermore, the stratiform processes clearly impact the SW feedback during La Niña in both the AMIP and CMIP runs, with correlation coefficients of −0.68 and −0.72, respectively (Figs. 13c,d). Under La Niña, the strengthened stratiform process in the models is primarily associated with the increased low-level moisture convergence or weakened moisture divergence (Seo and Wang 2010; Shen et al. 2011; also see Fig. 5), which is conducive to increasing low cloud (including LWP) and decreasing SW flux at the surface, and thus a positive SW feedback (Toniazzo et al. 2008; Brient and Bony 2012).

Fig. 13.
Fig. 13.

Scatterplots of (a) differences of SW feedback between CMIP runs and AMIP runs vs their differences of convective rainfall feedback and (b) their differences of stratiform rainfall feedback during El Niño; SW response vs stratiform rainfall response to La Niña in the (c) AMIP and (d) CMIP runs in the Niño-3 region. The value in parentheses on the top-right corner of (b) is the correlation coefficient between changes in SW responses and those in stratiform rainfall responses to El Niño for 17 CMIP5 models except for GISS-E2-R (marked with letter j).

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

The different changes in SW response to El Niño and La Niña in AMIP and CMIP are also related to the simulations of ENSO asymmetry. The ENSO amplitude (standard deviation of Niño-3 indices) simulation in most models (14 models except GISS-E2-R, INM-CM4.0, and MIROC5) falls within 25% of the observed value, whereas the observed positive skewness is underestimated by 15 models, except for CCSM4 and MIROC5, in which six models (CSIRO Mk3.6.0, GISS-E2-R, HadGEM2-ES, INM-CM4.0, IPSL-CM5A-LR, and NorESM1-M) produce a negative asymmetry (figure not shown). The small positive skewness in the coupled models primarily results from the weaker positive SSTA over the eastern Pacific and its westward shift (Zhang and Sun 2014). Therefore, compared with the uncoupled models, the weak El Niño in the coupled models contributes to its reduced rainfall anomalies (in particular convective rainfall) as well as the weakened negative SW responses. In contrast, the response to La Niña does not change much between the coupled and uncoupled simulations as the negative SSTA in the coupled models is close to the observation (Zhang and Sun 2014). In addition, the strong correlation between ENSO amplitude and αsw nonlinearity (the difference of SW response to El Niño and La Niña) is found as in Bellenger et al. (2014), which is mainly from the correlation (0.63, significant at 99%) between ENSO amplitude and αcldtot nonlinearity (the difference of total cloud fraction response to El Niño and La Niña) and partly from the relationship (0.49, significant at 95%) between ENSO amplitude and αlwp nonlinearity (the difference of total LWP response to El Niño and La Niña).

4. Summary and discussion

Previous studies suggested that the shortwave radiative feedback during ENSO could be decomposed into cloud fraction feedback, liquid water path feedback, and dynamics feedback. Based on this decomposition, the causes for the biases of shortwave radiative feedback over the eastern Pacific are analyzed using the three-dimensional cloud fraction and liquid water content (or LWP) provided by the 17 CMIP5 models. The relative roles of convective and stratiform rainfall feedbacks in the αsw are explored.

Following previous work, it is confirmed that 13 out of the 17 CMIP5 coupled models simulate a weak negative or even a positive αsw and a westward shift of the negative center in the equatorial Pacific; the other 4 models (CCSM4, CNRM-CM5, FGOALS-g2, and MIROC5) simulate relatively strong negative αsw, especially in the Niño-3 region. In the Niño-3 region, 12% of the CMIP5 models (two models) fall within 20% of the observation, comparable to 10% for CMIP5 (Bellenger et al. 2014) and to a less extent larger than 5% for CMIP3 (Lloyd et al. 2012). The positive and weak negative αsw in most models are primarily associated with the negative feedbacks of cloud fraction and liquid water path feedback in the boundary layers, together with insufficient middle or high cloud and weak dynamics feedbacks. The reasonable negative αsw in CCSM4, FGOALS-g2, and MIROC5 is largely a result of error compensation between the overestimated LWP feedback and the underestimated cloud fraction feedback. These biases remain significant in the uncoupled models and are independent of the linear regression method, indicating they have their sources in the AGCMs, especially in cloud condensation (convective and stratiform) processes.

The convective (stratiform) rainfall feedbacks are heavily overestimated (underestimated) in most of the AMIP runs, although their total rainfall feedbacks are closer to the observations. Compared with the AMIP runs, the convective rainfall feedback is largely reduced in the CMIP runs, because of excessive cold tongue in the CMIP (figure not shown; also see Sun et al. 2009; Chen et al. 2013), while the stratiform rainfall feedback does not change much. The changes in convective rainfall feedback contribute more than the stratiform counterparts to changes in αsw between the coupled and uncoupled simulations, and the changes in both convective rainfall feedback and SW feedback are mainly from the reduced responses to El Niño. The stratiform rainfall feedback plays a more important role in the αsw during La Niña.

The nonlinear analysis reveals that during El Niño, 5 of the 17 CMIP5 models produce a negative SW response larger than −4 W m−2 K−1 in magnitude, while there are only 2 out of 12 CMIP3 models; during La Niña, 13 of the 17 CMIP5 models versus 9 of the 12 CMIP3 models overestimate the observed positive SW response (Lloyd et al. 2012). The underestimate (overestimate) of αsw results from both the weak (strong) negative SW responses to El Niño and the large (small) (including reverse sign) positive response to La Niña, consistent with their biases of the responses of cloud fraction, LWP, and dynamics to El Niño and La Niña. Furthermore, the different changes in the atmospheric responses to El Niño and La Niña between the coupled and uncoupled models are also linked to the weak ENSO asymmetry simulations. This is consistent with the earlier work by Zhang and Sun (2014), who found that the simulated El Niño events are weaker than observed, whereas the simulated La Niña events are comparable to observations. They further showed that this might be a result of too weak precipitation responses to El Niño.

The errors in the simulations of feedbacks can often be traced to errors in the mean state (Guilyardi et al. 2009a). However, we did not find a significant statistical relationship between the feedbacks and mean state of the SW flux, LWP, cloud fraction, and precipitation, except for 500-hPa vertical velocity (figure not shown). Nonetheless, when compared to the models with strong negative SW feedback, the models with weak SW feedbacks produce “too few and too bright low clouds,” relatively less middle clouds and too much high clouds (Fig. 14), which may explain the too negative low-cloud and less midcloud feedback to El Niño. These low- and midcloud problems might be related to the stratiform processes: cloud fraction schemes (e.g., horizontal homogeneity and vertical maximum overlap), cloud microphysical schemes (underestimate of cloud effective radius or precipitation efficiency), and their tunable parameters (Shonk et al. 2010; Cole et al. 2011; Brient and Bony 2012; Nam et al. 2012). The development of too much high cloud is attributed to the deficiency of too frequent convection occurrence and weak stratiform condensation processes (Dai and Trenberth 2004; Zhang and Sun 2006; Li et al. 2014).

Fig. 14.
Fig. 14.

Vertical cross section of multiannual mean cloud fraction (color shading) and LWC (contours) for the ensemble mean of (a) models (CCSM4, FGOALS-g2, and MIROC5) with strong SW flux feedback, (b) models (except for CCSM4, FGOALS-g2, and MIROC5) with weak SW flux feedback, and (c) the differences (significant at 95%) between (b) and (a) from their historical runs. The contour lines have an interval of 5 × 10−3 g kg−1 in (a),(b) and 3 × 10−3 g kg−1 in (c), where the solid (dashed) lines represent positive (negative) values.

Citation: Journal of Climate 28, 24; 10.1175/JCLI-D-15-0276.1

Acknowledgments

This work was supported by the CAS Strategic Priority Research Program (Grant XDA05110304), the National 973 project (Grant 2015CB954102), and the National Natural Science Foundation of China (Grants 41205079 and 41305040).

REFERENCES

  • Bellenger, H., , E. Guilyardi, , J. Leloup, , M. Lengaigne, , and J. Vialard, 2014: ENSO representation in climate models: From CMIP3 to CMIP5. Climate Dyn., 42, 1999–2018, doi:10.1007/s00382-013-1783-z.

    • Search Google Scholar
    • Export Citation
  • Boyle, J., , and S. A. Klein, 2010: Impact of horizontal resolution on climate model forecasts of tropical precipitation and diabatic heating for the TWP-ICE period. J. Geophys. Res., 115, D23113, doi:10.1029/2010JD014262.

    • Search Google Scholar
    • Export Citation
  • Brient, F., , and S. Bony, 2012: How may low-cloud radiative properties simulated in the current climate influence low-cloud feedbacks under global warming? Geophys. Res. Lett., 39, L20807, doi:10.1029/2012GL053265.

    • Search Google Scholar
    • Export Citation
  • Chen, L., , Y. Yu, , and D.-Z. Sun, 2013: Cloud and water vapor feedbacks to the El Niño warming: Are they still biased in CMIP5 models? J. Climate, 26, 29472961, doi:10.1175/JCLI-D-12-00575.1.

    • Search Google Scholar
    • Export Citation
  • Cole, J., , H. Barker, , N. Loeb, , and K. von Salzen, 2011: Assessing simulated clouds and radiative fluxes using properties of clouds whose tops are exposed to space. J. Climate, 24, 27152727, doi:10.1175/2011JCLI3652.1.

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

  • Dai, A., , and K. E. Trenberth, 2004: The diurnal cycle and its depiction in the Community Climate System Model. J. Climate, 17, 930951, doi:10.1175/1520-0442(2004)017<0930:TDCAID>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., , P. Braconnot, , F. F. Jin, , S. T. Kim, , M. Kolasinski, , T. Li, , and I. Musat, 2009a: Atmosphere feedbacks during ENSO in a coupled GCM with a modified atmospheric convection scheme. J. Climate, 22, 56985718, doi:10.1175/2009JCLI2815.1.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., , A. Wittenberg, , A. Fedorov, , M. Collins, , C. Wang, , A. Capotondi, , G. J. van Oldenborgh, , and T. Stockdale, 2009b: Understanding El Niño in ocean–atmosphere general circulation models: Progress and challenges. Bull. Amer. Meteor. Soc., 90, 325340, doi:10.1175/2008BAMS2387.1.

    • Search Google Scholar
    • Export Citation
  • Ham, Y.-G., , and J.-S. Kug, 2014: Effects of Pacific Intertropical Convergence Zone precipitation bias on ENSO phase transition. Environ. Res. Lett., 9, 064008, doi:10.1088/1748-9326/9/6/064008.

    • Search Google Scholar
    • Export Citation
  • Kim, D., , J.-S. Kug, , I.-S. Kang, , F.-F. Jin, , and A. T. Wittenberg, 2008: Tropical Pacific impacts of convective momentum transport in the SNU coupled GCM. Climate Dyn., 31, 213226, doi:10.1007/s00382-007-0348-4.

    • Search Google Scholar
    • Export Citation
  • Kim, D., , Y.-S. Jang, , D.-H. Kim, , Y.-H. Kim, , M. Watanabe, , F.-F. Jin, , and J.-S. Kug, 2011: El Niño–Southern Oscillation sensitivity to cumulus entrainment in a coupled general circulation model. J. Geophys. Res., 116, D22112, doi:10.1029/2011JD016526.

    • Search Google Scholar
    • Export Citation
  • Kim, S. T., , W. Cai, , F.-F. Jin, , and J.-Y. Yu, 2014: ENSO stability in coupled climate models and its association with mean state. Climate Dyn., 42, 3313–3321, doi:10.1007/s00382-013-1833-6.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., , and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 15871606, doi:10.1175/1520-0442(1993)006,1587:TSCOLS.2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Li, G., , and G. J. Zhang, 2008: Understanding biases in shortwave cloud radiative forcing in the National Center for Atmospheric Research Community Atmosphere Model (CAM3) during El Niño. J. Geophys. Res., 113, D02103, doi:10.1029/2007JD008963.

    • Search Google Scholar
    • Export Citation
  • Li, L. J., , B. Wang, , and G. J. Zhang, 2014: The role of nonconvective condensation processes in response of surface shortwave cloud radiative forcing to El Niño warming. J. Climate, 27, 67216736, doi:10.1175/JCLI-D-13-00632.1.

    • Search Google Scholar
    • Export Citation
  • Lin, J. L., , B. Mapes, , M. Zhang, , and M. Newman, 2004: Stratiform precipitation, vertical heating profiles, and the Madden–Julian oscillation. J. Atmos. Sci., 61, 296309, doi:10.1175/1520-0469(2004)061<0296:SPVHPA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , E. Guilyardi, , H. Weller, , and J. Slingo, 2009: The role of atmosphere feedbacks during ENSO in the CMIP3 models. Atmos. Sci. Lett., 10, 170176, doi:10.1002/asl.227.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , E. Guilyardi, , and H. Weller, 2011: The role of atmosphere feedbacks during ENSO in the CMIP3 models. Part II: Using AMIP runs to understand the heat flux feedback mechanisms. Climate Dyn., 37, 12711292, doi:10.1007/s00382-010-0895-y.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , E. Guilyardi, , and H. Weller, 2012: The role of atmosphere feedbacks during ENSO in the CMIP3 models. Part III: The shortwave feedback. J. Climate, 25, 42754293, doi:10.1175/JCLI-D-11-00178.1.

    • Search Google Scholar
    • Export Citation
  • Ma, H. Y., , S. Xie, , J. S. Boyle, , S. A. Klein, , and Y. Zhang, 2013: Metrics and diagnostics for precipitation-related processes in climate model short-range hindcasts. J. Climate, 26, 15161534, doi:10.1175/JCLI-D-12-00235.1.

    • Search Google Scholar
    • Export Citation
  • Mauritsen, T., and et al. , 2012: Tuning the climate of a global model. J. Adv. Model. Earth Syst., 4, M00A01, doi:10.1029/2012MS000154.

    • 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
  • 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
  • Philander, S., , D. Gu, , G. Lambert, , T. Li, , D. Halpern, , N.-C. Lau, , and R. Pacanowski, 1996: Why the ITCZ is mostly north of the equator. J. Climate, 9, 29582972, doi:10.1175/1520-0442(1996)009<2958:WTIIMN>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., , D. E. Parker, , E. B. Horton, , C. K. Folland, , L. V. Alexander, , D. P. Rowell, , E. C. Kent, , and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Robertson, F. R., , D. E. Fitzjarrald, , and C. D. Kummerow, 2003: Effects of uncertainty in TRMM precipitation radar path integrated attenuation on interannual variations of tropical oceanic rainfall. Geophys. Res. Lett., 30, 1180, doi:10.1029/2002GL016416.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., , and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 22612288, doi:10.1175/1520-0477(1999)080<2261:AIUCFI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and R. A. Houze Jr., 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
  • Schumacher, C., , R. A. Houze Jr., , and I. Kraucunas, 2004: The tropical dynamical response to latent heating estimates derived from the TRMM Precipitation Radar. J. Atmos. Sci., 61, 13411358, doi:10.1175/1520-0469(2004)061<1341:TTDRTL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Seo, K. H., , and W. Q. Wang, 2010: The Madden–Julian oscillation simulated in the NCEP Climate Forecast System model: The importance of stratiform heating. J. Climate, 23, 47704793, doi:10.1175/2010JCLI2983.1.

    • Search Google Scholar
    • Export Citation
  • Shen, X., , Y. Wang, , and X. Li, 2011: Effects of vertical wind shear and cloud radiative processes on responses of rainfall to the large-scale forcing during pre-summer heavy rainfall over southern China. Quart. J. Roy. Meteor. Soc., 137, 236249, doi:10.1002/qj.735.

    • Search Google Scholar
    • Export Citation
  • Shonk, J., , R. Hogan, , G. Mace, , and J. Edwards, 2010: Effect of improving representation of horizontal and vertical cloud structure on the Earth’s global radiation budget. Part I: Review and parametrization. Quart. J. Roy. Meteor. Soc., 136, 11911204, doi:10.1002/qj.647.

    • Search Google Scholar
    • Export Citation
  • Sun, D.-Z., , J. Fasullo, , T. Zhang, , and A. Roubicek, 2003: On the radiative and dynamical feedbacks over the equatorial Pacific cold tongue. J. Climate, 16, 24252432, doi:10.1175/2786.1.

    • Search Google Scholar
    • Export Citation
  • Sun, D.-Z., and et al. , 2006: Radiative and dynamical feedbacks over the equatorial cold tongue: Results from nine atmospheric GCMs. J. Climate, 19, 40594074, doi:10.1175/JCLI3835.1.

    • Search Google Scholar
    • Export Citation
  • Sun, D.-Z., , Y. Yu, , and T. Zhang, 2009: Tropical water vapor and cloud feedbacks in climate models: A further assessment using coupled simulations. J. Climate, 22, 12871304, doi:10.1175/2008JCLI2267.1.

    • Search Google Scholar
    • Export Citation
  • Sundqvist, H., 1978: A parameterization scheme for non-convective condensation including prediction of cloud water content. Quart. J. Roy. Meteor. Soc., 104, 677690, doi:10.1002/qj.49710444110.

    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., , R. J. Stouffer, , and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485498, doi:10.1175/BAMS-D-11-00094.1.

    • Search Google Scholar
    • Export Citation
  • Toniazzo, T., , M. Collins, , and J. Brown, 2008: The variation of ENSO characteristics associated with atmospheric parameter perturbations in a coupled model. Climate Dyn., 30, 643656, doi:10.1007/s00382-007-0313-2.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and et al. , 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 29613012, doi:10.1256/qj.04.176.

  • Vannière, B., , E. Guilyardi, , G. Madec, , F. J. Doblas-Reyes, , and S. Woolnough, 2013: Using seasonal hindcasts to understand the origin of the equatorial cold tongue bias in CGCMs and its impact on ENSO. Climate Dyn., 40, 963981, doi:10.1007/s00382-012-1429-6.

    • Search Google Scholar
    • Export Citation
  • Vannière, B., , E. Guilyardi, , T. Toniazzo, , G. Madec, , and S. Woolnough, 2014: A systematic approach to identify sources of tropical SST errors in coupled models using the adjustment of initialised experiments. Climate Dyn., 43, 22612282, doi:10.1007/s00382-014-2051-6.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , M. Chikira, , Y. Imada, , and M. Kimoto, 2011: Convective control of ENSO simulated in MIROC. J. Climate, 24, 543562, doi:10.1175/2010JCLI3878.1.

    • Search Google Scholar
    • Export Citation
  • Webb, M., , C. Senior, , S. Bony, , and J. J. Morcrette, 2001: Combining ERBE and ISCCP data to assess clouds in the Hadley Centre, ECMWF and LMD atmospheric climate models. Climate Dyn., 17, 905922, doi:10.1007/s003820100157.

    • Search Google Scholar
    • Export Citation
  • Wu, R., , and B. P. Kirtman, 2007: Regimes of seasonal air–sea interaction and implications for performance of forced simulations. Climate Dyn., 29, 393410, doi:10.1007/s00382-007-0246-9.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., 2005: The shape of continents, air-sea interaction, and the rising branch of the Hadley circulation. The Hadley Circulation: Present, Past and Future, H. F. Diaz and R. S. Bradley, Eds., Advances in Global Change Research, Vol. 21, Kluwer Academic Publishers, 121–152.

    • Search Google Scholar
    • Export Citation
  • Xu, K.-M., , and D. A. Randall, 1996: A semiempirical cloudiness parameterization for use in climate models. J. Atmos. Sci., 53, 30843102, doi:10.1175/1520-0469(1996)053<3084:ASCPFU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zebiak, S. E., , and M. A. Cane, 1987: A model El Niño–Southern Oscillation. Mon. Wea. Rev., 115, 22622278, doi:10.1175/1520-0493(1987)115<2262:AMENO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, M., and et al. , 2005: Comparing clouds and their seasonal variations in 10 atmospheric general circulation models with satellite measurements. J. Geophys. Res., 110, D15S02, doi:10.1029/2004JD005021.

    • Search Google Scholar
    • Export Citation
  • Zhang, T., , and D.-Z. Sun, 2006: Response of water vapor and clouds to El Niño warming in three National Center for Atmospheric Research atmospheric models. J. Geophys. Res., 111, D17103, doi:10.1029/2005JD006700.

    • Search Google Scholar
    • Export Citation
  • Zhang, T., , and D.-Z. Sun, 2014: ENSO asymmetry in CMIP5 models. J. Climate, 27, 40704093, doi:10.1175/JCLI-D-13-00454.1.

Save