• Babić, K., B. Adler, N. Kalthoff, H. Andersen, C. Dione, F. Lohou, M. Lothon, and X. Pedruzo-Bagazgoitia, 2019: The observed diurnal cycle of low-level stratus clouds over southern West Africa: A case study. Atmos. Chem. Phys., 19, 12811299, https://doi.org/10.5194/acp-19-1281-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bony, S., and J.-L. Dufrense, 2005: Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models. Geophys. Res. Lett., 32, L20806, https://doi.org/10.1029/2005GL023851.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bretherton, C. S., P. N. Blossey, and C. R. Jones, 2013: Mechanisms of marine low cloud sensitivity to idealized climate perturbations: A single-LES exploration extending the CGILS cases. J. Adv. Model. Earth Syst., 5, 316337, https://doi.org/10.1002/jame.20019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brient, F., and S. Bony, 2013: Interpretation of the positive low-cloud feedback predicted by a climate model under global warming. Climate Dyn., 40, 24152431, https://doi.org/10.1007/s00382-011-1279-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • de Szoeke, S. P., K. L. Verlinden, S. E. Yuter, and D. B. Mechem, 2016: The time scales of variability of marine low clouds. J. Climate, 29, 64636481, https://doi.org/10.1175/JCLI-D-15-0460.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ek, M. B., and L. Mahrt, 1994: Daytime evolution of relative humidity at the boundary layer top. Mon. Wea. Rev., 122, 27092721, https://doi.org/10.1175/1520-0493(1994)122<2709:DEORHA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., 1985: Sea surface temperature anomalies, planetary waves, and air-sea feedback in the middle latitudes. Rev. Geophys., 23, 357390, https://doi.org/10.1029/RG023i004p00357.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., and K. Hasselmann, 1977: Stochastic climate models, Part II Application to sea-surface temperature anomalies and thermocline variability. Tellus, 29, 289305, https://doi.org/10.3402/TELLUSA.V29I4.11362.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., and R. W. Reynolds, 1983: Testing a dynamical model for mid-latitude sea surface temperature anomalies. J. Phys. Oceanogr., 13, 11311145, https://doi.org/10.1175/1520-0485(1983)013<1131:TADMFM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., N. Sennéchael, Y.-O. Kwon, and M. A. Alexander, 2011: Influence of the meridional shifts of the Kuroshio and the Oyashio Extensions on the atmospheric circulation. J. Climate, 24, 762777, https://doi.org/10.1175/2010JCLI3731.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 54195454, https://doi.org/10.1175/JCLI-D-16-0758.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hartmann, D. L., and D. A. Short, 1980: On the use of earth radiation budget statistics for studies of clouds and climate. J. Atmos. Sci., 37, 12331250, https://doi.org/10.1175/1520-0469(1980)037<1233:OTUOER>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Johnson, G. C., S. Schmidtko, and J. M. Lyman, 2012: Relative contributions of temperature and salinity to seasonal mixed layer density changes and horizontal density gradients. J. Geophys. Res., 117, C04015, https://doi.org/10.1029/2011JC007651.

    • Search Google Scholar
    • Export Citation
  • Kawamura, R., T. Murakami, and B. Wang, 1996: Tropical and mid-latitude 45-day perturbations over the Western Pacific during the northern summer. J. Meteor. Soc. Japan, 74, 867890, https://doi.org/10.2151/JMSJ1965.74.6_867.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Klein, S. A., 1997: Synoptic variability of low-cloud properties and meteorological parameters in the subtropical trade wind boundary layer. J. Climate, 10, 20182039, https://doi.org/10.1175/1520-0442(1997)010<2018:SVOLCP>2.0.CO;2.

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

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Klein, S. A., D. L. Hartmann, and J. R. Norris, 1995: On the relationships among low-cloud structure, sea surface temperature, and atmospheric circulation in the summertime Northeast Pacific. J. Climate, 8, 11401155, https://doi.org/10.1175/1520-0442(1995)008<1140:OTRALC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Klein, S. A., A. Hall, J. R. Norris, and R. Pincus, 2017: Low-cloud feedbacks from cloud-controlling factors: A review. Surv. Geophys., 38, 13071329, https://doi.org/10.1007/s10712-017-9433-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 548, https://doi.org/10.2151/JMSJ.2015-001.

    • Crossref
    • 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, https://doi.org/10.1175/JCLI-D-11-00478.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mace, G. G., Q. Zhang, M. Vaughan, R. Marchand, G. Stephens, C. Trepte, and D. Winker, 2009: A description of hydrometeor layer occurrence statistics derived from the first year of merged CloudSat and CALIPSO data. J. Geophys. Res., 114, D00A26, https://doi.org/10.1029/2007JD009755.

    • Search Google Scholar
    • Export Citation
  • Mauger, G. S., and J. R. Norris, 2010: Assessing the impact of meteorological history on subtropical cloud fraction. J. Climate, 23, 29262940, https://doi.org/10.1175/2010JCLI3272.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miyamoto, A., H. Nakamura, and T. Miyasaka, 2018: Influence of the subtropical high and storm track on low-cloud fraction and its seasonality over the South Indian Ocean. J. Climate, 31, 40174039, https://doi.org/10.1175/JCLI-D-17-0229.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moisan, J. R., and P. P. Niiler, 1998: The seasonal heat budget of the North Pacific: Net heat flux and heat storage rates (1950–1990). J. Phys. Oceanogr., 28, 401421, https://doi.org/10.1175/1520-0485(1998)028<0401:TSHBOT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mori, M., and M. Watanabe, 2008: The growth and triggering mechanisms of the PNA: A MJO-PNA coherence. J. Meteor. Soc. Japan, 86, 213236, https://doi.org/10.2151/jmsj.86.213.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Morioka, Y., T. Tozuka, and T. Yamagata, 2010: Climate variability in the southern Indian Ocean as revealed by self-organizing maps. Climate Dyn., 35, 10591072, https://doi.org/10.1007/s00382-010-0843-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Myers, T. A., and J. R. Norris, 2016: Reducing the uncertainty in subtropical cloud feedback. Geophys. Res. Lett., 43, 21442148, https://doi.org/10.1002/2015GL067416.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Myers, T. A., C. R. Mechoso, and M. J. DeFlorio, 2018: Coupling between marine boundary layer clouds and summer-to-summer sea surface temperature variability over the North Atlantic and Pacific. Climate Dyn., 50, 955969, https://doi.org/10.1007/s00382-017-3651-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Norris, J. R., and C. B. Leovy, 1994: Interannual variability in stratiform cloudiness and sea surface temperature. J. Climate, 7, 19151925, https://doi.org/10.1175/1520-0442(1994)007<1915:IVISCA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Norris, J. R., and S. F. Iacobellis, 2005: North Pacific cloud feedbacks inferred from synoptic-scale dynamic and thermodynamic relationships. J. Climate, 18, 48624878, https://doi.org/10.1175/JCLI3558.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ogawa, F., and T. Spengler, 2019: Prevailing surface wind direction during air–sea heat exchange. J. Climate, 32, 56015617, https://doi.org/10.1175/JCLI-D-18-0752.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pan, L.-L., and T. Li, 2008: Interactions between the tropical ISO and midlatitude low-frequency flow. Climate Dyn., 31, 375388, https://doi.org/10.1007/s00382-007-0272-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7, 952956, https://doi.org/10.1175/1520-0485(1977)007<0952:IMITUO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Platnick, S., P. Hubanks, K. Meyer, and M. D. King, 2015: MYD08 _ D3 - MODIS/Aqua Aerosol Cloud Water Vapor Ozone Daily L3 Global 1Deg CMG. NASA MODIS Adaptive Processing System, Goddard Space Flight Center, accessed 21 February 2017, https://doi.org/10.5067/MODIS/MYD08_D3.061.

    • Crossref
    • Export Citation
  • Qiu, B., and K. A. Kelly, 1993: Upper-ocean heat balance in the Kuroshio Extension region. J. Phys. Oceanogr., 23, 20272041, https://doi.org/10.1175/1520-0485(1993)023<2027:UOHBIT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Qu, X., A. Hall, S. A. Klein, and P. M. Caldwell, 2014: On the spread of changes in marine low cloud cover in climate model simulations of the 21st century. Climate Dyn., 42, 26032626, https://doi.org/10.1007/s00382-013-1945-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Qu, X., A. Hall, S. A. Klein, and A. M. DeAngelis, 2015: Positive tropical marine low-cloud cover feedback inferred from cloud-controlling factors. Geophys. Res. Lett., 42, 77677775, https://doi.org/10.1002/2015GL065627.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., N. A. Rayner, T. M. Smith, D. C. Stokes, and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 16091625, https://doi.org/10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, L., T. Li, and T. Zhou, 2012: Intraseasonal SST variability and air–sea interaction over the Kuroshio Extension region during boreal summer. J. Climate, 25, 16191634, https://doi.org/10.1175/JCLI-D-11-00109.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, L., T. Li, T. Zhou, and X. Rong, 2013: Origin of the intraseasonal variability over the North Pacific in boreal summer. J. Climate, 26, 12111229, https://doi.org/10.1175/JCLI-D-11-00704.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., B. R. Barkstrom, E. F. Harrison, R. B. Lee III, G. L. 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, https://doi.org/10.1175/1520-0477(1996)077<0853:CATERE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wood, R., 2012: Stratocumulus clouds. Mon. Wea. Rev., 140, 23732423, https://doi.org/10.1175/MWR-D-11-00121.1.

  • Wood, R., and C. S. Bretherton, 2006: On the relationship between stratiform low cloud cover and lower-tropospheric stability. J. Climate, 19, 64256432, https://doi.org/10.1175/JCLI3988.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, H., S. Xie, and Y. Wang, 2005: Subseasonal variability of the southeast Pacific stratus cloud deck. J. Climate, 18, 131142, https://doi.org/10.1175/JCLI3250.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yasunari, T., 1979: Cloudiness fluctuations associated with the Northern Hemisphere summer monsoon. J. Meteor. Soc. Japan, 57, 227242, https://doi.org/10.2151/JMSJ1965.57.3_227.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yu, L., X. Jin, and A. R. Weller, 2008: Multidecade global flux datasets from the Objectively Analyzed Air-Sea Fluxes (OAFlux) Project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. OAFlux Project Tech. Rep. OA-2008-01, Woods Hole Oceanographic Institution, Woods Hole, MA, 64 pp.

  • Zelinka, M. D., K. M. Grise, S. A. Klein, C. Zhou, A. M. DeAngelis, and M. W. Christensen, 2018: Drivers of the low-cloud response to poleward jet shifts in the North Pacific in observations and models. J. Climate, 31, 79257947, https://doi.org/10.1175/JCLI-D-18-0114.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Climatology of low-level cloud cover (LCC) derived from the Moderate Resolution Imaging Spectroradiometer (MODIS; shaded; %) and sea surface temperature (SST) derived from OISST [contours; °C; contour interval (CI) = 2°C; thick contour (TC) = 16°C] in the North Pacific in July from 2003 to 2016. (b) Standard deviation of deseasonal daily LCC data (shaded; %) and SST (contours; °C; CI = 0.5°C; TC = 1.5°C). The black rectangle shows the region where the standard deviation of LCC is large and the target meridional band from 30° to 40°N used in the present study. (c) Mean seasonal variation of the subseasonal standard deviation of LCC (shaded; %) and SST (contours; °C; CI = 0.3°C) for the target meridional band. The standard deviation for each variable on day i is calculated as the standard deviation of LCC and SST from day [i − 45] to day [i + 45] using a method similar to that of Wang et al. (2012). The two gray horizontal dashed lines show the season analyzed in the present study (JJASO). The black rectangle shows a target longitudinal band from 165° to 175°E where the standard deviation of LCC and SST are both large.

  • View in gallery

    (a) Mean power spectrum of the area-mean daily LCC in the target region (30°–40°N, 165°–175°E). The mean power spectrum was calculated as the mean value of all spectra for each JJASO season from 2003 to 2016. Solid and dashed lines indicate red noise and the 95% significance level, respectively. Green dots indicate when the power exceeds the 95% significance level line. (b) Mean squared coherency between LCC and SST (blue) and between LCC and surface horizontal temperature advection Tadv (orange). The horizontal dashed line indicates the 95% significance level determined by a chi-squared distribution.

  • View in gallery

    Histograms of (a) cycle period and (b) amplitude for each cycle of the intraseasonal variability of LCC in the target region. The total number of cases is 46 for JJASO from 2003 to 2016. The bin sizes of each histogram are 4 days in (a) and 2% in (b).

  • View in gallery

    Phase composite of the horizontal distribution of the LCC anomaly (shaded; %) and horizontal wind speed at 1000 hPa (vectors; m s−1). The black rectangle denotes the area for which the LCC anomaly index was calculated. The maximum LCC anomaly appears at a phase of 0°, while the minimum LCC anomaly occurs at phases of −180° and 180°.

  • View in gallery

    As in Fig. 4, but for SST (shaded; °C) and EIS (contours; CI = 0.1 K; TC = 0 K).

  • View in gallery

    Evolution of composite anomalies of LCC (black), SST (blue), Tadv (green), and EIS (red) averaged over the target area.

  • View in gallery

    Phase composite of a horizontal map of the anomalies of geopotential height (GPH) at 500 hPa (shaded; m), GPH at 250 hPa (contours; m; CI = 7.5 m), and horizontal wind speed at 500 hPa (vectors; m s−1). The black rectangle denotes the target area where the LCC anomaly index was calculated. The green outlined star at each phase indicates the position of minimum of GPH anomaly (i.e., cyclonic anomaly) at 500 hPa.

  • View in gallery

    Evolution of the vertical distributions of anomalies of relative humidity (RH; shaded; %), specific humidity q (contours; g kg−1; CI = 0.1 g kg−1; TC = 0 g kg−1), and vertical pressure velocity ω (vectors; hPa day−1). Red and blue arrows indicate updrafts and downdrafts, respectively.

  • View in gallery

    As in Fig. 8, but for anomalies of (a) RH tendency (dRH/dt; shaded; %/day) and RH (contours; %; CI = 1%; TC = 0%). The dRH/dt anomalies by (b) total adiabatic, (c) total diabatic, (d) horizontal adiabatic, and (e) vertical adiabatic processes. (f)–(h) The anomalies by horizontal adiabatic, vertical adiabatic, diabatic processes related to change in temperature (T-change; % day−1). (i)–(k) As in (f)–(h), but the processes related to change in specific humidity (q-change; % day−1). The horizontal dashed line in each panel indicates the pressure level at 700 hPa. Note that color spacing is not uniform in (b)–(k).

  • View in gallery

    Climatological seasonal-mean state in June–July–August (JJA) of a horizontal map of RH at 1000 hPa derived from ERA-Interim (%; CI = 2%) from 2003 to 2016.

  • View in gallery

    (a) Evolution of anomalies of SST tendency (dSST/dt; black bar) and surface heat flux term (red circles). (b) As in (a), but for anomalies of the total heat flux term (red bar), shortwave radiation (SW), longwave radiation (LW), sensible heat (SH), and latent heat (LH).

  • View in gallery

    (a) Evolution of anomalies of specific humidity at the sea surface qs (blue), specific humidity at the near surface qa (red), and wind speed at 10 m (WS10; green). (b) As in (a), but for anomalies of latent heat flux (LH; orange shading), LH by qs change [LH(qs); blue], LH by qa change [LH(qa); red], and LH by WS10 change [LH(WS10); green].

  • View in gallery

    Lag correlation coefficient between LCC and SST (blue) and between LCC and Tadv (green) derived from an unfiltered daily dataset for JJASO for 2003–16 in (a) the target region (30°–40°N, 165°–175°E) and (b) the northeastern subtropical Pacific region (20°–30°N, 140°–150°W). Lag correlation coefficients were calculated as mean values over 14 years. Horizontal dashed and solid lines indicate that the minimum correlation coefficient was exceeded at the 95% and 99% significance levels, respectively, calculated with an effective degree of freedom (Neff) of 152/50 × 12 = 36. Circles indicate when the coefficient at a certain lag day exceeds the 95% significance level.

  • View in gallery

    Horizontal map of correlation coefficient (a),(c) between LCC and SST and (b),(d) LCC and Tadv derived from (top) a deseasonal daily dataset and (bottom) filtered datasets of intraseasonal time scale (20–100 days) for JJASO for 2003–16. Correlation coefficients were calculated as mean values over 14 years. Numbers in grid box indicate the lasting (leading) days of significant correlation coefficients of SST (Tadv) based on the days of a maximum of the LCC anomaly with a 95% confidence level. Grid boxes without the number indicate any significant correlation coefficients do not appear in the day range for the calculation of the lag correlation (from −40 to 40 days).

  • View in gallery

    Cloud occurrence profiles based on estimated LCC with (a) random overlap assumption and (b) no assumption in the target region (30°–40°N, 165°–175°E) from 2007 to 2010 JJA, using daily data of MODIS grid data and level-2 profile data of the 2B-GEOPROF-lidar product.

  • View in gallery

    As in Fig. 6, but for the results based on LCC anomaly calculated without random overlap assumption.

  • View in gallery

    (a) Composited mean of each term in temperature budget integrated from 1000 to 850 hPa in the target region. (b) The phase mean of each term in the cooling phase (from phase −150° to −30°) is also displayed. Colors in the line and bar are corresponded within the two panels (black: total tendency, red: zonal advection term, blue: meridional advection term, green: vertical advection term, purple: adiabatic heating term associated with vertical pressure velocity, gray: diabatic heating term calculated as a residual).

All Time Past Year Past 30 Days
Abstract Views 133 125 0
Full Text Views 116 103 20
PDF Downloads 145 126 25

Air–Sea Interactions among Oceanic Low-Level Cloud, Sea Surface Temperature, and Atmospheric Circulation on an Intraseasonal Time Scale in the Summertime North Pacific Based on Satellite Data Analysis

View More View Less
  • 1 Center for Atmospheric and Oceanic Studies, Graduate School of Science, Tohoku University, Sendai, Japan
© Get Permissions
Free access

Abstract

Low-level cloud plays a key role in modulating air–sea interaction processes and sea surface temperature (SST) variability. The present study investigated the evolution process of oceanic low-level cloud cover (LCC) and related air–sea interaction processes on an intraseasonal time scale in the summertime (June–October) North Pacific (30°–40°N, 165°–175°E) based on satellite observational and reanalysis datasets from 2003 to 2016. The intraseasonal time scale (20–100 days) is dominant not only for the LCC, but also for LCC controlling factors, that is, SST, estimated inversion strength (EIS), and horizontal temperature advection (Tadv). To reveal the lead–lag relationship among these variables, we conducted phase composite analysis with a bandpass filter based on the intraseasonal variability (ISV) of LCC. It suggests that ISV of LCC leads to that of SST and that horizontal dry–cold advection from the poleward region leads to increasing LCC and decreasing SST. The increasing LCC corresponds to a positive relative humidity (RH) anomaly in the lower troposphere, which is due to adiabatic cooling with shallow convection, vertical moisture advection, and meridional RH advection associated with the anomalous cold Tadv. Heat budget analysis of the ocean mixed layer suggests the importance of anomalous dry–cold advection for cooling SST, not only via enhanced latent heat release but also via decreased downward shortwave radiation at the sea surface according to cloud radiative effect with a positive LCC anomaly. Determining the detailed lead–lag relationship between LCC and its controlling factor is a good approach to understand mechanisms of the local processes of both low-level cloud evolution and air–sea interaction.

Corresponding author: Naoya Takahashi, naoya.takahashi.d7@dc.tohoku.ac.jp

Abstract

Low-level cloud plays a key role in modulating air–sea interaction processes and sea surface temperature (SST) variability. The present study investigated the evolution process of oceanic low-level cloud cover (LCC) and related air–sea interaction processes on an intraseasonal time scale in the summertime (June–October) North Pacific (30°–40°N, 165°–175°E) based on satellite observational and reanalysis datasets from 2003 to 2016. The intraseasonal time scale (20–100 days) is dominant not only for the LCC, but also for LCC controlling factors, that is, SST, estimated inversion strength (EIS), and horizontal temperature advection (Tadv). To reveal the lead–lag relationship among these variables, we conducted phase composite analysis with a bandpass filter based on the intraseasonal variability (ISV) of LCC. It suggests that ISV of LCC leads to that of SST and that horizontal dry–cold advection from the poleward region leads to increasing LCC and decreasing SST. The increasing LCC corresponds to a positive relative humidity (RH) anomaly in the lower troposphere, which is due to adiabatic cooling with shallow convection, vertical moisture advection, and meridional RH advection associated with the anomalous cold Tadv. Heat budget analysis of the ocean mixed layer suggests the importance of anomalous dry–cold advection for cooling SST, not only via enhanced latent heat release but also via decreased downward shortwave radiation at the sea surface according to cloud radiative effect with a positive LCC anomaly. Determining the detailed lead–lag relationship between LCC and its controlling factor is a good approach to understand mechanisms of the local processes of both low-level cloud evolution and air–sea interaction.

Corresponding author: Naoya Takahashi, naoya.takahashi.d7@dc.tohoku.ac.jp

1. Introduction

Oceanic low-level cloud in the summertime North Pacific (NP) plays an important role in air–sea interaction processes because it has a strong cooling shortwave cloud radiative effect (SWCRE) at the sea surface and the low-level cloud properties are sensitive to sea surface state. The SWCRE of low-level cloud is derived from its high albedo and warm cloud top temperature (Hartmann and Short 1980). Low-level cloud cover (LCC) is an important parameter to determine SWCRE at the sea surface. Despite the importance of LCC and SWCRE, the projection quality of these parameters associated with climate variability in general circulation models (GCMs) is a major source of uncertainty in future climate projections (Brient and Bony 2013; Bony and Dufrense 2005; Qu et al. 2015, 2014; Myers and Norris 2016). One of the reasons for the large uncertainty in GCMs is a difficulty in representing of low-level cloud with turbulence within the boundary layer, cloud–radiative interaction, and cloud microphysical process in a coarse grid resolution. Additionally the evolution process of LCC is so complex because it is modulated by many environmental factors, which are called LCC controlling factors, that is, estimated inversion strength (EIS), horizontal temperature advection (Tadv), sea surface temperature (SST), and so on (see Klein et al. 2017 for a summary). In previous studies, the accuracy of a projection of the sensitivity of LCC and SWCRE to the controlling factors in GCMs was investigated and compared with observational results using a multilinear regression method to estimate the interannual variabilities of the monthly mean and annual mean states (Myers and Norris 2016; Qu et al. 2014, 2015). Previous studies have also investigated the relative contributions of the LCC response to each controlling factor. For example, EIS is the dominant LCC controlling factors on an interannual time scale. SST is also known as one of important LCC controlling factors and strongly connected with LCC variability through a positive feedback loop (Norris and Leovy 1994).

Time scale is one of key points to consider the evolution process of low-level cloud cover or its mechanism. For instance, de Szoeke et al. (2016) focused on the marine boundary layer cloud in the subtropics and showed the spectrum of LCC, SST, and EIS. Spectrum of LCC based on 6-hourly data, indicated by the blue line, has the strongest peak at diurnal cycle and secondary peaks at synoptic time scale (about 3–7 days). On the other hand, SST spectrum based on daily data has the strongest peak at annual cycle, which is much longer than that in LCC. Difference in typical time scales of cloud and SST is derived from their own specific heat capacities. Kubar et al. (2012) examined the correlation between low-level cloud fraction and SST in terms of time-scale dependence, using running mean filters with different time length window [e.g., 2 day (daily) to 90 day (seasonal)]. They showed that correlation coefficients dramatically increase with time from 1 day to approximately 2 weeks, then nearly stay at high value when passing after that point. It implies that atmospheric variability is more dominant to control the low-level cloud properties than SST variability on the synoptic time scale shorter than two weeks (i.e., synoptic atmospheric disturbances).

On a longer time scale (e.g., interannual time scale), SST is recognized as a one of LCC controlling factor and driver of local LCC variability through a stratification of boundary and suppression of entrainment process of dry-air from free troposphere (Bretherton et al. 2013). However, on short time scales (e.g., diurnal cycles and synoptic disturbance), the causal relationship between low-level cloud and SST might change. As described above, low-level cloud variability on synoptic time scale tends to be modulated by atmospheric variability rather than SST variability because typical time scale of SST is much longer than that of low-level cloud properties. Xu et al. (2005) investigated the subseasonal variability of cloud liquid water in the southeastern Pacific using satellite, buoy observation, and reanalysis datasets, and conducted a lead–lag composite analysis of low-level cloud properties and LCC controlling factors. They concluded that the low-level cloud variability is caused by changes in atmospheric circulation rather than by the underlying oceanic state and revealed an important role of cold advection in modulating low-level cloud variability. Other studies also reported the importance of horizontal Tadv near the sea surface for controlling LCC variation using both satellite and ship-based observations (Klein 1997; Norris and Iacobellis 2005; Xu et al. 2005; Mauger and Norris 2010; Miyamoto et al. 2018; Zelinka et al. 2018). Norris and Iacobellis (2005) investigated the horizontal distribution of the amount of each cloud type associated with the synoptic variability of Tadv in the NP. Their results showed that warm advection in summertime NP enhances optically thick and high-level cloud, fog, and stratus, as observed by satellite and ship-based observation; in contrast, the cold advection produces favorable conditions for low-level cloud, in particular stratocumulus and cumulus. The mechanism of LCC enhancement with cold advection over a warm sea surface is related with the processes that destabilizes lower atmosphere near the sea surface and enhances turbulent heat flux from ocean to atmosphere, meaning that advected cold air above the warm sea surface leads to unstable conditions in the marine boundary layer, and this destabilization can create stratocumulus or cumulus (Wood 2012). Therefore, horizontal cold advection is known to be a key parameter that enhances turbulent heat flux and the formation of low-level cloud in the subtropical and midlatitude ocean, especially in summertime. However, uncertainty remains regarding the causal relationship among low-level cloud, horizontal temperature advection, and SST variabilities on intermediate time scales (i.e., intraseasonal time scales).

Intraseasonal time scale is intermediate time scale between the typical time scales of cloud (e.g., synoptic time scale) and SST (e.g., seasonal to annual time scale). Thus, it is important to reveal relative contribution of atmospheric variability or oceanic process for trigger of feedback loop between LCC and SST. The intraseasonal variability (ISV) of atmospheric circulation is a prominent phenomenon, especially in summertime NP. Many previous studies observed the oscillation or variability of atmospheric circulation or precipitation on an intraseasonal scale and investigated the detailed mechanism of such a variability in summertime NP (Yasunari 1979; Kawamura et al. 1996; Wang et al. 2012, 2013). Cloud cover also varies with the ISV of atmospheric circulation related to remote forcing from the tropics and internal dynamics in the midlatitudes. Wang et al. (2012) investigated the ISV of the SST anomaly (SSTA) in the Kuroshio Extension region on an intraseasonal time scale as well as air–sea interaction using reanalysis and satellite observational datasets. They showed that the ISV of the large-scale atmospheric circulation anomaly induces an SSTA due to changes in the surface heat flux, including radiative and turbulent heat flux. On the other hand, the SSTA also induces the atmospheric circulation anomaly, therefore implying the two-way interactions, which in turn shows the importance of local air–sea interaction for the ISV of the SSTA. Wang et al. (2013) revealed the cause of the ISV of a large-scale atmospheric circulation anomaly in summertime NP from a different perspective, and found a two-way interaction between ISV and synoptic eddies in summertime NP. Therefore, there is a possibility that the SSTA plays an active role not only in atmospheric circulation but also in the evolution of low-level cloud on an intraseasonal time scale, similar to what occurs on an interannual time scale.

Here, we examined the air–sea interaction process, especially LCC versus SST, on an intraseasonal time scale, considering the time scale of LCC evolution. Additionally, lead–lag composite analysis was conducted based on the local LCC anomaly (LCCA) in order to investigate not only the causal relationship between LCC and LCC controlling factors, but also the evolution process of LCC with changing humidity field. One important question addressed by the present study is the following: “Which is the dominant controlling factor for the ISV of LCC, atmospheric circulation or local SST?”

The remainder of this paper is organized as follows. In section 2, the observational dataset used in the present study is described. This section also explains the filtering method used to extract the ISV of LCC and the phase composite analysis method used to describe the evolution process of LCC and other LCC controlling factors. Section 3 gives the results of the observed LCC evolution and air–sea interaction associated with the ISV of LCC. Section 3 is divided into four subsections: 1) dominant time scale of LCC variability in summertime western NP based on spectral analysis; 2) ISV of LCC and cloud controlling factors; 3) the mechanism of LCC evolution and humidity field; and 4) two-way interaction process associated with the ISV of LCC, especially for atmospheric forcing into oceanic mixed layer variability. The active role of the ocean in the properties of low-level cloud is discussed in section 4. A summary and the rest of the discussion are also presented in section 4.

2. Data and methods

a. Data

The L3 gridded cloud product (MYD08_D3) from the latest Moderate Resolution Imaging Spectroradiometer (MODIS) release (version 6) from MODIS instruments on board the Aqua satellite platform (Platnick et al. 2015) was used to estimate LCC. The daily mean MODIS dataset is produced by twice-daily observation at 0130 and 1330 local time. The daily cloud cover of each type in a 1° × 1° grid is calculated using the histogram of cloud top pressure and the number of observational grid points in the grid. Estimation of the LCC using passive sensors is difficult due to the problem of cloud overlapping, that is, the obscuring of mid- and high-level clouds. Therefore, to avoid the overlapping problem, the actual LCC was calculated assuming random overlap as LCC = fL/(1 − fMfH), where fL, fM, and fH are the fraction of the scene covered by each cloud type whose top is low [680 hPa ≤ cloud top pressure (CTP) < 1000 hPa], middle (440 hPa ≤ CTP < 680 hPa), and high altitude (CTP < 440 hPa), respectively. The missing LCC values for a certain grid point for daily data were linearly interpolated using the LCC values one day before and after the day of interest. The period from 2003 to 2016 was analyzed because all of the datasets described below were available for this time period.

Validity of the random overlap assumption applied to daily mean dataset of low-level cloud coverage from passive sensor is still needed to assess because daily mean LCC estimated with this assumption strongly depends on fM and fH. Thus, the 2B-GEOPROF-lidar product (Mace et al. 2009) derived from CloudSat and CALIPSO were used to confirm the validity of the random overlap assumption. These satellites follow the same orbits as Aqua satellite, which constitutes the afternoon satellite constellation (A-Train). They can measure the vertical distribution of cloud along the satellite track by their active sensors (vertical resolution of 240 m with a 1.4 km cross track × 1.8 km along the track footprint). To evaluate the cloud vertical distribution for estimated cloud fraction by MODIS, we collocated each profile of 2B-GEOPROF-lidar with each pixel of the MODIS product. The detailed result is described in appendix A. It suggested that the random overlap assumption is needed to avoid the bias derived from obscuration by mid- and high-level cloud cover. Our conclusion, however, does not strongly depend on its overlap assumption.

The National Oceanic and Atmospheric Administration (NOAA) optimum interpolation (OI) objectively analyzed SST version 2 (OISST; Reynolds et al. 2002) was used to analyze the variability of SST. The spatial and temporal resolution of OISST are 0.25° and daily, respectively. To investigate the meteorological field associated with the ISV of LCC and SST, we used ERA-Interim reanalysis data (Dee et al. 2011) with a horizontal resolution of 0.75° and 37 vertical levels. Using the 6-hourly ERA-Interim dataset, we obtained daily mean data for each variable. There are many other atmospheric reanalysis datasets [e.g., MERRA2; Gelaro et al. (2017); JRA-55; Kobayashi et al. (2015)], but we choose the ERA-Interim dataset for investigating humidity field in the atmospheric boundary layer because bias of low-level cloud cover in the summertime North Pacific is the smallest compared with the other reanalysis datasets while we focused on the monthly mean state (not shown). Daily data of temperature at 700 and 1000 hPa, horizontal wind speed at 1000 hPa, and sea level pressure were used to calculate the EIS and horizontal temperature advection, which was used as a proxy for LCC controlling factors. EIS was calculated following the method of Wood and Bretherton (2006). Additionally, Tadv was calculated as the product of the horizontal gradient of SST obtained from OISST and the horizontal wind speed at 1000 hPa obtained from ERA-Interim.

For the heat flux datasets [shortwave radiation (SW), longwave radiation (LW), sensible heat, and latent heat] at the sea surface, two observational datasets were used in this study. To estimate daily mean radiative fluxes (SW and LW), we used the Clouds and the Earth’s Radiant Energy System (CERES) Synoptic Radiative Fluxes and Clouds (SYN) product edition 4.0 (Ed4A) (Wielicki et al. 1996). The variables were calculated by a radiative transfer model initialized using satellite-based cloud and aerosol and meteorological assimilation data from reanalysis. The variables were also constrained by observed top-of-the-atmosphere radiative fluxes. The objectively analyzed air–sea fluxes (OAFlux) project at the Woods Hole Oceanographic Institution (WHOI) provided the global ocean-surface heat flux products from 1955 to present (Yu et al. 2008). We used the daily mean turbulent fluxes dataset. OAFlux products are constructed not from a single data source, but from an optimal blending of satellite retrievals and three atmospheric reanalysis datasets. The horizontal resolution of CERES and OAFlux product is 1° in longitude and latitude.

b. Methods

To extract the intraseasonal (20–100 days) signal from the original time series of each variable, we applied a Lanczos bandpass filter to the time series of all variables from 2003 to 2016. Before the filtering process, we removed the seasonal cycle signal from the original time series, which is the climatological mean value on each calendar day.

We applied a composite phase analysis method to the filtered LCCA in order to investigate how the LCCA and the controlling factors of LCC evolve within one cycle in the ISV of LCC. A based index which was used in the composite analysis was calculated as the area-mean LCCA in the target region (30°–40°N, 165°–175°E), where the standard deviation of the LCC is the largest in the NP. The other way to describe the evolution of LCCA is the lagged-day composite analysis method. However, this method is not suitable when the filtering time-range is broad (20–100 days), and a large error could occur for far days from a lag of 0. Then, we applied a composite phase analysis technique to the LCCA index in a similar way to SSTA in Wang et al. (2012). In the composite phase analysis method, each cycle of the ISV of LCC is divided into 12 phases with an interval of 30° (i.e., −180°, −150°, −120°, …, 150°, 180°). A phase of 0° corresponds to the time when the LCCA reaches a maximum value, and phases of −180° and 180° correspond to the time when the LCCA reaches a minimum value. The phase range from −180° to 0° (from 0° to 180°) can be defined as the “development” (“dissipation”) stage of the LCCA in the target region. To find the days with other phases, linear interpolation was conducted using the days with phases of −180°, 0°, and 180°. The composite cases include only days when the maximum and minimum LCCA exceed one standard deviation.

The target region was chosen because the variability of both LCC and SST on an intraseasonal time scale is exceptionally large in summertime NP. Figures 1a and 1b show the horizontal distributions of mean climatological conditions and the standard deviations of LCC and SST in summer NP (July) from 2003 to 2016. The standard deviation of the unfiltered daily LCC is large in the meridional band over 30°–40°N and 150°E–150°W (shown by the black rectangle in Fig. 1b); in contrast, the standard deviation in the meridional band over 40°–50°N and 150°E–150°W is relatively small. This difference in standard deviation is likely to correspond to the distribution of the climatological mean LCC shown in Fig. 1a, meaning that the standard deviation of the mean LCC in summertime is large. Figure 1c shows the mean seasonal variation of the subseasonal standard deviation of LCC (shaded) and SST (contours) in the southern flank of the SST frontal zone where meridional gradient of SST is strong around 40°N. The subseasonal standard deviation of each variable was calculated as the standard deviation for the period between 45 days before and 45 days after the day of interest. The degree of variability was found to have clear seasonality, with the variability being largest in summertime (Fig. 1c). In the present study, we define summertime as June–October (JJASO), the time when the LCC variability is largest in the midlatitude NP. Additionally, the standard deviation of SST is also large in the southern flank of the front and in summertime NP, although its peak of is located just over the SST front (Figs. 1b,c). It is because the local interaction between low-level cloud and SST is strong in summertime NP compared with wintertime. For the following analysis, we selected a target region where the variabilities of LCC and SST in summertime are especially large (30°–40°N, 165°–175°E; shown by the black rectangle in Fig. 1c). The domain size of 10° is determined by the typical spatial scale of ISV of LCC from the Hovmöller diagram result (not shown), and results shown in later do not strongly depend on the domain size.

Fig. 1.
Fig. 1.

(a) Climatology of low-level cloud cover (LCC) derived from the Moderate Resolution Imaging Spectroradiometer (MODIS; shaded; %) and sea surface temperature (SST) derived from OISST [contours; °C; contour interval (CI) = 2°C; thick contour (TC) = 16°C] in the North Pacific in July from 2003 to 2016. (b) Standard deviation of deseasonal daily LCC data (shaded; %) and SST (contours; °C; CI = 0.5°C; TC = 1.5°C). The black rectangle shows the region where the standard deviation of LCC is large and the target meridional band from 30° to 40°N used in the present study. (c) Mean seasonal variation of the subseasonal standard deviation of LCC (shaded; %) and SST (contours; °C; CI = 0.3°C) for the target meridional band. The standard deviation for each variable on day i is calculated as the standard deviation of LCC and SST from day [i − 45] to day [i + 45] using a method similar to that of Wang et al. (2012). The two gray horizontal dashed lines show the season analyzed in the present study (JJASO). The black rectangle shows a target longitudinal band from 165° to 175°E where the standard deviation of LCC and SST are both large.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

To evaluate the SST variability associated with the ISV of low-level cloud, an analysis of the mixed layer temperature budget was conducted. The mixed layer temperature budget equation (Moisan and Niiler 1998) is as follows:

SSTt=(QtotalQpen)ρCpH+R,

where Qtotal and Qpen represent the sum of the net downward heat fluxes at the ocean surface, and the sum of the penetrated shortwave radiative flux at the bottom of the mixed layer (Paulson and Simpson 1977), respectively; H is a constant mixed layer depth (= 25 m); ρ is the density of water; and Cp is the specific heat of water. The constant H was estimated as the climatological mean mixed layer depth in summertime (JJASO) in the target region as obtained from MIMOC (Johnson et al. 2012). The variable R is a residual term including an oceanic term (i.e., horizontal and vertical advection, entrainment processes, and vertical ocean mixing), some errors for observational data and the assumption with constant H. A prime denotes the filtered variables (20–100 days).

Relative humidity (RH) is an important factor for determining cloud evolution. To investigate the detailed mechanism of LCC evolution, we analyzed the temporal variation of RH and its tendency. RH is calculated by RH = e/es, where e is the water vapor pressure and es is the saturated water vapor pressure. These two variables are determined using temperature T and specific humidity q. Therefore, the tendency of RH is separated into tendencies of T change and q change (Ek and Mahrt 1994; Babić et al. 2019) as

RHt=pes0.622(0.378q+0.622)2qtqLυ(0.378q+0.622)RdT2Tt,

where p is pressure, Lυ is the latent heat of evaporation, and Rd is the constant of dry air. We conducted a quantitative investigation of RH tendency. To reveal what process is important for changes in RH tendency, we further decomposed tendencies of q and T into adiabatic and diabatic process, respectively. Thus, we used temperature and moisture budget equations described below:

Tt=(uTy+υTy+ωTp+ωαcp)+Q1cp,
qt=(uqy+υqy+ωqp)Q2Lυ,

where (u, υ, ω) is the three-dimensional wind, α is the specific volume, Q1 is the apparent diabatic heat source, and Q2 is the apparent diabatic moisture sink. The first four (three) terms in the right-hand side (rhs) of Eq. (3) [Eq. (4)] are related to adiabatic process and fifth (fourth) term is related to diabatic process, which is calculated as the residual of the Eqs. (3) and (4). Combining Eqs. (2)(4), we decomposed the RH tendency into three components; RH tendencies associated with 1) horizontal adiabatic process (i.e., zonal and meridional advections), 2) vertical adiabatic process (i.e., vertical advection and adiabatic cooling term), 3) diabatic process. In addition, we separated each term into processes associated with the tendencies of T and q. The formation and dissipation processes of low-level cloud associated with RH change are discussed in section 3c.

3. Results

a. Dominant time scale of LCC variability in summertime NP

We investigated variabilities of the dominant time scales of LCC, SST, and Tadv in the target region and conducted fast Fourier transfer (FFT) analysis of LCC variability. Figure 2a shows the 14-yr mean summertime spectrum of the area-mean LCC in the target region. The results indicate that there are two dominant time scales in the variability of LCC: 1) 2–3 days and 2) 10–50 days. Because the sampling period is once a day, the first time scale is almost equal to the Nyquist frequency, and is likely to correspond to synoptic disturbances. The spectral signal is significantly weaker for the second time scale than for the first time scale. The second spectral signal is likely to be associated with the ISV of LCC, atmospheric circulation, and SST in summertime NP, which is consistent with the findings of Wang et al. (2012). The squared coherency between the area-mean LCC and area-mean SST in the target region is shown in Fig. 2b. No significant coherency peak was observed on the synoptic time scale (2–3 days). However, a significant coherency peak was observed on the intraseasonal time scale (20–40 days), with a maximum at 30 days. The squared coherency between LCC and Tadv has similar characteristics on the intraseasonal time scale; therefore, we conducted further analysis focusing on the ISV of LCC using a bandpass filter and a phase composite analysis technique as described in section 2b. We define the intraseasonal time scale as 20–100 days, following Wang et al. (2012).

Fig. 2.
Fig. 2.

(a) Mean power spectrum of the area-mean daily LCC in the target region (30°–40°N, 165°–175°E). The mean power spectrum was calculated as the mean value of all spectra for each JJASO season from 2003 to 2016. Solid and dashed lines indicate red noise and the 95% significance level, respectively. Green dots indicate when the power exceeds the 95% significance level line. (b) Mean squared coherency between LCC and SST (blue) and between LCC and surface horizontal temperature advection Tadv (orange). The horizontal dashed line indicates the 95% significance level determined by a chi-squared distribution.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

b. ISV of LCC in summertime NP

Next, we investigated the characteristics of LCC evolution in the target region on an intraseasonal time scale using phase composite analysis method. From the phase composite analysis of the 14-yr JJASO LCC time series, we obtained 46 cycles of LCC ISV in the target region. Figure 3 shows the statistical information of the ISV of LCC, indicating the one-cycle period and amplitude (Figs. 4a,b, respectively). The amplitude was calculated as the difference between the maximum and minimum values of LCCA divided by 2. The mean cycle period and the mean amplitude were found to be 29.5 days and 11.3%, respectively. No cycle was observed with a period longer than 50 days, implying that, if such a cycle exists, its amplitude is small. The mean cycle period is similar to the largest period determined by the FFT analysis on an intraseasonal time scale (Fig. 2a), and the mean amplitude is about half of the standard deviation of LCC during JJASO (Fig. 1c).

Fig. 3.
Fig. 3.

Histograms of (a) cycle period and (b) amplitude for each cycle of the intraseasonal variability of LCC in the target region. The total number of cases is 46 for JJASO from 2003 to 2016. The bin sizes of each histogram are 4 days in (a) and 2% in (b).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

Fig. 4.
Fig. 4.

Phase composite of the horizontal distribution of the LCC anomaly (shaded; %) and horizontal wind speed at 1000 hPa (vectors; m s−1). The black rectangle denotes the area for which the LCC anomaly index was calculated. The maximum LCC anomaly appears at a phase of 0°, while the minimum LCC anomaly occurs at phases of −180° and 180°.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

Figure 4 shows the composite mean horizontal distribution of the 20–100-day filtered LCCA and the anomalous horizontal wind field at 1000 hPa at each phase. Note that, because the mean cycle period is 29.5 days (Fig. 2a), the length of time between each phase interval of 30° is about 2–3 days. The nature of the phase composite method means that the maximum (minimum) LCCA appears at a phase of 0° (phases of −180° and 180°) in the target region (shown by black rectangles in Fig. 4). The sign of the LCCA index changes at around a phase of −90° (negative to positive) and 90° (positive to negative). When the LCCA is minimum (at phases of −180° and 180°), anomalous lower-atmospheric circulation patterns are anticyclonic; these patterns weaken and disappear gradually as the phase increases from −180°, and circulation becomes cyclonic when the LCCA reaches its maximum value (phase of 0°). Similarly, as the phase decreases again, the cyclonic circulation patterns gradually disappear, and anticyclonic patterns appear at a phase of −180°. Another characteristic feature of atmospheric circulation in the target region is that anomalous equatorward advection occurs during the development stage of the LCCA from phases of −90° to 0°. Anomalous Tadv is likely to be an important controlling factor of LCC (e.g., Klein 1997). The evolution of other controlling factors of LCC in the cycle, namely EIS and SST, were also investigated using the same phase composite analysis technique (Fig. 5). As many studies have reported a strong relationship between LCC, SST, and EIS (e.g., Norris and Leovy 1994; Wood and Bretherton 2006), we also found that SST and EIS anomalies are positive and negative in the target region, respectively, when the LCCA reaches a maximum at a phase of 0°, and vice versa (Fig. 5).

Fig. 5.
Fig. 5.

As in Fig. 4, but for SST (shaded; °C) and EIS (contours; CI = 0.1 K; TC = 0 K).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

To qualitatively analyze the detailed lead–lag relationship between LCC and LCC controlling factors, the evolution of the area-mean anomalies of LCC, SST, EIS, and Tadv were calculated as shown in Fig. 6. Two prominent features of the lead–lag relationships were found. First, the minimum Tadv anomaly (i.e., cold advection) appears at a phase of −30°, before LCCA reaches a maximum at a phase of 0°. The phase lags of the minima/maxima of the LCC and Tadv anomalies imply that anomalous Tadv is a “trigger” of the LCCA variation within the cycle. Anomalous cold Tadv impacts LCC variation by modulating the characteristics of the atmospheric boundary layer, that is, a destabilization process or sensible heat flux (e.g., Wood 2012). Thus, we next tested the detailed mechanism of LCC evolution associated with Tadv in terms of changes to the humidity field, as detailed in the next section. The second prominent feature of the lead–lag relationships is that the minimum SSTA and the maximum EIS anomaly occur at phases of 30° and 60°, respectively, after LCCA reaches its maximum at a phase of 0°. Considering the physical connections between LCC and SST, this implies that the SST and EIS anomalies are likely to be “followers” of LCCA variation within the cycle, through the interaction between LCC and SST associated with the cloud radiative effect. While, it is doubtful that EIS anomaly is follower of LCCA because lead–lag relationship of EIS and LCC are different in the cases with and without random overlap assumption (appendix A). In the present study, the radiative impact of low-level cloud on the SST anomaly was investigated by applying the mixed layer temperature budget equation [Eq. (1)]. The lead–lag relationships between LCC, Tadv, and SST are consistent with the findings of Xu et al. (2005), who investigated the subseasonal variability of the southeast Pacific stratus cloud deck based on satellite, buoy, and reanalysis datasets. While SST and EIS are identified as “followers” of LCC inspired from Fig. 6, they have potentials to modulate LCC through processes of suppressed entrainment of dry-air from free troposphere (Bretherton et al. 2013). Thus, there is a possibility of the existences of the positive feedback process and the fast modulation process of negative SST and positive EIS for positive LCC anomaly around phase 0 (Fig. 6). However, robust conclusion about the relationship among Tadv, LCC, and SST is “Tadv is a trigger of LCC and SST” on intraseasonal time scale in midlatitude western NP.

Fig. 6.
Fig. 6.

Evolution of composite anomalies of LCC (black), SST (blue), Tadv (green), and EIS (red) averaged over the target area.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

To determine the phenomena that cause anomalous changes in atmospheric circulation on an intraseasonal time scale, we investigated the evolution of large-scale atmospheric circulation at the surface, midlevel (500 hPa), and upper altitude (250 hPa) to investigate the origin of the ISV of Tadv. Figure 7 shows the horizontal distribution of geopotential height (GPH) anomalies at 500 hPa (shaded) and 250 hPa (contours) and the corresponding horizontal wind field at 500 hPa (vectors). A similar feature with lower-atmospheric circulation anomalies is visible in Fig. 4, including cyclonic (anticyclonic) circulation patterns with a negative GPH anomaly that appears at a phase of 0° (phases of −180° and 180°) in both the mid- and upper troposphere. This suggests that changes in circulation patterns during the evolution of LCCA have a barotropic structure. GPH anomalies at 500 and 250 hPa are not stationary, but rather propagate westward (see green stars in Fig. 7). This westward propagation is likely associated with Rossby wave excitation around the Rocky Mountains, which is similar to the first mode of summertime GPH on an intraseasonal time scale (Wang et al. 2013). This anomalous atmospheric circulation pattern is mainly driven by internal atmospheric dynamics in the midlatitudes; however, the GPH anomaly at 250 hPa shows a weak wave train pattern with a cyclonic circulation in the middle and two anticyclonic circulations at the two sides from phases of −30° to 30°. Previous studies have shown that atmospheric circulation patterns in the midlatitudes of summertime NP are remotely forced by intraseasonal atmospheric oscillation in the tropics (Kawamura et al. 1996; Pan and Li 2008; Mori and Watanabe 2008). Although determining the contributions of remote forcing in the tropics and internal dynamics in the midlatitudes to the anomalous atmospheric circulation pattern in the NP is key to understanding the ISV of atmospheric circulation, this topic will not be investigated further here because it is outside the scope of the present study.

Fig. 7.
Fig. 7.

Phase composite of a horizontal map of the anomalies of geopotential height (GPH) at 500 hPa (shaded; m), GPH at 250 hPa (contours; m; CI = 7.5 m), and horizontal wind speed at 500 hPa (vectors; m s−1). The black rectangle denotes the target area where the LCC anomaly index was calculated. The green outlined star at each phase indicates the position of minimum of GPH anomaly (i.e., cyclonic anomaly) at 500 hPa.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

c. Mechanism of evolution of LCC and changes in humidity field associated with the ISV of LCC

Cold Tadv is known as a one of the indicators to promote the formation of low-level cloud through an enhanced turbulence within the atmospheric boundary layer due to the upward surface buoyancy flux from the sea surface (e.g., Klein et al. 1995; Klein 1997). In this subsection, we further investigated the LCC evolution process in terms of the RH field when anomalous Tadv occurs. The RH is an important atmospheric environmental controlling factor for LCC formation and cloud microphysics within the atmospheric boundary layer (rather than the specific humidity). Therefore, to determine the physical process of increasing (decreasing) LCCA from the Euler perspective, we further investigated the variations of the thermodynamic fields (i.e., T, q, and RH) associated with the cycle of the ISV of LCC using the same phase composite analysis method.

Figure 8 shows the evolution of the area-mean vertical distributions of q, RH, and vertical pressure velocity ω associated with the evolution of LCCA. When the LCCA reaches a maximum at a phase of 0°, the q anomaly is negative (dry) near the surface below 850 hPa, while the RH anomaly is positive (moist) there. The negative q anomaly results from the anomalous equatorward advection of dry air at a phase of 0° (Fig. 4). Compared with the near the surface, both the RH and q anomalies are positive between 700 and 850 hPa at a phase of 0°. When LCCA is minimum (at phases of −180° and 180°), the signs of both humidity anomalies are opposite to those at a phase of 0°. In terms of the atmospheric dynamics, updraft (downdraft) occurs when LCCA is positive (negative). The evolution of RH tendencies are shown in Fig. 9a, indicating that the tendencies are positive during the development stage of LCCA from a phase of −180° to 0° and negative in the dissipation stage from a phase of 0° to 180°. The variations of RH and its tendency corresponded well with the variation of LCCA.

Fig. 8.
Fig. 8.

Evolution of the vertical distributions of anomalies of relative humidity (RH; shaded; %), specific humidity q (contours; g kg−1; CI = 0.1 g kg−1; TC = 0 g kg−1), and vertical pressure velocity ω (vectors; hPa day−1). Red and blue arrows indicate updrafts and downdrafts, respectively.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

Fig. 9.
Fig. 9.

As in Fig. 8, but for anomalies of (a) RH tendency (dRH/dt; shaded; %/day) and RH (contours; %; CI = 1%; TC = 0%). The dRH/dt anomalies by (b) total adiabatic, (c) total diabatic, (d) horizontal adiabatic, and (e) vertical adiabatic processes. (f)–(h) The anomalies by horizontal adiabatic, vertical adiabatic, diabatic processes related to change in temperature (T-change; % day−1). (i)–(k) As in (f)–(h), but the processes related to change in specific humidity (q-change; % day−1). The horizontal dashed line in each panel indicates the pressure level at 700 hPa. Note that color spacing is not uniform in (b)–(k).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

We will further describe a dominant process that determines the variation of the RH tendency. Figures 9b–k show decomposed anomalies of the RH tendency into horizontal adiabatic, vertical adiabatic, and diabatic processes. The anomalies of horizontal adiabatic process are the sum of contributions from zonal and meridional advections, and those of vertical adiabatic process are the sum of contributions from vertical advection and adiabatic cooling process with updraft. The contribution from diabatic process is calculated as the residual in the budget equations of temperature and moisture, as described in section 2b. Each contribution is further decomposed into that associated with changes in temperature and moisture (T-change and q-change, respectively). Hereafter, we mainly focus on the positive RH tendency within the boundary layer (below 700 hPa) in the developing stage of LCCA before a phase of 0°. First, the decomposed anomalies of total adiabatic and diabatic processes are almost balanced (Fig. 9b versus Fig. 9c); however, the positive RH tendencies from −120° to 0° (from −180° to −120°) are derived from adiabatic (diabatic) process. Second, the decomposed anomalies of the adiabatic process indicated that the vertical adiabatic processes are dominant (Fig. 9d versus Fig. 9e). Additional temperature budget analysis within the boundary layer, shown in appendix B, suggested that the vertical processes are due to adiabatic cooling with updraft and vertical moisture advection related to shallow convection (Figs. 9g,j, B1). While the vertical adiabatic process is dominant in the developing stage, the horizontal adiabatic process contributes to some parts of the positive RH tendencies below 850 hPa (Fig. 9d), which is related to the T-change rather than the q-change (Figs. 9f,i). It implied that horizontal advection of positive RH anomaly associated with anomalous dry but cold advection is also important to increase LCCA. The RH advection is driven by meridional winds (i.e., northerly) rather than zonal winds (Fig. B1), which is also supported by large meridional gradient of seasonal-mean state of RH around 40°N in summertime NP (Fig. 10). We also expected that large RH in northern part from the target region is derived from low es induced by low temperature, not from high e (i.e., water vapor pressure) induced by large specific humidity. The diabatic processes from −180° to −120° seem to be related to cooling from sea surface (Fig. 9h) due to the anomalies of negative SST and positive Tadv (i.e., warm advection) during the period (Fig. 6). From the overall results, we concluded that not only vertical adiabatic process with shallow convection but also meridional advection of positive RH associated with cold Tadv near sea surface play key roles in formation of low-level cloud in the ISV cycle.

Fig. 10.
Fig. 10.

Climatological seasonal-mean state in June–July–August (JJA) of a horizontal map of RH at 1000 hPa derived from ERA-Interim (%; CI = 2%) from 2003 to 2016.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

d. Air–sea interaction process between low-level cloud and SST associated with the ISV of LCC

In this subsection, we discuss the air–sea interaction processes among LCC, surface heat flux, and SST that are associated with the ISV of LCC. In section 3b, we determined the lead–lag relationship between the LCC and SST within a cycle of the ISV of LCC, showing that SST variation follows to LCC variation, that is, the maximum LCCA at a phase of 0° induces the minimum SSTA at a phase of 60° (Fig. 6). Surface heat flux into the oceanic mixed layer is a key component for determining SST variation in summertime NP. In particular, the reflection of incoming SW at the sea surface by oceanic low-level cloud is an important factor in the decrease of SST in summertime NP. Besides the SW radiative heat flux, other factors that determine SST variability on an intraseasonal time scale include turbulent heat flux, oceanic meridional advection, and Ekman pumping (Wang et al. 2012). To qualitatively investigate the relative importance of these factors, we evaluated each term of the mixed layer temperature budget equation [Eq. (1)]. This equation requires the mixed layer depth (MLD) at each phase. Unfortunately, no observational daily dataset of MLD exists; therefore, we used a constant value for the MLD (25 m) based on the climatological JJASO-mean MLD in the target region from the MIMOC dataset. Although MLD variation is an important factor for determining the SST tendency (Morioka et al. 2010), we ignored such variation in the present study. As explained above, some terms in the mixed layer temperature budget equation are associated with oceanic processes (e.g., an advection term and an entrainment term) (Qiu and Kelly 1993). However, we also ignored these terms because their contributions to SST tendency on an intraseasonal time scale in summertime Kuroshio Extension region have been reported to be much smaller than the contribution of the surface heat flux term (Wang et al. 2012).

Figure 11a shows the evolutions of SST tendency and the surface heat flux term in Eq. (1) at each phase, based on the phase composite analysis for the ISV of LCC. To the first order, the heat flux term can explain most of the variation of SST tendency. Additionally, there is a slight lag between the observed SST tendency and the heat flux term. We expect that this lag is due either to the fact that oceanic terms were ignored or to the uncertainty in the observational dataset for the total heat flux or SST. Figure 11b shows the decompositions of the surface heat flux into SW, LW, sensible heat (SH), and latent heat (LH), and their evolutions. In addition to the term regarding the phase of SST cooling by surface heat flux (from a phase of −90° to 60°; see Fig. 11b), SW and LH are two dominant factors for SST cooling. The contributions of SW and LH to SST cooling are almost the same in the cooling phase. The contributions of SH and LW to SST tendency are negligible compared with the contributions of SW and LH. The contributions of SW and LH to SST tendency are also dominant during the SST warming phase (from phases of 0°–180° to 120° and from 120° to 180°, respectively; see Fig. 11b); however, the SW heating is slightly larger than the LH heating.

Fig. 11.
Fig. 11.

(a) Evolution of anomalies of SST tendency (dSST/dt; black bar) and surface heat flux term (red circles). (b) As in (a), but for anomalies of the total heat flux term (red bar), shortwave radiation (SW), longwave radiation (LW), sensible heat (SH), and latent heat (LH).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

As shown in Fig. 11, the radiative flux term of SW variation corresponds to the LCC variability, meaning that an increase (decrease) in LCC increases (decreases) the albedo of the entire target region and suppresses (enhances) the downward SW radiation at the sea surface. However, the reason why LH variation is associated with the ISV of LCC is still unclear. Thus, we continued to analyze in detail the ISV of LH by linear decomposition of the bulk formula of LH as expressed below. LH variation is caused by a change in specific humidity at the sea surface qs and near the sea surface qa, and a change in wind speed near the sea surface (WS10). The upward LH at the sea surface is represented by the bulk formula

LH=ρaLCEWS10(qsqa),

where ρa is the atmospheric density, L is the latent heat of vaporization, and CE is the bulk coefficient. We decomposed the LH anomaly into the contributions from qs, qa, and WS10 using the following equation:

LH=ρaLCE[WS10(qs¯qa¯)+WS10¯(qs)WS10¯(qa)+WS10(qsqa)],

where the overbar denotes the climatological mean and prime denotes the anomaly from the mean. The first, second, and third terms on the right side of the equation are the contributions to the LH anomaly from changes in WS10, qs, and qa, respectively. The fourth term on the right side represents a nonlinear term related to changes in wind speed and specific humidity. We ignored the fourth term because its value is significantly smaller than those of the other three terms, differing by two orders of magnitude. Figure 12 shows the evolutions of the anomalies of WS10, qs, and qa, (Fig. 12a), as well as their contributions to the LH anomaly (Fig. 12b). At a phase of 0°, the WS10 anomaly is positive and the qa anomaly is negative. Then, the contributions to the LH anomaly from WS10 and qa are both negative. The anomaly of qs is also negative following to the negative qa anomaly from a phase of −60°, but the qa anomaly is more negative than the qs anomaly (Fig. 12a). Negative contribution of qa and the lagged relationship between anomalies of qa and qs suggest that anomalous dry advection enhances latent heat release. The amplitude of the contribution to LH from WS10 is slightly larger than those of the contributions from qs and qa. Additionally, the contribution to LH from WS10 is at the same phase as the total LH anomaly. Therefore, we conclude that LH variation associated with the ISV of LCC is driven mainly by WS10 change. The results of “strong” northerly winds is opposite sense to the climatological meridional wind direction, i.e., mean southerly wind due to the North Pacific subtropical high in boreal summer. However, scalar wind speeds on shorter time scales than subweekly depend on not only the climatological mean state but also synoptic atmospheric variability (Miyamoto et al. 2018; Ogawa and Spengler 2019). The detailed mechanisms of high-frequency wind changes and its impact on ISV of SST still need for more discussion, but beyond the scope of this study. Based on the overall results from sections 3b to 3d, we also suggest that the ISV of anomalous Tadv plays a key role in controlling the surface heat flux via changes not only in turbulent heat flux but also in radiative flux by modulating the evolution of LCC.

Fig. 12.
Fig. 12.

(a) Evolution of anomalies of specific humidity at the sea surface qs (blue), specific humidity at the near surface qa (red), and wind speed at 10 m (WS10; green). (b) As in (a), but for anomalies of latent heat flux (LH; orange shading), LH by qs change [LH(qs); blue], LH by qa change [LH(qa); red], and LH by WS10 change [LH(WS10); green].

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

Finally, we further tested the regionality of the coupling process among LCC, SST, and Tadv described above in entire North Pacific. For this purpose, we performed a simple lead–lag correlation analysis without the filtering method, and without a phase composite analysis in not only the target region but also other region with the same domain size. Figure 13 shows the lead–lag correlation between the unfiltered but deseasonal time series of LCC and SST and Tadv in the target region and northeastern subtropical Pacific region (20°–30°N, 140°–150°W) where low-level cloud frequently appears. As suggested in section 3b, in the target region, the lag correlation between LCC and Tadv is negative and reaches a minimum at a lag of +1 day (Fig. 13a). This supports the result shown in Fig. 6 and indicates that anomalous Tadv plays a role in triggering LCC variation. In contrast, the lag correlation between LCC and SST is significantly negative from a lag of −8 to +1, and a negative peak appears at a lag of −2 (Fig. 13a). These two results imply that SST increases (decreases) after a decrease (increase) in LCC and the radiative impact of LCC on SST, which persists for about a week. As shown in Fig. 6, the LCC variability is a leading mode compared with the SST variability. The lag correlation coefficient if SST leads LCC (positive x axis) is also slightly negative; however, it is not significant except for a lag of +1 (Fig. 13a). The lead–lag relationship in northeastern subtropical region is quite similar to that in the target region, implying the similar air–sea interaction process described in the present study occurs in the other region. Figure 14 show horizontal maps of correlation coefficient at lag 0 of LCC and SST, and that of LCC and Tadv, respectively, using deseasonal time series in grid boxes with 10° resolution. We also displayed the SST persistent days and Tadv leading days in each grid defined as the continuous days with the significant correlations at the 95% level. Significant negative correlations of LCC and SST, and that of LCC and Tadv appears in two specific regions; 1) south of SST frontal region in western and central NP included in the target region and 2) northeastern (subtropical) NP, while SST persistent days are longer in the eastern NP than that in the western and central NP (Fig. 14a). The spatial feature of relationship is similar if we focused on only the intraseasonal time scale (Figs. 14c,d). It suggested that lead–lag relationship among LCC, SST, and Tadv is robust in the two specific regions through air–sea interaction process presented in the present study. Regionality of the SST persistent days seems to be related with the horizontal distribution of oceanic MLD, e.g., MLD in the specific regions is about 25 m in boreal summer.

Fig. 13.
Fig. 13.

Lag correlation coefficient between LCC and SST (blue) and between LCC and Tadv (green) derived from an unfiltered daily dataset for JJASO for 2003–16 in (a) the target region (30°–40°N, 165°–175°E) and (b) the northeastern subtropical Pacific region (20°–30°N, 140°–150°W). Lag correlation coefficients were calculated as mean values over 14 years. Horizontal dashed and solid lines indicate that the minimum correlation coefficient was exceeded at the 95% and 99% significance levels, respectively, calculated with an effective degree of freedom (Neff) of 152/50 × 12 = 36. Circles indicate when the coefficient at a certain lag day exceeds the 95% significance level.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

Fig. 14.
Fig. 14.

Horizontal map of correlation coefficient (a),(c) between LCC and SST and (b),(d) LCC and Tadv derived from (top) a deseasonal daily dataset and (bottom) filtered datasets of intraseasonal time scale (20–100 days) for JJASO for 2003–16. Correlation coefficients were calculated as mean values over 14 years. Numbers in grid box indicate the lasting (leading) days of significant correlation coefficients of SST (Tadv) based on the days of a maximum of the LCC anomaly with a 95% confidence level. Grid boxes without the number indicate any significant correlation coefficients do not appear in the day range for the calculation of the lag correlation (from −40 to 40 days).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

4. Discussion and summary

This study revealed the intraseasonal (20–100 days) variability (ISV) of oceanic low-level cloud cover (LCC) and associated LCC controlling factors, in particular SST and Tadv, in summertime NP using satellite observations and a meteorological reanalysis dataset. We applied a filtering method and a phase composite technique to the LCC anomaly (LCCA) in the western NP and extracted 46 cycles of the ISV of LCC using a 14-yr JJASO dataset, whose mean amplitude and period are 11.3% in cloud fraction and 29.5 days, respectively. Within the cycle of the ISV of LCC, an increase in the LCCA is likely to be triggered by adiabatic cooling with shallow convection and meridional advection of RH with anomalous cold air with a cyclonic atmospheric circulation pattern. Previous studies found that the ISV of the atmospheric circulation pattern in summertime NP is associated with internal atmospheric dynamics in the midlatitudes (Wang et al. 2013) and remote forcing from the tropics (Mori and Watanabe 2008). In the evolution process of increasing of LCC, anomalous cold advection and updraft within the boundary layer is important for inducing a positive RH anomaly, due to the decrease in saturated water vapor pressure by anomalous cold temperature. This process encourages favorable conditions for the formation of low-level clouds in the lower troposphere below 700 hPa, despite the fact that cold advection brings dry air from the polar region.

Although SST and EIS are important factors controlling LCC variation on seasonal and interannual time scales (e.g., Klein and Hartmann 1993; Klein et al. 2017), SST anomaly is a follower to LCCA, that is, SST reaches a minimum after LCC reaches its maximum. EIS is also likely to follower of LCCA, however, further investigation is needed for the relationship between LCC and EIS because the result of their lead–lag relationship depends on overlapping assumption (appendix A). Our thorough analysis of the mixed layer temperature budget emphasized the importance of anomalous dry and cold advection for not only the formation of low-level clouds but also for the decrease in the local SST due to the enhanced release of turbulent heat flux. During the cooling phase of the SSTA, the positive LCCA induced by the cold advection leads to SW radiative cooling of the sea surface. Furthermore, the dry advection enhances the latent heat release from the sea surface. Originality of the present study is to reveal causal relationship implied by the lead–lag composite analysis within the air–sea interaction associated with ISV of LCC [e.g., positive feedback between LCC and SST (Norris and Leovy 1994)] in the summertime western North Pacific. The method is similar to Wang et al. (2012), however, differences in based variables for composite analysis between the present study and the previous study brought to us new insights for the formation mechanism of low-level cloud evolution, different pathway of SST variation, and regionality of the robust coupling among LCC, SST, and Tadv on the intraseasonal time scale.

Previous studies investigating the lead–lag relationship between cloud and ocean generally suggested that the ocean plays a passive role under the atmospheric forcing via surface heat flux in summertime midlatitude ocean basins (Frankignoul and Hasselmann 1977; Frankignoul and Reynolds 1983; Frankignoul 1985). We also obtained similar results with it, indicating the passive role of cloud radiative forcing and LH on SST on an intraseasonal time scale. On the other hand, some studies have suggested that the ocean plays an active role in the ocean-to-atmosphere circulation pattern on intraseasonal and interannual time scales (Wang et al. 2012). Frankignoul et al. (2011) examined the atmospheric response to variability of the subarctic front associated with the cold Oyashio and warm Kuroshio, showing that a change in large-scale circulation is induced by a change in surface heat flux derived from the SSTA, especially in wintertime. Wang et al. (2012) proposed a two-way interaction process between the local SSTA and a change in atmospheric circulation during the transition phase of the change in vertical atmospheric structure from an anomalous barotropic to a baroclinic structure, related with the heating of the lower troposphere by SSTA. This suggests an active role of SSTA in promoting LCC formation in summertime NP. Mauger and Norris (2010) showed that SST has an active impact on the evolution of stratocumulus using the back-trajectory method with a meteorological reanalysis dataset. They also found that the horizontal distribution of SST along the pathway of low-level cloud is important to LCC evolution. Both these previous studies and the present study suggested that, when using local phase composite analysis, the Lagrangian method should be applied instead of the Eulerian method in order to determine the active impact of SST on the properties of oceanic low-level cloud.

Observational evidence of the impact of SST on LCC has been confirmed for interannual and longer time scales. However, the impact of SST on LCC has not been confirmed on shorter time scales, such as synoptic and intraseasonal time scales. This is due to the fact that the dominant time scales of cloud and SST variabilities are less than 3 days and more than 7 days, respectively (Fig. 2a). This was also investigated by de Szoeke et al. (2016). This difference between the time scales of cloud and SST variability complicates the characterization of air–sea interaction. To further understand the importance of the interaction between cloud and ocean in midlatitude oceanic basins for the time scale of cloud variability, it is necessary to consider and investigate the interaction process across different time scales; for example, the relationship between the monthly mean state and variability of LCC on a synoptic or intraseasonal time scale. In other word, large variability of LCC appears in the region where mean SST is high and mean LCC is small around the specific meridional zone from 30° to 40°N, which is implied in Figs. 1a and 1b. Further analysis is required using high temporal resolution cloud datasets (e.g., from geostationary satellites) for combined analysis of cloud variability on shorter time scale with SST variability on longer time scale.

Despite the difficulty in detecting the active role of SST in the evolution of LCC, the method used in the present study is useful for model output evaluation. Myers et al. (2018) evaluated the difference in the feedback processes of low-level cloud and SST in CMIP5 model outputs, showing that summer-to-summer SST variability depends on the representation of the feedback between low-level cloud and SST in each model. To reduce the uncertainty in the feedback process on an interannual time scale, it is necessary to reveal the complex process of low-level cloud evolution associated with many controlling factors. Additionally, evaluations of each air–sea interaction process on a time scale shorter than interannual are required. The method used in the present study is helpful to better understand the physical processes involved in the evolution of LCC and evaluate model performance of the air–sea interaction.

Acknowledgments

This work was supported by a Grant-in-Aid from the Japan Society for the Promotion of Science (JSPS), Research Fellow JP18J10606, and by a Grant-in-Aid for Scientific Research (B) 16H04046, also from the JSPS. The Aqua/MODIS Cloud Daily L3 Global 1 Deg CMG dataset was acquired from the Level-1 and Atmosphere Archive Distribution System (LAADS) Distributed Active Archive Center (DAAC), located in the Goddard Space Flight Center in Greenbelt, Maryland (https://ladsweb.nascom.nasa.gov/). CERES data were obtained from the NASA Langley Research Center CERES ordering tool at http://ceres.larc.nasa.gov/. OISST data were provided by the National Oceanic and Atmospheric Administration (NOAA). The global ocean heat flux and evaporation data were provided by the WHOI OAFlux project (http://oaflux.whoi.edu/last access: 6 November 2018) funded by the NOAA Climate Observations and Monitoring (COM) program. ERA-Interim data were downloaded from the European Centre for Medium-Range Weather Forecasts (ECMWF) data server at http://apps.ecmwf.int/datasets/. The author would like to thank Tim Li and Bo Qiu of the University of Hawai‘i at Mānoa and Toshio Suga of Tohoku University for their helpful discussion. Additionally, Joel Norris and Shang-Ping Xie, of the Scripps Institution of Oceanography, provided helpful insight and comments. Last, the authors are grateful for the thoughtful comments of three anonymous reviewers to improve the original manuscripts.

APPENDIX A

Random Overlap Assumption for Daily Mean Dataset of MODIS

To estimate the “true” low-level cloud cover based on passive sensor, we applied the random overlap assumption in the present study. Basically, high-level cloud tends to coappear with low-level cloud in the target region of western North Pacific. Thus, we investigated the vertical profile of cloud fraction observed by active sensor on CloudSat/CALIPSO corresponded with certain estimated LCC from passive sensor. We also compared the cloud occurrence profiles obtained from the 2B-GEOPROF-lidar product based on estimated LCC from MODIS with “random overlap assumption” and “non-overlap assumption.” Figure A1 show the composited-mean cloud occurrence profiles based on estimated LCC with and without the assumption. It shows that LCC estimated from random overlap assumption increases with increasing high-level cloud cover (HCC) (Fig. A1a). On the other hand, estimated LCC without any assumption increases with decreasing HCC (Fig. A1b). These results imply that LCC estimated from passive sensor in the western NP depends on middle- and high-level cloud covers, with large uncertainty due to the HCC variation. However, the estimated LCC without any assumption is significantly influenced by the obscuration by high-level cloud cover, which is an artificial problem. Thus, we applied the random overlap assumption to estimate LCC by the measurements using passive sensor.

Fig. A1.
Fig. A1.

Cloud occurrence profiles based on estimated LCC with (a) random overlap assumption and (b) no assumption in the target region (30°–40°N, 165°–175°E) from 2007 to 2010 JJA, using daily data of MODIS grid data and level-2 profile data of the 2B-GEOPROF-lidar product.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

To confirm the conclusion about the lead–lag relationship among LCC, SST, EIS, and Tadv, we conducted the same analysis based on anomaly of LCC without random overlap assumption (Fig. A2). Results about the evolution of the area-mean anomalies of LCC without any assumption, SST, and Tadv are qualitatively similar to that with random overlap assumption. In contrast, lead–lag relationship between EIS and LCC is slightly different, that is, there is no lag between their variabilities. Thus, further analysis is needed for the LCC–EIS relationship in the western NP where high-level cloud occurrence is relatively frequent compared with northeastern subtropical Pacific.

Fig. A2.
Fig. A2.

As in Fig. 6, but for the results based on LCC anomaly calculated without random overlap assumption.

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

APPENDIX B

Temperature Tendency Equation for Intraseasonal Variability in the Atmospheric Boundary Layer

The detailed process of cooling atmospheric boundary layer, which is important to modulate saturated water vapor pressure (section 3c), were investigated. We applied the temperature tendency equation [Eq. (3)] for filtered variables on intraseasonal time scale. Figure B1a displays the evolution of the area-mean anomalies of each term integrated in the boundary layer (from 850 to 1000 hPa) calculated from the phase composite analysis based on LCCA in the target region, similar to Fig. 6. It indicates that cooling and warming within the boundary layer occurs before and after a phase of 0°, respectively. In particular, in the phases from −150° to −30°, significant cooling occurs due to the meridional advection term and adiabatic cooling term (Fig. B1b). The adiabatic cooling term seems to be related with the shallow convection because the updraft occurs within the boundary layer when LCCA is positive (Fig. 8). Additionally, in terms of the phase relationship between total temperature tendency term and each term on the rhs, the variation of meridional advection corresponded well with that of the total temperature tendency term. Thus, the results imply the importance of meridional cold advection in the atmospheric boundary layer cooling process in the developing stage of LCCA (before a phase of 0°).

Fig. B1.
Fig. B1.

(a) Composited mean of each term in temperature budget integrated from 1000 to 850 hPa in the target region. (b) The phase mean of each term in the cooling phase (from phase −150° to −30°) is also displayed. Colors in the line and bar are corresponded within the two panels (black: total tendency, red: zonal advection term, blue: meridional advection term, green: vertical advection term, purple: adiabatic heating term associated with vertical pressure velocity, gray: diabatic heating term calculated as a residual).

Citation: Journal of Climate 33, 21; 10.1175/JCLI-D-19-0670.1

REFERENCES

  • Babić, K., B. Adler, N. Kalthoff, H. Andersen, C. Dione, F. Lohou, M. Lothon, and X. Pedruzo-Bagazgoitia, 2019: The observed diurnal cycle of low-level stratus clouds over southern West Africa: A case study. Atmos. Chem. Phys., 19, 12811299, https://doi.org/10.5194/acp-19-1281-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bony, S., and J.-L. Dufrense, 2005: Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models. Geophys. Res. Lett., 32, L20806, https://doi.org/10.1029/2005GL023851.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bretherton, C. S., P. N. Blossey, and C. R. Jones, 2013: Mechanisms of marine low cloud sensitivity to idealized climate perturbations: A single-LES exploration extending the CGILS cases. J. Adv. Model. Earth Syst., 5, 316337, https://doi.org/10.1002/jame.20019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brient, F., and S. Bony, 2013: Interpretation of the positive low-cloud feedback predicted by a climate model under global warming. Climate Dyn., 40, 24152431, https://doi.org/10.1007/s00382-011-1279-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • de Szoeke, S. P., K. L. Verlinden, S. E. Yuter, and D. B. Mechem, 2016: The time scales of variability of marine low clouds. J. Climate, 29, 64636481, https://doi.org/10.1175/JCLI-D-15-0460.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ek, M. B., and L. Mahrt, 1994: Daytime evolution of relative humidity at the boundary layer top. Mon. Wea. Rev., 122, 27092721, https://doi.org/10.1175/1520-0493(1994)122<2709:DEORHA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., 1985: Sea surface temperature anomalies, planetary waves, and air-sea feedback in the middle latitudes. Rev. Geophys., 23, 357390, https://doi.org/10.1029/RG023i004p00357.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., and K. Hasselmann, 1977: Stochastic climate models, Part II Application to sea-surface temperature anomalies and thermocline variability. Tellus, 29, 289305, https://doi.org/10.3402/TELLUSA.V29I4.11362.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., and R. W. Reynolds, 1983: Testing a dynamical model for mid-latitude sea surface temperature anomalies. J. Phys. Oceanogr., 13, 11311145, https://doi.org/10.1175/1520-0485(1983)013<1131:TADMFM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frankignoul, C., N. Sennéchael, Y.-O. Kwon, and M. A. Alexander, 2011: Influence of the meridional shifts of the Kuroshio and the Oyashio Extensions on the atmospheric circulation. J. Climate, 24, 762777, https://doi.org/10.1175/2010JCLI3731.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 54195454, https://doi.org/10.1175/JCLI-D-16-0758.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hartmann, D. L., and D. A. Short, 1980: On the use of earth radiation budget statistics for studies of clouds and climate. J. Atmos. Sci., 37, 12331250, https://doi.org/10.1175/1520-0469(1980)037<1233:OTUOER>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Johnson, G. C., S. Schmidtko, and J. M. Lyman, 2012: Relative contributions of temperature and salinity to seasonal mixed layer density changes and horizontal density gradients. J. Geophys. Res., 117, C04015, https://doi.org/10.1029/2011JC007651.

    • Search Google Scholar
    • Export Citation
  • Kawamura, R., T. Murakami, and B. Wang, 1996: Tropical and mid-latitude 45-day perturbations over the Western Pacific during the northern summer. J. Meteor. Soc. Japan, 74, 867890, https://doi.org/10.2151/JMSJ1965.74.6_867.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Klein, S. A., 1997: Synoptic variability of low-cloud properties and meteorological parameters in the subtropical trade wind boundary layer. J. Climate, 10, 20182039, https://doi.org/10.1175/1520-0442(1997)010<2018:SVOLCP>2.0.CO;2.

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

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Klein, S. A., D. L. Hartmann, and J. R. Norris, 1995: On the relationships among low-cloud structure, sea surface temperature, and atmospheric circulation in the summertime Northeast Pacific. J. Climate, 8, 11401155, https://doi.org/10.1175/1520-0442(1995)008<1140:OTRALC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Klein, S. A., A. Hall, J. R. Norris, and R. Pincus, 2017: Low-cloud feedbacks from cloud-controlling factors: A review. Surv. Geophys., 38, 13071329, https://doi.org/10.1007/s10712-017-9433-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 548, https://doi.org/10.2151/JMSJ.2015-001.

    • Crossref
    • 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, https://doi.org/10.1175/JCLI-D-11-00478.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mace, G. G., Q. Zhang, M. Vaughan, R. Marchand, G. Stephens, C. Trepte, and D. Winker, 2009: A description of hydrometeor layer occurrence statistics derived from the first year of merged CloudSat and CALIPSO data. J. Geophys. Res., 114, D00A26, https://doi.org/10.1029/2007JD009755.

    • Search Google Scholar
    • Export Citation
  • Mauger, G. S., and J. R. Norris, 2010: Assessing the impact of meteorological history on subtropical cloud fraction. J. Climate, 23, 29262940, https://doi.org/10.1175/2010JCLI3272.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miyamoto, A., H. Nakamura, and T. Miyasaka, 2018: Influence of the subtropical high and storm track on low-cloud fraction and its seasonality over the South Indian Ocean. J. Climate, 31, 40174039, https://doi.org/10.1175/JCLI-D-17-0229.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moisan, J. R., and P. P. Niiler, 1998: The seasonal heat budget of the North Pacific: Net heat flux and heat storage rates (1950–1990). J. Phys. Oceanogr., 28, 401421, https://doi.org/10.1175/1520-0485(1998)028<0401:TSHBOT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mori, M., and M. Watanabe, 2008: The growth and triggering mechanisms of the PNA: A MJO-PNA coherence. J. Meteor. Soc. Japan, 86, 213236, https://doi.org/10.2151/jmsj.86.213.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Morioka, Y., T. Tozuka, and T. Yamagata, 2010: Climate variability in the southern Indian Ocean as revealed by self-organizing maps. Climate Dyn., 35, 10591072, https://doi.org/10.1007/s00382-010-0843-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Myers, T. A., and J. R. Norris, 2016: Reducing the uncertainty in subtropical cloud feedback. Geophys. Res. Lett., 43, 21442148, https://doi.org/10.1002/2015GL067416.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Myers, T. A., C. R. Mechoso, and M. J. DeFlorio, 2018: Coupling between marine boundary layer clouds and summer-to-summer sea surface temperature variability over the North Atlantic and Pacific. Climate Dyn., 50, 955969, https://doi.org/10.1007/s00382-017-3651-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Norris, J. R., and C. B. Leovy, 1994: Interannual variability in stratiform cloudiness and sea surface temperature. J. Climate, 7, 19151925, https://doi.org/10.1175/1520-0442(1994)007<1915:IVISCA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Norris, J. R., and S. F. Iacobellis, 2005: North Pacific cloud feedbacks inferred from synoptic-scale dynamic and thermodynamic relationships. J. Climate, 18, 48624878, https://doi.org/10.1175/JCLI3558.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ogawa, F., and T. Spengler, 2019: Prevailing surface wind direction during air–sea heat exchange. J. Climate, 32, 56015617, https://doi.org/10.1175/JCLI-D-18-0752.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pan, L.-L., and T. Li, 2008: Interactions between the tropical ISO and midlatitude low-frequency flow. Climate Dyn., 31, 375388, https://doi.org/10.1007/s00382-007-0272-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7, 952956, https://doi.org/10.1175/1520-0485(1977)007<0952:IMITUO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Platnick, S., P. Hubanks, K. Meyer, and M. D. King, 2015: MYD08 _ D3 - MODIS/Aqua Aerosol Cloud Water Vapor Ozone Daily L3 Global 1Deg CMG. NASA MODIS Adaptive Processing System, Goddard Space Flight Center, accessed 21 February 2017, https://doi.org/10.5067/MODIS/MYD08_D3.061.

    • Crossref
    • Export Citation
  • Qiu, B., and K. A. Kelly, 1993: Upper-ocean heat balance in the Kuroshio Extension region. J. Phys. Oceanogr., 23, 20272041, https://doi.org/10.1175/1520-0485(1993)023<2027:UOHBIT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Qu, X., A. Hall, S. A. Klein, and P. M. Caldwell, 2014: On the spread of changes in marine low cloud cover in climate model simulations of the 21st century. Climate Dyn., 42, 26032626, https://doi.org/10.1007/s00382-013-1945-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Qu, X., A. Hall, S. A. Klein, and A. M. DeAngelis, 2015: Positive tropical marine low-cloud cover feedback inferred from cloud-controlling factors. Geophys. Res. Lett., 42, 77677775, https://doi.org/10.1002/2015GL065627.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., N. A. Rayner, T. M. Smith, D. C. Stokes, and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 16091625, https://doi.org/10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, L., T. Li, and T. Zhou, 2012: Intraseasonal SST variability and air–sea interaction over the Kuroshio Extension region during boreal summer. J. Climate, 25, 16191634, https://doi.org/10.1175/JCLI-D-11-00109.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, L., T. Li, T. Zhou, and X. Rong, 2013: Origin of the intraseasonal variability over the North Pacific in boreal summer. J. Climate, 26, 12111229, https://doi.org/10.1175/JCLI-D-11-00704.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wielicki, B. A., B. R. Barkstrom, E. F. Harrison, R. B. Lee III, G. L. 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, https://doi.org/10.1175/1520-0477(1996)077<0853:CATERE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wood, R., 2012: Stratocumulus clouds. Mon. Wea. Rev., 140, 23732423, https://doi.org/10.1175/MWR-D-11-00121.1.

  • Wood, R., and C. S. Bretherton, 2006: On the relationship between stratiform low cloud cover and lower-tropospheric stability. J. Climate, 19, 64256432, https://doi.org/10.1175/JCLI3988.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, H., S. Xie, and Y. Wang, 2005: Subseasonal variability of the southeast Pacific stratus cloud deck. J. Climate, 18, 131142, https://doi.org/10.1175/JCLI3250.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yasunari, T., 1979: Cloudiness fluctuations associated with the Northern Hemisphere summer monsoon. J. Meteor. Soc. Japan, 57, 227242, https://doi.org/10.2151/JMSJ1965.57.3_227.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yu, L., X. Jin, and A. R. Weller, 2008: Multidecade global flux datasets from the Objectively Analyzed Air-Sea Fluxes (OAFlux) Project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. OAFlux Project Tech. Rep. OA-2008-01, Woods Hole Oceanographic Institution, Woods Hole, MA, 64 pp.

  • Zelinka, M. D., K. M. Grise, S. A. Klein, C. Zhou, A. M. DeAngelis, and M. W. Christensen, 2018: Drivers of the low-cloud response to poleward jet shifts in the North Pacific in observations and models. J. Climate, 31, 79257947, https://doi.org/10.1175/JCLI-D-18-0114.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save