• Allan, R. P., , A. Slingo, , S. F. Milton, , and M. E. Brooks, 2007: Evaluation of the Met Office global model using Geostationary Earth Radiation Budget (GERB) data. Quart. J. Roy. Meteor. Soc., 133, 19932010.

    • Search Google Scholar
    • Export Citation
  • Bergman, J. W., , and M. L. Salby, 1996: Diurnal variations of cloud cover and their relationship to climatological conditions. J. Climate, 9, 28022820.

    • Search Google Scholar
    • Export Citation
  • Bergman, J. W., , and M. L. Salby, 1997: The role of cloud diurnal variations in the time-mean energy budget. J. Climate, 10, 11141124.

    • Search Google Scholar
    • Export Citation
  • Bloom, S. A., and Coauthors, 2005: Documentation and validation of the Goddard Earth Observing System (GEOS) data assimilation system–Version 4. Technical Report Series on Global Modeling and Data Assimilation, Vol. 26, NASA Tech. Memo. NASA/TM-25-104606, 181 pp.

  • Chung, E.-S., , B.-J. Sohn, , and J. Schmetz, 2009: Diurnal variation of outgoing longwave radiation associated with high cloud and UTH changes from Meteosat-5 measurements. Meteor. Atmos. Phys., 105, 109119, doi:10.1007/s00703-009-0041-8.

    • Search Google Scholar
    • Export Citation
  • Comer, R. E., , A. Slingo, , and R. P. Allan, 2007: Observations of the diurnal cycle of outgoing longwave radiation from the Geostationary Earth Radiation Budget instrument. Geophys. Res. Lett., 34, L02823, doi:10.1029/2006GL028229.

    • Search Google Scholar
    • Export Citation
  • Del Genio, A. D., , and J. Wu, 2010: The role of entrainment in the diurnal cycle of continental convection. J. Climate, 23, 27222738.

  • Duynkerke, P. G., , and J. Teixeira, 2001: Comparison of the ECMWF reanalysis with FIRE I observations: Diurnal variation of marine stratocumulus. J. Climate, 14, 14661478.

    • Search Google Scholar
    • Export Citation
  • Futyan, J. M., , J. E. Russel, , and J. E. Harries, 2005. Determining cloud forcing by cloud type from geostationary satellite data. Geophys. Res. Lett., 32, L08807, doi:10.1029/2004GL022275.

    • Search Google Scholar
    • Export Citation
  • Garreaud, R. D., , and R. Munoz, 2004: The diurnal cycle in circulation and cloudiness over the subtropical southeast Pacific: A modeling study. J. Climate, 17, 16991710.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., , and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105, 11711188.

  • Gruber, A., , and T. S. Chen, 1988: Diurnal variation of outgoing longwave radiation. Int. J. Climatol., 8, 116.

  • Harries, J. E., and Coauthors, 2005: The Geostationary Earth Radiation Budget project. Bull. Amer. Meteor. Soc., 86, 945960.

  • Hartmann, D. L., , and E. E. Recker, 1986: Diurnal variation of outgoing longwave radiation in the tropics. J. Climate Appl. Meteor., 25, 800812.

    • Search Google Scholar
    • Export Citation
  • Hartmann, D. L., , K. J. Kowalewsky, , and M. L. Michelsen, 1991: Diurnal variations of outgoing longwave radiation and albedo from ERBE scanner data. J. Climate, 4, 598617.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., , and K. Woodberry, 1993: The diurnal cycle of tropical convection. J. Geophys. Res., 98 (D9), 16 62316 637.

  • Hong, G., , G. Heygster, , and C. A. M Rodriguez, 2006: Effect of cirrus clouds on the diurnal cycle of tropical deep convective clouds. J. Geophys. Res., 111, D06209, doi:10.1029/2005JD006208.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., , P. A. Arkin, , and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122, 22962311.

    • Search Google Scholar
    • Export Citation
  • Liebmann, B., , and A. Gruber, 1988: Annual variation of the diurnal cycle of outgoing longwave radiation. Mon. Wea. Rev., 116, 16591670.

    • Search Google Scholar
    • Export Citation
  • Lin, X., , D. A. Randall, , and L. D. Fowler, 2000: Diurnal variability of the hydrologic cycle and radiative fluxes: Comparisons between observations and a GCM. J. Climate, 13, 41594179.

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

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , and E. F. Harrison, 1984a: Diurnal variability of regional cloud and clear-sky radiative parameters derived from GOES data. Part I: Analysis method. J. Climate Appl. Meteor., 23, 9931011.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , and E. F. Harrison, 1984b: Diurnal variability of regional cloud and clear-sky radiative parameters derived from GOES data. Part II: November 1978 cloud distributions. J. Climate Appl. Meteor., 23, 10121031.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , and E. F. Harrison, 1984c: Diurnal variability of regional cloud and clear-sky radiative parameters derived from GOES data. Part III: November 1978 radiative parameters. J. Climate Appl. Meteor., 23, 10321051.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , D. F. Young, , and E. F. Harrison, 1991: Examination of the relationship between outgoing infrared window and total longwave fluxes using satellite data. J. Climate, 4, 11141133.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., and Coauthors, 2011: CERES edition-2 cloud property retrievals using TRMM VIRS and Terra and Aqua MODIS data—Part I: Algorithms. IEEE Trans. Geosci. Remote Sens., 49, 43744400.

    • Search Google Scholar
    • Export Citation
  • Morcrette, J.-J., 1991: Evaluation of model-generated cloudiness: Satellite-observed and model-generated diurnal variability of brightness temperature. Mon. Wea. Rev., 119, 12051224.

    • Search Google Scholar
    • Export Citation
  • Nesbitt, S. W., , and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16, 14561475.

    • Search Google Scholar
    • Export Citation
  • Nowicki, S. M. J., , and C. J. Merchant, 2004: Observations of diurnal and spatial variability of radiative forcing by equatorial deep convective clouds. J. Geophys. Res., 109, D11202, doi:10.1029/2003JD004176.

    • Search Google Scholar
    • Export Citation
  • Randall, D. A., , Harshvardhan, , and D. A. Dazlich, 1991: Diurnal variability of the hydrologic cycle in a general circulation model. J. Atmos. Sci., 48, 4061.

    • Search Google Scholar
    • Export Citation
  • Raschke, E., , and W. R. Bandeen, 1970: The radiation balance of the planet earth from radiation measurements of the satellite Nimbus II. J. Appl. Meteor., 9, 215238.

    • Search Google Scholar
    • Export Citation
  • Rozendaal, M. A., , C. B. Leovy, , and S. A. Klein, 1995: An observational study of diurnal variation of marine stratiform cloud. J. Climate, 8, 17951809.

    • Search Google Scholar
    • Export Citation
  • Schmetz, J., , and Q. Liu, 1988: Outgoing longwave radiation and its diurnal variation at regional scales derived from Meteosat. J. Geophys. Res., 93 (D9), 11 19211 204.

    • Search Google Scholar
    • Export Citation
  • Short, D. A., , and J. M. Wallace, 1980: Satellite-inferred morning-to-evening cloudiness changes. Mon. Wea. Rev., 108, 11601169.

  • Slingo, A., , K. I. Hodges, , and G. J. Robinson, 2004: Simulation of the diurnal cycle in a climate model and its evaluation using data from Meteosat 7. Quart. J. Roy. Meteor. Soc., 130, 14491467.

    • Search Google Scholar
    • Export Citation
  • Smith, G. L., , and D. A. Rutan, 2003: The diurnal cycle of outgoing longwave radiation from Earth Radiation Budget Experiment measurements. J. Atmos. Sci., 60, 15291542.

    • Search Google Scholar
    • Export Citation
  • Soden, B. J., 2000: The diurnal cycle of convection, clouds, and water vapor in the tropical upper troposphere. Geophys. Res. Lett., 27, 21732176.

    • Search Google Scholar
    • Export Citation
  • Sui, C.-H., , K.-M. Lau, , Y. N. Takayabu, , and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54, 639655.

    • Search Google Scholar
    • Export Citation
  • Tian, B., , B. J. Soden, , and X. Wu, 2004: Diurnal cycle of convection, clouds, and water vapor in the tropical upper troposphere: Satellites versus a general circulation model. J. Geophys. Res., 109, D10101, doi:10.1029/2003JD004117.

    • Search Google Scholar
    • Export Citation
  • Turton, J. D., , and S. Nicholls, 1987: A study of the diurnal variation of stratocumulus using a multiple mixed layer model. Quart. J. Roy. Meteor. Soc., 113, 9691009.

    • Search Google Scholar
    • Export Citation
  • Wu, M.-L. C., , S. D. Schubert, , M. J. Suarez, , and P. J. Pegion, 2006: Seasonality and meridional propagation of the MJO. J. Climate, 19, 19011921.

    • Search Google Scholar
    • Export Citation
  • Yang, G.-Y., , and J. Slingo, 2001: The diurnal cycle in the tropics. Mon. Wea. Rev., 129, 784801.

  • Yang, S., , K.-S. Kuo, , and E. A. Smith, 2008: Persistent nature of secondary diurnal modes of precipitation over oceanic and continental regimes. J. Climate, 21, 41154131.

    • Search Google Scholar
    • Export Citation
  • Young, D. F., , P. Minnis, , D. R. Doelling, , G. G. Gibson, , and T. Wong, 1998: Temporal interpolation methods for the Clouds and the Earth’s Radiant Energy System (CERES) experiment. J. Appl. Meteor., 37, 572590.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Variance explained by the first diurnal cycle harmonic in the CERES-Terra SYN Ed2rev1 68-month, 3-hourly composite: (top) OLR, (middle) LWCF, and (bottom) OLRCLR.

  • View in gallery

    CERES-Terra SYN Ed2rev1 68-month, 3-hourly composite first harmonic amplitude and phase: (a) AOLR,clim, (b) ALWCF,clim, (c) AOLRCLR,clim (units: W m−2), (d) POLR,clim, (e) PLWCF,clim, and (f) POLRCLR,clim (units: h). Regions contoured in white are not statistically significant at the 95% confidence level.

  • View in gallery

    CERES-Terra SYN Ed2rev1 first harmonic amplitude seasonal variability expressed as the standard deviation over the 12-month annual cycle: (a) AOLR, (b) ALWCF, and (c) AOLRCLR (units: W m−2), and (d) POLR and (e) PLWCF (units: h). OLRCLR is not shown because the variability is less than 1 h in the tropics.

  • View in gallery

    Three-month season when (top) AOLR and (bottom) ALWCF are maximum: MAM (blue), JJA (green), SON (yellow), and DJF (red).

  • View in gallery

    Climatological and seasonal POLR frequency of occurrence for 1° × 1° grid boxes in the (top left) SPCZ, (top right) Indian Ocean, (bottom left) western Pacific, and (bottom right) Atlantic ITCZ regions.

  • View in gallery

    Climatological OLR diurnal evolution histograms in several regions: (a) central Australia (10.0°–30.0°S, 115.0°–150.0°E), (b) northern Africa (10.0°–30.0°N, 0.0°–30.0°E), (c) central South America (0.0°–25.0°S, 50.0°–70.0°W), (d) Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E), (e) western Pacific (10.0°N–10.0°S, 160.0°E–180.0°), and (f) Peruvian MSc (10.0°–25.0°S, 80.0°–95.0°W). The y axis is OLR (W m−2), and the x axis is local time.

  • View in gallery

    As in Fig. 6, but for climatological LWCF. The y axis is LWCF (W m−2), and the x axis is local time.

  • View in gallery

    Regional relationships between LWCF and TS for (top) central South America (0.0°–25.0°S, 50.0°–70.0°W) and (bottom) central Africa (0.0°–15.0°N, 15.0°–35.0°E) at time of (left) OLRMAX and (right) OLRMIN. The bin width of TS is 5 K. Dashed diagonal lines represent constant OLR, where OLRCLR is computed using TS and an effective transmittance of 0.6.

  • View in gallery

    Three-month seasonal OLR diurnal evolution histograms over central South America (0.0°–25.0°S, 50.0°–70.0°W): (top left) MAM, (top right) JJA, (bottom left) SON, and (bottom right) DJF. Contour interval is 1%. Positive (negative) frequency anomalies are contoured with solid (dashed) lines.

  • View in gallery

    Three-month seasonal LWCF diurnal evolution histograms over the Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E): (top left) MAM, (top right) JJA, (bottom left) SON, and (bottom right) DJF. Contour interval is 0.2%. Positive (negative) frequency anomalies are contoured with solid (dashed) lines.

  • View in gallery

    Annual cycle of POLR (solid) and OLR (dotted) in the (top left) SPCZ (10.0°–20.0°S, 170°E–150°W), (top right) Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E), (bottom left) western Pacific (10.0°N–10.0°S, 160.0°E–180.0°), and (bottom right) Atlantic ITCZ (0.0°–10.0°N, 20.0°–50.0°W). Diurnal cycle harmonic is computed after averaging over all 1° × 1° grid boxes in the region.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 254 254 18
PDF Downloads 224 224 26

Tropical Outgoing Longwave Radiation and Longwave Cloud Forcing Diurnal Cycles from CERES

View More View Less
  • 1 Climate Science Branch, NASA Langley Research Center, Hampton, Virginia
© Get Permissions
Full access

Abstract

The diurnal cycle is a fundamental earth system variability driven by daily variations in solar insolation. Understanding diurnal variability is important for characterizing top-of-atmosphere and surface energy budgets. Climatological and seasonal first diurnal cycle harmonics of outgoing longwave radiation (OLR) and longwave cloud forcing (LWCF) are investigated using the Clouds and the Earth’s Radiant Energy System (CERES) synoptic 3-hourly data. A comparison with previous studies indicates generally similar results. However, the results indicate that the CERES OLR diurnal cycle amplitudes are 10%–20% larger in desert regions than previous analyses. This difference results from the temporal interpolation technique overestimating the daily maximum OLR. OLR diurnal cycle amplitudes in other tropical regions agree with previous work. Results show that the diurnal maximum and minimum OLR variability contributes equally to the total OLR variance over ocean; however, over land the diurnal maximum OLR variance contributes at least 50% more to the total OLR variability than the minimum OLR. The differences in maximum and minimum daily OLR variability are largely due to differences in surface temperature standard deviations at these times, about 5–6 and 3–4 K, respectively. The OLR variance at diurnal maximum and minimum is also influenced by negative and positive correlations, respectively, between LWCF and clear-sky OLR. The anticorrelation between LWCF and clear-sky OLR at diurnal OLR maximum indicates smaller cloud fractions at warmer surface temperatures. The relationship between LWCF and clear-sky OLR at diurnal minimum OLR appears to result from a preference for deep convection, more high clouds, and larger LWCF values to occur with warmer surface temperatures driving a narrower diurnal minimum OLR distribution.

Corresponding author address: Patrick Taylor, NASA Langley Research Center, 21 Langley Blvd., Mail Stop 420, Hampton, VA 23681. E-mail: patrick.c.taylor@nasa.gov

Abstract

The diurnal cycle is a fundamental earth system variability driven by daily variations in solar insolation. Understanding diurnal variability is important for characterizing top-of-atmosphere and surface energy budgets. Climatological and seasonal first diurnal cycle harmonics of outgoing longwave radiation (OLR) and longwave cloud forcing (LWCF) are investigated using the Clouds and the Earth’s Radiant Energy System (CERES) synoptic 3-hourly data. A comparison with previous studies indicates generally similar results. However, the results indicate that the CERES OLR diurnal cycle amplitudes are 10%–20% larger in desert regions than previous analyses. This difference results from the temporal interpolation technique overestimating the daily maximum OLR. OLR diurnal cycle amplitudes in other tropical regions agree with previous work. Results show that the diurnal maximum and minimum OLR variability contributes equally to the total OLR variance over ocean; however, over land the diurnal maximum OLR variance contributes at least 50% more to the total OLR variability than the minimum OLR. The differences in maximum and minimum daily OLR variability are largely due to differences in surface temperature standard deviations at these times, about 5–6 and 3–4 K, respectively. The OLR variance at diurnal maximum and minimum is also influenced by negative and positive correlations, respectively, between LWCF and clear-sky OLR. The anticorrelation between LWCF and clear-sky OLR at diurnal OLR maximum indicates smaller cloud fractions at warmer surface temperatures. The relationship between LWCF and clear-sky OLR at diurnal minimum OLR appears to result from a preference for deep convection, more high clouds, and larger LWCF values to occur with warmer surface temperatures driving a narrower diurnal minimum OLR distribution.

Corresponding author address: Patrick Taylor, NASA Langley Research Center, 21 Langley Blvd., Mail Stop 420, Hampton, VA 23681. E-mail: patrick.c.taylor@nasa.gov

1. Introduction

The diurnal cycle is a fundamental mode of Earth system variability encompassing critical climate system processes, namely, radiation and convection. Forced by the 24-h period of solar insolation, the diurnal cycle represents externally forced system variability. Significant diurnal cycle signals are evident in many geophysical datasets, including temperature, water vapor, clouds, radiation, and convective precipitation (Minnis and Harrison 1984a,b,c; Hartmann and Recker 1986; Hartmann et al. 1991; Randall et al. 1991; Janowiak et al. 1994; Bergman and Salby 1996; Lin et al. 2000; Soden 2000; Yang and Slingo 2001; Slingo et al. 2004; Tian et al. 2004). The diurnal cycle structure influences the top-of-atmosphere (TOA) and surface energy budgets (Rozendaal et al. 1995; Bergman and Salby 1997; Loeb et al. 2009; Del Genio and Wu 2010). Therefore, understanding diurnal cycle behavior provides insight into fundamental physical processes important for understanding climate.

Initial interest in the outgoing longwave radiation (OLR) diurnal cycle stemmed from attempts to determine the Earth energy budget from sun-synchronous satellite measurements (Raschke and Bandeen 1970). Raschke and Bandeen (1970) present the first indications of OLR diurnal cycle characteristics using noon-minus-midnight differences, identifying large OLR diurnal cycles over land in desert and convective regions. Short and Wallace (1980) also used day–night differences to investigate diurnal variations in cloudiness, demonstrating early morning peaks in both high and low clouds over convective and marine stratocumulus (MSc) regions, respectively. Hartmann and Recker (1986) combined OLR data from several sun-synchronous satellites to move beyond day–night differences to investigate the tropical OLR 24-h harmonic. The results indicate 20–25 W m−2 OLR diurnal cycle amplitude over desert regions, a 10–15 W m−2 OLR diurnal cycle over land regions with frequent convection, and OLR diurnal cycle amplitudes less than 5 W m−2 over ocean. Using similar data, Gruber and Chen (1988) looked more closely at OLR first harmonic amplitude–phase relationships, finding significant seasonality over land regions and weak seasonality over ocean. Further analysis of the OLR first diurnal harmonic annual cycle suggests a link between the seasonal OLR variability and the OLR first diurnal cycle harmonic (Liebmann and Gruber 1988).

Several studies use geostationary (GEO) satellite data to investigate the tropical cloud and radiative diurnal cycle. Minnis and Harrison (1984a,b,c) investigate the cloud cover and TOA flux diurnal cycles in the Peruvian MSc and South American regions, identifying a robust diurnal cycle in low-cloud fraction. Schmetz and Liu (1988) reveal similar cloud diurnal cycle behavior over Namibian MSc and central African regions. More recently, the Geostationary Earth Radiation Budget (GERB; Harries et al. 2005) instrument is used to elucidate the OLR diurnal cycle over Africa (Nowicki and Merchant 2004; Futyan et al. 2005; Comer et al. 2007) and to validate weather forecast models (Allan et al. 2007).

Previous studies quantify the influence of diurnal cycle evolution on the TOA time-mean energy budget, indicating significant impacts. Bergman and Salby (1997) quantify cloud diurnal cycle radiative energy budget impacts, showing a 5–15 W m−2 TOA shortwave (SW) and 1–5 W m−2 TOA longwave (LW) effect on the time-mean energy budget. Rozendaal et al. (1995) indicate similar diurnal cycle impacts in MSc cloud regions. Loeb et al. (2009) find up to 30 W m−2 regional biases in annual-mean TOA irradiances measured from a sun-synchronous radiometer in MSc and tropical land convective regions due to neglecting diurnal variations. Moreover, a 1-h shift in the MSc cloud amount diurnal cycle can lead to a time-mean energy budget error larger than neglecting diurnal variations altogether (Bergman and Salby 1997). These studies demonstrate the importance of cloud to the TOA flux diurnal cycle; however, few studies analyze the diurnal cycle of longwave cloud forcing (LWCF).

The diurnal cycle of the cloud radiative impacts on the OLR diurnal cycle can be quantified using LWCF. LWCF is defined as the difference between OLR and clear-sky OLR (OLRCLR), as shown:
e1
Nowicki and Merchant (2004) use LWCF and shortwave cloud forcing (SWCF) determined from GERB to elucidate the radiative impact of the deep convective cloud diurnal cycle on TOA flux over central Africa and the equatorial Atlantic Ocean. The analysis reveals a sensitivity of the net cloud forcing to the deep convective cloud diurnal cycle phase; a +2-h shift can change the net cloud forcing by +10 W m−2. Futyan et al. (2005) using GERB and coregistered imager data show that large diurnal variability in LWCF over convective regions of Africa is due to high cloud fraction. The LWCF definition is sensitive to OLRCLR. For instance, a cloud with the same area coverage, optical depth, and temperature has a larger LWCF if the surface temperature is higher. It is not straightforward to determine OLRCLR instantaneously from satellite measurements.

In the present study, Clouds and the Earth’s Radiant Energy System (CERES) data are used to investigate diurnal variations in OLR and LWCF. CERES is in sun-synchronous orbit; therefore, to resolve the diurnal cycle, it is merged with GEO satellite data. Harmonic analysis is employed to characterize the climatological and 3-month seasonal diurnal cycle amplitude and phase. The dataset and harmonic analysis methodology are described in section 2. Section 3 summarizes the first diurnal cycle harmonic analysis results from a tropics-wide and regional perspective. A comparison of results with previous investigations is presented in section 4. Last, section 5 provides a summary and conclusions.

2. Data and methodology

CERES-Terra synoptic (SYN) edition 2 revision 1 (Ed2rev1) OLR and LWCF data contain a 3-hourly time resolution on a 1° × 1° grid spanning the time interval from March 2000 to October 2005. Because of CERES orbit, an enhanced temporal interpolation scheme is developed (Young et al. 1998) to account for the diurnal variation using global GEO satellite visible (VIS; 0.65 μm) and infrared (IR; 11.0 μm) channels. The LW flux technique is described here; the SW diurnal cycle is not described, as it is beyond the present scope. Further details for treating diurnal cycles are discussed in Young et al. (1998) and D. R. Doelling (2012, personal communication).

Briefly, the LW GEO-enhanced time interpolation technique (Young et al. 1998; D. R. Doelling 2012, personal communication; GEO-interp) proceeds in three steps: 1) GEO 11.0-μm radiance is converted to broadband radiance using an empirical regression relationship, 2) GEO broadband radiance is converted to irradiance, and 3) derived broadband irradiance is normalized to CERES. GEO 11.0-μm radiance I11μm is converted to a broadband radiance Rbroadband using
e2
The coefficients a0,1,2,3 are determined monthly by land surface type. In (2), rh is the column-averaged relative humidity obtained from Goddard Earth Observing System Data Assimilation System reanalysis version 4 (GEOS-4; Bloom et al. 2005). Broadband OLR is computed using angular integration with the form
e3
In (3), γ(θ) is a limb-darkening correction (Minnis et al. 1991) at viewing zenith angle θ that converts the irradiance to its equivalent nadir value; 6.18 results from the hemispheric integration of the anisotropic empirical limb-darkening function (Minnis et al. 1991; Young et al. 1998). Last, regional normalization to CERES-observed flux adjusts for regional variability in the narrowband-to-broadband regression relationships (Young et al. 1998), preserving CERES calibration. The GEO-interp process yields a global-mean monthly flux bias of 0.1 W m−2 and regional root-mean-square errors less than 1 W m−2 (D. R. Doelling 2012, personal communication). The largest uncertainty in 5-yr global-mean CERES OLR is from absolution calibration: 3.7 W m−2 (Loeb et al. 2009). Absolute calibration uncertainty does not affect this analysis because the mean value is removed to analyze diurnal anomalies.

Defining OLRCLR from satellite radiometer observations (e.g., CERES) is inherently different than during radiative transfer model simulations. For a 20-km CERES footprint, an OLR value is considered to be OLRCLR only when a scene is 99% cloud free as determined by the CERES–Moderate Resolution Imaging Spectroradiometer (MODIS) cloud retrieval algorithm (Minnis et al. 2011). As a result, CERES OLRCLR represents a different temperature and water vapor vertical profile than all-sky OLR observations affecting LWCF. OLRCLR at non-CERES overpass times is determined using a half-sine fit, as described by Young et al. (1998) and D. R. Doelling (2012, personal communication), that follows the climatological diurnal variation of surface temperature (Minnis and Harrison 1984a). In cases where no 99% cloud-free pixels are observed OLRCLR is not reported, implying potential sampling problems. Therefore, spatial and temporal averaging is required to overcome limited OLRCLR sampling. CERES SYN data also contain OLRCLR computed from GEOS-4 reanalysis temperature and humidity profiles. Monthly-mean computed OLRCLR is used to evaluate observed OLRCLR.

To a first order, the diurnal cycle is well characterized by the two parameters from a first harmonic decomposition (Hartmann and Recker 1986; Yang et al. 2008): 1) amplitude and 2) phase. As in previous analysis, Fourier decomposition of 3-hourly diurnal cycle composites are used to compute the first diurnal cycle harmonic amplitude and phase (e.g., Yang et al. 2008). Here,
e4
where X′(t) is the diurnal anomaly, defined as the difference between the daily mean and the 3-hourly quantity; t is time in hours, A represents the amplitude, and P represents the phase angle. For clarity, the phase is discussed as the local solar time (LST) of diurnal cycle maximum. For the remaining discussion, Avariable and Pvariable will refer to the first diurnal harmonic amplitude and phase of the subscripted variable, respectively.

The first diurnal cycle harmonic is a valid expression for the diurnal cycle (e.g., Hartmann and Recker 1986; Gruber and Chen 1988; Yang and Slingo 2001). Figure 1 shows the variance explained by the first harmonic fit to the 3-hourly, 68-month climatological composites of OLR, LWCF, and OLRCLR. The first harmonic over tropical (30°N–30°S) land regions explains more than 70% of the variance for OLR, LWCF, and OLRCLR. The fraction of explained variance of the OLR and LWCF first harmonic is less than 50% in many equatorial ocean convective regions [e.g., Pacific intertropical convergence zone (ITCZ), South Pacific convergence zone (SPCZ), and Indian Ocean]. Statistical significance of the first harmonic is determined using a Fisher statistical significance test (F test) as in Minnis and Harrison (1984b). At the 95% confidence level, an explained variance greater than 67% indicates statistical significance (Fig. 1, black line). The OLR and LWCF harmonic fits are statistically significant in most tropical land regions, but the statistical significance over ocean is regionally dependent.

Fig. 1.
Fig. 1.

Variance explained by the first diurnal cycle harmonic in the CERES-Terra SYN Ed2rev1 68-month, 3-hourly composite: (top) OLR, (middle) LWCF, and (bottom) OLRCLR.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

3. Results

a. Tropical TOA LW diurnal cycle amplitude and phase climatological spatial distribution

Climatological and 3-month seasonal OLR, LWCF, and OLRCLR first harmonic diurnal cycle amplitudes and phases computed from CERES-Terra SYN Ed2rev1 data are described in this section. To facilitate discussion, four traditional diurnal cycle categories are used: land convective (e.g., central Africa and central South America), land nonconvective (e.g., Sahara and southern Africa), ocean convective (e.g., Indian Ocean and western Pacific Ocean), and ocean nonconvective (e.g., Peruvian and Namibian MSc regions).

1) Land nonconvective

Significant and robust OLR diurnal cycles are observed in land nonconvective regions. Diurnal evolution in these regions is in response to large diurnal variation of surface temperature indicated by the presence of large AOLR,clim and AOLRCLR,clim with weak ALWCF,clim (Fig. 2). The largest AOLR,clim, greater than 45 W m−2, are observed in the Atacama Desert (15.0°–25.0°S, 70.0°–80.0°W). The northern and southern regions of the African continent, central Australia, and Arabian Peninsula exhibit 25–30 W m−2 AOLR,clim magnitudes with local values exceeding 35 W m−2. Seasonal variations in land nonconvective AOLR magnitudes are 10%–20% (Fig. 3). In Figs. 3 and 4, first harmonic decomposition is applied to monthly, 3-hourly diurnal composites; standard deviations are computed using the 12-month annual cycle. Land nonconvective POLR is consistent across most land regions occurring between 1400–1600 LST (Fig. 2) and indicates little seasonal variation (Fig. 4).

Fig. 2.
Fig. 2.

CERES-Terra SYN Ed2rev1 68-month, 3-hourly composite first harmonic amplitude and phase: (a) AOLR,clim, (b) ALWCF,clim, (c) AOLRCLR,clim (units: W m−2), (d) POLR,clim, (e) PLWCF,clim, and (f) POLRCLR,clim (units: h). Regions contoured in white are not statistically significant at the 95% confidence level.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

Fig. 3.
Fig. 3.

CERES-Terra SYN Ed2rev1 first harmonic amplitude seasonal variability expressed as the standard deviation over the 12-month annual cycle: (a) AOLR, (b) ALWCF, and (c) AOLRCLR (units: W m−2), and (d) POLR and (e) PLWCF (units: h). OLRCLR is not shown because the variability is less than 1 h in the tropics.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

Fig. 4.
Fig. 4.

Three-month season when (top) AOLR and (bottom) ALWCF are maximum: MAM (blue), JJA (green), SON (yellow), and DJF (red).

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

The season of maximum land nonconvective diurnal cycle amplitude varies regionally. In North Africa and the Arabian Peninsula, the largest AOLR values occur in local summer [June–August (JJA); Fig. 5]. The largest AOLR values in southern Africa, the Atacama Desert, and central Australia do not occur in local summer but in September–November (SON) or JJA. In these Southern Hemisphere regions AOLR seasonality results from statistically significant ALWCF during local-summer-reducing AOLR. In the Arabian Peninsula and North Africa ALWCF is insignificant for all seasons, indicating that different mechanisms drive AOLR seasonality between the Northern and Southern Hemisphere arid land regions. The regional variations and AOLR seasonality agree with previous results (e.g., Hartmann and Recker 1986; Liebmann and Gruber 1988).

Fig. 5.
Fig. 5.

Climatological and seasonal POLR frequency of occurrence for 1° × 1° grid boxes in the (top left) SPCZ, (top right) Indian Ocean, (bottom left) western Pacific, and (bottom right) Atlantic ITCZ regions.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

2) Land convective

Significant OLR and LWCF diurnal cycles are observed in tropical land convective regions (Fig. 2). The strongest land convective diurnal cycles are found in equatorial regions: central South America, central Africa, and the Maritime Continent. Off-equatorial land regions possess statistically significant convective cloud diurnal cycles with weaker amplitudes. Land convective regions differ from nonconvective regions because they exhibit ALWCF,clim signals, generally 10–15 W m−2, attributed to strong diurnal variations in high cloud (Futyan et al. 2005); ALWCF,clim and PLWCF,clim (Fig. 2) appear consistent with features of diurnally forced convection that suggest maximum cloudiness in the evening to early morning (Lin et al. 2000; Yang and Slingo 2001; Futyan et al. 2005), which is phase lagged with respect to late afternoon or early evening diurnally forced convective precipitation; POLR,clim, 1000–1400 LST, is consistent with low-cloud development in late morning, causing the daily OLR maximum (OLRMAX) to occur earlier than in land nonconvective regions; and PLWCF,clim, 1800–0200 LST, is driven by the diurnal variation in high cloudiness and is consistent with the evening maximum of convective precipitation and deep convective clouds in the evening over tropical land regions (Janowiak et al. 1994; Yang and Slingo 2001; Nowicki and Merchant 2004; Futyan et al. 2005). Tian et al. (2004) speculate that the behavior of the high cloud and LWCF phase is determined in response to the earlier intense latent heating associated with maximum convective precipitation resulting in a phase lag.

Land convective regions exhibit greater seasonality than land nonconvective regions (Fig. 3). In equatorial land convective regions, seasonality is weak and the maximum AOLR and ALWCF values occur during MAM and SON. Seasonality is larger at 10°N and 10°S in ALWCF; the strongest amplitudes occur during local summer, and the LWCF diurnal cycle becomes insignificant during local winter. In equatorial regions, POLR remains relatively constant with season; however, at 10°N and 10°S, POLR tends to occur earlier during local summer due to increases in late-morning and early-afternoon cloud formation.

Regional variability in land convective diurnal cycle features is partly attributed to propagating convective features driven by cloud effects. In central Africa, PLWCF,clim occurs from 1800 to 2000 LST near the Ethiopian highlands (Comer et al. 2007) and west of the highlands PLWCF,clim occurs in the early morning (Fig. 2). A complex feature related to convective propagation is also evident in northeastern Brazil where coastal PLWCF,clim occurs near 1800 LST, whereas inland PLWCF,clim occurs in the early morning. Complex regional OLR and LWCF diurnal cycle structures are observed (Fig. 2) around the islands of the Maritime Continent (Short and Wallace 1980; Hartmann and Recker 1986; Yang and Slingo 2001), which tend to spread over adjacent ocean (Yang and Slingo 2001). In these oceanic regions, ALWCF,clim tends to be at least 25% larger than over open ocean. Propagating features are evident in POLR and PLWCF (Fig. 2) from the eastern coast of India into the Bay of Bengal and also along the western Mexican coast, noted by Short and Wallace (1980) and Yang and Slingo (2001). The present results confirm that propagating convective features are evident in OLR and LWCF diurnal cycles as well as in cloud fraction.

3) Ocean convective

Many ocean convective regions exhibit a statistically robust diurnal cycle despite smaller AOLR,clim and ALWCF,clim than over land. Observed AOLR,clim and ALWCF,clim range from 2 to 5 W m−2, in line with previous analysis (Hartmann and Recker 1986; Gruber and Chen 1988; Schmetz and Liu 1988). A smaller oceanic AOLR,clim diurnal cycle signal results from a generally weak diurnal sea surface temperature (SST) response evident in AOLRCLR,clim (Fig. 2). Further, AOLR and ALWCF are nearly identical over ocean, indicating that AOLR is determined almost entirely by diurnal cloud evolution. Previous results (e.g., Sui et al. 1997; Chung et al. 2009) suggest this result by demonstrating weak sea surface temperature and water vapor diurnal variations. Values of POLR,clim range from 0400 to 1400 LST in ocean convective regions, indicating more variability than over land. Contrary to previous reports (Hartmann and Recker 1986; Gruber and Chen 1988), many regions of oceanic convection (e.g., Pacific ITCZ, SPCZ, and Indian Ocean) do not exhibit a statistically significant OLR diurnal cycle at 95% confidence, discussed further in section 4; PLWCF,clim exhibits a range of values from evening to early morning, coinciding with the oceanic convective precipitation diurnal cycle (Gray and Jacobson 1977; Janowiak et al. 1994; Nesbitt and Zipser 2003). Few previous analyses of LWCF diurnal cycle are available; however, Nowicki and Merchant (2004) find similar diurnal cycle amplitude and phase behavior over the Atlantic ITCZ.

OLR and LWCF seasonal diurnal cycle amplitude and phase variability differ. Figure 3 indicates a significant POLR seasonality exceeding 3 h in the SPCZ and 2 h in most other ocean convective regions. Larger POLR seasonality in SPCZ is attributed to a larger annual cycle in solar insolation. Seasonal PLWCF variability is larger than POLR in oceanic convective regions. Alternatively, AOLR and ALWCF seasonality is weak with little spatial variability and does not exceed about 1 W m−2. Hartmann and Recker (1986) suggest that seasonal POLR variability in oceanic convective regions exhibits two preferred values related to convective intensity: 1) 0600 LST POLR corresponding to intense convection and 2) a local-noon POLR corresponding to weaker convection. A range of POLR values is evident in Fig. 2, not two preferred phases, especially in the western Pacific and SPCZ. This is better illustrated in regional probability distributions (Fig. 5) where the POLR distribution appears Gaussian in all of the ocean convective regions. The variability of POLR,clim differs among ocean convective regions and no region exhibits a two-preferred-phase behavior, illustrating a difference with Hartmann and Recker (1986).

4) Ocean nonconvective

Robust diurnal cycles are found in ocean nonconvective regions typically populated by MSc and trade cumulus cloud types. The present results are consistent with previous studies (e.g., Turton and Nicholls 1987; Rozendaal et al. 1995; Duynkerke and Teixeira 2001; Garreaud and Munoz 2004). Across MSc regions AOLR,clim and ALWCF,clim are similar, exhibiting values from 3 to 7 W m−2 (Fig. 2). Similarly, POLR,clim and PLWCF,clim are generally uniform across MSc regions with values of 1400–1600 and 0600–0800 LST, respectively. A slightly later POLR,clim, 1600–1800 LST, is observed in the Namibian MSc region but with the same PLWCF,clim. Regional variations in POLR, but not PLWCF, suggest that the timing of the morning cloud maximum may be more robust across MSc regions than the afternoon cloud minimum, which exhibits more variability.

Weak AOLR and ALWCF seasonal variations are evident in ocean nonconvective regions (Fig. 3). The largest AOLR and ALWCF values are found during local summer (Fig. 4); ALWCF exceeds 10 W m−2 in Peruvian and Namibian MSc regions during December–February (DJF). More variability in the season of maximum AOLR and ALWCF is found in the Californian and Canarian MSc; weak POLR seasonal variability, less than 1 h, is also evident in these regions (Fig. 3).

b. Regional focus: Diurnal cycle evolution histograms

This section investigates OLR and LWCF diurnal cycle structure and variability in selected regions that represent one of the four traditional diurnal cycle regimes. The primary analysis method in this section is the diurnal cycle evolution histogram. This tool describes the likelihood of a particular OLR or LWCF value to occur at a given local time relative to the daily probability distribution function (PDF) described in Morcrette (1991) and Yang and Slingo (2001). Evolution histograms are constructed by binning 3-hourly OLR and LWCF data into 5 W m−2 bins. Frequency anomalies are computed as the difference between the 3-hourly and daily PDFs. Positive (negative) frequency anomalies indicate a higher (lower) probability of a particular OLR or LWCF value occurring at a specific time relative to the daily PDF. OLR and LWCF seasonal mean-state variations are removed to isolate seasonal variations in diurnal cycle structure.

The OLR diurnal cycle is determined by the convolution of surface temperature and cloud evolution illustrated by diurnal evolution histograms (Figs. 6 and 7). OLR diurnal cycle behavior in desert regions is driven primarily by surface temperature TS. Central Australia (10.0°–30.0°S, 115.0°–150.0°E; Fig. 6a) and North Africa (10.0°–30.0°N, 0.0°–30.0°E; Fig. 6b) OLR increases from an early morning minimum to a late afternoon maximum following increases in TS. Near 1400–1600 LST POLR occurs, coinciding with the TS diurnal cycle peak. The dry desert atmosphere does not support significant cloud development (Figs. 7a and 7b) to alter the TS-driven OLR diurnal cycle. Desert LWCF values are especially weak, less than 10 W m−2, at the time of OLRMAX.

Fig. 6.
Fig. 6.

Climatological OLR diurnal evolution histograms in several regions: (a) central Australia (10.0°–30.0°S, 115.0°–150.0°E), (b) northern Africa (10.0°–30.0°N, 0.0°–30.0°E), (c) central South America (0.0°–25.0°S, 50.0°–70.0°W), (d) Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E), (e) western Pacific (10.0°N–10.0°S, 160.0°E–180.0°), and (f) Peruvian MSc (10.0°–25.0°S, 80.0°–95.0°W). The y axis is OLR (W m−2), and the x axis is local time.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

Fig. 7.
Fig. 7.

As in Fig. 6, but for climatological LWCF. The y axis is LWCF (W m−2), and the x axis is local time.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

Clouds play a key role in the diurnal evolution in all other diurnal cycle regions. The TS diurnal cycle in land convective regions evolves similar to land nonconvective regions with an afternoon maximum (1400–1600 LST). However, the central South American land convective region (0.0°–25.0°S, 50.0°–70.0°W; Fig. 6c) OLR diurnal cycle peaks about 2 h earlier due to late morning and afternoon cloud formation, illustrated by LWCF (Fig. 7c). Similar regional OLR diurnal cycle evolution is found in other land convective regions as well (e.g., central Africa, 0.0°–15.0°N, 15.0°–35.0°E). Morning to afternoon cloud development in land convective regions is in response to the deepening of the boundary layer and enhanced turbulent mixing associated with increased TS from solar heating. The role of clouds in determining the land convective OLR diurnal cycle structure is also apparent at the time of minimum OLR (OLRMIN). Land convective region OLRMIN does not coincide with minimum TS (0400–0600 LST) but with maximum LWCF (LWCFMAX) and maximum high cloud fraction (Futyan et al. 2005). LWCF diurnal evolution (Fig. 7c) indicates that between 1800 and 2200 LST typical LWCF values exceed 50 W m−2, coinciding with OLRMIN. This LWCF maximum directly corresponds to the maximum diurnal high cloud fraction (Futyan et al. 2005). Both LWCFMAX and high cloud fraction occur in the evening to early morning and are phase lagged with respect to maximum convective precipitation (Yang and Slingo 2001; Tian et al. 2004; Hong et al. 2006). This suggests cloud fraction is a more important determinant of LWCF diurnal evolution as opposed to cloud-top height. The highest convective clouds occur when updraft speeds and precipitation are maximum; however, a significant radiative impact of convective storms appears later in the life cycle (Fig. 7c) after significant cirrus cloud formation (Nowicki and Merchant 2004). Radiative impacts of deep convective clouds are extremely sensitive to the position within the diurnal cycle due to the relative timing of maximum solar insolation and cloud fraction. Nowicki and Merchant (2004) indicate that a 2-h shift in deep convective phase changes the net cloud forcing by 10 W m−2 mainly through impacts on reflected solar radiation.

Ocean convective and nonconvective regions are nearly entirely determined by diurnal cloud variability as from a climatological perspective, SST exhibits little diurnal evolution. Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E; Fig. 6d) and western Pacific (10.0°N–10.0°S, 160.0°E–180.0°; Fig. 6e) OLR diurnal evolution is similar, exhibiting OLRMAX between 1000 and 0200 LST and OLRMIN between 1800 and 0400 LST. This OLRMAX and OLRMIN behavior is consistent with diurnal cloud evolution in these regions. The oceanic convective region diurnal cloud evolution is characterized by three cloud modes: 1) early-morning low cloud, 2) afternoon scattered convection, and 3) nocturnal organized convection (Sui et al. 1997). Ocean convective LWCF diurnal evolution (Figs. 7d and 7e) exhibits a LWCF minimum (LWCFMIN) from 1000 to 0200 LST, coinciding with OLRMAX and low clouds. Into early evening, LWCF values increase and exceed 50 W m−2 by 1600 LST, associated with afternoon scattered convection. LWCFMAX can occur between 1600 and 2000 LST in both regions. A secondary LWCFMAX is evident between 0200 and 0600 LST, associated with nocturnal organized convection (Figs. 7d and 7e) corresponding to increased probabilities of OLRMIN during this period due to increased high cloudiness. Indian Ocean and western Pacific OLR and LWCF diurnal evolution histograms are consistent with the trimodal ocean convective cloud diurnal evolution.

Nonconvective ocean regions exhibit a different OLR diurnal evolution than ocean convective regions but are also mainly driven by cloud evolution. Cumulus and MSc cloud types populate ocean nonconvective diurnal cycle regions. However, only MSc exhibits a robust diurnal cycle. The Peruvian region (10.0°–25.0°S, 80.0°–95.0°W; Fig. 6f) represents a typical MSc diurnal evolution. OLRMIN occurs near sunrise, corresponding to a maximum in LWCF and cloud fraction. Through the day, OLR increases due to the dissipation of MSc clouds in response to solar heating of the cloud layer (Turton and Nicholls 1987), not changes in SST. OLRMAX coincides with LWCFMIN and a minimum in cloudiness occurring in the afternoon. The rapid decay of cloud corresponds to the decoupling of the cloud layer from the surface mixed layer: its moisture source (Turton and Nicholls 1987). This explanation is consistent with our understanding of low-cloud evolution in the absence of SST variations.

1) OLR diurnal cycle maximum and minimum behavior

Contributions to OLR variability can be distributed unevenly throughout the day. Understanding this variability can help identify physical processes that contribute to overall OLR variability. For instance, processes important to determining OLRMAX—for example, shallow convective cloud area (reductions in cloud area relate to increased OLR) and surface temperature—tend to be different from those that determine diurnal OLRMIN, for example, deep convection and high cloud area. Regionally dependent behavior is indicated in OLRMAX and OLRMIN distributions.

Climatological OLR diurnal cycle evolution histograms (Fig. 6) illustrate variability in the OLR PDF as a function of local time as well as differences in OLRMAX and OLRMIN variability. Visually, the vertical packing of contours represents OLRMAX and OLRMIN variability: tight (loose) packing yields smaller (larger) standard deviation. In central Australia (Fig. 6a), larger frequency anomalies and a tighter vertical packing of contours are observed for OLRMIN (Fig. 6a) compared to OLRMAX, 8% and 5%, respectively. Similar behavior is observed in other land nonconvective regions, for example, North Africa (Fig. 6b), but with smaller frequency anomalies. It is evident in Fig. 6a that OLRMAX exhibits more variability than OLRMIN in central Australia; however, this is not found in all regions. Comparing OLRMAX and OLRMIN standard deviations, OLRMAX contributes twice as much, about 20%, to the climatological, regional OLR variance σ2OLR as OLRMIN, about 10%, in North Africa and central Australia. In land convective regions (central South America, Fig. 6c), the OLRMAX and OLRMIN exhibit similar magnitude frequency anomalies, but a looser packing of contours is shown at OLRMAX. Stronger OLRMAX variability is verified by the regional OLRMAX standard deviation, 22 W m−2 compared to 16 W m−2 for OLRMIN; OLRMAX contributes about 15% to σ2OLR, which is 50% more than OLRMIN, about 10%, in both central South America and central Africa. In both land convective and nonconvective regions, the additional variability is attributed to larger TS standard deviations at OLRMAX than at OLRMIN, about 5–6 and 3–4 K, respectively. Ocean convective regions (Figs. 6d and 6e) exhibit similar vertical packing of contours at OLRMAX and OLRMIN, indicating equal contributions to σ2OLR, about 13%. Nonconvective ocean regions (Peruvian MSc, Fig. 6f) possess a very robust diurnal cycle evolution exhibiting similar OLRMAX and OLRMIN variability. These results indicate larger variability in OLRMAX than OLRMIN over land, which suggests that the physical processes driving OLRMAX variability, afternoon TS and cloud fraction, contribute more to σ2OLR. Over ocean, however, climatological OLRMAX and OLRMIN contribute equally to σ2OLR.

The role of clouds in driving climatological OLRMAX and OLRMIN variations is illustrated by regional LWCF climatological diurnal cycle evolution histograms (Fig. 7). In ocean convective (Figs. 7d and 7e) and nonconvective regions (Fig. 7f), LWCFMAX and LWCFMIN possess similar frequency anomalies and vertical packing of contours and contribute equally to OLRMAX and OLRMIN variability. Over land convective and nonconvective regions, LWCFMIN possesses a tighter vertical packing of contours at LWCFMAX, indicating more variability. This behavior in LWCFMAX and LWCFMIN is opposite of the OLRMAX and OLRMIN behavior, respectively. Therefore, a compensating relationship between LWCF and OLRCLR must exist where larger OLRCLR values accompany larger (smaller) LWCF values in the evening (afternoon) to reduce OLRMIN (enhance OLRMAX) variability. Statistically significant (95% confidence) regional correlations—negative and positive at the time of OLRMAX and OLRMIN, respectively—are found in central South America between monthly-mean OLRCLR and LWCF. Similar behavior is also evident in central Africa, central Australia, and North Africa. A certain level of correlation between LWCF and OLRCLR is expected from (1). Further, OLRCLR sampling inconsistencies could influence this correlation. To further investigate this relationship, an independent TS dataset (GEOS-4; Bloom et al. 2005) is used to represent OLRCLR (Fig. 8) showing similar relationships at OLRMAX and OLRMIN. The diagonal dashed lines in Fig. 8 represent lines of constant OLR computed using TS and an effective transmittance of 0.6 to determine OLRCLR. It is hypothesized that in the evening, this relationship stems from more intense convection and larger LWCF preferentially occurring when TS is larger (Fig. 8, right), reducing OLRMIN variability. A feedback between convective intensity and OLRCLR may also be operating where more intense convection generates enhanced subsidence in the surrounding regions, contributing to clear-sky warming and drying, further increasing OLRCLR. The negative correlation at OLRMAX is less obvious (Fig. 8, left); however, the majority of the points correspond to the negative slope portion of the curve, greater than 300 K. This relationship indicates that when TS is higher (high OLRCLR), lower cloud fraction and LWCF are present, which increases OLRMAX variability.

Fig. 8.
Fig. 8.

Regional relationships between LWCF and TS for (top) central South America (0.0°–25.0°S, 50.0°–70.0°W) and (bottom) central Africa (0.0°–15.0°N, 15.0°–35.0°E) at time of (left) OLRMAX and (right) OLRMIN. The bin width of TS is 5 K. Dashed diagonal lines represent constant OLR, where OLRCLR is computed using TS and an effective transmittance of 0.6.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

2) Seasonal variations in OLR and LWCF diurnal cycle structure

Overall, the diurnal cycle OLR evolution structure for land convective and nonconvective regions is statistically robust with limited seasonal variability. Most seasonal diurnal cycle variability occurs in AOLR and ALWCF, and no changes are evident in POLR and PLWCF for either region. The largest seasonal variations in OLR diurnal cycle structure relate to changes in the LWCF diurnal cycle. In central Australia, the smallest AOLR and weakest frequency anomalies are found in DJF coinciding with the strongest ALWCF, 6.0 W m−2—2 times larger than any other season. In central South America, seasonal variability is indicated in the OLRMIN and OLRMAX frequency anomalies (Fig. 9). Frequency anomalies are largest around OLRMAX in March–May (MAM) and SON, and largest around OLRMIN in JJA and DJF. These variations in diurnal cycle structure indicate seasonality in the contributions of OLRMAX and OLRMIN to OLR variance.

Fig. 9.
Fig. 9.

Three-month seasonal OLR diurnal evolution histograms over central South America (0.0°–25.0°S, 50.0°–70.0°W): (top left) MAM, (top right) JJA, (bottom left) SON, and (bottom right) DJF. Contour interval is 1%. Positive (negative) frequency anomalies are contoured with solid (dashed) lines.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

The seasonality of the OLR diurnal cycle structure is most evident in ocean convective regions for OLRMAX. Over the Indian Ocean, significant seasonal OLR variations are evident in the positive frequency anomaly range near local noon. The largest OLRMAX range occurs during DJF and MAM. This range is reduced by about 40% during JJA and SON. This OLRMAX seasonality is evident in other ocean convective regions (i.e., western Pacific), but it is most notable in the Indian Ocean. Contributions from OLRMAX to regional OLR variance are 12% in JJA and SON and 20% in DJF and MAM. Wu et al. (2006) demonstrate that off-equatorial development of Madden–Julian oscillation (MJO) convection is favored during local summer in the Indian Ocean region. An increased frequency of off-equatorial convection over the Indian Ocean could increase the range and variability of OLRMAX through increased cloudiness. Figure 10 demonstrates that the seasonality in OLRMAX is forced by changes in cloud evolution variability. During DJF and MAM, a wider range in LWCF occurs at 1200 LST with enhanced probability of LWCF values up to about 40 W m−2. Alternatively, JJA and SON show LWCF values around 1200 LST to be about 20 W m−2. This suggests seasonal differences in the intensity and frequency of afternoon scattered convection diurnal cloud mode contribute to the seasonality of OLRMAX variability in ocean convective regions.

Fig. 10.
Fig. 10.

Three-month seasonal LWCF diurnal evolution histograms over the Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E): (top left) MAM, (top right) JJA, (bottom left) SON, and (bottom right) DJF. Contour interval is 0.2%. Positive (negative) frequency anomalies are contoured with solid (dashed) lines.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

3) Regional variability in OLR and LWCF diurnal cycle structure

Regional variations are largest in land convective regions. Similar evolution histogram structures and limited regional variability are observed in all analyzed land nonconvective regions: northern Africa (Fig. 6b), southern Africa, Chile, and the Arabian Peninsula (not shown). Regional variability in OLR and LWCF diurnal evolution, however, is observed between central South America and central Africa (not shown). The main differences are smaller OLR and LWCF frequency anomalies in central Africa and a slight shift in the LWCFMAX and OLRMIN from 2000 LST in central South America to 0000–0300 LST in central Africa. It is speculated that these differences in the timing and diurnal cycle consistency are linked to the prevalence of propagating convection in central Africa (Comer et al. 2007), which is sensitive to atmospheric conditions, especially wind shear. Further, the variability in propagating convection contributes to the observed variability in OLRMAX in land convective regions.

Ocean convective regions exhibit similar OLR diurnal cycle evolution; however, differences between the Indian Ocean and western Pacific LWCF diurnal evolution are evident in the basic structure. A small ALWCF is observed in the Indian Ocean LWCF evolution histogram but significant only at the 90% level (Fig. 7d). The LWCF diurnal cycle evolution histogram exhibits a higher probability of LWCF from 5 to 10 W m−2 near 1200 LST and LWCF values, greater than 60 W m−2, frequently occurring from 1600 to 2000 LST. The LWCF first diurnal cycle harmonic in the western Pacific (Fig. 7e) does not exhibit a statistical significance at the 90% level. The western Pacific LWCF diurnal cycle structure possesses a semidiurnal cycle harmonic pattern. Western Pacific LWCF values between 20 and 40 W m−2 occur frequently at 0800–1400 and 2200–0200 LST, whereas LWCF values greater than 60.0 W m−2 occur at 1600–2000 and 0200–0600 LST. These two regions demonstrate significant regional variability in the oceanic convective LWCF diurnal evolution structure.

4. Discussion: A comparison with previous results

Overall, the present results show consistency with previous investigations. The strongest OLR diurnal cycles occur over arid land regions, and significant diurnal cycles are found in land regions with frequent convection. Further, land–sea breeze circulations evident in 11-μm brightness temperatures (Yang and Slingo 2001) are also evident in broadband flux and LWCF. Ocean convective region AOLR generally ranges from 2 to 5 W m−2, consistent with Hartmann and Recker (1986) and Schmetz and Liu (1988). Further, MSc cloud regions show small but robust OLR and LWCF diurnal cycles (Minnis and Harrison 1984b; Schmetz and Liu 1988; Gruber and Chen 1988; Rozendaal et al. 1995).

Several differences with previous results are evident. First, the absolute value of AOLR tends to be larger in this study than in most previous investigations. Previous results indicate that AOLR ranges from 18 to 25 W m−2 across North Africa (Hartmann and Recker 1986; Gruber and Chen 1988; Smith and Rutan 2003). The present results indicate AOLR values between 24 and 29 W m−2 averaged over the North African region (10°–30°N, 0°–30°E); several 1° × 1° regions exceed 35 W m−2 over a 3-month season. These results agree with Schmetz and Liu (1988), reporting a monthly-mean diurnal range of 62 W m−2 and maximum values of 75 W m−2, which correspond to an AOLR of about 31 and 37.5 W m−2, respectively. Besides Schmetz and Liu (1988), AOLR from the current study is larger than previous studies by 10%–20%. D. R. Doelling (2012, personal communication) corroborate this finding. They evaluate the GEO-interp method over North Africa against GERB, indicating a 20% AOLR overestimate. They indicate that CERES overestimates OLRMAX due to the GEO-interp method, which causes AOLR to be too large. The results of AOLR in other land nonconvective regions (e.g., central Australia and the Atacama Desert) are also 10%–20% larger than detailed in previous work (Hartmann and Recker 1986; Gruber and Chen 1988; Smith and Rutan 2003), indicating that this problem in the GEO-interp method extends to other desert regions as well. This amplitude bias, however, appears only over arid land. D. R. Doelling (2012, personal communication) show amplitude differences less than 10% between CERES and GERB in the Namibian MSc region and central Africa.

Second, the OLR and LWCF first diurnal cycle harmonics are not statistically significant at the 95% confidence level in many 1° × 1° oceanic deep convective regions. Hartmann and Recker (1986) and Gruber and Chen (1988) indicate that OLR diurnal cycles in the western Pacific, Atlantic ITCZ, and SPCZ are statistically significant at the 95% level. Averaging over areas larger than 1° × 1°, AOLR becomes statistically significant in oceanic convective regions with a 2–3 W m−2 climatological value. Previous work in these regions describes a robust cloud diurnal cycle (Hartmann and Recker 1986; Hendon and Woodberry 1993; Janowiak et al. 1994). The Indian Ocean and Atlantic ITCZ LWCF first diurnal cycle harmonic becomes statistically significant after averaging over a larger area. The LWCF first diurnal harmonic in the western Pacific and SPCZ, however, explains less than 20% of the variance after including a larger area. Mentioned previously, the second LWCF diurnal cycle harmonic represents a statistically significant fit with 95% confidence in the western Pacific (Fig. 7e) and the SPCZ regions (not shown). The LWCF diurnal cycle behavior in the western Pacific and SPCZ regions resembles a combined land and oceanic convective diurnal cycle, because high clouds frequently occur around 1800 and 0400 LST. The results indicate that regional LWCF diurnal cycle variability among oceanic convective regions stems from the importance of the second diurnal harmonic. Sui et al. (1997) demonstrated that morning low clouds, late-afternoon scattered convection, and nocturnal organized convection characterize the cloud diurnal cycle in ocean convective regions. A strong nocturnal LWCF maximum and weaker late-afternoon LWCF magnitudes are inferred from this cloud diurnal structure. The observed regional variation in the LWCF diurnal cycle and the semidiurnal structure suggests regional differences in relative magnitude of afternoon scattered convection to nocturnal organized convection. This can potentially lead to significant differences in regional energy budgets through reflected radiation differences in the afternoon.

Third, Hartmann and Recker (1986) identify a POLR dependence on convective intensity also suggested by the results of Janowiak et al. (1994). They suggest that intense oceanic convective regions, identified by lower OLR values, exhibit a POLR early in the morning, whereas weaker convective regions, higher OLR values, exhibit a local-noon POLR. Figure 2 suggests this behavior in the SPCZ. A region near 20°S, 150°W possesses a POLR of 0800–1000 LST and moving outward from this location POLR increases to about 1200 LST. A weak positive statistically significant correlation between POLR and OLR is suggested in the SPCZ regional mean annual cycle (Fig. 11). The POLR–OLR relationship through the annual cycle, however, does not exhibit significant relationships in the western Pacific, Atlantic ITCZ, and Indian Ocean regions (Fig. 11). This seasonal difference between the SPCZ and other oceanic convective regions is attributed to a larger annual cycle in solar radiation due to its latitude range. All analyzed ocean convective regions exhibit statistically significant positive correlations (95% confidence level) between POLR and OLRclim, in agreement with Hartmann and Recker (1986). This behavior occurs because regions with stronger convection begin forming clouds earlier, indicated by larger LWCF values and a shift in POLR to the late morning. Therefore, a relationship between convective intensity and POLR likely exists in many ocean convective regions comparing 1° × 1° grid boxes; however, this relationship is obscured at the larger scale, about 20° × 20°.

Fig. 11.
Fig. 11.

Annual cycle of POLR (solid) and OLR (dotted) in the (top left) SPCZ (10.0°–20.0°S, 170°E–150°W), (top right) Indian Ocean (0.0°–15.0°S, 55.0°–95.0°E), (bottom left) western Pacific (10.0°N–10.0°S, 160.0°E–180.0°), and (bottom right) Atlantic ITCZ (0.0°–10.0°N, 20.0°–50.0°W). Diurnal cycle harmonic is computed after averaging over all 1° × 1° grid boxes in the region.

Citation: Journal of the Atmospheric Sciences 69, 12; 10.1175/JAS-D-12-088.1

5. Summary and conclusions

In summary, the climatological and seasonal OLR and LWCF tropical diurnal cycles are analyzed using CERES SYN Ed2rev1 data. These data provide detailed characterization of OLR diurnal cycle in the tropics, and a unique opportunity to analyze the tropical LWCF diurnal cycle. This analysis reveals similar tropical OLR diurnal cycle features as previous analysis. Over arid land regions, the OLR diurnal cycle amplitude is largest with a maximum value occurring between 1400 and 1600 LST. In land regions with frequent convection, OLR diurnal cycle amplitude is statistically significant and generally greater than 10 W m−2. The OLR diurnal cycle phase occurs near 1200 LST in these regions, which is slightly earlier than arid land regions due to the afternoon increase in LWCF. Robust OLR diurnal cycles are identified over ocean and possess weaker amplitudes than over land, typically less than 10 W m−2. The oceanic OLR diurnal cycle exhibits an afternoon maximum where marine stratocumulus clouds dominate and there is significant variability in ocean regions of climatological convection (e.g., western Pacific and SPCZ).

The LWCF diurnal cycle amplitude possesses significant spatial variability. Arid land regions tend not to exhibit a significant climatological LWCF diurnal cycle; however, individual 3-month seasons show statistical significance. Southern Hemispheric arid land regions (e.g., central Australia, and the Kalahari and Atacama Deserts) indicate a robust LWCF diurnal cycle during local summer that reduces the OLR diurnal cycle amplitude. The result is a stronger OLR diurnal cycle in transition seasons. Central African and South American convective regions possess significant LWCF diurnal cycle amplitudes generally greater than 10 W m−2 in all seasons. Maximum LWCF values are found to coincide with the convective precipitation maximum in the evening to early morning, 1800–0200 LST. Regional variability is evident between central South America and central African convective regions, attributed to propagating convection.

A significant LWCF diurnal cycle is also observed over ocean regions. The climatological LWCF diurnal cycle in the Peruvian and Namibian marine stratocumulus regions exhibits a 3–6 W m−2 amplitude, with values reaching 10 W m−2 during DJF. The observed diurnal cycle phase indicates maximum LWCF near sunrise, around 0600 LST, consistent across all marine stratocumulus regions. Oceanic convective regions possess a LWCF diurnal cycle amplitude similar to marine stratocumulus regions but with weaker statistical significance. Statistical significance of the first diurnal cycle harmonic is observed in the Indian Ocean and Atlantic ITCZ regions when averaging over many grid points. This is not the case for the western Pacific and SPCZ regions; the analysis revealed a statistically significant semidiurnal cycle in LWCF in these regions. The LWCF semidiurnal cycle is related to the diurnal phasing of different cloud types, in particular the relative importance of late-afternoon scattered convection and nocturnal organized convection. One implication is that regional differences in the LWCF diurnal cycle of oceanic convective regions can influence the regional TOA energy budget. The expectation based on previous work showing the inability of global models to simulate middle-level clouds is that the LWCF semidiurnal harmonic will likely not be represented in oceanic convective regions, causing biases in regional energy budgets.

Regional analysis using diurnal cycle evolution histograms indicates the relative importance of understanding physical processes driving the diurnal maximum and minimum OLR values. Evident in land convective and nonconvective regions, the maximum OLR value contributes more to the total OLR variance than the diurnal minimum OLR, by more than 50%. Maximum and minimum daily OLR variability differences are explained by differences in surface temperature variability, about 5–6 and 3–4 K at maximum and minimum OLR, respectively. This implies that the processes controlling diurnal maximum OLR and surface temperature (e.g., shallow-cloud evolution and dust loading) are more important to OLR variability than at OLR minimum. Over land, diurnal minimum OLR variance is reduced by a covariance between LWCF and OLRCLR: larger LWCF values occur with larger OLRCLR. The results indicate that this behavior results from the tendency of stronger deep convection and more high clouds to occur when surface temperatures are warmer. Diurnal maximum and minimum OLR variability over ocean convective and nonconvective regions contribute equally to overall OLR variances. However, significant seasonality is found in OLR variance around local noon in ocean convective regions (e.g., Indian Ocean and western Pacific) due to seasonal variation in diurnal cloud evolution. In closing, stronger OLR variations tend to occur in the afternoon. Thus, representing variability in the maximum OLR may be more important for accurately representing OLR variability in models than the minimum OLR.

Acknowledgments

I would like to thank Dr. Seiji Kato and the anonymous reviewers for their constructive and extremely helpful comments on this manuscript. The data used in this study are stored at the Atmospheric Science Data Center at NASA Langley Research Center.

REFERENCES

  • Allan, R. P., , A. Slingo, , S. F. Milton, , and M. E. Brooks, 2007: Evaluation of the Met Office global model using Geostationary Earth Radiation Budget (GERB) data. Quart. J. Roy. Meteor. Soc., 133, 19932010.

    • Search Google Scholar
    • Export Citation
  • Bergman, J. W., , and M. L. Salby, 1996: Diurnal variations of cloud cover and their relationship to climatological conditions. J. Climate, 9, 28022820.

    • Search Google Scholar
    • Export Citation
  • Bergman, J. W., , and M. L. Salby, 1997: The role of cloud diurnal variations in the time-mean energy budget. J. Climate, 10, 11141124.

    • Search Google Scholar
    • Export Citation
  • Bloom, S. A., and Coauthors, 2005: Documentation and validation of the Goddard Earth Observing System (GEOS) data assimilation system–Version 4. Technical Report Series on Global Modeling and Data Assimilation, Vol. 26, NASA Tech. Memo. NASA/TM-25-104606, 181 pp.

  • Chung, E.-S., , B.-J. Sohn, , and J. Schmetz, 2009: Diurnal variation of outgoing longwave radiation associated with high cloud and UTH changes from Meteosat-5 measurements. Meteor. Atmos. Phys., 105, 109119, doi:10.1007/s00703-009-0041-8.

    • Search Google Scholar
    • Export Citation
  • Comer, R. E., , A. Slingo, , and R. P. Allan, 2007: Observations of the diurnal cycle of outgoing longwave radiation from the Geostationary Earth Radiation Budget instrument. Geophys. Res. Lett., 34, L02823, doi:10.1029/2006GL028229.

    • Search Google Scholar
    • Export Citation
  • Del Genio, A. D., , and J. Wu, 2010: The role of entrainment in the diurnal cycle of continental convection. J. Climate, 23, 27222738.

  • Duynkerke, P. G., , and J. Teixeira, 2001: Comparison of the ECMWF reanalysis with FIRE I observations: Diurnal variation of marine stratocumulus. J. Climate, 14, 14661478.

    • Search Google Scholar
    • Export Citation
  • Futyan, J. M., , J. E. Russel, , and J. E. Harries, 2005. Determining cloud forcing by cloud type from geostationary satellite data. Geophys. Res. Lett., 32, L08807, doi:10.1029/2004GL022275.

    • Search Google Scholar
    • Export Citation
  • Garreaud, R. D., , and R. Munoz, 2004: The diurnal cycle in circulation and cloudiness over the subtropical southeast Pacific: A modeling study. J. Climate, 17, 16991710.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., , and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105, 11711188.

  • Gruber, A., , and T. S. Chen, 1988: Diurnal variation of outgoing longwave radiation. Int. J. Climatol., 8, 116.

  • Harries, J. E., and Coauthors, 2005: The Geostationary Earth Radiation Budget project. Bull. Amer. Meteor. Soc., 86, 945960.

  • Hartmann, D. L., , and E. E. Recker, 1986: Diurnal variation of outgoing longwave radiation in the tropics. J. Climate Appl. Meteor., 25, 800812.

    • Search Google Scholar
    • Export Citation
  • Hartmann, D. L., , K. J. Kowalewsky, , and M. L. Michelsen, 1991: Diurnal variations of outgoing longwave radiation and albedo from ERBE scanner data. J. Climate, 4, 598617.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., , and K. Woodberry, 1993: The diurnal cycle of tropical convection. J. Geophys. Res., 98 (D9), 16 62316 637.

  • Hong, G., , G. Heygster, , and C. A. M Rodriguez, 2006: Effect of cirrus clouds on the diurnal cycle of tropical deep convective clouds. J. Geophys. Res., 111, D06209, doi:10.1029/2005JD006208.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., , P. A. Arkin, , and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122, 22962311.

    • Search Google Scholar
    • Export Citation
  • Liebmann, B., , and A. Gruber, 1988: Annual variation of the diurnal cycle of outgoing longwave radiation. Mon. Wea. Rev., 116, 16591670.

    • Search Google Scholar
    • Export Citation
  • Lin, X., , D. A. Randall, , and L. D. Fowler, 2000: Diurnal variability of the hydrologic cycle and radiative fluxes: Comparisons between observations and a GCM. J. Climate, 13, 41594179.

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

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , and E. F. Harrison, 1984a: Diurnal variability of regional cloud and clear-sky radiative parameters derived from GOES data. Part I: Analysis method. J. Climate Appl. Meteor., 23, 9931011.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , and E. F. Harrison, 1984b: Diurnal variability of regional cloud and clear-sky radiative parameters derived from GOES data. Part II: November 1978 cloud distributions. J. Climate Appl. Meteor., 23, 10121031.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , and E. F. Harrison, 1984c: Diurnal variability of regional cloud and clear-sky radiative parameters derived from GOES data. Part III: November 1978 radiative parameters. J. Climate Appl. Meteor., 23, 10321051.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., , D. F. Young, , and E. F. Harrison, 1991: Examination of the relationship between outgoing infrared window and total longwave fluxes using satellite data. J. Climate, 4, 11141133.

    • Search Google Scholar
    • Export Citation
  • Minnis, P., and Coauthors, 2011: CERES edition-2 cloud property retrievals using TRMM VIRS and Terra and Aqua MODIS data—Part I: Algorithms. IEEE Trans. Geosci. Remote Sens., 49, 43744400.

    • Search Google Scholar
    • Export Citation
  • Morcrette, J.-J., 1991: Evaluation of model-generated cloudiness: Satellite-observed and model-generated diurnal variability of brightness temperature. Mon. Wea. Rev., 119, 12051224.

    • Search Google Scholar
    • Export Citation
  • Nesbitt, S. W., , and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16, 14561475.

    • Search Google Scholar
    • Export Citation
  • Nowicki, S. M. J., , and C. J. Merchant, 2004: Observations of diurnal and spatial variability of radiative forcing by equatorial deep convective clouds. J. Geophys. Res., 109, D11202, doi:10.1029/2003JD004176.

    • Search Google Scholar
    • Export Citation
  • Randall, D. A., , Harshvardhan, , and D. A. Dazlich, 1991: Diurnal variability of the hydrologic cycle in a general circulation model. J. Atmos. Sci., 48, 4061.

    • Search Google Scholar
    • Export Citation
  • Raschke, E., , and W. R. Bandeen, 1970: The radiation balance of the planet earth from radiation measurements of the satellite Nimbus II. J. Appl. Meteor., 9, 215238.

    • Search Google Scholar
    • Export Citation
  • Rozendaal, M. A., , C. B. Leovy, , and S. A. Klein, 1995: An observational study of diurnal variation of marine stratiform cloud. J. Climate, 8, 17951809.

    • Search Google Scholar
    • Export Citation
  • Schmetz, J., , and Q. Liu, 1988: Outgoing longwave radiation and its diurnal variation at regional scales derived from Meteosat. J. Geophys. Res., 93 (D9), 11 19211 204.

    • Search Google Scholar
    • Export Citation
  • Short, D. A., , and J. M. Wallace, 1980: Satellite-inferred morning-to-evening cloudiness changes. Mon. Wea. Rev., 108, 11601169.

  • Slingo, A., , K. I. Hodges, , and G. J. Robinson, 2004: Simulation of the diurnal cycle in a climate model and its evaluation using data from Meteosat 7. Quart. J. Roy. Meteor. Soc., 130, 14491467.

    • Search Google Scholar
    • Export Citation
  • Smith, G. L., , and D. A. Rutan, 2003: The diurnal cycle of outgoing longwave radiation from Earth Radiation Budget Experiment measurements. J. Atmos. Sci., 60, 15291542.

    • Search Google Scholar
    • Export Citation
  • Soden, B. J., 2000: The diurnal cycle of convection, clouds, and water vapor in the tropical upper troposphere. Geophys. Res. Lett., 27, 21732176.

    • Search Google Scholar
    • Export Citation
  • Sui, C.-H., , K.-M. Lau, , Y. N. Takayabu, , and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54, 639655.

    • Search Google Scholar
    • Export Citation
  • Tian, B., , B. J. Soden, , and X. Wu, 2004: Diurnal cycle of convection, clouds, and water vapor in the tropical upper troposphere: Satellites versus a general circulation model. J. Geophys. Res., 109, D10101, doi:10.1029/2003JD004117.

    • Search Google Scholar
    • Export Citation
  • Turton, J. D., , and S. Nicholls, 1987: A study of the diurnal variation of stratocumulus using a multiple mixed layer model. Quart. J. Roy. Meteor. Soc., 113, 9691009.

    • Search Google Scholar
    • Export Citation
  • Wu, M.-L. C., , S. D. Schubert, , M. J. Suarez, , and P. J. Pegion, 2006: Seasonality and meridional propagation of the MJO. J. Climate, 19, 19011921.

    • Search Google Scholar
    • Export Citation
  • Yang, G.-Y., , and J. Slingo, 2001: The diurnal cycle in the tropics. Mon. Wea. Rev., 129, 784801.

  • Yang, S., , K.-S. Kuo, , and E. A. Smith, 2008: Persistent nature of secondary diurnal modes of precipitation over oceanic and continental regimes. J. Climate, 21, 41154131.

    • Search Google Scholar
    • Export Citation
  • Young, D. F., , P. Minnis, , D. R. Doelling, , G. G. Gibson, , and T. Wong, 1998: Temporal interpolation methods for the Clouds and the Earth’s Radiant Energy System (CERES) experiment. J. Appl. Meteor., 37, 572590.

    • Search Google Scholar
    • Export Citation
Save