1. Introduction
One of the more important problems in climate research is understanding the role of clouds in modifying the earth’s radiation balance. Considering that radiative imbalance is the ultimate driving force of the earth’s general circulation, clouds can have a direct impact on the global circulation when they produce transient effects on the established radiative gradients. Even though convectively generated latent heating within clouds is often considered to be the most important diabatic influence of clouds on the heating field, a number of modeling studies have pointed out that cloud-related radiative heating–cooling is as important as latent heat release in terms of a forcing mechanism for large-scale circulations and deep cumulus convection (e.g., Gray and Jacobson 1977; Slingo and Slingo 1988; Randall et al. 1989; Sherwood et al. 1994). These effects are generally referred to as a result of “cloud-radiative forcing” (CRF) that can be defined as the impact of cloud variations on the radiation balance at the top of the atmosphere (TOA) and the surface. Although CRF cannot be measured directly, TOA CRF has been determined from satellite estimates including measurements from the Earth Radiation Budget Experiment (ERBE; Ramanathan et al. 1989) and from the Nimbus-7 ERB experiment (Ardanuy et al. 1991; Sohn and Smith 1992).
The importance of cloud-generated radiative heating–cooling stems from the fact that the major heat sources and sinks arise from the large-scale coupling between the meridional and zonal components of the mass circulation (Zillman 1972). Under certain conditions, cloud-induced variations in radiative heating are able to provide a basic forcing for atmospheric motions and thus can be considered to be an essential element maintaining the global circulation. Cloud influence on the global circulation can be achieved through modification of the available potential energy from cloud-induced differential heating and the energy conversion from potential to kinetic energy, as described by Stuhlmann and Smith (1988). Consistent with this notion, Sohn and Smith (1992) demonstrated that one important global aspect of a cloud’s overall role is its modification of north–south gradients in net radiation and thus modification of the mean meridional circulation.
Thermally induced circulations are not just a function of the horizontal distribution of heating, but are also dependent on the vertical distribution of heating (e.g., Hartmann et al. 1984; Lau and Peng 1987). Thus, in order to understand how changes in cloudiness can alter the general circulation and thus the global climate, it is necessary to estimate three-dimensional distributions of cloud-generated diabatic heating. Obviously such information is not directly obtainable from satellite measurements. Therefore, in order to determine both horizontal and vertical distributions of cloud-induced radiative heating, we use a radiative transfer model in conjunction with inputs of global satellite observations of cloudiness along with temperature–moisture data from the European Centre for Medium-Range Weather Forecasts (ECMWF). Even though this approach brings in a more accurate flux calculation compared with the climate model approach, uncertainties in the calculated fluxes are dominated by the quality of the input data and limited by the quality of available validation datasets (Rossow and Lacis 1990). However, the recent advent of more accurate global observations of clouds and radiation fluxes at the top of the atmosphere such as the International Satellite Cloud Climatology Project (ISCCP) and ERBE data allows us to calculate more reliable fluxes on a three-dimensional basis. Using this updated information it is now possible to separate the effects of clouds on TOA and surface radiation balance from other factors and to determine three-dimensional distribution of radiative heating caused by individual cloud types.
The main objective of this study is to understand the role of three-dimensional cloud-induced heat sources and sinks in forcing planetary-scale circulations that are coupled with major regions of diabatic heating and cooling. In general, clouds have a large influence on surface heating through the interactions with shortwave radiation. However, how much shortwave atmospheric absorption is induced by clouds has been a subject of much debate (Stephens and Tsay 1990; Chou et al. 1995; Li and Moreau 1996; among many others). Even though some observational studies report much higher cloud absorption referred to as the cloud absorption anomaly, contemporary radiation models used in general circulation models and in the retrieval of surface and atmospheric solar radiation budgets suggest virtually negligible impact of clouds (Randall et al. 1989; Cess et al. 1995). Current radiative transfer models tend to produce much smaller cloud absorption of solar radiation within the atmosphere, compared with longwave absorption (e.g., Pinker and Laszlo 1992; Chou et al. 1998). Thus, in this analysis, we focus on the longwave diabatic heating component based on the assumption that the atmospheric radiative cooling is dominated by infrared processes and, in general, the major effect of clouds on solar radiation is to bring about surface cooling. By examining the prominent features of the cloud-driven heating distribution, we seek to explain the potential influence of clouds on underlying atmospheric circulations over the globe.
2. Data sources
Calculation of radiative cooling profiles throughout the atmospheric column requires physical parameters within the atmosphere and on the surface, including atmospheric temperature-mixing ratio profiles for radiatively active vertically heterogeneous atmospheric gases (H2O and O3), surface temperature, emissivity, and the optical properties of clouds. In this study, atmospheric temperature and water vapor mixing ratio are obtained from ECMWF global analyses. The ECMWF analyses provide temperature and relative humidity at seven designated pressure levels (1000, 850, 700, 500, 300, 200, and 100 mb) with a horizontal resolution of 2.5° × 2.5° latitude–longitude. This study uses only the seasonal means for the summer of 1988 (June, July, and August 1988; hereafter referred to as JJA88) and the winter of 1989 (December 1988, January and February 1989; hereafter referred to as DJF89), obtained from twice-daily analyses (0000 and 1200 UTC).
Generally, the accuracy of radiative transfer calculations using a theoretical model mostly relies upon the quality of the input parameters. To reduce the inherent uncertainties in input parameters related to clear-sky flux calculations, the ECMWF profiles of humidity and temperature have been adjusted following the methodology given in Sohn (1994). In this process the ECMWF total precipitable water values are bias adjusted to those derived from Special Sensor Microwave/Imager (SSM/I) measurements, and the ECMWF temperature fields over oceans are modified to produce brightness temperatures equal to those from the National Oceanic and Atmospheric Administration Microwave Sounding Unit (MSU) channel 2 (53.74 GHz) measurements. For the stratosphere, temperature distributions at 50 and 10 mb are specified as 5° zonal means for each season, based on the Geophysical Fluid Dynamics Laboratory temperature climatology over a 15-yr period (Oort 1983).
Vertical distributions of ozone concentrations are obtained from the Klenk et al. (1983) parameterization, based on results from the Nimbus-4 Backscattered Ultraviolet (BUV) experiment for the upper layer and on balloon measurements for the lower stratosphere and troposphere. Since this parameterization expresses the ozone profile as a function of time of year and latitude, zonal means for each season are specified. The vertical distribution of CO2 concentration is kept constant at 350 ppm throughout all atmospheric layers. Surface temperature is obtained from ISCCP products (Rossow and Schiffer 1991).
For the radiative transfer calculations within a cloudy atmosphere, the vertical profiles of cloud optical parameters are needed. In this study, we use six cloud types (cirrus, cirrostratus, deep convection, altocumulus, nimbostratus, and low clouds) as shown in Fig. 1, a classification similar to that done by Ockert-Bell and Hartmann (1992). Monthly mean cloud amount and cloud-top pressure for the six cloud types are extracted from ISCCP archives.
For a given cloud top, cloud-base pressure is assigned according to the given cloud type. Here, bases of all stratiform and cirrus clouds (cirrus, cirrostratus, altocumulus, and low clouds) are assigned 50 mb below their respective tops, which is representative of midtropospheric clouds according to climatological analyses (e.g., Sasamori et al. 1972). This climatological value, however, is not representative of nimbostratus and deep convective–type clouds, whose thickness can be extended over several kilometers. For these deep convective-type clouds (deep convection and nimbostratus), we assign the base height between the lifting condensation level (LCL) and the freezing level. This is to account for the fact that cloud bases tend to move upward to levels above the LCL during the evolution of a cloud convective system from cumulus stage to dissipation stage (Doswell 1985).
3. Calculation of radiative heating rates
Many techniques exist for solving the radiative transfer equation in clear and cloudy atmospheres. In this study, we use a nonscattering, medium-resolution, random band model described in Smith and Shi (1992), which divides the thermal infrared spectrum between 2 and 250 μm (5000–40 cm−1) into 27 spectral intervals (see also Mehta and Smith 1997).
a. Design of infrared radiative transfer model
For the radiative transfer calculations in the cloudy (or overcast) atmosphere, each cloud is assumed to exist in only one layer. The plane-parallel assumption is made for parameterizing cloud influence on the upward and downward flux profiles. This is an appropriate assumption from a space perspective since it is not straightforward to distinguish underlying clouds or retrieve cloud geometrical properties from current global satellite datasets. Even though plane-parallel theory is not complete for conditions of broken clouds, it is applicable when the three-dimensional cloud geometry is not available. This is the main reason behind its use in all current operational general circulation models.
b. Validation of radiative transfer model
Fluxes calculated from the longwave transfer model used in this study have been compared with the line-by-line model and other participating model results in the Intercomparison of Radiation Codes for Climate Models (ICRCCM) study (Ellingson et al. 1991). By comparing net flux at the tropopause and net flux difference of the troposphere for specified atmospheric conditions, the medium band model used in this study has been evaluated as one comparable to line-by-line models for clear sky, with an agreement within a few watts per square meter.
For the cloudy atmosphere case in the ICRCCM study, the ranges of fluxes for six different cloud types from different models are provided, without line-by-line benchmarks. Therefore, we can only intercompare our model results relative to the flux ranges found among the ICRCCM infrared models. For the specified atmospheric and cloud conditions, the fluxes calculated by the radiative transfer model used in this study lie within the central range bars of the ICRCCM models. Although this agreement does not guarantee the absolute accuracy of the model, it benchmarks the model to a number of standard radiative transfer codes used for climate research.
c. Cloud-induced atmospheric heating rate
4. Discussion of results
a. Longwave radiative transfer calculations
The TOA longwave CRF over the globe is obtained for JJA88 and DJF89 with the radiative transfer model using the ISCCP cloud climatology in conjunction with ECMWF analyses and is compared with the ERBE-derived CRF. The distributions of the ERBE and modeled CRFs for JJA88 are given in Figs. 2a and 2b, respectively. Shaded areas in Fig. 2 indicate values greater than 40 W m−2 of longwave CRF. It is noted that the radiative transfer calculations reproduce the general patterns noted in the ERBE observations, in particular the locations of maxima over the Indian Ocean and the western Pacific, and lower values over the eastern Pacific subtropical highs. Noticeable differences in magnitude between ERBE and the modeled values are found over convectively active regions such as the tropical rain forest areas, the tropical eastern Pacific ITCZ, and the western Pacific warm pool region. Differences up to 15 W m−2 are found in some tropical regions.
For the DJF89 case, the CRFs from the ERBE observation and radiative transfer model are presented in Figs. 3a and 3b, respectively. Similar to the summer case, the two sets of patterns shown in Fig. 3 are in good agreement, although the model values are smaller than ERBE over most of the tropical latitudes, in particular, convectively active regions. Also substantial differences are found along midlatitude storm tracks.
Discrepancies over the convectively active regions are more evident in the zonal mean distributions of cloud forcing given in Fig. 4. For JJA88, the largest differences (up to 10 W m−2) are found in the latitude belts around 10° and 60°N. These are associated with the mean positions of the ITCZ and Northern Hemisphere midlatitude storm tracks, respectively. Similar patterns are found in DJF89, with the locations of maximum difference situated over the Tropics and midlatitude zones of 40°N and 60°S, corresponding to the convectively active regions during the winter season.
These discrepancies may arise from modeling deficiencies, in particular, from how clouds are parameterized. However, they can also arise due to the different definitions of clear-sky radiation flux between the ERBE and ISCCP datasets, based on differing clear-sky scene identification methods. For example, a more restrictive definition of clear-sky is used in the ERBE data analysis, generating relatively larger clear-sky fluxes and thus larger CRF. This issue has been addressed by Sohn and Robertson (1993), noting that differences in CRF from various methodologies and datasets are as large as 20 W m−2 in the zonal mean sense, with the ERBE-derived CRF being largest. In addition, Weare (1995) has suggested that ERBE longwave clear-sky fluxes (and thus the associated CRFs) are overestimated by 2–10 W m−2. Although it is difficult to diagnose the exact reasons of the discrepancies, the sensitivity tests discussed in the appendix may provide some insight.
Together with the TOA radiation flux, the surface flux determines atmospheric radiative heating and cloud radiative forcing; therefore, precise estimates of surface flux are a prerequisite for credible estimates of radiatively induced diabatic heating. In general, it is difficult to directly determine the surface downwelling radiance from satellite measurements because of the ill-posed retrieval problem at infrared wavelength in a cloudy atmosphere. In recent years, however, some progress has been made. Here, we intercompare the model-generated CRF at the surface with that from the surface radiation budget calculations of Darnell et al. (1992), in which longwave surface flux estimates are obtained from the method of Gupta (1989). Figure 5 shows the distributions of zonal mean CRF for JJA88 at the surface. Agreement is close except at midlatitudes around 50° latitude of both hemispheres and 10°N where flux differences of several watts per square meter are found. However, since flux uncertainties of the Gupta algorithm are ∼15 W m−2 (Darnell et al. 1992), the discrepancies shown in Fig. 5 are not surprising.
b. Total cloud-induced atmospheric heating
To help interpret the cloud-induced heating distributions and their interseasonal variation in this section we first briefly describe regional mean total cloud cover, cirrus, and low clouds. For the Northern Hemisphere summer during 1988 (JJA88) the total cloud amount reaches 70%–80% over the western Pacific, the Asian monsoon region, and the ITCZ. Cirrus clouds are most abundant over the western Pacific and over the Asian monsoon region, with mean fraction ranging from 20% to 30%. In contrast to the abundant cirrus clouds in warm oceans, the eastern Pacific cold oceans off the west coasts of California and Peru show dominant low cloud amount reaching 70%. Low cloud amounts over the western Pacific and Indian Oceans are less than 10%. Between the equator and 30°S, the Southern Hemisphere subtropics and the descending branch of the Hadley cell are indicated by cloud-free regions. In the Northern Hemisphere, relative cloud-free subtropics are located in the central Pacific.
During the winter season of 1989 (DJF89), substantial cloud cover exceeding 80% is related to the winter monsoon over the Indonesian archipelago, and over the South Pacific Convergence Zone where abundant cirrus clouds are also noted. Interestingly nearly the same amount of cirrostratus as that of cirrus clouds is found at the same locations. Over the eastern Pacific cold oceans the low cloud amount is smaller by up to 40% to 50%, and the maximum locations are found in somewhat higher latitudes, compared with the summer.
As expected, the bulk longwave heating (negative cooling) rates of the clear-sky atmosphere for JJA88 (given in Fig. 6a) show atmospheric cooling everywhere. The maximum cooling in the clear atmosphere takes place in the equatorial belt with magnitudes exceeding 2.1° day−1. This is mainly due to the large water vapor paths over the low-latitude warm oceanic areas.
Radiative heating (negative cooling) rates for the summer of 1988 are presented in Fig. 6b. The overall radiative cooling effect of clouds appears to be largely dependent upon geographical location and cloud type. For example, the magnitudes of the maximum cooling rates over the western Pacific and Indian Oceans are smaller than those observed for the clear sky, indicating that clouds over these regions generate radiative heating. By contrast, the eastern Pacific off the west coast of the United States and South America loses more infrared radiation, compared to the cooling distributions for the clear-sky atmosphere. This is because the radiative effects of clouds increase cooling above the cloud tops and decrease cooling below the bases. Since cloud-top altitudes are generally higher in the western Pacific and Indian Oceans due to abundant high-level clouds associated with deep convection, cloud-induced warming over these regions should be dominant. By contrast, over the eastern Pacific where marine stratus cloud decks are predominant and thus cloud-top altitudes are low, the cloud-top temperature is not much different from the surface temperature (Albrecht et al. 1995). Therefore, cloud forcing at the TOA due to low clouds appears to be minor. However, downward cloud forcing at the surface should be large because of the near unit emissivity of clouds and high cloud-base temperature. The cloud forcing at the surface solely induced by low clouds over the eastern Pacific is over 40 W m−2, indicating a strong energy supply (relative to clear skies) to the ocean surface by clouds. Since under these conditions there is strong net divergence of longwave radiation, the overall effect of clouds is to cool radiatively.
In order to examine potential cloud-radiative effects on atmospheric heating, purely cloud-induced heating rates are determined by taking the differences between heating rates for the total cloudy-sky and the clear-sky atmosphere, as expressed in Eq. (7). The distribution of vertically averaged cloud-induced radiative heating for JJA88 is presented in Fig. 6c, using a contour interval of 0.2° day−1. The salient features of this diagram include cloud-induced relative cooling over the Northern Hemisphere continents, polar latitudes, much of the subtropical latitudes, and the Southern Hemisphere midlatitudes. By contrast, prominent cloud-induced relative heating is found over the western Pacific and Indian Oceans, tropical rainforest areas, and along the ITCZ, resulting in a broad heating regime between the equator and 20°N, except in the eastern Pacific off the west coast of the North American continent. Maximum relative heating over the western Pacific exceeds 0.4° day−1, while relative cooling over the colder eastern Pacific exceeds 0.4° day−1. Such cloud-induced heating and cooling distributions are similar in pattern to the total diabatic heating distributions from ECMWF assimilated data (Christy 1991; Schaack and Johnson 1994), suggesting that clouds serve to intensify the north–south and east–west heating gradients found in the total diabatic heating field. It is important to recognize that the magnitudes of cloud-induced longwave heating are about 20% of the total heating suggested by Schaack and Johnson (1994).
For DJF89, the vertically averaged distributions of heating rate of clear sky, cloudy sky, and their difference (cloud-induced heating rate) are given in Figs. 7a–c, respectively. The tropical region heating rates, where radiative cooling exceeds 2.1° day−1 (as seen in Fig. 7a), are much smaller compared to JJA88, indicating a significant interseasonal change in the diabatic heating pattern. The cloud-induced radiative heating distribution (Fig. 7c) is also much different from that found in JJA88; for example, the tropical heating is centered south of the equator with the area of maxima (>0.2° day−1) spanning from South Africa to east of the date line, resulting in a dominant north–south heating gradient across subtropical latitudes in the Southern Hemisphere. Also, there is a considerable gradient of radiative heating from the equatorial central Pacific to the eastern Pacific off the Peruvian coast in the Southern Hemisphere.
In addition to the meridional heating gradients between the Tropics and the subtropics, it is noted that substantial east–west gradients of relative heating also exist between the western Pacific and the eastern Pacific off the American continent, with magnitudes comparable to meridional gradients across the Indian Ocean. Since the structure of tropical atmospheric circulations is uniquely related to the response of the atmosphere to direct thermal forcing (Webster 1994; Schaack and Johnson 1994), we can deduce the influences of cloud-induced radiative heating on the atmospheric circulation. These should be an ascending motion over the warm pool regions of the Indian and Pacific Oceans, in conjunction with sinking motion over the Tibetan Plateau and south Indian Ocean, as well as over the cold oceans in the eastern Pacific during the summer. Therefore, cloud-radiative forcing can enhance the north–south Hadley circulation as well as the east–west zonal circulation through longwave radiative processes. These features are very similar to those found in the role of latent heating distributions (Webster 1994), indicating that latent and cloud-induced heating gradients are of the same sign. This is because the initial cloud formation is accompanied by latent heat release and the resultant cloud mass produces general warming through the longwave mechanism. However, latent heat release is directly tied to upward vertical motion or precipitation whose area is a small portion of total cloudy area (about 15% of the total cloudiness at any one time; Liu et al. 1995) and latent heat release does not last longer than the cloud-existing time. Since radiative heating requires only cloud mass, the entire western Pacific and Indian Oceans may be subject to a slower but broader ascent associated with cloud forcing, compared to the effect of latent heat release. A similar interpretation can be applied for the winter, that is, ascending motion around the 10°–20°S regions over the Indian Ocean and western Pacific, and sinking motion over the surrounding area.
c. Atmospheric heating by cloud types
As suggested in Eqs. (4) and (5), the flux calculations with the ISCCP dataset make it possible to diagnose the atmospheric heating influences of different cloud types on the total heating distribution. Although the diagnostic study is performed for a total of six cloud types, only the heating contributions by cirrus and low clouds are presented. This is because they are the two cloud types inducing the largest flux changes at the TOA and surface, respectively. Figures 8a and 8b illustrate the global distributions of atmospheric heating due to cirrus and low cloud for the summer of 1988, respectively. As expected, cirrus clouds contribute most to longwave-induced heating over the Indian Ocean and the western Pacific because of the high cloud tops and their abundance over these regions. In contrast, atmospheric cooling is mostly produced by low clouds, especially in the marine stratus regions over the eastern Pacific and high-latitude oceans of both hemispheres. Thus, the large-scale distribution of cirrus and low clouds is the main factor in establishing differential heating gradients, as noted in Fig. 6c.
To examine the possible link of the vertical heating structure to the mean Hadley circulation, latitude–height cross sections of cloud-induced heating averaged over the longitude belt from 60°E to 180° have been constructed. The vertical cross sections are shown for JJA88 (Fig. 9a) and DJF89 (Fig. 9b). Pronounced heating contrasts are found between the tropical deep convective areas and the Southern Hemispheric subtropical latitudes, with dipolelike patterns of radiative heating in the upper troposphere and cooling about 1 K day−1 in the lower troposphere. These patterns strongly suggest that cloud influences on the Hadley-type circulations are much stronger in the Southern Hemisphere. It is partly due to the high topography in the Northern Hemisphere since strong cooling in association with low-level clouds is confined within the cold oceanic area. The feature of the dipolelike pattern (Ackerman et al. 1988) is a signature of the optically thick convective clouds, which warm near their bases and cool their tops. By cooling aloft, the clouds (anvil and cirrus clouds) over the Tropics tend to destabilize the atmosphere, exerting on the atmospheric environment, and thus leading to a positive cloud feedback mechanism favoring deep convection. This destabilizing effect due to clouds is consistent with findings that tropical clouds can strengthen the precipitation maxima at low latitudes through longwave radiative processes (Webster and Stephens 1980; Slingo and Slingo 1988; Randall et al. 1989). By contrast, there is some indication that this destabilizing mechanism on precipitation has little impact on the precipitation on the storm scale (Tao et al. 1996).
Figure 10 shows the vertical cross sections of cloud-induced heating averaged over the latitudinal belt from 30°N to 10°S for JJA88 (Fig. 10a), and from 10°N to 30°S for DJF89 (Fig. 10b). Although the average was taken over most of the tropical latitudes and thus a much smoother heating contrast between the western Pacific and the eastern Pacific is shown, the cloud-induced Walker-type circulation can be expected from thermally driven upward motion over the western Pacific due to the heating and downward motion over the eastern Pacific due to the low-level cooling.
In order to examine in detail the suggested link of the vertical heating structure to the large-scale east–west circulation, we select two areas that compose the ascending and descending branches of the Walker circulation. Figure 11 shows the vertical distributions of longwave radiative heating for JJA88, averaged over the western Pacific (10°S–20°N, 125°–155°E) for total cloud as well as other six types of cloud (Fig. 11a), and over the eastern Pacific (0°–30°N, 80°–110°W) for total cloud and low cloud (Fig. 11b). The profile of total heating indicates that most of the troposphere, except at cloud top just below the tropopause level, is heated by total cloud. This feature is indicative of the destabilizing effects that clouds can exert on the atmospheric environment, same as shown in latitude–height cross sections of cloud-induced heating. Total cloud cooling near 250 mb is achieved by the imbalance between strong cooling by deep convective cloud and cirrus warming. On the other hand, most of the radiative heating in the lower troposphere is produced by clouds with high-level cloud tops (cirrus, cirrostratus, and deep convective clouds). Thus, this analysis suggests that deep convective clouds are the main element to shape cooling in the upper troposphere and heating in the lower troposphere shown in the total over the western Pacific.
By contrast, cooling profiles due to total and low clouds confined within the lower boundary layer over the eastern Pacific (Fig. 11b) exhibit nearly identical contributions, suggesting that low clouds provide the predominant cloud-forcing effect over that region. Overall, strong cooling with a maximum of up to 3° day−1 at 800 mb and near-zero heating below 800 mb would induce thermodynamic destabilization effects in the surface boundary layer without some type of adiabatic or diabatic compensation.
5. Summary and conclusions
In this paper, we have focused on the cloud-induced longwave heating within the atmosphere and its potential impact on large-scale atmospheric circulation in the Tropics. In doing so, we calculated cloud-induced radiative heating rates for the summer of 1988 and the winter of 1989 using a radiative transfer model. The major input data to the model are humidity and temperature profiles from ECMWF analyses and cloud amount and cloud-top pressure of six different types from the ISCCP dataset. The major difference of this study from others is to use the most updated cloud and radiation budget information, which allows us to examine the impact of various cloud types on atmospheric radiative heating on three-dimensional basis.
The prominent features found in the distribution of cloud-induced longwave heating for JJA88 consist of maximum radiative heating over the western Pacific and the Indian Oceans, and maximum radiative cooling over the eastern Pacific subtropical regions. These distributions induce significant differential heating gradients in both north–south and east–west directions. The cloud-induced heating field for DJF89 exhibits a considerably different pattern from that found in the summer; that is, the maximum heating is located south of the equator with a relatively weaker magnitude and a strong north–south heating gradient around 30°S over the Indian and Pacific Oceans.
The importance of these differential heating fields stems from the fact that the tropical circulation is characterized by direct response of the atmosphere to heat sources and sinks, and from the fact that the global circulation is achieved through the generation of available potential energy due to differential heating and its conversion to kinetic energy. It has been shown that clouds are capable of generating and modifying the differential heating through longwave radiative heating processes. Thus, clouds can provide basic forcing to the large-scale atmospheric circulation.
During the summer, the strongest heating gradients take place along the north–south direction from tropical convective regions such as the Indian monsoon region, the eastern Pacific ITCZ, and the continental rain forest areas. Considerable heating gradients are also found longitudinally between the western Pacific and the eastern Pacific off the Californian and Peruvian coasts, and between the western Pacific and the region from the Indian Ocean to central Africa.
A finding of particular importance in this study is the relationship of heating gradients to the tropical circulation. Since clouds build up differential heating over the Tropics and thus amplify kinetic energy generation by north–south and east–west overturning (i.e., mass responses to the heat sources and sinks), cloud forcing is a direct factor in maintaining or reinforcing both the Hadley and Walker circulations. This notion is consistent with the findings from studies based on general circulation models that thermodynamically direct tropical circulation systems such as the Hadley circulation are strengthened through the longwave radiation processes of clouds (Slingo and Slingo 1988; Randall et al. 1989; Sherwood et al. 1994). Therefore, as considered here, the role of clouds on the general circulation is through direct thermodynamic forcing. If we include indirect effects such as increased convergence of water vapor and latent heat release produced by a modified circulation, the impact of CRF could be even more significant.
Radiative transfer calculations using a set of cloud types generate individual diabatic heating profiles associated to each cloud type. In tropical convective regions more than 50% of the total heating is contributed by upper-level heating due to cirrus clouds. Thus, the amount of cirrus is critical in determining the thermodynamic and radiative properties of the tropical atmosphere as found in the vertical distribution of heating given in Fig. 9. By contrast, in marine stratus regions and at high latitudes, low clouds are of the greatest importance for cooling.
The foremost result of this study is that longwave CRF, which is generally a localized heat source, can induce a widespread impact on the large-scale tropical circulation. Since the radiative heating process consists of both longwave and shortwave effects, we would have to examine the additional influence of shortwave CRF before drawing final conclusions related to cloud-generated radiative heating and its influence on the atmosphere’s general circulation. It would be expected that such analysis would raise questions concerning uncertainties in cloud parameterization, radiative transfer modeling, and input data. However, as indicated in the sensitivity test presented in the appendix, the conclusions drawn from this study may remain unaffected.
Acknowledgments
The author wishes to thank Professor Eric A. Smith of The Florida State University for valuable comments and suggestions. This study was supported by the Basic Research Institute Program, within the Ministry of Education of Korea. A portion of the research was performed while the author was visiting the NASA/Marshall Space Flight Center under support from the Universities Space Research Association.
REFERENCES
Ackerman, S. A., and S. K. Cox, 1987: Radiative energy budget estimates for the 1979 southwest summer monsoon. J. Atmos. Sci.,44, 3052–3078.
Ackerman, T. P., K.-N. Liou, F. P. J. Valero, and L. Pfister, 1988: Heating rates in tropical anvils. J. Atmos. Sci.,45, 1606–1623.
Albrecht, B. A., M. P. Jensen, and W. J. Syrett, 1995: Marine boundary layer structure and fractional cloudiness. J. Geophys. Res.,100, 14 209–14 222.
Ardanuy, P. E., L. L. Stowe, A. Gruber, and M. Weiss, 1991: Shortwave, longwave and net cloud-radiative forcing as determined from Nimbus 7 observations. J. Geophys. Res.,96, 18 537–18 549.
Cess, R. D., and Coauthors, 1995: Absorption of solar radiation by clouds: Observations versus models. Science,267, 496–499.
Christy, J. R., 1991: Diabatic heating rate estimates from European Centre for Medium-Range Weather Forecasts analyses. J. Geophys. Res.,96, 5123–5135.
Chou, M.-D., A. Arking, and W. L. Ridgway, 1995: The effect of clouds on atmospheric heating. Geophys. Res. Lett.,22, 1885–1888.
——, W. Zhao, and S.-H. Chou, 1998: Radiation budgets and cloud radiative forcing in the Pacific warm pool during TOGA COARE. J. Geophys. Res.,103, 16 967–16 978.
Cox, S. K., and K. T. Griffith, 1979: Estimates of radiative divergence during Phase III of the GARP Atlantic Tropical Experiment: Part I. Methodology. J. Atmos. Sci.,36, 576–585.
Darnell W. L., W. F. Staylor, S. K. Gupta, N. A. Ritchey, and A. C. Wilber, 1992: Seasonal variation of surface radiation budget derived from International Satellite Cloud Climatology Project C1 data. J. Geophys. Res.,97, 15 741–15 760.
Doswell, C. A., III, 1985: The operational meteorology of convective weather. Volume II: Storm scale analysis. NOAA Tech. Memo. ERL ESG-15, 240 pp. [Available from U.S. Government Printing Office, Washington, DC 20402.].
Ellingson, R., J. Ellis, and S. Fels, 1991: The intercomparison of radiation codes used in climate models: Longwave results. J. Geophys. Res.,96, 8929–8953.
Gray, W. M., and R. W. Jacobson, 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev.,105, 1171–1188.
Griffith, K., S. K. Cox, and R. G. Knollenberg, 1980: Infrared properties of tropical cirrus clouds inferred from aircraft measurements. J. Atmos. Sci.,37, 1077–1087.
Gupta, S. K., 1989: A parameterization for longwave surface radiation from sun-synchronous satellite data. J. Climate,2, 305–320.
Hartmann, D. L., M. H. Hendon, and R. A. House, 1984: Some implication of the mesoscale circulations in tropical cloud clusters for large-scale dynamics and climate. J. Atmos. Sci.,41, 113–121.
Heymsfield, A. T., and C. M. R. Platt, 1984: A parameterization of the particle size spectrum of ice clouds in terms of the ambient temperature and the ice water content. J. Atmos. Sci.,41, 846–855.
Klenk, K. F., P. K. Bhartia, E. Hilsenrath, and A. J. Fleig, 1983: Standard ozone profile from balloon and satellite data sets. J. Climate Appl. Meteor.,22, 2012–2022.
Lau, K.-M., and L. Peng, 1987: Origin of low-frequency (intraseasonal) oscillations in the tropical atmosphere. Part I: Basic theory. J. Atmos. Sci.,44, 950–972.
Li, Z., and L. Moreau, 1996: Alteration of atmospheric solar absorption by clouds: Simulation and observation. J. Appl. Meteor.,35, 653–670.
Liou, K. N., 1986: Influence of cirrus clouds on weather and climate processes: A global perspective. Mon. Wea. Rev.,114, 1167–1199.
——, and G. D. Wittman, 1979: Parameterization of the radiative properties of clouds. J. Atmos. Sci.,36, 1261–1273.
Liu, G., J. A. Curry, and R.-S. Sheu, 1995: Classification of clouds over the western Pacific Ocean using combined infrared and microwave satellite data. J. Geophys. Res.,100, 13 811–13 826.
Manabe, S., and R. T. Wetherald, 1967: Thermal equilibrium of the atmosphere with a given distribution of the relative humidity. J. Atmos. Sci.,24, 241–259.
Mehta, A. V., and E. A. Smith, 1997: Variability of radiative cooling during the Asian summer monsoon and its influence on intraseasonal waves. J. Atmos. Sci.,54, 941–966.
Ockert-Bell, M. E., and D. L. Hartmann, 1992: The effect of cloud type on earth’s energy balance: Results for selected regions. J. Climate,5, 1157–1171.
Oort, A. H., 1983: Global atmospheric circulation statistics, 1958–1973. NOAA Prof. Paper 14, NOAA/Geophysical Fluid Dynamics Laboratory, Princeton, NJ, 180 pp.
Paltridge, G. W., and C. M. R. Platt, 1981: Aircraft measurements of solar and infrared radiation and the microphysics of cirrus cloud. Quart. J. Roy. Meteor. Soc.,107, 367–380.
Pinker, R., and I. Laszlo, 1992: Modeling surface solar irradiance for satellite applications on a global scale. J. Appl. Meteor.,31, 194–211.
Platt, C. M. R., 1976: Infrared absorption and liquid water content in stratocumulus clouds. Quart. J. Roy. Meteor. Soc.,102, 553–561.
Ramanathan, V., R. D. Cess, E. F. Harrison, P. Minnis, B. R. Barkstrom, E. Ahmad, and D. Hartmann, 1989: Cloud-radiative forcing and climate: Results from the Earth Radiation Budget Experiment. Science,243, 57–63.
Randall, D. A., Harshvardhan, D. D. Dazlich, and T. G. Corsetti, 1989: Interactions among radiation, convection, and large-scale dynamics in a general circulation model. J. Atmos. Sci.,46, 1943–1970.
Rossow, W. B., and A. A. Lacis, 1990: Global, seasonal cloud variations from satellite radiance measurements. Part II: Cloud properties and radiative effects. J. Climate,3, 1204–1253.
——, and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer Meteor. Soc.,72, 2–20.
Sasamori, T. J., J. London, and D. V. Hoyt, 1972: Radiation budget of the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 35, Amer. Meteor. Soc., 9–24.
Schaack, T. K., and D. R. Johnson, 1994: January and July global distributions of atmospheric heating for 1986, 1987, and 1988. J. Climate,7, 1270–1285.
Sherwood, S. C., V. Ramanathan, T. P. Barnett, M. K. Tyree, and E. Roeckner, 1994: Response of an atmospheric general circulation model to radiative forcing of tropical clouds. J. Geophys. Res.,99, 20 829–20 845.
Slingo, A., and J. M. Slingo, 1988: The response of a general circulation model to cloud longwave radiative forcing. I: Introduction and initial experiments. Quart. J. Roy. Meteor. Soc.,114, 1027–1062.
Smith, E. A., and L. Shi, 1992: Surface forcing of the infrared cooling profile over the Tibetan Plateau. Part I: Influence of relative longwave radiative forcing at high latitude. J. Atmos. Sci.,49, 805–822.
Sohn, B. J., 1994: Temperature–moisture biases in ECMWF analyses based on clear sky longwave simulations constrained by SSMI and MSU measurements and comparisons to ERBE estimates. J. Climate,7, 1707–1718.
——, and E. A. Smith, 1992: The significance of cloud–radiative forcing to the general circulation on climate time scales—A satellite interpretation. J. Atmos. Sci.,49, 845–860.
——, and F. R. Robertson, 1993: Intercomparison of observed cloud radiative forcing: A zonal and global perspective. Bull. Amer. Meteor. Soc.,74, 997–1006.
Stephens, G. L., 1978: Radiation profiles in extended water clouds. II: Parameterization scheme. J. Atmos. Sci.,35, 2123–2132.
——, and S.-C. Tsay, 1990: On the cloud absorption anomaly. Quart. J. Roy. Meteor. Soc.,116, 671–704.
Stuhlmann, R., and G. L. Smith, 1988: A study of cloud-generated radiative heating and its generation of available potential energy. Part II: Results for a climatological zonal mean January. J. Atmos. Sci.,45, 3928–3943.
Tao, W.-K., S. Lang, J. Simpson, C.-H. Sui, B. Ferrier, and M.-D. Chou, 1996: Mechanisms of cloud–radiation interaction in the Tropics and midlatitudes. J. Atmos. Sci.,53, 2624–2651.
Weare, B. C., 1995: Factors controlling ERBE longwave clear sky and cloud forcing fluxes. J. Climate,8, 1889–1899.
Webster, P. J., 1994: The role of hydrological processes in ocean–atmosphere interactions. Rev. Geophys.,32, 427–476.
——, and G. L. Stephens, 1980: Tropical upper-tropospheric extended clouds: Inferences from Winter MONEX. J. Atmos. Sci.,37, 1521–1541.
Zhang, Y.-C., W. B. Rossow, and A. A. Lacis, 1995: Calculation of surface and top of atmosphere radiative fluxes from physical quantities based on ISCCP data sets. 1. Method and sensitivity to input data uncertainties. J. Geophys. Res.,100, 1149–1165.
Zillman, J. W., 1972: Isentropically time-averaged mass circulations in the Southern Hemisphere. Ph.D. thesis, University of Wisconsin—Madison, 205 pp.
APPENDIX
Sensitivity Test
To assess the credibility of the flux calculations in conjunction with how input parameters are specified we have carried out a number of sensitivity tests concerning how the results are affected by uncertainties in the cloud input parameters. Table A1 summarizes changes in the diabatic heating rate (QR) and its gradient (ΔQR), the global mean of TOA longwave radiation (TOA LW), and the cloud radiative forcing at both TOA (TOA CRF) and surface (SFC CRF), in response to variations in cloud-top and cloud-base height, cloud liquid water distribution, and cloud amount. Tests have been made for the summer of 1988. For the calculation of heating gradients (ΔQR), differences are taken between (0°, 90°E) and (30°S, 90°E), and between (10°N, 140°E) and (30°N, 130°W). They are assumed to be representative of the equatorial and the south Indian Ocean, and the western and eastern Pacific, respectively.
An increase of cloud-top height by 25 mb (p25), holding all other parameters constant, produces increased diabatic heating over the equatorial Indian and western Pacific (warm) oceans, but decreased cooling over the subtropical (cold) oceans due to elevated cloud bases. Under these changes, the perturbed gradients are not much different from the present results. For the cloud-base determination, we assume that the cloud-base pressure of optically thin clouds (cirrus, cirrostratus, altocumulus, and low clouds) is 50 mb higher than the respective cloud-top pressure, leading to cloud thickness of 50 mb. For the sensitivity test, 100-mb cloud layer thickness is assigned to those same four cloud types (d100). The test indicates that the thicker cloud layer (or lower cloud base) also induces increased heating over the warm tropical oceans and increased cooling over the cold subtropical oceans. Although increased heating gradients are produced, as well as increased CRF at both TOA and surface, the overall global atmospheric heating due to clouds is weak. An increase of cloud liquid water content by 10% (clw110) leads to an increase in diabatic heating over the western Pacific and Indian Ocean areas, but reduced cooling over the cold oceans. Overall, the sensitivity of in the cloud-induced heating and its gradient to altered input parameters is about 10% of the present results.
In order to examine the influence of ISCCP satellite-retrieved cloud amounts on the radiative flux calculations, a cloud overlapping test has been conducted. This is because satellite retrieval in an overlapped multiple cloud layer environment only generates a cloud fraction for the uppermost layer (i.e., clouds in the lower layers are partly or completely blocked by those above them). Since the approach that is used in this study assumes no cloud layers where they are not directly observed, we have conducted a sensitivity test related to this assumption by employing random overlapping in proportion to the fraction of each cloud layer (e.g., Manabe and Wetherald 1967). Results suggest that overlapping causes less warming over the warm oceans and more cooling over the cold oceans, but only slight changes in the overall radiative heating gradients. It is also noted that TOA values of CRF are not very sensitive to cloud overlapping, while, as expected, surface CRF is relatively sensitive. However, although the changes in heating gradients between the selected regions are measurable, they are small. Thus, for all four test cases, our interpretation of a cloud-induced relative radiative heating mechanism at work remains unchanged since tropical direct circulations are linked to differential heating rather than the absolute heating magnitudes. On the other hand, much extensive discussion on the sensitivities of the calculated radiative flux to the uncertainties in the ISCCP input parameter is found in Zhang et al. (1995).
Six cloud types used in this study. Classification is based on the cloud-top pressure and cloud optical thickness determined from ISCCP (Rossow and Schiffer 1991).
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Longwave cloud-radiative forcing at the top of the atmosphere (a) observed by the ERBE and (b) calculated by a radiative transfer model for the summer of 1988 (Jun, Jul, and Aug 1988). Contour interval is 20 W m−2. Shading indicates values greater than 40 W m−2.
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Same as Fig. 2 except for the winter of 1989 (Dec 1988, Jan and Feb 1989).
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Zonal mean distribution of the TOA cloud-radiative forcing for (a) the summer of 1988 and (b) the winter of 1989.
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Zonal mean distribution of surface cloud-radiative forcing for the summer of 1988.
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Global distributions of vertically averaged longwave radiative heating for (a) clear-sky atmosphere, (b) cloudy atmosphere, and (c) difference for the summer of 1988. Values are multiplied by 10 with units given in °C day−1.
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Same as in Fig. 6 except for the winter of 1989.
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Global distributions of vertically averaged longwave radiative heating induced by (a) cirrus and (b) low clouds for the summer of 1988. Values are multiplied by 10 with units given in °C day−1.
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Vertical sections of cloud-induced heating (0.1°C day−1) averaged over (a) the longitudinal belt from 60°E to 180° for (a) the summer of 1988, and (b) the winter of 1989. Shaded area represents negative heating (cooling).
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Vertical sections of cloud-induced heating (0.1°C day−1) averaged over (a) the latitudinal belt from 30°N to 10°S for the summer of 1988, and (b) from 10°N to 30°S for the winter of 1989. Shaded area represents negative heating (cooling).
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Fig. 11: Vertical distributions of the longwave radiative heating (°C day−1) induced by total cloud and six types of cloud for the western Pacific (10°S–20°N, 125°–155°E), and by total and low clouds over the eastern Pacific (0°–30°N, 80°–110°W).
Citation: Journal of the Atmospheric Sciences 56, 15; 10.1175/1520-0469(1999)056<2657:CIIRHA>2.0.CO;2
Table A1. Changes in cloud-induced diabatic heating rate, heating gradient, global means of TOA longwave radiation, TOA, and surface cloud-radiative forcing due to changes in cloud properties. The definition of the symbols is given in appendix.
