Enhanced Low-Level Stratus in the FSU Coupled Ocean–Atmosphere Model

David R. Bachiochi Center for Ocean–Atmosphere Prediction Studies, The Florida State University. Tallahassee, Florida

Search for other papers by David R. Bachiochi in
Current site
Google Scholar
PubMed
Close
and
T. N. Krishnamurti Department of Meteorology, The Florida State University, Tallahassee, Florida

Search for other papers by T. N. Krishnamurti in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

An empirical planetary boundary layer cloud parameterization has been developed for The Florida State University coupled ocean–atmosphere model to improve low-level clouds in the model. The scheme diagnoses the clouds by combining the PBL depth, ground wetness, and the relative humidity. Winter and summer simulations between 1987 and 1989 suggest an improvement in the low cloud representation compared to the International Satellite Cloud Climatology Project analysis. When implemented in the model, this parameterization results in positive impacts on shortwave fluxes and low-level circulation, particularly along the west coasts of the North and South American continents. Enhanced mechanical forcing at the ocean surface improves the SST representation in the eastern Pacific Ocean basin. Warm versus cold phase ENSO variability of the east Pacific SSTs are also improved during the seasonal simulations. Furthermore, the near-coast diurnal fluctuation of the low cloud is comparable to observations.

Corresponding author address: Dr. David Bachiochi, COAPS, The Florida State University, 2035 E. Paul Dirac Road, R. M. Johnson Building, Suite 200, Tallahassee, FL 32310.

Email: bachi@met.fsu.edu

Abstract

An empirical planetary boundary layer cloud parameterization has been developed for The Florida State University coupled ocean–atmosphere model to improve low-level clouds in the model. The scheme diagnoses the clouds by combining the PBL depth, ground wetness, and the relative humidity. Winter and summer simulations between 1987 and 1989 suggest an improvement in the low cloud representation compared to the International Satellite Cloud Climatology Project analysis. When implemented in the model, this parameterization results in positive impacts on shortwave fluxes and low-level circulation, particularly along the west coasts of the North and South American continents. Enhanced mechanical forcing at the ocean surface improves the SST representation in the eastern Pacific Ocean basin. Warm versus cold phase ENSO variability of the east Pacific SSTs are also improved during the seasonal simulations. Furthermore, the near-coast diurnal fluctuation of the low cloud is comparable to observations.

Corresponding author address: Dr. David Bachiochi, COAPS, The Florida State University, 2035 E. Paul Dirac Road, R. M. Johnson Building, Suite 200, Tallahassee, FL 32310.

Email: bachi@met.fsu.edu

1. Introduction

One problem that has followed climate modelers has been a poor representation of low-level planetary boundary layer (PBL) clouds (Palmer and Anderson 1994; Delecluse et al. 1998). Observations tell us that they largely reside over the oceans off the west coasts of the major continents and are most prevalent in the summer hemisphere. These thin clouds alter the radiation characteristics of the lower atmosphere and hence alter the surface fluxes of heat and momentum. Pertinent near-cloud balances are between radiational cooling and subsidence warming as well as convective moistening and subsidence drying. Observed seasonal variability of east Pacific low-level stratus includes a summer maximum in the Northern Hemisphere and an October maximum in the Southern Hemisphere (Klein and Hartmann 1993), where the Southern Hemisphere maximum of low cloud corresponds to the minimum sea surface temperature. Klein and Hartmann (1993) indicate that in most low stratus cloud regions of the world, the above-cloud stability is crucial to the existence of PBL clouds. They also suggest that low-level temperature is a primary factor in the Southern Hemisphere east Pacific. The sea surface temperatures off the coast of Peru are more variable than off the coast of regions such as California for both interannual and seasonal timescales. Proper representation of low clouds in coupled model simulations may improve the predicted sea surface temperatures (SST) and their subsequent feedback on the atmosphere.

Philander et al. (1996) and Ma et al. (1996) show the strong impact of the stratus-type clouds in coupled global models. Philander et al. (1996) used a simple stability-based cloud that impacts the shortwave radiation into the ocean. Results suggested a strong cooling effect on the ocean (about 2°C) between 30°S and 30°N. Heat fluxes into the ocean were reduced by over 100 W m−2 over the eastern ocean basins where the largest stratus amounts typically reside. Ma et al. (1996) show that the SSTs off the coast of Peru are cooled more realistically with prescribed PBL stratus clouds over the eastern Pacific Ocean. The resulting enhanced circulation isolated the intertropical convergence zone (ITCZ) north of the equator. Li and Philander (1996) also show enhanced variability of Pacific Ocean SSTs in a coupled model that includes tropical marine stratus. Frey et al. (1997) found that improving the low-level stratus helped to reduce a Southern Hemisphere eastern Pacific Ocean warm bias during a 10-yr integration of the ECHO-2 coupled ocean–atmosphere model. A key to the relationship between low clouds and surface temperatures in these studies was the impact of the clouds on low-level winds and coastal upwelling. Changes in the low-level horizontal thermal gradient between regions with and without the low clouds may partially explain the enhancement of the low-level wind. The reduced low-level solar fluxes below the PBL clouds enhance the thermal gradients from the cloud regions toward the ITCZ, resulting in an enhanced trade wind circulation over the eastern Pacific Ocean.

The development of PBL clouds is related to the existence of a moist, buoyant layer below the PBL top, capped by a more stable layer of air (Lilly 1968). A problem with diagnosing such a situation in most climate models is the lack of vertical resolution to resolve such thin layers of moisture and thermal profiles without going to a higher-resolution model (Bechtold et al. 1996). Limitations also exist due to low horizontal resolutions. Most coupled models are currently run with 20–40 levels in the vertical. The models with a smaller number of vertical levels are lacking the resolution to properly represent the vertical gradients observed in the atmospheric boundary layer. Examples of physical parameters used in simple PBL cloud parameterizations are relative humidity, mixing-line slope, and lapse rate among others (Slingo 1987; Betts and Boers 1990; Albrecht 1989; Sundqvist et al. 1989).

The Florida State University (FSU) global spectral model uses a parameterization of clouds based on relative humidity (RH) thresholds (Slingo 1980). Slingo (1987) developed a scheme that uses the vertical temperature lapse rate along with RH applied to the European Centre for Medium-Range Weather Forecasts (ECMWF) global model and later applied to the National Center for Atmospheric Research (NCAR) Community Climate Model (Slingo and Slingo 1991). Results from these documented model changes suggest substantial improvement to global radiation, humidity, and temperature. Klein and Hartmann (1993) showed a strong correlation between low-level stratus and stability based on the difference between the surface temperature and the above-inversion temperature (700 hPa). Philander et al. (1996) used this temperature measure for stability in their coupled model stratus study. These simple schemes attempt to capture the basic horizontal PBL cloud variability. Much of the current advances in prognostic and statistical cloud schemes are spawned from the works of Sommeria and Deardorff (1977) among others. These studies derived mean cloud fractions for grid volumes based on assumed statistical representations (Gaussian for instance) of conservative thermodynamic variables about their volume mean. Empirical relationships for the clouds were derived in terms of saturated volumes and mean relative humidity. These ideas have been used in conjunction with prognostic cloud water schemes in more advanced global models and higher-resolution mesoscale models (Smith 1990; Tiedke 1993; Sundqvist et al. 1989).

Motivation for the development of the PBL cloud scheme presented here was derived from the low-level cloud deficiency in the FSU coupled ocean–atmosphere model (LaRow and Krishnamurti 1998, hereafter referred to as LK98). Investigation into potential variables for a cloud parameterization led to the observation that there is a similar spatial structure between PBL heights and low-level clouds. PBL height is a natural parameter for a simple PBL cloud scheme since these clouds exist just below the top of the PBL and the PBL height diagnostic scheme used for this work depends on low-level stability, which is a key parameter in cloud formation. Certain care has been taken to generate a simple, computationally feasible scheme for low-level clouds using the height of the PBL top.

The remainder of the work discusses the model, PBL cloud parameterization, and results. Section 2 presents the coupled ocean–atmosphere model used for the work. The PBL height parameterization and PBL stratus cloud parameterization are presented in section 3. Results, including seasonal cloud impacts on the ocean and atmosphere, are presented in section 4, and a brief discussion of interannual impacts is presented in section 5. A concluding discussion follows in section 6.

2. Model

The model used for this work is the FSU coupled ocean–atmosphere model (LK98). The atmospheric component of the model consists of the FSU global spectral model (Krishnamurti et al. 1991). It is coupled synchronously to an ocean model that is similar to the model used by Latif (1987).

The atmospheric model was run with a horizontal, spectral resolution T42, using 14 sigma layers in the vertical (approximately between 30 and 1000 hPa). The model is very similar to the model used by Krishnamurti et al. (1991) with recent changes discussed in LK98, as well as changes discussed in section 3. The current work requires a brief discussion of some of the model physics pertaining to the radiation and lower atmosphere. The model contains a wideband radiation code used to calculate shortwave solar, near-infrared solar, and thermal longwave radiation separately (Chang 1979). Multiple Rayleigh scattering is applied to the shortwave radiation and an emissivity method is used for longwave fluxes. The near-infrared radiation accounts for water vapor absorption, while the solar zenith angle includes seasonal and diurnal variations. The low, middle, and high clouds are based on threshold relative humidity following Slingo (1980). Surface fluxes of heat, moisture, and momentum are estimated using Monin–Obukhov similarity theory (Businger et al. 1971). A local, first-order closure scheme is used for the vertical mixing of heat, moisture, and momentum (Louis 1979). The turbulent mixing coefficient is a function of mixing length, local wind shear, and local stability. The vertical profile of momentum, heat, and moisture for stable and unstable boundary layers are linear and functions of PBL height. Cumulus convection is parameterized by a modified Kuo scheme (Kuo 1965, 1974; Krishnamurti et al. 1983).

The ocean model is a global primitive equation model with realistic bottom topography. Boussinesq and hydrostatic approximations are used. It is run with a horizontal, staggered Arakawa and Lamb (1977) E grid at a resolution of 5° longitude and variable in the meridional direction (0.5° between 10°N/S latitude, increasing to 5.0° out to 70°N/S). There are 17 irregularly spaced levels in the ocean model; the first 10 layers are within 300 m of the surface. The model is dynamically active between 55°N and 55°S and outside this region the model is relaxed to the climatology of Levitus (1982).1 The mixing of heat, momentum, and salinity in the model uses a Richardson number–dependent mixing scheme based on Pacanowski and Philander (1981) and a mixed layer parameterization following Latif et al. (1994).

The atmospheric model is integrated for six 20-min time steps before the ocean model is integrated for one time step of 2 h. The ocean uses the fluxes of momentum, solar radiation, freshwater, and net heat from the atmosphere, whereas the atmosphere uses the SST field generated by the ocean model. Bilinear interpolation is used to transfer information between the ocean and atmosphere. Furthermore, no flux correction or anomaly coupling is applied. Therefore, there is no artificial forcing applied at the ocean–atmosphere interface to contain model climate drift. The Dewey and Heim [(1982), see footnote 1] National Environmental Satellite, Data and Information Service monthly mean snow cover fields are used to define the southerly extent of the snow line in the Northern Hemisphere for the time periods of this study. In the Southern Hemisphere, Antarctica is assumed fully covered with snow. Snow-covered land is given a maximum albedo of 0.6, but elsewhere the albedo is held constant through the season of simulation (ECMWF analysis dataset).

3. Low cloud scheme

a. Planetary boundary layer height

Older versions of the FSU global spectral model contain parameterizations for the vertical distribution of fluxes of moisture, temperature, and momentum that do not take into account the variation of the planetary boundary layer height. Those versions either fix the height unconditionally or use two values, for a stable and unstable PBL. This feature of prior versions of the FSU global spectral model is not up to standard when compared to other climate models that use diagnostic, prognostic, or specific sigma levels to define the PBL top. Differences in surface characteristics throughout the globe contribute to variations in the stability properties of the PBL and therefore in its height. These variations need to be accounted for in the PBL to appropriately represent the lower atmosphere.

Holtslag and Boville (1993) developed a diffusion scheme using the PBL height of Troen and Mahrt (1986). The same iterative procedure for PBL heights is applied to the FSU model vertical diffusion. Troen and Mahrt (1986) define the PBL height as
i1520-0493-128-9-3083-e1
The critical, bulk Richardson number, Ricr, is set to 0.5. This value departs from the theoretical value of 0.25 because of the atmospheric model’s coarse resolution. The horizontal wind components at hPBL are u(hPBL) and υ(hPBL), θυ(hPBL) is the virtual potential temperature at the PBL top, and θs is the near-surface potential temperature. The ratio g/θs is a buoyancy parameter with g the gravitational acceleration. The solution of θs follows Troen and Mahrt (1986) by defining it as a function of the surface layer virtual temperature (10-m value) and the virtual heat flux at the surface. The temperature at 10 m is calculated following the procedure of Geleyn (1988). The boundary layer height is determined by calculating the bulk Richardson number between the level of θs and subsequent levels in the vertical until a value greater than Ricr is found. The final value for the PBL height is determined by linear interpolation of the hPBL at the levels above and below where the bulk Richardson number would be Ricr. A minimum value for hPBL is defined following Koracin and Berkowicz (1988).

The diagnosed June and December mean PBL height from a coupled simulation without the PBL stratus is shown in Fig. 1. The PBL stratus clouds were not used during these 1-month simulations to show the preexistence of low PBL heights in regions where PBL stratus clouds are observed (each 1-month simulation starts from a state initialized as described in section 4). The PBL heights are monthly means of daily averages for 1987 and 1988. The plot illustrates the parameterization’s ability to represent the PBL height magnitude as well as its seasonal variability. The boundary layer heights compare well with those of Holtslag and Boville (1993). Overland boundary layer heights in the FSU simulations are not as high as those obtained by Holtslag and Boville (1993). Diurnal variations (not shown) are representative of the global surface heating variability. Large seasonal variations are noted over extratropical oceans and desert regions. The lowest PBL heights are located over the eastern oceans, poles, and over winter hemisphere landmasses.

b. Stratus cloud parameterization

A spatial relationship between the PBL depth and the low-level stratus cloud is observed between diagnosed PBL heights (Fig. 1) and the International Satellite Cloud Climatology Project (ISCCP) D2 low clouds (Fig. 2). The ISCCP D2 dataset is a revision of the ISCCP C2 dataset (Rossow and Schiffer 1991; Rossow et al. 1989; Rossow and Garder 1993). The D2 and C2 are monthly means derived from 3-hourly higher horizontal resolution data (D1 and C1). Revisions in the D2 set include satellite radiance calibrations; improved cloud detection, primarily at high latitudes; improved radiation model; and increased number of variables in the dataset. The largest amount of ISCCP D2 low cloud (Fig. 2) is in the eastern ocean basins, particularly in the month of June compared to December. Clouds are most abundant in the Southern Hemisphere compared to the Northern Hemisphere, with minimal cloud north or south of 60° in either month. There is also an indication of more clouds along the equator in the eastern Pacific. Warren et al. (1988; see also Fig. 1 in Norris and Leovy 1994) suggest a greater amount of low cloud poleward of 45° latitude compared to the ISCCP D2 data. The Warren et al. (1988) cloud climatology is based on physical observations and includes stratus, stratocumulus, and sky-obscuring fog for low clouds. These data are discussed here to suggest that there may be more clouds over the northern and southern reaches of the major ocean basins.

We observe a relationship between the low clouds and low PBL heights over the oceans and to a lesser extent over the land. The relationship between the stratus clouds and the PBL height makes sense since the depth of the PBL is associated with the stability near the surface and aloft. The stability plays an important role in the development and maintenance of stratus at the top of the PBL. The Klein and Hartmann (1993) correlations between low-level clouds and stability are recalled. Philander et al. (1996) also show a strong correlation between low-level stability (850-mb temperature − SST) and ISCCP C2 low clouds in the eastern Atlantic and Pacific Ocean basins and used this relationship for their simple stratus parameterization. Looking at Eq. (1), it is noted that the PBL height as parameterized here is inversely related to the low-level stratus since the parameterized PBL height is inversely proportional to the lapse rate over the PBL (following the two aforementioned papers). Therefore, a cloud fraction based on PBL height [Eq. (1)] is controlled by the PBL lapse rate, surface temperature (near-surface buoyancy), and the square of the magnitude of the horizontal wind at the PBL top. Prior research and currently utilized low-cloud parameterizations require a certain amount of moisture available for clouds to exist. This is also true for some of the more complex cloud schemes mentioned above, where a probability distribution is used. Therefore, a parameterization defined by the PBL height should also include moisture constraints.

Based on the observed spatial correspondence between the PBL depth and low clouds, we introduce an empirical parameterization for PBL stratus cloud fraction. The basic structure of the PBL stratus cloud parameterization is defined as
i1520-0493-128-9-3083-e2
The highest PBL depth, hmax allowing nonzero clouds was selected as 900 m, while hmin was set to 150 m. The damping coefficient, CF, is selected to be 0.75. A value of 1.0 better represents the ISCCP C2 data over the eastern ocean basins, but a more representative value to match the ISCCP D2 data was used here. Besides the sinusoidal representation (selected for its smooth properties at the endpoints), the primary shape controlling factors are hmax and hmin. A number of 3-month summer cases were run with observed weekly SSTs (Reynolds and Smith 1994) in which the values of hmax and hmin were varied and best values were selected visually by the horizontal shape of the resulting clouds compared with the ISCCP clouds. The minimum value for hPBL was varied between 150 and 300 m. The larger values gave broader regions of large cloud amount. Similarly, hmax was varied between 900 and 1200 m for values of hmin of 150 and 300 m. The larger values produced cloud amounts off the west coasts of the Americas that were spatially too broad compared to the ISCCP C2 and D2 analysis. PBL heights for regions off the western continental coasts with high stratocumulus incidence are roughly between 500 and 1200 m, which suggests that our value for hmax is a reasonable selection. Much more rigorous selection techniques exist, but were not utilized for this initial version of the cloud scheme. We set the cloud fraction to zero for RHPBL < RHcrit to assure that there is an ample supply of moisture at the near-cloud level. The critical relative humidity, RHcrit, is selected as 80%. Slingo (1987) and others use similar values for low-level stratus critical RH. Here, RHPBL is the greater of the relative humidity at either the level defining the PBL top or the model level just below the PBL top.
To help restrict the marine stratus over land and ice, the parameterization requires sufficient near-surface moisture. The model ground wetness, gw, is parameterized by
σ2
for σ (surface albedo) < 0.25 and otherwise set to zero. A simple dependence on this ground wetness was defined for the PBL cloud parameterization as follows:
i1520-0493-128-9-3083-e4
The value gwmax is selected as the maximum value that the ground wetness can attain (over the ocean). The minimum ground wetness, gwmin below which stratus are not allowed to form was selected based on an assumption that it is unlikely that PBL clouds would form over a desertlike area. This value was selected to be 0.1. Figure 3 is a plot of the PBL cloud given by Eq. (4) for an appropriate range of PBL height and ground wetness. Linear and exponential functions were tested for Eqs. (2) and (4), but failed to give spatially desirable results compared to the ISCCP C2 data.
The stratus clouds are incorporated into the radiation scheme by defining the new low cloud fraction as
i1520-0493-128-9-3083-e5
where Cold is the existing low cloud fraction. The stratus clouds are defined to be one sigma layer thick and the layer is defined as the layer containing the PBL top. Also, a cloud is not allowed at the lowest sigma layer. The cloud fraction defined by Eq. (5) is used in place of the old low cloud. The dependence of the PBL height on the surface fluxes of heat and the near-surface vertical temperature gradients and the coupling of the clouds with the radiation make this scheme reasonably well integrated into the model. The parameterization could be easily applied to other models that parameterize PBL heights and require low cloud fractions. It could also be applied as a separate PBL cloud layer instead of combining it with the low cloud scheme as done here. The main limitation to application of the scheme is related to the need to select the shape parameters, hmax and hmin. A variational technique could be used to determine the shape parameters by minimizing a function that depends on a cloud analysis dataset such as ISCCP.

The diurnal variability of the low-level clouds may strongly influence the radiational, thermal, and circulation properties of the lower coastal atmosphere. For this particular scheme, the diurnal behavior is dictated by the PBL heights. The diurnal behavior of the PBL height parameterization is consistent with diurnal surface heating patterns of the landmasses. Diurnal variability is smaller over the ocean due to the near-constant surface temperatures on a diurnal timescale.

The PBL stratus clouds impact the heating of the atmospheric column by absorbing shortwave radiation while absorbing and reemitting longwave radiation from the surface. The resulting changes in the temperature profile destabilize the upper PBL in the cloud layer, decoupling it from the lower boundary layer (Wang and Albrecht 1994). PBL stratus tend to increase at night in response to PBL convection and outgoing longwave radiation, which cools the boundary layer (Rogers et al. 1995). An example of the model PBL stratus diurnal variability is shown in Fig. 4 for Los Angeles during June 1987. Cloud amounts for Los Angeles are determined by linear interpolation of the model gridpoint values in both the latitudinal and longitudinal directions to the appropriate coordinates. Clearly the model low cloud has a strong diurnal component. Besides the first day of integration, the model has near 0% cloud amount at around 0600 UTC (2100 local time) and then the cloud amount increases to around 50% around 1800 UTC (1100 local time). Also embedded in the diurnal variation is an interweekly variability associated with the passage of synoptic-scale features. These results are consistent with those found for longer model simulations. Figure 5 shows the ISCCP D1 low cloud amount for the first five days of July 1987 for the Los Angeles region. This period adequately represents the typical summer diurnal variability of the ISCCP low clouds. The mismatch in time between Figs. 4 and 5 dissuades any direct temporal comparison between the model results and the analysis. The ISCCP and model data have a similar variability and time of minimum and maximum cloud amount. These results suggest that the use of boundary layer heights to define PBL clouds capture the diurnal variability of the clouds.

4. Seasonal results

The 1987 and 1988 Northern Hemisphere summer and 1987/88 and 1988/89 Northern Hemisphere winter seasons are used to investigate the cloud scheme’s impact on the coupled model. This selection not only allows for study of the intraannual variability, but also the interannual variability of El Niño and La Niña years. To obtain the initial state of the coupled system, the ocean model was first spun up for an extended period and then the coupled system was initialized using a Newtonian relaxation approach. Between January 1980 and May 1986, FSU wind stresses (Stricherz et al. 1992) and SSTs (Reynolds and Smith 1994) were used as forcing, and prior to that a 10-yr ocean spinup was performed using climatology only forcing (LK98). The coupled initialization was conducted continuously for 6 months prior to each initial forecast time. This differs from LK98, as that study used a 12-month assimilation. Differences between 6- and 12-month assimilation are minimal. The coupled assimilation uses Newtonian relaxation of daily 1200 UTC ECMWF analysis with weekly SST analysis (Reynolds and Smith 1994) following LK98. The relaxation coefficients used are 1 × 10−5 s−1 for vorticity, 5 × 10−6 s−1 for divergence and temperature, 1 × 10−6 s−1 for dewpoint depression, and 5 × 10−6 s−1 for SST. These values are close to the values used by LR98. The dewpoint depression coefficient is about 20% of the LR98 value to allow the moisture more freedom during initialization. The SST relaxation coefficient is an order of magnitude smaller than the LR98 value since it was found that the lower value allowed a smoother transition into the simulation phase. Each set of coefficients is further linearly reduced to 70% of its original value over the 6-month assimilation. Each 6-month period is initialized consecutively from the previous 6-month assimilation. For each season, the model is run with the old low cloud parameterization (hereafter referred to as the control) and the new low cloud parameterization (hereafter referred to as the stratus run). Each 3-month simulation is allowed to run freely as a coupled system. The only external data constraints are those described in section 2. Furthermore, there are no climatology or model bias corrections done during postsimulation analysis.

a. Global low clouds

The enhancement in model clouds is presented in Fig. 6, which shows the low cloud amount for the stratus run (panels a and b), control (panels c and d), and ISSCP D2 (panels e and f) for the summer and winter seasons. Seasonal means are calculated over the months of June, July, and August (JJA) for the summers of 1987 and 1988 and over the months of December, January, and February (DJF) for the winters of 1987/88 and 1988/89. Considerable improvements are noted over the entire globe. In general, the cloud amount is increased by 5%–20%. The new cloud parameterization better represents the zonal and meridional distribution of low cloud at all latitudes. Midlatitude cloud cover, particularly for cooler ocean temperature regions, has been increased significantly, yet poleward of 75°N/S the cloud amount is not changed much from the control. The largest and most significant change is for the west coast low-level clouds. Percent increases as high as 500% with the new cloud scheme are evident off the North American, South American, and African coasts. The enhanced low-level clouds are not as large in magnitude as the ISCCP clouds around 30°S, but still improved compared to the control. The larger amounts in the stratus run and the control in the region poleward of 60°N in DJF are not representative of the ISCCP D2 data. A large portion of this cloud disparity between the model and the ISCCP data in the polar region is due to the old cloud parameterization and perhaps other model deficiencies.

b. East Pacific

The importance of the eastern Pacific in interannual variability of weather features around the globe and the existence of extensive, persistent low clouds in the east Pacific region are well known. The remainder of the discussion is focused on that region. Prior studies indicate the importance of SSTs as well as low-level temperatures in controlling the possible existence of PBL clouds (Philander et al. 1996). The importance of meridional winds along coastal regions of the eastern ocean basins in upwelling subsurface cold water is also well known. Ma et al. (1996) illustrated the impact of coastal low clouds on offshore surface stresses in response to the reduced solar fluxes into the lower atmosphere and ocean. Increased low clouds reduce the low-level solar fluxes, initiating cooling near and at the surface. The cooler low-level temperatures in the continental offshore regions enhance the temperature gradients toward the equator, therefore enhancing the trades and hence coastal upwelling. The changes in the low-level temperatures and winds have a positive feedback on the low-level stratus.

1) Summer

A large impact due to the enhanced stratus during the summer can be seen in the spatial representation of surface shortwave radiation compared to the ISCCP C2 analysis (Fig. 7). Reductions in the shortwave fluxes for the stratus run compared to control are as high as 50 W m−2, particularly within 20° longitude of the eastern ocean boundaries. The lower fluxes over the land in the model are likely due to the overrepresentation of the low clouds over land, particularly over the United States. With and without the new stratus the model tends to produce solar fluxes that are too large by up to 50 W m−2 in regions such as the Gulf of Mexico and the western subtropical Atlantic region. Some deficiencies in the model’s surface solar radiation flux may be due to middle- and upper-level clouds and possibly in the representation of the low cloud optical properties.

SST, surface heat flux, and rain rates are compared in Fig. 8 for the stratus and control for the summer (JJA) averaged over 1987 and 1988. Figure 8a is a difference of the SST between the stratus and control run. There are significantly colder SSTs off the east (>0.75 K) and west (>1.5 K) coasts of North America. These are regions in which the low cloud amount was significantly increased (over 200% off the East Coast and 500% off the West Coast) for the stratus run (Fig. 6). Smaller differences are evident in the Southern Hemisphere and along the equator (<0.5 K). The stratus run is slightly cooler along the equator toward the central Pacific, and warmer off the coast of Central America. The differences themselves do not ensure that the SSTs for the stratus run are an improvement over the control. Therefore, a ratio between the difference in the stratus run and the Reynolds and Smith (1994) SSTs to the difference in the control and Reynolds and Smith (1994) SSTs is plotted in Fig. 8b. Shading occurs for a ratio less than 1, which indicates that the stratus run is closer to the observed SSTs. During JJA, SST improvements with the enhanced PBL stratus occur mainly in the Northern Hemisphere to about 20° south of the equator. The ratio values are very near 1 along the equator indicating little impact. Positive values indicate that the control did better than the stratus run. These mainly occur when differences between the stratus run and control (Fig. 8a) are small. The difference in the total heat flux into the ocean (Fig. 8c) is not an obvious indicator of why the SSTs are cooler in the Northern Hemisphere for the stratus run. This suggests that advective properties of the ocean may be very important in explaining the differences in SST between the stratus and control runs. There is likely greater upwelling along the coastal regions with horizontally advected cool surface water away from the coast. During this season, the subtropical high in the Northern Hemisphere Pacific is strongest with northerly low-level coastal flow that is conducive to coastal upwelling. Figure 8d is a plot of the stratus minus control rain rates for JJA. Rain rates are significantly reduced in the stratus run over Central America and westward off the Mexican coast, but increased in north-central Mexico and into the midwest United States. Large rain rates occur along 10°N in the stratus run, located south of its location in the control. This is likely due to an increased low-level convergence due to enhanced northerly winds off the North American coast in the stratus run. Furthermore, the heat fluxes from the low levels into the ocean are significantly larger in the stratus run off the Mexican coast, which may indicate a tapping of the low-level atmospheric thermal energy in the stratus run for that region.

Thermal properties of the vertical column are also influenced by the low-level stratus. Figure 9 suggests that the vertical column of longwave and shortwave heating may be influenced by the PBL stratus. The plots are for a strong episode (daily averaged) of PBL stratus off the coast of California in the model with PBL stratus compared with the control. The average PBL height is between 1000 and 900 hPa. The mean PBL height is 480 m for the region and time of the profile. The larger optical depth of the cloud near the surface incurs longwave heating (Fig. 9a) below the 900-hPa level in the PBL stratus case with a strong transition to cooling toward the cloud top. Above about 600 hPa, the column becomes more radiatively homogeneous in the PBL stratus case, hence the smaller IR cooling rates (i.e., the IR fluxes at adjacent levels are similar) compared to the control. The shortwave heating in Fig. 9b shows that the PBL stratus increases the solar warming in the layers above the cloud, up to about 500 hPa. The high albedo of the low cloud reflects the incident solar radiation at the cloud top. It is difficult to determine an exact cause for these differences since the system is nonlinear. Differences between the stratus and control heating rates at higher altitudes may be due to different synoptic conditions.

The meridional wind between 1000 and 700 hPa is illustrated in Fig. 10 for the stratus simulation, control simulation, and ECMWF analysis. The mean JJA (1987 and 1988) wind is averaged between 30° and 35°N and plotted between 150° and 90°W. This cross section intersects the North American continent near southern California. Compared to the control, the stratus simulation develops more robust meridional winds. The magnitudes for the stratus simulation compare well to the ECMWF analysis. In particular, the northerly winds below 750 hPa between 140° and 110°W (ocean) for the stratus run are between 1 and 2 m s−1 stronger than the control. Below 850 hPa for the same longitude range, the ECMWF analysis winds are greater than 4 m s−1 toward the south. These magnitudes are close to the stratus simulation values. For both the stratus run and ECMWF analysis the southerly meridional winds between 105° and 90°W are above 4 m s−1 below 800 hPa. Winds for the control are mostly less than 4 m s−1. These results not only have implications for the ocean circulation and SSTs (Fig. 8), but may also be relevant to the North American monsoon. Figure 8 suggested a decrease in North American monsoon rains with the inclusion of PBL stratus. Since much of the monsoon rain is related to mechanical forcing by the Rocky Mountains and related ranges in the United States and Mexico, a higher-resolution model would likely give different results.

Figure 11 suggests a similar improvement to the meridional wind due to the enhanced low clouds off the South American coast. This plot is a cross section for the average between 10° and 30°S and between 120° and 45°W. There is an enhanced southerly flow in the stratus run that is in better agreement with the ECMWF analysis over the ocean (approximately west of 75°W). Both the stratus and control simulations represent the winds over land poorly compared to the ECMWF analysis.

2) Winter

In the winter, the downward surface flux of solar radiation is reduced over the oceans in the stratus run compared to the control, but particularly along the coastal regions and the Northern Hemisphere (Fig. 12). Fluxes over the southeast United States are lower for the stratus simulation compared to the control, which is consistent with the larger low cloud amount in that region for the stratus simulation (Figs. 6b and 6d). The solar flux over much of the western half of South America is larger for the stratus simulation compared to the control. The control solar fluxes are too low compared to the ISCCP C2 analysis product over the South American continent. The western portion of South America is covered by a slightly higher amount of low clouds in the control (Fig. 6d) compared to the stratus run (Fig. 6b) and the ISCCP D2 analysis (Fig. 6f).

Climatologically speaking, the boreal winter months are a time when the warm ocean water in the Northern Hemisphere diminishes in magnitude in the east Pacific, and the warm SSTs of the Southern Hemisphere migrate toward the east. The convection and accompanying wind systems follow the progression of the warmer SSTs. Figure 13a and 13b represent the low-level cloud impact on the SSTs. The stratus run tends to be warmer north of the equator and along the equator to about 90°W. East of 90°W along the equator and also south of about 5°S the stratus simulation SSTs are cooler (Fig. 13a). Therefore, the SSTs are cooler in the Southern Hemisphere with the stratus. Figure 13a implies a stronger SST gradient from the Southern Hemisphere to the Northern Hemisphere for the stratus simulation compared to the control, particularly between 120° and 90°W. The SSTs are simulated better (shaded regions in Fig. 13b) in the stratus run compared to the control for nearly the whole domain, except for along the equator, east of about 120°W, and a few other small areas in the Southern Hemisphere. The rain rates in Fig. 13d indicate that the ITCZ is enhanced north of the equator around 5°N and there is a reduction of a double ITCZ in the Southern Hemisphere. Rainfall rate plots for DJF (not shown) indicate a double ITCZ in the control that is much less pronounced in the stratus run. The shift in the rainfall asymmetry about the equator is likely related to the warmer SSTs along and north of the equator for the stratus simulation compared to control (Fig. 13a).

Figure 14 shows a DJF meridional wind cross section for a meridional average between 30° and 10°S. As in the summer case there is a more robust representation of the meridional winds with the enhanced low clouds compared to the control. Winds over both the ocean and land are closer to the ECMWF analysis for the stratus run, particularly up to 700 hPa. The primary reason for the better representation over the ocean is the enhanced flow around the subtropical high in the eastern Southern Hemisphere Pacific. This configuration along the coast is conducive to coastal upwelling. These stronger southerly winds in the stratus case, along with the reduced radiation in the region (Fig. 12), help explain the colder SSTs in the stratus simulation (Fig. 13a). The enhanced northerly winds east of 75°W in the stratus case are associated with the northwesterly winds of the South American monsoon (Zhou and Lau 1998). Figure 15 shows the difference between the stratus and control runs for the temperature (left panel) and meridional winds (right panel) averaged between the 1000- and 850-hPa surfaces over the South American domain. The lower-atmospheric levels of the model over the ocean are cooler with the new stratus parameterization. Differences are more than 1 K about 20° longitude offshore. Furthermore, the temperature over the mountainous region of Peru is warmer by greater than 1 K in the stratus run. The strong thermal difference between the two runs is likely contributing to the enhanced (greater than 2 m s−1) meridional winds in the stratus case off the coast. The stronger southerly winds enhance convergence near the equator, helping to concentrate the precipitation north of the equator in the eastern Pacific Ocean region (Fig. 13d). The control run lacks a well-defined convection region north of the equator, but instead has more of a double ITCZ with a precipitation maximum both north and south of the equator as indicated by the positive in the rainfall difference in Fig. 13. This double ITCZ feature in the work of Philander et al. (1996) was partially corrected with proper incorporation of the stratus over the eastern ocean boundary of the South Pacific Ocean.

5. Interannual results

Much work in the literature revolves around model simulation of interannual behavior related to the El Niño–Southern Oscillation (ENSO) for seasonal and longer simulations. Therefore, to address this subject as related to the enhanced low-level clouds, differences between the warm and cold phase (1987 and 1988) are presented for JJA.

Figure 16 shows the JJA 1987 minus 1988 low clouds over the eastern Pacific Ocean region. The most significant difference noted is the difference near the equator, east of 150°W. The stratus simulation plot (Fig. 16a) suggests a greater cloud amount during the cold phase that agrees well with this observed (ISCCP D2) difference. This behavior is attributed to the colder eastern Pacific SSTs during the cold phase. The control cloud ENSO signal in the equatorial East Pacific is the opposite sign of the signal for the stratus simulation and observation.

The surface temperatures in Fig. 17 suggest that the variability between the cold and warm season are better represented by the stratus case over both land and ocean when compared to ECMWF analysis. The warm temperature anomaly in the equatorial region extends farther to the east for the stratus simulation as in the analysis, with a slightly broader expanse of higher temperatures around 90°W. The warm signal in the stratus case extends across the South American continent, similar to the ECMWF analysis. The cold anomaly off the coast of Peru around 90°W is also better represented with the enhanced stratus compared to the control. The temperature differences west of 150°W are similar for both the stratus and control. The reason for this is twofold. First, the thermocline depth in the central and western Pacific Ocean is deeper than in the east. Second, the larger low cloud amounts in the eastern ocean basin for the stratus simulation are likely to have a greater impact on the ocean surface through low-level winds in the east Pacific. As shown in the winter and summer sections, the enhanced low-level clouds strongly impact the alongshore winds, which are important to the eastern Pacific Ocean variability.

During DJF the impacts due to the stratus clouds are not as evident. The stratus simulation is slightly better (about 1 K larger anomaly between 120° and 90°W) in representing the differences in SSTs over the eastern equatorial Pacific Ocean and over North America compared to Reynolds and Smith (1994) analysis (not shown). There are also small improvements in the low-level wind differences, especially off the South American coast where the 1987 minus 1988 anomalies were more southeasterly compared to the control. The southeasterly wind anomaly in that region contributes to greater convergence in the equatorial central Pacific Ocean in the stratus run compared to the control.

6. Discussion

A new PBL cloud parameterization has been introduced into the FSU coupled ocean–atmosphere model. As defined, the PBL clouds depend on the PBL column thermal structure, low-level stability, wind magnitude at the PBL top, relative humidity, and surface wetness. The global representation of the surface solar radiation fluxes with this cloud parameterization is better than the simulated solar fluxes with the old cloud parameterization. The parameterization greatly improves the eastern ocean basin cloud amount during seasonal simulations and hence the solar radiation at the surface. Due to these improvements, both the low-level winds and thermal structure over the ocean and land regions are improved. The greatest improvements are in and over the eastern tropical Pacific Ocean basin. These results follow other, prior work including Slingo (1987), Philander (1996), and Ma et al. (1996). The results presented here suggest, as in these prior works, the importance of proper representation of low clouds in the eastern ocean basins, particularly for coupled model experiments. Feedback between the ocean and atmosphere are very important in the eastern Pacific Ocean where the thermocline is shallow.

The improved low clouds enhanced the eastern Pacific Ocean basin low-level atmospheric circulation. This was seen off the western coast of both North and South America. Changes in ocean heating due to solar radiation, longwave trapping within the PBL, and land–sea thermal contrasts all may contribute to the development of the stronger low-level wind. The alongshore winds are improved with the PBL stratus (Figs. 10, 11, and 14), enhancing the coastal upwelling and Ekman transport as represented in the surface temperature plots (Figs. 8 and 13). Cooler coastal surface temperatures help to further stabilize the PBL, reinforcing the existence of the stratus. The stronger southeasterly trades in the stratus simulation compared to the control along the South American coast enhance the magnitude of the ITCZ rainfall in the eastern Pacific and reduce the presence of a double ITCZ, especially during DJF. The stratus simulation SSTs are closer to the analysis SSTs in most locations where there is a large difference between the stratus case and control (Figs 8a, 8b, 13a, and 13b).

A summary of some of the model improvements is displayed in Table 1. The global and tropical low cloud amounts are vastly improved as suggested in the spatial plots (Fig. 6). Most of the enhancement is for tropical low clouds. The control global surface solar flux is closer to the ISCCP C2 global average than the stratus run, while the reverse is true in the Tropics. Rainfall correlations are considerably better with the stratus. The stratus run does 10% better than the control for the global domain and 7% better in the Tropics. A large amount of the difference is due to the winter seasons, where the stratus run is 14% better than the control (not shown). Rms errors of the SST versus Reynolds and Smith (1994) are, again, better for the stratus run compared to the control. The east Pacific domain errors are more in favor of the stratus run than in the Niño-3.4 region. Much of the error for both runs is due to the winter seasons (not shown), where rms values in the east Pacific domain were 1.5 K for the stratus run and 2.0 K for the control. Initial rms errors following the initialization process were 0.65 and 0.7 for the stratus and control runs respectively.

The interannual SST variability between the cold and warm ENSO phases of 1987 and 1989 is improved with the inclusion of the PBL clouds, particularly in the JJA. Diurnal variability of the PBL clouds is well represented along the North American coast. The changes in the land–sea temperature gradients by including the new low cloud scheme may be important in the representation of features such as the low-level jet off the Californian coast or the North American monsoon. Furthermore, it is suggested that enhancement of the northerly flow over the northern South American continent (Fig. 15) indicates an enhanced South American monsoon. Further work is needed to expound upon the existence of these features in the model and their relationship to the representation of the low-level status. Higher resolutions in the horizontal and vertical directions may be necessary.

The results in this paper are based on a new empirical parameterization that enhances the stratus clouds in seasonal simulations within the FSU coupled ocean–atmosphere model. The stratus cloud scheme depends on the Troen and Mahrt (1986) PBL height parameterization as well as the ground wetness parameterization and relative humidity. Other physical properties are likely necessary to make the stratus clouds more realistic and interactive with the model environment, including an entrainment parameterization. This initial representation of the PBL stratus is simple, efficient, and effective for seasonal simulations.

Acknowledgments

The research reported here was supported by NOAA Grants NA86GP0031 and NA76GP0521, and NASA Grant NAG5-4729. The authors would like to thank Dr. Timothy LaRow of the Center for Ocean–Atmosphere Prediction Studies and The Florida State University for his helpful suggestions and discussion. His prior and current work developing the coupled model used here made this work possible. The authors also wish to thank William B. Rossow, and the Goddard Institute for Space Studies, New York, New York, for the production of the ISCCP datasets, and the Distributed Active Archive Center (Code 902) at the Goddard Space Flight Center, Greenbelt, Maryland, for putting these data in their present format and distributing them. These distribution activities were sponsored by NASA’s Mission to Planet Earth program. Further thanks go to ECMWF and NCAR for the public availability of the ECMWF daily analyses.

REFERENCES

  • Albrecht, B. A., 1989: Aerosols, cloud microphysics, and fractional cloudiness. Science,245, 1227–1230.

  • Arakawa, A., and V. R. Lamb, 1977: Computational design of the basic dynamical processes for the UCLA general circulation model. J. Comput. Phys.,16, 173–263.

  • Bechtold, P., S. K. Krueger, and W. S. Lewellen, 1996: Modeling a stratocumulus-topped PBL: Intercomparison among different one-dimensional codes and with large eddy simulation. Bull. Amer. Meteor. Soc.,77, 2033–42.

  • Betts, A. K., and R. Boers, 1990: A cloudiness transition in a marine boundary layer. J. Atmos. Sci.,47, 1480–1497.

  • Businger, J. A., J. C. Wyngaard, Y., Izumi, and E. F. Bradly, 1971: Flux profile relationship in the atmospheric surface layer. J. Atmos. Sci.,28, 181–189.

  • Chang, C. B., 1979: On the influence of solar radiation and diurnal variation of surface temperatures on African disturbances. Rep. No. 79-3, 157 pp. [Available from Dept. of Meteorology, The Florida State University, Tallahassee, FL 32306.].

  • Delecluse, P., M. K. Davey, Y. Kitamura, S. G. H. Philander, M. Suarez, and L. Bengtsson, 1998: Coupled general circulation modeling of the tropical Pacific. J. Geophys. Res.,103, 14 357–14 373.

  • Dewey, K. F., and R. Heim, 1982: Digital archive of Northern Hemisphere snow cover, November 1966 through December 1980. Bull. Amer. Meteor. Soc.,63, 1132–1141. (Extended dataset available from Data Support Section, National Center for Atmospheric Research, Boulder, CO 80307.).

  • Frey, H., M. Latif, and T. Stockdale, 1997: The coupled GCM ECHO-2. Part I: The tropical Pacific. Mon. Wea. Rev.,125, 703–720.

  • Geleyn, J.-F., 1988: Interpolation of wind, temperature, and humidity values from model levels to the height of measurements. Tellus,40A, 347–351.

  • Holtslag, A. A. M., and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate,6, 1825–1842.

  • Klein, S. A., and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate,6, 1587–1606.

  • Koracin, G. C., and R. Berkowicz, 1988: Nocturnal boundary-layer height: Observations by acoustic sounders and predictions in terms of surface-layer parameters. Bound.-Layer Meteor.,43, 65–83.

  • Krishnamurti, T. N., S. Low-Nam, and R. Pasch, 1983: Cumulus parameterization and rainfall rates II. Mon. Wea. Rev.,111, 816–828.

  • ——, J. Xue, H. S. Bedi, K. Ingles, and D. Oosterhof, 1991: Physical initialization for numerical weather prediction over the tropics. Tellus,43A, 53–81.

  • Kuo, H. L., 1965: On formation and intensification of tropical cyclones through latent heat release by cumulus convection. J. Atmos. Sci.,22, 40–63.

  • ——, 1974: Further studies of the parameterization of the influence of cumulus convection on large scale flow. J Atmos. Sci.,31, 1232–1240.

  • LaRow, T. E., and T. N. Krishnamurti, 1998: Initial conditions and ENSO prediction using a coupled ocean–atmosphere model. Tellus,50A, 76–94.

  • Latif, M., 1987: Tropical ocean circulation experiments. J. Phys. Oceanogr.,17, 246–263.

  • ——, T. Stockdale, J. Wolff, G. Burgers, E. Maire-Reiner, M. Junge, K. Arpe, and L. Bengtsson, 1994: Climatology and variability in the ECHO coupled GCM. Tellus,46A, 351–366.

  • Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper No. 13, U.S. Government Printing Office, Washington DC, 173 pp.

  • Li, T., and S. G. H. Philander, 1996: On the annual cycle of the eastern equatorial Pacific. J. Climate,9, 2986–2998.

  • Lilly, D. K., 1968: Models of cloud-topped mixed layers under a strong inversion. Quart. J. Roy. Meteor. Soc.,94, 292–309.

  • Louis, J.-F., 1979: A parametric model of vertical eddy fluxes in the atmosphere. Bound.-Layer Meteor.,17, 187–202.

  • Ma, C.-C., C. R. Mechoso, A. W. Robertson, and A. Arakawa, 1996:Peruvian stratocumulus clouds and the tropical Pacific circulation: A coupled ocean–atmosphere GCM study. J. Climate,9, 1635–1645.

  • Norris, J. R., and C. B. Leovy, 1994: Interannual variability in stratiform cloudiness and sea surface temperature. J. Climate,7, 1915–1925.

  • Pacanowski, R., and G. Philander, 1981: Parameterization of vertical mixing in numerical models of tropical oceans. J Phys. Oceanogr.,11, 1443–1451.

  • Palmer, T. N., and D. L. T. Anderson, 1994: The prospects for seasonal forecasting—A review paper. Quart. J. Roy. Meteor. Soc.,120, 755–793.

  • Philander, S. G. H., D. Gu, D. Halpern, D. Lambert, N.-C. Lau, T. Li, and R. C. Pacanowski, 1996: Why the ITCZ is mostly north of the equator. J. Climate,9, 2958–2972.

  • Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analysis using optimal interpolation. J. Climate,7, 929–948.

  • Rogers, D. P., X. Yang, and P. M. Norris, 1995: Diurnal evolution of the cloud topped marine boundary layer. Part I: Nocturnal stratocumulus development. J. Atmos. Sci.,52, 2953–2966.

  • Rossow, W. B., and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer. Meteor. Soc.,72, 2–20.

  • ——, and L. C. Garder 1993: Validation of ISCCP cloud detections. J. Climate,6, 2370–2393.

  • ——, ——, and A. A. Lacis, 1989: Global seasonal cloud variations from satellite radiance measurements. Part I: Sensitivity of analysis. J. Climate,2, 419–458.

  • Slingo, A., and J. M. Slingo, 1991: Response of the National Center for Atmospheric Research Community Climate Model to improvements in the representation of clouds. J. Geophys. Res.,96, 15 341–15 357.

  • Slingo, J. M., 1980: A cloud parameterization scheme derived from GATE data for use with a numerical model. Quart. J. R. Meteor. Soc.,106, 747–770.

  • ——, 1987: The development and verification of a cloud prediction model for the ECMWF model. Quart. J. Roy. Meteor. Soc.,113, 899–927.

  • Smith, R. N. B., 1990: A scheme for predicting layer clouds and their water content in a general circulation model. Quart. J. Roy. Meteor. Soc.,116, 435–460.

  • Sommeria, G., and J. W. Deardorff, 1977: Subgrid-scale condensation in models of nonprecipitating clouds. J. Atmos. Sci.,34, 344–355.

  • Stricherz, J. N., J. J. O’Brien, and D. M. Legler, 1992: Atlas of Florida State University Winds for TOGA 1966–1985. The Florida State University, 250 pp. [Available from Center for Ocean–Atmospheric Prediction Studies, The Florida State University, Tallahassee, FL 32306-2840.].

  • Sundqvist, H. E., E. Berge, and J. E. Kristjansson, 1989: Condensation and cloud parameterization studies with a mesoscale NWP model. Mon. Wea. Rev.,117, 1641–1657.

  • Tiedke, M., 1993: Representation of clouds in large-scale models. Mon. Wea. Rev.,121, 3040–3061.

  • Troen, I., and L. Mahrt, 1986: A simple model of the atmospheric boundary layer: Sensitivity to surface evaporation. Bound.-Layer Meteor.,37, 129–148.

  • Wang, Q., and B. A. Albrecht, 1994: Observations of cloud-top entrainment in marine stratocumulus clouds. J. Atmos. Sci.,51, 1530–1547.

  • Warren, S. G., C. J. Hahn, J. London, R. M. Chervin, and R. L. Jenne, 1988: Global distribution of total cloud cover and cloud type amounts over the ocean. NCAR Tech. Note NCAR/TN-317+STR and Department of Energy Tech. Rep. DEO/ER-0406, 42 pp. plus 170 maps.

  • Xie, P., and P. A. Arkin, 1996: Analysis of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate,9, 840–858.

  • Zhou, J., and K. M. Lau, 1998: Does a monsoon climate exist over South America? J. Climate,11, 1029–1040.

Fig. 1.
Fig. 1.

Monthly mean daily averages of PBL heights (m) for (a) Jun and (b) Dec averaged for 1987 and 1988. The contour interval is 200 m and shading is above 1000 m

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 2.
Fig. 2.

Monthly mean of ISCCP D2 low cloud amount (%) for (a) Jun and (b) Dec (lower) averaged for 1987 and 1988. The contour interval is 20% and shading levels are 25% and 50%

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 3.
Fig. 3.

Fractional PBL stratus amount for a range of PBL heights (m) and ground wetness fraction

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 4.
Fig. 4.

Diurnal variability of the coupled model PBL stratus (%) plotted every 6 h for Los Angeles in Jun 1987. The first plot time is for 1800 UTC

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 5.
Fig. 5.

Diurnal variability of the First ISCCP Regional Experiment (FIRE) D1 stratus cloud amount (%) plotted every 6 h for Los Angeles between 1 and 5 Jul 1987

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 6.
Fig. 6.

Low cloud amount (%) for the (a) stratus run for the JJA mean averaged over 1987 and 1988, (b) stratus run for the DJF mean averaged over 1987/88 and 1988/89, (c) control run for the JJA mean averaged over 1987 and 1988, (d) control run for the DJF mean averaged over 1987/88 and 1988/89 (e) ISCCP D2 for the JJA mean averaged over 1987 and 1988, and (f) ISCCP D2 for the DJF mean averaged for 1987/88 and 1988/89. The contour interval is 20% and shading levels are 25% and 50%

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 7.
Fig. 7.

JJA mean surface shortwave radiation (W m−2) for (a) the stratus run, (b) the control, and (c) ISCCP C2 averaged over 1987 and 1988. The contour interval is 30 W m−2 and shading is above 270 W m−2

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 8.
Fig. 8.

JJA mean comparison plots between the stratus run and control for (a) SST difference (K) between the stratus run and control with an interval of 0.25 K, excluding zero; (b) SST error ratio of stratus run to control against the Reynolds and Smith (1994) SST, (c) heat flux difference (W m−2) between the stratus run and control with interval of 10 W m−2, excluding zero; and (d) rain-rate difference (mm month−1) between the stratus run and control with interval of 25 mm month−1, excluding zero. Plots are averaged over 1987 and 1988. Shading is for |SST| > 0.5 K, ratio < 1, |heat flux| > 20 W m−2, and |rain rate| > 25 mm month−1

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 9.
Fig. 9.

Heating rate profiles (K day−1) for the stratus run (dashed) and control (solid) averaged over the domain bounded by 124°W, 131°W, 32°N, and 39°N for (a) longwave and (b) shortwave radiation

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 10.
Fig. 10.

JJA mean meridional wind (m s−1) cross section for (a) the stratus run, (b) the control, and (c) ECMWF analysis averaged over 1987 and 1988. The contour levels are −4, −2, −1, 1, 2, and 4 m s−1 and shading is for |υ| > 2 m s−1. The plot is averaged between 30° and 35°N

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 11.
Fig. 11.

JJA mean meridional wind (m s−1) cross section for (a) the stratus run, (b) the control, and (c) ECMWF analysis averaged over 1987 and 1988. The contour levels are −4, −2, −1, 1, 2, and 4 m s−1 and shading is for |υ| > 2 m s−1. The plot is averaged between 30° and 10°S

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 12.
Fig. 12.

DJF mean downward surface shortwave radiation (W m−2) for (a) the stratus run, (b) the control, and (c) ISCCP C2 averaged over 1987 and 1988. The contour interval is 30 W m−2 and shading is above 270 W m−2

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 13.
Fig. 13.

DJF mean comparison plots between the stratus run and control for (a) SST difference (K) between the stratus run and control with an interval of 0.25, excluding zero; (b) SST error ratio of stratus run to control against the Reynolds and Smith (1994) SST; (c) heat flux difference (W m−2) between the stratus run and control with interval of 10, excluding zero; and (d) rain rate difference (mm month−1) between the stratus run and control with interval of 25, excluding zero. Plots are averaged over 1987 and 1988. Shading is for |SST| > 0.5, ratio < 1, |heat flux| > 20, and |rain rate| > 25

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 14.
Fig. 14.

DJF mean meridional wind (m s−1) cross section for (a) the stratus run, (b) the control, and (c) ECMWF analysis averaged over 1987 and 1988. The contour levels are −4, −2, −1, 1, 2, and 4 m s−1 and shading is for |υ| > 2 m s−1. The plot is averaged between 30°S and 10°N

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 15.
Fig. 15.

DJF difference between the stratus run and control for (a) temperature (K) and (b) meridional wind (m s−1). The plots are averaged over 1987 and 1988 and 1000- and 850-hPa levels. The contour interval is 0.2 K, excluding zero and 0.5 m s−1, excluding zero for the winds. The shading levels are for |T| > 0.6, 1 K and |υ| > 1, 2 m s−1

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 16.
Fig. 16.

JJA mean low clouds (%) difference between 1987 and 1988 for (a) the stratus run, (b) the control, and (c) ISCCP D2. The contour interval is 5% and shading is for |cloud| > 10%

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Fig. 17.
Fig. 17.

JJA mean surface temperature (K) difference between 1987 and 1988 for (a) the stratus run, (b) the control, and (c) ECMWF analysis. The contour levels are −4, −3, −2, −1, −0.5, 0.5, 1, 2, 3, and 4 K and shading is for |SST| > 1, 3 K

Citation: Monthly Weather Review 128, 9; 10.1175/1520-0493(2000)128<3083:ELLSIT>2.0.CO;2

Table 1.

Comparison of stratus and control runs against analysis averaged for all seasons used in the experiments. Low cloud amount is in %, surface solar fluxes are in W m−2, rainfall correlations are in %, and rms SST errors are in K. The rainfall correlations are against the dataset of Xie and Arkin (1996) and the rms SST errors are against Reynolds and Smith (1994)

Table 1.

1

An extended dataset is available from Data Support Section, NCAR, Boulder, CO 80307.

Save
  • Albrecht, B. A., 1989: Aerosols, cloud microphysics, and fractional cloudiness. Science,245, 1227–1230.

  • Arakawa, A., and V. R. Lamb, 1977: Computational design of the basic dynamical processes for the UCLA general circulation model. J. Comput. Phys.,16, 173–263.

  • Bechtold, P., S. K. Krueger, and W. S. Lewellen, 1996: Modeling a stratocumulus-topped PBL: Intercomparison among different one-dimensional codes and with large eddy simulation. Bull. Amer. Meteor. Soc.,77, 2033–42.

  • Betts, A. K., and R. Boers, 1990: A cloudiness transition in a marine boundary layer. J. Atmos. Sci.,47, 1480–1497.

  • Businger, J. A., J. C. Wyngaard, Y., Izumi, and E. F. Bradly, 1971: Flux profile relationship in the atmospheric surface layer. J. Atmos. Sci.,28, 181–189.

  • Chang, C. B., 1979: On the influence of solar radiation and diurnal variation of surface temperatures on African disturbances. Rep. No. 79-3, 157 pp. [Available from Dept. of Meteorology, The Florida State University, Tallahassee, FL 32306.].

  • Delecluse, P., M. K. Davey, Y. Kitamura, S. G. H. Philander, M. Suarez, and L. Bengtsson, 1998: Coupled general circulation modeling of the tropical Pacific. J. Geophys. Res.,103, 14 357–14 373.

  • Dewey, K. F., and R. Heim, 1982: Digital archive of Northern Hemisphere snow cover, November 1966 through December 1980. Bull. Amer. Meteor. Soc.,63, 1132–1141. (Extended dataset available from Data Support Section, National Center for Atmospheric Research, Boulder, CO 80307.).

  • Frey, H., M. Latif, and T. Stockdale, 1997: The coupled GCM ECHO-2. Part I: The tropical Pacific. Mon. Wea. Rev.,125, 703–720.

  • Geleyn, J.-F., 1988: Interpolation of wind, temperature, and humidity values from model levels to the height of measurements. Tellus,40A, 347–351.

  • Holtslag, A. A. M., and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate,6, 1825–1842.

  • Klein, S. A., and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate,6, 1587–1606.

  • Koracin, G. C., and R. Berkowicz, 1988: Nocturnal boundary-layer height: Observations by acoustic sounders and predictions in terms of surface-layer parameters. Bound.-Layer Meteor.,43, 65–83.

  • Krishnamurti, T. N., S. Low-Nam, and R. Pasch, 1983: Cumulus parameterization and rainfall rates II. Mon. Wea. Rev.,111, 816–828.

  • ——, J. Xue, H. S. Bedi, K. Ingles, and D. Oosterhof, 1991: Physical initialization for numerical weather prediction over the tropics. Tellus,43A, 53–81.

  • Kuo, H. L., 1965: On formation and intensification of tropical cyclones through latent heat release by cumulus convection. J. Atmos. Sci.,22, 40–63.

  • ——, 1974: Further studies of the parameterization of the influence of cumulus convection on large scale flow. J Atmos. Sci.,31, 1232–1240.

  • LaRow, T. E., and T. N. Krishnamurti, 1998: Initial conditions and ENSO prediction using a coupled ocean–atmosphere model. Tellus,50A, 76–94.

  • Latif, M., 1987: Tropical ocean circulation experiments. J. Phys. Oceanogr.,17, 246–263.

  • ——, T. Stockdale, J. Wolff, G. Burgers, E. Maire-Reiner, M. Junge, K. Arpe, and L. Bengtsson, 1994: Climatology and variability in the ECHO coupled GCM. Tellus,46A, 351–366.

  • Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper No. 13, U.S. Government Printing Office, Washington DC, 173 pp.

  • Li, T., and S. G. H. Philander, 1996: On the annual cycle of the eastern equatorial Pacific. J. Climate,9, 2986–2998.

  • Lilly, D. K., 1968: Models of cloud-topped mixed layers under a strong inversion. Quart. J. Roy. Meteor. Soc.,94, 292–309.

  • Louis, J.-F., 1979: A parametric model of vertical eddy fluxes in the atmosphere. Bound.-Layer Meteor.,17, 187–202.

  • Ma, C.-C., C. R. Mechoso, A. W. Robertson, and A. Arakawa, 1996:Peruvian stratocumulus clouds and the tropical Pacific circulation: A coupled ocean–atmosphere GCM study. J. Climate,9, 1635–1645.

  • Norris, J. R., and C. B. Leovy, 1994: Interannual variability in stratiform cloudiness and sea surface temperature. J. Climate,7, 1915–1925.

  • Pacanowski, R., and G. Philander, 1981: Parameterization of vertical mixing in numerical models of tropical oceans. J Phys. Oceanogr.,11, 1443–1451.

  • Palmer, T. N., and D. L. T. Anderson, 1994: The prospects for seasonal forecasting—A review paper. Quart. J. Roy. Meteor. Soc.,120, 755–793.

  • Philander, S. G. H., D. Gu, D. Halpern, D. Lambert, N.-C. Lau, T. Li, and R. C. Pacanowski, 1996: Why the ITCZ is mostly north of the equator. J. Climate,9, 2958–2972.

  • Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analysis using optimal interpolation. J. Climate,7, 929–948.

  • Rogers, D. P., X. Yang, and P. M. Norris, 1995: Diurnal evolution of the cloud topped marine boundary layer. Part I: Nocturnal stratocumulus development. J. Atmos. Sci.,52, 2953–2966.

  • Rossow, W. B., and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer. Meteor. Soc.,72, 2–20.

  • ——, and L. C. Garder 1993: Validation of ISCCP cloud detections. J. Climate,6, 2370–2393.

  • ——, ——, and A. A. Lacis, 1989: Global seasonal cloud variations from satellite radiance measurements. Part I: Sensitivity of analysis. J. Climate,2, 419–458.

  • Slingo, A., and J. M. Slingo, 1991: Response of the National Center for Atmospheric Research Community Climate Model to improvements in the representation of clouds. J. Geophys. Res.,96, 15 341–15 357.

  • Slingo, J. M., 1980: A cloud parameterization scheme derived from GATE data for use with a numerical model. Quart. J. R. Meteor. Soc.,106, 747–770.

  • ——, 1987: The development and verification of a cloud prediction model for the ECMWF model. Quart. J. Roy. Meteor. Soc.,113, 899–927.

  • Smith, R. N. B., 1990: A scheme for predicting layer clouds and their water content in a general circulation model. Quart. J. Roy. Meteor. Soc.,116, 435–460.

  • Sommeria, G., and J. W. Deardorff, 1977: Subgrid-scale condensation in models of nonprecipitating clouds. J. Atmos. Sci.,34, 344–355.

  • Stricherz, J. N., J. J. O’Brien, and D. M. Legler, 1992: Atlas of Florida State University Winds for TOGA 1966–1985. The Florida State University, 250 pp. [Available from Center for Ocean–Atmospheric Prediction Studies, The Florida State University, Tallahassee, FL 32306-2840.].

  • Sundqvist, H. E., E. Berge, and J. E. Kristjansson, 1989: Condensation and cloud parameterization studies with a mesoscale NWP model. Mon. Wea. Rev.,117, 1641–1657.

  • Tiedke, M., 1993: Representation of clouds in large-scale models. Mon. Wea. Rev.,121, 3040–3061.

  • Troen, I., and L. Mahrt, 1986: A simple model of the atmospheric boundary layer: Sensitivity to surface evaporation. Bound.-Layer Meteor.,37, 129–148.

  • Wang, Q., and B. A. Albrecht, 1994: Observations of cloud-top entrainment in marine stratocumulus clouds. J. Atmos. Sci.,51, 1530–1547.

  • Warren, S. G., C. J. Hahn, J. London, R. M. Chervin, and R. L. Jenne, 1988: Global distribution of total cloud cover and cloud type amounts over the ocean. NCAR Tech. Note NCAR/TN-317+STR and Department of Energy Tech. Rep. DEO/ER-0406, 42 pp. plus 170 maps.

  • Xie, P., and P. A. Arkin, 1996: Analysis of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate,9, 840–858.

  • Zhou, J., and K. M. Lau, 1998: Does a monsoon climate exist over South America? J. Climate,11, 1029–1040.

  • Fig. 1.

    Monthly mean daily averages of PBL heights (m) for (a) Jun and (b) Dec averaged for 1987 and 1988. The contour interval is 200 m and shading is above 1000 m

  • Fig. 2.

    Monthly mean of ISCCP D2 low cloud amount (%) for (a) Jun and (b) Dec (lower) averaged for 1987 and 1988. The contour interval is 20% and shading levels are 25% and 50%

  • Fig. 3.

    Fractional PBL stratus amount for a range of PBL heights (m) and ground wetness fraction

  • Fig. 4.

    Diurnal variability of the coupled model PBL stratus (%) plotted every 6 h for Los Angeles in Jun 1987. The first plot time is for 1800 UTC

  • Fig. 5.

    Diurnal variability of the First ISCCP Regional Experiment (FIRE) D1 stratus cloud amount (%) plotted every 6 h for Los Angeles between 1 and 5 Jul 1987

  • Fig. 6.

    Low cloud amount (%) for the (a) stratus run for the JJA mean averaged over 1987 and 1988, (b) stratus run for the DJF mean averaged over 1987/88 and 1988/89, (c) control run for the JJA mean averaged over 1987 and 1988, (d) control run for the DJF mean averaged over 1987/88 and 1988/89 (e) ISCCP D2 for the JJA mean averaged over 1987 and 1988, and (f) ISCCP D2 for the DJF mean averaged for 1987/88 and 1988/89. The contour interval is 20% and shading levels are 25% and 50%

  • Fig. 7.

    JJA mean surface shortwave radiation (W m−2) for (a) the stratus run, (b) the control, and (c) ISCCP C2 averaged over 1987 and 1988. The contour interval is 30 W m−2 and shading is above 270 W m−2

  • Fig. 8.

    JJA mean comparison plots between the stratus run and control for (a) SST difference (K) between the stratus run and control with an interval of 0.25 K, excluding zero; (b) SST error ratio of stratus run to control against the Reynolds and Smith (1994) SST, (c) heat flux difference (W m−2) between the stratus run and control with interval of 10 W m−2, excluding zero; and (d) rain-rate difference (mm month−1) between the stratus run and control with interval of 25 mm month−1, excluding zero. Plots are averaged over 1987 and 1988. Shading is for |SST| > 0.5 K, ratio < 1, |heat flux| > 20 W m−2, and |rain rate| > 25 mm month−1

  • Fig. 9.

    Heating rate profiles (K day−1) for the stratus run (dashed) and control (solid) averaged over the domain bounded by 124°W, 131°W, 32°N, and 39°N for (a) longwave and (b) shortwave radiation

  • Fig. 10.

    JJA mean meridional wind (m s−1) cross section for (a) the stratus run, (b) the control, and (c) ECMWF analysis averaged over 1987 and 1988. The contour levels are −4, −2, −1, 1, 2, and 4 m s−1 and shading is for |υ| > 2 m s−1. The plot is averaged between 30° and 35°N

  • Fig. 11.

    JJA mean meridional wind (m s−1) cross section for (a) the stratus run, (b) the control, and (c) ECMWF analysis averaged over 1987 and 1988. The contour levels are −4, −2, −1, 1, 2, and 4 m s−1 and shading is for |υ| > 2 m s−1. The plot is averaged between 30° and 10°S

  • Fig. 12.

    DJF mean downward surface shortwave radiation (W m−2) for (a) the stratus run, (b) the control, and (c) ISCCP C2 averaged over 1987 and 1988. The contour interval is 30 W m−2 and shading is above 270 W m−2

  • Fig. 13.

    DJF mean comparison plots between the stratus run and control for (a) SST difference (K) between the stratus run and control with an interval of 0.25, excluding zero; (b) SST error ratio of stratus run to control against the Reynolds and Smith (1994) SST; (c) heat flux difference (W m−2) between the stratus run and control with interval of 10, excluding zero; and (d) rain rate difference (mm month−1) between the stratus run and control with interval of 25, excluding zero. Plots are averaged over 1987 and 1988. Shading is for |SST| > 0.5, ratio < 1, |heat flux| > 20, and |rain rate| > 25

  • Fig. 14.

    DJF mean meridional wind (m s−1) cross section for (a) the stratus run, (b) the control, and (c) ECMWF analysis averaged over 1987 and 1988. The contour levels are −4, −2, −1, 1, 2, and 4 m s−1 and shading is for |υ| > 2 m s−1. The plot is averaged between 30°S and 10°N

  • Fig. 15.

    DJF difference between the stratus run and control for (a) temperature (K) and (b) meridional wind (m s−1). The plots are averaged over 1987 and 1988 and 1000- and 850-hPa levels. The contour interval is 0.2 K, excluding zero and 0.5 m s−1, excluding zero for the winds. The shading levels are for |T| > 0.6, 1 K and |υ| > 1, 2 m s−1

  • Fig. 16.

    JJA mean low clouds (%) difference between 1987 and 1988 for (a) the stratus run, (b) the control, and (c) ISCCP D2. The contour interval is 5% and shading is for |cloud| > 10%

  • Fig. 17.

    JJA mean surface temperature (K) difference between 1987 and 1988 for (a) the stratus run, (b) the control, and (c) ECMWF analysis. The contour levels are −4, −3, −2, −1, −0.5, 0.5, 1, 2, 3, and 4 K and shading is for |SST| > 1, 3 K

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 193 113 5
PDF Downloads 42 12 1