Abstract

Using a global nonhydrostatic model with explicit cloud processes, upper-cloud changes are investigated by comparing the present climate condition under the perpetual July setting and the global warming condition, in which the sea surface temperature (SST) is raised by 2°. The sensitivity of the upper-cloud cover and the ice water path (IWP) are investigated through a set of experiments. The responses of convective mass flux and convective areas are also examined, together with those of the large-scale subsidence and relative humidity in the subtropics. The responses of the IWP and the upper-cloud cover are found to be opposite; that is, as the SST increases, the IWP averaged over the tropics decreases, whereas the upper-cloud cover in the tropics increases. To clarify the IWP response, a simple conceptual model is constructed. The model consists of three columns of deep convective core, anvil, and environmental subsidence regions. The vertical profiles of hydrometers are predicted with cloud microphysics processes and kinematically prescribed circulation. The reduction in convective mass flux is found to be a primary factor in the decrease of the IWP under the global warming condition. Even when a different and more comprehensive cloud microphysics scheme is used, the reduction in the IWP due to the mass flux change is also confirmed.

1. Introduction

One of the fundamental issues in climate projection research is to clarify mechanisms of cloud changes associated with future expected global warming. Clouds are the most ambiguous factor in the uncertainty of climate projection (Soden and Held 2006). Although the albedo effect of lower clouds is most variable for the climate sensitivity (Bony and Dufresne 2005), there is no consensus as to how upper clouds will change in the future. Two opposing views are proposed to explain upper-cloud changes under a warming condition. The thermostat hypothesis postulates that upper clouds increase as the sea surface temperature (SST) increases (Ramanathan and Collins 1991), whereas the iris effect suggests that upper clouds decrease as the SST increases (Lindzen et al. 2001). These opposing views have been investigated through observations and modeling studies (Del Genio and Kovari 2002; Del Genio et al. 2005; Clement and Soden 2005; Rondanelli and Lindzen 2008).

Several current climate models have shown a general decrease in the upper-cloud cover in the future climate (Ringer et al. 2006; Zelinka and Hartmann 2010). On the one hand, the manner in which the proposed mechanisms work in the climate models and whether the model results are robust remain unclear. On the other hand, recent results obtained using a high-resolution global nonhydrostatic model with explicit cloud microphysics revealed that upper clouds increase under the global warming condition (Collins and Satoh 2009, hereafter CS). Using the Nonhydrostatic Icosahedral Atmospheric Model (NICAM; Tomita and Satoh 2004; Satoh et al. 2008), Miura et al. (2005) examined the cloud changes between the control experiment (CTL) and an experiment in which the SST was increased under the aquaplanet condition. The results indicated that clouds in high latitudes increase with SST. Similar results were obtained by Wyant et al. (2006), who used a Multiscale Modeling Framework (MMF), or superparameterization, to estimate climate sensitivity. With respect to the cloud response in the low latitudes, Miura et al. (2005) showed that upper clouds extend farther into the subtropics and that the cloud fraction at the equator decreases, which is a result of the control experimental configuration of the aquaplanet experiment, in which the latitudinal width of the warm region is smaller than the observed SST profile (Neale and Hoskins 2001). Iga et al. (2007) analyzed quasi-equilibrated states of more realistic experiments under the perpetual July condition of the NICAM simulation and reproduced cloud distributions comparable to the observations. Iga et al. (2011) further analyzed the properties of the upper clouds simulated by the sensitivity experiments using NICAM, and showed a relation between the ice water path (IWP) and the upper-cloud fraction.

Following Iga et al. (2007, 2011), CS investigated cloud changes for the global warming designed by the Cloud Feedback Model Intercomparison Program (CFMIP) through a NICAM simulation in which the mesh size was approximately 7 km. CS examined the change in the simulated cloud fraction categorized by the International Satellite Cloud Climatology Project (ISCCP) (Rossow and Schiffer 1999) and showed that the upper-cloud fraction is more widespread, particularly from the tropics to the subtropics, under the warm SST condition (+2K-exp). This is in contrast to the results obtained by coarser-resolution general circulation models (GCM), which generally indicate a decrease in the upper-thin-cloud fraction in the active convective region (Ringer et al. 2006). Thus, the reason why this difference occurs and the mechanism that determines the upper-cloud response under the warmer SST condition are of interest. NICAM uses a cloud microphysics scheme and explicitly simulates deep convective circulation, whereas GCMs use conventional cumulus parameterization schemes to represent the statistical effects of deep convection. Although the different implementations of cloud schemes are thought to result in the different responses of upper clouds, no detailed mechanism has yet been reported in this regard.

In the present study, we focus on changes in upper clouds by a series of CFMIP-type global warming experiments using NICAM and investigate possible mechanisms of the upper-cloud changes. Comparison of the results obtained from NICAM and the coarser-resolution GCM will be discussed in a separate paper (Y. Tsushima et al. 2011, unpublished manuscript). In section 2, the model and experimental setup are described. In section 3, the results obtained from the numerical experiments are presented, and the relations between the upper-cloud properties, the hydrological cycle, the tropical overturning circulations, and the convective properties are discussed. We further investigate possible mechanisms of the change in the IWP by constructing a simple conceptual model in section 4. A summary and discussions are presented in section 5.

2. Model and experimental design

We use a global nonhydrostatic model with an explicit cloud microphysics scheme to investigate the responses of upper clouds. Numerical simulations are performed using NICAM (Tomita and Satoh 2004; Satoh et al. 2008), which explicitly calculates deep convective circulation without using a cumulus parameterization scheme. The experiments conducted in the present study follow the CFMIP. The control experiment (CTL-exp) adopts the SST distribution of the present condition of July climatology (perpetual July), and the warming experiment (+2K-exp) adopts the SST with an increase of 2° over the SST of CTL-exp (Cess et al. 1990). The basic properties of climatology and energy budget have been reported by Iga et al. (2007, 2011) along with the sensitivities to the experimental parameters, as described below. In the present study, we performed four pairs of experiments, as listed in Table 1, where each pair of experiments consists of CTL and +2K experiments. The sensitivities to the cloud microphysics parameter, the horizontal resolution, and the planetary boundary layer (PBL) scheme are investigated. The microphysics parameter Cs, which is the coefficient of snow sedimentation rate, is changed from the default value Cs = 4 to Cs = 3 (denoted by CS4 and CS3, respectively), where CS3 leads to slow sedimentation of snowfall. The PBL scheme developed by Mellor and Yamada (1974) and Smith (1997) is used as a control, in which the master length is changed from the default value of L = 100 m to 200 m [L100 and L200, respectively; the experiment with L100 is used by Iga et al. (2011)]. An improved version of the Mellor–Yamada scheme developed by Nakanishi and Niino (2004, 2006; MYNN), in which the mixing length is internally calculated diagnostically, is also used. The horizontal grid spacing is approximately 14 km for three cases (CS3L200, CS4L200, and CS4MYNN) and is approximately 7 km for CS3L200dx7.

Table 1.

Summary of the experiments. All of the experiments are conducted using two cases of SST, namely, CTL and +2K. Here, L0 is the mixing length used for the PBL scheme. For CS4MYNN, the level-2 turbulence scheme of Nakanishi and Niino (2006) was used (MYNN scheme). The horizontal grid interval is denoted by dx.

Summary of the experiments. All of the experiments are conducted using two cases of SST, namely, CTL and +2K. Here, L0 is the mixing length used for the PBL scheme. For CS4MYNN, the level-2 turbulence scheme of Nakanishi and Niino (2006) was used (MYNN scheme). The horizontal grid interval is denoted by dx.
Summary of the experiments. All of the experiments are conducted using two cases of SST, namely, CTL and +2K. Here, L0 is the mixing length used for the PBL scheme. For CS4MYNN, the level-2 turbulence scheme of Nakanishi and Niino (2006) was used (MYNN scheme). The horizontal grid interval is denoted by dx.

The integration is performed for 90 days, and the last 30 days are used for the analysis. Since the experiment is conducted under the perpetual July condition, no seasonal change in the monsoon circulations was observed. We focus on the statistical properties of clouds using the time-averaged fields and do not discuss the time variability of clouds. Although the simulation contains the signal of the intraseasonal oscillation, these variations are smoothed by averaging over a period of one month. Time sequences of the precipitable water show that robust differences of the precipitable water are confirmed between each pair of the CTL and +2K experiments (figures not shown).

As described above, we use relatively coarser mesh sizes of approximately 7 and 14 km for the sensitivity experiments. At the highest resolution, NICAM simulations with a mesh size of approximately 3.5 km have been conducted (Tomita et al. 2005; Miura et al. 2007). The cloud properties of the NICAM simulations are evaluated by comparing satellite data (Inoue et al. 2008, 2010; Masunaga et al. 2008; Satoh et al. 2010; Noda et al. 2010). The general characteristics of the cloud distribution are found to be similar, even for the case in which the mesh size is coarsened to approximately 7 or 14 km, although the cloud thickness and the size of the mesoscale convective system depend quantitatively on resolution. These results are encouraging because nonhydrostatic models are generally used for mesh sizes of less than approximately 5 km. Based on the above considerations and on the limitation of the computer resources, we herein conduct NICAM simulations, the mesh size of which is approximately 7 and 14 km, for the sensitivity experiments shown below.

3. Numerical results

First, to provide a general view of the results, Fig. 1 shows the geographical distribution of the change in the ISCCP upper-cloud fraction for the highest-resolution simulation (CS3L200dx7). The experimental data are continuously processed to provide ISCCP cloud fractions using the ISCCP simulator (Klein and Jakob 1999). The ISCCP project classifies clouds according to cloud-top height and cloud optical thickness. Figure 1 shows the change in the upper-cloud fraction integrated over the entire cloud optical depth range. Under the warm SST condition (+2K-exp), the NICAM results show that the upper cloud is spread more widely, particularly in the region from the tropics to the subtropics, as described in CS.

Fig. 1.

Difference in upper-cloud fraction classified by the ISCCP simulator between the CTL-exp and +2K-exp (CS3L200dx7) (%).

Fig. 1.

Difference in upper-cloud fraction classified by the ISCCP simulator between the CTL-exp and +2K-exp (CS3L200dx7) (%).

The latitudinal distribution of the change in the zonal and time mean upper-cloud fractions and the column-integrated IWP for each experiment are shown in Figs. 2 and 3, respectively. The upper-cloud fractions of Fig. 2 are defined by the ISCCP classification; that is, the cloud-top pressure is between 50 and 440 hPa and the cloud total thickness is larger than 0.1 (Rossow and Schiffer 1999; Klein and Jakob 1999). The sensitivity to the parameters for CTL-exp is described by Iga et al. (2007, 2011). For example, as the snow sedimentation becomes slower (from CS4L200 to CS3L200), the upper-cloud fraction increases. The effect of the ice sedimentation on the upper cloud is also investigated in previous papers (Wu et al. 1999; Grabowski 2000, 2003; Satoh and Matsuda 2009). Figure 2 shows that the upper clouds—as classified by ISCCP—generally become wider under the warmed condition. The level of change in the upper clouds is similar for CS4L200, CS3L200, and CS3L200dx7, namely, approximately 10%–20% in the low latitudes; however, the change for CS4MYNN is more modest, that is, less than approximately 5%. In the subtropics near the latitude belts around 30°S and 30°N, the change for CS3L200dx7 is most pronounced, so that the geographical distribution shown in Fig. 1 is regarded as the most extreme case.

Fig. 2.

Latitudinal distribution of ISCCP upper-cloud fraction for (top) the control experiments, CTL, and (bottom) the differences between the results of the control experiments and the +2K experiments (SST+2K − CTL).

Fig. 2.

Latitudinal distribution of ISCCP upper-cloud fraction for (top) the control experiments, CTL, and (bottom) the differences between the results of the control experiments and the +2K experiments (SST+2K − CTL).

Fig. 3.

Latitudinal distribution of (top) the IWP and (bottom) the difference between the results of the control experiments and the +2K experiments (kg m−2).

Fig. 3.

Latitudinal distribution of (top) the IWP and (bottom) the difference between the results of the control experiments and the +2K experiments (kg m−2).

Figure 3 shows the latitudinal distributions of the IWPs and their changes. This figure indicates that the IWP decreases as the SST increases near the tropics, except for CS4MYNN. Intuitively, the IWP might be expected to increase as the upper cloud becomes wider, but the results of the present study indicate the opposite tendency. Iga et al. (2011) examined the sensitivities of the IWP for the set of CTL-exp and found that the IWP generally decreases as the upper-cloud cover increases. According to Fig. 7 of Iga et al. (2011), the upper-cloud areas shrink to deep convective regions for the case in which the IWP is more abundant. Iga et al. (2011) speculated that as the IWP increases, hydrometeors of heavier categories increase and fall more rapidly. The hydrometeors of snow category are heavier than the hydrometeors of the cloud ice category in the present cloud microphysics scheme (Grabowski 1998, hereafter G98). Thus, the state in which the IWP is more abundant corresponds to smaller upper-cloud fractions. A similar sensitivity was reported by Satoh and Matsuda (2009), in which the upper-cloud cover decreases as the contribution of graupel to the IWP increases. In general, if the ratio of heavier categories of hydrometeors increases, then the precipitation efficiency increases and upper-cloud areas become smaller. The above results are summarized in Table 2, which shows the upper-cloud fractions averaged over 30°S–30°N and IWP averaged over 20°S–20°N. Ice water path generally decreases under the warmer condition, whereas the ISCCP upper-cloud fractions increase.

Table 2.

Comparisons of upper-cloud fractions and IWP for CTL-exp and +2K-exp. Upper-cloud fractions are averaged between 30°S and 30°N, and IWPs are averaged between 20°S and 20°N (kg m−2).

Comparisons of upper-cloud fractions and IWP for CTL-exp and +2K-exp. Upper-cloud fractions are averaged between 30°S and 30°N, and IWPs are averaged between 20°S and 20°N (kg m−2).
Comparisons of upper-cloud fractions and IWP for CTL-exp and +2K-exp. Upper-cloud fractions are averaged between 30°S and 30°N, and IWPs are averaged between 20°S and 20°N (kg m−2).

Figure 4 shows vertical profiles of relative humidity averaged over 30°S–30°N and their differences between CTL-exp and +2K-exp. For CTL-exp, Fig. 4a, together with Figs. 2a and 3a, indicates that the relative humidity in the free troposphere generally increases as the upper-cloud cover increases and the IWP decreases. Note that relative humidity is related more closely to the upper-cloud cover than to the IWP. The response under the warming condition (Fig. 4b) indicates that the relative humidity in the free troposphere generally increases under the more humid condition and decreases near the upper troposphere. This vertical profile is consistent with Sherwood et al. (2010), who showed that climate model projection simulations indicate that tropospheric relative humidity patterns are similar in different models.

Fig. 4.

Vertical profiles of relative humidity for the control experiments averaged over (left) 30°S–30°N and (right) changes in relative humidity between the control experiments and the +2K experiments (%).

Fig. 4.

Vertical profiles of relative humidity for the control experiments averaged over (left) 30°S–30°N and (right) changes in relative humidity between the control experiments and the +2K experiments (%).

Figure 4 also indicates that the tropopause height shifts upward as the SST becomes warmer. The relative humidity near the tropopause is close to the saturated humidity, and thus the upward shift of the tropopause leads to an increase in relative humidity above 100 hPa and a decrease in relative humidity below 100 hPa. The relative humidity profiles are also plotted with respect to temperature in Fig. 5, together with the ice water content (IWC) and cloud fraction (defined by the threshold value of total hydrometeors of 0.005 g kg−1; Tompkins and Craig 1998). Figure 6 shows their changes between CTL-exp and +2K-exp. For the case of CS4MYNN, the profiles of the three variables are very similar between CTL-exp and +2K-exp. For the other cases, both the IWC and the upper-cloud fraction become larger around 200–210 K, while the IWC becomes smaller around 240 K. The profiles of the changes in relative humidity are similar to the changes in cloud fraction, but not to the changes in the IWC. Thus, in a broad sense, these profiles are almost unchanged between CTL-exp and +2K-exp except for 200–220 K. These results, especially for CS4MYNN, are consistent with the fixed anvil temperature (FAT) hypothesis proposed by Hartmann and Larson (2002), with an additional factor of the response of upper thin clouds in the layer 200–220 K.

Fig. 5.

Vertical profiles for (left) relative humidity (%), (middle) IWC (kg m−3), and (right) cloud fraction (%) as a function of temperature (unit of vertical axis: K).

Fig. 5.

Vertical profiles for (left) relative humidity (%), (middle) IWC (kg m−3), and (right) cloud fraction (%) as a function of temperature (unit of vertical axis: K).

Fig. 6.

As in Fig. 5, but for changes between the control experiments and the +2K experiments.

Fig. 6.

As in Fig. 5, but for changes between the control experiments and the +2K experiments.

Figure 7 shows the vertical distributions of omega velocity averaged over 30°–10°S, which corresponds to the subsidence region of the southern branch of the Hadley circulation. The sensitivity of the large-scale circulation for the set of CTL-exp was reported by Iga et al. (2011). As the upper cloud prevails, the Hadley circulation weakens, as inferred from the energy balance constraint. The response to the SST increase (Fig. 7b) indicates that the subsidence generally decreases, which is consistent with previous studies (Held and Soden 2006; Vecchi and Soden 2007).

Fig. 7.

As in Fig. 4, but for vertical profiles of omega velocity averaged over 30°–10°S.

Fig. 7.

As in Fig. 4, but for vertical profiles of omega velocity averaged over 30°–10°S.

Another view of the circulation change is shown in Fig. 8, which shows the vertical distributions of convective mass flux. This is averaged over the columns for which the vertical velocity exceeds 1 m s−1 at an altitude of 5 km. The analyzed region is between 30°S and 30°N, which covers the intertropical convergence zone (ITCZ). Figure 8 implies that the convective mass flux decreases as the Hadley subsidence decreases, which is consistent with the findings of Vecchi and Soden (2007). This tendency is valid for both the sensitivity for the set of CTL-exp and the difference between CTL- and +2K-exp. Figure 9 shows the vertical distributions of the area coverage of intense convection, the vertical velocity of which exceeds 1 m s−1 at an altitude of 5 km. The area fraction of the deep convective core is approximately 0.1% at the maximum in the layer around 5–7 km. This frequency of the strong vertical wind of the NICAM data is consistent with observational data of the equatorial atmosphere radar (H. Kubokawa et al. 2011, personal communication). Figures 8 and 9 indicate that the mass flux change is approximately explained by the change in the fractional area of convection. This tendency remains the same if the threshold value of deep convection is relaxed to 0.1 m s−1, in which case the fractional area becomes approximately 2% in the middle of the troposphere. It is generally thought that the large-scale overturning weakens as the SST increases because the static stability effect overwhelms the radiative cooling effect. The decrease in the large-scale overturning corresponds to the decrease in the fractional area of the intensive convective core. This relation is first obtained in the present study using the global nonhydrostatic model by explicitly resolving deep convective circulations and is consistent with the results obtained by a smaller domain cloud-resolving simulation (Cohen and Craig 2006).

Fig. 8.

As in Fig. 4, but for convective mass flux averaged over 30°S–30°N (unit of horizontal axis: kg m−2 s−1).

Fig. 8.

As in Fig. 4, but for convective mass flux averaged over 30°S–30°N (unit of horizontal axis: kg m−2 s−1).

Fig. 9.

As in Fig. 4, but for a fractional area of intense convection averaged over 30°S–30°N.

Fig. 9.

As in Fig. 4, but for a fractional area of intense convection averaged over 30°S–30°N.

It is interesting to note that the response of the upper clouds and convective mass flux are different by the choice of the PBL scheme. For CS4MYNN, as the SST increases, the upper-cloud fraction becomes slightly smaller and the IWP does not change robustly (Fig. 3). The Hadley subsidence decreases in the middle troposphere (Fig. 7b), whereas the convective mass flux and the fractional area of convection are almost unchanged in the middle of the troposphere, but increases in the upper troposphere (Figs. 8b and 9b). The reason for this difference between the experiments with MYNN and the other PBL scheme is likely related to the treatment of the subgrid moist process and the mixing length assumed in the PBL scheme. Such behavior of the MYNN has been reported by Noda et al. (2010) and Iga et al. (2011).

Figures 2 and 3 show that CS4MYNN has the largest IWP and the smallest upper-cloud fractions near the tropics under the present climate condition. Generally, the turbulent flux contributes to the vertical heat and moisture fluxes in the free troposphere above the PBL and may affect the convective mass flux and upper-cloud ice. The magnitude of the turbulent flux depends on the mixing length of the turbulent scheme. In the upper layer, as the mixing length becomes longer, the turbulent flux becomes enhanced, and both the convective mass flux and the explicit water convective flux decrease (figures not shown).

As for the change between the CTL and +2K experiments, Fig. 8 shows that the mass flux of CS4MYNN is almost unchanged or slightly increased. This suggests that the turbulent flux of CS4MYNN is reduced in the free troposphere in order for the explicit convective mass flux to be almost unchanged. Since the MYNN scheme uses a diagnosed mixing length (Nakanishi and Niino 2004), the turbulent flux of CS4MYNN is more sensitive to the environmental condition than that of the other cases. One possible reason for the reduction of the turbulent flux is the increase in the static stability under the +2K experiment in the MYNN scheme. Since the upper layer is close to the saturation condition, the mixing length of CS4MYNN tends to be more variable in the upper layer.

Figure 7 shows that the omega subsidence decreases in the middle of the troposphere for CS4MYNN as for the other cases. Figures 7 and 8 imply that the weaker updrafts decrease in the case of CS4MYNN. In fact, the contribution of the weaker updraft with the threshold value of w > 0.1 m s−1 to the convective mass flux is much more than the contribution of the strong updraft with the threshold value of w > 1 m s−1 [see Fig. 5 of Iga et al. (2011)]. We do not yet resolve the mechanism why such a different response in the updraft spectrum is caused by the use of the different PBL schemes. Further study is required.

4. Column model analysis

We attempt to clarify the response of the upper-cloud changes described in the previous section using a simple conceptual model. We specifically attempt to clarify the possible mechanisms behind the decrease in the IWP in the tropics as the SST increases. A number of factors affect the IWP as the SST becomes higher. The assumption here is that the tropospheric temperature profile approximately follows a moist adiabatic profile starting from the SST specified at the surface.

  • As the temperature increases, the water content in the air might increase according to the Clausius–Clapeyron relation. This is true at least for the upward motion branch of deep convection, but it remains unclear as to whether the water content in the downward branch and ice/liquid airborne clouds increases.

  • As the SST increases, the altitude of the melting level increases, which leads to a decrease in the IWP.

  • As the SST increases, the tropopause height generally becomes higher according to the radiative–convective equilibrium constraint, which leads to an increase in the IWP.

  • Cloud microphysical processes might change with temperature because some processes of cloud microphysics schemes are directly related to temperature. Although possible mechanisms have been proposed (Tompkins and Craig 1999), the effects of changes in the cloud microphysics processes on the IWP remain unclear.

  • Changes in circulation may affect the IWP. However, this has not yet been explicitly discussed.

Tompkins and Craig (1999) used a three-dimensional nonhydrostatic model to examine cloud responses to the SST changes under the radiative–convective equilibrium condition. They revealed changes in the tropopause height and the melting level as described above, but they did not discuss how these changes contribute to the changes in the IWP. To evaluate the relative contributions of each factor, we construct a conceptual model, which consists of three columns, as described below. We do not intend to discuss every process that occurs, such as the life cycle of cloud systems or the effects of large-scale circulations (e.g., shear). Instead, we use a simple model to examine the relative importance of the aforementioned processes. A conceptual model of similar complexity was proposed by Del Genio et al. (2005), who showed that the IWC in a layered model becomes smaller as the circulation becomes weaker. However, they did not take into account the change in the vertical distributions of the IWC.

The conceptual model is schematically shown in Fig. 10. Column cumulus (C) is the region of deep convective cores, and column anvil (A) is the region of convective anvils. Column subsidence (S) is the surrounding environmentally clear region. The area fractions of C, A, and S are denoted as fc, fa, and 1− fafc, respectively. The planetary boundary layer is not considered explicitly, and the tropopause height H is specified. The circulation structure is specified by the upward mass flux in the deep core region (C) as

 
formula

for z > H/2, where the exponent n = 1 is used as the default and for zH/2 (i.e., no entrainment is considered in the lower troposphere). The subsidence speed is assumed to be the same for A and S. Thus, the subsidence mass flux in A is , and that in S it is . The SST is specified at Ts over the entire region, and the temperature profile is assumed to be moist adiabatic starting from Ts at the surface. The water vapor is saturated in the deep convective column C. Cloud microphysics processes are calculated using the same scheme used in NICAM. We first use the scheme developed by G98, which was used for the NICAM simulation in the previous section. Then, we use a more comprehensive scheme, namely, the NICAM single-momentum scheme with six categories of water (NSW6; Tomita 2008). Here, Ts = 300 K, Mc = 0.002 kg m−2 s−1, H = 15 km, fc = 0.001, and fa = 0.2 for the control (CTL case). We then examine the sensitivities to Ts, Mc, H, and fc. The global warming case (GW case) corresponds to the set of parameter values of Ts = 302 K, H = 17 km, Mc = 0.0015 kg m−2 s−1, and fc = 0.000 75. These values are chosen based on the results obtained through the NICAM simulations in the previous section.

Fig. 10.

Schematic diagram of the column model. Here, C, A, and S represent deep convective core, anvil, and surrounding environmentally clear regions, respectively, the respective fractional areas of which are denoted as fc, fa, and 1 – fafc. The tropopause height is denoted as H, and M(z) is the convective mass flux.

Fig. 10.

Schematic diagram of the column model. Here, C, A, and S represent deep convective core, anvil, and surrounding environmentally clear regions, respectively, the respective fractional areas of which are denoted as fc, fa, and 1 – fafc. The tropopause height is denoted as H, and M(z) is the convective mass flux.

The results of the CTL case are shown in Table 3, in which column-integrated water species in the C, A, and S regions are shown for both cloud microphysics schemes, G98 and NSW6. As shown in Table 3, the idealized model has a number of limitations. The ice water path is the sum of cloud ice and snow categories for G98 and is the sum of cloud ice, snow, and graupel for NSW6. The ice water path in the cumulus area is approximately 1 kg m−2 for G98, and it is a typical value of deep convective regions simulated by a global nonhydrostatic model (Iga et al. 2011). However, the IWP in the outer regions of the convective core is very small and unrealistic compared to the actual value for anvil clouds (e.g., IWP = 0.025 kg m−2, which corresponds to the radiatively active clouds). This presumably is due to the assumption of the steady circulation and the lack of consideration of the time evolution of cloud life cycle in the simple model. Notwithstanding this limitation, we regard the average value between the cumulus and anvil regions to be a representative value of the IWP for this simple model. If time evolution is considered, then the generation of the IWP in deep convection and its supply to stratiform clouds is crudely represented by this averaged value. Ice water paths of the CTL case listed in Table 3 are denoted as IWP0, and the sensitivities to the parameters are examined in the following.

Table 3.

Column-integrated water species in the three regions for the two cloud microphysics schemes, G98 and NSW6 (kg m−2).

Column-integrated water species in the three regions for the two cloud microphysics schemes, G98 and NSW6 (kg m−2).
Column-integrated water species in the three regions for the two cloud microphysics schemes, G98 and NSW6 (kg m−2).

Figure 11 shows the sensitivities to surface temperature Ts for G98, whereas the other parameters (H, Mc, and fc) are fixed. For the CTL case (denoted by diamonds), the IWP first increases until Ts = 302 K and then decreases as the temperature increases. The Clausius–Clapeyron effect dominates until Ts = 302 K, but the increase in the melting level becomes a dominating factor at Ts > 302 K. For the case of a higher tropopause height of H = 17 km, the IWP is generally larger than in the CTL case. Moreover, the IWP continues to increase for H = 17 km until Ts = 306 K, above which the IWP begins to decrease (data not shown). Figure 11 shows that the convective mass flux has a significant effect on the IWP. For a slow mass flux with Mc = 0.0015 kg m−2, the IWP is generally smaller than in the CTL case. For a smaller cumulus area (fc = 0.000 75), the vertical velocity becomes larger, resulting in a larger IWP. The decrease in the IWP for smaller vertical velocity is related to the decrease in the water supply, in addition to the increase of sedimentation of hydrometeors (Del Genio et al. 2005).

Fig. 11.

Dependency of IWP on Ts for the G98 cloud microphysics scheme. IWP is normalized by the IWP of the CTL-exp at Ts = 300 K (IWP0). The parameters of the CTL-exp are H = 15 km, Mc = 0.002 kg m−2 s−1, and fc = 0.001.

Fig. 11.

Dependency of IWP on Ts for the G98 cloud microphysics scheme. IWP is normalized by the IWP of the CTL-exp at Ts = 300 K (IWP0). The parameters of the CTL-exp are H = 15 km, Mc = 0.002 kg m−2 s−1, and fc = 0.001.

We now compare the IWP for the CTL case with the reference temperature Ts = 300 K (IWP0) and for the GW case (denoted by squares) of H = 17 km, Mc = 0.0015 kg m−2, and fc = 0.000 75. As shown by Fig. 11, the IWP for the GW case with Ts =302 K is smaller than that for the CTL case with Ts = 300 K, and IWP/IWP0 = 0.85.

Figure 12 shows the dependency of the IWP on surface temperature Ts for NSW6. The above-mentioned dependencies on the parameters are approximately the same, even if a different cloud microphysics scheme NSW6 is used. For the CTL case (denoted by diamonds), the effect of the increase in the melting level is always dominant over the Clausius–Clapeyron effect, so that the IWP decreases as the SST increases. For the total effect in the GW case (Ts = 302 K, H = 17 km, Mc = 0.0015 kg m−2, and fc = 0.000 75), the IWP is smaller than in the CTL case with Ts = 300 K, and IWP/IWP0 = 0.89.

Fig. 12.

As in Fig. 11, but for the NSW6 cloud microphysics scheme.

Fig. 12.

As in Fig. 11, but for the NSW6 cloud microphysics scheme.

Note that the above-mentioned result is the same as for the case in which the vertical profile of the convective mass flux is changed by n = 3 in (1) as well as for different cloud microphysics processes, such as an increase in the sedimentation of cloud ice in NSW6.

Even though a highly idealized simple model is used, we have shown that each of the processes considered herein might contribute to the change in the IWP as the SST increases. The total change in the IWP depends on the relative role of each process. The change in the convective mass flux has a significant impact on the decrease in the IWP. This decrease is moderated by the effect of the increase in the tropopause height, which has an impact on the increase in the IWP. The difference in cloud microphysics processes appears to be secondary.

Note, however, that the PBL scheme would have an impact on the IWP change. As for the sensitivities for the changes between the CTL and +2K experiments, as shown in Fig. 8, CS4MYNN does not exhibit a decrease in the convective mass flux. Based on the analysis using the simple column model, if the decrease in the mass flux is small, then the decrease in the IWP becomes small, and thus the increase in the upper-cloud cover becomes smaller. These results are consistent with Figs. 2 and 3, including the results for the CS4MYNN scheme.

5. Summary and discussion

The present study examines the responses of the upper-cloud cover and the ice water path (IWP) to surface temperature using a global nonhydrostatic model with explicit cloud processes (NICAM). Both the present perpetual July case (CTL-exp) and the global warming case, in which the sea surface temperature is increased by 2K (+2K-exp), are examined. Parameters related to the cloud microphysics scheme and the planetary boundary schemes are varied. The horizontal mesh intervals of approximately 7 and 14 km are also compared. This investigation follows the studies by Iga et al. (2007, 2011), in which the sensitivities under the present climate condition (CTL-exp of Table 1) are examined in detail. The numerical simulations show that, in general, the upper clouds prevail under the warmer condition, whereas the IWP decreases as temperature increases, especially over the tropics and subtropics, except for the case CS4MYNN, in which both the changes of the IWP and the upper-cloud fractions are small. Such a relation between the upper-cloud cover and the IWP is similar to the relation obtained under the present climate condition. Iga et al. (2011) showed that the upper-cloud cover becomes smaller as the IWP increases. The present study reveals that the upper clouds generally become more widespread, whereas the IWP generally becomes smaller in +2K-exp as compared to CTL-exp. Under the warmer condition, the relative humidity in the subtropics generally increases in the free troposphere and decreases in the upper troposphere. The increase in relative humidity is enhanced when the upper-cloud cover is more abundant.

The profiles of relative humidity, IWC, and cloud fraction as a function of temperature are almost unchanged between CTL-exp and +2K-exp except for 200–220 K. Then, our results, especially for CS4MYNN, are consistent with the FAT hypothesis proposed by Hartmann and Larson (2002), with an additional factor of the response of high thin clouds in 200–220 K (Fig. 6).

The changes in the intensity of the large-scale circulation and the changes in the convective circulation are also analyzed. The intensity of the large-scale overturning Hadley circulation is defined by the zonal mean omega velocity in the middle troposphere (500 hPa) in the subtropics, and the convective mass flux is defined by the threshold of vertical velocity of 1 m s−1 at an altitude of 5 km. It is demonstrated that both the Hadley circulation and the convective mass flux generally decrease in +2K-exp as compared to CTL-exp (again, except for CS4MYNN). These results are consistent with the results of previous studies (Held and Soden 2006; Vecchi and Soden 2007), in which the convective mass flux is diagnosed through a cumulus parameterization scheme used in a GCM. The explicit evaluation of the convective mass flux is achieved for the first time by the global nonhydrostatic model with explicit cloud processes. We also evaluated the fractional area of deep convective cores and found that the convective area decreases approximately proportionally with the decrease in the convective mass flux. The convective mass flux is estimated by a crude approximation as Mc = Q/Lq, where Q is the radiative cooling in the free atmosphere and Lq, which represents static stability, is the surface specific humidity multiplied by the latent heat. The convective mass flux is thought to become weaker because the static stability increases more than the radiative cooling under a warmer condition (Vecchi and Soden 2007). Although the expected tendency of the change in the convective mass flux relation is based on a crude approximation and could be affected by other factors, such as dissipation (Satoh and Hayashi 1992), the results obtained by the model considered herein confirmed the theoretical expectation and revealed that the response of the convective mass flux behaves in a manner similar to the large-scale overturning circulation. In particular, the change in the fractional area of the deep convective core is a major source of the change in the convective mass flux, rather than the change in speed of the vertical velocity.

In this paper, we point out the importance of the subgrid turbulent schemes on both the sensitivities under the present climate condition and the responses between the present climate condition and the global warming condition. We found that the subgrid turbulent scheme also affects the water vapor flux in the free troposphere. The relative contributions of explicit convection and the subgrid turbulence on the water vapor flux depend on the turbulent schemes and on the details of the implementation of the schemes (Noda et al. 2010). To clarify why these responses are generated, we must perform additional experiments focusing on the contribution of explicit convection and subgrid flux. The role of the subgrid turbulence will be an interesting topic for future studies.

To clarify the mechanism causing the decrease in the IWP with the increase in SST in the simulations, we constructed a simple conceptual model that consists of three columns of deep convective core, anvil, and surrounding environmental clear regions. In this model, the circulation speed was externally specified, and the effects of changes in tropopause height, convective mass flux, and fractional area of convection together with SST were examined. The results revealed that the decrease in convective mass flux has a significant impact on the decrease in the IWP under a warmer condition. These results were confirmed through the use of a different (or more comprehensive) cloud microphysics scheme (NSW6). The change in tropopause height has an effect on the increase in the IWP and moderates the effect of the convective mass flux.

Although the simulation results of the present study are not conclusive, the simple model used herein suggests the importance of the effect of the circulation change on the properties of upper clouds. Since the change in convective mass flux is approximately proportional to the change in large-scale overturning, it is reasonable that small-domain experiments without large-scale forcing, such as those conducted by Tompkins and Craig (1999) and Kuang and Hartmann (2007), will not reveal the response in the present study. In the future, the effect of the large-scale circulation on the change in upper clouds must be taken into account. The FAT mechanism leads to positive feedback of long wave radiation (Zelinka and Hartmann 2010). In our study, the extension of high thin clouds for +2L-exp might enhance positive feedback as described by CS. This mechanism will be further examined in a separate paper (Y. Tsushima et al. 2011, unpublished manuscript).

The results obtained using the simple model indicate a similar response of IWP under a warmer condition, even for very different cloud microphysics schemes. Recent NICAM simulations using NSW6 revealed that the IWP also decreases under a warmer condition, which is similar to the results obtained using G98 (Kodama et al. 2011, manuscript submitted to J. Geophys. Res.). However, these results do not imply that the response of the IWP is completely independent of the cloud microphysics schemes, and it is necessary to understand how cloud microphysics regulates upper clouds using cloud microphysics schemes.

The NICAM results indicate that the upper-cloud cover increases, whereas the IWP decreases as the SST increases (except for CS4MYNN). We speculate that, as the convective mass flux decreases, the IWP decreases, and the representative size of ice hydrometeors becomes smaller; this results in slower sedimentation of the ice phase and more detrainment of ice clouds from deep convection in upper layers (Satoh and Matsuda, 2009; Iga et al. 2011). Although the effect of the representative size of ice-phase particles on statistical values of the IWP remains unknown, an investigation of the effect of the treatment of ice-phase size in cloud schemes used in climate models on climate sensitivity may be useful.

Of course, we do not discount the importance of other factors, which might control the upper-cloud cover. The present results also indicate that the relative humidity in the free troposphere generally increases as the upper-cloud cover increases and the IWP decreases (Fig. 5). However, detailed analysis of the mechanism between the changes in upper clouds and relative humidity in the free troposphere is beyond the scope of the present study. A more humid condition is preferable for longer cirrus clouds’ lifetimes, so that the change in cirrus clouds might be viewed as a consequence of the change in the environmental condition. In contrast, more upper clouds would supply humidity in the upper troposphere and result in more humid conditions. To resolve these speculations, it is necessary to evaluate the relative roles of various processes, such as detrainment from deep convection, cloud microphysics of upper ice clouds, vertical mixing by sporadic convection, radiatively driven subsidence, and adiabatic mixing by extratropical eddies.

Clarification of the relation among the upper-cloud cover, IWP, and convective mass flux based on observations is not a straightforward task. Del Genio et al. (2005) showed the sensitivities of upper clouds and the IWP to SST using Tropical Rainfall Measuring Mission (TRMM) observations. Note, however, that the available observations are not used in a straightforward manner for the interpretation of the future change of the relation between the IWP and the circulation. For convective mass flux, observational data for radar sites (e.g., Fukao et al. 2003) as well as the data obtained by an indirect method (e.g., Machado and Laurent 2004) have recently become available. Moreover, a large scatter of IWP data still exists between different observational methods that use recent satellite data (Waliser et al. 2009). Despite such limitations, we must investigate a possible link among the upper-cloud cover, IWP, and convective mass flux from observational evidences. In addition to these efforts, it is necessary to use numerical models with cloud microphysics of various complexities to infer statistical relations in upper clouds. Simulated cloud properties must be evaluated using high-resolution and high-frequency observational data. Inoue et al. (2008, 2010) and Satoh et al. (2010) have taken one step toward such a goal for evaluations of simulated cloud properties by satellite remote sensing. Additional simulations and experience with cloud-resolving models involving cloud microphysics schemes are necessary to obtain more robust conclusions.

Acknowledgments

The authors thank the NICAM development members for their helpful discussions. The present study was supported by the Core Research for Evolutional Science and Technology (CREST) program of the Japan Science and Technology Agency (JST), and by the Innovation Program of Climate Projection for the 21st Century of the Ministry of Education, Culture, Sports, Science and Technology (MEXT). The Earth Simulator at the Japan Agency for Marine-Earth Science and Technology was used for the simulations.

REFERENCES

REFERENCES
Bony
,
S.
, and
J.-L.
Dufresne
,
2005
:
Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models
.
Geophys. Res. Lett.
,
32
,
L20806
,
doi:10.1029/2005GL023851
.
Cess
,
R. D.
, and
Coauthors
,
1990
:
Intercomparison and interpretation of climate feedback processes in 19 atmospheric general circulation models
.
J. Geophys. Res.
,
95
,
16 601
16 615
.
Clement
,
A. C.
, and
B.
Soden
,
2005
:
The sensitivity of the tropical-mean radiation budget
.
J. Climate
,
18
,
3189
3203
.
Cohen
,
B. G.
, and
G. C.
Craig
,
2006
:
Fluctuations in an equilibrium convective ensemble. Part II: Numerical experiments
.
J. Atmos. Sci.
,
63
,
2005
2015
.
Collins
,
W. D.
, and
M.
Satoh
,
2009
:
Simulating global clouds: Past, present, and future
.
Clouds in the Perturbed Climate System: Their Relationship to Energy Balance, Atmospheric Dynamics, and Precipitation, J. Heintzenberg and R. J. Charlson, Eds., Strüngmann Forum Reports, Vol. 2, The MIT Press, 469–486
.
Del Genio
,
A. D.
, and
W.
Kovari
,
2002
:
Climatic properties of tropical precipitating convection under varying environmental conditions
.
J. Climate
,
15
,
2597
2615
.
Del Genio
,
A. D.
,
W.
Kovari
, and
M.-S.
Yao
,
2005
:
Cumulus microphysics and climate sensitivity
.
J. Climate
,
18
,
2376
2387
.
Fukao
,
S.
, and
Coauthors
,
2003
:
Equatorial atmosphere radar (EAR): System description and first results
.
Radio Sci.
,
38
,
1053
,
doi:10.1029/2002RS002767
.
Grabowski
,
W. W.
,
1998
:
Toward cloud resolving modeling of large-scale tropical circulations: A simple cloud microphysics parameterization
.
J. Atmos. Sci.
,
55
,
3283
3298
.
Grabowski
,
W. W.
,
2000
:
Cloud microphysics and the tropical climate: Cloud-resolving model perspective
.
J. Climate
,
13
,
2306
2322
.
Grabowski
,
W. W.
,
2003
:
Impact of cloud microphysics on convective–radiative quasi equilibrium revealed by cloud-resolving convection parameterization
.
J. Climate
,
16
,
3463
3475
.
Hartmann
,
D. L.
, and
K.
Larson
,
2002
:
An important constraint on tropical cloud - climate feedback
.
Geophys. Res. Lett.
,
29
,
1951
,
doi:10.1029/2002GL015835
.
Held
,
I. M.
, and
B. J.
Soden
,
2006
:
Robust responses of the hydrological cycle to global warming
.
J. Climate
,
19
,
5686
5705
.
Iga
,
S.-I.
,
H.
Tomita
,
Y.
Tsushima
, and
M.
Satoh
,
2007
:
Climatology of a nonhydrostatic global model with explicit cloud processes
.
Geophys. Res. Lett.
,
34
,
L22814
,
doi:10.1029/2007GL031048
.
Iga
,
S.-I.
,
H.
Tomita
,
Y.
Tsushima
, and
M.
Satoh
,
2011
:
Sensitivity of Hadley circulation to physical parameters and resolution through changing upper-tropospheric ice clouds using a global cloud-system resolving model
.
J. Climate
,
24
,
2666
2679
.
Inoue
,
T.
,
M.
Satoh
,
H.
Miura
, and
B.
Mapes
,
2008
:
Characteristics of cloud size of deep convection simulated by a global cloud resolving model over the western tropical Pacific
.
J. Meteor. Soc. Japan
,
86A
,
1
15
.
Inoue
,
T.
,
M.
Satoh
,
Y.
Hagihara
,
H.
Miura
, and
J.
Schmetz
,
2010
:
Comparison of high-level clouds represented in a global cloud system–resolving model with CALIPSO/CloudSat and geostationary satellite observations
.
J. Geophys. Res.
,
115
,
D00H22
,
doi:10.1029/2009JD012371
.
Klein
,
S. A.
, and
C.
Jakob
,
1999
:
Validation and sensitivities of frontal clouds simulated by the ECMWF model
.
Mon. Wea. Rev.
,
127
,
2514
2531
.
Kuang
,
Z.
, and
D. L.
Hartmann
,
2007
:
Testing the fixed anvil temperature hypothesis in a cloud-resolving model
.
J. Climate
,
20
,
2051
2057
.
Lindzen
,
R. S.
,
M.-D.
Chou
, and
A. Y.
Hou
,
2001
:
Does the earth have an adaptive infrared iris?
Bull. Amer. Meteor. Soc.
,
82
,
417
432
.
Machado
,
L. A. T.
, and
H.
Laurent
,
2004
:
The convective system area expansion over Amazonia and its relationships with convective system life duration and high-level wind divergence
.
Mon. Wea. Rev.
,
132
,
714
725
.
Masunaga
,
H.
,
M.
Satoh
, and
H.
Miura
,
2008
:
A joint satellite and global cloud-resolving model analysis of a Madden-Julian oscillation event: Model diagnosis
.
J. Geophys. Res.
,
113
,
D17210
,
doi:10.1029/2008JD009986
.
Mellor
,
G.
, and
T.
Yamada
,
1974
:
A hierarchy of turbulence closure models for planetary boundary layer
.
J. Atmos. Sci.
,
31
,
1791
1806
.
Miura
,
H.
,
H.
Tomita
,
T.
Nasuno
,
S.-I.
Iga
,
M.
Satoh
, and
T.
Matsuno
,
2005
:
A climate sensitivity test using a global cloud resolving model under an aqua planet condition
.
Geophys. Res. Lett.
,
32
,
L19717
,
doi:10.1029/2005GL023672
.
Miura
,
H.
,
M.
Satoh
,
T.
Nasuno
,
A. T.
Noda
, and
K.
Oouchi
,
2007
:
A Madden-Julian oscillation event realistically simulated by a global cloud-resolving model
.
Science
,
318
,
1763
1765
.
Nakanishi
,
M.
, and
H.
Niino
,
2004
:
An improved Mellor-Yamada level-3 model with condensation physics: Its design and verification
.
Bound.-Layer Meteor.
,
112
,
1
31
.
Nakanishi
,
M.
, and
H.
Niino
,
2006
:
An improved Mellor-Yamada level-3 model: Its numerical stability and application to a regional prediction of advection fog
.
Bound.-Layer Meteor.
,
119
,
397
407
.
Neale
,
R. B.
, and
B. J.
Hoskins
,
2001
:
A standard test for AGCMs including their physical parameterizations. I: The proposal
.
Atmos. Sci. Lett.
,
1
,
101
107
,
doi:10.1006/asle.2000.0022
.
Noda
,
A. T.
,
K.
Oouchi
,
M.
Satoh
,
H.
Tomita
,
S.-I.
Iga
, and
Y.
Tsushima
,
2010
:
Importance of the subgrid-scale turbulent moist process: Cloud distribution in global cloud-resolving simulations
.
Atmos. Res.
,
96
,
208
217
,
doi:10.1016/j.atmosres.2009.05.007
.
Ramanathan
,
V.
, and
W.
Collins
,
1991
:
Thermodynamic regulation of ocean warming by cirrus clouds deduced from observations of the 1987 El Nino
.
Nature
,
351
,
27
32
.
Ringer
,
M. A.
, and
Coauthors
,
2006
:
Global mean cloud feedbacks in idealized climate change experiments
.
Geophys. Res. Lett.
,
33
,
L07718
,
doi:10.1029/2005GL025370
.
Rondanelli
,
R.
, and
R. S.
Lindzen
,
2008
:
Observed variations in convective precipitation fraction and stratiform area with sea surface temperature
.
J. Geophys. Res.
,
113
,
D16119
,
doi:10.1029/2008JD010064
.
Rossow
,
W. B.
, and
R. A.
Schiffer
,
1999
:
Advances in understanding clouds from ISCCP
.
Bull. Amer. Meteor. Soc.
,
80
,
2261
2287
.
Satoh
,
M.
, and
Y.-Y.
Hayashi
,
1992
:
Simple cumulus models in one-dimensional radiative convective equilibrium problems
.
J. Atmos. Sci.
,
49
,
1202
1220
.
Satoh
,
M.
, and
Y.
Matsuda
,
2009
:
Statistics on high-cloud areas and their sensitivities to cloud microphysics using single-cloud experiments
.
J. Atmos. Sci.
,
66
,
2659
2677
.
Satoh
,
M.
,
T.
Matsuno
,
H.
Tomita
,
H.
Miura
,
T.
Nasuno
, and
S.
Iga
,
2008
:
Nonhydrostatic Icosahedral Atmospheric Model (NICAM) for global cloud resolving simulations
.
J. Comput. Phys.
,
227
,
3486
3514
.
Satoh
,
M.
,
T.
Inoue
, and
H.
Miura
,
2010
:
Evaluations of cloud properties of global and local cloud system resolving models using CALIPSO and CloudSat simulators
.
J. Geophys. Res.
,
115
,
D00H14
,
doi:10.1029/2009JD012247
.
Sherwood
,
S. C.
,
W.
Ingram
,
Y.
Tsushima
,
M.
Satoh
,
M.
Roberts
,
P. L.
Vidale
, and
P. A.
O’Gorman
,
2010
:
Relative humidity changes in a warmer climate
.
J. Geophys. Res.
,
115
,
D09104
,
doi:10.1029/2009JD012585
.
Smith
,
R. K.
, Ed.,
1997
:
The Physics and Parameterization of Moist Atmospheric Convection
.
NATO Science Series C, Vol. 505, Kluwer Academic Publishers, 498 pp
.
Soden
,
B. J.
, and
I. M.
Held
,
2006
:
An assessment of climate feedbacks in coupled ocean–atmosphere models
.
J. Climate
,
19
,
3354
3360
.
Tomita
,
H.
,
2008
:
New microphysics with five and six categories with diagnostic generation of cloud ice
.
J. Meteor. Soc. Japan
,
86A
,
121
142
.
Tomita
,
H.
, and
M.
Satoh
,
2004
:
A new dynamical framework of nonhydrostatic global model using the icosahedral grid
.
Fluid Dyn. Res.
,
34
,
357
400
.
Tomita
,
H.
,
H.
Miura
,
S.
Iga
,
T.
Nasuno
, and
M.
Satoh
,
2005
:
A global cloud-resolving simulation: Preliminary results from an aqua planet experiment
.
Geophys. Res. Lett.
,
32
,
L08805
,
doi:10.1029/2005GL022459
.
Tompkins
,
A. M.
, and
G. C.
Craig
,
1998
:
Radiative-convective equilibrium in a three-dimensional cloud ensemble model
.
Quart. J. Roy. Meteor. Soc.
,
124
,
2073
2097
.
Tompkins
,
A. M.
, and
G. C.
Craig
,
1999
:
Sensitivity of tropical convection to sea surface temperature in the absence of large-scale flow
.
J. Climate
,
12
,
462
476
.
Vecchi
,
G. A.
, and
B. J.
Soden
,
2007
:
Global warming and the weakening of the tropical circulation
.
J. Climate
,
20
,
4316
4340
.
Waliser
,
D. E.
, and
Coauthors
,
2009
:
Cloud ice: A climate model challenge with signs and expectations of progress
.
J. Geophys. Res.
,
114
,
D00A21
,
doi:10.1029/2008JD010015
.
Wu
,
X.
,
W. D.
Hall
,
W. W.
Grabowski
,
M. W.
Moncrieff
,
W. D.
Collins
, and
J. T.
Kiehl
,
1999
:
Long-term behavior of cloud systems in TOGA COARE and their interactions with radiative and surface processes. Part II: Effects of ice microphysics on cloud–radiation interaction
.
J. Atmos. Sci.
,
56
,
3177
3195
.
Wyant
,
M. C.
,
M.
Khairoutdinov
, and
C. S.
Bretherton
,
2006
:
Climate sensitivity and cloud response of a GCM with a superparameterization
.
Geophys. Res. Lett.
,
33
,
L06714
,
doi:10.1029/2005GL025464
.
Zelinka
,
M. D.
, and
D. L.
Hartmann
,
2010
:
Why is longwave cloud feedback positive?
J. Geophys. Res.
,
115
,
D16117
,
doi:10.1029/2010JD013817
.