1. Introduction
Large-scale disturbances in the Tropics are often coupled with moist convection. A typical example is the Madden–Julian oscillation (MJO; Madden and Julian 1971, 1972) that is characterized by horizontal scales of 40 000 km and time scales of 40–60 days. The MJO has been investigated observationally and analytically since its discovery (Madden and Julian 1994, 2005; Yanai et al. 2000; Wang 2005; Zhang 2005).
Observational studies have revealed the multiscale structure of the MJO (Nakazawa 1986, 1988; Takayabu and Murakami 1991; Takayabu 1994a, b; Chen et al. 1996; Wheeler and Kiladis 1999; Houze et al. 2000; Straub and Kiladis 2002). Nakazawa (1988) analyzed Geostationary Meteorological Satellite (GMS) data and showed that super cloud clusters (SCCs) with scales of several thousand kilometers comprise the MJO event. These SCCs consist of cloud clusters (CCs) with scales of a few hundred kilometers. Cloud clusters are a common tropical cloud system (Leary and Houze 1979; Houze and Betts 1981; Takayabu and Murakami 1991; Mori 1992, 1995; Satoh et al. 1995; Chen et al. 1996; Houze et al. 2000). Studies using data from the Global Atmospheric Research Program’s (GARP) Atlantic Tropical Experiment (GATE) showed that a CC is comprised of cloud systems organized on the mesoscale with a horizontal scale of O(10 km) and a lifetime of several to ten hours (e.g., Zipser 1969, 1977; Houze 1977; Leary and Houze 1979; Zipser et al. 1981; Mori 1992, 1995; Satoh et al. 1995; Fujiyoshi and Geng 1995; Takahashi and Uyeda 1995; Chen et al. 1996; Houze et al. 2000). A variety of terms have been used to define various mesoscale cloud systems, such as mesoscale precipitating feature (MPF; Leary and Houze 1979), mesoscale convective system (MCS; Houze 1993, p. 334), rainband, squall line, and sometimes cloud cluster. This paper uses the term mesoscale cloud system (MC) to represent the basic category of mesoscale organized convection and includes all specific types. The MC is similar to mesoscale convection as defined by Yamasaki (1983) based on the explicit simulation of tropical cyclones. The MC includes cloud elements with a horizontal scale of around 1 km (e.g., Houze and Betts 1981; Mori 1992, 1995; Satoh et al. 1995; this scale is referred to as cloud scale for this paper). Such cloud elements correspond to individual echo cells in radar observations.
Houze et al. (2000) analyzed the MJO event that was observed during the Tropical Ocean Global Atmospheric Coupled Ocean–Atmosphere Response Experiment (TOGA-COARE) and stressed the importance of momentum transport in MCs (MCSs) in maintaining the large-scale dynamical structure. Yanai et al. (2000) argued that the key to the dynamics of the MJO is coupling between deep convection and large-scale wind perturbations, where the MC is the primary form of organized deep convection (Chen et al. 1996; Houze et al. 2000).
Conventional large-scale models (e.g., general circulation models; GCMs) do not usually explicitly represent mesoscale and cloud-scale processes. Rather, net effects are incorporated using an idealized formulation (i.e., parameterized). Slingo et al. (2005) reviewed GCM intercomparisons in this decade and concluded that cumulus parameterization is a major source of ambiguity in current GCMs, especially in reproducing the MJO. Numerical experiments that explicitly resolve clouds are one way to elucidate forcing mechanisms in multiscale convective systems and the interactions between different scales. Explicit cloud simulations using nonhydrostatic models over planetary-scale domains (Oouchi 1999; Oouchi and Yamasaki 2001; Grabowski and Moncrieff 2001, 2002; Tulich et al. 2007) have successfully reproduced the multiscale structure of large-scale convective disturbances and shed light on the mechanisms of convective organization within a range of idealized two-dimensional setups.
Grabowski (2001) proposed a new parameterization technique that extends the explicit approach. In this technique, cloud-resolving simulations operate in place of conventional cloud parameterization. GCMs that used this parameterization showed improved representation of tropical convection (Grabowski 2001, 2004; Khairoutdinov and Randall 2001; Khairoutdinov et al. 2005). Moncrieff (2004) and Biello and Majda (2005) developed new frameworks that explicitly treat the multiscale organization of convection with simple formulations. These methods have merit in their lower computational costs; thus long-time simulations, sensitivity experiments, and ensemble runs can be conducted.
A global cloud-resolving simulation is another promising step. Tomita et al. (2005, hereafter T05) and Satoh et al. (2005, hereafter S05) described an aquaplanet experiment using a nonhydrostatic model (NICAM; Tomita and Satoh 2004) with a horizontal mesh size down to 3.5 km. Experiment results included spontaneous organization of eastward-propagating large-scale cloud systems with zonal sizes and multiscale structures similar to observed SCCs. The vertical structure and propagation speed of the simulated SCCs (about 15 m s−1) were similar to observed convectively coupled Kelvin waves (Wheeler et al. 2000). New cloud clusters were successively generated ahead of the eastward migrating SCC. The new cloud clusters traveled westward and were incorporated into the SCC. Simulated CCs have sizes and lifetimes (about 2 days) similar to observed CCs. The present study seeks to understand the properties of SCCs shown by T05 and S05 from a multiscale perspective. Particular attention is given to mechanisms of MCs and CCs that comprised the SCCs, and to the effects of planetary-scale (e.g., Kelvin waves) and synoptic-scale disturbances on the organization of the convective systems. A separate paper (Nasuno et al. 2007, manuscript submitted to J. Atmos. Sci.) describes the coupling of convection and equatorial waves.
Results in this study are from a single experiment with a limited integration period. This experiment is the first simulation of planetary-scale disturbances using a model that marginally resolves clouds over the entire globe. Advantages of such a model over two-dimensional cloud-resolving models (Oouchi 1999; Oouchi and Yamasaki 2001; Grabowski and Moncrieff 2001, 2002; Tulich et al. 2007) include the full representation of the three-dimensional properties, such as meridional variability and cross-latitudinal interactions, of the multiscale convective organization. The aquaplanet setup is used as a basic test to understand the mechanisms of global convective behavior (Grabowski 2001; Hayashi and Sumi 1986; Hayashi and Golder 1997; Numaguti and Hayashi 1991; Flatau et al. 1997; Innes et al. 2001; Woolnough et al. 2001, 2004; Nakajima et al. 2004). As such, it is a vital first step to understanding real phenomena under highly complex conditions.
This paper is organized as follows. Section 2 briefly describes the numerical model and the experimental design. Section 3 describes the multiscale organization observed in the simulation results. Sections 4 and 5 contain a discussion and concluding remarks, respectively.
2. Numerical model and experimental design
Numerical experiments used the NICAM (Tomita and Satoh 2004). The model was designed for fine-mesh [O(1 km)] global simulations. It includes explicit moist processes in a nonhydrostatic framework.
The model equations guarantee conservation of total (air and water) mass and total (kinetic and internal) energy (Satoh 2002, 2003). This model formulation is suitable for long-term simulations such as the climate problem. Computations at present can be executed only on very efficient massively parallel computers, such as the Earth Simulator. The NICAM uses an icosahedral grid system, instead of a conventional grid system or a spectral domain, to boost computer performance. An advantage of icosahedral grids is quasi-uniformity.
Physical processes in the model are primarily the same as those in the Center for Climate System Research (CCSR)/National Institute for Environmental Studies (NIES)/Frontier Research Center for Global Change (FRCGC) atmospheric GCM (Numaguti et al. 1997; K-1 Model Developers 2004). A notable exception is the cloud physics. The NICAM does not include a cloud parameterization. Instead, cloud processes are calculated explicitly using a cloud microphysical scheme that includes prognostic variables for precipitating and nonprecipitating water condensates (Grabowski 1998). Water condensates are diagnostically partitioned into liquid and solid phases based on temperature.
The aquaplanet configuration in this study was the same as that in the control run by Neale and Hoskins (2000). Sea surface temperatures were fixed by a zonally uniform value with a meridional distribution defined by an idealized formula (Neale and Hoskins 2000). T05 describes the NICAM runs in detail. The model was run with horizontal grid spacings of 14, 7, and 3.5 km for 90, 40, and 10 model days, respectively. The same vertical grid was used in all experiments. The model layer depth increased from 35 m at the surface to 750 m in the upper troposphere. The top of the domain was near 40 km. Initial conditions for the 3.5-km (7-km) mesh were interpolated from 7-km (14-km) mesh results on the 20th (60th) day. A mesh size less than 1 km is desirable to represent cloud-scale processes accurately. However, the 3.5-km mesh allows a qualitatively adequate simulation of mesoscale processes as shown in the following sections.
3. Results
a. Large-scale structure
Figure 1 shows a snapshot of outgoing longwave radiation (OLR) at the top of the atmosphere in the 3.5-km mesh run. Convection was organized in clusters with O(1000 km) scale along the equatorial zone: around 130°–90°W, 50°–20°W (A), 20°–70°E (B), and 90°–110°E. Low OLR (<120 W m−2) was generally more frequent from 60°W–120°E than from 120°E–60°W, especially in the eastern part of the O(1000 km)-scale convective regions. The well-organized cloud systems A and B are considered to be SCCs for this paper. Off-equator regions of low OLR have roots in the active portion of the equatorial convection (e.g., around SCC B). This off-equator convection was associated with a Rossby wave response (as shown by Nasuno et al. 2007, manuscript submitted to J. Atmos. Sci.). A cloud system with a very high cloud top (OLR < 90 W m−2 over a diameter of 500 km) is considered a super convective system (SCS), following an analysis of GMS data (Chen et al. 1996). The SCS is characterized by a large area of deep cloud cover. Chen et al. (1996) required the continuous region of cold infrared (IR) temperatures (<208 K) to exceed 9 × 104 km2.
Figure 2 shows a vertical cross section of zonal velocity (colors) and total condensate (contours) corresponding to Fig. 1. Low-level easterlies and upper-level westerlies were enhanced from 100°E–140°W, to the east of the active hemisphere (60°W–120°E). A strong signal formed above SCC A in the lower stratosphere. Circulations with zonal extents around 10 000 km (low-level convergence and upper-level divergence) accompanied the O(1000 km)-scale convective regions (e.g., SCCs) and were superimposed on this 40 000-km scale structure. Circulations associated with the SCCs had tilted phase lines: westward with height in the troposphere and reversed above. Observations of convectively coupled Kelvin waves have shown this boomerang-shaped structure (Wheeler et al. 2000). The SCS was embedded in the region of strongest vertical shear of the zonal wind and was characterized by distinct local circulations including inflow/outflow branches in the midtroposphere.
Figure 3a shows the time–longitude section of 1°S–1°N averaged OLR from the 40-day run with the 7-km mesh. Super cloud clusters A and B were the most prominent convective signals during the entire period, and both propagated eastward at approximately 17 m s−1. Day 25 (Figs. 1 and 2) was a time of rapid weakening (growth) of SCC A (B) when the SCS was well established. Chen et al. (1996) showed that observed SCSs had lifetimes ranging from several hours to a few days without any specified propagation; the SCC was a special example of a SCS. Multiscale structural differences between the modeled SCCs and SCS were clear. The SCCs consisted of westward-propagating components (cloud clusters) of O(100 km). The SCS, in contrast, was a single eastward moving 500-km scale signal. The SCCs persisted for 20 days. The SCS lasted 2 days. The SCS was the most vigorous of all the simulated cloud systems, as characterized by the duration of exceptionally low OLR (<90 W m−2) that traveled eastward in the low-level easterlies. Strong vertical shear (Fig. 2) and deepening surface pressure (Fig. 3 b) preceding this special event were the keys to its occurrence, as discussed in section 3c.
Figures 3b and 3c show results from the 3.5-km run. Output from the 7-km run captured the general evolution of 3.5-km output, including the multiscale organization of SCCs. From days 24 to 27, the SCS and SCC B had sharper configurations in the 3.5-km mesh than in the 7-km mesh. Enhanced low-level easterlies to the east of the active hemisphere (Fig. 2) and the pressure anomaly with a horizontal scale of 40 000 km had Kelvin wave structures. SCCs A and B were also characterized by Kelvin wave structures with horizontal scales of 10 000 km. The wave structures1 of these two scales had the same phase near the SCS on day 25. The semidiurnal cycle that appeared in surface pressure is due to atmospheric tides (Woolnough et al. 2004).
The propagation speed and structure (Figs. 2 and 3) of the Kelvin-wave responses that accompanied the simulated SCCs resembled observed convectively coupled Kelvin waves (Nakazawa 1988; Takayabu and Murakami 1991; Takayabu et al. 1999; Wheeler and Kiladis 1999; Wheeler et al. 2000; Straub and Kiladis 2002, 2003c). Figure 4 shows the thermodynamic structure around SCC B (and the SCS in the right part of the panels). Observed characteristics that were reproduced include negative temperature anomalies in the lower troposphere of the deep convective region and positive moisture anomalies in the boundary layer to the east of the deep convective phase (Straub and Kiladis 2002, 2003c; Majda et al. 2004). Cool and dry anomalies near the surface suggest evaporation of precipitation. Surface cooling lagged the highest cloud tops in SCC B by at least by 5° (to the west). Surface cooling and highest cloud tops were nearly coherent in the SCS. This difference implies different roles for the cold pools in the eastward propagation of the SCC and the SCS. Total condensate and temperature data suggest that the cloud top in the SCS reached 18 km above sea level (ASL). The cloud top in SCC B decreased from 16 km ASL at the eastern end of SCC B to 12 km ASL at the western end. The OLR plot (Fig. 1) provides a three-dimensional view; the SCS included a round deep convective core with a diameter of 500 km. SCC B spread meridionally westward with lowered cloud tops.
The eastward increase in cloud top height (Fig. 4) suggests a zonal gradient in moist instability. In fact, the low-level moisture anomalies showed a large-scale increase to the east (Fig. 4c). Figure 5 shows a time–longitude section of moist static energy. The difference between values at z = 35 m and 12 km is plotted as a measure of instability. In the 40-day simulation, anomalously high (low) instabilities to the east (west) generally accompanied the deep convective phase of the SCCs (e.g., black contour lines showing OLR = 120 W m−2). The instability was especially enhanced near day 25 to the east of the SCS, reflecting the large-scale wave structures (i.e., 10 000- and 40 000-km-scale Kelvin wave responses). Large-scale waves, once established, selectively excited convection to the east through destabilization (Fig. 5) and large-scale circulations (Fig. 2).
b. Multiscale structure in SCC B
An important question is whether the large-scale dynamics are the major control to which convection only passively responds. The multiscale organization of convection in SCC B is considered in this section to answer that question.
Figure 6 displays ice condensates at 10 km ASL at 90-min intervals. Groups of clouds at 40°–48°E, 50°–57°E, 60°–64°E, and 65°–70°E are defined as CCa, b, c, d. The CCs were comprised of O(10 km)-scale cores with large concentrations of condensates surrounded by O(100 km)-scale regions with less condensates. The CCs persisted for 1.5–2 days and propagated westward at 9–10 m s−1 (Figs. 3c and 6). These modeled CCs thus have similar characteristics to observed CCs (Nakazawa 1988; Takayabu 1994b; Takayabu et al. 1996; Chen et al. 1996). The O(10 km)-scale convective regions are identified as mesoscale cloud systems. The MCs formed, expanded, merged, and disappeared within 3 h as shown in Fig. 6, during which time the CCs’ configurations changed only slightly. Figures 4 and 6 suggest that CCc was developing in a highly unstable environment, CCb was at the mature stage with cold pools and large amounts of condensates, and CCa was decaying in a stabilized environment. CCd was in the early stage of formation; only a few MCs were recognized at 0000 UTC day 25.
The distribution about the equator of active MCs in the CCs was asymmetric. They were aligned along 1°–2°S in CCb. The north-northwest (NNW)–south-southeast (SSE) orientation in CCa was pronounced by 0300 UTC. CCc and CCd were elongated south-southwest (SSW)–north-northeast (NNE) and NNW–SSE, respectively. The line of active MCs thus formed a meandering pattern. Figure 7 shows time series of temperature deviations at z = 2 km and meridional wind at z = 1.5 km. Deviations were calculated after removing the meridional gradient of the basic state by subtracting zonal mean values. The SCC-scale (from 40°–66°E) zonal gradients were then removed by subtracting the meridional average (over 4°N–4°S). Wavy patterns are evident at first glance in these variables. Regions of deep convection (e.g., contour lines showing OLR = 100 W m−2) appeared along the edges of negative temperature anomalies where meridional convergence was present in the lower troposphere. The wavy patterns had zonal wavelengths of approximately 1000 km and migrated westward, keeping pace with the CCs, suggesting that synoptic-scale wave disturbances influence the organization of CCs (Takayabu and Murakami 1991; Takayabu 1994b; Mori 1995).
Synoptic-scale waves are clearly present in the low-level meridional wind along the equator (Fig. 7). Thus, longitude–time sections of zonal and meridional velocities (z = 1.5 km) were examined (Fig. 8). Zonal velocity yields information on mechanisms controlling the westward propagation of CCs, which is the next topic considered. The meridional velocity (Fig. 8c) shows that westward-propagating signals with zonal wavelengths around 1000 km were abundant during 10 days of the simulation. These signals were especially prominent around SCC B from days 24 to 28 (including the period shown in Figs. 6 and 7). The signals came from the east, amplified in the SCC, and caused a few days of enhanced wind. The plot of meridional velocity for the 40-day simulation on the 7-km mesh (Fig. 9) shows several such amplifications of synoptic-scale signals around the SCCs.
Westward propagation of small scale (<1000 km) anomalies was widely observed in precipitation fields and in low-level meridional velocity fields (Fig. 8). Propagation speeds were generally consistent with zonal velocity, suggesting that advection by the easterlies was the primary forcing causing the movement. The westward migration of weak precipitation was particularly pronounced to the east of the SCCs, a region characterized by large OLR (Fig. 3), large moist instability (Fig. 4), and strong low-level easterlies (Fig. 8b). Many clouds with low tops that were generated in this unstable region were subsequently advected by easterlies. This convection deepened rapidly at the leading end of the SCCs. To the rear of the SCCs, where low-level easterlies were weak (<5 m s−1), precipitation did not move much. In this region, westward propagation was more distinct in OLR fields (Fig. 3). Figure 10 shows time–longitude plots of OLR and zonal wind at z = 12 km around CCa–d. The decaying and mature CCs (CCa,b) were accompanied by upper-level easterliesof 6–12 m s−1, consistent with advection of well-developed upper-level clouds by the divergent flow. Easterlies in the upper troposphere were much weaker around developing CCs (CCc,d), reflecting the SCC-scale structure (Figs. 4 and 10b); advection of upper-level clouds by easterlies does not account for the westward migration of these CCs.
c. Mesoscale cloud systems and the super convective system
This section describes the organization of MCs and the SCS. The SCS was the most vigorous cloud system during the entire simulation. These systems were allowed to form by the explicit treatment of clouds.
Figure 11 presents the 1°S–1°N averaged time–longitude sections of cloud condensates and temperature near the surface in CCa–d. The evolution of major MCs in the CCs can be traced. MCs in the developing CCs (CCc,d) showed high concentrations of ice condensates. Such high concentrations were preceded, by approximately 10 min, by the formation of liquid condensates. Cold pools were formed by precipitation (Fig. 11c), and persisted for a few hours. A new MC formed to the west of the preexisting one (i.e., at the western edge of the cold pools) in CCc,d. This successive formation is a major mechanism forcing the westward propagation of developing CCs. Mature and decaying CCs (CCa,b) were characterized by large amounts of liquid condensates that followed the formation of ice condensates. MCs often merged together and persisted for more than a few hours. CCs developed cold pools of O(100 km) through the aggregate effects of repeatedly formed MCs. Several MCs were aligned over the broad cold pool associated with CCb; some of the MCs originated from 1°–2°S (Figs. 6, 7), but MCs were more often generated along the edges (western in particular) of the cold pool than near the middle.
Figure 12 shows time series of CCc from 0–3 h around the equator (Fig. 11) at 20-min intervals. CCc is in the eastern half of the presented domain. The eastern end of CCb appears in the western part of the panels. Contour lines indicate condensate values exceeding 4 g kg−1. MCs were characterized by a maximum diameter of the enclosed region of around 100 km and were organized by successive formation of deep convective cells with horizontal scales of less than 10 km (close to the grid size in this model), or cloud-scale elements. Cloud-scale elements evolved within an hour and tended to initiate near preexisting elements. Individual MCs moved little during their lifetime of a few hours. However, the location of active MCs shifted westward by approximately 1° during the 3 h presented in Fig. 12, consistent with the time–longitude sections (Fig. 11). Selective development of an MC on the west side of preexisting MCs was evident at 60°–62°E on the equator. A group of MCs that developed between CCb and CCc eventually merged with CCc (Fig. 11).
Figure 13 shows the low-level structure for the last 40 min of Fig. 12. Cold pools were generated by the MCs (Fig. 13a, shading; Figs. 13b–d, thick contour lines). Low-level clouds repeatedly formed along the edges of the cold pools (Figs. 13b–d, shading), with frequent replacement of small (about grid size) peak regions. Low-level flow in this region was generally easterly (Figs. 13b–d, vectors). Moisture advection by the low-level easterlies partly accounts for the selective formation of MCs to the west.
Figure 14 shows low-level structures in typical MCs to highlight the organization of clouds in MCs. The isolated MCs that formed in the leading region of SCC B were sampled. Evaporation of precipitation generated MC-scale [O(10 km)] cold pools (Fig. 14c). The cold pools forced downward motion and divergent outflow; upward motions occurred along their peripheries (Fig. 14b). In regions of intense upward motion, clouds formed at 1 km ASL (Fig. 14a) by the same well-known mechanism that forces thunderstorms (Byers and Braham 1949) and mesoscale cloud systems (e.g., Zipser 1969). Tompkins (2001) described a radiative–convective equilibrium experiment with horizontal grid scales down to 200 m and highlighted the roles in cloud formation of moisture gain near the surface and increase of convective available potential energy. Moist static energy in the boundary layer (Fig. 14d) showed a qualitatively similar effect in the present study; there were negative anomalies within the cold pools and positive anomalies on the outer edges of the cold pools; these anomalies mostly reflected moisture distributions. The MC organization was primarily controlled in this way by local processes of O(<10 km) within the MC; in other words, self-organization dominated. Large-scale disturbances (e.g., Kelvin waves and synoptic-scale waves) affected the location, movement, and duration of MCs.
The simulated SCS was the most prominent example of an organized cloud system in this experiment, especially with respect to the role of cold pools and vertical wind profiles on the organization of convection. Figure 15 shows a time–longitude section (Fig. 15a) and a horizontal plot (Fig. 15b) of the cold pool generated by the SCS. The front edge of the cold pool, where cloud water was successively generated, is clearly defined. The arc-shaped configuration of the leading edge (Fig. 15b) is typical of a squall line. Modeled features that are part of the basic structure of a squall line include eastward propagation in an environment with strong westerly shear (Figs. 2 and 3), vertical structure with a convective leading edge, a trailing anvil cloud to the rear, and midlevel rear inflow (Fig. 4). Thus, the simulated SCS is regarded as a squall cluster (Houze and Betts 1981).
4. Discussion
This section compares the multiscale organization of simulated convection, shown in the previous section, with results from previous observational and modeling studies. The main focuses are on propagation mechanisms of the convective systems, the effects of mean zonal wind (vertical shear), and the impacts of the aquaplanet setup. Roles of the cold pools at different scales of convective organization are also discussed.
a. Eastward propagation of the SCCs
The eastward velocity of the simulated SCCs was approximately 17 m s−1 (Fig. 3). This propagation is comparable to those observed in convectively coupled Kelvin waves (11–22 m s−1 for an equivalent depth of 12–50 m; Wheeler and Kiladis 1999), although it is substantially faster than those typically observed in the warm pool (about 5 m s−1; Chen et al. 1996; Yanai et al. 2000). Kelvin waves are often coupled to SCCs (Nakazawa 1986, 1988; Takayabu and Murakami 1991), especially during El Niño events (Takayabu et al. 1999) and in the eastern Pacific (Straub and Kiladis 2002).
Nakazawa (1988) showed that eastward propagation of a SCC is preceded by successive formation of CCs 1000–2000 km east of the SCC. The formation mechanisms of the CCs are relevant to the propagation. Recent observational studies have argued that vertical moisture transport in shallow convection preconditions the atmosphere before the onset of the deep convective phase of the MJO (Kemball-Cook and Weare 2001; Kikuchi and Takayabu 2004). In fact, Bladé and Hartmann (1993) and Waliser et al. (1999) proposed that an increase in moist static energy and moisture in the lower troposphere prior to the onset of active convection was key to the onset of MJO. Kiladis et al. (2005) showed that such structures were commonly observed in a variety of convectively coupled equatorial waves, regardless of wavelength and propagation direction. Results from the present study support the above studies (Figs. 4c, 5 and 8). Moreover, in the present aquaplanet setup, where clouds can be abundant along the equator (Fig. 1), the large-scale structure more significantly affected the deepening of CCs (e.g., by removing the suppression due to subsidence and by subsequent upward motion) than midlevel moistening by the shallow convection. Oouchi (1999) used results from a two-dimensional cloud-resolving simulation to show that gravity waves propagating from an SCC can trigger CCs 1000–2000 km distant. The global simulation in this study contained many disturbances at various scales, including disturbances that originated in the midlatitudes. The generation of clouds rather quickly responded to large-scale conditions. Mechanisms described by Oouchi (1999) may have occurred in some cases, and further investigation is warranted.
b. Westward propagation of the CCs
The present study showed that the major mechanisms forcing westward propagation of CCs vary with the life stages of CCs, reflecting the different dynamic and thermodynamic conditions present during different phases of the SCC (section 3b). During the early formation and decaying stages of CCs, advection by easterlies (in the lower and upper troposphere, respectively) accounted for the propagation speed. In developing CCs, the formation and movement of MCs that comprised the CC controlled the movement of the CC as a whole. Mori (1995) examined the mesoscale organization of convection in a CC that was observed during TOGA-COARE. The CC was in an environment with low-level westerlies. New MCs (MPFs) formed west (i.e., upwind) of the CC and propagated eastward. Developing CCs simulated in the present study were in low-level easterlies with relatively weak vertical shear, and new MCs selectively formed west (i.e., downwind) of the CCs. The MCs were nearly stationary during their lifetime. Thus, properties of MCs vary with environmental flow and thermodynamic conditions (Mori 1995; Houze et al. 2000), although self-organization mechanisms through cold-pool formation (section 3c) are similar.
Results from this study also show that synoptic-scale disturbances exert a considerable influence on CC organization; there was good correspondence between CC configurations and westward-propagating waves with zonal wavelengths of approximately 1000 km (Figs. 7 and 8). Observational studies have shown a wide variety in the organization and propagation of tropical cloud clusters (Takayabu and Murakami 1991; Takayabu 1994b; Mori 1995; Chen et al. 1996; Takayabu et al. 1996; Straub and Kiladis 2002). Several analyses have related the convective signal corresponding to CCs to wave disturbances including easterly waves (Takayabu and Murakami 1991) and inertio-gravity waves (Takayabu 1994b; Takayabu et al. 1996; Chen et al. 1996). Mori (1995) showed that a wave disturbance with a horizontal scale of 1500–2000 km that propagated westward significantly affected convective activity in MPFs that constituted the CC. The phase velocity of the wave was consistent with an inertio-gravity wave (Takayabu 1994b).
c. Impacts of the aquaplanet setup
The present study shows propagation speeds of SCCs that are similar to those observed during an El Niño event (Takayabu et al. 1999) when the zonal SST contrast was small. The SST contrast was not considered in the present aquaplanet setup. It is possible that the slower propagation speeds (about 5 m s−1) of SCCs over the warm pool are controlled by factors including a zonal SST contrast, an interactive sea surface, and monsoonal circulations that were omitted from the present experimental design. The absolute value of the SST can also influence propagation speed. Oouchi (1999) and Grabowski and Moncrieff (2001) described explicit, two-dimensional simulations of planetary-scale tropical convective systems. SCC phase speeds were 3–6, and 6–8 m s−1 when uniform SST values were 302 and 303 K, respectively. Linear theoretical analyses (Wang 1988; Wang and Rui 1990) showed that growth rates for the unstable Kelvin mode are greatest for SSTs of 28.5°–29°C. Growth rates are substantially smaller for cooler SSTs. The present study included a relatively narrow region of SSTs exceeding 25°C (10°N–10°S); the maximum value was approximately 27°C on the equator. SSTs exceeding 28°C critically affect the organization and propagation of convective systems. In addition, Flatau et al. (1997) and Waliser et al. (1999) argued that interactive sea surfaces in a model forced a stronger signal and slower phase speed for the MJO.
A number of aquaplanet experiments (Hayashi and Sumi 1986; Numaguti and Hayashi 1991; Flatau et al. 1997, in a case without air–sea interaction; Woolnough et al. 2001, control case; Nakajima et al. 2004) have revealed a dominant planetary-scale Kelvin wave structure with a phase velocity of around 15 m s−1 that accompanied the convective signal. Such results are therefore not unique to the present study. The key question is why and to what extent are the simulated convective disturbances in the present case similar to previous GCM runs. How do the different treatments of cloud processes (parameterization in conventional GCMs and explicit calculation in NICAM) influence the results? Answering this important question is left for future study.
Low-level environmental flow in the present experiment was generally easterly and in geostrophic balance with the meridional pressure gradient. The pressure gradient reflects the meridional SST distribution. Low-level easterlies also dominated in some observations (Takayabu and Murakami 1991; Takayabu et al. 1999). The behaviors of SCCs in those cases were similar to modeled results in the present study, as mentioned earlier. Houze et al. (2000) discussed the effects of vertical wind shear on momentum transport by mesoscale organized convection (i.e., a super convective system; SCS). Convective momentum transport associated with the MJO and with the mesoscale organization of convection has recently been highlighted in observational studies (Houze et al. 2000; Tung and Yanai 2002) and in a multiscale analytical framework (Moncrieff 2004). In the present experiment, a well-organized squall-type cloud system (SCS) spontaneously emerged in a region of high instability where strong vertical shear was present in the horizontal wind associated with large-scale structure. Scale interaction through convective momentum transport is another topic to be examined in the future using this global, explicit, cloud-modeling approach.
d. The role of cold pools
There are differences in the low-level structures of the simulated SCS, SCCs, and CCs. The coolest surface temperature in the SCCs was under the western portions of the SCCs (Figs. 4b and 11) and did not cause eastward propagation by a squall-type mechanism, although cold anomalies coincided with precipitation (Fig. 11). The CCs were similar to the SCS in size, lifetime (Figs. 1, 3 and 4), and in the formation of new constituents on the front edge of the cold pool (e.g., CCc). However, the vertical profile of the zonal wind around the CCc (Fig. 2) was not favorable for vigorous squall-type organization because low-level easterlies had to pass through a precipitating region before reaching the cloud-formation region on the western end. In addition, CC organization was more complex than SCS organization. Figure 16 shows the low-level structure around CCa, b. Two scales of cold pools are present, a CC scale [O(100 km)] and an MC scale [O(10 km)]. The maxima of the CC-scale cold pools were over the northern parts of CCa,b, about 300 km distant from the region of maximum low-level convergence. Similarly, 300 km separated regions of active convection and major precipitation; these regions coincided with maximum low-level convergence and cold pools, respectively. Much of the domain below the CCs was dominated by northerly flow that may not have been only cold-pool outflow. Cold pools associated with each MC eroded after a few hours, and their effects did not extend beyond a few tens of kilometers. Synoptic-scale waves (Fig. 7) also affected the structure and organization of the CCs, as mentioned earlier.
5. Conclusions
This study investigated the multiscale organization of tropical convection in an aquaplanet simulation with horizontal mesh size down to 3.5 km. Numerical experiments used the nonhydrostatic icosahedral atmospheric model (NICAM; Tomita and Satoh 2004) with explicit cloud physics to eliminate uncertainties that arise from convective parameterizations.
Simulation results revealed spontaneous organization of large-scale (several thousand kilometers) cloud systems with realistic multiscale structure: eastward-propagating super cloud clusters (SCCs) that consist of westward-propagating cloud clusters (CCs) of O(100 km). Mesoscale cloud systems (MCs) of O(10 km) and cloud-scale elements (<10 km) were also simulated. Modeled MCs represents a basic category of mesoscale organized convection, including mesoscale convective systems (MCSs) or mesoscale precipitating features (MPFs) that are known to be constituents of observed CCs (Houze 1993; Leary and Houze 1979).
Kelvin wave responses with zonal wavelengths of 10 000 and 40 000 km appeared in the output dynamics. These waves were originally excited by organized convection in the spinup run. Once excited, they subsequently had significant effects on the large-scale organization of convection. The simulated 10 000-km-scale wave was tightly coupled to an SCC and resembled observed convectively coupled Kelvin waves (Wheeler et al. 2000; Straub and Kiladis 2002, 2003c; Takayabu et al. 1999) in phase velocity (17 m s−1) and vertical structure. The Kelvin-wave structure amplified the moist instability to the east of the SCC and thereby facilitated the generation of a new CC at the eastern end of the SCC. The convective system consequently showed eastward migration. The 40 000-km-scale wave kept the location of the most active SCC in phase with the low-level convergence throughout the simulation.
CCs evolved as they propagated westward within the SCC. Propagation mechanisms varied depending on the CC life stages. Vertical shear of the zonal wind associated with the large-scale circulation had significant influences on the westward propagation of CCs, as well as the thermodynamic structure that affected the convective activity. For example, in the shallow convective phase in the eastern part of the SCC, advection by enhanced low-level easterlies was evident. During the rapid deepening phase in the middle of the SCC, where vertical wind shear was weak, propagation arose through the selective formation of new MCs on the west side of CCs. In the mature CCs under the western part of the SCC, propagation was forced mostly by advection of well-developed upper-level clouds by easterlies associated with SCC-scale divergence.
The SCC was exclusively coupled to a Kelvin wave in this simulation. In contrast, the organization of the CC was affected to similar degrees by multiple processes. This is consistent with the wide variety of mechanisms shown to influence observed CCs (Takayabu and Murakami 1991; Takayabu 1994b; Takayabu et al. 1996; Mori 1995; Chen et al. 1996; Wheeler and Kiladis 1999). Synoptic-scale waves and cold pools were relevant factors in the present study.
Synoptic-scale wave disturbances that were characterized by westward propagation and asymmetry about the equator emerged intermittently in the model results, especially in the output of low-level meridional velocity. These waves generally had large amplitudes around SCCs, and in some instances coherently moved with the active convective regions in the CCs. Quantification of the mechanisms linking the wave and the CCs are beyond the scope of this study. However, such spontaneous occurrence of synoptic-scale disturbances showing meridional structure and cross-scale interaction was simulated for the first time in this experiment.
Cold anomalies near the surface coincided well with precipitation and coalesced into cold pools of O(100 km) below mature CCs. This suggests a contribution of evaporation of precipitation to the O(100 km) structure. However, temperature gradients associated with large-scale dynamics (i.e., wave disturbances) also affected the CCs, and the relative importance of cold pools warrants further investigation. For example, active convection in a CC was 300 km distant from the major cold anomalies, and cold-pool formation mechanisms were not simple.
Only some of the various processes affecting CC organization were examined in this study. Future studies will examine the effects of the diurnal cycle (Chen et al. 1996), easterly waves (Takayabu and Murakami 1991), inertio-gravity waves (Takayabu 1994b; Takayabu et al. 1996; Mori 1995; Oouchi 1999), equatorial wave disturbances (Wheeler and Kiladis 1999; Wheeler et al. 2000; Clayson et al. 2002; Straub and Kiladis 2002, 2003a; Yang et al. 2003; Roundy and Frank 2004), and cold pools and downdrafts due to precipitation.
Simulated MCs were maintained by systematic formation of cloud-scale elements along the edges of cold pools, in agreement with previous studies on the mesoscale organization of convection (Yamasaki 1975, 1983; Oouchi and Yamasaki 2001). The MCs were clearly defined in developing CCs and had a typical lifetime of 2–3 h. MCs in mature and decaying CCs often merged together or aligned in banded configurations and therefore persisted longer. Thus, large-scale conditions modified modeled MC properties just as they modify properties in observed MCSs and MPFs (Mori 1995; Satoh et al. 1995; Takahashi and Uyeda 1995; Fujiyoshi and Geng 1995; Chen et al. 1996; Houze et al. 2000). However, the primary organizing mechanism for individual MCs was not essentially altered by large-scale conditions. Rather, local processes were of primary importance. This organization mechanism contrasts with that of CCs.
The simulation also reproduced a squall cluster (Houze and Betts 1981) with an area of high cloud top (z > 16 km) of O(100 km) in a region of strong instability and pronounced wind shear at a time when the 10 000- and 40 000-km-scale waves overlapped. This was the most active convective system in the entire simulation and was categorized as a super convective system (SCS) as in Chen et al.’s (1996) analysis of satellite data. The modeled SCS formed a well-defined cold pool and arc-shaped leading edge. The convective organization was successfully simulated using the explicit approach in the model.
Critical issues in the numerical experiments remain that are related to computational capacity. The mesh sizes of 7- and 3.5-km cloud marginally resolved deep clouds but not shallow clouds in the lower troposphere. The shallow convection presented in this experiment is a very crude approximation. Simulations with higher resolution that improve the representation of boundary layer processes are pending. In addition, only output from a single 40-day run (for the 7-km mesh run) was available for the present study. The generality of the results can only be examined using output from longer simulations. In particular, the ability of the explicit global model to reproduce the MJO must be investigated, preferably with a realistic setup. The simplified setup used in this study should be considered when results are compared to observations. Observational studies have noted the interaction between tropical convection and extratropical disturbances (e.g., Straub and Kiladis 2003b) for which global explicit simulations are a promising approach. The present study should be considered a first step toward realistic, global, cloud-resolving climate simulations.
Acknowledgments
The authors acknowledge Drs. Yukari Takayabu, Tetsuo Nakazawa, and Taroh Matsuno for valuable discussions and comments. They also thank three anonymous reviewers and the assigned editor for suggestions to improve the manuscript. Simulations were conducted on the Earth Simulator. This research was supported by Core Research for Evolutional Science and Technology, Japan Science and Technology Agency.
REFERENCES
Biello, J. A., and A. J. Majda, 2005: A new multiscale model for the Madden–Julian oscillation. J. Atmos. Sci., 62 , 1694–1721.
Bladé, I., and D. L. Hartmann, 1993: Tropical intraseasonal oscillations in a simple nonlinear model. J. Atmos. Sci., 50 , 2922–2939.
Byers, H. R., and R. R. Braham, 1949: The thunderstorm project. U.S. Weather Bureau Rep., Washington, DC, 287 pp. [NTIS PB234515.].
Chen, S. S., R. A. Houze Jr., and B. E. Mapes, 1996: Multiscale variability of deep convection in relation to large-scale circulation in TOGA COARE. J. Atmos. Sci., 53 , 1380–1409.
Clayson, C. A., B. Strahl, and J. Schrage, 2002: 2–3-day convective variability in the tropical western Pacific. Mon. Wea. Rev., 130 , 529–548.
Flatau, M., P. J. Flatau, P. Phoebus, and P. P. Niiler, 1997: The feedback between equatorial convection and local radiative and evaporative processes: The implications for intraseasonal oscillations. J. Atmos. Sci., 54 , 2373–2386.
Fujiyoshi, Y., and B. Geng, 1995: Dual doppler radar observation of a tropical rainband developed from two convective clouds. J. Meteor. Soc. Japan, 73 , 471–490.
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., 2001: Coupling cloud processes with the large-scale dynamics using the Cloud-Resolving Convection Parameterization (CRCP). J. Atmos. Sci., 58 , 978–997.
Grabowski, W. W., 2004: An improved framework for superparameterization. J. Atmos. Sci., 61 , 1940–1952.
Grabowski, W. W., and M. Moncrieff, 2001: Large-scale organization of tropical convection in two-dimensional explicit numerical simulations. Quart. J. Roy. Meteor. Soc., 127 , 445–468.
Grabowski, W. W., and M. Moncrieff, 2002: Large-scale organization of tropical convection in two-dimensional explicit numerical simulations: Effect of interactive radiation. Quart. J. Roy. Meteor. Soc., 128 , 2349–2375.
Hayashi, Y., and D. G. Golder, 1997: United mechanisms for the generation of low- and high-frequency tropical waves. Part I: Control experiments with moist convective adjustment. J. Atmos. Sci., 54 , 1262–1276.
Hayashi, Y-Y., and A. Sumi, 1986: The 20–30-day oscillations simulated in an “aqua planet” model. J. Meteor. Soc. Japan, 64 , 451–467.
Houze Jr., R. A., 1977: Structure and dynamics of a tropical squall-line system. Mon. Wea. Rev., 105 , 1540–1567.
Houze Jr., R. A., 1993: Cloud Dynamics. Academic Press, 573 pp.
Houze Jr., R. A., and A. K. Betts, 1981: Convection in GATE. Rev. Geophys. Space Phys., 19 , 541–576.
Houze Jr., R. A., S. S. Chen, D. E. Kingsmill, Y. Serra, and S. E. Yuter, 2000: Convection over the Pacific warm pool in relation to the atmospheric Kelvin–Rossby wave. J. Atmos. Sci., 57 , 3058–3089.
Innes, P. M., J. M. Slingo, S. J. Woolnough, R. B. Neale, and V. D. Pope, 2001: Organization of tropical convection in a GCM with varying vertical resolution: Implications for the simulation of the Madden–Julian Oscillation. Climate Dyn., 17 , 777–793.
Kemball-Cook, S. R., and B. C. Weare, 2001: The onset of convection in the Madden–Julian oscillation. J. Climate, 14 , 780–793.
Khairoutdinov, M., and D. Randall, 2001: A cloud resolving model as a cloud parameterization in the NCAR Community Climate Model: Preliminary results. Geophys. Res. Lett., 28 , 3617–3620.
Khairoutdinov, M., D. Randall, and C. DeMott, 2005: Simulations of the atmospheric general circulation using a cloud-resolving model as a superparameterization of physical processes. J. Atmos. Sci., 62 , 2136–2154.
Kikuchi, K., and Y. N. Takayabu, 2004: The development of organized convection associated with the MJO during TOGA COARE IOP: Trimodal characteristics. Geophys. Res. Lett., 31 .L10101, doi:10.1029/2004GL019601.
Kiladis, G. N., K. H. Straub, and P. T. Haertel, 2005: Zonal and vertical structure of the Madden–Julian oscillation. J. Atmos. Sci., 62 , 2790–2809.
K-1 Model Developers, 2004: K-1 coupled GCM (MIROC) description. Tech. Rep. 1, Center for Climate System Research, University of Tokyo, Tokyo, Japan, 34 pp.
Leary, C. A., and R. A. Houze Jr., 1979: The structure and evolution of convection in a tropical cloud cluster. J. Atmos. Sci., 36 , 437–457.
Madden, R. A., and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702–708.
Madden, R. A., and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29 , 1109–1123.
Madden, R. A., and P. R. Julian, 1994: Observations of the 40–50-day tropical oscillation—A review. Mon. Wea. Rev., 122 , 814–837.
Madden, R. A., and P. R. Julian, 2005: Historical perspective. Intraseasonal Variability in the Atmosphere-Ocean Climate System, W. K.-M. Lau and D. E. Waliser, Eds., Springer-Praxis, 1–18.
Majda, A. J., B. Khouider, G. N. Kiladis, K. H. Straub, and M. G. Shefter, 2004: A model for convectively coupled tropical waves: Nonlinearity, rotation, and comparison with observations. J. Atmos. Sci., 61 , 2188–2205.
Moncrieff, M. W., 2004: Analytic representation of the large-scale organization of tropical convection. J. Atmos. Sci., 61 , 1521–1538.
Mori, K., 1992: Internal structure and time evolution of a cloud cluster in the western tropical Pacific region observed by Keifu Maru. J. Meteor. Soc. Japan, 70 , 1111–1123.
Mori, K., 1995: Equatorial convection observed by the research vessel Keifu Maru during the TOGA COARE IOP, November 1992. J. Meteor. Soc. Japan, 73 , 491–508.
Nakajima, K., E. Toyoda, M. Ishiwatari, S. Takehiro, and Y-Y. Hayashi, 2004: Initial development of tropical precipitation patterns in response to a local warm SST area: An aqua-planet ensemble study. J. Meteor. Soc. Japan, 82 , 1483–1504.
Nakazawa, T., 1986: Mean features of 30–60 day variations as inferred from 8-year OLR data. J. Meteor. Soc. Japan, 64 , 777–786.
Nakazawa, T., 1988: Tropical super clusters within intraseasonal variations over the western Pacific. J. Meteor. Soc. Japan, 66 , 823–839.
Neale, R. B., and B. J. Hoskins, 2000: A standard test for AGCMs including their physical parameterizations. I: The proposal. Atmos. Sci. Lett., 1 , 101–107.
Numaguti, A., and Y-Y. Hayashi, 1991: Behavior of cumulus activity and the structures of circulations in an “aqua planet” model. Part I: The structure of the super clusters. J. Meteor. Soc. Japan, 69 , 541–561.
Numaguti, A., M. Takahashi, T. Nakajima, and A. Sumi, 1997: Description of CCSR/NIES atmospheric general circulation model. CGER’s Supercomputer, Monogr. Rep., Vol. 3, National Institute for Environmental Studies, 1–48.
Oouchi, K., 1999: Hierarchical organization of super cloud cluster caused by WISHE, convectively induced gravity waves and cold pool. J. Meteor. Soc. Japan, 77 , 907–927.
Oouchi, K., and M. Yamasaki, 2001: An MJO-like gravity wave and superclusters simulated in a two-dimensional cumulus-scale-resolving model. J. Meteor. Soc. Japan, 79 , 201–218.
Roundy, P. E., and W. M. Frank, 2004: A climatology of waves in the equatorial region. J. Atmos. Sci., 61 , 2105–2132.
Satoh, M., 2002: Conservative scheme for the compressible nonhydrostatic models with the horizontally explicit and vertically implicit time integration scheme. Mon. Wea. Rev., 130 , 1227–1245.
Satoh, M., 2003: Conservative scheme for a compressible nonhydrostatic model with moist processes. Mon. Wea. Rev., 131 , 1033–1050.
Satoh, M., H. Tomita, H. Miura, S. Iga, and T. Nasuno, 2005: Development of a global cloud-resolving model—A multi-scale structure of tropical convections. J. Earth Sim., 3 , 11–19.
Satoh, S., A. Kinoshita, and H. Uyeda, 1995: Doppler radar observations on the structure and characteristics of tropical clouds during the TOGA-COARE IOP in Manus, Papua New Guinea: Dual-Doppler analysis of mesoscale convective systems composing a cloud cluster. J. Meteor. Soc. Japan, 73 , 443–459.
Slingo, J. M., P. M. Innes, and K. R. Sperber, 2005: Modeling. Intraseasonal Variability in the Atmosphere-Ocean Climate System, W. K.-M. Lau and D. E. Waliser, Eds., Springer-Praxis, 361–388.
Straub, K. H., and G. N. Kiladis, 2002: Observations of a convectively coupled Kelvin wave in the eastern Pacific ITCZ. J. Atmos. Sci., 59 , 30–53.
Straub, K. H., and G. N. Kiladis, 2003a: Interactions between the boreal summer intraseasonal oscillation and higher-frequency tropical wave activity. Mon. Wea. Rev., 131 , 945–960.
Straub, K. H., and G. N. Kiladis, 2003b: Extratropical forcing of convectively coupled Kelvin waves during Austral winter. J. Atmos. Sci., 60 , 526–543.
Straub, K. H., and G. N. Kiladis, 2003c: The observed structure of convectively coupled Kelvin waves: Comparison with simple models of coupled wave instability. J. Atmos. Sci., 60 , 1655–1668.
Takahashi, N., and H. Uyeda, 1995: Doppler radar observation of the structure and characteristics of tropical clouds during the TOGA-COARE IOP in Manus, Papua New Guinea—Three case studies on November 23 and December 16, 1992. J. Meteor. Soc. Japan, 73 , 427–442.
Takayabu, Y. N., 1994a: Large-scale cloud disturbances associated with equatorial waves. Part I: Spectral features of the cloud disturbances. J. Meteor. Soc. Japan, 72 , 433–449.
Takayabu, Y., 1994b: Large-scale cloud disturbances associated with equatorial waves. Part II: Westward-propagating inertio-gravity waves. J. Meteor. Soc. Japan, 72 , 451–465.
Takayabu, Y., and M. Murakami, 1991: The structure of super cloud clusters observed in 1–20 June 1986 and their relationship to easterly waves. J. Meteor. Soc. Japan, 69 , 105–125.
Takayabu, Y. N., K-M. Lau, and C-H. Sui, 1996: Observation of a quasi-2-day wave during TOGA COARE. Mon. Wea. Rev., 124 , 1892–1913.
Takayabu, Y. N., T. Iguchi, M. Kachi, A. Shibata, and H. Kanazawa, 1999: Abrupt termination of the 1997–98 El Ninõ in response to a Madden–Julian oscillation. Nature, 402 , 279–282.
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., 2001: Organization of tropical convection in low vertical wind shears: The role of water vapor. J. Atmos. Sci., 58 , 529–545.
Tulich, S. N., D. Randall, and B. Mapes, 2007: Vertical-mode and cloud decomposition of large-scale convectively coupled gravity waves in a two-dimensional cloud-resolving model. J. Atmos. Sci., 64 , 1210–1229.
Tung, W-W., and M. Yanai, 2002: Convective momentum transport observed during the TOGA COARE IOP. Part I: General features. J. Atmos. Sci., 59 , 1857–1871.
Waliser, D. E., K. M. Lau, and J-H. Kim, 1999: The influence of coupled sea surface temperatures on the Madden–Julian oscillation: A model perturbation experiment. J. Atmos. Sci., 56 , 333–358.
Wang, B., 1988: Dynamics of tropical low-frequency waves: An analysis of the moist Kelvin wave. J. Atmos. Sci., 45 , 2051–2065.
Wang, B., 2005: Theory. Intraseasonal Variability in the Atmosphere-Ocean Climate System, W. K.-M. Lau and D. E. Waliser, Eds., Springer-Praxis, 307–360.
Wang, B., and H. Rui, 1990: Dynamics of the coupled moist Kelvin–Rossby wave on an equatorial β-plane. J. Atmos. Sci., 47 , 397–413.
Wheeler, M., and G. N. Kiladis, 1999: Convectively coupled equatorial waves: Analysis of clouds and temperature in the wavenumber–frequency domain. J. Atmos. Sci., 56 , 374–399.
Wheeler, M., G. N. Kiladis, and P. J. Webster, 2000: Large-scale dynamical fields associated with convectively coupled equatorial waves. J. Atmos. Sci., 57 , 613–640.
Woolnough, S. J., J. M. Slingo, and B. J. Hoskins, 2001: The organization of tropical convection by intraseasonal sea surface temperature anomalies. Quart. J. Roy. Meteor. Soc., 127 , 887–907.
Woolnough, S. J., J. M. Slingo, and B. J. Hoskins, 2004: The diurnal cycle of convection and atmospheric tides in an aquaplanet GCM. J. Atmos. Sci., 61 , 2559–2573.
Yamasaki, M., 1975: A numerical experiment of the interaction between cumulus convection and larger-scale motion. Pap. Meteor. Geophys., 26 , 63–91.
Yamasaki, M., 1983: A further study of the tropical cyclone without parameterizing the effects of cumulus convection. Pap. Meteor. Geophys., 34 , 221–260.
Yanai, M., B. Chen, and W-W. Tung, 2000: The Madden–Julian oscillation observed during the TOGA COARE IOP: Global view. J. Atmos. Sci., 57 , 2374–2396.
Yang, G-Y., B. Hoskins, and J. Slingo, 2003: Convectively coupled equatorial waves: A new methodology for identifying wave structures in observational data. J. Atmos. Sci., 60 , 1637–1654.
Zhang, C., 2005: Madden-Julian oscillation. Rev. Geophys., 43 .RG2003, doi:10.1029/2004RG000158.
Zipser, E. J., 1969: The role of organized unsaturated convective downdrafts in the structure and rapid decay of an equatorial disturbance. J. Appl. Meteor., 8 , 799–814.
Zipser, E. J., 1977: Mesoscale and convective-scale downdrafts as distinct components of squall-line structure. Mon. Wea. Rev., 105 , 1568–1589.
Zipser, E. J., R. J. Meitin, and M. A. LeMone, 1981: Mesoscale motion fields associated with a slowly moving GATE convective band. J. Atmos. Sci., 38 , 1725–1750.
Horizontal plots of OLR at the top of the atmosphere (W m−2) at 0000 UTC, day 25 in the 3.5-km mesh run.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Vertical cross sections (1°N–1°S average) of zonal velocity (m s−1) at 0000 UTC, day 25. Contour lines show total condensate = 0.5 g kg−1.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Hovmöller diagram of (a) OLR (W m−2) in the 7-km run, (b) surface pressure (hPa), and (c) OLR in the 3.5-km mesh run, averaged between 1°N and 1°S. Black lines indicate eastward velocity of 17 m s−1.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Vertical cross sections (1°N–1°S average) of SCC B and SCS at 0000 UTC, day 25: (a) zonal velocity (m s−1), (b) temperature (K), and (c) water vapor content (g kg−1). Deviations from zonal mean values are drawn. Contour lines show total condensate = 0.5 g kg−1.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
As in Fig. 3a, but for the moist static energy difference between altitudes of 35 m and 12 km. The unit is 105 J. Contour lines show OLR = 120 W m−2.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
CCa–d at 90-min intervals. The sum of cloud ice and snow content (g kg−1) at z = 10 km is shown.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
(left) Temperature deviations (K) at z = 2 m and (right) meridional velocity (m s−1) atz = 1.5 km with wind vectors at 6-h intervals. Thick contour lines show OLR = 100 W m−2.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Hovmöller diagram of (a) surface precipitation rate (mm h−1), (b) zonal, and (c) meridional velocity (m s−1) at z = 1.5 km in the 3.5-km mesh run, averaged between 1°N and 1°S. Contour lines show OLR = 120 W m−2.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
As in Fig. 3a, but for meridional velocity at z = 1.5 km. Contour lines show OLR = 120 W m−2.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Hovmöller diagram of CCa–d (1°N–1°S average): (a) OLR (W m−2) and (b) zonal velocity at z = 12 km. The ordinate indicates the time (hour) at 0000 UTC, day 25. Contour lines show OLR = 120 W m−2.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
As in Fig. 10, but for (a) ice water path (kg m−2), (b) liquid water path (kg m−2), and (c) temperature at z = 2 m (K) for a subdomain. Contour lines in (a) and (b) show OLR = 120 W m−2, and in (c) surface precipitation rate = 5 mm h−1. In (c) deviation from 300 K is drawn.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Sum of cloud ice and snow content (g kg−1) at z = 10 km in CCc at 20-min intervals. Contour lines are drawn where values exceed 4 g kg−1 (at 1 g kg−1 intervals). Wind vectors at this level are also plotted.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
As in Fig. 12, but for (a) potential temperature at z = 35 m (K) and snow and cloud water content at z = 2 km at (b) 140, (c) 160, and (d) 180 min. Contour lines in (a) indicate the sum of cloud ice and snow content (z = 10 km) of 2 and 4 g kg−1. (b)–(d) Thick contour lines show temperature of 297.5 K at z = 35 m. In (a) deviation from 300 K is drawn.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
Horizontal cross sections of MCs near the boundary layer at 0000 UTC, day 25. (a) Cloud water content at z = 1 km (g kg−1), (b) vertical velocity at z = 35 m (10−2 m s−1) with horizontal wind vectors at this level, (c) potential temperature at z = 35 m (K), and (d) moist static energy at z = 500 m (105 J kg−1). In (c) deviation from 300 K is drawn.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
(a) Hovmöller diagram (1°N–1°S average) and (b) horizontal plots of SCS. In (a) shading shows the liquid water path (kg m−2), and thin, thick solid, and thick broken lines show temperatures (z = 35 m) of 299, 298, and 297 K, respectively. In (b) shading shows potential temperature (z = 35 m) at 0000 UTC, day 25, and contour lines indicate cloud water content (z = 1 km) of 0.1 g kg−1. In (a) deviation from 300 K is drawn.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
As in Fig. 14, but for CCa, b.
Citation: Journal of the Atmospheric Sciences 64, 6; 10.1175/JAS3948.1
