1. Introduction
Tropical cyclones distinguish themselves from other convective systems by circularly organized vigorous deep convection. Tropical cyclone (TC) formation, to some extent, can be regarded as a process in which deep convection becomes stronger and better organized. Latent heat release of organized convection near the storm center drives the transverse circulation, and the associated low-level convergence leads to intensification of the system-scale circulation via the vortex-stretching effect. Convective evolution and organization is thus a critical issue for better understanding tropical cyclone formation.
There are two distinct types of precipitation processes in the tropics: stratiform and convective processes (e.g., Houze 1997). The former are characterized by midlevel convergence and divergence in the upper and lower troposphere, and the latter are characterized by low-level convergence and divergence at the cloud top. Tory et al. (2006) suggested that tropical cyclone formation may be involved with the atmospheric transition from a mean stratiform profile to a mean convective profile as the low-level divergence associated with stratiform processes tends to spin down the low-level circulation. In a numerical model simulation of pre–Hurricane Felix (2007), Wang et al. (2010b) showed that both stratiform and convective precipitation rates increase with time, and that the mean convergence profile becomes dominantly convective near the wave pouch1 center because of the relatively large increase in convective precipitation. A TC forms near the pouch center via the system-scale low-level convergence and vorticity aggregation (Wang et al. 2010a; Montgomery et al. 2010; Fang and Zhang 2010). Wang (2012) further showed that stratiform precipitation induces modest midlevel inflow but very weak low-level outflow due to the weak evaporative cooling in the moist lower troposphere. It thus contributes to the midlevel spinup without substantially spinning down the low-level circulation. Stratiform processes were also emphasized by Rappin and Nolan (2012) for their role in moistening the lower troposphere via the showerhead mechanism.
Using Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) field data, Johnson et al. (1999) highlighted the trimodal distribution of tropical convection: shallow cumulus clouds with radar echo tops between 0 and 4 km, midlevel congestus clouds with radar echo tops between 5 and 9 km, and deep cumulonimbus clouds with radar echo tops between 11 and 16 km (e.g., Riehl and Malkus 1958). The important role of deep convection in TC formation has been emphasized in many previous studies (e.g., Tory and Frank 2010 and references therein). The vortical hot tower (VHT) theory (Montgomery et al. 2006; Hendricks et al. 2004) suggests that VHTs or rotating deep cumulonimbus clouds are the preferred convective structure during the incipient formation stage, and that the merger, segregation, and axisymmetrization of VHTs are the fundamental processes for TC formation. It was also suggested that VHTs collectively act as a quasi-steady source of diabatic heating to drive the transverse circulation and intensify the system-scale circulation. Deep convection was also suggested to play a critical role in the two-stage conceptual model proposed by Zehr (1992). In the first stage of the conceptual model, convective bursts lead to the formation of a mesoscale vortex. In the second stage, a deep convective burst developing near the low-level circulation center leads to the intensification of the low-level circulation and the formation of a tropical cyclone.
Besides deep convection, congestus clouds are another primary type of clouds in the tropics. While deep convection may be suppressed by midlevel dry air, congestus clouds are often abundant in regions of large-scale subsidence (e.g., Takayabu et al. 2010). It has been estimated that congestus clouds constitute more than 50% of the precipitating convective clouds in the TOGA COARE experiment over the tropical western Pacific (Johnson et al. 1999) and make a substantial contribution (~20%–40%) to the tropical precipitation (Houze and Cheng 1977; Petty 1999; Lau and Wu 2003; Stephens et al. 2002). Congestus clouds are characterized by shallow heating in the lower troposphere (Schumacher et al. 2008; Takayabu et al. 2010) and play an important role in the tropical large-scale circulation and energy budget (Mapes 2000; Wu 2003; Khouider and Majda 2008). Moisture preconditioning by congestus clouds for deep convection has been emphasized in many previous studies (e.g., Brown and Zhang 1997; Johnson et al. 1999; Takemi et al. 2004; Khouider and Majda 2008; Waite and Khouider 2010). In particular, it is regarded as a critical process for the evolution of the Madden–Julian oscillation (MJO) and convectively coupled equatorial waves (CCEWs) (e.g., Kiladis et al. 2009; Khouider and Majda 2006). Motion systems of different spatiotemporal scales, such as the MJO, CCEWs, and mesoscale convective systems, display a large degree of self-similarity in cloud morphology, in which deep convection is preceded by shallow convection and followed by stratiform anvils (Mapes et al. 2006). However, the role of congestus clouds in the life cycle of tropical cyclones is largely overlooked.
In a companion paper, Wang (2014) examines the statistics of convective processes and vertical vorticity in a high-resolution simulation of Tropical Cyclone Fay (2008) from the tropical wave stage to the TC stage. It is shown that the intensity of vertical motion in the model simulations approximately follows a truncated lognormal distribution and that updrafts at the pregenesis stage are weaker than those in a mature tropical storm or in a typical midlatitude thunderstorm. For example, only about 5% of the upward velocities at 6 km exceed 1 m s−1 prior to genesis. On the other hand, the top 5% of updrafts contribute 40%–50% to the total upward mass flux2 and 50%–60% to the total upward moisture flux and are associated with about 70% of the total condensation. The analysis of the numerical model simulations showed that updrafts and downdrafts both intensify approaching genesis, but the former are much stronger than the latter. The mean vertical motion and vertical mass flux thus are upward and increase with time.
In this study, we will examine the roles of different types of convection in tropical cyclone formation, or more specifically, whether deep convective clouds make a major contribution to TC formation or different types of clouds collectively contribute to the formation and intensification of a tropical cyclone protovortex. We will address this question through the analyses of a cloud-resolving numerical model simulation. A brief description of the model simulation is given in section 2. The trimodal distribution of convective updrafts in the model simulation and the evolution of a congestus cloud cell are examined in section 3. The roles of cumulus congestus and deep convection are analyzed in section 4. Convective organization and transition to sustained deep convection near the wave pouch center are investigated in section 5, followed by a summary in section 6 and discussion in section 7. A conceptual model for TC formation is proposed in section 7.
2. Model description
A high-resolution model simulation of Tropical Cyclone Fay (2008) was analyzed in this study. The storm evolution and the numerical model configuration are described in detail in Fritz and Wang (2013) and Wang (2014). Only a brief description is provided below. The storm was simulated using the Weather Research and Forecasting model (WRF), version 3.2.1 (Skamarock et al. 2008). To resolve convection explicitly near the pouch center, a four-grid nested configuration with 27-, 9-, 3-, and 1-km resolutions was adopted, and the inner two grids moved automatically with the wave pouch center using a vortex-tracking algorithm described in Wang (2014). The Kain–Fritsch cumulus scheme (Kain 2004) was applied to the outermost grid, and cumulus convection was resolved explicitly at the grid scale at the three inner grids (9-, 3-, and 1-km resolutions). Other physics parameterizations include the WRF single-moment, six-class microphysics scheme (WSM6; Hong and Lim 2006), the Yonsei University (YSU) planetary boundary layer scheme (Hong et al. 2006), the Rapid Radiative Transfer Model (RRTM) longwave radiation scheme, and Dudhia’s (1989) shortwave radiation scheme. The model was initialized at 0000 UTC 13 August 2008 using Interim European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-Interim) 6-hourly data. A tropical depression formed at 0600 UTC 15 August (t = 54 h), and a tropical storm formed at 1200 UTC 15 August (t = 60 h) in the model simulation. The former is referred to as the genesis time.
The output from the innermost grid with 1-km resolution is analyzed in the following sections. The sensitivity tests in Wang (2014) showed that the simulation with 1-km resolution resolves the statistics of vertical motion reasonably well compared to a simulation with 250-m resolution. Our analysis will focus on the area within a 100-km radius of the pouch center if not specified otherwise. Wang (2012) showed that the thermodynamic and dynamic conditions in this inner pouch region are particularly favorable for deep convection and tropical cyclogenesis.
3. Trimodal distribution of convective updrafts and evolution of a congestus cloud cell
Johnson et al. (1999) identified the trimodal distribution of tropical convective clouds using radar data. However, it is not clear whether the same distribution also exists in tropical cyclones and their precursor disturbances and whether a numerical model can capture such a distribution. Instead of using the derived reflectivity field, we will directly examine the vertical velocity field. Figure 1 shows the frequency distribution of the updraft-top height (UTH) for all the convective grid columns within a 100-km radius of the pouch center. A grid column is designated as convective if (i) the layer-mean vertical velocity between 1 and 5 km is greater than zero and (ii) the maximum vertical velocity between 1 and 14 km exceeds 1 m s−1. The UTH of a convective column is defined as the maximum height of vertical velocity equal to or greater than 0.5 m s−1. To exclude detached, weak upward motion in the upper troposphere associated with stratiform anvils or gravity waves, it is also required that w is positive between the UTH and 3 km if the UTH is above 3 km. As shown in Fig. 1, a clear trimodal distribution of UTHs exists in the inner pouch region both prior to and after genesis. The distribution has one peak at 1.5–2.0 km, one at 5 km, and a third one at 13–14 km, corresponding to the shallow cumulus clouds, midlevel congestus clouds, and deep cumulonimbus clouds (Johnson et al. 1999). Comparison between the pregenesis and postgenesis periods shows that the three types of convection all occur more frequently with time, but deep convection increases most significantly. The histograms of UTHs suggest that shallow cumulus and congestus convection exist in tropical cyclones and their precursor disturbances just as in ordinary tropical convective systems. In the rest of the paper, shallow cumulus and congestus convection together will be referred to as cumulus congestus for brevity.
Distribution of the updraft-top height for convective grid columns for a pregenesis time period of 36–48 h (black bars) and a postgenesis time period of 60–72 h (gray bars) within the innermost model domain.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
Before examining the contributions of deep convection and cumulus congestus to tropical cyclone formation, it is informative to first look at the life cycle of a congestus cloud cell in the model simulation. Figure 2 shows the evolution of a congestus cloud cell located less than 100 km west-southwest of the pouch center. During the 40-min sequence shown in Fig. 2, the cloud cell moves northeastward with respect to the pouch center, and the domain of plot in each panel follows the cloud cell.
Evolution of relative vorticity and vertical velocity from (a) 0400 to (f) 0440 UTC 15 Aug. At each time, (top) a 20 km × 16 km horizontal domain, which is approximately centered at the cloud cell of interest, with shading representing the relative vorticity (10−3 s−1) at 1 km and contours indicating the maximum vertical velocity between 1 and 6 km (contours start at 1 m s−1 with an interval of 2 m s−1); and (bottom) the north–south cross section of relative vorticity (shading) and vertical velocity along the longitude of the maximum updraft of the cloud cell of interest (indicated by the dashed vertical line in the corresponding top panel). The purple curve is the 17-dBZ contour of reflectivity, and upward (downward) motion is represented by solid (dashed) contours with an interval of 1.0 (0.5) m s−1. Note that the domain moves with the convective cell of interest.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
At 0400 UTC 15 August (2 h prior to genesis), the convective cell (the one near the center of the domain) starts to grow in the lower troposphere. At this time, w > 1 m s−1 is confined below 1 km (the bottom panel in Fig. 2a), and the shallow updraft is associated with a northwest–southeast-orientated vorticity dipole pattern (the top panel in Fig. 2a). The magnitude of the cyclonic and anticyclonic vorticity anomalies is about 0.5 × 10−3 s−1. The convective cell grows rapidly in the next 16 min. The maximum vertical velocity exceeds 9 m s−1, and the contour 1 m s−1 extends beyond 5 km (Fig. 2c). Meanwhile, the vorticity dipole has also significantly intensified. The cyclonic relative vorticity is amplified by about 8 times and exceeds 4 × 10−3 s−1, and the anticyclonic vorticity exceeds 2.5 × 10−3 s−1. The updraft is still located between the cyclonic and anticyclonic vorticity centers, suggesting that the tilting term plays a dominant role in amplifying vorticity at this stage. It is also interesting to note that the updraft is surrounded by an annulus of downdraft of 1.0–1.5 m s−1 between 2 and 4 km. The downdraft is likely dynamically driven by the rapidly growing updraft.
From 0416 to 0424 UTC (Figs. 2c,d), the convective updraft continues growing upward. The cyclonic vorticity remains strong and is now located directly below the updraft, suggesting that the stretching term has become the dominant player in amplifying vorticity. Also note that the updraft has a slight southward vertical tilt, and downward motion has started developing near the surface below the updraft core, which can be attributed to evaporative cooling and water loading. This suggests that the convective cell has reached its mature stage.
At 0432 UTC (Fig. 2e), the maximum updraft remains more than 9 m s−1, but it now resides above a thin layer of downward motion near the surface. Downward motion also develops to the south of the updrafts, which peaks around 5 km with a maximum of 1.5 m s−1. Due to the low-level divergence associated with the downward motion, the cyclonic vorticity below the updraft core weakens and the cyclonic vorticity center shifts to the southeast of the updraft 8 min later. At 0440 UTC (Fig. 2f), the updraft has weakened from 9 to 5 m s−1 with a more pronounced vertical tilt, and the downward motion has expanded to a larger area. Meanwhile, another congestus cloud cell has attained a stronger intensity southeast of the decaying cell. The convective cyclonic vortices merge together 40 min later (not shown).
The time sequence in Fig. 2 suggests that congestus convection can effectively amplify the low-level ambient vorticity, which is consistent with the idealized numerical model simulations by Wissmeier and Smith (2011) and Kilroy and Smith (2013). It is also consistent with Fang and Zhang (2011), who suggested that vorticity anomalies resulting from different modes of moist convection contribute to the development of Hurricane Dolly (2008). The evolution of “vortical congestus clouds” illustrated in Fig. 2 is similar to the evolution of VHTs (Hendricks et al. 2004; Montgomery et al. 2006) except that the strong cyclonic vorticity of congestus clouds is confined to a shallow layer in the lower troposphere most of the time. It is also worth noting that the upward motion of a congestus cloud, despite the limited depth, can be rather strong (~10 m s−1). In other words, strong convection is not necessarily deep. The development of downdrafts and vertical tilt at the decay stage of congestus clouds resembles the development of stratiform anvils during the decay of deep convective clouds.
4. Deep convection versus cumulus congestus
a. Partitioning of cumulus congestus and deep convection
To examine their contributions to TC formation, we classify deep convection and cumulus congestus based on the following algorithm (VHTs and convective bursts following the definitions used in some previous studies are examined in the appendix). (i) A convective grid column (see the definition in section 3) is designated as deep convection if its UTH is above 10 km and is designated as cumulus congestus otherwise. (ii) To take into account vertically tilted updrafts, convective grid points with w ≥ 2.0 m s−1 that are contiguous in the 3D space are defined as an updraft feature. If one grid column of an updraft feature has been designated as deep convection in the first step, all the grid columns of the updraft feature and its neighboring grid columns will be classified as deep convection, and they form a deep convective feature. All the other convective features are designated as cumulus congestus. (iii) The UTH of an updraft feature is defined by the maximum UTH of the convective columns of that feature. If the UTH of a cumulus congestus feature is above 6 km and the virtual temperature at the UTH is at least 0.5 K greater than the average over its ambient environment (40 × 40 neighboring grid points), the cumulus congestus feature is regarded as a transient congestus feature (i.e., a developing deep convective feature; Luo et al. 2009), and all the grid columns of the feature are designated as deep convection.
Figure 3a shows the mean vertical velocity profiles for deep convection, cumulus congestus, and nonconvective areas averaged over 24–54 h. The mean upward motion of cumulus congestus is confined below 8 km, with a maximum of about 1 m s−1 around 2 km, and is weakly negative or close to zero above 8 km. In contrast, deep convection has a deep layer of upward motion extending above 14 km, with one maximum of about 1.6 m s−1 between 4 and 5 km and a secondary maximum of about 1.4 m s−1 around 11 km. The mean vertical velocity of deep convection, however, is weaker than that of cumulus congestus below 2 km. The vertical velocity averaged over all the nonconvective grid columns is characterized by downward motion below 6 km and upward motion above. The profile is similar to that of stratiform precipitation, but the magnitude (less than 0.1 m s−1) is much weaker as the nonconvective areas also include nonprecipitating grid columns.
(a) The mean vertical velocity profiles and (b) the areal fractions of cumulus congestus (black), deep convection (red), and nonconvective areas (blue) within a 100-km radius. The mean vertical velocities in (a) are averaged over 24–54 h. The ordinate for the frequency of nonconvective areas is on the right in (b).
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
The time series of the areal fraction of each cloud type is shown in Fig. 3b. The areal fraction of cumulus congestus fluctuates around 4%. It is modulated by the diurnal cycle and has a weak increasing trend. The areal fraction of deep convection has a clear positive trend, increasing from less than 1% at 24 h to about 8% by 72 h. A prominent peak occurs around 60 h. It is not clear whether this peak is due to the diurnal cycle or land impacts, or both. Prior to genesis, the frequency of occurrence of cumulus congestus is higher than or comparable to that of deep convection; after genesis, deep convection occurs more frequently. Nonconvective areas are dominant all the time, covering more than 82% of the inner pouch region, and have an evident decreasing trend.
To examine the cloud properties, the mean reflectivity and the mean mixing ratio of total condensate (including cloud water, rainwater, ice, snow, and graupel) are shown in Fig. 4. The mixing ratio of cumulus congestus is less than 0.2 g kg−1, and the contour of 0.1 g kg−1 is confined below 6 km prior to genesis (Fig. 4a). The maximum mean reflectivity of cumulus congestus is less than 30 dBZ and peaks around 1 km. These features are consistent with a mix of midlevel congestus clouds and shallow cumulus clouds. After 54 h, the mixing ratio and reflectivity both increase significantly above 4 km, and the mixing ratio contour of 0.2 g kg−1 and the reflectivity contour of 20 dBZ extend up to 6 km between 60 and 66 h. In contrast to cumulus congestus, deep convection is characterized by a high mixing ratio and strong reflectivity throughout a deep layer at both the tropical wave and tropical cyclone stages (Fig. 4b). A mixing ratio greater than 0.1 g kg−1 extends up to 14 km, and the reflectivity contour of 20 dBZ also extends to the upper troposphere. The cloud properties illustrated in Fig. 4 suggest that the convection partitioning algorithm separates deep convection and cumulus congestus reasonably well.
The mean reflectivity (contours; dBZ) and mixing ratio of the total condensate (shading; g kg−1) for (a) cumulus congestus and (b) deep convection.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
b. Roles of cumulus congestus and deep convection in TC formation
To examine the roles of cumulus congestus and deep convection in TC formation, the vertical mass flux, the vertical moisture flux, and the net condensation associated with cumulus congestus and deep convection are calculated. The condensation field is an output variable from WRF. A positive value means that the rate of condensation and deposition exceeds the rate of evaporation and sublimation in a grid cell. To examine the relative contribution by each type of updrafts, the vertical fluxes and net condensation are summed over all the corresponding grid columns within a 100-km radius.
The vertical mass fluxes of cumulus congestus and deep convection are shown in Figs. 5a and 5b, respectively. The upward mass flux of cumulus congestus has a shallow structure, largely confined below 6 km throughout the time period of calculation. It increases sharply with height near the surface and peaks at 2–3 km, implying a strong low-level inflow. Consistent with the frequency of occurrence, there is a discernible diurnal cycle peaking around 34 h (0600 local time). Compared to cumulus congestus, the mass flux of deep convection (Fig. 5b) has a much deeper structure but has strong fluctuations prior to genesis. A remarkable feature is the substantial increase around 54 h. After this time the vertical mass flux of deep convection becomes much stronger than that of cumulus congestus and also appears more persistent. The vertical mass fluxes averaged over a pregenesis time period (30–54 h) and a postgenesis time period (54–72 h) are shown in Fig. 5c. Prior to genesis, the vertical mass flux of cumulus congestus exceeds that of deep convection below 3 km; above 3 km, the vertical mass flux of cumulus congestus is weaker than that of deep convection but is still appreciable. Also note that cumulus congestus has a more bottom-heavy profile, indicating a stronger low-level mass convergence. The vertical mass flux of cumulus congestus remains nearly the same after genesis, while the vertical mass flux of deep convection more than doubles.
(a)–(c) The total upward mass flux (g m−2 s−1), (d)–(f) the total upward moisture flux (10−2 g m−2 s−1), and (g)–(i) the total net condensation (g kg−1 day−1) associated with (left) cumulus congestus and (center) deep convection within a 100-km radius. (right) The pregenesis averages (30–54 h; solid curves) and postgenesis averages (54–72 h; dashed curves) of the corresponding variables associated with cumulus congestus (black) and deep convective (red) updrafts.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
The vertical moisture fluxes associated with cumulus congestus and deep convection are shown in Figs. 5d–f. The vertical moisture flux of deep convection extends to a higher altitude than that of cumulus congestus, but the latter has a more steady vertical transport in the lower troposphere prior to genesis. After genesis, the vertical moisture flux of deep convection increases significantly and becomes much stronger than that of cumulus congestus. The comparison between the two types of convection is clearly illustrated by Fig. 5f. The pregenesis averages over 30–54 h show that cumulus congestus has a stronger vertical moisture transport than does deep convection below 3 km but the latter extends to a deep layer. While the vertical moisture flux of cumulus congestus does not change appreciably after genesis, the deep convective flux more than doubles and becomes dominant throughout the troposphere.
The net condensation associated with cumulus congestus and deep convection is shown in Figs. 5g–i. The net condensation of cumulus congestus is mainly confined below 6 km and does not show a clear trend from the tropical wave to the tropical cyclone stage. It has two maxima: one around 2 km and the other around 4 km. The former is associated with the enhanced static stability by the trade wind inversion layer, and the latter is associated with the freezing-level stability and/or the midlevel dry air (e.g., Johnson et al. 1999; Takemi et al. 2004; Takayabu et al. 2010; Brown and Zhang 1997; Zuidema 1998; Redelsperger et al. 2002). The net condensation of deep convection has a deeper structure with a maximum around 4–5 km. Similar to the vertical mass flux and vertical moisture flux, it has large fluctuations prior to genesis but becomes much stronger and more persistent after genesis. The pregenesis and the postgenesis averages (Fig. 5i) show that the net condensation of cumulus congestus is slightly stronger than that of deep convection below 3 km prior to genesis and that the contribution of deep convection increases significantly and becomes dominant after genesis. The implication of the associated diabatic heating profiles to intensification of the disturbance will be discussed later in this section.
Figure 6a shows the rhs terms in Eq. (3) integrated between 2 and 8 km for the cumulus congestus and deep convective areas within a 100-km radius. Here, 2 km is the altitude of the maximum upward moisture flux of cumulus congestus; 8 km is the height of the mean upward motion of cumulus congestus (Fig. 3a), and above this level the net condensation associated with cumulus congestus is negligible (Fig. 5). Throughout the time period of calculation, the net vertical moisture influx exceeds the net condensation for cumulus congestus, suggesting that cumulus congestus contributes to both condensational heating and moistening of the free troposphere. By contrast, the net condensation of deep convection exceeds the vertical moisture influx most of the time, especially after genesis. This suggests that deep convection will either reduce the local moisture content in the lower and middle troposphere or needs to get the additional moisture supply from the ambient environment via entrainment or convergence (Holloway and Neelin 2009). This is consistent with Wu’s (2003) finding that the moisture convergence driven by deep heating is not sufficient to sustain itself while shallow heating is more effective in driving the low-level convergence. It also explains why the high moisture content in the lower to middle troposphere is important for the development of deep convection. Also shown in Fig. 6 is the radial moisture flux across a 100-km radius between 2 and 8 km. It fluctuates around zero prior to genesis, suggesting that horizontal moisture convergence above the boundary layer does not make a significant contribution to the moisture supply in the inner pouch region. After genesis the inward moisture transport increases significantly due to a deep inflow layer (Fritz and Wang 2014, manuscript submitted to J. Atmos. Sci., hereafter FW). In summary, Fig. 6a suggests that cumulus congestus plays a more important role than deep convection in moistening the lower to middle troposphere.
The radial moisture flux (green dashed; positive values for inward flux), the net vertical moisture influx (solid curves), and the net condensation (dotted) within a 100-km radius (a) between 2 and 8 km and (b) between 3 and 14 km. Black curves are for cumulus congestus regions, red curves are for deep convective regions, and blue curves in (a) are for nonconvective regions. All units are 107 kg s−1.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
Figure 6b shows the net condensation and net vertical moisture influx integrated between 3 and 14 km: 3 km is the altitude of maximum upward moisture flux of deep convection, and 14 km is the height of the mean deep convective upward motion (Fig. 3a; replacing 14 km by 15 km does not induce any significant changes in Fig. 6b). Over this deep layer, the net vertical moisture influx exceeds the net condensation for both cumulus congestus and deep convection. Together with Fig. 6a, it suggests that deep convection moistens the upper troposphere, and cumulus congestus plays a dominant role in moistening the lower to middle troposphere. Although deep convection contributes to the net moistening of the deep troposphere, the moisture surplus (net vertical moisture influx minus net condensation) is not as large as that associated with cumulus congestus.
The PV production associated with cumulus congestus (Fig. 7a) is characterized by an intermittent positive PV tendency between 600 and 900 hPa and a layer of persistent positive tendency below 900 hPa. The latter feature is consistent with the strong vertical gradient of net condensation near the surface (Fig. 5g). The PV production by deep convection has a different vertical profile. It peaks between 600 and 700 hPa most of the time prior to genesis. Despite a stronger maximum magnitude, it is weaker than the PV production by cumulus congestus near the surface. Consistent with the decreasing net condensation with height in the upper troposphere (Fig. 5h), deep convection contributes to a negative PV tendency above 500 hPa. After genesis, the PV production by deep convection becomes stronger and more persistent, and large positive values extend down to the surface. Meanwhile, the negative PV tendency above 500 hPa also becomes more prominent. To examine the accumulative contribution by each type of convection, the PV production is integrated in time for a pregenesis time period (30–54 h) and a postgenesis time period (54–72 h; see Fig. 7c). Prior to genesis, cumulus congestus makes a larger contribution than deep convection below 900 hPa and a weaker but still appreciable contribution above 900 hPa. After genesis, the PV generation associated with deep convection increases significantly and exceeds that of cumulus congestus throughout the troposphere.
The time–height cross section of the total potential vorticity tendency (10−5 PVU s−1, where 1 PVU = 10−6 K kg−1 m2 s−1) associated with (a) cumulus congestus and (b) deep convection within a 100-km radius. (c) The accumulated PV change (PVU) associated with cumulus congestus (black) and deep convection (red) over 30–54 (solid) and 54–72 h (dashed).
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
In summary, cumulus congestus plays a dominant role in moistening the lower to middle troposphere and spinning up the near-surface circulation prior to genesis, while deep convective clouds play a key role in moistening the upper troposphere and amplifying the cyclonic circulation over a deep layer. It is also notable that the transition from the tropical wave stage to the TC stage is marked by a substantial increase in the vertical mass flux, vertical moisture flux, and net condensation by deep convection and is accompanied by a rapid drop of sea level pressure [see Fig. 1 in Wang (2014)]. This transition will be further investigated in section 5.
c. Contribution of the nonconvective processes
It is necessary to point out that the convection partitioning algorithm in section 4a only identifies cumulus congestus and deep convection at the active phase of their life cycles. As shown in Fig. 2 and previous studies (e.g., Houze 1997), downdrafts develop at the decay stage of convective clouds, and evaporation of cloud hydrometeors can contribute to additional moistening. Here, we will briefly examine the collective role of nonconvective processes in tropical cyclone formation. The nonconvective areas are the grid columns that are not identified as convective (see section 3), including stratiform anvil, decaying cumulus congestus, and nonprecipitating areas. Decaying cumulus congestus and decaying deep convection (i.e., stratiform anvils) are not examined separately because the latter has been examined in some previous studies (Wang et al. 2010b; Wang 2012; Rappin and Nolan 2012) and also because it is not clear whether and how decaying cumulus congestus can be distinguished from stratiform processes based on the vertical velocity profile. Nonprecipitating areas are also included in the calculations, but they are not expected to make a significant contribution to the net condensation or vertical moisture transport.
Figure 8a shows the time–height cross section of the vertical mass flux associated with nonconvective processes. It is characterized by upward mass flux in the upper troposphere (above 5–6 km) and downward mass flux in the lower troposphere, suggesting the preponderance of stratiform precipitation. Both the upward mass flux and downward mass flux have an increasing trend, and their fluctuations are roughly in phase. The comparison between the pregenesis average and postgenesis average (Fig. 8b) shows that the vertical mass transport after genesis is significantly larger than prior to genesis, which is consistent with Wang et al. (2010b). The level of zero vertical mass flux rises slightly after genesis. Also note that the magnitude of the upward mass flux in the upper troposphere is close to the magnitude of the downward mass flux in the lower troposphere. Given the decrease in density with height, this indicates that the upward motion in the upper troposphere is stronger than the downward motion in the lower troposphere (Wang et al. 2010b).
Time–height cross sections of (a) vertical mass flux, (c) vertical moisture flux, (e) net condensation, and (g) PV production associated with the nonconvective processes. (b),(d),(f),(h) The pregenesis averages (30–54 h; solid) and postgenesis averages (54–72 h; dashed) of the corresponding variables associated with the nonconvective processes.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
The vertical moisture flux is shown in Figs. 8c and 8d. Similar to the vertical mass flux, the vertical moisture flux is characterized by upward moisture flux in the upper troposphere and downward moisture flux in the lower troposphere. The former is much weaker than the latter due to the low moisture content in the upper troposphere. The upward moisture flux peaks around 7-km altitude, and the downward moisture flux peaks around 2-km altitude. The increase in the vertical moisture flux with height between the two levels indicates divergence of the vertical moisture flux, which contributes to a negative tendency of water vapor content in this layer.
Figures 8e and 8f show the net condensation. While the profile is consistent with the typical heating profile of stratiform precipitation, it is worth pointing out that deep downdrafts in nonprecipitating regions may enhance the evaporative cooling in the lower troposphere and weaken the heating in the upper troposphere. Consistent with the net condensation profile, the nonconvective processes have positive potential vorticity production between 600 and 350 hPa and weak negative PV production near the surface (Figs. 8g and 8h).
Stratiform processes can moisten the lower troposphere via the showerhead mechanism (Bister and Emanuel 1997; Rappin and Nolan 2012), as indicated by Fig. 8e and 8f, but the vertically integrated net condensation (blue dotted curve in Fig. 6a) suggests that evaporation associated with nonconvective processes only makes a small contribution to the water vapor budget. Also note that evaporation cannot increase the equivalent potential temperature (E. Zipser 2013, personal communication). Besides, the vertically integrated net vertical moisture flux (blue solid curve in Fig. 6a) contributes to a much larger negative tendency, suggesting that the net effect of the nonconvective processes is to reduce the water vapor content in the lower to middle troposphere.
5. Convective organization and transition to deep convection near the pouch center
Wang (2012) showed that the thermodynamic conditions in the inner pouch region are different from those in the outer pouch region. The former are characterized by a larger saturation fraction and a short incubation time scale, which are believed favorable for deep convection and genesis (e.g., Raymond 2000; Rappin et al. 2010; Smith and Montgomery 2012). In this section, we will examine how the thermodynamic conditions within the wave pouch affect the cloud population and evolution.
As shown in Fig. 9a, the mean vertical velocity profiles for cumulus congestus and deep convection within the 150–250-km annulus, or the outer pouch region, are similar to those in the inner pouch region. The frequencies of occurrence of cumulus congestus and deep convection, however, are both strongly modulated by the diurnal cycle in the outer pouch region, and neither has an evident increasing trend (Fig. 9b). Both frequencies of occurrence are less than 4% most of the time, and the frequency of deep convection is slightly but consistently lower than that of cumulus congestus. Nonconvective grids account for more than 92% of the total area, varying out of phase with cumulus congestus and deep convection. We also examined the vertical mass flux, vertical moisture flux, and net condensation associated with cumulus congestus and deep convection in the outer pouch region. Due to the low frequencies of occurrence of the two types of convection in the outer pouch region, these variables are smaller than their counterparts in the inner pouch region despite being summed over a much larger area (not shown). Similar to the time series of areal coverage, the fluxes and net condensation have a strong diurnal cycle and do not have an evident increasing trend.
As in Fig. 3, but for the outer pouch region (between 150- and 250-km radii).
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
The weak net condensation in the outer pouch region indicates a low precipitation rate. Previous studies have suggested that column water vapor (CWV) or column relative humidity (also known as saturation fraction) has strong control over precipitation rate (Raymond 2000; Bretherton et al. 2004). More recently, Peters and Neelin (2006), Neelin et al. (2009), and Peters et al. (2009) examined the transition to strong convection using satellite data and showed that the power-law increase of ensemble-average precipitation rate with CWV occurs above a critical value of CWV. Analysis of sounding and satellite data (Sherwood and Wahrlich 1999; Mapes et al. 2006; Holloway and Neelin 2010) showed that column water vapor was elevated a few hours before the onset of heavy precipitation and peaked either during or after convection, and suggested that high column water vapor is not only a consequence of heavy precipitation but also a precondition. Masunaga (2012) showed that the short-term relationship between tropospheric humidity and precipitation rate is highly nonlinear and reflects the growth and decay of convection in various humidity environments. The exponential increase of precipitation with CWV (or saturation fraction) was attributed to the fact that highly organized convection with heavy precipitation becomes more frequent and intense as the environment moistens.
Figure 10 shows a scatterplot of precipitation rate and saturation fraction (SF; column relative humidity between 300 and 1000 hPa) averaged over the inner and outer pouch regions. The precipitation rate was derived from hourly accumulated precipitation, and no temporal average was applied to the saturation fraction. In comparison to the outer pouch region, the inner pouch region has a higher saturation fraction and a larger precipitation rate. Due to the hourly data used here [note that daily or monthly mean data were used in Bretherton et al. (2004)], the plot has a high scatter. Nevertheless, the rapid increase in precipitation with saturation fraction above a certain critical value (~87%) is similar to what was depicted in previous observational studies (e.g., Raymond et al. 2007; Bretherton et al. 2004; Neelin et al. 2009). The 5-h-running-mean precipitation and SF are also shown in Fig. 10. When the precipitation rate is 5 in. day−1 (12.7 cm day−1) or lower in the inner pouch region, the orbit of the precipitation rate and SF follows a clockwise loop: SF increases before the peak in precipitation rate and reaches its maximum after the peak, suggesting that precipitation processes moisten the air column. With higher precipitation rate (>5 in. day−1), the orbit takes a counterclockwise turn, indicating a drop in SF after the maximum precipitation.
Scatterplot of saturation fraction (%) and precipitation rate (in. day−1; 1 in. = 2.54 cm) averaged over the inner pouch region (circles) and the outer pouch region (triangles) from 24 to 72 h (5-h running mean). The green marks indicate the conditions at t = 24 h, and the dashed curves represent the evolution of the 5-h-running-mean precipitation rate and saturation fraction.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
Figure 10 suggests that the low precipitation rate in the outer pouch region can be attributed to the low saturation fraction and that the rapid development of deep convection prior to genesis can be explained by the criticality of CWV or SF (Peters and Neelin 2006; Neelin et al. 2009). That is, the transition to strong convection occurs only after the column is moistened sufficiently. Such a transition fails to take place in the outer pouch region due to the different dynamic and kinematic characteristics of the flow between the inner and outer pouch regions. The flow in the outer pouch region has a large strain rate, and moisture lofted by convection tends to be filamented and deep convection may be suppressed (Rozoff et al. 2006). By contrast, the inner pouch is characterized by weak relative flow (note that the pouch center is a stagnation point) and weak deformation, where moisture is more likely to accumulate and convection can more effectively bring an air column close to saturation (Wang 2012). The low-level moisture convergence driven by diabatic heating and the boundary layer friction also helps to moisten the inner pouch region. Additionally, the inner pouch region is shielded from hostile environmental impacts by the surrounding “surf zone” [i.e., a ring of strong deformation acting as an effective material boundary (B. Rutherford 2012, personal communication)], while the outer pouch region is subject to dry-air entrainment from outside the pouch (Fritz and Wang 2013). All these factors make the inner pouch region more favorable for deep convection and genesis.
The convective evolution in the inner pouch region suggests the following picture for TC formation. At the early stage of TC formation, cumulus congestus clouds play a dominant role in moistening the lower to middle troposphere and intensifying the low-level circulation. Owing to the dynamic and kinematic structures of the wave pouch, convective moistening is most effective near the pouch center. After the inner pouch region is moistened sufficiently (i.e., CWV exceeding a critical threshold), deep convection develops rapidly near the pouch center. The thermodynamic condition in the inner pouch region thus plays an important role in organizing deep convection. The concentrated deep convection in the inner pouch region, with strong vertical and radial gradients of diabatic heating rate, significantly strengthens the secondary circulation, intensifies the primary circulation over a deep layer, and leads to the formation of a tropical cyclone.
6. Summary
The roles of cumulus congestus and deep convection in TC formation were examined in a high-resolution numerical model simulation of Tropical Cyclone Fay (2008). Cumulus congestus and deep convection are classified based on the updraft-top height of updraft features. A convective updraft feature is defined as convective grid points with w ≥ 2 m s−1 contiguous in the 3D space, which takes into account the vertical tilt of convective updrafts. The algorithm also identifies transient congestus clouds and designates them as deep convection. Examination of reflectivity and mixing ratio of total condensate suggests that the partitioning algorithm separates deep convection and cumulus congestus reasonably well.
It was shown that cumulus congestus makes a slightly larger contribution to the vertical mass flux and net condensation than deep convection below 3 km prior to genesis. Previous studies (e.g., Fritz and Wang 2013) showed that net condensation is a major water vapor sink and the vertical moisture flux is a major source for the free troposphere. A simple water vapor budget analysis reveals that the net vertical moisture influx of deep convection is less than the column net condensation in the lower to middle troposphere while the net vertical moisture influx by cumulus congestus exceeds the associated net condensation. Cumulus convection thus plays the major role in moistening the lower to middle troposphere. Although evaporation of hydrometeors also moistens the lower troposphere, it only makes a small contribution to the water vapor budget, and due to the divergence of the vertical moisture flux, the net effect of nonconvective processes is to reduce the water vapor content in the lower and middle troposphere.
The PV budget analysis shows that the PV production by cumulus congestus is larger than that by deep convection below 900 hPa prior to genesis and is weaker but still appreciable above 900 hPa. Therefore, cumulus congestus plays an important role in spinning up the near-surface circulation while deep convection is critical for intensification of the TC vortex over a deep layer. The transition from the tropical wave stage to the TC stage is marked by a substantial increase in the vertical mass and moisture fluxes and net condensation associated with deep convection, and deep convection becomes the dominant contributor after genesis. The rapid development of deep convection near the pouch center around the genesis time can be explained by the power-law increase of precipitation rate with CWV (e.g., Raymond 2000; Bretherton et al. 2004). Recent studies (Peters and Neelin 2006; Neelin et al. 2009; Peters et al. 2009) suggested that the transition to strong convection occurs above a critical threshold of CWV. The high CWV at the inner pouch region, together with the strong-rotation and weak-deformation flow and the system-scale low-level convergence, plays an important role in convective organization.
The distribution and evolution of deep convection and cumulus congestus in the inner pouch region (within a 100-km radius) were compared to those in the outer pouch region (between 150- and 250-km radii). It was found that the deep convection and cumulus congestus both occur less frequently in the outer pouch region than in the inner pouch region and do not have an evident increasing trend. The transition to sustained deep convection does not occur in the outer pouch region due to the relatively low saturation fraction and strong deformation in the outer pouch region.
7. Discussion
The evolution of a congestus cloud feature showed that cumulus congestus can effectively amplify the low-level ambient vorticity and that the life cycle of a “vortical congestus cloud” cell resembles that of a vortical hot tower. As pointed out by Raymond and Sessions (2007), the vertical mass flux of deep convection in the tropics typically peaks in the upper troposphere. Raymond et al. (2011) suggested that a midlevel vortex is conducive for TC genesis by changing the vertical mass flux profile of deep convection from top heavy to bottom heavy. This study suggests that cumulus congestus, owing to its bottom-heavy heating and vertical mass flux profiles, provides a simpler and more direct pathway to the formation and intensification of the TC protovortex near the surface. On the other hand, this does not mean that deep convection is not important for TC formation. Due to the shallow heating profile, the strong PV generation by cumulus congestus is confined to a thin layer near the surface. The balanced response to the cumulus congestus heating solved by the Sawyer–Eliassen equation (Eliassen 1951) has a shallow overturning circulation with outflow above 3 km (not shown), which tends to spin down the midlevel circulation. Cumulus congestus alone thus cannot produce a deep cyclonic vortex, and only after deep convection becomes dominant does a tropical cyclone vortex extend throughout the troposphere [see vorticity evolution in Fig. 9a in Wang (2014)]. With a stronger vortex and larger inertial stability, latent heat release can more efficiently be converted to the kinetic energy of the primary circulation (Hack and Schubert 1986; Nolan 2007; Fang and Zhang 2011).
There are two different views on TC formation: it is (i) an abrupt process, in which genesis is triggered by certain physical conditions or mechanisms (Nolan 2007) or (ii) a gradual process, which was succinctly described by Ooyama (1982): “It is unrealistic to assume that the formation of an incipient vortex is triggered by a special mechanism or mechanisms… it is far more natural to assume that genesis is a series of events, arising by chance from quantitative fluctuations of the normal disturbances, with the probability of further evolution gradually increasing as it proceeds.” This study suggests that the two views are not mutually exclusive and that TC formation can be regarded as the combination of the two processes or two stages. In the first stage, cumulus congestus gradually moistens the lower and middle troposphere and intensifies the low-level circulation. This process is punctuated by intermittent deep convection and modulated by the diurnal cycle. Owing to the dynamic and kinematic structures of the wave pouch, convective moistening is most effective near the pouch center. The second stage starts after CWV exceeds a certain critical value, and is marked by the transition to sustained deep convection in the inner pouch region, the rapid intensification of the TC vortex throughout the troposphere, and a sudden drop in sea level pressure. The first stage, at which cumulus congestus plays a dominant role in moisture preconditioning and low-level spinup, is a gradual process, but the second stage is a relatively abrupt process. It consummates the TC formation process and amplifies the disturbance into the TC intensity. This two-stage conceptual model differs from Zehr’s (1992) version in that we emphasize the role of cumulus congestus in preconditioning the atmosphere for the transition to sustained deep convection and the role of the wave pouch in convective organization, while Zehr (1992) emphasized the role of a convective maximum occurring about 3 days, on average, prior to genesis.
The importance of midlevel or column moistening for tropical cyclone formation has been documented by previous studies (e.g., Bister and Emanuel 1997; Raymond et al. 1998; Nolan 2007; Wang 2012; Smith and Montgomery 2012; Komaromi 2013). This study suggests that the column moistening is important for genesis because it is critical for the transition to sustained deep convection, instead of eliminating or reducing downdrafts. The convective transition after column moistening is consistent with the findings of Nolan (2007). Using idealized numerical model simulations, Nolan (2007) showed that a midlevel vortex intensifies and contracts after the relative humidity in the core exceeds 80% over most of the troposphere. After the midlevel vortex becomes sufficiently strong, a small surface-concentrated vortex (SSCV) is created by a long-lived convective updraft, which evolves into the core of the TC. The development of long-lived convective updrafts and the SSCV in Nolan’s simulations likely occurs during the second stage. On the other hand, the spinup of the low-level circulation in this simulation and in the numerical simulation in Wang (2012) occurs along with moisture preconditioning and precedes organized deep convection and the spinup of the midlevel circulation.4 Also note that the transition to sustained deep convection does not necessarily occur immediately after CWV exceeds the critical value due to the stochastic nature of convection (e.g., Neelin et al. 2009; Holloway and Neelin 2010; Stechmann and Neelin 2011; Masunaga 2012). In fact, the analysis of dropsonde data and a numerical model simulation in Wang (2012) showed that the column near the pouch center becomes quite moist 1–2 days prior to genesis, while the rapid increase in deep convection occurs just a few hours prior to genesis in the simulation of Fay. It is also worth emphasizing that moisture alone is not a sufficient condition for tropical cyclogenesis in the real atmosphere. The dynamic, thermodynamic, and kinematic structures of the wave pouch are all critical for convective organization and TC formation. A familiar example is the ITCZ. It is a region of frequent deep convection and abundant moisture, where midlevel mesoscale convective vortices are prevalent but tropical cyclogenesis is not (Dunkerton 2006).
Our emphasis on moisture preconditioning by cumulus congestus seems to contradict the recent study by Hohenegger and Stevens (2013a,b). Hohenegger and Stevens (2013a) estimated the time scale of moistening by congestus clouds based on the local surface latent heat flux, and showed that this time scale is longer than the time scale for the transition from cumulus congestus to deep convection. They thus concluded that cumulus congestus plays a less important role than large-scale ascent in moistening the column. It is worth pointing out that cumulus congestus draws moisture not only from local evaporation but also from the boundary layer convergence (FW). Neglecting the latter will lead to an overestimate of the moistening time scale by cumulus congestus.5 Besides, we are more interested in the transition between different cloud regimes, which requires column moistening over a finite area and is slower than the transition of individual clouds. On the other hand, we caution that the presence of cumulus congestus by itself does not provide a sufficient condition for the transition to deep convection. The moisture lofted by congestus convection needs to be retained in a certain region and isolated from the relatively dry environment so that the middle troposphere can be moistened sufficiently for deep convection. The inner pouch region, with its near-solid-body rotation and the surrounding “surf zone,” provides such an environment. In contrast to the vast area affected by the MJO’s moisture preconditioning, the inner pouch region is small and moistens quickly. Lack of horizontal confinement helps explain why there are many regions of congestus clouds with scarce deep convection in the tropics (e.g., Takayabu et al. 2010). Furthermore, cumulus congestus should not be regarded as the only process to initiate deep convection. The boundary layer dynamics, including the density current, gravity waves, vorticity filaments, and, in particular the low-level vortex, likely plays a more immediate role in convective triggering and convective organization. This issue merits further study.
This study is admittedly based on the numerical simulation of a single tropical cyclone originating from a tropical easterly wave. Although the simulation of 1-km resolution resolves the statistics of vertical motion reasonably well (Wang 2014), large-eddy simulations (LESs) may be necessary to study the evolution of individual convective clouds. Observational studies are also desirable to examine whether the convective transition suggested by the numerical model simulation occurs in a real tropical cyclone and whether the two-stage conceptual model can be generalized for tropical cyclone formation associated with other types of precursor disturbances.
Acknowledgments
This research was supported by National Science Foundation Grants AGS-1016095 and AGS-1118429. The author would like to thank Tim Dunkerton and three anonymous reviewers for helpful comments on an earlier version of the manuscript, Isaac Hankes for proofreading the manuscript, and Ed Zipser, Chris Davis, and Steve Nesbitt for stimulating discussions. Computing resources were provided by the Climate Simulation Laboratory at NCAR’s Computational and Information Systems Laboratory, sponsored by the National Science Foundation and other agencies.
APPENDIX
Vortical Hot Towers and Convective Bursts
The vertical fluxes and net condensation associated with convective bursts and vortical hot towers (VHTs) are examined here following the definitions used in some previous studies. In Montgomery et al. (2006), a VHT is identified if the vertical velocity in a grid column is no less than 1 m s−1 between 1 and 15 km. We adopted a less stringent definition and required that the vertical velocity in a grid column be no less than 1 m s−1 between 1 and 10 km for a VHT; in addition, no vorticity threshold was applied, following Montgomery et al. (2006). As shown in Fig. A1b, the areal fraction of VHTs is nearly zero before 30 h, remains less than 0.2% prior to genesis, and is less than 0.5% most of the time, even after genesis, which is much smaller than that of deep convective features (Fig. 3b). Despite the weak threshold used in the definition, the mean vertical velocity of VHTs averaged within a 100-km radius over 30–54 h has a deep layer of strong upward motion, with a maximum close to 8 m s−1 around 8 km (Fig. A1a). The time–height cross section of the mean vertical velocity in the inner pouch region (not shown) reveals episodes of strong upward motion of up to 16 m s−1 prior to genesis. However, the contributions of VHTs to the upward mass flux, upward moisture flux, and net condensation are quite limited due to their low frequency of occurrence. As shown by the pregenesis and postgenesis averages in Figs. A2a–c, the magnitudes of the total upward mass flux, upward moisture flux, and net condensation by VHTs are about 5% of those associated with cumulus congestus features (Fig. 5) prior to genesis and no more than ⅓ after genesis (Fig. 5).
(a) The mean vertical velocity profiles (averaged over 30–54 h) and (b) areal fractions (%) of VHTs (solid) and convective bursts (dashed) within a 100-km radius.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
The total contribution of (a)–(c) VHTs and (d)–(f) convective bursts to (left) the vertical mass flux (g m−2 s−1), (center) vertical moisture flux (10−2 g m−2 s−1), and (right) net condensation (g kg−1 day−1) within a 100-km radius. The solid lines represent the pregenesis (30–54 h) averages and the dashed lines represent the postgenesis (54–72 h) averages.
Citation: Journal of the Atmospheric Sciences 71, 5; 10.1175/JAS-D-13-0257.1
The frequency of occurrence of convective bursts was examined following Reasor et al. (2009). They defined convective bursts as the regions with the 2–6-km layer-mean vertical motion exceeding 5 m s−1 and 2-km reflectivity greater than 30 dBZ. Here, we only adopted the vertical velocity criterion to identify a convective burst. As shown in Fig. A1b (dashed curve), the areal fraction of convective bursts is 2–3 times as large as that of VHTs but is still much lower than the areal fraction of deep convection (Fig. 3b). The mean vertical velocity of the convective bursts peaks around 4.5 km with a maximum of about 8 m s−1 and is weaker than that of VHTs above 6 km or below 2 km (Fig. A1a). The contributions of convective bursts to the upward mass flux, upward moisture flux, and net condensation are larger than those of VHTs but are still lower than the contributions by deep convective features and are also lower than the contributions by cumulus congestus prior to genesis (Fig. 5).
There is no generally accepted algorithm to identify a VHT or a deep convective cell in a numerical model simulation. Given a background cyclonic rotation, deep convective clouds will attain a rotating structure as the associated low-level convergence amplifies the ambient vorticity. VHTs should thus constitute the major, if not the complete, population of deep convective clouds in the inner wave pouch region. Comparison between Figs. 3 and A1 suggests that the frequency of occurrence of VHTs is likely underestimated by Montgomery et al.’s (2006) definition or our modification thereof. This is mainly due to the implicit assumption in Montgomery et al.’s definition that convective updrafts all have an upright structure, which is not true when convective cells are subject to the impacts of strong vertical shear. A similar comment applies to the definition of convective bursts following Reasor et al. (2009), as deep convection is not necessarily associated with strong updrafts (Fierro et al. 2012), and strong upward motion in the lower troposphere does not necessarily have a deep structure (Fig. 2). The comparisons suggest that the vertical tilt structure needs to be considered to identify deep convection properly and that intense convection is not equivalent to deep convection.
REFERENCES
Bister, M., and K. Emanuel, 1997: The genesis of Hurricane Guillermo: TEXMEX analyses and a modeling study. Mon. Wea. Rev., 125, 2662–2682.
Braun, S. A., 2006: High-resolution simulation of Hurricane Bonnie (1998). Part II: Water budget. J. Atmos. Sci., 63, 43–64.
Bretherton, C. S., M. E. Peters, and L. E. Back, 2004: Relationships between water vapor path and precipitation over the tropical oceans. J. Climate, 17, 1517–1528.
Brown, R. G., and C. Zhang, 1997: Variability of midtropospheric moisture and its effect on cloud-top height distribution during TOGA COARE. J. Atmos. Sci., 54, 2760–2774.
de Rooy, W. C., and Coauthors, 2013: Entrainment and detrainment in cumulus convection: An overview. Quart. J. Roy. Meteor. Soc., 139, 1–19.
Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 3077–3107.
Dunkerton, T. J., 2006: A tale of two ITCZs—The Jim Holton perspective. Bull. Amer. Meteor. Soc., 87, 1492–1495.
Dunkerton, T. J., M. T. Montgomery, and Z. Wang, 2009: Tropical cyclogenesis in a tropical wave critical layer: Easterly waves. Atmos. Chem. Phys., 9, 5587–5646.
Eliassen, A., 1951: Slow thermally or frictionally controlled meridional circulation in a circular vortex. Astrophys. Norv., 5, 19–60.
Fang, J., and F. Zhang, 2010: Initial development and Genesis of Hurricane Dolly (2008). J. Atmos. Sci., 67, 655–672.
Fang, J., and F. Zhang, 2011: Evolution of multiscale vortices in the development of Hurricane Dolly (2008). J. Atmos. Sci., 68, 103–122.
Fierro, A. O., E. J. Zipser, M. A. LeMone, J. M. Straka, and J. (Malkus) Simpson, 2012: Tropical oceanic hot towers: Need they be undilute to transport energy from the boundary layer to the upper troposphere effectively? An answer based on trajectory analysis of a simulation of a TOGA COARE convective system. J. Atmos. Sci., 69, 195–213.
Fritz, C., and Z. Wang, 2013: A numerical study of the impacts of dry air on tropical cyclone formation: A development case and a non-development case. J. Atmos. Sci., 70, 91–111.
Hack, J. J., and W. H. Schubert, 1986: Nonlinear response of atmospheric vortices to heating by organized cumulus convection. J. Atmos. Sci., 43, 1559–1573.
Hendricks, E. A., M. T. Montgomery, and C. A. Davis, 2004: The role of “vortical” hot towers in the formation of Tropical Cyclone Diana. J. Atmos. Sci., 61, 1209–1232.
Hohenegger, C., and B. Stevens, 2013a: Preconditioning deep convection with cumulus congestus. J. Atmos. Sci., 70, 448–464.
Hohenegger, C., and B. Stevens, 2013b: Reply to “Comments on ‘Preconditioning deep convection with cumulus congestus.’” J. Atmos. Sci., 70, 4155–4156.
Holloway, C. E., and J. D. Neelin, 2009: Moisture vertical structure, column water vapor, and tropical deep convection. J. Atmos. Sci., 66, 1665–1683.
Holloway, C. E., and J. D. Neelin, 2010: Temporal relations of column water vapor and tropical precipitation. J. Atmos. Sci., 67, 1091–1105.
Hong, S.-Y., and J.-O. J. Lim, 2006: The WRF single-moment 6-class microphysics scheme (WSM6). J. Korean Meteor. Soc., 42, 129–151.
Hong, S.-Y., Y. Noh, and J. Dudhia, 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 2318–2341.
Houze, R. A., Jr., 1997: Stratiform precipitation in regions of convection: A meteorological paradox? Bull. Amer. Meteor. Soc., 78, 2179–2196.
Houze, R. A., Jr., and C. Cheng, 1977: Radar characteristics of tropical convection observed during GATE: Mean properties and trends over the summer season. Mon. Wea. Rev., 105, 964–980.
Johnson, R. H., T. M. Rickenbach, S. A. Rutledge, P. E. Ciesielski, and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12, 2397–2418.
Kain, J. S., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170–181.
Khouider, B., and A. J. Majda, 2006: A simple multicloud parameterization for convectively coupled tropical waves. Part I: Linear analysis. J. Atmos. Sci., 63, 1308–1323.
Khouider, B., and A. J. Majda, 2008: Multicloud models for organized tropical convection: enhanced congestus heating. J. Atmos. Sci., 65, 895–914.
Kiladis, G. N., M. C. Wheeler, P. T. Haertel, K. H. Straub, and P. E. Roundy, 2009: Convectively coupled equatorial waves. Rev. Geophys., 47, RG2003, doi:10.1029/2008RG000266.
Kilroy, G., and R. K. Smith, 2013: A numerical study of rotating convection during tropical cyclogenesis. Quart. J. Roy. Meteor. Soc., 139, 1255–1269.
Komaromi, W. A., 2013: An investigation of composite dropsonde profiles for developing and nondeveloping tropical waves during the 2010 PREDICT field campaign. J. Atmos. Sci., 70, 542–558.
Lau, K. M., and H. T. Wu, 2003: Warm rain processes over tropical oceans and climate implications. Geophys. Res. Lett., 30, L2290, doi:10.1029/2003GL018567.
Luo, Z., G. Y. Liu, G. L. Stephens, and R. H. Johnson, 2009: Terminal vs transient cumulus congestus: A CloudSat perspective. Geophys. Res. Lett., 36, L05808, doi:10.1029/2008GL036927.
Mapes, B. E., 2000: Convective inhibition, subgrid-scale triggering energy, and stratiform instability in a toy tropical wave model. J. Atmos. Sci., 57, 1515–1535.
Mapes, B. E., S. Tulich, J. Lin, and P. Zuidema, 2006: The mesoscale convection life cycle: Building block or prototype for large-scale tropical waves? Dyn. Atmos. Oceans, 42, 3–29.
Martin, J. E., 2006: Mid-Latitude Atmospheric Dynamics: A First Course. Wiley Press, 324 pp.
Masunaga, H., 2012: Short-term versus climatological relationship between precipitation and tropospheric humidity. J. Climate, 25, 7983–7990.
Montgomery, M. T., M. E. Nicholls, T. A. Cram, and A. B. Saunders, 2006: A vortical hot tower route to tropical cyclogenesis. J. Atmos. Sci., 63, 355–386.
Montgomery, M. T., Z. Wang, and T. J. Dunkerton, 2010: Coarse, intermediate and high resolution numerical simulations of the transition of a tropical wave critical layer to a tropical storm. Atmos. Chem. Phys., 10, 10 803–10 827.
Neelin, J. D., O. Peters, and K. Hales, 2009: The transition to strong convection. J. Atmos. Sci., 66, 2367–2384.
Nicholls, M. E., and M. T. Montgomery, 2013: An examination of two pathways to tropical cyclogenesis occurring in idealized simulations with a cloud-resolving numerical model. Atmos. Chem. Phys., 13, 5999–6022.
Nolan, D. S., 2007: What is the trigger for tropical cyclogenesis? Aust. Meteor. Mag., 56, 241–266.
Ooyama, K. V., 1982: Conceptual evolution of the theory and modeling of the tropical cyclone. J. Meteor. Soc. Japan, 60, 369–379.
Peters, O., and J. D. Neelin, 2006: Critical phenomena in atmospheric precipitation. Nat. Phys., 2, 393–396.
Peters, O., J. D. Neelin, and S. W. Nesbitt, 2009: Mesoscale convective systems and critical clusters. J. Atmos. Sci., 66, 2913–2924.
Petty, G. W., 1999: Prevalence of precipitation from warm-topped clouds over eastern Asia and the western Pacific. J. Climate, 12, 220–229.
Rappin, E. D., and D. S. Nolan, 2012: The effect of vertical shear orientation on tropical cyclogenesis. Quart. J. Roy. Meteor. Soc., 138, 1035–1054.
Rappin, E. D., D. S. Nolan, and K. A. Emanuel, 2010: Thermodynamic control of tropical cyclogenesis in environments of radiative–convective equilibrium with shear. Quart. J. Roy. Meteor. Soc., 136, 1954–1971.
Raymond, D. J., 2000: Thermodynamic control of tropical rainfall. Quart. J. Roy. Meteor. Soc., 126, 889–898.
Raymond, D. J., and S. L. Sessions, 2007: Evolution of convection during tropical cyclogenesis. Geophys. Res. Lett., 34, L06811, doi:10.1029/2006GL028607.
Raymond, D. J., C. Lopez-Carrillo, and L. Lopez Cavazos, 1998: Case-studies of developing east Pacific easterly waves. Quart. J. Roy. Meteor. Soc., 124, 2005–2034.
Raymond, D. J., S. L. Sessions, and Z. Fuchs, 2007: A theory for the spinup of tropical depressions. Quart. J. Roy. Meteor. Soc., 133, 1743–1754.
Raymond, D. J., S. L. Sessions, and C. López Carrillo, 2011: Thermodynamics of tropical cyclogenesis in the northwest Pacific. J. Geophys. Res., 116, D18101, doi:10.1029/2011JD015624.
Reasor, P. D., M. D. Eastin, and J. F. Gamache, 2009: Rapidly intensifying Hurricane Guillermo (1997). Part I: Low-wavenumber structure and evolution. Mon. Wea. Rev., 137, 603–631.
Redelsperger, J.-L., D. B. Parsons, and F. Guichard, 2002: Recovery processes and factors limiting cloud-top height following the arrival of a dry intrusion observed during TOGA COARE. J. Atmos. Sci., 59, 2438–2457.
Riehl, H., and J. S. Malkus, 1958: On the heat balance in the equatorial trough zone. Geophysica, 6, 503–538.
Rozoff, C. M., W. H. Schubert, B. D. McNoldy, and J. P. Kossin, 2006: Rapid filamentation zones in intense tropical cyclones. J. Atmos. Sci., 63, 325–340.
Schumacher, C., P. E. Ciesielski, and M. H. Zhang, 2008: Tropical cloud heating profiles: Analysis from KWAJEX. Mon. Wea. Rev., 136, 4289–4300.
Sherwood, S. C., and R. Wahrlich, 1999: Observed evolution of tropical deep convective events and their environment. Mon. Wea. Rev., 127, 1777–1795.
Skamarock, W. C., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN–475+STR, 113 pp.
Smith, R. K., and M. T. Montgomery, 2012: Observations of the convective environment in developing and non-developing tropical disturbances. Quart. J. Roy. Meteor. Soc., 138, 1721–1739.
Stechmann, S., and J. D. Neelin, 2011: A stochastic model for the transition to strong convection. J. Atmos. Sci., 68, 2955–2970.
Stephens, G. L., and Coauthors, 2002: The CloudSat mission and the A-Train: A new dimension to space-based observations of clouds and precipitation. Bull. Amer. Meteor. Soc., 83, 1771–1790.
Takayabu, Y. N., S. Shige, W.-K. Tao, and N. Hirota, 2010: Shallow and deep latent heating modes over tropical oceans observed with TRMM spectral latent heating data. J. Climate, 23, 2030–2046.
Takemi, T., O. Hirayama, and C. Liu, 2004: Factors responsible for the vertical development of tropical oceanic cumulus convection. Geophys. Res. Lett., 31, L11109, doi:10.1029/2004GL020225.
Tory, K. J., and W. M. Frank, 2010: Tropical cyclone formation. Global Perspectives on Tropical Cyclones, 2nd ed., J. Chan and J. D. Kepert, Eds., World Scientific, 55–92.
Tory, K. J., M. T. Montgomery, and N. E. Davidson, 2006: Prediction and diagnosis of tropical cyclone formation in an NWP system. Part I: The critical role of vortex enhancement in deep convection. J. Atmos. Sci., 63, 3077–3090.
Waite, M. L., and B. Khouider, 2010: The deepening of tropical convection by congestus preconditioning. J. Atmos. Sci., 67, 2601–2615.
Wang, Z., 2012: Thermodynamic aspects of tropical cyclone formation. J. Atmos. Sci., 69, 2433–2451.
Wang, Z., 2014: Characteristics of convective processes and vertical vorticity from the tropical wave to the tropical cyclone stage in the high-resolution numerical model simulations of Tropical Cyclone Fay (2008). J. Atmos. Sci., 71, 896–915.
Wang, Z., M. T. Montgomery, and T. J. Dunkerton, 2010a: Genesis of pre-Hurricane Felix (2007). Part I: The role of the easterly wave critical layer. J. Atmos. Sci., 67, 1711–1729.
Wang, Z., M. T. Montgomery, and T. J. Dunkerton, 2010b: Genesis of pre-Hurricane Felix (2007). Part II: Warm core formation, precipitation evolution, and predictability. J. Atmos. Sci., 67, 1730–1744.
Wissmeier, U., and R. K. Smith, 2011: Tropical-cyclone convection: The effects of ambient vertical vorticity. Quart. J. Roy. Meteor. Soc., 137, 845–857.
Wu, Z., 2003: A shallow CISK, deep equilibrium mechanism for the interaction between large-scale convection and large-scale circulation in the tropics. J. Atmos. Sci., 60, 377–392.
Zehr, R. M., 1992: Tropical cyclogenesis in the western North Pacific. NOAA Tech. Rep. NESDIS 61, 181 pp.
Zuidema, P., 1998: The 600–800-mb minimum in tropical cloudiness observed during TOGA COARE. J. Atmos. Sci., 55, 2220–2228.
A wave pouch is a meso-α-scale region of approximately closed Lagrangian circulation within the critical layer of a synoptic-scale wave. Recent studies (Dunkerton et al. 2009; Wang et al. 2010a,b; Montgomery et al. 2010) have shown that the wave pouch provides favorable environment for vorticity aggregation and protects a protovortex within from the generally hostile tropical environment (such as dry air and shear deformation).
Note that the percentage here is with respect to the total upward mass flux, which excludes downward mass flux. If the contribution is evaluated with respect to the net vertical mass flux, the percentage will be 2–3 times larger. A similar statement holds for the upward moisture flux and condensation.
The PV production is a Galilean invariant and independent of the translation speed of the coordinate system.
Both scenarios were simulated by Nicholls and Montgomery (2013) under a variety of pregenesis conditions, and the SSCV was noted in a minority of cases. The SSCV is distinguished by its remarkably small radius (a few kilometers) and antecedent midlevel vortex, which is in some ways reminiscent of tornadogenesis.
The separation of large-scale ascent/convergence from the local convectively driven ascent/convergence is a tricky issue. Wang (2012) showed that the azimuthally averaged transverse circulation associated with a wave pouch is broadly consistent with the balanced response to diabatic heating derived using the Sawyer–Eliassen equation. The mean vertical motion near the pouch center [see Fig. 7a in Wang (2014)] thus is more likely driven by the local convective heating than by a large-scale forcing.