Compared to the eastern Pacific and eastern Atlantic, tropical cyclogenesis associated with breakdown of the intertropical convergence zone (ITCZ) was less documented over the western North Pacific (WNP). A typical case in boreal summer 2006 over the WNP, in which tropical cyclogenesis is induced by ITCZ breakdown in association with synoptic-scale wave train (SWT), is examined through observation analysis and numerical simulations. Observational analysis displays that a northwest–southeast-oriented SWT developed and propagated northwestward several days prior to ITCZ breakdown. Furthermore, the comparisons of simulation results reveal the role of the SWT in ITCZ breakdown. On one hand, the SWT within the ITCZ region is conducive to break down the ITCZ by enhancing the potential vorticity (PV), strengthening the meridional PV gradient and producing an evident sign reversal of PV gradient. On the other hand, the anomalous anticyclonic circulation related to the SWT north of ITCZ can reduce the meridional scale of PV band and thus increase the meridional PV gradient to accelerate ITCZ breakdown.
On the synoptic scale, the intertropical convergence zone (ITCZ) is observed to occasionally undulate and break down into several tropical disturbances (DBs), which is referred to as ITCZ breakdown (Guinn and Schubert, 1993, hereafter GS93; Nieto Ferreira and Schubert, 1997, hereafter NFS97). Following the ITCZ breakdown, tropical cyclones (TCs) are likely to form within the favorable large-scale environment (Agee, 1972). The mechanisms by which the ITCZ breaks down and subsequently facilitates the TC formation involve interior and exterior dynamic processes. On one hand, a low-level positive potential vorticity (PV) anomaly produced by ITCZ convection tends to form a reversal of the meridional PV gradient, which may satisfy the necessary condition for combined barotropic and baroclinic instability (Charney and Stern, 1962), resulting in ITCZ breakdown into a series of DBs (Charney, 1963; Hack et al., 1989; GS93; NFS97; Wang and Magnusdottir, 2005, hereafter WM05). NFS97 and WM05 simulated the ITCZ breakdown with different idealized models. The former successfully simulated the breakdown processes of different ITCZ shapes in a nonlinear shallow-water model, whereas the latter examined ITCZ breakdowns associated with the shallow and deep convections in three-dimensional flows using a primitive equation model. On the other hand, ITCZ breakdown over the eastern Pacific can also be triggered by interacting with westward-propagating disturbances that either originate from the Atlantic (Avila and Clark, 1989) or are generated by the interaction of the flow with the terrain (Zehnder et al., 1999; Wang and Magnusdottir, 2006, hereafter WM06).
Although the two aforementioned mechanisms produce the same number of ITCZ breakdown events over the eastern Pacific (WM06), the ITCZ breakdown triggered by external synoptic disturbances is more effective in producing named TCs compared to the ITCZ breakdown caused by its own internal dynamic process, because these external disturbances may provide extra environmental vorticity which is favorable for TC formation.
ITCZ breakdown events have been studied observationally in the eastern Pacific and eastern Atlantic from a statistical viewpoint (Dickinson and Molinari, 2000; WM06). However, so far there is no systematic investigation of ITCZ breakdown over the western North Pacific (WNP), which is to some extent attributed to poorly defined structure of ITCZ over the WNP in contrast to other ocean basins (Sobel and Bretherton, 1999, hereafter SB99; Krouse and Sobel, 2010). SB99 did not document any sign reversal of absolute vorticity gradient in the tropical Pacific Ocean using the 2.5°× 2.5° gridded reanalysis data, whereas WM05 and WM06 found that the 1.1° × 1.1° horizontal resolution is required to produce strong zonal wind shear to initiate the breakdown process. In addition, one salient feature over the WNP is the frequent occurrence of synoptic-scale wave trains (SWT) during the summertime (Fu et al., 2007; Chen and Huang 2009). The SWT is oriented in a northwest–southeast direction and generally propagates northwestward with a period of 3–8 days (Lau and Lau, 1990, hereafter LL90; Li, 2006). The origin of SWT is related to barotropic instability of the tropical easterly flow (Nitta and Yanai, 1969), equatorial-to-off-equatorial transition of mixed Rossby-gravity waves (Takayabu and Nitta, 1993), Rossby wave energy dispersion of a preexisting mature TC (Holland, 1995; Li and Fu, 2006), and scale contraction and energy accumulation of easterly waves (Kuo et al., 2001). Therefore, several intriguing questions naturally arise: does the ITCZ breakdown lead to TC formation over the WNP? Can the presence of SWT be conducive to the ITCZ breakdown and the subsequent development of the resultant disturbances? What is the role of the SWT in the ITCZ breakdown? To address these questions, this study examines a case of tropical cyclogenesis associated with ITCZ breakdown over the WNP and further investigates the role of SWT in the ITCZ breakdown using numerical simulation.
A case in August of 2006 provides a direct observational evidence of an ITCZ breakdown over the WNP. Figure 1 shows the evolution of infrared images (left), unfiltered PV and 3- to 8-day filtered wind fields at 850 hPa (right) obtained from the 1°×1° NCEP Final (FNL) Operational Global Analysis data. At 1800 UTC 10 August, an elongated and slightly northeast–southwest-tilted cloud band was discerned and stretched over nearly 30° in longitude (Figure 1(a)). Corresponding to the satellite image, a long PV strip appeared between 10°N and 20°N (Figure 1(b)). Subsequently, the cloud band and corresponding PV strip became unstable at the eastern end (Figure 1(c)–(f)), and gradually rolled up to form a comma-shaped pattern poleward of the ITCZ, indicative of an ITCZ breakdown (Figure 1(g)–(h)). Meanwhile, Joint Typhoon Warning Center (JTWC) issued the first warning of a TC named Wukong located at 20.4°N, 141.4°E (labeled as W) at 0600 UTC 12 August. The remnant of ITCZ was observed to further break down into several DBs, and one of them developed into the TC Sonamu (labeled as S) at 17°N, 126.5°E at 0600 UTC 13 August.
Several days prior to the TC Wukong genesis, a well-defined southeast–northwestward propagating synoptic wave train was present over the WNP (Figure 1(b)), which had alternating cyclonic and anticyclonic circulations and covered a broad region between 5°–30°N, 120°–160°E (Figure 1(d)). In addition, the cyclonic circulation in the SWT embedded with the large-scale environmental circulation associated with the ITCZ was tilted southwest–northeast, indicating that the eddy flow covariance () is positive (Figure 4(b)). In the presence of a large-scale cyclonic shear of mean flow caused by ITCZ (), the eddy disturbances tended to amplify through the barotropic conversion from mean to eddy kinetic energy (LL90). The anomalous northeasterly flows on the western side of DB Wukong around 20°N, 135°E penetrated the ITCZ band and rolled cyclonically up PV strip in the eastern part of ITCZ into a concentrated PV volume (Figure 1(f)). Afterwards, the eastern PV rolling-up from the PV horizontal strip was strengthened and evolved eventually into the TC Wukong (Figure 1(h)). Meanwhile, it is noteworthy that the formation of TC Sonamu is linked closely to a remnant vortex from the ITCZ breakdown in the western side of ITCZ and TC Sonamu has no direct relation to the tilted wave train. Figure 2 depicts the evolution of meridional wind associated with the SWT along the northwest–southeast orientation. The wave train has a dominant wavelength of about 2000–2500 km and a northwestward propagation of 5 m s−1, consistent with the previous study (LL90). It is noteworthy that the group velocity seems to be zero in Figure 2, suggesting that the wave train does not directly contribute to the TC formation via southeastward energy dispersion as emphasized in some previous studies (Li, 2006; Li and Fu, 2006), but it may play an important role in promoting the ITCZ breakdown.
To summarize, the observational results evidently illustrate this ITCZ breakdown in association with the SWT, which induces two TC formations over the WNP. In order to examine the effect of SWT on the ITCZ breakdown, a model will be employed to simulate the different ITCZ behaviors with or without the SWT in the next section.
3. Results from model simulations
3.1. Model and experiment design
The numerical experiments are conducted with the Advanced Research Weather Research Forecasting (ARW-WRF) model version 3.1 developed by the National Center for Atmospheric Research. Because the ITCZ experiences breakdown on a synoptic scale, the numerical simulations include only one domain with the horizontal resolution of 30 km. The WRF model is initialized at 0000 UTC 9 August and 120-h integration is carried out. The 6-h NCEP FNL data are used as the initial and lateral boundary conditions. The sea surface temperature (SST) data are derived from the NCEP analysis (at 0.5° × 0.5°), and the SST is fixed during the model integration. Three numerical experiments are designed to compare different evolutions of the ITCZ. In the control run (hereafter as CTL), the initial fields from FNL are not adjusted. In one sensitivity experiment (hereafter as NOSWT-ALL), the Lanczos filter is performed on the whole dynamical and thermodynamical fields to remove the 3- to 8-day filtered signals from the model initial and lateral boundary fields (Duchon, 1979). Because the removal of 3- to 8-day anomalies from the whole fields can modify the ITCZ structure, the second sensitivity experiment (hereafter as NOSWT-PART) is conducted to assess the impacts of 3- to 8-day filtered signal outside of ITCZ region on ITCZ breakdown, in which 3- to 8-day signal is kept intact within the ITCZ region (0–15°N, 105°–160°E) but removed outside this region. The comparisons of these runs show that, in the NOSWT-ALL run the ITCZ has a relative weak strength and a decreased zonal extent at the initial time compared to that in the CTL run (Figure 3(a) and (b)). While the ITCZ structure remains unchanged in the NOSWT-PART run and the wind and vorticity fields outside of ITCZ region become weak (Figure 3(a), (c)).
Figure 3 shows the evolutions of PV and wind fields at 850 hPa in the three experiments during the period of 1800 UTC 10 August-0600 UTC 12 August at the interval of 12 h. Although the detailed structure in the CTL run differs somewhat from the observed, the simulation captures well the ITCZ breakdown in agreement with the observation as shown in Figure 1. In particular, the ITCZ breaks down into two organized vortex at 1800 UTC 11 August (Figure 3(j)). The eastern one evolves into TC Wukong at 0600 UTC 12 August, the minimum sea level pressure (SLP) of which is centered at 20.1°N, 141.5°E close to the observed position of 20.4°N, 141.4°E. And the other DB developing into the TC Sonamu is also reproduced well (Figure 3(m)). A careful comparison indicates that the simulated breakdown occurs about 6 h earlier than the observed, which may be attributed to the stronger simulated zonal wind shear around the PV strip that can accelerate the ITCZ breakdown similar to the NFS97 idealized simulation (Figure 3(d)).
By comparison with the CTL run, the ITCZ strip does not experience an evident breakdown in the NOSWT-ALL run. Several small-scale PV patches are scattered within the PV strip which has a broader meridional scale than that in the CTL at 1800 UTC 10 August (Figure 3(d) and (e)). Subsequently, PV intensity is increased due to the merging of small-scale PV patches, and ultimately the ITCZ-like PV strip also forms (Figure 3(h) and (k)). However, an elliptic gyre-like structure emerges without experiencing an ITCZ breakdown at 0600 UTC 12 August (Figure 3(n)), which is similar to the monsoon gyre (MG) defined by Chen et al. (2004). This gyre evolves eventually into a large-sized cyclone with the minimum SLP comparable to that of TC Wukong in the CTL, but the center position of the minimum SLP is displaced northwestward (Figure 3(m) and (n)). Prior to the ITCZ breakdown, one salient difference is that the PV strip in the CTL run exhibits a wavy characteristic with a strong anticyclonic circulation located to the north (Figure 3(g)), suggesting that the ITCZ is undergoing breakdown and forming a filament structure. In contrast, the PV strip in the NOSWT-ALL run still remains a single cyclonic center, accompanied by a zonally elongated vorticity band (Figure 3(h)).
One may doubt that the ITCZ inherent synoptic anomalies at the initial time can impact the ITCZ breakdown by altering the ITCZ structure. The NOSWT-PART run can quantify the relative roles of the SWT cell within and outside of the ITCZ domain in ITCZ breakdown. Different from the wavy ITCZ structure in the CTL, the ITCZ almost maintains a zonally elongated PV strip until 0600 UTC 12 August (Figure 3(f), (i), (l) and (o)). Subsequently, the eastern part of ITCZ also experiences a breakdown, but at the time about 48 h later than that in the CTL run. The sensitivity experiments indicate that the SWT cell in the eastern part of ITCZ can play a decisive role in the ITCZ breakdown. In conjunction with the SWT cell outside of ITCZ domain, the ITCZ breakdown can be accelerated through the increased PV gradient caused by the PV enhancement due to the SWT within the ITCZ region and the contraction of meridional extent due to the SWT north of the ITCZ discussed below.
In order to explain the ITCZ breakdown related to dynamic instability, the PV meridional gradients averaged between 110°E and 160°E in the three runs at 1800 UTC 9 August are shown in Figure 4. The increased magnitude and evident sign reversal of PV meridional gradient in the CTL run are found (Figure 4(a)), suggesting that the presence of the SWT is more favorable for the development of barotropic instability compared with the other two experiments. The presence of the SWT can contribute to the following processes. On one hand, the cyclonic flow associated with the SWT between 135°–160°E can enhance the positive PV of the ITCZ increasing the possibility of barotropic instability because an abrupt reversal of PV meridional gradient is likely to occur due to the decreased PV on the southern and northern flanks of ITCZ strip. On the other hand, the anticyclonic circulation in association with the SWT north of ITCZ can reduce the meridional scale of PV band, which can speed up ITCZ breakdown by steeping the meridional PV gradient. As shown in the CTL run (Figure 3(g)), when the SWT propagates across the eastern part of ITCZ, the anomalous northerly winds associated with the synoptic cyclonic circulation in the central part of ITCZ are increased and penetrate the PV strip to help form a wavy pattern resembling the structure of midlatitude baroclinic wave (Thorncroft et al., 1993). Accompanied by the roll-up and the increase of PV volume in the CTL run, the cyclonic circulation is intensified such that the wind-induced surface heat exchange (WISHE; Emanuel, 1986) mechanism begins to spin up the vortex by self-sustaining and positive feedback process, and accelerate the TC development.
4. Summary and discussion
Previous studies on the ITCZ breakdown are emphasized over the eastern Pacific (WM06) or employed with idealized models (GS93; NFS97). In this study, a typical case in boreal summer 2006 is studied through the observational and modeling examinations, in which two TC formations are associated with the ITCZ breakdown. The observational results show that the ITCZ breakdown in association with the SWT and the related tropical cyclogenesis can be well detected from the satellite images and the high-resolution data over the WNP. Furthermore, three numerical experiments with the ARW-WRF model confirm that the SWT plays a crucial role in the ITCZ breakdown. The SWT cell within the ITCZ region can increase the vorticity anomalies, causing the enhancement of the positive PV of the ITCZ to facilitate dynamic instability. On the other hand, the enhanced anticyclonic circulation north of the ITCZ can assist in condensing the PV strip which can lead to steepening of the meridional PV gradient favorable for the ITCZ breakdown.
It should be noted that this study focuses only on the role of SWT in the ITCZ breakdown. In this case, a coherent northwestward propagating wave train had emerged near the equator about one week prior to the ITCZ breakdown. In fact, the ITCZ background environment can also exert an influence on the SWT circulation through their interaction. Therefore, more elaborate examination of relative contribution from the ITCZ breakdown and the synoptic wave to the TC genesis is desirable in the future study. Besides, in the NOSWT-ALL run without the ITCZ breakdown, the scattered PV patches within the ITCZ band may be concentrated to spin up a large-scale organized monsoon depression like a MG structure, i.e. also likely to develop into a large-sized TC. What mechanism is responsible for this process also needs to be investigated in the further study.
The authors would like to give thanks to Drs. Xuyang Ge and Ping Huang and two anonymous reviewers for their helpful suggestions. This work was supported jointly by the National Natural Science Foundation of China (Grant 41275001) and the Special Scientific Research project for Public Welfare (Grant GYHY201006021).