1. Introduction
The western North Pacific (WNP) in boreal summer is characterized by multispatiotemporal scale circulation and convection (Holland 1995; Hsu 2005), for example, the intraseasonal oscillation (ISO) and tropical cyclones (TC). The well-explored 30–60-day intraseasonal oscillation is noted to propagate northward and northwestward in the South China and Philippine Seas (e.g., Lau and Chan 1986; Nitta 1987; Chen et al. 1988; Wang and Rui 1990; Hsu and Weng 2001; Tsou et al. 2005). Some of these meridionally propagating intraseasonal perturbations could be triggered by the eastward-propagating Madden–Julian oscillation (MJO) along the equator (Madden and Julian 1971; Wang and Xie 1997; Hsu 2005). Hsu and Weng (2001) documented the evolution of the northwestward propagating 30–60-day disturbances and proposed a convection–circulation mechanism to explain this northwestward propagation tendency. The frictional convergence near the center of the cyclonic circulation located to the northwest of the deep convection provided a favorable condition for the system to move northwestward. Another branch of interest for the ISO resides in the higher frequency (e.g., 10–25 or 7–30 day) band of the ISO. The higher-frequency ISO, which often propagated westward in the tropical WNP, was often explained as westward-propagating tropical waves (e.g., Chen and Chen 1995; Chen and Weng 1999; Straub and Kiladis 2003; Hsu 2005).
Tropical cyclones are the most energetic atmospheric phenomena in this region. Many studies have reported large-scale circulation modulation (e.g., the monsoon trough, gyre, ISO) of the TCs (e.g., Nakazawa 1986; Liebmann et al. 1994; Harr and Elsberry 1995; Chen et al. 2000; Maloney and Dickinson 2003; Camargo et al. 2007a,b). Both the genesis and movement of TCs are affected by the large-scale circulation and often appear in clusters (e.g., Liebmann et al. 1994; Gray 1998; Elsberry 2004). Holland (1995) pointed out that the confluence zone provided an excellent environment for tropical cyclone formation in the summer season in the WNP. Harr and Elsberry (1995) and Camargo et al. (2007a,b) classified TC tracks into different clusters and found that these TC clusters were closely associated with different configurations of the large-scale circulation (e.g., monsoon trough and anticyclonic ridge).
Nakazawa (1986) found that the TC tended to occur during the convective phase of the ISO during the 1979 summer. Based on multiyear data, Liebmann et al. (1994) documented this relationship between the MJO and the TCs over the Indian Ocean and the WNP. They confirmed that the TCs preferentially occurred during the convective phase of the MJO and clustered around the cyclonic vorticity and convergence anomalies that appeared poleward and westward of the large-scale convective anomaly along the equator. Maloney and Dickinson (2003) investigated the modulating effect of the MJO on tropical-depression-type disturbances. Their findings indicated larger synoptic (2.5–12.5 day) variance and more favorable conditions for the growth of the tropical disturbances in the westerly phase of the MJO than in the easterly phase.
Other westward- or northwestward-propagating wavelike perturbations were also observed to be closely associated with TCs. Ko and Hsu (2006, hereafter referred to as KH06) identified a 7–30-day wave pattern propagating north-northwestward from the northeast of Papua New Guinea to the oceanic area between Taiwan and Japan. When the cyclonic circulation of the wave pattern moved into this area, more than 70% of the wave patterns were associated with at least one recurving tropical cyclone. This wave pattern recurred at a 7–30-day time scale during mid to late summer (July–August). Moreover, the cyclonic/anticyclonic phase of this wave pattern was also related to the fluctuation of the monsoon trough/subtropical high system in the WNP. The cyclonic circulation of the wave pattern created a favorable region for recurving tropical cyclones to form, develop, and move along with this wave pattern. Another similar wavelike pattern moving toward the South China Sea on a shorter time scale (3–8 days) was also identified by Lau and Lau (1990) and Chang et al. (1996). Straub and Kiladis (2003) suggested that TCs often developed from the high-frequency, westward-propagating mixed Rossby gravity–tropical depression-type disturbances. Although these wavelike patterns may modulate the TCs, it has also been proposed that TCs may excite waves at the spatial scale of the synoptic-scale circulation associated with the cyclone, through Rossby wave energy dispersion (Holland 1995; Sobel and Bretherton 1999; Li and Fu 2006; Li et al. 2006; Krouse et al. 2008).
Most of the previous studies focused on the links between one type of large-scale circulation and TCs. Less attention has been paid to the possible multiscale links among the ISO, the submonthly wave pattern, and TCs. This study extends the KH06 study and demonstrates the existence of such a triscale linkage in the tropical WNP. The objectives of this study are 1) to examine the evolution of the submonthly wave patterns during different ISO phases, 2) to show the changes in TC statistics during different ISO phases, and 3) to study the possible effect of ISO on TCs and the submonthly wave pattern. Section 2 describes the data and analysis procedures. The TC case selection criteria and statistics are given in section 3. Sections 4 and 5 discuss the composites of the ISO and the submonthly wave pattern along with their relationship with the tropical cyclones. Concluding remarks are presented in section 6.
2. Data and analysis procedures
The circulation data used in this study were extracted from the 40-yr ECMWF Re-Analysis (ERA-40) (Uppala et al. 2005). The ERA-40 contains 6-hourly (0000, 0600, 1200, and 1800 UTC) temperature, humidity, horizontal winds, vertical p velocity, and geopotential on a 2.5° × 2.5° latitude–longitude grid. The best-track tropical cyclone data were obtained from the Joint Typhoon Warning Center (JTWC) at Guam. This study analyzed the 23-yr data for June–October (JJASO) 1979–2001.
The moving spectral method described by KH06 was used in this study to identify the 7–30-day signals in the oceanic region between Taiwan and Japan. This method, which has been applied in several previous studies (e.g., Hurrell and van Loon 1997; KH06), was designed to detect the dominant periodicity and the temporal evolution of the spectra in a manner similar to computing running means. In this study, the 60-day period was chosen as a “working window” to produce a series of spectra calculated by shifting the 60-day time series to the following time step. A Butterworth bandpassed filter (Kaylor 1977; Hamming 1989), following Hsu and Weng (2001) and KH06, was applied to extract the 7–30 (submonthly) and 30–80 day (intraseasonal) fluctuations. The composite technique as in KH06 was then used to reveal the characteristics of the circulation on these two frequency bands.
Since TCs are usually embedded in the cyclonic circulation of the submonthly wave pattern in the chosen events, one may wonder how much of the cyclonic circulation is contributed by the TCs. There is no clear answer to this question due to the nonlinear nature of the phenomenon. On the other hand, an artificial TC-removing technique can be applied to assess the potential contribution from TCs. Hsu et al. (2008) constructed a dataset by removing the TCs from the ECMWF analyses to assess the potential contribution of the TCs to the low-frequency variability, based on the technique developed by Kurihara et al. (1993, 1995) and Wu et al. (2002) that has proved to be useful in improving TC track simulation and forecast. The methodology is briefly described as follows. [Readers are referred to Hsu et al. (2008) and Kurihara et al. (1993, 1995) for details.] The wind fields were decomposed into the basic and disturbance fields. TC winds were then detected and removed from the disturbance field to create the non-TC component, which was then added back to the basic field to form the environmental flow. Hsu et al. demonstrated the effectiveness of the methodology in decomposing the total wind into TC and non-TC components. This study compares the results derived from the original and TC-removed wind fields to assess the contribution of the TC winds to the composited wind field.
3. Mean state characteristics and case selection criteria
The 23-yr climatological mean 850-hPa flow pattern and the corresponding horizontal moisture flux (and convergence) in the WNP from July to September are shown in Fig. 1a. A monsoon trough extends southeastward from southwestern China toward 15°N, 140°E. The moisture flux convergence area extends farther southeastward to around 5°N, 170°E. The wind speed variance at 850 hPa is shown in Fig. 1b to identify the major fluctuating regions. The variance maximum north of 35°N was mainly associated with the midlatitude synoptic disturbances, while a smaller variance in the central South China Sea was associated with the ISO. The latter one will be discussed later. The variance maximum located over the oceanic region between Taiwan and Japan is the focus of this study. As demonstrated in KH06, the high variance in this region could be attributed to the repeated passages of the 7–30-day wave pattern, which propagated north-northwestward in the Philippine Sea, and the recurving TCs, which were often embedded in the anomalously cyclonic wave pattern circulation. The area-averaged wind speed in this high-variance region (20°–35°N, 125°–140°E) between Taiwan and Japan is defined as an index for the following analysis.
The moving spectrum analysis results applied to the index are presented in Fig. 2, which shows only energy power that exceeds the 99.9% confidence limit. A spectral maximum between 7 and 30 days emerged in early July and persisted through late September. The maximum power was centered around 12 days in July and August and shifted slightly to a shorter period ∼10 days in September. It was not until mid-October when another maximum power period ∼5.5–6.7 days did emerge. The aforementioned results indicated that the 7–30-day signals were the dominant fluctuations in the oceanic region between Taiwan and Japan. Since the 7–30-day signals occurred mostly in the July–September period, the following analysis will focus on this period.
The ISO signal distribution can be seen from the variance analysis of the 850-hPa zonal wind. Figure 3 shows the variance ratios (30–80-day filtered/total) for the zonal wind. The maximum variance ratios (exceeding 45%) were located in the region (5°–15°N, 100°–125°E), which was also the area where the strong southwesterly wind was often observed along the southern flank of the monsoon trough (Fig. 1a). Two areas where the ISO contributed more than 30% of total variance were identified. The western one extended eastward from the eastern Indian Ocean to 140°E along the latitudinal band between the equator and 15°N. The eastern one extended eastward from 140°E to the date line roughly between 15° and 25°N. The western one happened to coincide with the southwesterly in the southern flank of the monsoon trough, while the eastern one resided in the southern flank of the Pacific anticyclonic ridge. These two features appeared to reflect the intraseasonal fluctuation of the monsoon trough and the anticyclonic ridge where the southeasterly prevailed. The box (5°–15°N, 100°–125°E) shown in Fig. 3 was chosen as a base region to calculate the area-averaged zonal wind speed, defined as the index to represent the fluctuation of the ISO in the tropical WNP. A time series comparison (not shown) indicated that this ISO index was qualitatively similar to the MJO index proposed by Wheeler and Hendon (2004), which was defined based on the eastward-propagating MJO along the equator. Since the ISO in this area often occurred north of the equator in the summer and not all the ISO cycles in the WNP were preceded by the MJO originating in the Indian Ocean and propagating eastward along the equator (e.g., Wang and Rui 1990; Hsu and Weng 2001), this local index was chosen to better represent the ISO in the WNP.
Cases were chosen for the composites to reveal the essential features of the submonthly and ISO events. The case selection criteria were twofold: one for the submonthly cases and the other for the ISO events. The submonthly cases were selected based on the 7–30-day filtered time series averaged over the base region (20°–35°N, 125°–140°E) shown in Fig. 1b. The submonthly cases were selected when the positive anomalous maxima were greater than 1 m s−1, which was equivalent to 0.75 standard deviation of those 23-yr July–September time series. A submonthly/TC case was further selected when at least one TC appeared in the base region (with a 1.5° latitude–longitude buffer zone just outside the edge) between Day −1 and Day +1 of a selected submonthly case. It has been reported (e.g., Maloney and Dickinson 2003) that the synoptic perturbations behave differently in different phases of the MJO. To examine the possible effect that the ISO had on the submonthly wave pattern and TCs, the submonthly/TC cases were further classified into two categories: ISO easterly and westerly phases. A submonthly/TC case was classified as a case in the westerly (easterly) phase if the anomalous 30–80 day filtered zonal wind averaged over the ISO base region (5°–15°N, 100°–125°E) at the occurrence time of the submonthly case was greater than 1.5 m s−1 (less than −1.5 m s−1). The ±1.5 m s−1 thresholds were approximately ±0.5 standard deviation of the 30–80-day filtered zonal wind time series. Figure 4 shows an example to illustrate how the cases were selected. The time series for 1983 in which a strong ISO and active submonthly cases (two in westerly and one in easterly) were seen. The horizontal bars represented the time when the TCs passed through the base region. The submonthly cases had to be strong enough and were associated with TCs to be selected. Additionally, in order to be selected as a westerly (or easterly) case, the peak time of each submonthly case had to occur when the corresponding 30–80-day filtered zonal wind exceeded the thresholds. Within those 23 years, the number of selected cases for each year ranged from 1 (1988) to 5 (1990).
Before discussing the composite results, it is informative to describe the statistic characteristics of the selected cases. There were 124 submonthly cases of which 97 (78.23%) were related to TCs (Fig. 5a). In other words, at least one TC was associated with each of those 97 submonthly cases. The remaining 27 (21.77%) cases were not related to any TCs, and 14 (51.85%) of them were found in July (Fig. 5b). Most of the no-TC cases occurred before mid-July and appeared to be related to the mei-yu frontal system (KH06). The remaining 13 no-TC cases were in August (5) and September (8).
There were 401 TCs in the western North Pacific during July–September 1979–2001. Among these TCs, 108 (26.93%) were related to those 97 submonthly cases (Fig. 5c). Additionally, these TCs were usually recurving TCs (KH06; Schnat et al. 1998) that passed through the base region near the east coast of China. Another 77 TCs (19.2%) passed through the base region but were not related to the selected submonthly cases (or the cases were too weak to be selected). The remaining 216 TCs (53.87%) occurred absolutely outside the base region (including the buffer zone) and most of them were straight moving cyclones related to 3–8-day synoptic waves (Lau and Lau 1990; Chang et al. 1996; KH06). Overall, about half of the TCs in the western North Pacific passed through the base region, and more than half of them were associated with the chosen submonthly cases.
The total numbers of the westerly and easterly ISO events were 28 and 23 (Fig. 5d), respectively. Among those 97 submonthly/TC cases, 40 (24) were in the westerly (easterly) phase. Additionally, there were 45 (27) TCs associated with those 40 submonthly cases in the westerly phase (24 submonthly cases in the easterly phase). On average, there were about 1.43 submonthly/TC cases for each westerly ISO event, and one for each easterly ISO event. This information indicated that the submonthly/TC activity was more frequent in the westerly phase than in the easterly phase.
4. ISO and submonthly wave pattern composites
The composite ISO evolution, based on the events associated with submonthly/TC cases, is presented in Figs. 6 and 7 for the westerly and easterly phases, respectively. In the westerly phase, a cyclonic center was located near 15°N, 125°E on Day −10 (Fig. 6). This cyclonic circulation moved northwestward and intensified. On Day 0, the cyclonic center was located south of Taiwan, resulting in a large area of strong southwesterly anomalies south of the cyclonic circulation. After reaching maximum phase, the cyclonic circulation moved farther northwestward toward southeastern China and decayed. The evolution and moving patterns of the ISO westerly phase were similar to that in Hsu and Weng (2001). In the easterly phase, there existed an anticyclonic circulation near 15°N, 130°E on Day −10 (Fig. 7). This anticyclonic circulation then moved slowly northward mainly in the South China Sea and became stronger, creating northeasterly anomalies in the ISO base region on Day 0. By Day +10, this anticyclonic circulation weakened quickly while reaching the northern South China Sea. The circulation pattern in the westerly phase was more coherent and stronger, as indicated by more wind vectors, shown only when they were statistically significant.
To demonstrate the links between the ISO and submonthly wave pattern, the composite streamfunction for the 7–30-day filtered anomalies during the ISO westerly and easterly phases are shown in Figs. 8 and 9, respectively. Also shown in the figures are the synoptic views of the ISO by adding the climatological mean (July–September averaged over 23 years) streamfunction to the intraseasonal anomalies, that is, adding Fig. 1a to Figs. 6 and 7. The evolution in Figs. 8 and 9 was contributed solely by Figs. 6 and 7. The purpose of adding the climatological mean was to clearly demonstrate how the intraseasonal signals affected the monsoon trough and anticyclonic ridge. The climatological flow across the ISO base region, which was located in the southern flank of the monsoon trough, was predominantly southwesterly. The cyclonic anomaly and southwesterly anomaly seen in the westerly phase indicated a strengthened southwesterly and monsoon trough, while the anticyclonic anomaly and northeasterly anomaly in the easterly phase indicated a weakened southwesterly and monsoon trough. A comparison between Fig. 1a and Figs. 8 and 9 demonstrated the ISO effect on the monsoon trough, as reported in previous studies.
The reason for plotting the submonthly wave pattern on the ISO synoptic flow is to isolate the slower ISO evolution as a background flow and study the possible effect of the ISO on the submonthly wave pattern. In the ISO westerly phase, a northwest–southeast wave pattern was embedded in the southeasterly and the confluent region between the monsoon trough and the subtropical anticyclone on Day −6. The wave pattern, with averaged wavelength ∼4000 km, continued moving toward the base region in the southeasterly region, and recurved northeastward toward the extratropics along the periphery of the subtropical anticyclone. When the wave pattern propagated within the background flow set up by the ISO, the background flow itself evolved at a much slower pace, characterized by the slow northward movement of the monsoon trough. However, the slow evolution of the ISO hardly altered the background flow on which the submonthly wave pattern propagated. The wave pattern in the westerly phase was located in the confluent zone, which appeared to be a favorable region for the wave pattern to grow and develop. The ISO in the westerly phase seemed to be related to the enhancement of the monsoon trough, which favored development of the submonthly wave pattern. This issue will be further explored in a later section.
In contrast to the westerly phase, the wave pattern in the easterly phase was much less coherent and exhibited an averaged wavelength of ∼5500 km. The wave pattern was located in an area between a southward-shifted subtropical anticyclone and a weak monsoon trough, which retreated to the northern South China Sea. Such a background configuration yielded a less favorable environment for the development of the wave pattern in the Philippine Sea. Compared to that in the easterly phase, the wave pattern in the westerly phase was better organized. Although not shown, there existed more statistically significant wind vectors of the submonthly wave patterns in the westerly phase than in the easterly phase.
The submonthly wave pattern shown in the present study (and in KH06) was distinctly different from the 3–8-day wave reported in previous studies (e.g., Lau and Lau 1990; Chang et al. 1996). As revealed above and below, the submonthly wave pattern propagated toward the east China Sea, accompanied mostly by recurving TCs, while the 3–8-day wave propagated toward the South China Sea and was accompanied mostly by the straight-moving TCs. In addition, the former exhibited a lower frequency and larger spatial scale than the latter.
5. Relationship with TCs
The close relationship between the submonthly wave pattern and TCs is demonstrated in Fig. 10, which presents the submonthly wave pattern with the superposition of TC tracks in the ISO westerly and easterly phases. The complete track of a selected TC was plotted, even though it might stay in the base region for just 1–2 days. More than 85% of the TCs associated with the wave pattern were recurving (87% for the westerly and 89% for the easterly). In comparison with the ISO westerly phase (Fig. 10a), the wave pattern in the ISO easterly phase (Fig. 10b) exhibited weaker strength and less coherent structure as well as fewer TCs associated with the wave pattern than its counterpart in the westerly phase. More coherent wave pattern and more recurving TC cases were found in the ISO westerly phase. Submonthly variance in the ISO westerly and easterly phases was also shown to reveal the propagating tendency of an active submonthly wave pattern. In the ISO westerly phase, a maximum variance area extended from the base region south-southeastward into the northeastern coast to Papua New Guinea and northeastward to the extratropical North Pacific. The area was also the region where TC tracks were clustered. This reflected the collocation of the propagation route of the submonthly wave pattern and the TC tracks. In comparison, the maximum variance area in the ISO easterly phase was confined in the vicinity of the base region and was not as consistent with TC tracks as in the ISO westerly phase. The wave pattern in the ISO easterly phase was weak and the strong signals seemed to appear mostly near the base region. This indicates that the wave pattern was a feature appearing predominantly in the ISO westerly phase. It is interesting to note that the anomalous flow pattern and the associated TC tracks in the ISO westerly phase resembled those in cluster A, defined by Camargo et al. (2007b), which occurred frequently during July–September (Camargo et al. 2007a). However, the flow pattern in the ISO easterly phase seemed to be a combination of clusters C, E, and G, defined by Camargo et al. (2007b), reflecting the large case-to-case variability in the ISO easterly phase.
To further investigate the links between the intraseasonal circulation anomalies and TCs, the composite 7–30-day streamfunction for both phases and the corresponding TC locations were shown in Fig. 11. A statistical test based on the Student’s t test was applied to all grids for each time frame of the composites. It was clearly seen that most of the maximum/minimum areas exceeded the 95% confidence level in the ISO westerly phase. However, in the ISO easterly phase, only areas near the base region passed the t test. In other words, there existed large case-to-case variability for the composite cases in the ISO easterly phase so that the composite pattern was weak and not well organized. The corresponding TC locations also showed that a more scattered pattern existed in the ISO easterly phase, whereas TCs were clustered in the cyclonic anomaly of the wave pattern in the ISO westerly phase. To study the possible effect of the ISO on the strength of TCs, those TCs with maximum wind speed greater than 64 kt (equivalent to hurricane stage) were marked as bigger dots on the composite maps. The ratio of stronger TCs increased from 43.6% (17/39) on Day −3 to 55.6% (25/45) on Day 0 in the ISO westerly phase; whereas in the ISO easterly phase, the ratios (8/21 = 38.1% to 13/27 = 48.1%) at both times were smaller. This result suggests that more TCs associated with the wave pattern in the ISO westerly phase tended to be stronger. The genesis of TCs in both phases was also examined. The genesis locations were marked on the Day −4.5 composites for both phases shown in Figs. 11c,f. The TC genesis tended to occur near the statistically significant cyclonic circulation in the ISO westerly phase, whereas the genesis in the ISO easterly occurred in a more scattered area, located across the cyclonic and anticyclonic anomalies. Further examination yielded that this area happened to be the climatological confluent zone (Fig. 1a) to the west of the subtropical high. It follows that the TC genesis might not necessarily be associated with the wave pattern in the ISO easterly phase.
The TC genesis in either ISO phase raised a question: How does the ISO and submonthly wave pattern affect tropical cyclogenesis? The horizontal moisture flux might provide some clues. Figure 12 shows the horizontal moisture flux and convergence in both ISO phases for the unfiltered fields, the 30–80-day filtered and 7–30-day filtered components averaged over the active periods (Day −10 to +10 for the unfiltered and 30–80-day filtered fields; Day −3 to +3 for the 7–30-day component). The unfiltered moisture flux convergence areas were in good agreement with the TC genesis locations. Moreover, in the ISO westerly phase, the enhanced unfiltered southwesterly flow supplied more moisture to the area where the submonthly cyclonic circulation in the early stage was located. Additionally, the TC genesis positions were collocated with the moisture flux convergence area in the ISO westerly phase. However, in the ISO easterly phase, the southwesterly flow and moisture flux convergence were weak, while TC genesis locations were scattered along the northern edge of the elongated convergence belt, which coincided with the climatological confluent region (Fig. 1a).
A comparison of the unfiltered moisture flux convergence with the intraseasonal and submonthly convergence anomalies yielded information about the relative contribution from the circulation on different time scales. The convergence anomaly corresponding to the intraseasonal signals, shown in Fig. 12, in the ISO westerly phase exhibited a maximum moisture flux convergence near the one seen in the unfiltered field. The submonthly wave pattern also contributed a moisture flux convergence region at a similar location. This result suggests that both the westerly phase ISO and submonthly wave pattern enhanced the probability of TC genesis in the Philippine Sea. However, this agreement failed to exist in the ISO easterly phase. Instead of a moisture flux convergence anomaly, a divergence anomaly existed in the ISO easterly phase. The convergence anomaly in the ISO easterly phase occurred only in the region near Taiwan, while the TC genesis occurred in a widespread region. Therefore, the submonthly wave pattern could be of help to the TC genesis only near the base region. On the whole, the TC genesis/tracks, ISO, and submonthly wave pattern were all in agreement in the ISO westerly phase, whereas in the ISO easterly phase, a close relationship among those three features was not found.
To delineate whether the cyclonic circulation in the submonthly wave pattern is simply the manifestation of TCs, a TC-removed routine was applied to the 850-hPa wind field of the ECMWF analysis to remove the TC wind from the global analysis (Hsu et al. 2008). Owing to the spatial limit of the dataset, only the area (5°S–50°N, 100°E–180°) is shown. The 7–30-day filtered flow pattern and the corresponding vorticity are shown in Fig. 13. A comparison between the original (Figs. 13a,b) and TC-removed (Figs. 13c,d) fields yielded a significant decrease (about half) in strength of the TC-removed wind and vorticity for the cyclonic circulation in which TCs were embedded. However, the wave patterns and the cyclonic circulation still existed in both phases. It is interesting to note that the submonthly cyclonic circulation in the ISO easterly phase almost disappeared after removing TCs, indicating a stronger influence of TCs on the submonthly wave pattern. This contrast was consistent with the fact that the wave pattern was less coherent in the ISO easterly phase. On the contrary, the wave pattern in the westerly phase was a robust feature that coexisted with TCs and was not simply a manifestation of TC circulation. However, the cause-and-effect relationship between the submonthly wave pattern and TCs is not clear in this study. While the submonthly wave pattern provided a favorable background for the TCs, the latter might enhance the wave pattern through Rossby wave energy dispersion, as proposed by several previous studies. Whether such a coupling exists needs further study. No matter what the mechanism is, the results presented here suggest that TCs and the large-scale wave pattern should be treated as one integrated system. A comparison (not shown) conducted for the ISO also indicated that removing TCs yielded a weaker intraseasonal cyclonic circulation but hardly altered the characteristics of the ISO.
6. Conclusions
In a previous study, KH06 identified a submonthly (7–30 day) wave pattern, which propagated north-northwestward from the southeastern Philippine Sea to the east China Sea. Recurving TCs were often embedded in the cyclonic circulation of the wave pattern. This study extended the KH06 study and found that the coupled wave–TC pattern exhibited distinct characteristics between westerly and easterly phases of the ISO in the tropical WNP. The ISO in the westerly phase provided a favorable background for the development of the wave–TC pattern, while the ISO in the easterly phase provided a less favorable environment. This ISO modulation likely resulted in the stronger and more coherent wave–TC pattern in the westerly phase than in the easterly phase.
The westerly phase ISO was characterized by strong 850-hPa westerly anomalies in the South China Sea and the Philippine Sea and was associated with a strong monsoon trough in the lower troposphere. On the other hand, the easterly phase ISO was characterized by an easterly anomaly in the same area: It was associated with a weak monsoon trough, which retreated to the northern South China Sea, and a southward-shifted Pacific subtropical ridge. The westerly phase ISO moved northwestward in the Philippine Sea, indicating a slow northwestward movement of an enhanced monsoon trough. The easterly phase ISO moved northward but mainly in the South China Sea, indicating the northward movement of a weaker monsoon trough.
In the ISO westerly phase, the submonthly wave pattern propagated north-northwestward along the southeasterly/confluent zone between the monsoon trough and the subtropical anticyclonic ridge. TCs were found to cluster near the cyclonic circulation of the wave pattern, move northwestward along with the wave pattern, and recurve northeastward after reaching the east China Sea. In the ISO easterly phase, the submonthly wave pattern was less coherent and varied among cases. As a result, a statistical significance test failed to support the existence of the wave pattern in the Philippine Sea except in the region close to the east China Sea. There were fewer and weaker TCs associated with the easterly phase submonthly wave pattern. In addition, TC genesis locations were scattered in the Philippine Sea along the climatologically confluent zone in the easterly phase, while they were more clustered in the cyclonic circulation of the wave pattern during the westerly phase.
As a whole, the coupled wave–TC pattern was well organized and stronger in the ISO westerly phase, but weaker and poorly organized in the easterly phase. The reason for this contrast could be attributed to the distinct configuration of the background flow, which fluctuated along with the ISO. The following hypothesis is proposed based on the aforementioned diagnostic results. In the westerly phase, the strong monsoon trough warrants the strong southwesterly moisture transport and the enhanced moisture convergence in the confluent zone. The submonthly wave pattern is presumably more likely to be sustained in the Philippine Sea in such an environment. The stronger submonthly wave pattern consists of strong moisture convergence and positive vorticity anomalies (negative streamfunction) in the lower troposphere, which further enhance the favorable environment for TC development. In the easterly phase, the monsoon trough is weaker and the subtropical anticyclonic ridge stronger and shifted southward in the Philippine Sea. The weaker moisture transport and convergence, as reflected in the easterly anomaly, was observed in the Philippine Sea. These characteristics all indicated a less favorable environment for the genesis and development of the TCs in the Philippine Sea. Instead, most activity of the coupled wave–TC pattern occurred mainly in the northwestern corner of the Philippine Sea and the east China Sea, where the base region for the choice of composite cases was located.
The results presented above demonstrated the multiscale nature, from synoptic to intraseasonal time scales, of the atmospheric flow in the tropical WNP. The ISO fluctuation in the 30–80-day scale modulated the monsoon trough and the subtropical anticyclonic ridge, which provided a favorable or less favorable background flow for the submonthly wave pattern and TCs. This study provides a general picture of the background flow in both ISO phases and their relationship with the coupled wave–TC pattern. However, the cause and effect relationship between the large-scale circulation (e.g., the ISO and the submonthly wave pattern) and TCs remains unsolved. Several studies suggested that TCs may induce wavelike disturbances to the southeast (e.g., Holland 1995; Sobel and Bretherton 1999; Li and Fu 2006; Li et al. 2006; Krouse et al. 2008). One may suspect whether the submonthly wave pattern and the ISO were partially forced by TCs. This possibility could not be ruled out. However, there was no clear evidence to support it. For example, no time lag existed between the evolution of the wave pattern and the occurrence of TCs, and there was no indication of the southeastward development of the wave pattern in the form of Rossby wave energy dispersion. Moreover, the submonthly wave pattern exhibited lower frequency and larger spatial scale (4000–5000 km) than the Rossby wave energy dispersion (2000–3000 km) forced by TCs, as shown in previous studies (e.g., Li and Fu 2006; Krouse et al. 2008). These results suggest that the wave pattern is likely to have its own identity and is not merely a response to TC forcing. It is possible that the existence of the large-scale circulation pattern provides a favorable condition for the genesis and development of TCs, which in turn enhance the large-scale circulation pattern.
Since there were always TCs embedded in the cyclonic circulation, the TC winds were removed to assess the TC effect on the large-scale circulation. A comparison between the composites of the original and TC-removed winds indicated that both the ISO and the submonthly wave pattern remain intact after removing TC winds, although their cyclonic circulations became weaker. This result suggests that, although the TCs may contribute notably to the cyclonic circulation and moisture flux convergence, the large-scale circulation can also contribute significantly to set up different background flows in the ISO westerly and easterly phases.
To understand the dynamics for the mutual relationship between the large-scale circulation and TCs, more diagnostics and numerical simulation studies are needed. For example, a calculation of the energy conversion between the ISO-background flow and the coupled wave–TC pattern is needed to pinpoint how the coupled wave–TC pattern grew at the expense of the ISO-background flow. Such a study is underway to contrast the effect of the westerly and easterly phase ISO on the higher frequency fluctuations.
Acknowledgments
The authors thank Dr. Yi-Chiang Yu for graphic assistance and Mr. Yen-Min Lee and Miss Yan-Lan Wang for preparing the data. Comments from three anonymous reviewers greatly improved this manuscript, and the authors thank them for pointing out several items that were not described clearly in the earlier version of the manuscript. This study was supported by the National Science Council of Taiwan under Grant NSC95-2111-M-017-001-AP4 issued to Ken-Chung Ko and Grant NSC 95-2111-M-002-004-AP4 to Huang-Hsiung Hsu.
REFERENCES
Camargo, S. J., A. W. Robertson, S. J. Gaffney, P. Smyth, and M. Ghil, 2007a: Cluster analysis of typhoon tracks. Part I: General properties. J. Climate, 20 , 3635–3653.
Camargo, S. J., A. W. Robertson, S. J. Gaffney, P. Smyth, and M. Ghil, 2007b: Cluster analysis of typhoon tracks. Part II: Large-scale circulation and ENSO. J. Climate, 20 , 3654–3676.
Chang, C-P., J. M. Chen, P. A. Harr, and L. E. Carr, 1996: Northwestward-propagating wave patterns over the tropical western North Pacific during summer. Mon. Wea. Rev., 124 , 2245–2266.
Chen, T-C., and J-M. Chen, 1995: An observational study of the South China Sea monsoon during the 1979 summer: Onset and life cycle. Mon. Wea. Rev., 123 , 2295–2318.
Chen, T-C., and S-P. Weng, 1999: Interannual and intraseasonal variations in monsoon depressions and their westward-propagating predecessors. Mon. Wea. Rev., 127 , 1005–1020.
Chen, T-C., M-C. Yen, and M. Murakami, 1988: The water vapor transport associated with the 30–50 day oscillation over the Asian monsoon regions during 1979 summer. Mon. Wea. Rev., 116 , 1983–2002.
Chen, T-C., M-C. Yen, and S-P. Weng, 2000: Interaction between the summer monsoons in East Asia and the South China Sea: Intraseasonal monsoon modes. J. Atmos. Sci., 57 , 1373–1392.
Elsberry, R. L., 2004: Monsoon-related tropical cyclones in East Asia. East Asian Monsoon, C.-P. Chang, Ed., World Scientific Series on Meteorology of East Asia, Vol. 2, World Scientific, 463–498.
Gray, W. M., 1998: The formation of tropical cyclones. Meteor. Atmos. Phys., 67 , 37–69.
Hamming, R. W., 1989: Digital Filters. Prentice-Hall, 284 pp.
Harr, P. A., and R. L. Elsberry, 1995: Large-scale circulation variability over the tropical western North Pacific. Part I: Spatial patterns and tropical cyclone characteristics. Mon. Wea. Rev., 123 , 1225–1246.
Holland, G. J., 1995: Scale interaction in the western Pacific monsoon. Meteor. Atmos. Phys., 56 , 57–79.
Hsu, H-H., 2005: East Asian and western North Pacific summer monsoon region. Intraseasonal Variability in the Atmosphere–Ocean System, K.-M. Lau and D. Waliser, Eds., Springer-Praxis, 65–98.
Hsu, H-H., and C-H. Weng, 2001: Northwestward propagation of the intraseasonal oscillation in the western North Pacific during the boreal summer: Structure and mechanism. J. Climate, 14 , 3834–3850.
Hsu, H-H., C-H. Hung, A-K. Lo, and C-W. Hung, 2008: Influence of tropical cyclones on the estimation of climate variability in the tropical western North Pacific. J. Climate, 21 , 2960–2975.
Hurrell, J. W., and H. van Loon, 1997: Decadal variations in climate associated with the North Atlantic Oscillation. Climatic Change, 36 , 301–326.
Kaylor, R. E., 1977: Filtering and decimation of digital time series. Institute of Physical Science Technology, Tech. Note BN 850, 42 pp.
Ko, K-C., and H-H. Hsu, 2006: Sub-monthly circulation features associated with tropical cyclone tracks over the East Asian monsoon area during July–August season. J. Meteor. Soc. Japan, 84 , 871–889.
Krouse, K. D., A. H. Sobel, and L. M. Polvani, 2008: On the wavelength of the Rossby waves radiated by tropical cyclones. J. Atmos. Sci., 65 , 644–654.
Kurihara, Y., M. A. Bender, and R. J. Ross, 1993: An initialization scheme of hurricane models by vortex specification. Mon. Wea. Rev., 121 , 2030–2045.
Kurihara, Y., M. A. Bender, R. E. Tuleya, and R. J. Ross, 1995: Improvements in the GFDL hurricane prediction system. Mon. Wea. Rev., 123 , 2791–2801.
Lau, K-H., and N-C. Lau, 1990: Observed structure and propagation characteristics of tropical summertime synoptic scale disturbances. Mon. Wea. Rev., 118 , 1888–1913.
Lau, K-M., and P. H. Chan, 1986: Aspects of the 40–50 day oscillation during the northern summer as inferred from outgoing longwave radiation. Mon. Wea. Rev., 114 , 1354–1367.
Li, T., and B. Fu, 2006: Tropical cyclogenesis associated with Rossby wave energy dispersion of a preexisting typhoon. Part I: Satellite data analysis. J. Atmos. Sci., 63 , 1377–1389.
Li, T., X. Ge, and B. Wang, 2006: Tropical cyclogenesis associated with Rossby wave energy dispersion of a preexisting typhoon. Part II: Numerical simulations. J. Atmos. Sci., 63 , 1390–1409.
Liebmann, B., H. H. Hendon, and J. D. Glick, 1994: The relationship between tropical cyclones of the western Pacific and Indian Oceans and the Madden–Julian oscillation. J. Meteor. Soc. Japan, 72 , 401–411.
Madden, R. A., and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702–708.
Maloney, E. D., and M. J. Dickinson, 2003: The intraseasonal oscillation and the energetics of summertime tropical western North Pacific synoptic-scale disturbances. J. Atmos. Sci., 60 , 2153–2168.
Nakazawa, T., 1986: Intraseasonal variations of OLR in the tropics during the FGGE year. J. Meteor. Soc. Japan, 64 , 17–34.
Nitta, T., 1987: Convective activity in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65 , 373–390.
Schnadt, C., A. Fink, D. G. Vincent, J. M. Schrage, and P. Speth, 1998: Tropical cyclone, 6–25 day oscillations, and tropical-extratropical interaction over the northwestern Pacific. Meteor. Atmos. Phys., 68 , 151–169.
Sobel, A. H., and C. S. Bretherton, 1999: Development of synoptic-scale disturbances over the summertime tropical Northwest Pacific. J. Atmos. Sci., 56 , 3106–3127.
Straub, K. H., and G. N. Kiladis, 2003: Interactions between the boreal summer intraseasonal oscillation and higher-frequency tropical wave activity. Mon. Wea. Rev., 131 , 945–960.
Tsou, C-H., P-C. Hsu, W-S. Kau, and H-H. Hsu, 2005: Northward and northwestward propagation of 30–60 day oscillation in the tropical and extra-tropical western North Pacific. J. Meteor. Soc. Japan, 83 , 711–726.
Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131 , 2961–3012.
Wang, B., and H. Rui, 1990: Synoptic climatology of transient tropical intraseasonal convection anomalies: 1975–1985. Meteor. Atmos. Phys., 44 , 43–61.
Wang, B., and X. Xie, 1997: A model for the boreal summer intraseasonal oscillation. J. Atmos. Sci., 54 , 72–86.
Wheeler, M. C., and H. H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132 , 1917–1932.
Wu, C-C., T-H. Yen, Y-H. Kuo, and W. Wang, 2002: Rainfall simulation associated with Typhoon Herb (1996) near Taiwan. Part I: The topographic effect. Wea. Forecasting, 17 , 1001–1015.
(a) Climatological mean streamfunction (unfiltered) and horizontal moisture flux and (b) wind speed variance at 850 hPa, averaged over 23 July–September seasons (1979–2001). The shaded areas for (a) are the moisture flux convergence (≦ −1.5 × 10−8g kg−1 s−1), and the contour interval is 10(105 m2 s−1) for the streamfunction. The interval for (b) is 4 (m s−1) 2 and areas > 22 (m s−1)2 are shaded. The box is the maximum variance area (20°–35°N, 125°–140°E).
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
The moving spectrum for the wind speed averaged over 20°–35°N, 125°–140°E. The x axis is the frequency (1/period in days) and the y axis is the day starting from 1 June. All spectra are ensemble averaged over 23 years (1979–2001). The 99.9% confidence limit has been subtracted. Areas greater than 50 m2 s−2 are shaded.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
The variance ratios (30–80 day/unfiltered) for the zonal wind averaged over July–September for 23 years (1979–2001). The interval is 0.05 and areas ≧ 0.45 are shaded. The box is the maximum area (5°–15°N, 100°–120°E).
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
Example of 1983 time series of 7–30-day filtered wind speed (shaded) averaged over the base region as in Fig. 1 and the time series of 30–80-day filtered zonal wind (solid) averaged over the base region as in Fig. 3. The horizontal bars mark the time when a TC enters the base region. See text for details.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
(a) Percentage of submonthly cases (TC and no-TC). (b) Percentage of no-TC cases in July–September. (c) Percentage of TCs associated with submonthly cases (TC_subm), TCs in the base region but no submonthly case (TC_nosubm), and TCs outside the base region. (d) Numbers of ISO westerly and easterly events (ISOw and ISOe), numbers of submonthly TC cases for the westerly and easterly phases (SubmTCw and SubmTCe), and numbers of TCs associated with submonthly TC cases (TC_submw and TC_subme).
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
Composite maps of 30–80-day filtered 850-hPa streamfunction and wind vectors for the westerly phase. The interval is 3(×10−5) m2 s−1 for streamfunction and only wind vectors that exceed the 95% confidence limit are shown.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
As in Fig. 6 but for the easterly phase.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
Composite maps of 7–30-day filtered 850-hPa streamfunction and the 30–80-day filtered streamfunction in the ISO westerly phase. Climatological mean streamfunction was added to the 30–80-day streamfunction perturbation to illustrate the intraseasonal fluctuation of the monsoon trough and anticyclone (see text for details). The interval is 3(105 m2 s−1) for the submonthly components and 10(105 m2 s−1) for the other.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
As in Fig. 8 but for the easterly phase.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
(a) Westerly and (b) easterly composite streamfunction (filtered) on Day 0 and the associated typhoon tracks over 23 July–September seasons (1979–2001). Variance from Day −6 to +6 is also shown and areas ≥80(1010 m4 s−2) are shaded.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
The 850-hPa composite filtered streamfunction (TC cases) on (a), (d) Day −3; (b), (e) Day 0; and (c), (f) Day −4.5 for 23 July–September summers (1979–2001). The interval is 3(×105) m2 s−1. Also shown for (a), (d) and (b), (e) are the typhoon positions (with the past one-day tracks, and big dots represent the wind speed ≥ 64 kt) associated with the composite cases and for (c), (f) the typhoon genesis positions between Day −1.5 and Day −7.5. The shaded areas represent streamfunction that exceeds the 95% confidence limit.
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
(top) Unfiltered horizontal moisture flux (vectors) and moisture flux convergence (shaded) averaged between −10 and +10 days of the peak intraseasonal events over 23 July–September seasons (1979–2001); (middle), (bottom) As in the upper panels except for the 30–80 filtered fields (−3 to +3 days for submonthly cases). Left (right) The westerly (easterly) phase. The interval for moisture flux convergence in the upper panels is 1 × 10−8g kg−1 s−1 and areas ≦ −1 × 10−8g kg−1 s−1 are shaded, and is 5 × 10−9g kg−1 s−1 and areas ≦ −5 × 10−9g kg−1 s−1 are shaded for the rest. Also shown in the upper panels are the typhoon genesis positions (triangles).
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
Composite 7–30-day filtered wind (vectors) and vorticity on Day 0 for (a) westerly and (b) easterly phases; (c), (d) as in (a) and (b) except for the TC-removed wind fields. The contour interval is 0.2 (×10−5 s−1) and dark shading is areas greater than 0.2; light shading is areas less than −0.2. (e), (f) The differences (a) minus (c), (b) minus (d).
Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2282.1
