The relationships between large-scale wintertime circulation and extratropical cyclones that develop explosively (the so-called bomb cyclones) over the western North Pacific are investigated using Japanese long-term reanalysis project data. On a monthly basis, the East Asian winter monsoon variability strongly modulates the bomb cyclone activity in terms of its geographical distribution. When the monsoon is strong, the bomb cyclone activity tends to concentrate in the vicinity of the Kuroshio Current and the Kuroshio Extension near Japan, while when the monsoon is weak, it disperses over the broader areas. The enhancement of the monsoon increases the heat and moisture supply from warm currents, facilitating unstable conditions within the atmospheric boundary layer and intensifying baroclinicity in the lower troposphere. These factors are believed to play a role in inducing bomb cyclones, particularly along the warm currents. On submonthly timescales, the stationary Rossby wave propagation along the South Asian waveguide serves as a prominent trigger for the rapid reinforcement of synoptic-scale disturbances around Japan. When a pronounced bomb cyclone comes to its mature stage northeast of Japan, it is capable of exciting stationary Rossby waves downstream from the Asian jet exit region as vorticity forcing. The stationary wave packets developing southeastward across the North Pacific Ocean basin induce surface cyclogenesis in the vicinity of the Hawaiian Islands by leading to the equatorward advection of higher potential vorticity from the midlatitudes, bringing about the occurrence of kona storms, which cause weather hazards in Hawaii.
 One of the major differences in large-scale atmospheric circulation between the northwestern Pacific and Atlantic regions is the presence of the East Asian monsoon system. In boreal winter, monsoonal northwesterly winds prevail over the East China Sea, the Japan Sea, and the northwestern Pacific Ocean, which is accompanied by the development of the Siberian high and Aleutian low systems. If the East Asian winter monsoon is enhanced, the associated northwesterly winds strengthen, further supplying heat and moisture to the overlying atmosphere from the warm currents along the northwestern periphery of the North Pacific. It appears that baroclinicity in the lower troposphere is also reinforced in the vicinity of Japan. We thus anticipate that these changes in large-scale environment strongly influence the statistical features of explosively developing extratropical cyclone activity, such as locations, tracks, and developing rates. It is not necessarily certain, however, how the East Asian winter monsoon circulation regulates such extratropical cyclone activities, especially explosive cyclone activity, over the warm westerly boundary currents of the North Pacific Ocean, although Nakamura and Sampe  and Nakamura et al.  showed a close relationship between the upper-level subtropical jet and storm track over the western North Pacific during northern winter in terms of storm track activity.
 It is also still uncertain what mechanisms govern the interannual variability of the explosive extratropical cyclone activity. Chen et al.  noted a possible correlation between explosive-cyclogenesis frequency off the East Asian coast and the El Niño-Southern Oscillation (ENSO), although they did not propose any physical mechanisms between the two. The ENSO-related tropical forcing significantly affects the East Asian winter monsoon activity [e.g., Zhang et al., 1996; Zhang et al., 1997; Wang et al., 2000; Sakai and Kawamura, 2009]. Wang et al.  clarified the role of the western North Pacific anticyclone in the East Asian winter monsoon through the lower troposphere. According to their study, low-level southerly (northerly) anomalies along the western periphery of the western North Pacific anticyclone (cyclone) contribute mainly to the lower tropospheric warming (cooling) in the southern part of East Asia. In contrast, Sakai and Kawamura  showed that ENSO-related anomalous convection can give rise to a change in the East Asian winter monsoon system through stationary Rossby wave propagation along the South Asian waveguide in the upper troposphere. These studies propose possible mechanisms that explain why the East Asian winter monsoon variability has an ENSO signal. To understand the physical relationship between the ENSO and explosive-cyclogenesis frequency over the western North Pacific via atmospheric bridges, it would be necessary to first clarify how the East Asian winter monsoon variability is related to the explosive cyclone activity.
 The so-called bomb cyclones develop rapidly because of the heat and moisture supply from the warm currents; on the other hand, the individual cyclones are usually accompanied by strong divergent motion in the upper troposphere in association with active latent heat release. It is conceivable that bomb cyclones play a vital role in generating significant vorticity anomalies within the upper-level subtropical jet (the so-called Asian jet) over the western North Pacific. As is well known, tropical cyclones have a warm core structure in the upper troposphere, also accompanied by outstanding upper-level divergence. Enomoto et al.  pointed out the importance of the remote effect of a tropical storm on a cutoff cyclone over Europe in the boreal summer. Kawamura and Ogasawara  and Yamada and Kawamura  also demonstrated that individual typhoons, which are a synoptic-scale convective heat source over the western North Pacific, can induce a barotropic Rossby wavetrain and significantly activate stationary fronts around Japan during the rainy season as remote forcing. It is quite possible, thus, that the Asian jet serves as a waveguide and the explosive extratropical cyclones have a significant impact on large-scale circulations through the downstream development of stationary wave propagations. In fact, Lau and Lau  noted the possibility that midlatitude disturbances accompanying cold-air outbreaks over East Asia contribute to downstream and equatorward development of low-frequency fluctuations, although they did not clearly specify its mechanism.
 In the present study, we will first show the climatology of the explosive extatropical cyclone activity over the western North Pacific region during the northern winter in the past three decades following the criteria of Yoshida and Asuma , who extracted explosively developing extratropical cyclones. As a next step, we will make composite analyses of various atmospheric and oceanic variables with respect to the strong and weak phases of the East Asian winter monsoon circulation using an appropriate winter monsoon index. As will be reported in a later section, the East Asian winter monsoon variability changes the baroclinicity, especially in the lower troposphere and the surface heat exchange between the Kuroshio Current and the Kuroshio Extension and the overlying atmosphere, playing an influential role in the concentration and the magnitude of the developing rate of explosive extratropical cyclones. Furthermore, we will demonstrate how explosively developing extratropical cyclones as pronounced vorticity sources contribute to the downstream and equatorward development of stationary Rossby wave propagations.
 The study is organized as follows. Section 2 contains a description of the data used and analysis procedures. Section 3 is dedicated to the overall statistical features of the explosive extratropical cyclone activity over the western North Pacific and examines how the East Asian winter monsoonal forcing regulates its cyclone activity. Section 4 is an examination of the effects of extratropical disturbances on the downstream development of large-scale circulations. The discussion and summary are presented in sections 5 and 6, respectively.
2. Data Used and Analysis Procedure
 The extraction of the explosive extratropical cyclones and associated winter monsoon circulation was made using 6-hourly data from the Japanese long-term Re-Analysis project (JRA-25) [Onogi et al., 2007] with a spatial resolution of 1.25° longitude by 1.25° latitude for the period of 1979–2004 and data from the Japan Meteorological Agency Climate Data Assimilation System (JCDAS) with the same resolution for the period of 2005–2007. Because JCDAS is operated with the same data assimilation system as JRA-25 to generate quasi-real-time products, comparatively homogeneous atmospheric circulation data are available over 29 years (1979–2007) using both JRA-25 and JCDAS data. Using the 6-hourly sea level pressure (SLP) data, we tried to extract the explosive extratropical cyclones within a specific region (110°–170°E, 20°–60°N) during 28 winters (December, January and February) from 1979 through 2007. Following the criteria of Yoshida and Asuma's , an explosively developing extratropical cyclone is identified as a bomb cyclone if the subsequent developing rate (ɛ) exceeds 1 hPa hr−1 (i.e., it is equivalent to 1 Bergeron, as named by Sanders and Gyakum ).
where p is the SLP at the center of cyclone, t is the time in hours, and ϕ is the latitude at the cyclone center. The initial appearance of an extratropical cyclone that develops and becomes a bomb cyclone is defined at any grid point with an SLP at least 1 hPa lower than the value at an adjacent grid point. Bomb cyclones that are of tropical storm origin are excluded from our analysis. The position of maximum development is taken to be location of the cyclone center at the middle of the 12-hour period of most rapid enhancement.
 Monthly outgoing longwave radiation (OLR) data are also used with a spatial resolution of 2.5° × 2.5° observed from the National Oceanic and Atmospheric Administration (NOAA) satellites for the period 1979–2007 and data sets from the Hadley Center Sea Ice and Sea Surface Temperature (HadISST) [Rayner et al., 2003] with a spatial resolution of 1° × 1° for the same period. In addition, we use daily mean surface sensible and latent heat flux data with a spatial resolution of 1° × 1° from the Japanese Ocean Flux data sets with Use of Remote sensing Observations (J-OFURO) version 2 [Kubota et al., 2002; H. Tomita, et al., An assessment of surface heat fluxes from J-OFURO2 at the KEO/JKEO sites, submitted to Journal of Geophysical Research, 2009] for the period 1988–2005, although the available period is shorter.
 To examine the relationship between the East Asian winter monsoon and explosive cyclogenesis over the warm currents of the western North Pacific, we used an index proposed by Hanawa et al.  and Watanabe  in this study as an appropriate measure of the East Asian winter monsoon intensity. This monsoon intensity index, which is simply defined as the difference in SLP between Irkutsk in Russia and Nemuro in Japan (Irkutsk minus Nemuro), demonstrates that northwesterly monsoonal winds in East Asia are regulated by a combination of both the Siberian high and Aleutian low systems. The validity of the index has been assessed by Sakai and Kawamura , and others and is very useful for understanding the monsoon variability as well as other indices [e.g., Jhun and Lee, 2004]. Hereafter, for convenience, we refer to the monsoon index as the MO index, as named by Hanawa et al.  and Watanabe . Since the reanalysis data are used in this study, we define the MO index (MOI) as the difference in SLP between two grid points of (105°E, 52.5°N) near Irkutsk and (145°E, 43.75°N) near Nemuro. After the monthly mean MOI series are standardized with regard to each winter month (December, January, and February), we defined a month for which the index is greater than +1.0 (less than −1.0) as a strong (weak) monsoon phase at which the East Asian winter monsoon circulation changes significantly and then constructed composite maps for the locations, tracks, and maximum developing rates of the bomb cyclones and other atmospheric and oceanic variables. This study focused specifically on winter monsoon variability with periods longer than about one month; however, it should be borne in mind that its variability contains both interannual and intraseasonal variability.
 As an additional analysis, we applied a combination of two simple low-pass filters (the 3-day-weighted running average and 31-day running average) to the daily data and extracted a high-frequency (HF) component (submonthly timescales). The 3-day filter is a 1-2-1 filter. The HF component is obtained as the difference of 3-day-filtered minus 31-day-filtered data, which is highlighted when examining the remote effects of explosive extratropical disturbances on large-scale anomalous circulations with submonthly scales.
3. East Asian Winter Monsoon Circulation and Associated Bomb Cyclone Activity
3.1. Climatology of the Bomb Cyclone Activity
 As indicated in Table 1, we extracted 506 cases considered as bomb cyclones over 84 months during 28 winters. On average, about 6 bomb cyclones per month are analyzed. The average in their maximum developing rates is 1.60 bergeron. These values are significantly different from the results of Yoshida and Asuma  because the analyzed domain, season, and period are not the same. In December and January, both the number of bomb cyclones and their maximum developing rates are very similar, but, in February, the number is small, and the maximum developing rate is somewhat larger than in the other winter months.
Table 1. Number of Bomb Cyclones and Monthly Average of Their Maximum Developing Rates (Unit: Bergeron) for December, January, and February
Maximum developing rate
Figure 1 is a display of the frequency distribution of cyclone tracks with respect to the bomb cyclone in December, January, and February during 28 winters from 1979 through 2007. It is noteworthy that the frequency is counted as the number of extratropical cyclones that pass through a respective grid point and the track is traced from the initial appearance through the disappearance of the cyclone that became a bomb cyclone. Two maxima of the frequency are located in December and January over the Japan Sea and the western North Pacific to the east of Japan, which is consistent with Figure 4 of Chen et al. . From January to February, the frequency over the Japan Sea diminishes and, conversely, becomes high in the vicinity of the Kuroshio Extension. Yoshida and Asuma  demonstrated that bomb cyclones developing rapidly over the Japan Sea and the Sea of Okhotsk tend to appear frequently in late fall and, in contrast, those formed to the south of Japan prevail in midwinter. Our results support their findings.
 As reported in section 2, we defined strong and weak monsoon phases when the standardized monthly mean MOI in each winter month is in excess of its threshold values, which resulted in the selection of 14 and 14 months as strong and weak monsoon categories, respectively, as shown in Table 2. The strong monsoon category consists of 3 times for December, 5 for January, and 6 for February, whereas the weak one consists of 6 times for December, 4 for January, and 4 for February. Consequently, these two categories are not strongly related to a particular winter month. Because the East Asian winter monsoon system has pronounced intermonthly variability, the strong and weak monsoon phases very often coexist in the same winter season (e.g., January and February 1994). That is why this definition seems reasonable. Figure 2 reveals the composite anomalies of 850-hPa temperature and wind vector at the strong monsoon phase (14 months) and weak monsoon phase (14 months). At the strong monsoon phase, northwesterly anomalies prevail over the eastern coast of China, Korea, and Japan. Significant negative temperature anomalies caused by cold advection cover almost the entire region of northern East Asia and expand toward the Kuroshio Current and the Kuroshio Extension. As for the weak monsoon phase, a mirror image with opposite signs is obtained. The temperature difference between the two phases reaches about 6°C around the Korean peninsula. It is thus confirmed that our criterion is successful to capture the strong and weak phases of the East Asian winter monsoon circulation. We also confirmed that, even though Jhun and Lee's index is used, the obtained features are very similar to those of Figure 2, indicating that the results do not depend on the choice of a particular monsoon intensity index.
Table 2. Strong and Weak Monsoon Categories Extracted on the Basis of the Standardized Monthly Mean MOI
 The numbers of bomb cyclones are 79 and 88 for the strong and weak monsoon phases, respectively. In other words, 5.64 and 6.29 bomb cyclones per month appear at the strong and weak phases of the East Asian winter monsoon activity, respectively. The average in maximum developing rates is 1.50 bergeron at the weak monsoon phase, which is smaller than a value of 1.61 bergeron at the strong monsoon phase.
3.2. Supply of Surface Sensible and Latent Heat From the Ocean
Figures 3a and 3b present the composite patterns of monthly mean surface turbulent heat (sensible and latent heat) fluxes and SST at the strong and weak phases of the East Asian winter monsoon activity, respectively. The locations of bomb cyclones when their rates of development reached the maximum are also exhibited. The size of circles denotes the magnitude of the maximum developing rate. Since surface heat flux data are available only for the period 1988–2005, eight and ten months are selected to make composite maps as the strong and weak monsoon phases, respectively (see Table 2). Comparing the two monsoon phases, as expected, the geographical distributions of the maximum developing rate are considerably different from each other.
 When the monsoon is strong, it turns out that the bomb cyclones with their maximum developing rates concentrate in the vicinity of the Kuroshio Current off the Pacific coast of Japan and the Kuroshio Extension just east of Japan. The corresponding ocean area, which is located south of a noticeable SST front, is characterized by surface turbulent heat flux in excess of 400 W m−2 because of the enhanced northwesterly monsoon flow, indicating a significant contribution of heat and moisture supply from the warm current to rapid intensification of extratropical cyclones. Likewise, the bomb cyclone activity is enhanced over the Japan Sea in association with increased heat fluxes. When the monsoon is weak, on the other hand, the ocean area in excess of 400 W m−2 shrinks because of the weakening of the overlying northwesterly winds, and there are few bomb cyclones with maximum developing rates in that vicinity. The difference in the surface turbulent heat flux between the strong and weak phases of the monsoon circulation exceeds 60 W m−2 in the vicinity of Japan, as indicated in Figure 3c. It is noteworthy that the heavy and light shades denote statistically significant regions with 95% and 90% confidence levels, respectively. The bomb cyclone activity is also weakened over the Japan Sea. The main area of such a cyclone activity shifts eastward and goes away from Japan. It is confirmed, thus, that both systems of the East Asian winter monsoon and the warm current characterize the bomb cyclone activity, especially in terms of its geographical distribution.
3.3. Static Stability and Baroclinicity in the Lower Troposphere
 In this subsection, we examine the vertical and horizontal gradients of equivalent potential temperature (θe) and the maximum Eady growth rate [Lindzen and Farrell, 1980] as useful measures of static stability and baroclinicity in the lower troposphere of the western North Pacific sector. Figures 4a and 4b and reveal the composite patterns of the θe difference between 925- and 1,000-hPa levels for the strong monsoon (14 months) and weak monsoon (14 months) phases, respectively. The tendency of the bomb cyclone activity is roughly the same as that shown in Figure 3. By increasing the composite numbers, the concentration of the bomb cyclone activity into the conjunction area between the Kuroshio Current and the Kuroshio Extension is more evident at the strong monsoon phase. It is interesting to note that, when the monsoon is strong, most of the bomb cyclones with their maximum developing rates over the western North Pacific are confined within a specific domain where the θe difference is less than 2 K. At the weak monsoon phase, static stability at the lowest level of the troposphere increases over the western North Pacific region. Associated with this change, the bomb cyclones with their maximum developing rates are dispersed widely. Figure 4c shows the difference in the near-surface static stability between the strong and weak monsoon phases. Significant areas coincide well with those of Figure 3c. As suggested from Figures 3 and 4, when a monsoon is strong (weak), the increase (decrease) in heat and moisture supply from the ocean into the atmospheric boundary layer facilitates unstable (stable) conditions within its boundary layer.
Figures 5a and 5b show the composite patterns of the horizontal gradient of θe at 925 hPa along with the bomb cyclone tracks for the strong and weak phases of the East Asian winter monsoon, respectively. A common feature between the two phases is the predominance of near-surface baroclinicity along the Kuroshio Current and the Kuroshio Extension. It turns out that the bomb cyclones originate and intensify rapidly along the baroclinic zone. When the monsoon is strong, the baroclinicity tends to be enhanced over the southern part of the East China Sea along the Kuroshio Current and the central Japan Sea as compared to the weak monsoon phase, while no significant changes in baroclinicity are found at about 35°–40°N around the Kuroshio Extension, as indicated in Figure 5c. The number of the bomb cyclones with maximum developing rates over the central Japan Sea appears to increase, in conjunction with the enhanced baroclinicity; however, over the southern part of the East China Sea near the Kuroshio Front, the bomb cyclones do not necessarily reach their maximum rates of development. Rather, enhanced near-surface baroclinicity over that region may contribute to the occurrence of extratropical disturbances or their development at early stages [e.g., Xie et al., 2002].
 Since the near-surface baroclinicity is almost unchanged in the Kuroshio Extension region just east of Japan, where the bomb cyclone activity is concentrated (dispersed) when the monsoon is strong (weak), we further examine the latitude-height section of the horizontal θe gradient along 142.5°–145°E across the Kuroshio Extension, as exhibited in Figure 6. Figures 6a and 6b show the strong and weak monsoon phases, respectively. A near-surface baroclinic zone lies between 36°–39°N, which coincides well with the SST front of the Kuroshio Extension, but the magnitude of the horizontal gradient is almost the same and is insensitive to the variability of the East Asian winter monsoon circulation. Apart from such a baroclinic zone, which is mostly confined to the atmospheric boundary layer, another salient baroclincity zone is observed in the middle and lower troposphere above the 850-hPa level. As for the lower troposphere, the maximum horizontal gradient is located at around 35°N in the 850- and 700-hPa layer. Figure 6c reveals the difference in the horizontal θe gradient between the strong and weak phases (strong minus weak) of the East Asian winter monsoon. In addition, the difference in zonal wind is denoted in order to see the meridional shift of the upper-level Asian jet that is intimately linked with the monsoon variability. A notable feature is the reinforcement of baroclinicity at about 35°N, especially between 850- and 700-hPa at the strong monsoon phase; in contrast, its weakening is apparent in the area about 40°–45°N. Such changes in baroclinicity are attributed to the equatorward displacement of the Asian jet. As already indicated in Figure 5 and other figures, bomb cyclones with maximum developing rates concentrate at around 35°N off the Pacific coast of the mainland of Japan when the monsoon is strong. The lower tropospheric baroclinicity is also enhanced in the area where the bomb cyclone activity is locally activated, being indicative of the influential role of the baroclinicity in the lower troposphere in the bomb cyclone activity over the Kuroshio Extension. Furthermore, different enhancement of baroclinicity is seen at around 23°–33°N, especially in the middle troposphere, which is accompanied by the equatorward migration of the upper-level jet, but a large part of the corresponding region is located to the south of the area where the bomb cyclone activity is observed. It is uncertain whether this baroclinicity is also important for the rapid intensification of extratropical cyclones.
 As another measure of the baroclinicity, we also estimated the maximum Eady growth rates (σ) for the layer between the 850- and 700-hPa levels as follows.
where f is the Coriolis parameter, N is the Brunt-Vaisala frequency, and ∂u/∂z is the zonal wind vertical shear. Figures 7a and 7b show the composite patterns of the maximum Eady growth rates at the strong and weak phases of the monsoon circulation, respectively. The locations of bomb cyclones when their rates of development reached the maximum are also shown. Focusing on the strong monsoon phase, the area where the baroclinicity is pronounced is located just off the Pacific coast of mainland Japan, and its magnitude exceeds 0.9 day−1. As expected, the lower tropospheric baroclinicity is enhanced (weakened) along the Kuroshio Current and the Kuroshio Extension when the monsoon is strong (weak). An opposite tendency can be seen to the north of about 40°N; that is, the baroclinicity is suppressed for the strong phase of the monsoon circulation, and vice versa (see Figure 7c). Although the maximum Eady growth rate does not consider the effect of the moist baroclinic process, these features are basically consistent with the results of Figures 5 and 6. We also computed the maximum Eady growth rates for the 850–500-hPa layer and found similar features.
 The results of this section demonstrate that the East Asian winter monsoon variability strongly regulates the bomb cyclone activity over the western North Pacific region. When the monsoon is strong, the bomb cyclone activity tends to concentrate in the vicinity of the Kuroshio Current and the Kuroshio Extension near Japan, whereas, when the monsoon is weak, it disperses over broader areas. We believe that the concentration (dispersion) of the bomb cyclone activity results from the increased (decreased) heat and moisture supply from the warm currents and associated enhanced (weakened) baroclinicity in the lower troposphere.
4. Relationships Between Bomb Cyclones and Large-Scale Circulations on Submonthly Scales
 In this section, we investigate the cross-relationships between bomb cyclones and large-scale atmospheric circulations on submonthly scales, although the response of bomb cyclone activity to the winter monsoon variability on the monthly basis is presented in the previous section. In particular, possible remote forcing of bomb cyclones to large-scale circulation is examined in the latter half of this section. To analyze the temporal evolutions of large-scale circulation patterns in detail, we highlighted the variability with the HF component, as mentioned in section 2.
4.1. Definition of an Index that Reflects the Rapid Intensification of Extratropical Cyclones
 Since the rapid development of extratropical cyclones is characterized by an extreme decrease in SLP, we use a daily SLP index as an appropriate measure of the cyclone activity. On the basis of the locations of bomb cyclones when their rates of developing reached the maximum, we obtain their mean position (151.25°E, 41.25°N), and the index is then concisely defined as the average of band-pass-filtered SLPs over a specific domain (141.25°–161.25°E, 31.25°–51.25°N) with a range of 20° latitude and 20° longitude and its center at their mean position. For convenience, we hereafter call it the maximum developing index (MDI) in this study. After the daily mean MDI series are standardized, we identify a day in which the index is less than a value of −2.0 and its minimum value is a specific event. Following the above criterion, the total number of the specific events is 64 over 84 months during 28 winters, and its number tends to increase from the 1980s through the 2000s, as indicated in Table 3. It is also worthwhile to note that such an increasing tendency is unchanged, even though we used a value of −1.5 as an alternate threshold. The number of cases in which a bomb cyclone contributes directly to the occurrence of a specific event is 54 out of 64 (approximately 84%), suggesting that the extracted events strongly reflect the behavior of a bomb cyclone.
Table 3. Number of Extreme Events Extracted on the Basis of the MDI With Sub-Monthly Timescalesa
A day for which the standardized index is less than a value of −2.0 and which has a minimum value is defined as a specific event. Also shown is the number of events when a value of −1.5 as an alternate threshold is used.
1979/80–1988/89 (10 winters)
1989/90–1998/99 (10 winters)
1999/2000–2006/07 (8 winters)
Less than −2.0
Less than −1.5
Figure 8a demonstrates the composite patterns (64 events) of the filtered SLP and isentropic potential vorticity (PV360) anomalies on the 360-K surface of potential temperature on day 0. Here, day 0 denotes the peak day when the MDI exhibits a minimum value. A remarkable negative SLP anomaly of less than −16 hPa can be seen to the south of the Kamchatka Peninsula, and its center shifts slightly northward from the mean position of bomb cyclones whose rates of development reached the maximum because bomb cyclones usually record a minimum SLP value after rapidly developing. Interestingly, a positive anomaly of SLP in excess of +4 hPa is observed around the Aleutian Islands, even though the composite is made with reference to MDI, which may be interpreted as the remote forcing of bomb cyclones. Looking at the PV360 distribution, a positive anomaly is located over northern Japan to the west of the center of the negative SLP anomaly. As is well known, synoptic-scale disturbances often develop rapidly, coupled with high potential vorticity anomalies at the upper level [e.g., Shapiro et al., 1999; Wang and Rogers, 2001]. Such a geographical relationship between the two suggests that an upper-level positive PV360 anomaly may trigger the rapid intensification of extratropical cyclones over the western North Pacific sector.
 In a similar fashion, the composite anomalies of the filtered 200-hPa velocity potential and OLR on day 0 are revealed in the middle. A strong anomalous divergence is appreciably evident in almost the same region as the negative SLP anomaly to the south of the Kamchatka Peninsula. Thus explosively developing extratropical cyclones, such as bomb cyclones, usually have strong divergent motion in the upper troposphere in association with the rapid decrease of SLP. On the northern half of the divergent anomaly, a negative OLR anomaly is zonally extended, which is a direct manifestation of a synoptic-scale cloud system that belongs to extratropical cyclones. Another distinctive negative OLR anomaly is seen just to the north of the Hawaiian Islands, accompanied by a positive PV360 anomaly, which will be discussed later.
Figure 8c shows the composite pattern of the filtered 200-hPa stream function anomalies on day 0. The 31-day running averages of daily zonal wind at 200-hPa level are also exhibited on the same day. An intriguing feature is the downstream development of stream function anomalies along the upper-level jet over the North Pacific. It is probable that bomb cyclones can excite downstream teleconnections as vorticity sources. Another feature is the presence of a wavetrain pattern along the South Asian waveguide [e.g., Hsu and Lin, 1992; Chang and Yu, 1999; Branstator, 2002; Watanabe, 2004]. Such a stationary wave pattern through the subtropical route may serve as a trigger for the generation of bomb cyclones in the vicinity of Japan. These issues are addressed in the next two subsections.
4.2. Role of the Subtropical Teleconnection in the Rapid Intensification of Extratropical Cyclones
Figure 9 demonstrates the composite anomalies of the filtered 200-hPa stream function from day −2 to +2 together with wave activity fluxes at the same level defined by Takaya and Nakamura . The horizontal component of their wave activity fluxes is defined as
where u is the wind velocity of horizontal basic flow (u, v), ψ is the stream function, and primes and subscripts denote perturbations and partial derivatives, respectively. Here the 31-day filtered data are used as the basic field in computing the wave activity fluxes. In a similar way, the composite patterns of the filtered PV360 anomalies are also shown in Figure 10. In the stream function field, a wavetrain pattern prevails from the Arabian Peninsula through the southern China on day −2, accompanied by eastward wave activity fluxes. From day −2 to day −1, northeastward fluxes are evident over southern China and the East China Sea, and a cyclonic circulation anomaly is also found over northern Japan. The stationary wave pattern over South Asia decays rapidly from day 0 to day +2. In the PV360 field, negative and positive anomalies are distributed along the South Asian waveguide from day −2 to day 0, corresponding to the above stationary wave pattern. On the other hand, a positive anomaly moves eastward from eastern Siberia and passes across northern Japan on day 0. This positive anomaly is expected to induce ascent motion in the lower troposphere to the east of its center, facilitating the rapid enhancement of extratropical cyclones in that vicinity. Judging from Figures 9 and 10, the positive PV360 anomaly is stimulated by a subtropical teleconnection along the South Asian waveguide. The stationary Rossby wave propagation over South Asia is considered to be a prominent trigger for the rapid reinforcement of extratropical cyclones, thus resulting in the generation of bomb cyclones over the western North Pacific sector.
 Similar ideas have already been proposed by Chang and Yu , Hoskins and Hodges , and Chang . Chang  postulated that a combination of two teleconnections along the subpolar and subtropical jets over the Eurasian continent contributes to the further development of extratropical cyclones over the western North Pacific. However, our results reveal less contribution of subpolar teleconnections across the Eurasian continent to the generation of bomb cyclones. Of the 64 extracted events, 18 are for December, 25, for January, and 21, for February, and the number of events is comparatively large in middle and late winters. As also indicated in Figure 1, bomb cyclones tend to originate and grow rapidly to the south of Japan, especially in middle and late winters. The development of such bomb cyclones might be more strongly affected by the subtropical teleconnection than by the subpolar one.
 It is not clear how the stationary Rossby waves propagating toward Japan are trapped in the subtropical waveguide over South Asia. Various types of forcing seem to excite those waves. Looking at Figure 8b again, a negative OLR anomaly is appreciably pronounced in the vicinity of the South China Sea on day 0, implying active tropical convection there. As an anomalous anticyclonic circulation at the 200-hPa level over southern China disappears from day 0 to day +2, the negative OLR anomaly also dissipates (figure not shown). Such tropical forcing appears to play a partial role in forming the upper-level anticyclonic circulation anomaly over southern China through the Matsuno-Gill type response [Matsuno, 1966; Gill, 1980] and hence modulates the subtropical teleconnection over the South Asian waveguide on submonthly scales.
4.3. Forcing of Bomb Cyclones to the Downstream Teleconnection
 Another noticeable feature of Figure 9 is the downstream development of a wavetrain pattern originating from the vicinity of Japan. The wavetrain pattern appears initially on day −1 and is well organized on day 0, also accompanied by remarkable southeastward wave activity fluxes. Further downstream development occurs from day 0 to day +2, resulting in the enhancement of an upper-level cyclonic circulation anomaly over the Hawaiian Islands. With a dominance of the stationary wave propagation, a positive PV360 anomaly gradually travels toward the Hawaiian Islands from day −1 to day +2. These features strongly indicate that the bomb cyclones excite stationary Rossby waves as synoptic-scale vorticity forcing, thereby giving rise to large-scale anomalous circulation over the central North Pacific region.
 As already stated in Figure 8b, a salient negative OLR anomaly is found near the Hawaiian Islands, which is accompanied by an upper-level divergent anomaly. A close relationship between the two persists until day +2 at least (figure not shown), indicating enhanced convective activity. The positive PV360 anomaly around the Hawaiian Islands, which is maintained by the stationary wave propagation across the central North Pacific, appears to induce active convection in that vicinity. For further confirmation of the above, Figure 11 shows the occurrence of a typical bomb cyclone as well as the presence of an MDI with an extreme value. Figure 11 shows the filtered OLR anomalies and PV360 from day −1 to day +2. Note here that day 0 corresponds to 8 January 2007. From day −1 to day +2, a high PV360 area in excess of 6 PVU intrudes equatorward over the central North Pacific Ocean basin and covers the entire region of the Hawaiian Islands. As the meridionally elongated high PV360 area approaches the islands from the northwest, a negative OLR anomaly rapidly becomes prominent there and reaches the minimum value of about −90 W m−2 on day +2. The feature is very similar to Figure 6 of Kiladis and Weickmann , who presented the case of 15 January 1987.
 It is well known that severe storms accompanied by subtropical cyclones, which are referred to as kona storms [Daingerfield, 1921; Simpson, 1952; Ramage, 1962], occasionally hit the Hawaiian Islands during the Northern Hemisphere cold season (October–March) and cause serious damage to crops, daily life, and economic activities in Hawaii. The rapid development of such surface subtropical cyclones results from the intrusion of upper-tropospheric disturbances of extratropical origin into the subtropics [e.g., Kiladis and Weickmann, 1992; Otkin and Martin, 2004]. Caruso and Businger  analyzed 70 upper-level lows that appeared during the 1980–2002 period and noted that 43 out of 70 developed into kona lows. Taken together with the findings of Caruso and Businger , the results of Figures 10 and 11 reflect a successive dynamic process in which the equatorward advection of higher potential vorticity from the midlatitudes triggers surface cyclogenesis in the vicinity of the Hawaiian Islands, eventually bringing about the genesis of kona storms. Lau and Lau  first pointed out the possibility that the origin of large-scale anomalous circulation related to kona storms might be traced upstream to cold-air outbreaks in East Asia several days earlier. Our results derived from this study not only strongly support their idea but also add more detailed features. When a particularly pronounced bomb cyclone comes to its mature stage to the northeast of Japan, its accompanying strong upper-level divergence has the ability to emanate stationary Rossby waves out of it as vorticity forcing. The stationary wave packets developing downstream and southeastward over the North Pacific Ocean basin lead to equatorward advection of higher potential vorticity from the midlatitudes. As the resulting upper-level positive potential vorticity anomaly approaches the Hawaii Islands, a synoptic-scale surface disturbance is induced and develops rapidly in that vicinity through a coupling process. It is similar to kona storms, which cause weather hazards in Hawaii. This successive dynamic process usually requires several days. Such a remote and time-lag relationship between the bomb cyclone prevailing near Japan and the kona storm in Hawaii indeed represents a good example of the so-called teleconnection phenomena.
 In section 3, we demonstrated that the East Asian winter monsoon variability strongly modulates bomb cyclone activity over the western North Pacific. The monsoonal forcing of the bomb cyclone activity, rather than the number and intensity of cyclones, is evident regarding its prevailing area. Once the monsoon becomes strong, lower-tropospheric baroclinicity is enhanced along the warm currents (i.e., the Kuroshio Current and the Kuroshio Extension) through air-sea heat exchange. Bomb cyclones with their maximum developing rates tend to concentrate within the enhanced baroclinic zone. Conversely, the cyclones disperse widely over the western North Pacific when the monsoon is weak. These results allow us to conclude that the change in the large-scale environment, which is equivalent to the East Asian winter monsoon circulation in this study, modulates a synoptic-scale disturbance activity in terms of its concentration and dispersion. We examined the horizontal gradient of equivalent potential temperature (∇θe) and the maximum Eady growth rate as measures of baroclinicity in the lower troposphere. The theory of Eady growth rate is well established [Lindzen and Farrell, 1980], but it does not consider the effect of moist baroclinic process. As an alternative indirect measure, ∇θe at 925-hPa is also used in this study to show the magnitude of the near-surface baroclinicity although it is not necessarily the most appropriate measure. Significant differences in the lower tropospheric baroclinicity between the strong and weak monsoon phases, based on these two measures, are evident particularly along the Kuroshio Current from the southern East China Sea to the Pacific coast of Japan, suggesting that enhanced lower-tropospheric baroclinicity along the Kuroshio Current, associated with the strong monsoon circulation, contributes to development of extratropical disturbances at early stages as a favorable background state [e.g., Xie et al., 2002]. However, such significant differences are not very obvious in the vicinity of the Kuroshio Extension region east of Japan. Alternatively, significant differences in static stability within the atmospheric boundary layer, associated with increased heat and moisture supply from the warm currents (see Figures 3c and 4c), may play an additional role in further development of the cyclones around the Kuroshio Extension region. One of the important and difficult questions is what physical processes are essential for regulating the wintertime variability of bomb cyclone activity over the western North Pacific, but this study could not necessarily present clear evidence of how themodynamic and dynamic processes, such as the heat and moisture supply into the atmospheric boundary layer and the baroclinicity in the lower troposphere, actually modulate the bomb cyclone activity. More physically and theoretically approaches may be required to address this issue.
 This study also focused on the role of warm currents in the bomb cyclone activity by examining the changes in surface sensible and latent heat fluxes. From our results, it could be concluded that the warm currents play a passive role in the bomb cyclone activity; however, such a conclusion would not be completely accurate. It should be noted that our composite maps were constructed on the basis of the strong and weak phases of the East Asian winter monsoon circulation in order to inspect the response of the bomb cyclone activity to the monsoon variability. It is quite possible that, as a result, the active role of the warm currents in the cyclone activity was masked. To clarify such an active role, further analyses are necessary on the basis of the variability in the warm currents.
 On submonthly timescales, both the response and forcing of bomb cyclones over the western North Pacific to the large-scale circulation are statistically detected by composite analyses, although only the monsoonal forcing of bomb cyclone activity is considered on interannual and intermonthly scales. We emphasize that the origin of large-scale anomalous circulation associated with bomb cyclones developing around Japan can be traced upstream along the South Asian waveguide several days earlier. It is certain that the stationary Rossby wave packets propagating eastward along the Asian jet over South Asia occasionally trigger the rapid growth of a synoptic-scale disturbance in the vicinity of Japan. Once it becomes a bomb cyclone that has pronounced upper-level divergence, it can, in turn, stimulate newly stationary Rossby waves downstream from the Asian jet exit region. One of the major findings in this study is the presence of a possible interaction on submonthly timescales between synoptic-scale disturbances over the western North Pacific and large-scale circulation. The bomb cyclone is indeed a key phenomenon for such a dynamic interaction. Alternatively, the behavior of bomb cyclones over the western North Pacific region is also important for weekly weather forecasting of downstream countries. As seen in Figures 10 and 11, the composite maps constructed by the 64 events obviously account for the remote impact of the bomb cyclones on the wintertime weather around the Hawaii Islands. If we examine each event in more detail, however, it can be recognized that there also exist cases in which stationary Rossby waves propagate toward the North American continent. The reason that stationary waves have a tendency to propagate toward the Hawaii Islands rather than the North American continent will require further detailed study.
 We investigated possible interactions between large-scale wintertime circulations and bomb cyclones prevailing over the western North Pacific Ocean using JRA-25, JCDAS, and J-OFURO2 data. Both the response and forcing of bomb cyclones over the western North Pacific to the large-scale circulation were examined on submonthly timescales, whereas the monsoonal forcing of the bomb cyclone activity was specifically highlighted on interannual and intermonthly timescales. The major findings in this study are briefly summarized as follows.
 1. The East Asian winter monsoon variability strongly modulates the bomb cyclone activity around the Japan Sea, the Kuroshio Current, and the Kuroshio Extension in terms of its concentration and dispersion rather than the number and intensity of bomb cyclones.
 2. At the strong monsoon phase, the bomb cyclone activity tends to concentrate in the vicinity of the Kuroshio Current and the Kuroshio Extension just near Japan, while, when the monsoon is weak, it disperses in a broader area.
 3. The predominance of the monsoon increases the heat and moisture supply from the Kuroshio Current and the Kuroshio Extension, facilitating unstable conditions within the atmospheric boundary layer and intensifying the baroclinicity (larger horizontal gradient of equivalent potential temperature and larger maximum Eady growth rate) in the lower troposphere. All these factors play an influential role in inducing bomb cyclones, particularly along the warm currents.
 4. On submonthly timescales, the stationary Rossby wave packets propagating eastward along the South Asian waveguide trigger the rapid growth of a synoptic-scale disturbance in the vicinity of Japan, resulting in the generation of bomb cyclones in that vicinity.
 5. An appreciably pronounced bomb cyclone near Japan is capable of exciting stationary Rossby waves downstream from the Asian jet exit region as synoptic-scale vorticity forcing. The downstream development of such a teleconnection induces surface cyclogenesis in the vicinity of the Hawaiian Islands by leading to the equatorward advection of higher potential vorticity from the midlatitudes, eventually bringing about the genesis of kona storms.
 The results derived from this study allow us to understand the dynamic interaction on submonthly timescales between synoptic-scale disturbances developing explosively over the western North Pacific Ocean and large-scale circulation. It is uncertain, however, whether a similar dynamic interaction is also dominant over the western North Atlantic Ocean. Although 506 bomb cyclones over 84 months during 28 winters were analyzed in this study, the long-term tendency of the bomb cyclone activity was not explicitly examined. The question of how the activity of bomb cyclones is associated with recent global warming also remains to be answered. Further intensive studies on these issues are required.
 We thank Satoshi Iizuka for his comments, which were very helpful in preparing an earlier version of this manuscript. Comments by two anonymous reviewers were extremely helpful. This research was supported by the Japan Science and Technology Agency, grants-in-aid (18540432) from the Japanese Ministry of Education, Science, Sports, and Culture, and the Mitsubishi Foundation for the Promotion of Science.