Journal of the Meteorological Society of Japan, Vol. 88, No. 3, pp. 313--324, 2010. 313 DOI:10.2151/jmsj.2010-304

Role of Large-Scale Circulation in Triggering Foehns in the Hokuriku District of Japan during Midsummer

Yukiko SHIBATA

Graduate School of Science and Engineering for Education, University of Toyama, Toyama, Japan

Ryuichi KAWAMURA

Section of Earth and Environmental Systems, Graduate School of Science and Engineering for Research, University of Toyama, Toyama, Japan

and

Hiroaki HATSUSHIKA

Toyama Prefectural Environmental Science Research Center, Toyama, Japan

(Manuscript received 6 August 2009, in final form 8 February 2010)

Abstract

The relationships between large-scale circulation and foehns observed during midsummer in Hokuriku dis- trict, located on the Japan Sea side of central Japan, are examined using Japanese long-term reanalysis project data, with additional data from the Japan Meteorological Agency climate data assimilation system. All foehn events are classified into two types: a tropical (TC)-induced foehn and an (EC)- induced foehn. The occurrence of the TC type is attributed to a combination of a and its induced tele- connection pattern, the Pacific-Japan (PJ) pattern, in the lower troposphere. Local intensification of the North Pacific high just east of Japan, accompanied by the dominance of the PJ pattern, can in turn force a typhoon track to shift westward. The northward migration of the typhoon along the western periphery of the locally en- hanced high strengthens a zonal pressure gradient across central Japan, thus producing a foehn. In contrast, an upper-level teleconnection along the Asian jet serves as a prominent trigger of the occurrence of an EC-type foehn. Stationary Rossby wave packets propagating eastward along the upper-level Asian waveguide facilitate not only the westward development of the North Pacific high but also the development of an extratropical cy- clone in the vicinity of the Japan Sea by leading to the equatorward advection of higher potential vorticity from high latitudes. Both developments are crucial for the reinforcement of a northwest-southeast pressure gradient in the lower troposphere around Japan, thus providing a favorable condition for a foehn event.

1. Introduction foehns depend on the configuration of the land, the pressure field at the level of mountain peaks, and The foehn phenomenon is one of the most famil- static stability (e.g., Yoshino 1976; Ikawa and Na- iar local winds. Previous studies have indicated that gasawa 1989; Saito and Ikawa 1991; Saito 1993). When a severe foehn occurs during midsummer, Corresponding author: Ryuichi Kawamura, Section of it not only damages a variety of crops, including Earth and Environmental Systems, Graduate School grains, vegetables, and fruits, but also harms peo- of Science and Engineering for Research, University ple’s daily lives and influences economic activities of Toyama, 3190 Gofuku, Toyama 930-8555, Japan. E-mail: [email protected] through an increase in daily electricity consump- 6 2010, Meteorological Society of Japan tion. Furthermore, much attention has been paid 314 Journal of the Meteorological Society of Japan Vol. 88, No. 3

East Asia should be considered to properly under- stand the dynamic processes associated with the foehns. Furthermore, the large-scale anomalous cir- culation in East Asia may regulate the activity of synoptic-scale disturbances around Japan to a large extent. If so, the anomalous circulation appears to indirectly influence the occurrence of foehns through the activity of synoptic-scale disturbances. It is also meaningful to inspect whether the synoptic-scale disturbances, such as the tropical cy- clones, can in turn a¤ect the behavior of the North Pacific high. In the present study, using the criteria of Inaba et al. (2002), we first extracted midsummer foehn events that occurred in Hokuriku district of Japan during the past three decades. Next, we conducted composite analyses of atmospheric variables with respect to -induced and extratropi- cal cyclone-induced foehns, and examined how the large-scale circulation contributes to the occurrence of foehns and whether there is a possible interaction between synoptic-scale disturbances and large-scale Fig. 1. Geographical map of the area of cen- circulation. tral Japan analyzed in this study. 2. Data and analysis To investigate large-scale circulation in associa- to the occurrence of foehns in order to prevent tion with the occurrence of foehns, we used 6-hour large fires caused by extraordinarily strong winds data from the Japanese Long-term Reanalysis proj- and low humidity. Since the track is in the vi- ect (JRA-25) (Onogi et al. 2007) based on a spatial cinity of Japan (e.g., Wallace et al. 1998), a number resolution of 1.25 longitude by 1.25 latitude for of extratropical or tropical tend to pass the period 1979–2004, as well as data from the Ja- through the Japan Sea. Although the center of the pan Meteorological Agency Climate Data Assimi- storm track shifts northward during midsummer, lation System (JCDAS) with the same resolution part of the cyclones still migrate across the Japan for the period 2005–2008. Since the JCDAS oper- Sea. When a well-organized cyclone approaches, ates with the same data assimilation system as the Toyama Plain in Hokuriku district of Japan JRA-25 to generate quasi-real-time products, com- often experiences a foehn because southerly winds, paratively homogeneous atmospheric circulation in association with a strong pressure gradient, pre- data are available for a period of over 30 years vail across the Hida Mountains and other high (1979–2008). To supplement in situ observations mountains (see Fig. 1) (e.g., Arakawa 1988). In near tropical cyclones, wind profile data, retrieved terms of synoptic-scale phenomena, it is well using operational tropical cyclone best-track data, known that an extratropical or tropical cyclone are used in the JRA-25 (Hatsushika et al. 2006). It can trigger a foehn in Hokuriku district, although, is thus expected that the representation of tropical of course, other factors such as static stability are cyclones by the JRA-25 is realistic, and the JRA- also necessary. We anticipate, on the other hand, 25 may be useful for examining the dynamic link that the behavior of the North Pacific high also between tropical cyclone activity and foehns. In plays a role in the occurrence of foehns during mid- addition, we used the typhoon track data compiled because the development of the high to- by the Japan Meteorological Agency (JMA). The ward Japan provides a favorable condition for en- JMA defines a typhoon as a tropical cyclone with hancing the zonal pressure gradient across central a maximum wind speed exceeding 17.2 m s1 in Japan. Thus, not only the synoptic-scale phenom- the vicinity of its center. ena but also large-scale anomalous circulation in Surface air temperature, wind velocity and direc- June 2010 Y. SHIBATA et al. 315

Fig. 2. Year-to-year variations in the number of foehn events () at the Toyama observatory in mid- summer (July–August) during the 30-year period 1979–2008. Heavy, light, and very light shadings denote the tropical cyclone (TC) type, the extratropical cyclone (EC) type, and others, respectively. Also exhibited are interannual variations in the July–August mean surface air temperature (solid line) at the same obser- vatory and its 30-year average (dashed line).

tion, and precipitation data were also used. These riety of crops. The fourth criterion was also im- data were derived from 266 AMeDAS (Automated proved because we wanted to highlight foehns that Meteorological Data Acquisition System) stations persist for a large portion of the day. of the JMA in the mainland of Japan to describe Following the above guidelines, we identified the regional features of foehns observed in Hokuriku number of foehn events during each midsummer district. Using JMA observational station data for period. The total is 55 foehn events. Figure 2 shows the period 1979–2008, we defined a distinctive the year-to-year variations in the frequency of foehn seen daily in Toyama (36 42.50 N, 137 foehn events observed in Toyama in midsummer 12.10 E) in midsummer (July–August), considering during the period 1979–2008. The July–August the following characteristics of a foehn: mean surface air temperatures are also shown. The frequency of foehns in midsummer di¤ers consider- i) daily maximum temperature greater than 33C, ably from year to year. It also turns out that the ii) daily minimum relative humidity less than 45%, number of foehn events is not very correlated with iii) daily mean wind speed greater than 3 m s1, the interannual variability of the mean air tempera- and ture. This may be because the ratio of hot days due iv) prevailing wind with southerly components to foehns to all the hot days exceeding 33C is only (SSW, S, SSE, and SE) persisting for 12 hours 13.3%. Since a tropical or extratropical cyclone is or more. expected to trigger a foehn, it is meaningful to ex- These criteria are similar to those proposed by amine the relationships between large-scale circula- Inaba et al. (2002), but the first and fourth criteria tion and foehn occurrence on the basis of the type were modified in this study. Although Inaba et al. of synoptic-scale disturbance. In this study, we adopted 35C as the daily maximum temperature thus classified all foehn events into two types: trop- threshold, we lowered the threshold somewhat be- ical cyclone (TC)-induced and extratropical cyclone cause their study focused specifically on severe (EC)-induced foehns. The TC type involves a ty- foehns with the potential to seriously damage a va- phoon or tropical cyclone located within a specific 316 Journal of the Meteorological Society of Japan Vol. 88, No. 3

Fig. 3. Composite patterns of daily mean surface air temperature, wind speed and direction, and daily total precipitation in central Japan for (a) the tropical cyclone (TC) type and (b) the extratropical cyclone (EC) type. The contour interval is 2C, and heavy shading denotes regions with temperatures above 28C. The reference arrow is 5 m s1. domain (120 –135E, 25 –45N) at 09 LT on a winds are observed in Hokuriku district. During foehn event. Twenty-eight events were identified as the TC type, daily total precipitation exceeding this type, whereas most of the remaining cases (27 10 mm or 15 mm occurs on the windward side of events) were regarded as the EC type (25 events), the Hida Mountains, but no remarkable precipita- except for two cases in which associated extratropi- tion occurs around that mountain range during the cal cyclones were of tropical storm origin. Figure 2 EC type. It appears that such a di¤erence reflects also demonstrates the number of each type in the the di¤erence in the synoptic-scale disturbance past three decades. The validity of this classification types that give rise to the occurrence of foehns. As is shown in the next section. indicated in Fig. 4, a typhoon travels along the western periphery of the North Pacific high; conse- 3. Results quently, northward moisture transport in the lower Figure 3 shows the composite patterns of daily troposphere becomes prominent to the east of the mean surface air temperature, wind speed and di- typhoon through an enhanced zonal pressure gradi- rection, and daily total precipitation in central Ja- ent. Reinforcement of the low-level moisture supply pan for the TC and EC types. For both types, hot inland from the Pacific Ocean with the approach of areas in excess of 28C expand along the entire the typhoon seems to contribute to substantial pre- coast of the Japan Sea, whereas relatively low tem- cipitation on the windward side of the Hida Moun- peratures are indicated on the Pacific Ocean side. tains during the TC type. The high temperatures, accompanied by southerly The left panel of Fig. 4 reveals the composite winds, cover not only the Toyama Plain but also pattern of the 850-hPa geopotential height on foehn the entire Hokuriku district, confirming that our events and the tracks of or tropical cy- criterion successfully captures the foehns prevailing clones over the western North Pacific Ocean for in that district. The TC type is more evident than the TC type. These track data were provided by the EC type in terms of the magnitudes of tempera- the JMA. The specific domain used in identify- ture and wind speed. In particular, strong southerly ing the TC type is also shown. The typhoons origi- June 2010 Y. SHIBATA et al. 317

Fig. 4. Composite patterns of the 850-hPa geopotential height on the day of foehn occurrence and foehn- related cyclone tracks for (a) the TC type and (b) the EC type. The contour interval is 20 m. Also exhibited is the specific domain used for identifying the TC type. The mean position of the typhoons at the foehn events is plotted by a white circle. Note that the track is traced from the initial appearance through to the disappearance of the cyclone that triggered a foehn. nate over the warm pool region of the western ysis. Many extratropical cyclones occur over the North Pacific to the east of the Philippine islands Asian continent and then travel eastward or north- and then migrate northward or northwestward eastward. A remarkable low at 850 hPa, together along the western periphery of the North Pacific with the North Pacific high, is located to the west high. When the foehns predominate in Hokuriku of the Sea of Okhotsk, thus forming a northwest- district, most of the typhoons are located within southeast pressure gradient around Japan. The the East China Sea and the Japan Sea. As for the pressure gradient in the lower troposphere in the TC type, the typhoons pass through the west of the vicinity of Japan during the EC type has a more mainland of Japan, suggesting that the westward meridional component than the TC type. The ridge development of the North Pacific high leads to the of the North Pacific high expands westward into westward displacement of the typhoon tracks. It is the south of Japan, in contrast to that during the apparent that part of the North Pacific high is lo- TC type. Ishizaki and Takayabu (2009) also em- cally enhanced to the east of Japan, which is also phasized the influence of large-scale circulation on identified with the Bonin high. The ridge of the foehn occurrence over the Japan Sea side. Their high further extends into northern Japan and the Fig. 3 resembles a mixed pattern of the TC and the Sea of Okhotsk. EC types observed in the current study, probably With respect to the EC type, the tracks of extra- because they do not classify foehn events in terms tropical cyclones, along with the composite of the of synoptic-scale disturbance types. 850-hPa geopotential height on foehn events, are Figure 5 shows the composite anomalies of the shown in the right panel of Fig. 4. The initial ap- 850-hPa and 200-hPa geopotential heights from pearance of an extratropical cyclone is defined at day 4 to day þ2 for the TC type. It is noteworthy any grid point with a sea-level pressure at least here that anomalies are relative to the 1979–2008 1 hPa lower than the value at an adjacent grid average and a three-day-weighted running average point. In addition, surface weather maps provided is applied to the daily data to remove day-to-day by the JMA are also used to determine the occur- variations. At the 850-hPa level, a salient negative rence, track, and disappearance of individual cy- height anomaly located to the northeast of the Phil- clones. As stated earlier, extratropical cyclones of ippines moves northward from day 4 and arrives tropical storm origin were excluded from our anal- over the East China Sea around southern Japan on 318 Journal of the Meteorological Society of Japan Vol. 88, No. 3

Fig. 5. Composite patterns of the 850-hPa and 200-hPa geopotential height anomalies on day 4, day 2, day 0, and day þ2 for the TC type. Day 0 denotes the day of foehn occurrence in the Hokuriku district. The shading interval for the 850-hPa geopotential height is 10 m, and the contour interval for the 200-hPa geopotential height is 20 m. day 0, corresponding to the typhoon tracks. An Likewise, an equivalent barotropic structure is anomalous positive height to the east of Japan is more obvious in the central and eastern parts of already organized on day 4, and its magnitude the North Pacific than in the western one. reaches a maximum value of þ30 m on day 0. Figure 6 exhibits the spatial distributions of 850- These two opposite anomalies form a PJ-like pat- hPa wave activity flux from day 4 to day þ2, de- tern (e.g., Nitta 1987; Huang and Li 1987; Kawa- fined by Takaya and Nakamura (2001), together mura et al. 1996; Kosaka and Nakamura 2006; with the composite stream function anomalies at Ogasawara and Kawamura 2007). Since the vicin- the same level. Also exhibited is the climatological ity of Japan is characterized by a pronounced zonal 850-hPa zonal wind. Fluxes in the vicinity of the pressure gradient on day 0, southerly geostrophic typhoon-related cyclonic circulation anomaly have winds are expected to be dominant over central Ja- a prominent northward component from day 4 pan, which is consistent with the foehn occurrence. to day 2. Northeastward fluxes are also evident A notable feature at 200-hPa, on the other hand, to the south of Japan, thus contributing to the en- is the presence of an appreciably positive height hancement of an anticyclonic circulation anomaly anomaly just over northern Japan on day 0, indi- east of Japan. On day þ2, the wavetrain pattern cating a baroclinic structure in the vicinity of north- accompanied by eastward fluxes becomes robust ern Japan. This feature is similar to that shown in along the low-level westerlies over the central Fig. 3 of Yamada and Kawamura (2007), who ex- North Pacific. These results indicate that such a amined typhoon-induced PJ patterns from early wavetrain pattern is a manifestation of stationary summer to . From day 0 to day þ2, the sys- Rossby wave packets propagating eastward, and tematic wavetrain patterns across the central North its origin can be traced upstream to typhoons to Pacific Ocean are well established at both levels. the south of Japan, which is consistent with the June 2010 Y. SHIBATA et al. 319

Fig. 6. Composite maps of the 850-hPa stream function anomalies and wave activity flux on day 4, day 2, day 0, and day þ2 for the TC type. Also indicated is the climatological 850-hPa zonal wind. Day 0 de- notes the day of foehn occurrence in the Hokuriku district. The shading interval for the 850-hPa stream function is 1 105 m2 s1. The contour interval for the 850-hPa zonal wind is 2 m s1. The reference ar- row is 50 m2 s2. Fluxes of less than 10 m2 s2 are suppressed.

findings of Kawamura and Ogasawara (2006) and day 2 until day þ2. Associated with these fea- Yamada and Kawamura (2007). Excitement of the tures, a negative 850-hPa height anomaly also PJ pattern due to typhoons implies the westward becomes evident to the north of the Japan Sea, cor- development of the North Pacific high. The local responding to the development of extratropical cy- reinforcement of the high just east of Japan, which clones. The anomaly has its minimum value on is part of the PJ pattern, can in turn play an influ- day 1 (figure is not shown). It is also interesting ential role in the tracks of typhoons; that is, the ty- to note that the intensification of an anomalous phoon tracks, which usually run along the western positive height at 850 hPa is observed to the east periphery of the North Pacific high, are forced to of Japan from day 2 to day 0. Because its geo- shift westward. As a typhoon travels northward, a graphical position is similar to that in the case of zonal pressure gradient strengthens in the lower tro- the TC type, such intensification also means the posphere between the typhoon and the locally en- westward development of the North Pacific high. hanced high east of Japan, triggering a foehn in A combination of these contrasting anomalies gives the Hukuriku district. rise to the occurrence of foehns in Hokuriku district Figure 7 shows the composite patterns of the on day 0. 850-hPa and 200-hPa geopotential height anoma- One may wonder why the positive height anom- lies from day 4 to day þ2 for the EC type. At aly is enhanced in the lower troposphere to the 200 hPa, a wavetrain pattern develops along the east of Japan. Figure 8 reveals the spatial patterns Asian jet in the upper troposphere from day 4to of the 200-hPa wave activity flux from day 4to day 0. A negative 200-hPa height anomaly to the day þ2 for the EC type, along with the composite west of the Japan Sea is almost stationary from stream function anomalies at the same level. Also 320 Journal of the Meteorological Society of Japan Vol. 88, No. 3

Fig. 7. Same as in Fig. 5, but for the EC type. shown is the climatological 200-hPa zonal wind. On be di¤erent from that of the silk road pattern or day 4, eastward fluxes accompany an outstanding the WJ pattern. anticyclonic circulation anomaly located to the To further demonstrate the relationship between south of the Lake Baikal. It is conceivable that, the upper-level teleconnection and the developing from day 4 to day 0, eastward propagation of sta- extratropical cyclone around the Japan Sea, we tionary Rossby waves through the upper-level show in Fig. 9 the composite patterns of isentropic waveguide forms a pronounced wavetrain pattern potential vorticity (PV350) on the 350-K surface of and contributes to the local enhancement of an potential temperature and its anomaly from day anomalous anticyclonic circulation to the south of 4 to day þ2 for the EC type. In the PV350 anom- the Okhotsk Sea. The locally enhanced anticyclonic aly field, negative and positive anomalies are dis- circulation anomaly at the 200-hPa level represents tributed along about 40–50N, corresponding to an approximately equivalent barotropic structure the upper-level stationary wave pattern. A notable with the underlying anomaly at the 850-hPa level, feature is that a positive anomaly becomes promi- suggesting that the low-level positive height anom- nent to the north of the Korean peninsula from aly to the east of Japan is dynamically induced by day 2 to day 0, which implies the equatorward upper-level stationary waves. The silk road pattern advection of higher potential vorticity from the (Enomoto et al. 2003; Enomoto 2004) or the West high latitudes. It is well known that synoptic-scale Asia-Japan (WJ) pattern (Wakabayashi and Kawa- disturbances often develop, coupled with high po- mura 2004; Ogasawara and Kawamura 2007), both tential vorticity anomalies at the upper level (e.g., of which have the potential to cause extraordinary Hoskins et al. 1985; Shapiro et al. 1999; Wang and summertime weather in Japan, is well known as an Rogers 2001). This information suggests that an upper-level teleconnection along the Asian jet. One upper-level high potential vorticity anomaly in- may think that the upper-level teleconnection seen duces ascent motion in the lower troposphere to in the EC type can be identified with such a telecon- the east of its center. As already indicated in Figs. nection pattern, but its center of action appears to 4, 7, the negative 850-hPa height anomaly to the June 2010 Y. SHIBATA et al. 321

Fig. 8. Composite maps of the 200-hPa stream function anomalies and wave activity flux on day 4, day 2, day 0, and day þ2 for the EC type. Also exhibited is the climatological 200-hPa zonal wind. Day 0 denotes the day of foehn occurrence in Hokuriku district. The shading interval for the 200-hPa stream func- tion is 2 105 m2 s1. The contour interval for the 200-hPa zonal wind is 10 m s1. The reference arrow is 200 m2 s2. Fluxes of less than 50 m2 s2 are suppressed. north of the Japan Sea on day 0 is a direct manifes- and showed that stationary Rossby wave packets tation of the growing extratropical cyclones. Com- propagating eastward along the South Asian wave- paring Fig. 9 with Fig. 7, one can see that the neg- guide trigger the rapid growth of a synoptic-scale ative height anomaly is geographically located to disturbance in the vicinity of Japan, resulting in the east of the distinctive positive PV350 anomaly, the generation of an explosive cyclone in that vicin- suggesting that the upper-level high potential vor- ity. What we would like to emphasize with regard ticity anomaly facilitates the development of extra- to the EC type is that the stationary Rossby wave tropical cyclones over northeastern Asia. It is thus packets propagating eastward through the upper- considered that the upper-level stationary Rossby level Asian waveguide not only trigger the develop- wave propagation along the Asian jet serves as a ment of extratropical cyclones to the north of the prominent trigger for the growth of extratropical Japan Sea but also induce the local reinforcement cyclones, eventually contributing to the occurrence of the North Pacific high to the east of Japan. As of foehns in Hokuriku district. a result, a combination of both developments pro- Similar ideas have been proposed by Chang and duces foehns in Hukuriku district through an en- Yu (1999), Hoskins and Hodges (2002), Chang hanced horizontal pressure gradient across central (2005), and Yoshiike and Kawamura (2009), al- Japan. though their studies focused mainly on the North- 4. Concluding remarks ern Hemisphere . For instance, Yoshiike and Kawamura (2009) examined the activity of explo- We classified foehns occurring in Hokuriku dis- sively developing extratropical cyclones over the trict of Japan during midsummer into two types: a western North Pacific during the northern winter tropical cyclone (TC)-induced foehn and an extra- 322 Journal of the Meteorological Society of Japan Vol. 88, No. 3

Fig. 9. Composite patterns of isentropic potential vorticity (PV350) on the 350-K surface of potential temper- ature and its anomaly from day 4 to day þ2 for the EC type. Day 0 denotes the day of foehn occurrence in Hokuriku district. The shading and black contour intervals are 1.0 PVU and 0.5 PVU, respectively.

tropical cyclone (EC)-induced foehn. Using JRA- For the EC type, stationary Rossby wave packets 25 and JCDAS data, we examined how large-scale propagating eastward along the Asian jet contrib- circulation plays a significant role in foehn occur- ute not only to the westward development of the rence through the activity of synoptic-scale distur- North Pacific high but also to the development bances. The major findings of the present study are of an extratropical cyclone in the vicinity of the briefly summarized as follows: Japan Sea by leading to the equatorward advec- tion of higher potential vorticity from the high The westward development of the North Pacific latitudes. This combined e¤ect intensifies the high is apparent to the east of Japan for both the northwest-southeast gradient of pressure in the TC and the EC types, and it plays an influential lower troposphere around Japan, bringing about role in enhancing the low-level horizontal pres- the occurrence of foehns. sure gradient across the mainland of Japan. For the TC type, the excitement of the PJ pattern Our findings provide information useful for due to a typhoon gives rise to the local reinforce- understanding the occurrence and magnitude of ment of the North Pacific high just east of Japan, foehns on the Japan Sea side. In particular, the de- which can in turn force a typhoon track to shift gree to which the North Pacific high is locally en- westward. As a typhoon migrates northward hanced to the east of Japan is a key factor in both along the western periphery of the locally intensi- the development and the magnitude of a tropical or fied high, a zonal pressure gradient strengthens extratropical cyclone-induced foehn. This point has across central Japan, thus triggering a foehn in not been considered in terms of foehn occurrence Hukuriku district on the Japan Sea coast. until now, except by Inaba et al. (2002). The results June 2010 Y. SHIBATA et al. 323 derived from this study also contribute to the un- in the Hokuriku district of Japan in the 1999 sum- derstanding of possible interactions between sum- mer. J. Meteor. Soc. Japan, 80, 579–594. mertime large-scale circulation and synoptic-scale Ishizaki, N., and I. Takayabu, 2009: Climatological fea- disturbances developing in the vicinity of Japan, ture of foehn and three heating factors over north but details of their interactions are still unclear. of Central Mountain range in Japan by using NHRCM. Submitted to SOLA. Further studies on these issues are required. Kawamura, R., and T. Ogasawara, 2006: On the role of typhoons in generating PJ teleconnection patterns Acknowledgments over the western North Pacific in late summer. SOLA, 2, 37–40. Comments by two anonymous reviewers were ex- Kawamura, R., T. Murakami, and B. Wang, 1996: Trop- tremely helpful. This research was supported by the ical and midlatitude 45-day perturbations over the Grants-in-Aid (20240075) from the Japanese Minis- western Pacific during the northern summer. J. try of Education, Sports, Culture, Science, and Meteor. Soc. Japan, 74, 867–890. Technology. Kosaka, Y., and H. Nakamura, 2006: Structure and dy- namics of the summertime Pacific-Japan (PJ) tele- connection pattern. Quart. J. Roy. Meteor. Soc., References 132, 2009–2030. Arakawa, S., 1988: An introductory survey of local circu- Nitta, T., 1987: Convective activities in the tropical west- lation. Note on Meteorological Research, 163,1– ern Pacific and their impact on the Northern Hemi- 22, in Japanese. sphere summer circulation. J. Meteor. Soc. Japan, Chang, E. K. M., and D. B. Yu, 1999: Characteristics 65, 373–390. of wave packets in the upper troposphere. Part I: Ogasawara, T., and R. Kawamura, 2007: Combined ef- Northern Hemisphere winter. J. Atmos. Sci., 56, fects of teleconnection patterns on anomalous sum- 1708–1728. mer weather in Japan. J. Meteor. Soc. Japan, 85, Chang, E. K. M., 2005: The impact of wave packets 11–24. propagating across Asia on Pacific cyclone devel- Onogi, K., J. Tsutsui, H. Koide, M. Sakamoto, S. Ko- opment. Mon. Wea. Rev., 133, 1998–2015. bayashi, H. Hatsushika, T. Matsumoto, N. Yama- Enomoto, T., B. J. Hoskins, and Y. Matsuda, 2003: The zaki, H. Kamahori, K. Takahashi, S. Kadokura, formation mechanism of the Bonin high in August. K. Wada, K. Kato, R. Oyama, T. Ose, N. Manoji, Quart. J. Roy. Meteor. Soc., 129, 157–178. and R. Taira, 2007: The JRA-25 reanalysis. J. Enomoto, T., 2004: Interannual variability of the Bonin Meteor. Soc. Japan, 85, 369–432. high associated with propagation of Rossby waves Saito, K., and M. Ikawa, 1991: A numerical study of the along the Asian jet. J. Meteor. Soc. Japan, 82, local downslope wind ‘‘Yamaji-kaze’’ in Japan. J. 1019–1034. Meteor. Soc. Japan, 69, 31–56. Hatsushika, H., J. Tsutsui, M. Fiorino, and K. Onogi, Saito, K., 1993: A numerical study of the local down- 2006: Impact of wind profile retrievals on the anal- slope wind ‘‘Yamaji-kaze’’ in Japan. Part 2: Non- ysis of tropical cyclones in the JRA-25 reanalysis. linear aspect of the 3-D flow over a mountain J. Meteor. Soc. Japan, 84, 891–905. range with a col. J. Meteor. Soc. Japan, 71, 247– Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 272. 1985: On the use and significance of isentropic po- Shapiro, M. A., H. Wernli, J.-W. Bao, J. Methven, X. tential vorticity maps. Quart. J. Roy. Meteor. Soc., Zou, J. Doyle, T. Holt, E. Donall-Grell, and P. 111, 877–946. Neiman, 1999: A planetary-scale to mesoscale per- Hoskins, B. J., and K. I. Hodges, 2002: New perspectives spective of the life cycles of extratropical cyclones: on the Northern Hemisphere winter storm tracks. The bridge between theory and observations, in J. Atmos. Sci., 59, 1041–1061. The Life Cycles of Extratropical Cyclones, edited Huang, R. H., and W. J. Li, 1987: Influence of the heat by M. Shapiro and S. Gronas, pp. 139–186, Amer. source anomaly over the western tropical Pacific Meteor. Soc., Boston, Mass. on the subtropical high over East Asia. Proc. inter- Takaya, K., and H. Nakamura, 2001: A formulation of a national conference on the general circulation of phase-independent wave-activity flux for stationary East Asia, 40–51. and migratory quasigeostrophic eddies on a zon- Ikawa, M., and Y. Nagasawa, 1989: A numerical study ally varying basic flow. J. Atmos. Sci., 58, 608– of a dynamically induced foehn observed in the 627. Abashiri-Ohmu area. J. Meteor. Soc. Japan, 67, Wakabayashi, S., and R. Kawamura, 2004: Extraction of 429–458. major teleconnection patterns possibly associated Inaba, H., R. Kawamura, T. Kayahara, and H. Ueda, with anomalous summer climate in Japan. J. 2002: Extraordinary persistence of foehn observed Meteor. Soc. Japan, 82, 1577–1588. 324 Journal of the Meteorological Society of Japan Vol. 88, No. 3

Wallace, J. M., G.-H. Lim, and M. L. Blackmon, 1998: pattern from early summer to autumn as revealed Relationship between cyclone tracks, by the JRA-25 reanalysis. SOLA, 3, 65–68. tracks and baroclinic waveguides. J. Atmos. Sci., Yoshiike, S., and R. Kawamura, 2009: Influence of win- 45, 439–462. tertime large-scale circulation on the explosively Wang, C.-C., and J. C. Rogers, 2001: A composite study developing cyclones over the western North Pacific of explosive in di¤erent sectors of the and their downstream e¤ects. J. Geophys. Res., North Atlantic. Part I: Cyclone structure and evo- 114, D13110, doi:10.1029/2009JD011820. lution. Mon. Wea. Rev., 129, 1481–1499. Yoshino, M., 1976: Local wind bora. University of Tokyo Yamada, K., and R. Kawamura, 2007: Dynamical link Press, 289 pp. between typhoon activity and the PJ teleconnection