Journal ong>ofong> Natural Disaster Science, Volume 25, Number 2, 2003, pp47-56ong>Anong> ong>Anong>alysis ong>ofong> ong>Typhoonong> ong>9807ong> (ong>Vickiong>) ong>Basedong> on Surface Meteorological RecordsObtained from Fire StationsYasuo OKUDABuilding Research InstituteHirohiko ISHIKAWADPRI, Kyoto UniversityTakeshi FUJIIKyoto Sangyo University(Received for 3. Mar., 2003 and in revised form 5 Jan., 2003)ABSTRACTong>Basedong> on weather observation data collected at fire stations in the Kinki and Chubu districts ong>ofong> Japan, the surfacestructures ong>ofong> typhoon ong>Vickiong> were investigated. There were four local, high wind areas in the Kinki district,the instantaneous wind velocity exceeding 50m/s. High winds occurred in two cases; almost simultaneouslywhen minimum atmospheric pressures were reached and about one hour after they were recorded. A comparisonong>ofong> the local high wind areas and combined radar echo charts showed that the latter local high winds occurredaround a rain band that formed to the rear ong>ofong> the center ong>ofong> the typhoon. Detailed surface structures ong>ofong> the typhoonconcerned with temperature, humidity, precipitation, and atmospheric pressure were obtained. Weather observationdata recorded at fire stations were shown to be very useful for analyzing the surface meteorological structureswithin a typhoon.1. INTRODUCTIONong>Typhoonong> ong>Vickiong> made landfall on the coastline north ong>ofong> GoboCity, Wakayama Prefecture at about 13:00JST on September 22nd1998, just after typhoon Waldo which had made landfall atWakayama Prefecture the previous day. ong>Typhoonong> ong>Vickiong> coursedthrough the Kinki district producing severe damage in manyplaces, especially Nara Prefecture. Because this typhoon passedthe central part ong>ofong> the Kinki district, surface weather data wasrecorded not only at ong>ofong>ficial meteorological stations but at manyfire stations, schools, power stations, and other facilities.In recent years, local governments in Japan have installedweather data acquisition systems for fire fighting. Many fire fightingheadquarters have been equipped with automatic weatherobservation systems to help the prevention ong>ofong> disasters in localareas. There are tens ong>ofong> fire fighting headquarters in each prefecturein the Kinki and Chubu districts, and surface weather data arecollected by data acquisition systems in most ong>ofong> them. The weath-Table 1Fire station responsesKEY WORDS: typhoon ong>Vickiong>, fire station, weather situation, combined radar echo
48Y. OKUDA, H. ISHIKAWA AND T. FUJIIer observation network that includes these fire fighting headquartershas a denser distribution than that ong>ofong> the JMA (JapanMeteorological Agency) which has only one or two stations perprefecture. The average distance between the AMeDAS(Automated Meteorological Data Acquisition System) observationsites is about 20km whereas the average distance between the fireTable 2Total observation pointsfighting headquarters is a few kilometers in urban areas.Moreover, indispensable weather elements for the typhoon analysiswhich are not measured in the AMeDAS, the instantaneouswind velocity, atmospheric pressure, and humidity, are recorded intime series at the fire fighting headquarters. This weather observationnetwork based on fire fighting headquarters has importancefor more detailed analyses ong>ofong> surface meteorological structureswithin a typhoon.Uematsu et al. (1999, 2002) published a database for theweather observation network covering Aomori, Akita, Miyagi,Yamagata, and Fukushima prefectures in the Tohoku district.There has been no report, however, that verifies the utility ong>ofong> aweather observation network that collects weather data recordedby weather data acquisition systems that cover an extensive area,because the purpose ong>ofong> these systems seems to be to provideweather information only for local areas.Fig. 1 (a) Path ong>ofong> ong>Typhoonong> ong>Vickiong> The times are JST. Fig. 1 (b) Weather chart at 9:00JST on September 22nd 1998.Fig. 2 Locations ong>ofong> fire stations, meteorological observatories,AMeDAS, and power stationsFig. 3 Path ong>ofong> ong>Typhoonong> ong>Vickiong> in the Kinki DistrictThe change in wind direction at observation sites is clockwiseat the closed circles, anticlockwise at the open circles.
AN ANALYSIS OF TYPHOON ong>9807ong> (VICKI) BASED ON SURFACE METEOROLOGICAL RECORDS OBTAINED FROM FIRE STATIONS49Okuda et al. (2000a) asked for weather data after typhoonong>Vickiong>s attack in order to analyze in detail the surface weather situationduring that typhoon. Many responses were received asshown in Table 1. The response rates for Nara and Aichi prefectures,in particular, exceeded 80%. In addition, weather recordswere obtained from government weather service stations (JapanMeteorological Agency, 1998a) and from power stations in theKinki district. Totally, data was received from 253 observationsites (Table 2). Okuda et al. (2000a) determined the meteorologicalcharacteristics ong>ofong> typhoon ong>Vickiong> from these observation data.Okuda et al. (2000b) also discussed the high wind generatedbehind the center ong>ofong> typhoon ong>Vickiong>. Here, findings on time serieschanges in the weather elements seen in the typhoon passagecharts are discussed and compared with rainfall strength figuresfrom combined radar echo charts. The utility ong>ofong> weather observationnetworks at the fire stations is verified from the findings.centers such as the large cities ong>ofong> Osaka and Nagoya. Few firefighting headquarter observation sites are located in ravine areas,but there are many power station observation sites in those andcoastal areas ong>ofong> the Kinki district. They complement the observationnetwork based on fire stations. In the northern region ong>ofong> NaraPrefecture, where serious damage was caused by the severe windsong>ofong> typhoon ong>Vickiong>, there is only one meteorological observatory andthree AMeDAS stations, but as the number ong>ofong> fire fighting headquartersis 11, more detailed information on surface meteorologicalconditions could be obtained. This figure shows that the denseweather observation network based on fire stations that mainly hasdeveloped in urban or suburban areas is very useful for understandingsevere storm disasters such as this typhoon, in urbanareas.4. METEOROLOGICAL ANALYSIS2. GENERAL FEATURES OF TYPHOON VICKIThe course ong>ofong> typhoon ong>Vickiong> and weather chart at 9:00JST onSeptember 22nd (Japan Meteorological Agency, 1998b) are shownin Fig.1 (a) and (b), before landfall on the Kii Peninsula. A weaktropical depression, in the western sea ong>ofong>f ong>ofong> the PhilippinesLuzon Island at 15:00JST on September 17th, developed intotyphoon ong>Vickiong> at 21:00JST the same day. Development was gradualover the course shown in Fig.1 (a), and the typhoon struck theKii Peninsula ong>ofong> Japan about 13:00JST on September 22nd. Theatmospheric pressure at its center was recorded as 960hPa at landfall.Maximum wind velocity was 35m/s in the vicinity ong>ofong> the center.The radius ong>ofong> the severe wind area in which the wind velocitywas more than 25m/s was 190km on the southeast and 150km onthe northwest sides. The radius ong>ofong> the strong wind area in whichthe wind velocity was more than 15m/s was 700km on the southeastand 220km on the northwest sides.After landfall, typhoon ong>Vickiong> moved northeast over the centerong>ofong> the Kinki district weakening gradually as it approached ToyamaBay. Moreover, it made re-landfall in Yamagata Prefecture before22:00JST the same day. The central atmospheric pressure at thesecond landfall was 990hPa, and the maximum wind velocity nearthe center 25m/s. The radius ong>ofong> the strong wind area with a windvelocity ong>ofong> more than 15m/s was 460km on the southeast and190km on the northwest sides. It passed through Honshu to thePacific Ocean, becoming an extra tropical cyclone at 9:00JST onSeptember 23rd. This was not an abnormal typhoon in size orintensity, but it moved very fast and was accompanied by heavyrainfall owing to the front to the rear ong>ofong> the typhoons center. Thespeed (Japan Meteorological Agency, 1998b) ong>ofong> this typhoon wasabout 50-55 km/h immediately before landfall on the KiiPeninsula, thereafter increasing to about 75-85 km/h.3. SURFACE METEOROLOGICAL DATALocations ong>ofong> fire fighting headquarters, government weather servicestations, and power stations from which the observation datawere obtained, are shown in Fig.2. AMeDAS stations are spatiallyarranged approximately 17km apart, and there are one or twometeorological observatories per prefecture. In contrast, the firefighting headquarters observation sites are concentrated in urban4.1 Course ong>ofong> ong>Typhoonong> ong>Vickiong>Fig.3 shows the distributions ong>ofong> veering and backing ong>ofong> thechanges in wind direction. The thick lines represent the borderbetween them. This border divides the Kinki district into approximatelyeast and west regions. The line is equivalent to the courseong>ofong> the wind center ong>ofong> the typhoon. As the times when wind directionchanged were recorded at some anemograms in some observationsites, the hourly position ong>ofong> the circulation center in thetyphoon is shown in Fig.3. The positions ong>ofong> the atmospheric pressurecenter ong>ofong> the typhoon, determined at 10-minute intervals byobjective pressure analysis (Fujii et al., 1999) are shown by doublecircles. Fujii et al. used atmospheric pressure records from JMAstations. Triangles mark the positions determined by the JMA forintervals ong>ofong> one hour. Each typhoon center course meanders. Thecourses in the circulation and atmospheric pressure centers closelyresemble each other, but the course ong>ofong> the circulation center isabout 10 to 20km to the left ong>ofong> that ong>ofong> the atmospheric pressurecenter.4.2 Local High Wind EventsMean wind velocity distributions for 10 minutes (average 50to 00an hour) are shown for the time series from 13:00 to16:00JST in Fig.4 (a)-(d). At 14:00JST, when the atmosphericpressure center ong>ofong> the typhoon reaches around the border betweenOsaka and Nara prefectures, a local high wind area, in which themean wind velocity exceeds 20m/s, appears in the rear region aswell as on the right side ong>ofong> the typhoon. At a fire station inWakayama City, the maximum recorded wind velocity was32.4m/s at 14:20JST. Thereafter the local high wind area passedOsaka, Nara, and Mie prefectures and reached the western part ong>ofong>Aichi Prefecture and the southwest part ong>ofong> Gifu Prefecture. Meanwind velocities exceeding 30m/s were recorded; maximum windvelocity 31.9m/s at 15:48JST in Kameyama City, Mie Prefectureand 32.7m/s at 15:57JST in Kaitsu Town, Gifu Prefecture. As thewind velocity vector ong>ofong> Fig.4 shows, wind direction mainly wassoutheast to south in the local high wind area on the right side ong>ofong>the typhoon, whereas it was mainly south to west in the local highwind area within the rear area ong>ofong> the typhoon.Fig.5 shows the distribution ong>ofong> the daily maximum instantaneouswind velocity at each observation site. The local high windarea, in which the maximum instantaneous wind velocity exceeds
50Y. OKUDA, H. ISHIKAWA AND T. FUJII40m/s, spreads from the southern part ong>ofong> the Kinki district to thewestern part ong>ofong> the Chubu district. The recorded daily maximuminstantaneous wind velocities were 51.0m/s at 13:13JST in TanabeCity, 50.0m/s at 14:13JST in Wakayama City, 50.8m/s at 14:07JSTin Sumoto City, 59.5m/s at 14:59JST in Shinjo Town, and 56.4m/sat 15:14JST in Ueno City.Fig.6 shows the difference in the times ong>ofong> the maximuminstantaneous wind velocity and minimum atmospheric pressure ateach observation sites; Area A, little difference in occurrencetimes, Area B, a time difference ong>ofong> more than one hour. In TanabeCity the maximum instantaneous wind velocity and minimumatmospheric pressure were recorded at the almost same time,Fig. 5 Daily maximum instantaneous wind velocity distributions(m/s)Circles show the position ong>ofong> the atmospheric pressure centerong>ofong> the typhoon obtained from an objective pressure analysis(Fujii et al., 1999).Fig. 4 Mean wind vector distribution (m/s)Circles show the position ong>ofong> the atmospheric pressure centerong>ofong> the typhoon at each time in an objective pressure analysis(Fujii et al., 1999).Fig. 6 Difference in hours between the time ong>ofong> the daily maximuminstantaneous wind velocity and daily minimum atmosphericpressureArea A: little difference, Area B: more than one hourCircles show the position ong>ofong> the atmospheric pressure centerong>ofong> the typhoon obtained from an objective pressure analysis(Fujii et al., 1999).
AN ANALYSIS OF TYPHOON ong>9807ong> (VICKI) BASED ON SURFACE METEOROLOGICAL RECORDS OBTAINED FROM FIRE STATIONS 51Fig. 7 Distribution ong>ofong> the difference in temperature from the dailymean temperature ()Small circles show the position ong>ofong> the atmospheric pressurecenter ong>ofong> the typhoon obtained from an objective pressureanalysis (Fujii et al., 1999). The large circle in Fig.7 (c)denotes the large temperature gradient region.Fig. 8 Distribution ong>ofong> hourly precipitation (mm/h)The atmospheric pressure center position ong>ofong> the typhoon isthe same as in Fig.7.
52Y. OKUDA, H. ISHIKAWA AND T. FUJIIFig. 9 Distribution ong>ofong> specific humidity (g/kg)The atmospheric pressure center position ong>ofong> the typhoon isthe same as in Fig.7.The large circle in Fig.9 (b) denotes the high specific humidityregion.Fig. 10 Distributions ong>ofong> differences from maximum atmospheric pressuresbefore daily minimum atmospheric pressure (-hPa)The atmospheric pressure center position ong>ofong> the typhoon is thesame as in Fig.7.
AN ANALYSIS OF TYPHOON ong>9807ong> (VICKI) BASED ON SURFACE METEOROLOGICAL RECORDS OBTAINED FROM FIRE STATIONS53whereas in Sumoto City, Wakayama City, Shinjo Town, and UenoCity maximum instantaneous wind velocity was recorded aboutone hour after minimum atmospheric pressure.4.3 TemperatureThe distributions ong>ofong> air temperature from 13:00 to 16:00JSTare shown in Fig.7 (a)-(d). The absolute temperature value cannotbe compared among the stations because it depends on the altitudeong>ofong> each observation site. The distribution ong>ofong> the temperature differencefrom the daily mean value at each observation site thereforeis shown.The cold area spreads out on the western side close to the centerong>ofong> the typhoon. From 14:00 to 15:00JST, this cold area is distributedlike a hook from west ong>ofong> the center, indicative that cold airentered the center ong>ofong> the typhoon from the northwest. At this time,the isotherm interval became dense from southern OsakaPrefecture to northern Nara Prefecture, and the temperature gradientin space became very large; 0.2/km. The cold air invadingfrom the northwest side ong>ofong> the typhoon therefore brought a rapiddecrease in temperature across the border ong>ofong> Osaka and Nara prefectures.4.4 Precipitation and Specific HumidityHourly precipitation distributions from 13:00 to 16:00JST areshown in Fig.8 (a)-(d). As typhoon ong>Vickiong> approached, precipitationin the southern Kinki district rapidly increased. From 14:00 to15:00JST, the precipitation distributions made ahookas therainfall area from the western side ong>ofong> the typhoon surrounded thecenter ong>ofong> the typhoon. In contrast, on the east side ong>ofong> the typhoon,rainfall was less than several mm/h, and some sites had none.Thereafter, the rainfall area moved northeast together with thetyphoon. The hook-shaped rainfall area became illegible, no rainfalling in the Kinki district, except in the north.Specific humidity distributions from 13:00 to 16:00JST areshown in Fig.9 (a)-(d). While typhoon ong>Vickiong> was passing throughthe Kinki district, dry air entering the Kii Peninsula from thesoutheast met the damp air on ong>ofong> the typhoons northwest side. Asa result, the damp air region was a swirl around the eye ong>ofong> thetyphoon from 14:00 to 15:00JST. The high specific humidity,exceeding 20g/kg around the border between Wakayama andOsaka prefectures at 14:00JST seems to have been brought by therain band that developed on the west side ong>ofong> the typhoon. In contrast,the specific humidity on the east side ong>ofong> the typhoon wasreduced to 10-15g/kg by the dry air flowing in from the southwestong>ofong> the Kinki district.4.5 Atmospheric PressureFig.10 (a)-(d) shows the atmospheric pressure distributionsfrom 13:00 to 16:00JST. Atmospheric pressure values observed atmost ong>ofong> the fire stations were not normalized to the sea level value.Moreover, the accuracy ong>ofong> pressure values recorded at some firestations were not considered highly confident because they had notbeen ong>ofong>ficially approved by the JMA. The pressure differencebetween the atmospheric pressure at each recording time and thepeak pressure before passage ong>ofong> the typhoon therefore are given.The pressure differences are distributed in a rotating ellipticalform.5. THE COMBINED RADAR ECHO AND HIGHWIND AREA RELATIONSHIP IN THE KINKIDISTRICTThe combined radar echo charts together with observations atMurotomisaki and Takayasuyama sites were used in the comparisonwith maximum wind distribution. Fig.11 (a)-(p) shows thecombined radar echo charts for every 7 minutes and 30 secondsfrom 14:00 to 16:00JST and maximum instantaneous wind velocitiesare displayed as vectors based on the times when the maximumwas recorded. The combined radar echo at 14:45JST lacksan image. In Fig.6, the severe storm area mainly is divided intotwo areas, in which the minimum atmospheric pressures and maximuminstantaneous wind velocities were recorded at the almostsame time or maximum instantaneous wind velocities were recordedabout one hour after the minimum atmospheric pressures. Thefirst area corresponds to the local high wind area present on theright side ong>ofong> the typhoon, whereas the second is the local high windarea that formed to the rear ong>ofong> the center ong>ofong> the typhoon.Yamamoto et al. (1963) reported severe storms accompanied by acold front in their study ong>ofong> the 2nd Muroto typhoon. The highwind formed behind typhoon ong>Vickiong> seems to have the same characteras that in the 2nd Muroto typhoon.The combined radar echo chart at 14:00JST shows a band ong>ofong>strong rainfall on the western side ong>ofong> the typhoon. High windswere recorded mainly inside the rain band. The daily maximuminstantaneous west wind around the north ong>ofong> the rain band wasrecorded in Sumoto City, Hyogo Prefecture at 14:15JST. In the 6minutes from 14:15 to 14:21JST, a southerly high wind wasrecorded at Matsuzaka City, Mie Prefecture, but on the right sideong>ofong> the typhoon. A westerly high wind was recorded in WakayamaCity. These high winds at Sumoto and Wakayama cities wereinside the same rain band indicated as A in Fig.11. Thereafter, thelocal high wind area moved in the east-northeast direction togetherwith the movement ong>ofong> the rain band. At 14:37JST, the local highwind area had moved from the north ong>ofong> Wakayama Prefecture tothe west ong>ofong> Nara Prefecture.Very high winds blew in the south ong>ofong> the Nara basin and thewest-southwest area along the Yoshino River. The inflow angleacross the isobars was large along the Yoshino River, whereas itwas small in the south ong>ofong> the Nara basin. There were high windsfrom the south across Mt. Kongo on the border between Osaka andNara prefectures. The maximum wind velocity was 59.5m/s inShinjo Town, Nara Prefecture at 15:00JST. Only the wind directionrecorded at Shinjo Town was south-southeast. At 15:07JST,westerly high winds were recorded around the rain band from thesouthern Osaka Prefecture to northern Nara Prefecture. Withinjust 15 minutes, from 14:52 to 15:07JST, the wind direction acrossMt. Kongo had changed from west to south and again from southto west. Our field study in Nara Prefecture turned up many firsthandtestimonies: at first a southerly wind blew then it shifted to awesterly one, and there was much evidence ong>ofong> the damage done tohouses and trees by the westerly wind, as well as by the southerlyone. It therefore is believed that the rain band that had a rapidchange in the wind direction had passed through.From an analysis ong>ofong> the Doppler radar echo obtained at KansaiInternational Airport at 15:30JST, Sakakibara et al. (1999) showedthat a converging shear line existed near the rain band over Nara
54Y. OKUDA, H. ISHIKAWA AND T. FUJIISumoto CityWakayama CityMatsuzaka CityMr. Kongo Shinjo TownYoshino RiverFig. 11 Combined radar echo and daily maximum instantaneous wind vectorsThe rain band formed behind the typhoon center is shown as a boxed A.
AN ANALYSIS OF TYPHOON ong>9807ong> (VICKI) BASED ON SURFACE METEOROLOGICAL RECORDS OBTAINED FROM FIRE STATIONS 55Ritto TownShigaraki TownRain [mm/h]Fig. 11 (continued)0 1 4 16 32 64
56Y. OKUDA, H. ISHIKAWA AND T. FUJIICity. They reported that this converging shear line caused the highwind on the ground. Kawano et al. (1999) observed 3 componentsong>ofong> wind velocity in typhoon ong>Vickiong> based on MU radar system datafrom Shigaraki Town, Shiga Prefecture. At 15:53JST, observationwas stopped by a blackout, but they recorded a strong down draftong>ofong> 4-5m/s, at height ong>ofong> 2-3 km above Shigaraki Town. At thattime, the rain band seen on the combined radar echo chartsapproached Shigaraki Town. Maximum instantaneous wind velocitieswere 36.4m/s at 15:45JST in Ritto Town and 41.6m/s at16:16JST in Yokaichi City, both in Shiga Prefecture. The convergingshear line, down draft, and local high wind on the groundare considered to have been related.6. CONCLUSIONSong>Basedong> mainly on weather observation data obtained at fire stations,we analyzed the surface structures ong>ofong> typhoon ong>Vickiong>:1) The weather observation networks ong>ofong> fire stations wereextremely useful for the analysis because they were muchdenser than the observation networks ong>ofong> the JMA.2) Three courses shown for this typhoon in the Kinki districtwere meandering ones. The meanderings ong>ofong> the wind circulationand atmospheric pressure centers, in particular, were wellcorrelated.3) The local high wind area in the typhoon was divided into twoparts, one formed on the right side ong>ofong> the typhoon, the other inits rear area as shown by the typhoon tracking. The first localhigh wind area occurred in Wakayama and Aichi prefectures,the second, in the central ong>ofong> Kinki district, causing widespreaddamage.4) The temperature distribution showed that cold air intruded thecenter from the typhoons northwest side. The temperaturegradient became very large across the border between Osakaand Nara prefectures at 15:00JST.5) The hourly precipitation distribution ong>ofong> showed a hook-shapedrainfall area within the typhoon.6) The specific humidity distribution showed that dry air fromthe southwest and damp air from the northeast flew in the centerong>ofong> the typhoon at 15:00JST.7) The isobar distribution in the typhoon ong>ofong>ten had a rotatingelliptical form.ACKNOWLEDGEMENTSA part ong>ofong> this study was supported by Grants-in-Aid forScientific Research from the Ministry ong>ofong> Education, Science,Sports and Culture (Research Representative; Prong>ofong>. MasaruMatsumoto). In our collection ong>ofong> the various weather data, wereceived much cooperation from stuff members ong>ofong> the JapanMeteorological Agency (JMA), Japan Weather Association(JWA), Japan Meteorological Business Support Center (JMBSC)as well as the fire fighting headquarters ong>ofong> the local governments,Kansai Electric Power Co. Inc. and Nihon Electric Instruments Co.Inc.REFERENCESFujii T., T. Hayashi and Y. Mitsuta, 1992. On an analysis and a wind distributionong>ofong> ong>Typhoonong> 9119 (Mireille), ong>Anong>nuals ong>ofong> Disaster PreventionResearch Institute, Kyoto University, 35 B-1, 183-191 (in Japanese).Fujii T. and Y. Mitsuta, 1999. Relation between wind velocity calculatedfrom atmospheric pressure field and observed wind velocity (Part 3):ong>Anong>alysis ong>ofong> ong>Typhoonong> ong>Vickiong>, ong>Anong>nuals ong>ofong> Disaster Prevention ResearchInstitute, Kyoto University, 42 B-1, 289-299 (in Japanese).Japan Meteorological Agency, 1998a. Ground weather observation data,area weather observation data and upper-air observation data,Meteorological Agency Monthly Report, September (in Japanese).Japan Meteorological Agency, 1998b. Heavy rain and severe storm causedby ong>Typhoonong>s Nos.6, 7, 8, 9, and fronts from 18th September to 2ndOctober, Disaster Weather Newsletter (in Japanese).Kawano N., M. Yamamoto and S. Fukao, 1999. Observation ong>ofong> ong>Typhoonong>ong>Vickiong> using an MU radar method, Preprint, ong>Anong>nual ScientificMeetings, the Meteorological Society ong>ofong> Japan, 223 (in Japanese).Nakajima C., Y. Mitsuta, Y. Gocho, M. Tanaka, T. Fujii and N. Monji,1980. On ong>Typhoonong> 7916, ong>Anong>nuals ong>ofong> Disaster Prevention ResearchInstitute, Kyoto University, 23 B-1, 87-111 (in Japanese).Okuda Y., J. Katsura, T. Fujii and T. Hayashi, 2000a. Meteorological characteristicsong>ofong> ong>Typhoonong> ong>9807ong> (ong>Vickiong>) observed at fire stations, Journalong>ofong> Wind Engineering, 82, 69-80 (in Japanese).Okuda Y., H. Ishikawa and T. Fujii, 2000b. High wind behind center ong>ofong>ong>Typhoonong> ong>Vickiong>, Proceedings ong>ofong> 16th National Symposium on WindEngineering, 19-24 (in Japanese).Sakakibara H. and Y. Tanaka, 1999. Severe storm caused by ong>Typhoonong>ong>Vickiong>, Preprint, ong>Anong>nual Scientific Meetings, the MeteorologicalSociety ong>ofong> Japan, 215 (in Japanese).Shaw N., 1918. The Travel ong>ofong> circular depressions and tornadoes and therelation ong>ofong> pressure to wind for circular isobars, MeteorologicalOffice, Geophysical Memoirs, 12, 19-44.Uematsu Y. and K. Sasaki, 1999. Wind measurements in the Tohoku district(Part 1 Aomori and Akita prefectures), Journal ong>ofong> WindEngineering, 78, 81-93 (in Japanese).Uematsu Y., T Moteki and M. Noguchi, 2002. Wind measurements in theTohoku district (Part 2 AMeDAS sites in Miyagi, Yamagata andFukushima prefectures), Journal ong>ofong> Wind Engineering, 91, 55-64 (inJapanese).Yamamoto R., Y. Mitsuta and K. Miyata, 1963. Some studies on ong>Typhoonong>Nancy (6118), ong>Anong>nuals ong>ofong> Disaster Prevention Research Institute,Kyoto University, 6, 113-127 (in Japanese).