Two typhoons Kaemi (200605) and Bopha (200609) crossing over Taiwan were continuously monitored by a wind profiler radar located on the lee side of the Central Mountain Range (CMR) of the island. Wind fields rotating systematically associated with the storm passage is pronounced for both typhoons. Nevertheless, significant wind shear takes place above and below the altitude of about 3.5 km owing to the mountain blocking before typhoon landing is reported in this paper for the first time. Winds deflected due to the orography behave diversely with respect to different typhoon tracks and strengths that may cause a secondary circulation to modify the original one. In this study, a moderate typhoon, Kaemi re-organizes the winds from the CMR blocking after landfall. On the other hand, weak typhoon Bopha shows a very complicated wind pattern structure from this typhoon-orography interaction. As the winds are disturbed dramatically, it dissipates over the sea soon after the typhoon center leaves Taiwan.
 There are about four typhoons encounter the high Central Mountain Range (CMR) of Taiwan per year. In Taiwan, a band of flat land locates along the west coast and the steep CMR orients north-northeast to south-southwest throughout the island. In summer, most of the invading typhoon comes from the east coast of Taiwan. With peaks higher than 3 km, the CMR usually deflects the tracks and modifies the structures of typhoons. Some observational studies [e.g., Brand and Blelloch, 1974; Wang, 1980; Tsay, 1994] have been reported on the interesting phenomena of typhoon-orography interactions. An extensive study was carried out by Wang  in which he examined the behavior of more than 120 typhoons (that threatened Taiwan between 1949 and 1979) and concluded that the low-level structure and the movement of typhoons experience significant modifications as they encounter Taiwan orography. Tsay  analyzed the hourly surface wind and pressure structure over the Taiwan area during two typhoons namely Betty and Nadine that invaded Taiwan in 1961 and 1971 respectively and found secondary circulation and a secondary low formed over western Taiwan when typhoon Betty moved close but still to the east of CMR while center of typhoon Nadine passed continuously over CMR. Moreover, several numerical modeling studies have also addressed the effect of the terrain on typhoons crossing Taiwan [e.g., Chang, 1982; Bender et al., 1987; Yeh and Elsberry, 1993a, 1993b; Wu and Kuo, 1999; Lin et al., 1999, 2005]. Recently, Lin et al.  proposed a conceptual model to explain track deflection and continuity for a westward-moving tropical cyclone (TC) encountering idealized topography representative of the Central Mountain Range of Taiwan.
 As the typhoon encounters the CMR from the south-eastern of Taiwan, the winds at the lower altitude levels have to deflect along the topography due to the mountain blocking. On the other hand, the wind fields at the upper altitudes are able to pass over the mountain range and appear mainly northeasterly winds owing to the counter-clockwise circulation of typhoon in the northern hemisphere. The blocking effect of the CMR to wind fields plays a dominant role in the formation of secondary circulation/low and devotes to the structure modifications of typhoon. However, the above studies do not provide concrete experimental evidence particularly how the winds changes with height owing to the CMR blocking as the typhoon approaches. The resolutions provided by numerical simulations are also not sufficient to investigate the detailed wind information near the TC center region.
 In this paper, we use the high height-time resolution wind-profiler (hereafter wind profiler radar or profiler) data to study the detailed winds and echo power variation associated with two typhoons crossing Taiwan. This profiler, operating at 1.3 GHz and located at the lee side of the CMR, is one of the most suitable instruments to investigate the wind fields of a tropical typhoon and which can operate quasi-continuously almost in all weather conditions. In Japan, a network of thirty-one L-band (1.3 GHz) wind profilers named Wind profiler Network and Data Acquisition System (WINDAS) [Kato et al., 2003] has been operated since June 2002. The observations of typhoon passages and tropical cyclone eye have been reported with that facility since then [Teshiba et al., 2004, 2005].
 An important issue associated with the observational studies of typhoons affecting Taiwan is the lack of good temporal and vertical/spatial resolution data. Most of the previous observational studies on this subject are based on traditional surface and radiosonde observations, which are highly inadequate in describing the important three-dimensional wind and structure changes associated with a typhoon as it approaches and attacks the island. We report in this study the significant influences of the topography on two typhoons pass over Taiwan by analyzing the height-time distributions of the horizontal winds and additionally it is possible to observe precipitating cloud information associated with the two typhoons by the signal-to-noise ratio profile measured by the vertical beam of the profiler.
2. Experimental Set-Up
 The DEGREWIND PCL1300 profiler was deployed at an industrial park (120.38°E, 22.6°N) close to Kao-Hsiung city in the southern Taiwan and operated continuously during the typhoon periods in order to observe the clear and precipitation echoes in the lower atmosphere. The main characteristics of this Doppler wind profiler radar are a 1290-MHz transmitted frequency, with a 4-kW peak pulse power, a 25-kHz pulse repetition frequency, and a 150-m pulse length. One vertical and four oblique beams, with an off-zenith angle of 17 degrees disposed every 90 degrees in azimuth, are swinging continuously to detect the winds. The altitude coverage is from 0.235 km to 6.543 km with a 150-m range resolution. 20 successive Doppler spectra obtained from a 128-point discrete Fourier transform are used to extract the first three moments of the atmospheric echoes. The information about the echo power intensity, radial winds as well as noise level is then deduced. We further incoherently average the derived products in the time domain to present the winds and the signal-to-noise ratio at every 30 minutes. From the profiles of the signal-to-noise ratio, the precipitation types are studied by the method of Williams et al. . For more details on technique information of this type of wind profiler, refer to Heo et al. .
3. Observational Results
Figure 1 depicts the topography of Taiwan and tracks (provided by the Central Weather Bureau of Taiwan) of two typhoons Kaemi (thick black line with star) and Bopha (thick black line with open circle) crossing Taiwan along with the wind profiler radar location (black color star). Due to the high altitude and complicated mesoscale topography associated with Taiwan's CMR, which has an average elevation higher than 2000 m and a dimension of 300 km × 100 km, significant variations in track and intensity occur as both typhoons approaches the island. According to these typhoon tracks, a clear turning of wind field from north-easterly to south-westerly associated with the passage of typhoon is anticipated from the wind profiler radar observations.
 The moderate typhoon Kaemi (200605) upgraded from a tropical depression on July 21, 2006. It then was made landfall at 1545 UTC (LT = UTC + 8) of July 24 near Cheng-Kung (22°94′N, 121°38′E), Taiwan and moved north-northwestward to dissipate over mainland China on July 25. We present in the Figure 2 the wind profiler observations starting from 0900 to 2400 UTC of July 24 that covered the time periods of typhoon approaching (from 0900 to 1544 UTC), landing (from 1545 to 2005) and leaving (from 2006 to 2400 UTC) Taiwan, respectively.
Figure 2 denotes the height-time distributions of the horizontal U-V component of the winds (Figure 2 (top)) and the signal-to-ratio (Figure 2 (middle)) observed by the wind profiler. The distance between the wind profiler location and the center of Typhoon is presented in Figure 2 (bottom) with respect to the progress of time and orography of the CMR where a positive (negative) value corresponds to the period when the Typhoon approaches (leaves). As we can see, the northeasterly wind was dominated above 3.5 km from 1030 to 1600 UTC but the northerly to northwesterly winds took place below 3.5 km while the typhoon center was about 200 km to 100 km away from the observation site. This demarcation of different wind directions at above and below 3.5 km altitude implies that the blocking effect of the CMR is significant below 3.5 km altitude even though the typhoon was still 200 km to the east of CMR and the dominance of blocking effect was prevailed till 1600 UTC of July 24 when the typhoon Kaemi was about to land.
 On the other hand, the horizontal distribution of precipitation associated with typhoon eyewall and rain bands (as measured by the signal-to-noise ratio shown in Figure 2 (middle)) shows there are two distinct rain bands existed from 1200 to 1300 UTC and 1330 to 1500 UTC respectively with a pronounced bright band at around 5 km altitude. After the landfall of typhoon center at about 1600 UTC, the bright bands are occasionally seen that indicates convective type of rainfall becomes dominant within the eyewall of typhoon Kaemi. From about 1630 UTC of July 24, the horizontal winds changed anti-clockwise to northwesterly and continued in this direction till 1830 UTC, when the typhoon center was closest (69 km) to the profiler, then changed to southwesterly direction from 1830 UTC. It is interested to notice that re-organization of typhoon wind structure and less rainfall indicates the weakening of blocking effect of the CMR after the typhoon center made landfall.
 The weak typhoon Bopha (200609) formed at about 1200 UTC of August 5, 2006 and also landed near Cheng-Kung at around 1920 UTC of August 8. It then proceeded westward until it decayed over the sea of Taiwan banks on August 9. Figure 3 shows the height-time distributions of horizontal winds, signal-to-noise ratio and the distance between observation site and Typhoon center as those shown in the Figure 2. Typhoon Bopha approached from 1200 to 1920 UTC, landfall from 1921 to 2320 UTC and moved from 2321 of 8 August to 0300 of August 9.
 Although the horizontal winds are more complicated than those of Typhoon Kaemi, it is clearly seen that the wind direction is northeasterly above 3.5 km altitude and mainly southwesterly below 3.5 km altitude from about 1530 to 1900 UTC of August 8. It is again evident from this observation that the blocking effect of CMR to the wind field is dominant below 3.5 km altitude. The signal-to-noise ratio profile belongs to approaching period of typhoon shows that a rain band from 1700 to 1830 UTC is observed with a bright band located near up to 5 km altitude.
 The center of typhoon Bopha landed at 1920 UTC and was estimated to have passed north of the wind profiler at around 2230 UTC with the minimum distance of about 22 km. It further weakened and left Taiwan at about 2320 UTC of August 8. From August 9, typhoon Saomai was approaching Taiwan and dominating the weather system.
 During 1700 UTC to 2300 UTC, the winds detected near surface are mainly from the east, which is consistent with the report of the Kao-Hsiung surface station of the Central Weather Bureau (not shown here). Those easterly winds seem “lift-up” the south-westerly winds below 3.5 km altitude where the blocking effect of CMR takes place before the typhoon landing. With the westerly jet-like winds centered at about 3 km altitude during 18 UTC to 22 UTC, a rotor-look structure with core at 700 m altitude or so is notified. A clear turning of wind fields as the typhoon reaches the closest distance of wind profiler around 2230 UTC is again observed. Nevertheless, this feature only takes place at the region above about 2 km altitude.
 The height-time distribution of signal-to-noise ratio shows two continuous rain bands from 1700 to 1830 UTC and 1900 to 2330 UTC, respectively, of 8 August with a continuous bright band at around 5 km altitude. Since both are characterized by a bright band at 5 km, these precipitation types are also stratiform ones [Williams et al., 1995]. For instance, the horizontal winds (belong to typhoon leaving period) after 2320 UTC behaved irregularly until the end of observations (0300 UTC). On the other hand, no continuous rain band is observed from 2330 UTC of 8 August, 2006 to 0300 UTC of 9 August, 2006.
 The CMR was found to play a key role for both wind structure and precipitation produced during the passage of two Typhoons. The height-time distributions of horizontal winds and signal-to-noise ratio associated with two typhoons crossed the central mountain range of Taiwan in 2006 are reported in this study. The two typhoons have crossed the CMR of Taiwan with different strengths and tracks. The typhoon-topography interaction has modified the wind fields of both typhoons during the different stages of typhoon invasion. With the wind profiler located at the lee side of CMR, we report the first study of the blocking effect due to the mountains during two typhoon passages. Following are the newly found characteristics investigated associated with this effect.
 Firstly, for both cases of typhoon, the blocking effect is siginificant up to the altitude of 3.5 km although the highest altitude of mountain that the typhoon has climbed over is about 2.5 km and 1.8 km, respectively, for typhoon Kaemi and Bopha. It may start from 200 km away from the observation site and then prevails utill the typhoon is about to land. Moreover, in the regions “shadowed” by the significant mountain blocking, the wind fields behave diversly depending on different typhoon. For the shadowed regions during typhoon Kaemi, it is suggested that the deflected winds are northerly to norther-westerly owing to be along with the mountains, which is seen in the Figure 2. This feature is not appliable to the typhoon Bopha since the winds are mainly south-westerly in the Figure 3.
 The secondary circulation/low is most apt to form when the storm is over the southeastern Taiwan and its adjacent ocean. Some secondary lows may even take over the primary circulation and makes the typhoon appears to jump forward through the island [Chang et al., 1993; Tsay, 1994]. Although detecting the secondary circulation/low by using only one wind profiler is difficult, cyclonic circulation over the west side of CMR at the approaching phase of these two typhoons do occur as usual according to the surface stations of Central Weather Bureau, Taiwan (not shown here). As typhoon Kaemi proceeds north-westward, the location of secondary circulation is further north and apart far from the wind profiler. For the westward moving typhoon Bopha with center almost pass over our observation site, the induced cyclone is to the west but is close to the wind profiler. It is suggested that the south-westerly winds are from the secondary circulation at the shadowed regions during typhoon Bopha approaching. Nevertheless, the near surface easterly winds from typhoon Bopha landing period and those strong south-easterly winds last as long as the end of the observation may be due to the interference from other weather systems that is beyond this investigation.
 Finally, due to the CMR blocking, some typhoons will tend to weaken or even dissipate over Taiwan after landing. It is interesting to examine the influence of structure modifications with these two typhoons. The winds split into two different parts above and below the altitude of 3.5 km for both typhoons. This makes a clear modification to the wind structure before typhoon landfall. For the case of typhoon Kaemi, significant wind shear due to blocking disappears right after it landed. The storm winds become systematically organized again and continue to proceed toward northwestward. Also from the height-time distribution of the signal-to-noise ratio, convective rainfall dominates during the landing period of the typhoon. This suggests the storm is still with strong convection and does not disrupt much with the topography. Precipitation is also indicated a stratiform rain type dominates even within the eyewall of the storm. This may imply the center of storm does not support with strong convection. It seems indicate that the typhoon is weakening as it approaches Taiwan owing to the interaction between typhoon and orography. The typhoon Bopha dissipates just 6 hours later over the sea after its center moves away from Taiwan.
 For the first time, typhoon-orography interaction is studied by a wind profiler radar located at the lee side of the Central Mountain Range of Taiwan with two typhoon passages. Profiler observations suggest that the interaction of the typhoons circulation with the CMR produces significant variations in wind structures and precipitation. Although the altitude coverage of the profiler is from 0.235 km to about 6.5 km only, we find the mountain blocking is up to about 3.5 km altitude.
 This wind profiler is operated by the Environmental Protection Association (EPA) of Taiwan and through grant EPA97-FA11-03-A023. CJP is supported by the NSC of Taiwan through grant NSC 95-2111-M-008-012-MY3.