Vol.16 No.2 JOURNAL OF TROPICAL METEOROLOGY June 2010

Article ID: 1006-8775(2010) 02-0125-09 A DYNAMICAL INTERPRETATION OF THE WIND FIELD IN TROPICAL CYCLONES WITH THE CONSIDERATION OF OROGRAPHIC FACTORS

1 1 2 3 HAO Shi-feng (郝世峰) , PAN Jin-song (潘劲松) , YUE Cai-jun (岳彩军) , CUI Xiao-peng (崔晓鹏) , YANG 1 Shi-fang (杨诗芳)

(1. Meteorology Observatory, Hangzhou 310017 China; 2. Typhoon Institute of China Meteorological Administration, Shanghai 200030 China; 3. LACS, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029 China)

Abstract: Based on the barotropic primitive equation in the polar coordinate system and the appropriate assumption, we obtained the mathematical equation of orographic forcing on unit mass air parcel. With the consideration of the frictional stress of the sea and land, supposing that parcel velocity in tropical cyclones is in linear variation and that the distribution of surface pressure is circular, a set of equations are derived, which describe the impact of orographic slope error, the central pressure error and position error of tropical cyclones on the wind field in the tropical cyclone. Typhoon Wipha (2007) is selected to verify the above interpretation method. The results show that the orographic slope, the frictional coefficient, the intensity and position of the cyclone are the important factors which have great influence on the interpretation of wind information about tropical cyclones. The dynamic interpretation method gives very good results, especially for the coastal area. It is applicable to improving the forecasts of the wind field in tropical cyclones.

Key words: tropical cyclone; orographic slope; frictional stress; strong wind; dynamic interpretation CLC number: P444 Document code: A doi: 10.3969/j.issn.1006-8775.2010.02.004

1 INTRODUCTION cyclones. In order to improve numerical forecasts, a lot of efforts have been made to refine model capability, Tropical cyclones are intense synoptic systems enhance initial fields, etc. However, forecast errors are which occur over the tropical ocean. Since the inevitable since the permanent existence of data errors conventional synoptic observations over the ocean are and model errors. Then how to interpret products of rare and unreliable, it is difficult to know the true numerical forecasts and lessen the influence of the structure of tropical cyclones and associated above mentioned errors becomes an important task. characteristics. Although there exist plenty of satellite Presently, methods for interpretation of products of observations, it is still difficult to get very good results numerical forecasts mainly include statistical due to some reasons, such as difficulties in assimilation interpretation (such as MOS, PP and Kalman filter), technology. They are the most important reasons that synoptic interpretation, artificial intelligence lead to the difficulties of forecasts of tracks, intensities, interpretation, etc. While no matter which method to wind and rainfall associated with tropical cyclones. select, the key factors influencing the errors must be The disasters caused by tropical cyclones are taken into account. According to the particularity of mainly induced by strong wind and heavy rainfall. typhoons’ structure, some reasonable typhoon models Torrential rainfall, especially typhoon-induced are set up by incorporating the main factors which torrential rainfall, has been studied extensively by [1-5] influence the wind and pressure field of tropical many scholars at home and abroad , but few studies cyclones, such as the Fujita formula, Myers formula have been carried out on the strong wind of tropical and a tangential wind velocity profile scheme by

Received date: 2009-12-11; revised date: 2010-03-02 Foundation item: National Basic Research Program of China (973 Program) (2009CB421505); major projects for science and technology development of Zhejiang province (2007C13G1610002); major promoting projects for new technology of China Meteorologycal Administration (09A13) Biography: HAO Shi-feng, senior engineer, M.S., mainly undertaking the research on numerical predication. E-mail for corresponding author: [email protected]

PDF created with pdfFactory trial version www.pdffactory.com 126 Journal of Tropical Meteorology Vol.16

Williams in 1987, and applied to practical operations. 2 [6] [7] dvr vθ ∂φ Zou and Xiao and Pu and Braun constructed −fvFθ−=−+r Bogus schemes by using typhoon models and dtrr∂ assimilated them into initial fields, which improved dvvv 1 ∂φ θθ+r+fvF=−+, (1) forecasts. These simplified typhoon models could  r θ  dtrr∂θ describe the main characteristics of typhoons since they ∂φ → have the key factors. While the wind fields of these +∇•(φ −=ghv)0 typhoon models are mainly symmetric, which are ∂t different from those of actual typhoons with apparent in which Fr and Fθ are the frictional stress, h is asymmetric features, Lei and Chen [8] showed that the the height of terrain, and all other notations here are distributions of the physical fields associated with conventional in the meteorological context. The typhoons are often asymmetric due to the influences of subscripts “θ ” and “ r ” indicate the tangential and the moving speed, frictional stress and environmental radial directions, respectively. Supposing that the synoptic systems. Many researches have focused on the geopotential equation of Eq.(1) is for stationary motion, asymmetric structures of typhoons at home and abroad. [9] ∂φ Zhang and Sui put forward a model of wind fields of = 0 , and the distribution of geopotential height in typhoons by considering frictions in stable state in ∂t 1989, and discussed it under ideal conditions. Hu et al. tropical cyclones is circular, namely the zero order [10] and Huang et al. [11] further studied the model, ∂φ approximation[13, 14], = 0 , the geopotential verified it by using wind fields of real typhoon cases r∂θ and pointed out that this model was applicable in equation can be expressed as: describing the asymmetric distribution of wind fields of ∂φ →→ tropical cyclones. vr =v•∇gh−()φ −∇hgv. (2) In the previous work, researches are mainly ∂r restricted to the wind field of tropical cyclones on the By introducing the equation of continuity, Eq.(2) sea level. Hao et al[12] studied the dynamical can be further written as interpretation of wind field of tropical cyclones over ∂φ → v=v•∇+ghgω (3) the sea. Due to not considering the orographical factors, rs∂r their researches show that the interpretation effect is where ω is the vertical velocity over the surface, not good on land, and the wind field on islands was not s which is the sum of vertical velocity caused by improved after the tropical cyclone landfalling. In this [15] paper, a dynamical interpretation method of wind fields orographic uplift and frictional effect , C → of tropical cyclones, in which orographical factors ωξ=d +vh•∇ , where ξ is the vorticity over were considered, is proposed by combining the method s f proposed by Hao with operational forecasts and by the surface, and C is the drag coefficient. By analyzing the key factors influencing wind fields in d tropical cyclones. substituting ωs into Eq.(3) we obtain ∂φ gCg∂∂hh =d ξ ++2()vv v ≠ 0 . (4) 2 DYNAMICAL INTERPRETATION ∂rvfvθ rr∂∂θ r r METHOD OF WIND FIELDS IN TROPICAL rr CYCLONES In Eq.(4), the second term expresses the radial orographic forcing on unit mass air parcel in tropical 2.1 Correction equation of numerical forecasts of cyclones on the condition that the atmosphere is wind fields in tropical cyclones barotropic, which is associated with the orographic slope, wind speed, and the angle between the wind By setting the center of tropical cyclones as the direction and orographic slope. Set the second term of origin of polar coordinates, the barotropical primitive Eq.(4) as equations of any air parcel in tropical cyclones could g∂∂hh be expressed as: frr=+2()vvθ . (5) vrrr∂∂θ Because the direction of forcing caused by the orographic slope and the orographic gradient are collinear, the tangential forcing caused by the orographic slope is

PDF crea126t ed with pdfFactory trial version www.pdffactory.com No.2 HAO Shi-feng (郝世峰), PAN Jin-song (潘劲松) et al. 127 g∂h∂∂hh suitable to study the wind field over land surface f=+2[vv()21()]− . (6) θθvr∂θθ∂∂rrr directly, they were introduced into the primitive r equation as modified terms as follows: As Eqs. (5 & 6) were obtained on the condition that the atmosphere is barotropic, which are not 2 dvrvθ12∂pg∂∂hh −fvθθ−=−−δ ()v++vFrr dtrρθ∂rvrr∂∂  r (7) dvvv 12∂pg∂h∂∂hh θθ+r+fv=−−δ[v()21()]−++vF  rrθθ dtrr∂θvrr∂θθ∂∂rr 1 ∂p tropical cyclones is circular, −=0 , considering where δ is the modified coefficient. r ∂θ Supposing that parcel velocity takes on linear the influence of the speed of movement of tropical variation in tropical cyclones, that is, [16] cyclones, vs , on the cyclonic curvature radius r , dvr dvθ = Cr , =−Cθ are constants, further and setting the counter-clockwise direction to be dt dt positive, the frictional stress to be proportional to the supposing that the distribution of sea level pressure in square of wind speed and k to be frictional coefficient. Eq.(7) could be further written as: 2 vθθvvscosα 12∂pg∂∂hh 22  ++fvθ=+δ ()vθθ+vr+kvrv++vCrr rrρθ∂rvrr∂∂  r (8) vvvv cosα2g∂h∂∂hh rθ+rs +fv=−δ[v()2()]−1+v−kvv22++vC  rθrrθθθ rrvrr∂θθ∂∂rr where α is the angle between the moving direction of ∂h Ng= 2δ , and Q=+vv22, Eq.(8) can be tropical cyclones and the tangential direction of parcels. ∂r θ r 1 ∂p ∂h By setting π = , Mg= 2δ , expressed alternatively as: ρ ∂r r∂θ 2 vθθvvscosα 1 ++fvθθ=π +()vM+vrN++kvrrQC rrv  r . (9) vvvv cosα 1 θr+rs +fv=−()vM21N− +vM−+kvQC  rrθθθ rrvr Since the orographic slope is constant, obtain the influence of the error of the orographic slope ∂NM on wind speed as follows: MN22+=const , we can obtain =− . ∂MN By differentiating Eq. (9) with respect to M , we ∂v−B[2MN2v−vM3−vN3]+−C()vN32MNv θθθθrr, (10) = 3 ∂+M()ACBDvNr 23332233 ∂vrBD[2MNvθ−vθM−vrN]−DC(vθθN−MNvrr)v(2)MN++MvN =−33. (11) ∂+MC()ACBDvrrNCvN By differentiating Eq. (9) with respect to π , we By differentiating Eq. (9) with respect to r, we obtain the influence of the change in pressure gradients obtain the influence of the change in the distance on wind speed as follows: between the observation points and the tropical cyclone ∂v C center on the wind speed as follows: θ = , (12) ∂+vvvBvC2 ∂+π vr ()ACBD θθθr = 2 , (14) ∂vD ∂+r()ACBDr r =− . (13) ∂+π vr ()ACBD

PDF created with pdfFactory trial version www.pdffactory.com 127 128 Journal of Tropical Meteorology Vol.16

2 2.2 Estimation of the correction term associated with ∂v vθvrr(AC+BD)−+D()vvθθBvC r= , (15) pressure gradients ∂+rC()ACBDr2 in which Since the correction term associated with sea level 2v+ vcosα M kvv pressure gradients could not be obtained directly, Af=θθsr+−− , estimation is needed. rvQr Supposing that the sea level pressure field of 2 tropical cyclones is determined by the Fujita formula. kvr vMθ B=kQ +−2 , 1 Qvr 2 2 p=p∞∞−(p−+p0 )/[12(rr/m )], 2 vθ+vsrcosα kvθθvvM where p∞ is the pressure in the periphery of tropical Cf=++−2 , rQvNr cyclones, p0 is the pressure in the center of tropical 2 2 kvv1 M cyclones, and by introducing the equation of state, we D=kQ +θ ++r . obtain QrvNr 3 2()p− pRTr − It can be known that all the terms on the right hand π=+∞ 0 [12(rr/)]2 2 sides of Eqs. (10-15) could be obtained from the 2 m rpm forecasts of numerical models. Thus the correction where R is a gas constant, T is the sea surface equation of numerical forecasts of wind fields in temperature, and r is the function of p , which tropical cyclones could be expressed as: m 0 ∂v∂∂vv must be estimated. According to Eq. (9), and by θθθ dvθ =dM++drdπ , (16) applying vv= cos β , vv= sin β , the maximum ∂Mr∂∂π θ r ∂v∂∂vv wind speed radius r could be expressed as: dv=dMr++drdrrπ , (17) m r ∂Mr∂∂π π−−v22cosβvvcosβαcos r =mmms (18) where dM , dω and dr represent the correction m 2 fvmcosβ−kvmrsin ββ−Mctg−−NC terms associated with orographic slope errors, pressure −3/2 gradient errors, and distance errors between the 2⋅−3RT00()pp∞ where π m =−1/2 , vm is the cyclone center and the observation point (determined p∞∞−−3()pp0 by the position of tropical cyclones), respectively. Thus maximum wind speed, T is the sea surface by incorporating the errors of the three terms in 0 operational applications, it is easy to correct the temperature in the center of tropical cyclones. By numerical forecasts of wind fields in tropical cyclones eliminating r in Eq. (9), we get the simple quadric dynamically. In the above formulas, dr and equation of vm . Due to the uncertainty of Cr and could be obtained by extrapolating the difference dM Cθ , the equation could not be solved directly. If vm between model forecasts and observations of tropical is replaced by the maximum wind speed of a stationary cyclones, or by comparing the predictions of ensemble tropical cyclone[9 - 11], namely, C and C are both members. As dπ could not be obtained directly, it r θ needs to be estimated. zero, the equation of vm reduces to cosα v4+v3v+Cv22+Cv−(ctgM/NM+=)0π (19) mmscos 21mmm in which vcosα(Mctgβ+N)−πf++(ctgβαM2 /NMv)cos C=tgβ sms 1 k Mcosβ+sinβN−πk++(ctgββM2/NM)cos C= m 2 k gradients could be where β is a complex function of P0 , vm , vs , α . [12] .(20) According to the research of Zhang et al. , in the dπ=ππ(p0+dp0,T0+dT0,r+−dr)(p00,Tr,) maximum wind speed region of tropical cyclones, β ＜ 5°, namely 0.996< cos β <1, so cos β and 3 REAL CASE VERIFICATION OF THE DYNAMICAL INTERPRETATION tgβ could be considered as constants. From Eqs. METHOD OF WIND FIELDS IN TROPICAL (17–19), the correction term associated with pressure CYCLONES

PDF crea128t ed with pdfFactory trial version www.pdffactory.com No.2 HAO Shi-feng (郝世峰), PAN Jin-song (潘劲松) et al. 129 3.1 Model, data, and methodologies northern Zhejiang is in an area of strong wind while the wind speed in the southern coastal area is relatively The numerical model used in this paper is WRF low, which is opposite from the observation (Fig. 1b & v2.2 with the horizontal resolution of 45 km and 41 1c). From Fig. 1d that gives the result of dynamic layers in vertical direction. The top model level is 50 interpretation with Scheme 2 that considers the hPa. The Betts-Miller-Janjic cumulus parameterization orographic slope and frictional effect, the wind field is scheme and the YSU boundary layer scheme are used shown to be improved obviously, especially near the in the integration. The data are from 1°×1° GFS data of landing site, where there is a strong wind area of 27 m the NCEP reanalysis. The tropical cyclone Wipha, s-1, which is close to the observation. Around Wenling coded 0713 in China, is selected. The model is in Taizhou City, although the wind intensity is weaker integrated from 0800 LST (same below) 17 September than that of the control experiment, the structure is very 2007 and the integration lasts for 72 h. Scheme 1 similar. Remarkably, there are four strong wind areas (control experiment) is a direct forecast by the model, in the coastal areas of northern Zhejiang and Shanghai, and Scheme 2 is the experiment of a dynamical which is relatively stronger against the observation, interpretation method, where δ = 0.1 , the frictional -5 especially in the coastal area near Ningbo, but the coefficient over the sea (k1) is 4.2×10 , and the -2 distribution is similar to the observation. From Fig. 1(b frictional coefficient on land (k2) is 6.9×10 . In & c), we could see that the forecasted wind speed is Scheme 3, there is no orographic slope effect on the clearly strong in the inland area, but the structure of the wind field, namely δ = 0 . In Scheme 4, there is no wind field is improved (see Fig. 1d): some small strong frictional effect on the wind field, namely k1 = k2 = 0. wind centers were interpreted, but the intensity is much In Scheme 5, there are no orographic or frictional stronger. For example, there is a wind area of 18 m s-1

effects. dP0 , dr and dM are obtained from the near Tiantai, while the measured velocity was only 9 m -1 observations and the forecasts of the Scheme 1 (control s . Generally speaking, compared with the controlled experiment), and the impact of the change in sea experiment, the results of the dynamic interpretation of surface temperature is not considered. The inner Scheme 2 is better than those of the control experiment, deflection angle of wind direction at the maximum especially in the coastal area. Fig. 1e shows the result wind speed radius, β , is set to be 1°. of Scheme 3, in which the modified coefficient is zero (δ = 0 ), and there are two strong falsely predicted 3.2 Results centers of wind in province, while the wind field was improved noticeably around the landing point, but Occurring over the ocean to the northeast of the not in the inland or the northern coastal area of at 0800 September 16 and at 0200 Zhejiang. Figure 1f shows the interpreted result of September 17, Wipha strengthened into a typhoon with Scheme 4, in which there is no frictional effect (k1 = a path towards the northwest. At 0230 September 19, it k2= 0), the wind speed in the coastal area of Zhejiang made its landfall at Xiaguan town in Cangnan county was strengthened remarkably, strong winds at the of Zhejiang province with a minimum central pressure speed of 27 m s-1 distribute widely in the -1 of 950 hPa and a maximum wind speed of 45 m s . It neighbourhood of the coastal area of Zhoushan and moved towards the northwest by west at a speed of Ningbo, and strong wind areas at the speed above 21 m -1 about 20 km h . Due to the influence of the typhoon, s-1 also appear in the inland area. Compared with the -1 the wind of Yuliao in Cangnan reached 55.3 m s observation, the error is very large. This result clearly (Level 16) at 0200 LST, and that of Xiaguan reached shows that the orographic slope is an important -1 39.1 m s (Level 13). The wind generally reached influencing factor for the wind field, which has Level 10–12 over the whole coastal sea surface. stronger effect than the frictional stress. Fig. 1g is the The control experiment reproduced the track of the interpreted result of Scheme 5, in which the modified typhoon very well, which was basically consistent with coefficient and friction coefficient are both zero, and the observed track to the east of 122°E. Deviations are the wind field of the typhoon is a rather symmetric a little bigger after that and the landfalling position is a circular structure, which means that the frictional stress little northward while the moving direction is basically and orographic slope are the important factors that the same as in the observations after the landfall, with determined the asymmetric distribution of the wind the track shifted a little eastward (Fig. 1a). Because of field of the typhoon. the high moving speed of typhoon in the control experiment (Scheme 1), the error of landfall time is much bigger. The landfalling position is northwards of that of observation and the distribution of the wind field is described as follows. The coastal area in

PDF created with pdfFactory trial version www.pdffactory.com 129 130 Journal of Tropical Meteorology Vol.16

(a)

(d)

(b) (e)

(c)

(f)

PDF crea130t ed with pdfFactory trial version www.pdffactory.com No.2 HAO Shi-feng (郝世峰), PAN Jin-song (潘劲松) et al. 131 to –9 m s-1 and the wind field near the landfalling point of the cyclone is also improved with an increment of up to 6 m s-1. In conclusion, the orographic slope is a very important interpretation factor for the wind field. Figure 2c & 2d show that the correction term associated with the position of the cyclone and the gradient of pressure exert little influence on the wind field of the inland area, but have great influence on the coastal area, especially near the center of the tropical cyclone, up to 15 m s-1 and 12 m s-1, respectively. In order to further verify the interpretation effect of the orographic factor in the whole forecast period of Wipha, a station at Nanji in southern Zhejiang and another one at in the inland area are chosen to compare the results of Scheme 1 (the controlled (g) experiment), Scheme 2 and Scheme 3. Figure 3 shows Fig.1 Observed and forecasted tracks of the typhoon in a) the time series of 72-h forecasts and observations of with ■ for the observed, ○ for the forecasted, with wind speed in the two stations. Nanji is an island time at 12-h intervals; observed wind speed in b), station located to the east of Xiaguan town (the and simulated wind speed of the control landfalling point). In Fig. 3a, the controlled experiment experiment in c). For Scheme 2(d), δ = 0.1, k1 = and observation are almost out-of-phase from 1400 4.2×10-5, k = 6.9×10-2; for Scheme 3(e), , k 2 δ = 0 1 September 17 till 0200 September 19, especially at = 4.2×10-5, k = 6.9×10-2, for Scheme 4(f), 2 1400 September 18 when the error is the biggest. δ = 0.1, k1 = k2 = 0 and for Scheme 5(g), δ = 0 , During this period, the error of Scheme 2 is most k1 = k2 = 0 at 0200 September 19, 2007. unstable and the effect of Scheme 3 is very weak, with Figure 2 gives the orographic slope and the wind the result almost the same as the control experiment speed correction in Scheme 2 from the three correction during the first 24 hours and getting somewhat better terms associated with the orographic slope, pressure from 0800 to 1400 September 18. After 0200 gradients and cyclone position on the right hand sides September 18, the result of Scheme 2 is much better of Eqs. 16 & 17. Figure 2a is the orographic slope in than the controlled experiment and Scheme 3, which is the model, which shows that the orographic slope is noticeably close to the observation, while the result of higher in the coastal area of Zhejiang and Fujian the Scheme 3 is almost the same as that of the control provinces and the contours are dense. It is important to experiment in the 48-h to 72-h forecast. Figure 3b is note that the orographic slope is lower inland, ranging the interpreted result of Wenzhou Station. Although the from 1×10-3 to 6×10-3, due to the low resolution of 45 results of the three schemes are similar, the result of km in the model. In other words, the orographic slope Scheme 2 is the best in the whole forecast period, while -1 in the model is quite different from the real one, which in Scheme 3, the wind speed is 15 m s with the error -1 is why the modified coefficient δ was introduced. It up to 10 m s at 1400 September 18. From the above is easy to note in Fig. 2b that the correction term results of the two stations, Scheme 2 is better than the associated with the orographic slope exerts the greatest controlled experiment and Scheme 3, which means that influence on wind fields in the coastal area of Zhejiang the orographic slope is a major factor affecting the and the boundary between Zhejiang and effect of dynamical interpretation of the wind field in a provinces with a velocity increment of up to 12 m s-1. tropical cyclone. Although the slope of Zhoushan island is very low, the influence is remarkable with a negative increment of up

PDF created with pdfFactory trial version www.pdffactory.com 131 132 Journal of Tropical Meteorology Vol.16

a b

c d

Fig.2 Orographic slope (a), difference fields of wind speed between the control experiment and Scheme 2 by reducing the orographic slope (b), pressure gradient (c) and cyclone position (d).

4 CONCLUSIONS AND DISCUSSION the orographic slope, the frictional coefficient, the intensity and position of the cyclone are the important (1) Based on the barotropic primitive equation in factors, which have great influence on the interpreted the polar coordinate system and the appropriate wind field of the tropical cyclone, Among them, the assumption, we obtained the mathematical equation of orographic slope is a very important factor. The scheme the orographic forcing on unit mass air parcel. With the considering orographic factors is noticeably better than equation and the consideration of the frictional effect of those excluding them and satisfies the correction of the the land and sea surface and supposing that parcel wind field. The improvement of wind field is much velocity in tropical cyclones is in linear variation and better in the coastal area than in the inland area, being that the distribution of surface pressure is circular, a set related to the exclusion of regional difference of land of equations are derived, which describe the impact of friction coefficients and the regional difference of errors of the orographic slope, central pressure and modified coefficient of the orographic slope. This is a position on the wind field of the tropical cyclone. defect of the current method and could be improved and (2) By using the above dynamical interpretation perfected in the future work. method of numerical forecasts of wind fields in tropical In routine weather forecast operations, based on cyclones, a real case experimental study with Typhoon the analysis of numerical forecasts in the initial period, Wipha (0713) was performed. The results show that or based on the comparison among predications of

PDF crea132t ed with pdfFactory trial version www.pdffactory.com No.2 HAO Shi-feng (郝世峰), PAN Jin-song (潘劲松) et al. 133 ensemble members, it is easy to correct the numerical [2] CUI Xiao-peng, GAO Shou-ting, WU Guo-xiong. predications of wind fields in tropical cyclones Up-sliding slantwise vorticity development and the complete dynamically only by aiming at the three factors of vorticity equation with mass forcing [J]. Adv. Atmos. Sci., 2003, 20(5): 825-836. central pressure, position, and orographic slope. The [3] CUI Xiao-peng. Quantitative diagnostic analysis of surface method presented in this paper is simple and applicable rainfall processes by surface rainfall equation [J]. Chin. J. in routine operation. Atmos. Sci.(in Chinese), 2009, 33(2): 375-387. [4] ZHANG Hai-xia, CUI Xiao-peng, KANG Feng-qin, et al. Observational analysis for a torrential rain caused by a landfall 30 typhoon [J]. Plateau Meteor. (in Chinese), 2007, 26(5): 980-991. 25 [5] ZHAO Yu, CUI Xiao-peng, WANG Jian-guo. A study on a heavy rainfall event triggered by inverted typhoon trough in 20 Shandong province [J]. Acta Meteor. Sinica (in Chinese), 2008, 66(3): 423-436. 15 [6] ZOU X L, XIAO Q N. Studies on the initialization and simulation of a mature hurricane using a variational bogus 10 data assimilation scheme [J]. J. Atmos. Sci., 2000, 57(7): 836-860. 5 [7] PU Z, BRAUN S A. Evaluation of bogus vortex techniques with four dimensional variational data assimilation [J]. Mon. 0 Wea. Rev., 2001, 129(8): 2023-2039. 08111417202302050811141720230205081114172023020508 [8] LEI Xiao-tu, CHEN Lian-shou. An overview on the interaction between tropical cyclone and mid-latitude weather (a) systems [J]. J. Trop. Meteor. (in Chinese), 2001, 17(4): 452-461. 15 [9] ZHANG Jian-lin, SUI Shi-feng. CHGS method for numerical forecasting of typhoon wave [J]. Trop. Oceanol. (in Chinese), 1989, 8(1): 58-66. [10] HU Bang-hui, TAN Yan-ke, ZHANG Xue-min. 10 Distribution of wind speed of tropical cyclone over the sea [J]. Chin. J. Atmos. Sci. (in Chinese), 1999, 23(3): 316-322. [11] HUANG Xiao-gang, FEI Jian-fang, ZHANG Gen-sheng, et al. Construction of wind speed in typhoon on ocean surface 5 [J]. J. Trop. Meteor. (in Chinese). 2004, 20(2): 130-136. [12] HAO Shi-feng, CUI Xiao-peng, PAN Jin-song, et al. A dynamical interpretation of the wind field tropical cyclones [J]. J. Trop. Meteor., 2009, 15(2): 210-216. 0 [13] LIU Shi-kuo, LIU Shi-da. Inner and outer Structures of 08111417202302050811141720230205081114172023020508 tropical cyclone. I: Barotropic modle [J]. J. Trop. Meteor. (in (b) Chinese) 1994, 10(3): 193-203. Fig.3 Time list of wind speed from 0800 on September 17 [14] LIU Shi-kuo, LIU Shi-da. Inner and outer Structures Of to 0800 September 20, 2007, for Nanji station (a) Tropical Cyclone II Baroclinic Model [J] . J. Trop. Meteor. and Wenzhou station (b). Observation: blue line; (in Chinese) 1994, 10(4): 290-299. control experiment: pink line; Scheme 2: green line; [15] ZHU Qian-gen, LIN Jin-rui, SHOU Shao-wen, et al. Scheme 3: red line. (Unit: ms-1). Principles and Methods of Synoptic Meteorology [M] (in Chinese). Beijing: China Meteorological Press, 1992: 320-324. [16] LV Mei-zhong, Peng Yong-qing. Tutorial of Dynamical REFERENCES: Meteorology [M]. Beijing: China Meteorological Press, 1990: 92. [1] CHEN Lian-shou, LUO Zhe-xian, LI Ying. Research advances on tropical cyclone landfall process [J]. Acta Meteor. Sinica(in Chinese), 2004, 62(5): 541-549.

Citation: HAO Shi-feng, PAN Jin-song, YUE Cai-jun et al. A dynamical interpretation of the wind field in tropical cyclones with the consideration of orographic factors. J. Trop. Meteor., 2010, 16(2): 125-133.

PDF created with pdfFactory trial version www.pdffactory.com 133