2 Public Weather Service Center, China Meteorological Administration, Beijing 100081;
3 Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029
Oceans occupy two-thirds of the world’s surface, and the influence of air-sea interaction on weather process and climate change is much more significant thanthat of l and -atmosphere interaction(Ola et al., 2005).The disastrous weather and climate phenomena in marineareas are more severe than those in l and areas.However,due to the limitation of observational conditions,it is difficult to make measurements of themarine-atmospheric boundary layer. Thus,to date,observations and studies of the marine-atmosphericboundary layer remain inadequate,especially for situationsinvolving strong winds. Most existing work concerningthe marine-atmospheric boundary layer hasinvolved applying the results obtained in laboratory-based fluid dynamics experiments and atmosphericboundary layer station observations. The laws governingthe characteristics of the marine-atmosphericboundary layer under windy conditions have not yetbeen clearly revealed. Actually,strong wind occursmuch more frequently above the oceanic surface thanabove the l and surface. Driven by the strong wind,theoceanic surface current and breaking waves togetherwith the atmospheric motion form a coupled marineatmosphericboundary layer,constituting a complicatedair-sea system.
In the past two decades,due to increased interestin tropical storms,hurricanes, and their formation and evolution,as well as the significant impactof strong winds on marine engineering and operation,atmospheric and oceanographic scientists,along withstructural engineers,have made some important observations and analyses of strong marine winds,especiallythose associated with hurricanes and typhoons.Although such observations have largely been madeon a case-by-case basis,they are still extremely valuable.Observations of this kind are mostly performedby dropsonde,aircraft,Doppler radar, and so on(e.g.,Franklin et al., 2002; Powell et al., 2003; Schroeder and Douglas, 2003; Knupp et al., 2005; Aberson et al., 2006; French et al., 2007; Kudryavtsev and Makin, 2007; Sanford et al., 2007; Zhang et al., 2008; Zedler et al., 2009). Some valuable directly observed datahave also been obtained during typhoon l and falls and when tropical cyclones passing over isl and s or marineobservational platforms(Sparks,2001; Xu and Zhan, 2001; Song et al., 2005,2010; Harper,2008; Cao et al., 2009; Li et al., 2010; Chen et al., 2011; Liu et al., 2011;Peng et al., 2012; Xiao et al., 2012). In particular,theworks of Song et al.(2005,2010)were based on strongtyphoons such as Damrey,Nuri,Chunchi,Prapiroon,Hagupit, and so on. They focused on studying the engineeringproblems faced by structures such as buildings,bridges, and wind power installations.
Using the data of Song et al.(2005,2010),inthis paper we focus on the marine-atmospheric boundarylayer characteristics,such as those concerning turbulence and wind gusts, and their parameterization.We will employ our own specially developed method,which has already been successfully applied to theprevious studies of the boundary layer during strongwinds related to s and storms and cold surges(Zeng,2006; Cheng et al., 2007; Zeng et al., 2010; Cheng et al., 2011,2012a,b). We divide the fluctuations(orturbulences)into two parts,one being high-frequencyturbulent fluctuation(frequency higher than 1/60 Hz), and the other being gusty wind disturbance(frequencybetween 1/600 and 1/60 Hz). The two parts have differentstructures and play different roles,such as inthe transport of heavy aerosol particles(Zeng,2006;Zeng et al., 2010; Cheng et al., 2012b). Meanwhile,by comparing with strong cold-surge cases(Cheng et al., 2014),we show that there are many common featuresin the marine-atmospheric boundary layer understrong winds.2. Data and instruments
In China,there are many marine meteorologicalobservational stations in the coastal areas of GuangdongProvince and over the isl and s of the South ChinaSea. Certain stations play host to periods of denser and one-off observational campaigns during the typhoonseason to monitor the l and falling of particulartyphoons of interest. The station on Zhizi Isl and is onesuch station. The data used in this study were mainlyobtained from Zhizi(additional data from some othercoastal meteorological stations were also used). For adetailed description of this station and its instruments,readers can refer to Liu et al.(2011) and Xiao et al.(2012). However,for convenience,a brief overview isprovided as follows.
Zhizi Isl and is located off the coast of Bohe,Maoming City. The shortest offshore distance is 4.6km, and the exposed part of the isl and is about 90 mlong and 40 m wide(Fig. 1). The observation toweron Zhizi is 100 m high at 21◦27'23''N,111◦22'28''E, and the base altitude is 10 m above sea level. Thesurrounding water depth is 6–10 m. There are 6 setsof cup anemometers(NRG-Symphonie type)installedat heights of 10,20,40,60,80, and 100 m on the tower, and 3 sets of wind direction observation instrumentsat heights of 10,60, and 100 m. In the typhoon season from August 2008 to August 2010,there was alsoone ultrasonic anemometer at 60-m height(Gill-Windmaster Pro.; sampling frequency 10 Hz).3. Typhoon Hagupit and its structure and development
Typhoon Hagupit formed on 19 September 2008in the western Pacific Ocean near the Philippines.It moved westward and strengthened into a strongtyphoon, and l and ed at 0645 BT 24 Septembernear Chen Village,Bohe,Maoming City,GuangdongProvince(Liu et al., 2011; Xiao et al., 2012). The trajectoryof Hagupit is shown in Fig. 1. The shortestdistance between the typhoon center and Zhizi stationis 8.5 km. During the typhoon’s l and ing,the 10-minaveraged wind speed reached force 15(48.5 m s-1)atBohe meteorological station near Zhizi, and the 3-s instantmaximum gust reached 63.9 m s-1,as observedat 60-m level of the Zhizi tower. The lowest surface pressure was 956 hPa,which lasted for 8 min(Song et al., 2010; Liu et al., 2011; Xiao et al., 2012).
There have been many detailed analyses of the10-min averaged velocity characteristics and turbulentstatistics of Hagupit(Song et al., 2010; Liu et al., 2011;Xiao et al., 2012),but the goals of these studies wereto examine the typhoon’s effects on engineering structures.Based on these works,we reanalyzed the originaldata with added quality control tests. The 10-minaveraged wind velocities(u,w)that we reanalyzed thistime were the same as those in the previous studies.However,we now employ a different method in turbulencedecomposition to analyze the characteristics offluctuations, and hope to obtain some new results.
Figure 2 shows the 10-min averaged horizontalwind speed u recorded by the cup anemometers at 6levels and the ultrasonic anemometer at 1 level on thetower. Figure 2 captures the whole process of Hagupit,as observed by Zhizi station from 0000 BT 22 to 2400BT 27 September,including the region of the typhoon eye,two regions of eye wall squall, and the outsidefield of the typhoon before(24 September) and after(25 September)its l and ing. Figure 3 shows u and wobserved by the ultrasonic anemometer,but only for24 September 2008.
These two figures indicate that the results of thecup anemometers and the ultrasonic anemometer areconsistent. There are some clear characteristics in theaverage flow(u,w). Firstly,in the 20–110-m layer(above sea surface),u is almost independent of height(note that by adding the data of the highest velocityfrom Bohe station during the typhoon’s l and ing,wecould conclude that u from sea level to 110 m is thesame). Only for a very short time during the period ofstrongest wind(> 35 m s-1)is there an exception—the difference between u at the highest two levels isabout 10 m s-1. Secondly,at the 60-m level(70 mabove the sea surface),there is upward motion(w >0); specifically,w reaches 2–4 m s-1 at the squall wall, and 0.5–1 m s-1 in the typhoon eye region. The variablew increases with u. Only for a very short timeafter the typhoon’s passage is w reduced to nearly 0m s-1. This is an important point, and why w is >0 m s-1 is to be discussed in the last section of thispaper. It should be noted that the observed w is correct,because before and after the typhoon’s passagethe difference in wind direction is about 180◦,but theobserved w does not change sign, and it is unlikelythat the inclinometer has produced any error.
Figure 4 shows the pairs(u,w) and their regression.It can be seen that the accuracy of the regres-sion is high. Note that here u is considered as the windvelocity at 10 m above sea surface,as suggested in theliterature. 4. Characteristics of the turbulent fluctuations and gusty wind disturbances
The methods by Zeng et al.(2010) and Chenget al.(2011,2012a,2014)are employed here to analyzethe characteristics of fluctuations(u',v',w')ofthe ultrasonic anemometer data,where u' = u − u;v' = v,¯v = 0; w' = w − w, and u is along the downwinddirection. The averaging interval is 10 min. Wefurther divide the fluctuations into two types: turbulentfluctuations(ut,vt,wt) and gusty disturbances(ug,vg,wg). The former consists of components withfrequency > 1/60 Hz, and the latter with frequencybetween 1/600 and 1/60 Hz. The variable ui(i = t,g)is along the 10-min averaged wind direction. The dataare analyzed for the whole period,i.e.,22–27 September2008,but with special attention paid to the strongwind period(u≥10ms-1).
Figure 5 shows the gusty disturbances ug and turbulentfluctuations ut for the eye wall squall period(0400–0500 BT 24 September 2008),as denoted bythe two vertical lines in Fig. 3. It can be seen thatthe gusty disturbances and turbulent fluctuations areall quite strong. By using u as the abscissa,the turbulencekinetic energy(Et) and the gusty amplitudeof the horizontal component(Agh)are given in Figs.6 and 7,respectively. Three direction components of turbulence energy(Etu,Etv,Etw) and three componentsof gusty amplitude(Agu,Agv,Agw)are alsogiven in Figs. 6 and 7,where , and the kinetic energy of gusts. It can be seen that duringthe strong wind period,the anisotropy of gusty disturbances(Agu> Agv> Agw)is significant, and theturbulent fluctuations are also anisotropic to some extent,but weaker than that of gusts. This is a commoncharacteristic of windy atmospheric layers,no matterif they are above the l and surface(Zeng et al., 2010; Cheng et al., 2011)or above the oceanic surface(Cheng et al., 2014); however,during strong typhoons,this feature is the strongest.
The equivalent period of gusts(Tg,another usefulindex in practical applications) and its regressionare given in Fig. 8. The frequency of the gusty winddisturbance is between 1/600 and 1/60 Hz. There aredifferent periods of gusts in the atmospheric boundary layer. The equivalent period of gusts refers to themain period of gusts by analyzing their power spectra.Note that the points in the region with u < 10 m s-1are scattered due to the r and omness and spontaneityof gust occurrences during weak wind cases,but thisdoes not lead to a serious problem in the practicalapplication of the regression because gusts are weak and present only a small influence on heat and masstransport in such situations.5. Friction velocity and vertical transport ofmomentum
Figure 9 is the superposition of ug and wg during0400–0500 BT 24 September 2008. It shows verystrong coherence at most moments: ug > 0 is accompaniedby wg < 0, and ug < 0 by wg > 0. Therefore,the downward vertical flux of momentum,contributedby gusty disturbances,is large. The vertical flux of momentum contributed by turbulent fluctuations isalso downward. The absolute values of both partsof the momentum fluxes are equal to u2g* and u2t*,respectively. Figures 10a and 10b show respectivelythe gusty friction velocity(ug*) and turbulent frictionvelocity(ut*). Figures 10c and 10d also show u*(thefriction velocity in the conventional sense) and um*,where (Cheng et al., 2007),um* iscalled the average flow friction velocity, and the relatedmomentum flux is u · w,i.e.,.In the case of Hagupit,fluctuations cause downwardfluxes of momentum,but the average flow causes upwardflux. The upward momentum flux by the averageflow is at least one order of magnitude larger than thedownward momentum flux by fluctuations,althoughthe latter is very much larger than that under weakwind conditions. It seems that the 10-min averagemotion is related to the transport of momentum from ocean to the atmosphere during the typhoon period.Note that in strong wind situations,ug*,ut*,u*, and even um* can be parameterized by u,as shown by thehigh accuracy of the regressions in Fig. 10.
The sensible heat flux and latent heat flux are notgiven in this paper because the heat and water vaporexchanges between the atmosphere and the sea spray and spume generated by breaking waves and entrainedinto the atmospheric boundary layer are special subjectsof research, and will thus be discussed in futurework.
6. Summary and discussion
Analyses of the marine-atmospheric boundarylayer during strong Typhoon Hagupit in September2008 and the windy atmospheric boundary layerrelated to a cold surge in March 2012 over the South China Sea(Cheng et al., 2014)indicate that in themarine-atmospheric boundary layer during strongwinds,the characteristics of both the average flow(u,w) and fluctuations(turbulences and gusts)possessmany similarities but also some differences withthose during the strong wind cases over the l and surface(Zeng et al., 2010; Cheng et al., 2011). The majorfindings of the study can be summarized as follows.
(1)In the low-latitude marine-atmosphericboundary layer,during strong wind periods associatedwith tropical storms,hurricanes,or cold surges,the horizontal average velocity u is almost independentof height below about 100 m, and the verticalvelocity w is greater than 0 m s-1. This is differentfrom l and surface cases related to cold air mass outbreaks,where u increases rapidly with height, and wmay be negative(descending cold air).
(2)During the strong wind periods,in the lower part of the atmospheric layer,whether over the oceanor l and ,gusty wind disturbances are anisotropic and coherent, and turbulent fluctuations are alsoanisotropic to some extent(with vertical kinetic energysignificantly less than the horizontal one)butwith weak coherency.
(3)During strong wind periods,the energies ofgusts(Egi) and turbulences(Eti)(i = u,v,w),as wellas the corresponding friction velocities(ug* and ut*),are all much larger than those in weak-wind situations.The vertical fluxes of momentum contributedby gusts and turbulences are all downward. They canbe parameterized by using u as the controlling factorin marine boundary layer cases because of the independenceof u to height. However,the top height ofour observations is not sufficient for obtaining theirvertical profiles.
(4)According to our analysis of observational data,strong wind is accompanied by upward verticalvelocity in the marine-atmospheric boundary layer, and w is≥0.25 m s-1 when u is≥10 m s-1, and wis≥1.0 m s-1 when u≥30 m s-1. This means thatthe upward transport of horizontal momentum,u · w,is≥2.5 and ≥30.0 m2 s−2,respectively. These valuesare at least one order of magnitude larger than thedownward fluxes due to the fluctuations(u*2= 0.04 and 0.6 m2 s−2,respectively). This fact is important and should be further studied because strong wind occursfrequently above the marine surface. Generally speaking,we can imagine that strong wind(average atmosphericflow) and the superimposed fluctuations driveoceanic surface currents and generate large breakingwaves. Thus,a marine-atmosphere coupled boundarylayer is formed, and on the atmospheric side,theboundary layer is different from that over the l and , and some portion of the sea momentum can be fedback to the atmosphere from the ocean. This may bethe reason why u is independent of height. Furthermore,the sea spray and spume droplets make specialexchanges of heat and mass transport between theocean and atmosphere.
From the reanalysis data of the ECMWF,NCEP, and GFDL(Geophysical Fluid Dynamics Laboratory),we can see that in the middle and high latitudes w ≥0(i.e.,ω ≡ dP/dt ≤ 0)at 10 m over a broad areaof the oceanic surface when there is a mature cyclonepassing over, and w ≤ 0(i.e.,ω > 0)occurs only inthe cold frontal region near the center of the cyclone.If this can be confirmed by observations,it might betrue that w≥0 at 10 m occurs very often and coversmost of the oceanic surface areas during strong winds.
Another concern is about the 10-m wind. Accordingto the Beaufort scale,when the 10-m wind isforce 8(17.2–20.7 m s-1),force 10(24.5–28.4 m s-1), and force 12(hurricane; 32.7–36.9 m s-1),the significantwave heights are 5.5(usual)to 7.5 m(highest),9.0(usual)to 12.5 m(highest), and 14.0(usual)to> 16.0 m(highest),respectively. Therefore,the socalled10-m wind above the oceanic surface is actuallymeaningless. Fortunately,u is independent of heightin the strong wind situation.
Acknowledgments.We are very grateful to the Guangdong Meteorological Bureau for providingsome data.
|Aberson, S. D., M. L. Black, R. A. Black, et al., 2006: Thirty years of tropical cyclone research with the NOAA P-3 aircraft. Bull. Amer. Meteor. Soc., 87, 1039-1055.|
|Cao, S. Y., Y. Tamura, N. Kikuchi, et al., 2009: Wind characteristics of a strong typhoon. J. Wind Eng. Indust. Aerodyn., 97, 11-21.|
|Chen, W. C., L. L. Song, S. Q. Zhi, et al., 2011: Analysis on gust factor of tropical cyclone strong wind over different underlying surfaces. Sci. China (Tech. Sci.), 54, 2576-2586.|
|Cheng, X. L., Q. C. Zeng, and F. Hu, 2011: Charac-teristics of gusty wind disturbances and turbulent fluctuations in windy atmospheric boundary layer behind cold fronts. J. Geophys. Res. Atmos., 116, D06101, doi: 10.1029/2010JD015081.|
|—–, —–, and F. Hu, 2012a: Parameterizations of some important characteristics of turbulences and gusts in the atmospheric boundary layer. J. Geophys. Res. Atmos., 117, D08113.|
|—–, —–, and —–, 2012b: Stochastic modeling of the effect of wind gust on dust entrainment during sand storms. Chinese Sci. Bull., 57, 3595-3602.|
|—–, J. Huang, L. Wu, et al., 2014: Structures and char-acteristics of windy atmospheric boundary layer in the South China Sea region during cold surge. Adv. Atmos. Sci. (in press)|
|Cheng Xueling, Zeng Qingcun, Hu Fei, et al., 2007: Gustiness and coherent structure of strong winds in the atmospheric boundary layer. Climatic Environ. Res., 12, 227-243. (in Chinese)|
|Franklin, J. L., M. L. Black, and K. Valde, 2002: GPS dropwindsonde wind profiles in hurricanes and their operational implications. Wea. Forecasting, 18, 32-44.|
|French, J. R., W. M. Drennan, J. A. Zhang, et al., 2007: Turbulent fluxes in the hurricane boundary layer. Part I: Momentum flux. J. Atmos. Sci., 64, 1089-1102.|
|Haper, B. R., 2008: Wind speed time averaging conver-sions for tropical cyclone conditions. 28th Confer-ence on Hurricanes and Tropical Meteorology, Amer. Meteor. Soc., Orlando, FL, No. 4B1.|
|Knupp, K. R., J. Walters, and M. Biggerstaff, 2005: Doppler profiler and Radar observations of bound-ary layer variability during the landfall of tropical storm Gabrielle. J. Atmos. Sci., 63, 234-251.|
|Kudryavtsev, V. N., and V. K. Makin, 2007: Aerody-namic roughness of the sea surface at high winds. Boundary-Layer Meteor., 125, 289-303.|
|Li Qiusheng, Dai Yimin, Li Zhengnong, et al., 2010: Sur-face layer wind field characteristics during a severe typhoon "hagupit' land falling. J. Building Struc-tures, 31, 54-61. (in Chinese)|
|Liu Donghai, Song Lili, Li Guoping, et al., 2011: Anal-ysis of the extreme loads on offshore wind turbines by strong Typhoon Hagupit. J. Trop. Meteor., 27, 317-326. (in Chinese)|
|Ola, P., G. Persson, J. E. Hare, et al., 2005: Air-sea interaction processes in warm and cold sectors of extratropical cyclonic storms observed during FAS-TEX. Quart. J. Roy. Meteor. Soc., 131, 877-912.|
|Peng Zhen, Song Lili, Hu Fei, et al., 2012: Multi-scale analysis on momentum flux of Typhoon Chanchu during its landfall. J. Trop. Meteor., 28, 61-67. (in Chinese)|
|Powell, M. D., P. J. Vickery, and T. A. Reinhold, 2003: Reduced drag coefficient for high wind speeds in tropical cyclones. Nature, 422, 279-283.|
|Sanford, T. B., J. F. Price, J. B. Girton, et al., 2007: Highly resolved observations and simulations of the ocean response to a hurricane. Geophys. Res. Lett., 34, L13604, doi: 10.1029/2007GL029679.|
|Schroeder, J. L., and S. Douglas, 2003: A hurricane bonnie wind flow characteristics as determined from WEMITE. J. Wind Eng. Indust. Aerodyn., 91, 767-789.|
|Song Lili, Mao Huiqin, Huang Haohui, et al., 2005: Anal-ysis on boundary layer turbulent features of land-falling typhoon. Acta Meteor. Sinica, 63, 915-921. (in Chinese)|
|—–, Pang Jiabin, Jiang Chenglin, et al., 2010: Field measurement and analysis of turbulence coherence for Typhoon Nuri at Macao Friendship Bridge. Sci. China (Tech. Sci.), 53, 2647-2657.|
|Sparks, P. R., 2001: Wind speeds in tropical cyclones and surface-to-gradient wind-speed ratios in tropical cy-clones. J. Wind Eng. Indust. Aerodyn., 89(11-12), 1047-1058.|
|Xiao Yiqing, Li Lixiao, Song Lili, et al., 2012: Study on wind characteristics of Typhoon Hagupit based on offshore sea surface measurements. Acta Aerodyn. Sinica, 30, 380-387. (in Chinese)|
|Xu, Y. L., and S. Zhan, 2001: Field measurements of Diwang Tower during Typhoon York. J. Wind Eng. Indust. Aerodyn., 89, 73-93.|
|Zedler, S. E., P. P. Niiler, D. Stammer, et al., 2009: Ocean's response to Hurricane Frances and its im-plications for drag coefficient parameterization at high wind speeds. J. Geophys. Res. Oceans, 114, C04016, doi: 10.1029/2008JC005205.|
|Zeng Qingcun, 2006: Gigantic Yellow Cloud—–he Dust Storm in Eastern Asia. Science Press, Beijing, 228 pp. (in Chinese)|
|—–, Cheng Xueling, Hu Fei, et al., 2010: Gustiness and coherent structure of strong winds and their role in the dust emission and entrainment. Adv. Atmos. Sci., 27, 1-13.|
|Zhang, J. A., P. G. Black, J. R. French, et al., 2008: First direct measurements of enthalpy flux in the hurricane boundary layer: The CBLAST re-sults. Geophys. Res. Lett., 35, L14813. doi: 10.1029/2008GL034374.|