᭧2002American Meteorological Society
surfaceflow over cool tropical regions where SSTs are below the convective thresholds(Lindzen and Nigam 1987).In reality,temperature variations in both the free troposphere and PBL contribute to tho in SLP and hence surface wind in the Tropics(Wang and Li1993; Chiang et al.2001).
牛津词典在线
In addition to the SLP driving for surface wind, Wallace et al.(1989)propo that vertical mixing of momentum near the surface is important in the eastern equatorial Pacific on asonal,interannual,and longer timescales.A sharp SST front forms around2ЊN from June to December,parating the cold upwelled water on the equator from the warmer water(ϳ27ЊC)to the north.Wallace et al.hypothesize that when the south-easterly trade winds cross this equatorial front to warmer a surface,vertical mixing intensifies,bring-ing fast-moving air down and thereby accelerating the surfaceflow.They suggest that the enhanced vertical mixing on the warmer side of the equatorial front is responsible for the maximum in meridional wind speed there.1Paluch et al.(1999)describe their air-craft passage across the equatorial front:‘‘The tran-sition to the warmer a surface was associated with a sudden appearance of numerous whitecaps(there were no whitecaps before this time),which suggests an increa in surface wind speed.’’Making turbu-lence measurements on board the aircraft,Paluch et al.obrve the intensity of eddy vertical velocity and h
ence vertical mixing increasing(decreasing)over the warm(cold)side of the SST front.
In order for Wallace et al.’s(1989)mechanism to work,vertical shear with upward-increasing speed is required in the mean wind.Few vertical soundings of the atmosphere exist in the remote eastern Pacific.Drop-sonde measurements were made in the equatorial Pacific during the First Global Atmospheric Rearch Program Global Experiment(FGGE)in spring1979.An earlier version of FGGE data does not give wind readings be-low900mb(Kloel and Albrecht1989),while a recent analysis suggests an upward-decreasing wind shear in the eastern equatorial Pacific(Yin and Albrecht2000), opposite to the shear required for the Wallace et al. mechanism.Analyzing soundings obtained in October and November1989,Bond(1992)reports an upward-increasing shear between the top of the mixed layer and a surface that decreas north of the equatorial front, consistent with the Wallace et al.(1989)hypothesis. This difference in wind shear between the FGGE and Bond soundings may be due to asonal variations in wind.Long-term wind profiler obrvations at the Ga-lapagos Islands(0.9ЊS,89.61ЊW)reveal a southerly jet at400m that intensifies during the colder half of the year(June–November)and weakens during the warmer half(December–May;Hartten and Gage2000).Using soundings from September1998,Anderson(2001)re-1An alternative advective mechanism is propod by Tomas et al. (1999).See also Mahrt(1972).ports that temperature is stabl
y stratified to the south of the equatorial front while a mixed layer500m deep develops to the north.He further reports a rapid increa in surface wind speed as the ship cross the front from the south.
b.Tropical instability waves
出版社英文
While the southward SLP gradient,with either linear (Lindzen and Nigam1987)or nonlinear(Mahrt1972; Tomas et al.1999)PBL dynamics,contributes at least partially to the acceleration of the southerly winds across the equatorial front,the Wallace et al.(1989) vertical mixing mechanism appears dominant in month-ly variability on the equatorial front(Hayes et al.1989; Xie et al.1998).During the colder half of the year (June–December),the equatorial front often displays large cusp-shaped meanders of typical periods of1 month and typical zonal wavelengths of1000km(Le-geckis1977;Chelton et al.2000),in association with oceanic TIWs that grow on shears of rapid equatorial currents(Philander1978;Yu et al.1995).Obrvations show coherent covariability in surface wind with am-plitudes of1–2m sϪ1.The southeasterly trades accel-erate(decelerate)in the warm(cold)pha of the month-ly SST waves,a phasing that is consistent with the ver-tical mixing mechanism but not with the Lindzen and Nigam(1987)SLP mechanism(Hayes et al.1989).The latter would predict a90Њpha difference between wind and SST anomalies near the equator where the
Coriolis effect is small.
Global measurements of vector wind by satellite scat-terometers allow the determination of space–time struc-ture of TIW-induced wind variability,as demonstrated by Xie et al.(1998)with the European Remote Sensing scatterometer.Launched into space in June1999,the SeaWinds scatterometer on the QuikSCAT satellite of-fers a higher space resolution and daily near-global cov-erage that can sample TIWs adequately.Applying var-ious statistical technique to the QuikSCAT data,Liu et al.(2000),Chelton et al.(2001),and Hashizume et al. (2001)show that the trade wind acceleration is more or less in pha with SST variability,in support of the vertical mixing mechanism(e also Wentz et al.2000; Thum et al.2002).
自学考试证书查询While buoy and satellite measurements both show the dominance of the vertical mixing mechanism for TIW-induced wind variability,it is unclear,physically,why the SLP mechanism should not be more important.Sup-po that SST-induced changes in air temperaturefill a 1-km-deep PBL as in Lindzen and Nigam(1987).A simple calculation to be prented in ction6gives zonal wind anomalies ten times larger than obrvations.
A tenfold reduction in the depth over which the SST effect on air temperature extends can bring the
wind anomaly estimate clor to the obrved amplitudes,but then the pha relative to SST still does not agree with obrvations.
c.Vertical structure
This influence depth of SST,though important for SLP adjustment,has never been obrved directly. TIWs induce coherent changes in the amount of stra-tus clouds north of the equatorial front(Der et al. 1993).While the cloud top is not determined in Der et al.(1993),we suggest an influence depth of at least 400m,the typical mixed-layer depth that often is also the cloud ba(Kloel and Albrecht1989).This cloud respon to TIWs is confirmed by a recent anal-ysis of passive microwave measurements from the Tropical Rain Measuring Mission(TRMM)and Spe-cial Sensor Microwave Imager(SSM/I)satellites (Hashizume et al.2001).Furthermore,Hashizume et al.(2001)detect coherent rainfall anomalies north of TIWs in the southern portion of the intertropical con-vergence zone(ITCZ),suggesting that the depth of the atmospheric respon exceeds the PBL depth at least over the warm convective zone.
The vertical structure of the atmosphere does not only hold the key to the puzzle of why the SLP mechanism is unimportant for the TIW phenomenon,but it can also provide direct evidence for the ve
rtical mixing mech-anism.Under the vertical mixing mechanism,we expect to e a couplet of wind acceleration and deceleration in the vertical as en in a general circulation model simulation(Xie et al.1998).In this model,SST-induced air temperature anomalies extend over a depth of1km and the SLP effect contributes equally to zonal wind anomalies as the vertical mixing,further indicating that there is no a priori reason why the SLP mechanism should be small.
Vertical soundings of air temperature,humidity,and wind velocity were obtained along2ЊN from140Њto 110ЊW in September1999,on board the Japane re-arch vesl Shoyo-maru.This paper reports results from this crui that cut across veral SST waves on its way to the east.To our knowledge,this is thefirst time that the vertical structure of TIW-induced atmo-spheric waves has been measured.We u the in situ measurements to investigate the thermal and dynamic respon of the atmosphere to the slow undulation of SST.The SST’s influence depth is a key parameter to dynamic adjustment as discusd above and will be a focus of this study.We show that the TIW effects pen-etrate the whole depth of the PBL,causing large vertical displacement of the temperature inversion that caps the PBL.Our wind velocity measurements indicate a ver-tical shear adjustment that is consistent with the Wallace et al.(1989)hypothesis.The obrved thermal respon is associated with greatly reduced pressure anomalies at the a surface.
The rest of the paper is organized as follows.Section 2describes the crui and measurements on board, which is followed by a brief analysis of satellite mea-surements that sample the Pacific both in space and time (ction3).Section4investigates the thermal respon,and ction5examines the vertical structure of wind velocity variability.Section6considers the vertical pro-file of pressure anomalies and address the question of why the SLP signal is smaller than one might expect. Section7is a summary.
2.Crui data
The equatorial Pacific was in a weak La Nin˜a state in1999,with SST in the east slightly below the normal. TIW-induced SST waves started developing in late May, reached large amplitudes in June and July,and remained strong until the end of the year.On16September1999, the rearch vesl Shoyo-maru of the Japan Fisheries Agency left Honolulu on its way to the equatorial Pa-cific.It arrived at2ЊN,140ЊW on21September and began a week-long crui along2ЊN,reaching2ЊN, 110ЊW on28September.We cho2ЊN for the survey becau it is the latitude at which TIW-induced SST variance reaches a maximum(Hashizume et al.2001). Encountering the strong south equatorial current that exceeded1m sϪ1,the Shoyo-maru shifted its cour slightly northward to3ЊN between110Њand105ЊW,and sailed farther northeastward thereafter to occupy sta-tions for a squid
fisheries survey(Fig.1;Shiotani et al. 2000).
Throughout the crui,Vaisala RS-80GPS radio-sondes that measured air temperature,relative humidity, pressure,and wind velocity were launched four times a day.The sondes were relead from a U-shaped wind screen made of strong vinyl on the rear deck of the Shoyo-maru,which was generally moving eastward(ϳ6 m sϪ1)against the wind.During an intensive obrva-tion period that was centered on a cold cusp at125ЊW, the launch frequency was incread to eight times a day.
A total of36soundings launched between140Њand 105ЊW are ud in this study(Fig.1).The soundings are binned at a10-m vertical resolution using the linear interpolation.
During the crui,surface meteorological data were also measured at1-min intervals.Continuous surface wind velocity and SLP measurements were made at17-and10-m heights,respectively.SST were also contin-uously measured from the intake water at the bottom of the ship.Five-minute averages are shown in this paper for the surface measurements.
Once every day,an ozonesonde(Science Pump ECC ozonesonde)was attached to the GPS radiosonde that measured ozone mixing ratio up to25-km height.De-tails of ozone measurements and their results are pre-nted in Shiotani et al.(2002).
3.Satellite measurements
We u concurrent satellite measurements to supple-ment and put the Shoyo-maru survey into a large-scale perspective.Specifically,the TRMM Microwave Imager (TMI)makes measurements of SST through clouds ex-
F IG.1.Sounding sites(triangle symbols)during the Shoyo-maru crui from16Sep to9Oct1999.Clod triangles(numbered from1 to36)refer to the sounding sites ud in this paper.Superimpod is the TMI SST distribution averaged for22–24Sep.Heavy(light)shade denotes23ЊCϽSSTϽ24ЊC(24ЊCϽSSTϽ25ЊC).The meanders of the equatorial front along2ЊN are caud by tropical instability waves.
cept under raining conditions and improves the sampling in cloudy regions as compared with traditional infrared radiometers(Wentz et al.2000).We u a datat for column integrated water vapor(WV)and cloud liquid water(CLW)derived by combining the TMI product and measurements by three co-orbiting(F-11,F-12,F-14)SSM/Is(Hashizume et al.2001).The original TMI and SSM/I data were procesd at Remote Sensing Sys-tems(Wentz,1997)and obtained via FTP.The TMI and SSM/I data are originally gridded at a0.25Њϫ0.25Њand twice-daily resolution and we regridded the
m as3-day means at the0.5Њϫ0.5Њresolution.This3-day averagingfills nearly all the oceanic grid points.The QuikSCAT wind velocity data are gridded by a succes-sive correction method(Liu et al.1998)at a0.5Њϫ0.5Њand twice-daily resolution.
Both satellite and in situ measurements sample high-frequency atmospheric weather events such as the east-erly waves that develop along the ITCZ.Taking advan-tage of the large disparity in typical wavelength between atmospheric easterly waves(ϳ6000km)and oceanic TIWs(ϳ1000km),we remove12Њmoving averages in the zonal direction from satellite data to extract TIW signals.Hashizume et al.(2001)show that this high-pass zonalfilter removes weather noi quite effectively (their Fig.3).A similar spatialfilter of zonal moving average with a Gaussian-type weight(effective radius 5Њ)is also applied to the radiosonde data along2ЊN. Hereafter,high-passfiltered variables are referred to as TIW-induced anomalies,or simply anomalies.
Figure2a shows the longitude–time ction of high-passfiltered TMI SST at2ЊN,with the straight solid line denoting the track of the Shoyo-maru crui,which sampled three major SST waves between140Њand 110ЊW.The typical amplitude of SST waves is1ЊC.TMI SST tracks the ship SST measurements quite well,par-ticularly west of125ЊW but showing a cold bias of0.5Њ–1ЊC to the east(Fig.3b).
baby中文歌词
The combined TMI and SSM/I data indicate that TIWs exert a clear influence over WV and to a lesr degree,over CLW(Figs.2b,c).In order to further sharp-en TIW signals,we make moving averages along the westward-propagating TIW pha line at each grid point for a13-day period centered at the time of the Shoyo-maru’s passage.The resultant composite distributions reinforce the notion that WV and CLW are largely in pha with SST(Fig.2d),confirming the results of Hashizume(1998),Liu et al.(2000),and Hashizume et al.(2001).
The longitude–time ctions of high-passfiltered wind velocities also show a clear westward copro-pagation with SST(Figs.2e,f).The13-day compos-ites reaffirm this association(Fig.2g);both wind com-ponents are nearly in pha with local SST anomalies, but the maximum easterly anomaly tends to shift slightly to the east of the SST maximum,a pha difference Hashizume et al.(2001)attributed to SLP effects.
The association between wind velocity and SST is quite apparent in the longitude–time ctions bad on satellite measurements.Detecting TIW signals becomes more difficult,however,with a one-time tranct on board the Shoyo-maru,which contains high-frequency variability unrelated to TIWs.Figure3a shows5-min zonal wind speed measured on board the Shoyo-maru (solid line)that contains more variability than the sat-ellite measurements.The Shoyo-maru and satellite mea-surem
ents are similar on the TIW scale,except over the SST minimum centered at135ЊW.The QuikSCAT cap-tures the reduced wind speed over the125ЊW SST min-imum but shows little covariability with SST between 122Њand115ЊW.Regarding the meridional wind veloc-
F IG.2.Time–longitude ctions of satellite measurements along2ЊN:(a)SST(ЊC),(b)column-integrated water vapor(WV;mm),(c)cloud liquid water(CLW;mm),(e)zonal wind(U;m sϪ1), and(f)meridional wind(V;m sϪ1).Positive values are shaded and the straight lines denote the Shoyo-maru crui track.(d)The composite distribution of SST(ЊC;solid line),WV(mm;dashed line),and CL
W(mm;dotted line)by moving average along the westward-propagating TIW pha line during13days before and after each point along the crui line.(g)Same as(d)except for U(m sϪ1;dashed line)and V(m sϪ1;dotted line).
QuikSCAT data indicates few TIW signals on track(Fig.2f),partly becau of the
high-frequency weather noi(e ap-四六级考试时间12月
reason,we will not discuss meridional
further in this paper.
respon to TIW-induced SST
three ctions analyze the Shoyo-maru
results will be related to satellite mea-
appropriate.We begin with tempera-humidity respon.
demerolstructure
wj
temperature generally decreas with
rate between the dry adiabatic and moist
inversion layer where temperature in-height often caps the PBL over cold ocean
统考英语b
descending branch of the Hadley circu-Hastenrath1995).Such a temperature in-obrved in all the Shoyo-maru soundings at
a strongly stratified layer at1–1.6km
ction of virtual potential temperature
4).Thefilled star symbols mark the
(e the definition in thefigure caption).
Figure5a shows a sounding in the Above a hydrostatically unstable layer near the a surface is a400-m-deep mixed both the virtual potential temperature specific humidity(dotted line)are constant tical.Temperature becomes weakly stably specific humidity starts to decrea slightly, ative of a clou
d layer from500m all inversion that is located at1550m sounding.The prence of clouds is further by high relative humidity(dashed line)
The sharp inversion,only100m thick, top,above which humidity decreas vertical structure—a thin unstable surface layer topped by a cloud layer,and an ping—is often obrved in the eastern (Kloel and Albrecht1989;Bond 2001).
b.Temperature
maincourFrom the Fig.4top,it is quite apparent influence reaches the full depth of the爱国主义教育主题班会
2In the Tropics poleward of the ITCZ and the clouds often ri above a weaker inversion over Bretherton and Pincus1995).
F I
G .3.(a)Surface zonal wind anomaly (m s Ϫ1).(b)SST anomaly (ЊC).Solid lines and open circles show the ship obrvation and the satellite obrvation,
respectively.
F I
G .4.(top)Longitude–height ction of virtual potential temper-ature (K).The star symbol marks the main inversion defined as where virtual potential temperature gradient (evaluated over 100-m layers)is largest in the part of the atmosphere with specific humidity greater than 8g kg Ϫ1.The triangle symbol marks the condary stable layer defined at the lowest level where virtual potential temperature in-creas by more than 0.5K with a 100-m increa in height.(bottom)SST (ЊC)obrved on board the Shoyo-maru.The numerals at the bottom denote the sounding numbers (e Fig.1)and the shading marks nighttime.
layer temperature follows the SST ,with its minima clearly locked to the cusps of TIWs.The inversion layer varies its height (star symbols)from 1000to 1600m,again following the underlying SST variation
s.For ex-ample,SST increas by 2ЊC from 136Њto 131ЊW ,and the inversion height increas by 300m.In respon to the SST minimum centered at 125ЊW ,the inversion drops its height as much as 500m.
The correspondence between the inversion height and local SST is not perfect.Between 123Њand 110ЊW,the inversion height experiences one single depression (except a small dip at 112ЊW)while SST shows two waves.At 115ЊW,the inversion is lowered to 1000m despite that the SST reaches a weak local maximum.A clo look into the large-scale structure of the SST waves (Figs.1and 2a)suggests that the SST maximum at 115ЊW is short lived in a failed attempt at developing a cond trough of the SST front within the major wave that spans 123Њand 110ЊW.In this n,the lowered inversion height at 115ЊW can be viewed as a nonlocal atmospheric respon to this major SST wave.In the mixed layer below 400m,the effect of the condary SST wave on air temper-ature is still clear with local maxima of air temper-ature and SST collocated.This suggests that the mixed layer responds to local SST changes while the height of the main PBL-capping inversion is controlled more by larger-scale SST waves.
The sounding in Fig.5a,obtained at the warm pha of TIWs,is characterized by a smooth transition from the mixed layer to a cloud layer above.At TIW’s cold pha (115ЊW),by contrast,the mixed lay
er is parated from the upper PBL by a weak,condary inversion (Fig.5b).Such a local maximum in static stability at the top of the mixed layer is commonly en in the Shoyo-maru soundings over the colder ctors of TIWs (triangle symbols in Fig.4).We define a condary sta-ble layer as the first level from the surface where virtual potential temperature increas by 0.5K in a 100-m layer.
Diurnal SST variations are small in this oceanic re-gion,but the cloud-topped PBL can display significant diurnal cycle due to radiative forcing,with enhanced mixing and cloudiness in the predawn hours (Der and Smith 1998;Ciesielski et al.2001).However,the diurnal cycle does not em to dominate our soundings.For example,we identified 14soundings with a condary stable layer at the top of the mixed layer.Among them,six were taken in day and eight at night (Fig.4).Instead,the condary stable layer almost never occurs over the warmer ctor of TIWs but frequents the Humidity and clouds
Dry convection takes place in the mixed layer where temperature decreas with height rapidly at the adiabatic lap rate.At the top of the mixed layer,the temperature cools down to a level that allows clouds to form.By releasing latent heat and emitting longwave radiation to space,the stratus clouds,capped by the main inversion,are the main agent for mixing above the mixed layer.At the warm pha of TIWs,the main inversion ris.In a thick layer be-tween this inversion and mixed-laye
r top,relative hu-midity is high and virtual potential temperature in-creas slightly with height (Fig.5a),both suggestive of the prence of clouds.
