《上层海流对台风的反应》由会员分享，可在线阅读，更多相关《上层海流对台风的反应（31页珍藏版）》请在金锄头文库上搜索。
1、Upper Ocean Response to a Upper Ocean Response to a hurricanehurricane (James F.Price 1981) (James F.Price 1981)Zhu ShunliPukyong National UniversityContents1. Introduction2. Review of observations3. A model of the upper ocean response4. Specification of the EB-10/Eloise case5. Overview of the respo
2、nse6. Comparison of the model solution with the EB-10 data7 Numerical experiments8. Conclusion1. IntroductionQuestion1.What is the reason for SST response to a hurricane and rightward bias?2. How does the response depend upon such factors as the hurricane translation speed, the ocean initial conditi
3、on, etc.?3. What role do nonlocal dynamics play in the upper ocean response4. Is there evidence that air-sea transfer coefficients increase significantly under hurricane conditions?2. Review of observationsa. Hydrographic surveysTable 1 Hydrographic studies of the sea surface temperature response to
4、 Table 1 Hydrographic studies of the sea surface temperature response to hurricanehurricaneFig.1 Response of SST across the track of hurricane Ella &Response of Fig.1 Response of SST across the track of hurricane Ella &Response of subsurface temperature across the track of hurricane Ellasubsurface t
5、emperature across the track of hurricane EllaFig 2 SST around the track of hurricane Tess&Fig 2 SST around the track of hurricane Tess&Temperature section AB made across the track of TessTemperature section AB made across the track of Tessb. Buoy/hurricane encountersa. Hydrographic surveysFig.3 Nati
6、onal Hurricane Center storm tracks of hurricanes Eloise (1975) Fig.3 National Hurricane Center storm tracks of hurricanes Eloise (1975) and Belle(1976)and Belle(1976)1)EB-10/Eloise1)EB-10/EloiseFig 4 Meteorological measurements form EB-10 during the passage of Fig 4 Meteorological measurements form
7、EB-10 during the passage of hurricane Eloise hurricane Eloise Fig 5. Ocean temperature from EB-10 Fig 5. Ocean temperature from EB-10 during and during and after the passage of Hurricane Eloiseafter the passage of Hurricane Eloise2)EB-04/Eloise2)EB-04/EloiseFig 6. Air-sea measurement (upper/lower pa
8、nels) from EB-04 Fig 6. Air-sea measurement (upper/lower panels) from EB-04 during the passage of Eloiseduring the passage of Eloise2)EB-15/Eloise2)EB-15/EloiseFig 7 Air-sea measurement (upper/lower panels) from EB-15 during Fig 7 Air-sea measurement (upper/lower panels) from EB-15 during the passag
9、e of Bellethe passage of Belle3. A model of the upper ocean response3. A model of the upper ocean responsea. Scales and physical approximationsa. Scales and physical approximationsFig 8.Schematic plan view of the numerical modelFig 8.Schematic plan view of the numerical modelPhysical approximations
10、are as follows:Physical approximations are as follows:1)The Coriolis parameter f is constant, pressure is hydrostatic, the 1)The Coriolis parameter f is constant, pressure is hydrostatic, the Boussinesp approximation and diffusive processes are excludedBoussinesp approximation and diffusive processe
11、s are excluded2)Density :2)Density :Where Where =-3.3=-3.3* *1010-4-4 and and =7.6=7.6* *1010-4 -4 are the appropriate thermal and haline are the appropriate thermal and haline expansion coefficients and expansion coefficients and 0=1.02350=1.0235* *10103 3 kg m kg m-3 -3 is the reference is the ref
12、erence densitydensity3)The sea surface is treated as a rigid lid (excludes the barotropic 3)The sea surface is treated as a rigid lid (excludes the barotropic mode).mode).4)The subthermocline ocean is taken as infinity deep and unable to 4)The subthermocline ocean is taken as infinity deep and unabl
13、e to sustain a pressure gradientsustain a pressure gradient。5 5）The temperature and salinity in layer 2 are assumed to have depth The temperature and salinity in layer 2 are assumed to have depth dependencedependence。The velocity within layer 2 is assumed constant. The initial The velocity within la
14、yer 2 is assumed constant. The initial thickness of layer is chosen so that after the hurricane passage, h2h1thickness of layer is chosen so that after the hurricane passage, h2h16)Vertical density gradients in layer 2 and 3 are held constant since the 6)Vertical density gradients in layer 2 and 3 a
15、re held constant since the stretching is smallstretching is smallb. Budgetsb. BudgetsLayer thickness :Layer thickness :Entrainment velocity We0 Heat and salt balance in the ML :Heat and salt balance in the ML :The heat flux Q=Qs+QL + R is the sum of sensible , latent and radiative Heat exchange acro
16、ss the sea surface F is the mass flux, equal to evaporation minus precipitationThe term TWe is called the entrainment heat flux and is the bulk Representation of the turbulent heat flux Heat and salt balance in the ML :Heat and salt balance in the ML :is understood to be vertical gradient of T withi
17、n layer 2Momentum balances for the layers:Momentum balances for the layers:Where f is the Coriolis parameter times the vertical unit vector, is wind stress,P is the hydrostatic pressure perturbationThe pressure gradients:The pressure gradients:The horizontal density gradient in the ML:The horizontal
18、 density gradient in the ML:And in the lower layers C. Turbulent-flux parameterizationsC. Turbulent-flux parameterizationsSensible heat fluxSensible heat fluxLatent heat fluxLatent heat fluxDensity of air divided by density of seawaterBulk transfer coefficient,=1.3*10-3Wind speed at 10m heightB heat
19、 capacity of air divided by heat capacity of seaweater q10,ss specific humidity at 10m height, and at the sea surface assuming satuation at the sea surface temperatureK latent heat of vaporation divided by the heat capacity of seawaterC. Turbulent-flux parameterizationsC. Turbulent-flux parameteriza
20、tionsC. Turbulent-flux parameterizationsC. Turbulent-flux parameterizationsWind stress :Where the drag coefficient0, if Rv1,If Rv1is a bulk Richardson number Parameterize entrainmentFig 9 Schematic plan view of the numerical modelFig 9 Schematic plan view of the numerical modeld. Implementationd. Im
21、plementationlThe time step t = x/mUH ,Where m is the smallest Integer that gives t 2*103s lThe ocean is moved down the grid one row after everym steps and the first row of the grid points is updated with the ocean initial condition.lSpace derivative are estimated with a second-order centered Form, a
22、nd time-stepping is carried out using the three-time-level ,two-step, leapfrog trapezoidal method. Fig 10 Upwelling computed by the numerical model(left Fig 10 Upwelling computed by the numerical model(left side) and Geislers analytical solution(right side)side) and Geislers analytical solution(righ
23、t side)A check of the entrainment calculations Comparing with the quasi-steady entrainment Where n=1/2 for a linearly straitified fluid , and n=1 for a two-layer stratified fluid, Where R=(gh)/(p0) is the external Richardson Number。Table 4 EB-10/Eloise ocean initial conditionTable 4 EB-10/Eloise oce
24、an initial conditionSpecification on the EB-10/Eloise casea. Ocean initial conditionb. Model hurricaneFig 12 Radial and tangential wind-speed radial profile computed from the EB-10 Fig 12 Radial and tangential wind-speed radial profile computed from the EB-10 wind data(points), and the model of the
25、hurricane used in the simulation of the wind data(points), and the model of the hurricane used in the simulation of the EB-10/Eloise(Solid line)EB-10/Eloise(Solid line)Fig 13.ML transport V1h1 computed from the directly Fig 13.ML transport V1h1 computed from the directly observed winds of Eloise(sol
26、id line) and from the model observed winds of Eloise(solid line) and from the model hurricane (dashed line)hurricane (dashed line)the ML transports computed from a local momentum balance :Fig 15 Fig 15 Plan view of the model-predicted y component of ML velocity& The SST response Plan view of the mod
27、el-predicted y component of ML velocity& The SST response shown as the change in sea surface temperature from the initial condition shown as the change in sea surface temperature from the initial condition Rightward biasFig 16 A schematic that shows the rotation of a wind-stress vector & the cosine
28、of the angle Fig 16 A schematic that shows the rotation of a wind-stress vector & the cosine of the angle between the wind stress and the ML velocity(a measure of how efficiently the wind stress is between the wind stress and the ML velocity(a measure of how efficiently the wind stress is working to
29、 increase the ML kinetic energy )working to increase the ML kinetic energy )5. Overview of the responseb. Horizontal advection of SSTHorizontal advection of SSTFig 15b The SST response shown as the change in sea surface temperature Fig 23 a Observed and predicted sea surface temperature for the EB-1
30、0/EloiseReason: the y-component of velocity in the ML combined with the strong y-gradients in C. Upwelling, pressure gradients and velocity response in the C. Upwelling, pressure gradients and velocity response in the interiorinteriorFig 18bPumping & Fig 19bUpwelling below the base of the MLTwo dyna
31、mical consequences of the upwellingTwo dynamical consequences of the upwelling1)The first half of the upwelling cycle is upward and thus tends to cause the ML thickness to 1)The first half of the upwelling cycle is upward and thus tends to cause the ML thickness to decrease.decrease.2)Upwelling sets
32、 up a time-dependent pressure gradient which couples the ML with the interior 2)Upwelling sets up a time-dependent pressure gradient which couples the ML with the interior and causes the ML velocity to rotate 5% faster than that in a free inertial motionand causes the ML velocity to rotate 5% faster
33、 than that in a free inertial motiond. The subsurface response of temperatured. The subsurface response of temperatureFig 21 Predicted subsurface temperature response shown as a cross-track section for Fig 21 Predicted subsurface temperature response shown as a cross-track section for the EB-10 Eloi
34、se case & Initial(solid) and final(dashed) temperature profiles aty=-100,0,+100kmthe EB-10 Eloise case & Initial(solid) and final(dashed) temperature profiles aty=-100,0,+100km6. Comparison of the model solution with the EB-10 6. Comparison of the model solution with the EB-10 datadataa. Upwellinga.
35、 UpwellingFig 22 Observed(solid) and predicted(dashed ) upwelling below the base of Fig 22 Observed(solid) and predicted(dashed ) upwelling below the base of the ML during the EB-10/Eloise eventthe ML during the EB-10/Eloise eventT(t) is the corrected temperature, T0 is the initial time-averaged tem
36、perature, Is local temperature gradientb. Sea surface temperatureb. Sea surface temperatureFig 23 Observed(solid) and predicted(dashed)sea surface temperature for the EB-Fig 23 Observed(solid) and predicted(dashed)sea surface temperature for the EB-10/Eloise event & The ML heat balance for the simul
37、ation of the EB-10/Eloise event10/Eloise event & The ML heat balance for the simulation of the EB-10/Eloise eventb. Sea surface temperatureb. Sea surface temperatureAlong the track, the net air-sea exchange is 0.2*10Along the track, the net air-sea exchange is 0.2*102 2 m(2K cal cmm(2K cal cm-2-2);t
38、he net entrainment heat flux is );the net entrainment heat flux is 1.3*101.3*102 2 m(13K cal cmm(13K cal cm-2-2), or 85% of the total irreversible flux into the ML.), or 85% of the total irreversible flux into the ML.Entrainment is important in the ML momentum balance. Because layer 2 is quiescent d
39、uring the Entrainment is important in the ML momentum balance. Because layer 2 is quiescent during the entrainment process, entrainment acts to increase the ML thickness and decrease Vl. The increase entrainment process, entrainment acts to increase the ML thickness and decrease Vl. The increase in
40、ML depth by entrainment is crucial in determining the maximum ML velocity.in ML depth by entrainment is crucial in determining the maximum ML velocity.C CQQ is difficult to infer from simulations because the air-sea exchange is small. A constant drag is difficult to infer from simulations because th
41、e air-sea exchange is small. A constant drag coefficient CD=1.5*10coefficient CD=1.5*10-3-3 gives greatly reduced entrainment and an SST response40% below that gives greatly reduced entrainment and an SST response40% below that observed.observed.b. Sea surface temperatureb. Sea surface temperaturel
42、lAlong the track, the net air-sea exchange is 0.2*10Along the track, the net air-sea exchange is 0.2*102 2 m(2K cal m(2K cal cmcm-2-2);the net entrainment heat flux is 1.3*10);the net entrainment heat flux is 1.3*102 2 m(13K cal cmm(13K cal cm-2-2), or ), or 85% of the total irreversible flux into t
43、he ML.85% of the total irreversible flux into the ML.l lEntrainment is important in the ML momentum balance. Because Entrainment is important in the ML momentum balance. Because layer 2 is quiescent during the entrainment process, entrainment layer 2 is quiescent during the entrainment process, entr
44、ainment acts to increase the ML thickness and decrease Vl. The increase in acts to increase the ML thickness and decrease Vl. The increase in ML depth by entrainment is crucial in determining the maximum ML ML depth by entrainment is crucial in determining the maximum ML velocity.velocity.l lC CQQ i
45、s difficult to infer from simulations because the air-sea is difficult to infer from simulations because the air-sea exchange is small. A constant drag coefficient CD=1.5*10exchange is small. A constant drag coefficient CD=1.5*10-3-3 gives gives greatly reduced entrainment and an SST response40% bel
46、ow that greatly reduced entrainment and an SST response40% below that observed.observed.7.7.Numerical experimentsNumerical experimentsa. Parametric dependencea. Parametric dependenceHurricane: strength UHurricane: strength U10max10max, translation speed and , translation speed and size(the radius at
47、 which the wind speed eaquals 1/2 size(the radius at which the wind speed eaquals 1/2 U U10max10max ) )Ocean : initial ML depth, the temperature gradient in Ocean : initial ML depth, the temperature gradient in layer 2, and local inertial period.layer 2, and local inertial period. The ocean response
48、 is described by The ocean response is described by （the y-the y-average average SST along section ASST along section A）, , V, , Vlmax lmax (the (the maximum ML current)maximum ML current)The depenence of VThe depenence of Vlamx lamx on all on all parameters is roughly linear and can be parameters i
49、s roughly linear and can be summarized by the derivatives summarized by the derivatives Variation of the oceanic response with variation in air-Variation of the oceanic response with variation in air-sea parameterssea parametersTwo results are surprising:(1)The maximum ML current is insensitively to
50、 everything except (1)The maximum ML current is insensitively to everything except hurricane strengthhurricane strength(2) The response of (2) The response of SSTmax is insensitive to hurricane size and to SSTmax is insensitive to hurricane size and to local inertial period.local inertial period.b.
51、Dynamical experimentsb. Dynamical experimentsMethod: switch off nonlocal terms in the model equationMethod: switch off nonlocal terms in the model equationFig 24 SST response for hurricane translating at UFig 24 SST response for hurricane translating at UHH =4 and 16 m s =4 and 16 m s-1-1 Fig 25 Fig
52、 25 Upwelling and change in ML depth due to entrainment Upwelling and change in ML depth due to entrainment for hurricanes moving at UH =4 and 16 m sfor hurricanes moving at UH =4 and 16 m s-1-1Fig 26 Cross-track(section A) profile of Fig 26 Cross-track(section A) profile of SST for different SST fo
53、r different hurricane translation speeds hurricane translation speeds 8. Conclusions1)Entrainmnet is the primary mechanism that lowers the SST 1)Entrainmnet is the primary mechanism that lowers the SST beneath a moving hurricane.beneath a moving hurricane.2)The SST response is a lively function of h
54、urricane strength and 2)The SST response is a lively function of hurricane strength and translation speed , and of the initial ML depth and upper thermocline translation speed , and of the initial ML depth and upper thermocline temperature gradient temperature gradient 3)Upwelling causes a significa
55、nt enhancement of the SST response 3)Upwelling causes a significant enhancement of the SST response to a slowly moving hurricane. Horrizontal advection is important in to a slowly moving hurricane. Horrizontal advection is important in the pointwize balances.the pointwize balances.4)The EB-10/Eloise case provides evidence that Garratts wind-4)The EB-10/Eloise case provides evidence that Garratts wind-speed-dependent drag coefficient and the entrainment law are speed-dependent drag coefficient and the entrainment law are appropriate for hurricane conditionsappropriate for hurricane conditions
