04/19/23 1

Air-Sea Interaction:Physics of air-surface interactions and coupling to

ocean/atmosphere BL processes

• Emphasize surface fluxes

• Statement of problem

• Present status

• Parameterization issues

• An amusing case

04/19/23 2

Flux Definitions

plsnlnsnet HHHRRHHeatNet :

)(:

ˆ''ˆ'':

'':

'':

wetspwp

yaxa

eal

paas

TTPcHHeatRain

juwiuwStress

TwLHHeatLatent

TwcHHeatSensible

ealw LHPEPFWater /:

)'(')()'(':

'':

)(/:

/61.0/:

rnwrnwrnwFExchangeParticle

rwFExchangeGas

PEgcgHFWater

LHTcHFAir

sgn

xx

pwwnetb

ealpaasb

04/19/23 3

Present Status of Surface Flux Parameterizations

• P: No dependence on surface variables• Radiation: Depends on albedo, emissivity, and Ts

but real problem is clouds• Turbulent Fluxes: Bulk Parameterization

),,,,(

;)()(:

.'':

)('':

* slopebreakingwaveuUfFF

rnrVFParticles

solXkxwFluxGas

XUCXXUCxwFluxMet

whitecapsource

ddeposition

xx

xrsx

04/19/23 4

Physically-Based Parameterizations

2/)]/)()(

(1][)(4

9exp[)(

3

4:

]2/)ln(5[)]/ln([/

]/['':

)]/()/)][ln(/()/[ln(

)]()(*98.0[])()(['':

3/43

2/12/12/1*

2/122*

222

u

gs

cadcaaxwrwcwwaw

arxwraxx

qoquo

ssatmesoyx

SloperWhUerf

ra

l

frn

rDropletSpraySea

SCShzSh

XXurwFluxGas

LzzzLzzz

zqTqUWUUqwFluxMet

Old Days: CE=1E-3 and k=0.003*U2 and spray=S(r)*fwhitecap

04/19/23 5

Historical perspective on turbulent fluxes:Typical moisture transfer coefficients

Algorithms of UA (solid lines), COARE 2.5 (dotted lines), CCM3 (short-dashed lines), ECMWF (dot-dashed lines), NCEP (tripledot-dashed lines), and GEOS (long-dashed lines) .

04/19/23 6

04/19/23 7

Air-Sea transfer coefficients as a function of wind speed: latent heat flux (upper panel) and momentum flux (lower panel). The red line is the COARE algorithm version 3.0; the circles are the average of direct flux measurements from 12 ETL cruises (1990-1999); the dashed line the original NCEP model.

04/19/23 8

04/19/23 9

CO2 Flux: Transfer velocity versus wind speedHare, McGillis, Edson, Fairall

Work under way on DMS and Ozone

04/19/23 10

Particle Fluxes

• Optically relevant (.1 – 10 micron):– Principally whitecap-bubble production– Measurement and interpretation problems– Some dependence on laboratory work– No consensus

• Thermodynamically relevant (50-500 micron)– Principally breaking-wave spume production– No measurements at high winds– Order of magnitude uncertainty

04/19/23 11

Progress in Last 5ish Years

• Conventional turbulent fluxes: – Greatly expanded data base– 5% 0-20 m/s– Progress on wind-wave-stress models– M-O stability functions, light-wind convective & stable

• Gas Fluxes: – Ship-based covariance measurements– Physically-based parameterization

• Particle Fluxes: – Expanded modeling efforts

04/19/23 12

Flux Parameterization Issues

• Representation in GCM– Except for P, most observations are point time averages– Concept of gustiness sufficient? – Mesoscale variable? Precip, convective mass flux, …

• Strong winds– General question of turbulent fluxes, flow separation, wave momentum input– Sea spray influence

• Waves– Stress vector vs wind vector (2-D wave spectrum)– zo vs wave age & wave height

• Breaking waves– Gas and particle fluxes– Distribution of stress and TKE in ocean mixed layer (P. Sullivan)

• Gas fluxes– Bubbles– Surfactants (physical vs chemical effects)– Extend models to chemical reactions

• Particle fluxes– Interpretation of measurements– Source vs deposition

04/19/23 13

Turbulent Fluxes at High Winds

04/19/23 14

Strong wind turbulent fluxes

• Direct turbulent fluxes– Cd or Charnock coeff– Ch/Ce or zot/zoq=f(Rr)

• Droplet mediated fluxes– Momentum <ρwu>– Mass flux <ρw>

• Enthalpy flux; partitioning Qs and Ql

04/19/23 15

Evidence • Strom surge models• Cd/Ck ratio, Emanuel• Powell drop sonde profiles• Price ocean mixed layer integrations• Laboratory simulations

Explanations

Slippery young waves (direct Cd) – Moon et alDroplet mass effect (ρ<w’u’>– AndreasDroplet stability effect (<w’ ρ’> - Makin

04/19/23 16

EM-APEX163416331636

day 250

GOES SST imagery. Daily composites made from hourly images. GOES seemsto be the most prolific SST imaging system, though at the expense of accuracy and noise level.

SST cooling in these images exhibits:1) Significant horizontal structure, i) a marked rightward bias, ii) along-track variability that is not correlated with intensity, and,2) A rapid relaxation back toward pre-storm SST, e-folding approx 10 days.

04/19/23 17

The numerical ocean model is Price et al., '94; grid-level, high resolution, closed with PWP upper ocean mixing algorithm. The ocean IC is from pre-Frances EM-APEX. The single most important thing is the hurricane stress field: a fit to HWINDS for the wind field and Powell et al. for the drag coefficient. The implicit assumption is that stressocean = stressair and so this is the null model with respect to some of the most interesting effects of surface waves.

A numerical simulation of the UO response

04/19/23 18

04/19/23 19

A Sea-Spray Thermodynamic Parameterization Including Feedback

C. W. Fairall *, J-W. Bao, and J. WilczakNOAA Environmental Technology Laboratory (ETL)

Boulder, CO

1. Background

2. Source strength

3. Feedback

4. Sensitivities

5. Model tests

04/19/23 20

04/19/23 21

Original Droplet EquationsFairall/Andreas circa 1990

H c C U T Ts a pa H o a' ( )

H L C U q T ql a e E s o a' ( ( ) )

Q c F T Ts w pw v o a' ( )

Q L Fl w e E'

F r S r drv n 4 3 3 / ( )

F r S r drE

f

rn 4 3

33

/ ( )

04/19/23 22

Sn Surface Source Strength for Sea Spray Droplets

04/19/23 23

Droplet Source Functions

4

3

9

41 2

34 3

rS r

f

l rerf

U h V Slopen

f

u

( )P r

exp [ ( ) ] * [ (( ) /

) ] //

S f U S rn no ( ) ( )

f U W Ub( ) . . 3 8 1 0 6 3 4

P energy wave breaking σ surface tensionr droplet radiusη Kolmogorov microscale f fraction of P going into droplet productionVf=droplet mean fall velocity

Fairall et al. 1994

Fairall, Banner, Asher Physical Model

04/19/23 24

Feedback

04/19/23 25

Partitioning of Droplet Contribution:Stages of cooling/evaporation

• Simplification: consider large droplets that are ejected, cool to wet bulb temperature and re-enter ocean with negligible change in mass

• Stages:– Cool from To to Tair = Qs– Cool from Tair to Twet = Ql_a– Evaporation while at Twet = Ql_b

• Total droplet enthapy transfer Qse=Qs+Ql_a• Enthalpy Bowen ratio = Qs/Ql_a=(To-Ta)/(Ta-Twet)• Qs=Qse*bowen/(1+bowen)

04/19/23 26

Feedback Characterization δTa

H L C U q T q T Tl a e E s o s d d [ ( ) ( ) ]

H L C U T T Ts a e E o a a [ ( ) ]

Q c F T Tse w pw v o w ( )

Q L G U h T q T T q T Tlb w e o s a a s d d ( ) ( )[ ( ) ( )]

T T feed T T feed sa d a w

1 1

1* ( ) * ( )

feedQ

Q H Q H feed tunel

l s se s

( ) / _

Effect on the fluxes:

04/19/23 27

Turbulent Fluxes Above the Droplet Evaporation Layer

H H Q Q H L C U q Q Ql to t l la lb l e E a la l_ ' '

H H Q H Q H c C U T Q H Qs to t s s s lb s p H f s s l_ ' '

04/19/23 28

Direct Transfer Coefficients Assumed in Parameterization

10 15 20 25 30 35 40 45 50 55 60

1

1.5

2

2.5

3

3.5

4

x 10-3

U10 (m/s)

Cd,

Ck

04/19/23 29

Ratio of Transfer Coefficients With Droplet Enthalpy Flux

10 15 20 25 30 35 40 45 50 55 600.8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

U10

(m/s)

Cd/C

k

Feedtune=0.3, 1.0, 3.0

source=1.0

source=0.1

04/19/23 30

Feedback Sensitivity:Source Strength=0.3

04/19/23 31

Model Tests (Bao and Ginis)

• IVAN, ISABEL– GFDL operational– GFDL new zo, zt– WRF

• PLANS– HWRF at high resolution – matrix of tune values– Explicit droplet model (Kepert/ Fairall) in HWRF– Coordinate with Penn State LES work

04/19/23 32

Simulation with GFDL Operational Model: Isabel

04/19/23 33

But:Simulations with New Cd and Ce/Ch

New Cd Ce/Ch Old Cd Ce/Ch