william drennan and erik sahlée · william drennan and erik sahlée. rsmas, university of miami,...

William Drennan and Erik Sahlée

RSMAS, University

of Miami, FloridaMeteorology, Uppsala Universitet, Uppsala, Sweden

On air-sea fluxes in high winds-

a view from the trenches (troughs)

Photo: ASIS direct flux mooring during TS Ernesto

From Fairall

et al 2009, OceanObs, Venice.

Very good agreement (in the mean) for data in range 4 –

16 m/s

Very little data for U > 18 m/s. The few existing data points are from isolated events, and may not be representative of more general conditions.

UM-ASIS datapre-2005:

Only 5 out of 1200 hours for U >18 m/s.

Why so few flux data for U> 18m/s ? This plots shows maximum annual SAR wind speeds in 3x3 deg degrees boxes.

Scatterometer daily winds: High wind events are often transient events

Another reason for the paucity of high wind fluxes are that platforms aren’t reliable in these conditions.

Sensors are also a problem …

ASIS drift track in blue (8 days, 200km)

0

5

10

15

20

U10

(m s

-1)

R/V BrownASIS

71 72 73 74 75 76 77 7890

180

270

360

Win

d D

ir (°

)

Year day 2008

0.05

0.1

0.15

0.2

0.25

0.3Fr

eque

ncy

[Hz]

Wave Spectral Energy(Evolution of 1D spectrum)

2

4

6

Hs [m

]

Significant Wave Height

71 72 73 74 75 76 77 78

0.5

1

1.5

U10

/cp

Year day, 2008

Inverse wave ageFully Developed

0.4 Hz

0.2 Hz

U10=4.2

5

10

15

20

25

Spec

tral e

nerg

y de

nsity

, [m

2 /Hz]

0.4 Hz

0.2 Hz

U10=15.1

Southern Ocean waves: Unimodal, with no spectralseparation between windsea

and swell

0

5

10

15

20

U10

(m s

-1)

R/V BrownASIS

A

A

B

B

Drag coefficient at high winds

0 5 10 15 200

0.5

1

1.5

2

2.5

3

3.5

4x 10

-3

U10 (ms-1)

CD

N

CDN during SoGasex

E1/E2<0.2

0.2<=E1/E2<1

1<=E1/E2<2

2<=E1/E2<4

4<=E1/E2

COARE 3.0

E1 = swell energy No significant

effect of swell on CD E2 = windsea

energy in unimodal

seas of Southern Ocean

Aircraft data

- Eddy correlation- Flight altitudes 35 –

45m- Flights in surface layer-

Fluxes corrected for height using Donelan

1990

GFD CDs are similar to ASIS-SOGasex

CDs, showing enhancement above COARE3 and Smith curves around 16-

20m/s. Above 20m/s, there is no clear trend.

1) Group GPS dropsonde profiles by wind speed at mean BL height2) Assuming mean log profile, extrapolate mean profiles to find zo, thence CD

CD levels off by 40m/s, then decreases at 50m/s

U. Miami ASIST wind wave flume15m x 1m x 1m

Drag levels off at high winds

Figure 4. Vorticity contours obtained via Digital Particle Image Velocimetry (DPIV) in the air flow over wind driven waves [Reul, 1998]. Both wave and air flow are from left to right. (Top) waves of gentle slope – non-separated flow. (Bottom) waves of steep slope – separated flow.

Z(cm)

0 10 20 30 40x (cm)

0

0

6

6

Hypothesis: Flow separation over steep waves reduces effective roughness and drag

BAMS, 2007 “CBLAST”. Also:

French et al 2007 (JAS)

Drennan et al 2007 (JAS)

Zhang et al 2008 (GRL and BLM)

Zhang et al 2009 (JAS)

A: Calibration

B: Eye penetration

Stepped Descents C2

C3 C1

IsabelNOAA GOES-12 VIS image14 Sept 2003, 17:45

St Croix, USVI(colour~ time)

CBLAST data

- Eddy correlation fluxes- Flight altitudes 60 –

400m- Flights in BL, above surface layer-

Momentum fluxes corrected to surface layer using Donelan

1990

Asterisks: CBLAST . Asterisks: CBLAST . FairallFairall et al. (2003); et al. (2003);

------ Smith (1980); Smith (1980); ------ DonelanDonelan et al. (2004); et al. (2004); -------- Powell et al. (2003)Powell et al. (2003)

CBLAST data

- Eddy correlation fluxes- Flight altitudes 60 –

400m- Flights in BL, above surface layer-

Momentum fluxes corrected to surface layer using Donelan

1990 -

Surface winds from SFMR (microwave radiometer)

Again, CD levels off at high U

Drag coefficients in hurricanes٭ CBLAST: French et al 2007◊

Tank: Donelan et al 2004□

GPS Profile: Powell et al 2003

The three high wind results are in qualitative agreement: CD

levels off at high winds. But all are lower than GFD/SOGasex

at 20m/s. These results need to be confirmed with in situ field measurements. Even then, they may not be representative of high latitude conditions, where the wave fields are very different.

0 5 10 15 200.5

1

1.5

2

2.5x 10

-3

Cd1

0n

GFD/SOGasex

What about scalar fluxes ?

Dalton number at high winds

O AGILE (Donelan & Drennan 1995) X HEXOS (Decosmo

et al 1996) ◊

GASEX-ASIS (McGillis

et al 2004)SOWEX (Banner et al 1999)

□

SWADE (Katsaros

et al 1993)∆

CBLAST (Drennan et al 2007)

∇

→

CE

(and CH

) independent of wind to 30 m/s

“Extra”

scatter attributable to sampling

Dalton number at high winds

???

GFD data(Petersen, Renfrew)

●

ASIST flume▲ CBLAST aircraftX

HEXOS tower

CK

~ CE

~ CH

Geophys. Res. Lett, in press

qLTqcqck vlpd ++−= ])1([

Laboratory data indicate constant bulk heat coefficients to 40m/s.No evidence for sea spray enhancement

Ck

= bulk coefficient for moist enthalpy, k

So, what about sea spray ? These are sea spray fluxes (source functions)from the Uni. Leeds CLASP (Compact Lightweight Aerosol Spectrometer –

Norris et al. 2008) mounted near the ASIS sonic anemometer during SOGasex.

10-2 10-1 102 103

How to get fluxes at high winds ? We will be deploying 2 direct

flux moorings off Taiwan during the “Impact of Typhoons of the Pacific”

experiment, June-November 2010. The buoys will measure direct fluxes of momentum, latent and sensible heat, and sea spray, SW and LW radiation, waves, currents, TS, biochemistry). Assuming we have anything left after November, we plan on a long term (1 year) deployment in the Southern Ocean…