Potential of profiling floats to enhance NASA’s mission

Outline:

What are profiling floats?

Studies to date involving optics and profiling floats.

Apex float 0005.

Emmanuel BossUniversity of Maine

Collaborators: L. Taylor, P. Brickley, D. Swift, R. Zaneveld, S. Riser, C. Moore, MJ Perry,P. Strutton

‘Lagrangian’ profiling platforms (Davis, 90s):

Autonomous Lagrangian Circulation Explorer (ALACE)

•Global coverage without need for acoustic sources.

•At the surface, typically ½ to 1 day (ARGOS link). Improved with Irridium.

•Can obtain a profile on the way (PALACE).

•Lifetimes of 5 years or 200 cycles to 1000 m.

•full vertical control allows complex depth vs. time missions and active isopycnal following.

There are several domestic and foreign commercial manufacturers.

Floats with optics:K-SOLO float (G. Mitchell, M. Kahru, J. Sherman, 2000)

3-wavelengh downwelling irradiance (Ed) sensor (380, 490 and 555 nm)

K = -1/Ed dEd/dz

Vertical profiles of temperature (__) and irradiance at three wavelengths (380 m - ; 490 nm - ; 555 nm - ) transmitted by KSOLO for A. March 11 and B. May 12, 2000.

A B

Some results:

•Vernal bloom

•Satellite validation kd(490)

Floats with optics:

SOLO float (J. Bishop, R. Davis, J. Sherman) – Carbon explorers

Beam-c used as a proxy of POC.

Results:

Two blooms of phytoplankton in North Pacific following dust deposition event.

SeaWIFS [chl] and POC covary.

Bishop et al., Science, 2002

Measure beam-c in and out of an Iron seeded patch.

Use the transmission due to particles that settled on optical window at depth to quantify carbon flux.

SOLO floats (J. Bishop, R. Davis, J. Sherman) – Carbon explorers

Bishop et al., Science, 2004

Upcomming Project: PROBIO

12 ARGO floats + radiometer + transmissiometer

+ (Chla & CDOM) fluoresence + bb meter + irridium

• Deployment planned in 2007:

• Med Sea, South Pacific Gyre, North Pacific Gyre, North Atlantic

• Iridium capability : two-way communication (dialogue) :

to adaptively sample in time, and as well along the vertical, in order to resolve some events (e.g. storms, bloom, matchups with sensors…) as accurately as possible.

Funding by CNES, Operational Agency (Coriolis), and ANR

(Hervé Claustre, PI)

Radiometer (412 490,555)

Transmissiometerc(660)

Chla fluorescence, CDOM fluorescence, bb(540)

For PROBIO

Added bio-optical sensors

Radiometer (3λ)C-meterChla fluorescenceCDOM fluorescencebb-meter

APEX Float 0005:

New auxiliary board.New (Lithium) batteries.Chl-LSS combined sensorO2 (failed after 6 months).To date: 206 profiles, one every 5days.

-60 -55 -50 -45 -40 -3545

50

55

60

65

Longitude

Lat

itude

09/10/04

12/19/04

03/29/05

07/07/05

10/15/05

01/23/06

05/03/06

08/11/06

11/19/06

02/27/07

Greenland

Newfoundland

-60 -55 -50 -45 -40 -3545

50

55

60

65

Longitude

Lat

itude

4

6

8

10

12

14

16

18Greenland

Calibration:

•Vicarious calibration with of Fchl-LSS sensor with three Fchl and bb(440) sensors deployed by P. Strutton on a mooring at the same time as the float.

•The mooring drifted away from its anchor and we used 2.5months of the deployment for the calibration.

•Strutton’s Fchl calibration was done with local in-water samples (factor of 3 difference relative to laboratory phytoplankton culture).

Note:

•Fchl while an IOP varies with species composition, light history, amount of pigment per cell.

•Float data used here was taken around midnight to avoid NPQ.

27.6

27.65

27.7

27.75

27.8

dens

ity-1

000

[kg]

Jun04 Jun05 Jun06 Jun070

0.5

1

1.5

2x 10

-3

back

sca

tter

[m-1

]

Jun04 Jun05 Jun06 Jun070

0.02

0.04

0.06

0.08

0.1

chlo

roph

yll [

mg/

m3 ]

34.86

34.88

34.9

34.92

34.94

34.96

salin

ity [

psu]

Stability of sensors, values at depths (975-1000m):

For reference: bbsw(440)~2.2x10-3 m-1

0 5 10 15 20 250.55

0.6

0.65

0.7

0.75

0.8

0.85

Distance from float (km)

Cor

rela

tion

coef

fici

ent c

hlfl

oat v

s. c

hlsa

t

Modis LinearModis NearestSeawifs LinearSeawifs Nearest

0 5 10 15 20 2550

100

150

200

250

300

350

400

Distance from float (km)

Num

er o

f m

atch

up p

oint

s

What is the appropriate horizontal scale to compare the satellite and float data over?

Note: we use 4km SeaWIFS data and 1km MODIS data

Consistent with the (Rossby) deformation radius.

Hereon we use L<7.5km averaging for all comparison of remote-sensing with float.

Smith et al., 2000, JPO

0 2 4 6 80

2

4

6

8

Float chlorophyll

MO

DIS

/Sea

WIF

S ch

loro

phyl

l

0 0.002 0.004 0.006 0.0080

0.002

0.004

0.006

0.008

0.01

Float bb(440)

MO

DIS

/Sea

WIF

S In

vers

ion:

b b(440

)

MODISSeaWIFS

Jun04 Jun05 Jun06 Jun070

0.005

0.01

0.015

b b(4

40)

Jun04 Jun05 Jun06 Jun070

2

4

6

8

Chl

orop

hyll

[mg

m-3

]

Float 0005MODISSeaWIFS

Stability of sensors: comparison with remotely derived surface IOPs.

bb based on semi-analytical inversion (IOCCG, 2006, ch. 8)

Nov06‘Bloom’

R=0.76, 0.82

R=0.7, 0.86

An attractive feature of floats is that they can sample under clouds.

How important is it for the region under consideration?

2004 2005 2006 2007 20080

5

10

15

FloatSeaWIFS

2004 2005 2006 2007 20080

5

10

15

20

FloatModis

Clouds mask the region over extensive periods each year.

Jun04 Jun05 Jun06 Jun07

0

0.5

1

0

10

20Jun04 Jun05 Jun06 Jun07

0

0.5

1

0

10

20

Indeed, coverage from space correlates with normalized variability:

Sta

ndard

devia

tion/m

ed

ian o

f float’

s ch

loro

phyll

Num

ber o

f match

up

s per m

onth

SeaWIFS

MODIS

Apr04 Jul04 Oct04 Jan05 Apr05 Jul05 Oct05 Jan06 Apr06 Jul06 Oct06 Jan07 Apr070.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Inte

grat

ed b

b(440

) [

]

20

40

60

80

100

120

140

160

180

The ‘November 2006 event’:

Within a month following the float’s passage to subtropical gyre an coherent increase is observed in bb at depth associated with a small surface bloom.

Note: integrated properties vary by a factor < 4 between summer and winter.

Integrated chlorophyll [mg m

-2]

0

0.5

1

1.5

2

chlo

roph

yll

Aug06 Oct06 Dec06 Feb070

0.5

1

1.5

x 10-3

b b(440

)

24.5

25

25.5

26

26.5

27

27.5

28

dens

ity

Aug06 Oct06 Dec06 Feb0733

33.5

34

34.5

35

35.5

36

36.5

salin

ity

0-50m80-130m200-250m600-650m>950m

Anticyclonic, low density eddy with enhanced bb down to depth.

Deep data suggests LARGE ‘flux’:

Surface data suggests small ‘bloom’:

-60 -50 -40 -3045

50

55

60

65

Longitude

Lat

itude

09/20/06

10/10/06

10/30/06

11/19/06

12/09/06

12/29/06

Float moves in association with an anticyclonic (light) eddy

Altimeter data (surface anomaly) suggests float is caught in eddy:

Hypothesis:

1. E. huxleyi bloom (But liths need to be repackaged to provide depth coherence).

2. Lateral input (not consistent with T/S).

3. Concentration by eddy. (How?)

4. Aerosol deposition (No anomalous observed).

Association of flux events with mesoscale processes (eddies) has been observed before.

Similarities with the ‘Christmas bloom’ at BATS (Conte et al., 2003):

Evidence on 2-D processes in floats profiles:

~10% of the profiles show chlorophyll profiles that cannot be explained by 1-D models (e.g. multiple chl peaks).

Likely processes: slantwise convection and interleaving of water masses.

0 0.2 0.4-300

-200

-100

0chlorophyll [mg m-3]

dept

h [d

b]

12 13 14-300

-200

-100

0Temperature [C]

dept

h [d

b]

35.5 35.6 35.7 35.8-300

-200

-100

0Salinity

0 2 4 6

x 10-4

-300

-200

-100

0

bb(440) [m-1]

Value of floats consistent with NASA’s mission:

1. Sampling under cloud (is it a problem?).

2. Sub-surface structure (in particular when lateral processes are at work).

3. Validation – magnitude of uncertainties in products for given horizontal scales.

4. Ability to collect cheaply independent data to be used in for improvement of algorithms.

Where do we go from here:

Currently a strong community push of addition of O2 sensors (N. Gruber).

We (NASA (?)) should spearhead the push to add optical sensors.

As of April 9th 2007, 2820 floats are profiling the oceans.

Imagine how much we could learn if 10% of them had biogeochemical sensors…

How biased are averages obtained from space at the location of the float and during its 3yrs deployment due to clouds?

0 2 4 6 8 10 12-100

-50

0

50

0 2 4 6 8 10 12 14 16 18-300

-200

-100

0

100

Numbers of matchups for a given month

% bias in monthly median value of (chlfloat-chlsat)/chlfloat:

In months when clouds are significant there is little variability little bias.