Radiative Coupling in the Oceansusing MODIS-Aqua Ocean Radiance Data

Watson Gregg, Lars Nerger Cecile Rousseaux

NASA/GMAO

Assimilate MODIS-Aqua Water-Leaving Radiances to:

1) Improve understanding of the propagation of light in the surface layer Primary ProductionHeat transfer -- Changes in ocean density structure and circulation

2) Invert the radiance assimilation into new information on model phytoplankton groups, CDOM, and other internal model variables.

Chlorophyll, Phytoplankton GroupsPrimary ProductionNutrientsDOC, DIC, pCO2

Spectral Irradiance/Radiance

Outputs:

RadiativeModel

(OASIM)

CirculationModel

Winds SSTShortwave Radiation

Layer DepthsIOP, Eu

Ed

Es

Sea Ice

NASA Ocean Biogeochemical Model (NOBM)

Winds, ozone, humidity, pressure, precip. water, clouds, aerosols

Dust (Fe)

Advection-diffusion

Temperature, Layer DepthsBiogeochemicalProcesses Model

Ed

Es

Ozone

Molecules, aerosols

OxygenWater vaporCO2

air

sea

Ed

Es

(1 - )

EdEs

(1 - )EdEsEu

Ed, Es Ed, Es

LwN

Total Surface Irradiance (direct+diffuse; spectrally integrated; clear/cloudy): bias=1.6 W m-2 (0.8%)RMS=20.1 W m-2 (11%)r=0.89 (P<0.05)

OASIM

Surface Clear Sky Spectral Irradiance (PAR wavelengths):RMS=6.6%Integrated PAR:RMS=5.1%

mW

cm

-2 u

m-1 s

r-1

Modeling Water-Leaving Radiances (with assimilated chlorophyll) Tropical

Rivers(CDOM)

Cocco-lithophores

ΔLwN = LwNassim - LwNmodel

tFo bbi

LwN = ---- gi ------------- (15)n2 a + bb

∑or

450 475 500 600 625 650 675

a(λ), bb(λ)

525 550 575 700350 375 400 425

Chlorophyll components:diatomschlorophytescyanobacteriacoccolithophores

CDOM

water

aw(λ), bbw(λ)

aCDOM(λ)

ap(λ), bbp(λ)

OASIM Upwelling Irradiance(Forward Model)

detritus

ad(λ),bbd(λ)

443 488412 550 667

MODIS-Aqua Upwelling Radiance(Inverse Model)

Objective 2: Invert the radiance assimilation into new information on model phytoplankton groups, CDOM, and other internal model variables.

Consistent Ocean Chlorophyll from NPP/VIIRS NPP Science Team for Climate Data Records

Watson W. Gregg and Nancy W. Casey

Investigate the ability of an established approach to improve the consistency of VIIRS ocean color data.

Utilizes completely processed and gridded ocean color data (Level-3 Environmental Data Records):

1) applies in situ data for a posteriori correction, and then 2) applies data assimilation to correct sampling problems.

In situ Data

NODCNASA AMT

ESRIDEmpirical Satellite

Radiance-In situ Data

Assimilated Data

ESRID-Assimilated VIIRS Ocean Color Data

Derive statistical relationships

to reduce bias

Global couplednumerical model

Assimilate to reduce sampling biases

Standard processingSatellite Data

VIIRS EDRs

ESRIDReduces bias in satellite observations,and sensor-to-sensor differences.Including those due to:

Radiometric calibrationBand locations and widthsBand misregistrationOut-of-band effects/crosstalk (mostly)AlgorithmsOrbit

Gregg, et al., 2009, Remote Sensing of Environment

Satellite-Weighted Error: Global Ocean

-5%

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%P

erce

nt

Err

or

SeaWiFS 28.0% 40.6%

ESRID 0% w/h 0.7% 37.3%

ESRID 50% w/h -0.3% 38.9%

1 2

(0.130) (0.298)

(0.020)

(0.028)

(0.291)

(0.291)N=2922

N=5994

N=5920

Bias Uncertainty

w/h = withholding of in situ datalog bias and uncertainty in parentheses

ESRIDEmpirical Satellite Radiance-In situ Data Algorithm

0.150

0.155

0.160

0.165

0.170

0.175

0.180

0.185

0.190

0.195

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Ch

loro

ph

yll (

mg

m-3

)

SeaWiFS-standard

MODIS-standardSeaWiFS-ESRID

MODIS-ESRID

Mean Difference MODIS-SeaWiFS:

Standard -12.2% ESRID -0.6%

Global Annual Median Chlorophyll circa 2007 Processing

ESRID improves consistency between disparate data setsUnifies the representation of ocean biology from ships and satellitesGregg and Casey, Geophysical Research Letters, 2010

North Indian

Antarctic

EquatorialAtlantic

mg

m-3

Sampling Differences/Biases

Aerosol optical thickness limit = 0.4 globally

= 0.3North/Equatorial Indian

= 0.3North Central/South Atlantic

Data eliminated Equatorial Atlantic

Two-Step Process to Remove Sampling Differences:1) Remove high aerosol regions and tropical river regions (Equatorial Atlantic)

ESRID-AssimilatedGlobal Annual Median Chlorophyllfor MODIS-Aqua

mg

m-3

Global Annual Median Chlorophyllfor MODIS-Aqua

Step 2: Assimilate to backfill missing regions/seasons