High Resolution MODIS Ocean ColorFred Patt1, Bryan Franz1, Gerhard Meister2, P. Jeremy Werdell3

NASA Ocean Biology Processing Group

1Science Applications International Corporation2Futuretech Corporation

3Science Systems and Applications, Inc.Concept

• The goal is to investigate the utility of the 250 and 500-meter land bands for ocean color applications, and to provide that capability to SeaDAS users.

• A second goal is to investigate the use of the short-wave IR bands for use in atmospheric correction (Wang 2005), especially for coastal applications, and to provide that capability to SeaDAS users.

• The approach is to define a virtual sensor which includes all the 250 and 500-meter channels, as well as the 1000-meter channels in the visible and near IR, at a consistent resolution.

• We call this virtual sensor HIRES MODIS, with Aqua and Terra variants known as HMODISA & HMODIST.

Band Wavelength Resolution SNR Ltypical

(nm) (meters) (W cm -2 m-1 sr-1)

8 412 1000 880 44.9

9 443 1000 838 41.9

3 469 500 243 35.3

10 488 1000 802 32.1

11 531 1000 754 27.9

12 551 1000 750 21.0

4 555 500 228 29.0

1 645 250 128 21.8

13 667 1000 910 9.5

14 678 1000 1087 8.7

15 748 1000 586 10.2

2 859 250 201 24.7

16 869 1000 516 6.2

5 1240 500 74 5.4

6 1640 * 500 275 7.3

7 2130 500 110 1.0

* band 6 failing on MODIS/Aqua

Characteristics of MODIS Visible and NIR Bands

* band 6 failing on MODIS/Aqua

RGB Image

Standard MODISvs

HIRES MODIS

250 m

75%

1 km

Progress to Date

• The processing software has been modified to read the additional MODIS bands, and to allow processing at a user-specified spatial resolution (250, 500, or 1000-meters).

• Interpolation is used to raise the resolution of the lower-resolution bands when working at higher resolutions. When processing higher-

resolution bands at reduced resolution, the Level-1B aggregated radiance fields are utilized.

• Geolocation and path geometries are appropriately interpolated for the defined processing resolution.

• Standard Rayleigh tables have been created for the full 16-band suite, and aerosol models have been spectrally interpolated.

• A full set of aerosol model tables, specific to the band suite, has also been generated, with testing now underway.

• Polarization sensitivities have been determined, and standard polarization corrections have been implemented for all bands.

• An initial vicarious calibration to MOBY has been performed.

Processing Example

Using fixed aerosol type, with aerosol concentration from 869-nm band at 250-meter resolution

OC2 chlorophyll algorithm using 469 and 555-nm bands at 500-meter resolution.

Future Plans• Improved implementation into MSL12

– add support for thermal bands to allow co-registered SST retrievals

– update aerosol model tables– finalize vicarious calibration

• Investigate use of SWIR bands for atmospheric correction– NIR Lw correction– aerosol model selection

• Enhance SeaDAS to support HIRES MODIS – processing and display

Subframe Striping

• Subframe striping was found in band 3 ocean data

• Analysis of median DN of L1B ocean scenes shows subframe striping in bands 1-3, 5 and 7

• MCST indicates effect has been qualitatively known

Possible Correction Approach

• Look-up table of correction vs. radiance to be applied to the L1B radiances

• The best method of determining the correction is still being worked

AcknowledgementsThe work presented here is supported by the Ocean Biology Group (OBPG) at NASA Goddard Space Flight center, within the NASA Biogeochemistry Program.We also wish to acknowledge our ongoing collaboration with the Naval Research Laboratory (NRL) and the MODIS Characterization Support Team (MCST).

Band 1 Band 2

Band 4Band 3

nLw at 645nm chlorophyll using OC2 algorithm

Wang, M. and W. Shi, “Estimation of ocean contribution at the MODIS near-infrared wavelengths along the east coast of the U.S.: Two case studies”, Geophys. Res. Letters, 32, 2005.