Introduction

The Ocean Biology Processing Group (OBPG) completed a full-mission ocean color reprocessing of the COMS Geostationary Ocean Color Imager (GOCI) dataset in May 2016. This reprocessing is part of a multi-mission effort to update common algorithms, product suites, and data formats across all supported missions. Sensor-independent changes are detailed in the R2014.0 Ocean Color Reprocessing General Description.

Sensor-specific Processing Details

Source Data

The source data is a Level-1B hdf 5 format dataset containing top-of-atmosphere radiances in 8 visible and NIR bands (similar to those for SeaWiFS) for an approximately 2500 x 2500 km portion of the Earth around the Korean peninsula. The ground sampling distance (GSD) is 500 m and due to the COMS geostationary orbit, GOCI samples the same area 8 times per day, once every hour from 0900 to 1600 local time. More description of GOCI can be found in Faure et al (2008)

Note that this reprocessing coincides with the formal start of the distribution of Level-1B and Level-2 GOCI data by the OBPG.

Vicarious_Calibration

The vicarious calibration for this processing was determined by a cross-calibrating the GOCI to VIIRS, performed by NRL.

Wavelength (nm)	412	443	490	555	660	680	745	865
Gain	0.9726	0.9520	0.9258	0.8974	0.9007	0.8719	0.9430	1.0

This set of vicarious calibration coefficients improves the comparison of GOCI chlorophyll-a to that of the MODIS Aqua instrument:

The 3 chlorophyll-a images above are of the region covered by GOCI on 19 April, 2016 at around local noon. The first is the GOCI image made using the older vicarious calibration, the second is taken by Aqua, and the third is from GOCI but with the revised vicarious calibration. Generally, the chlor-a was increased using the new calibration.

The 2 plots above show histograms of the pixel-to-pixel ratio of chlor-a from GOCI to Aqua. The first is using the old GOCI vicarious calibration while the second uses the current GOCI calibration. The old calibration results in a lower chlorophyll estimate - about 0.6 that for Aqua. The new calibration brings the ratio to about 0.9 of Aqua.

Level-2 Processing

The Level-2 GOCI processing is still in the early phases. The quality of the products are expected to improve as the knowlege of the data matures.

The following improvements are included in this reprocessing:

1. The COMS satellite orbit information was interpreted better. See this poster for more information.
2. A better determination of the timing of portions of the data (slots) was implemented. This improved the determination of the solar geometry, leading to better image quality.

   

   The top 2 panels of the image on the left shows the true color and chlorophyll-a imagery for a GOCI granule at 1600 local time on 10 December, 2013. A single time was used as the dataset time. The bottom 2 panels show the same scene when the improved timing was used. Ocean features appear more natural and the chlor-a retrievals are more numerous with the better timing. The plots on the right show the average Rrs throughout the day for the 8 GOCI acquisition times from 0900 - 1600 local time for an area covered by the box drawn on the image set. The solid line is the average using the old timing and the dashed line is for the improved timing. The Rrs has greater stability through the day due to the improved timing.

3. The GOCI data still has a spectral artefact affecting data quality around clouds. This was mentioned by Fukushima et al (2015). No repairs have been implemented but they are being looked into.
4. There is also an un-corrected radiance signal that occurs in a portion of the GOCI slots which can be seen as a discontinuity in the slots in the imagery. A correction is in the works (Kim et al 2015), but has not been applied to the Level-1B data.

References

Faure, François, Pierre Coste, Gmsil Kang 2008. “The GOCI Instrument on COMS Mission - the first Geostationary Ocean Color Imager.” in Proceedings of the 7th ICSO (International Conference on Space Optics) 2008, Toulouse, France, Oct. 14-17. http://www.ioccg.org/sensors/GOCI-Faure.pdf.

Fukushima, Hajime, K. Ogata, M. Toratani, J. Ahn, W. Kim, Y. Park 2015. "Cloud-affected pixel identification on COMS/GOCI Ocean Color Imagery in Consideration to Fast-moving Cloud Fragments." in Proceedings of the International Symposium on Remote Sensing (ISRS), April, 2015.

Wonkook Kim, Jae-Hyun Ahn, and Young-Je Park, 2015, "Correction of Stray-Light-Driven Interslot Radiometric Discrepancy (ISRD) Present in Radiometric Products of Geostationary Ocean Color Imager (GOCI)." IEEE Transactions on Geoscience and Remote Sensing, 53, (10) DOI: 10.1109/TGRS.2015.2422831 http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7101246