MODIS-TERRA Ocean Color Reprocessing 2022.0

Introduction

The Ocean Biology Processing Group (OBPG) intiated a full-mission ocean color reprocessing of the MODIS-Terra dataset in November 2022. This reprocessing is part of a multi-mission effort to update the instrument calibrations, vicarious calibration, and product and algorithm changes. Sensor-independent changes are detailed in the R2022.0 Ocean Color Reprocessing General Description. Here we describe the MODIS-Terra sensor-specific details of the reprocessing, and provide an assessment of data quality and impact relative to the previous R2018 MODIS-Terra reprocessing.

Sensor-specific Processing Details

Instrument Calibration

In this reprocessing, the instrument calibration was updated for the VIIRS instrument to utilize the additional on-board (solar/lunar) calibration measurements collected since the R2018.0 reprocessings, and to incorporate advancements in modeling of sensor radiometric degradation from the on-board calibrators.

Vicarious Calibration

MODIS-Terra utilizes a vicarious calibration to the Marine Optical Buoy (MOBY) to minimize residual error in instrument absolute calibration and systematic bias in the atmospheric correction algorithm. As part of the R2018 reprocessing, the MOBY time-series was also reprocessed to incorporate an updated protocol for calculating the sea-air transmission coefficient that is used to calculate water leaving radiances (Lw) from MOBY measurements.

Historical data processing protocols recommended assuming a constant transmission coefficient for calculating water leaving radiances (Lw) from underwater measurements of upwelling radiance (Lu0-). The constant value 0.543 had been an accepted simplification, calculated using inputs of 500 nm, 25 C, and 35 ppt. However, using a fixed value regardless of wavelength resulted in a small spectral bias across Lw spectra, causing MOBY data to be approximately 1.3% too high at 380 nm, decreasing inversely with wavelength to be approximately 1.1% too low at 700 nm. Note that this was a general data processing issue, not a problem with MOBY.

This issue and its specific impacts on MOBY data were described by Voss and Flora (2017), and the refined advice for calculating the sea-air transmission coefficient was also included in the most recent update to community-based AOP protocols (Zibordi et al. 2019). Based on those recommendations, the MOBY time-series was reprocessed. The R2018 reprocessing included an updated sea-air transmission coefficient calculation that incorporated wavelength-specific inputs as well as temperature measurements to remove the described bias and improve the accuracy of satellite retrievals through the updated vicarious calibration.

Product and Algorithm Changes

Changes in products and algorithms have been made for all the missions including Terra in this reprocessing, including: chlorophyll, particulate inorganic carbon (PIC), raman scattering, absorption and backscattering spectra for water, ancillary data sources (for ozone, sea surface temperature, land/water mask and bathymetery), multi-scattering Epsilon aerosol model and optical properties determination.,

Impact and Quality Assessment

Impact of Reprocessing on Timeseries

To assess the impact of this R2022.0 reprocessing, a comparative timeseries analysis was perfomed. The impact of all calibration, product and algorithm, and ancillary data changes on chlorophyll and remote sensing reflectance of global deep water (>1000m) is shown in Fig. 1 and Fig. 2. Chlorophyll and Rrs time series generated for MODIS-Terra and MODIS-Aqua using the R2022 configuration is shown in Fig. 1 (left panels). Their ratios are shown on the right panels. The comparison of results between R2022 and R2018 reprocessing are shown in Fig. 2. Overall, the differences in Chlorophyll and remote sensing reflectance between MODIS-Terra and MODIS-Aqua R2022 are within 10%, and these products from these two most recent reprocessings are consistent.

Deep water time series comparisons of MODIS-Terra R2022 to MODIS-Aqua R2022 — Figure 1: Global deepwater time series comparisons of MODIS-Terra R2022 to MODIS_Aqua R2022

Deep water time series comparisons of MODIS-Terra R2022 to MODIS-Terra R2018 — Figure 2: Global deepwater time series comparisons of MODIS-Terra R2022 to MODIS_Terra R2018

Comparison with In Situ Measurements

Validation of the chlorophyll, apparent optical properties (Rrs, Kd490), inherent optical properties (a_giop, aph_giop, adg_giop, bbp_giop), particulate organic/inorganic carbon (POC/PIC), and photosynthetically available radiation (PAR) retrievals was performed relative to all available match-ups from SeaBASS and the Aerosol Robotic Network - Ocean Color (AERONET-OC). Statistical analysis, scatter plots and frequency distribution comparisons of the satellite to in situ match-ups are provided below.

Chlorophyll a (chlor_a)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	MODIS-Terra Range	In situ Range
chlor_a	2267	1.07929^*	1.71417^*	0.02212, 72.85196	0.00900, 151.30000

Apparent Optical Properties

Remote Sensing Reflectance (Rrs)

Product Name	#	Mean Bias	Mean Absolute Error (MAE)	MODIS-Terra Range	In situ Range
Rrs412	6183	-0.00031	0.00116	-0.00516, 0.03121	-0.00000, 0.03927
Rrs443	6507	-0.00016	0.00088	-0.00443, 0.03954	0.00007, 0.04488
Rrs488	6127	-0.00057	0.00088	-0.00005, 0.04859	0.00039, 0.05464
Rrs531	3725	-0.00060	0.00086	0.00083, 0.04133	0.00113, 0.04763
Rrs547	5606	-0.00064	0.00091	0.00088, 0.03752	0.00080, 0.04210
Rrs555	5650	-0.00085	0.00102	0.00064, 0.03339	0.00050, 0.04211
Rrs667	5746	-0.00020	0.00036	-0.00069, 0.01970	-0.00000, 0.02154
Rrs678	516	-0.00018	0.00037	-0.00053, 0.01046	0.00004, 0.00965

Diffuse Attenuation Coefficients (kd490)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	MODIS-Terra Range	In situ Range
kd490	626	1.06968^*	1.29197^*	0.02180, 6.40000	0.01383, 1.59180

Inherent Optical Properties

Total Absorption Coefficients (a_giop)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	Aqua Range	In situ Range
a_giop412	199	0.80596^*	1.56804^*	0.01565, 1.57042	0.01492, 1.48437
a_giop443	199	0.78931^*	1.41989^*	0.01599, 1.23717	0.01553, 1.12670
a_giop488	199	0.78390^*	1.34566^*	0.01961, 0.78620	0.01956, 0.60427
a_giop531	199	0.84811^*	1.23294^*	0.04024, 0.48792	0.04553, 0.37732
a_giop547	199	0.85589^*	1.21759^*	0.02744, 0.42236	0.05789, 0.29862
a_giop555	199	0.89161^*	1.16857^*	0.03510, 0.38199	0.06102, 0.27676
a_giop667	199	0.99748^*	1.04826^*	0.16395, 0.96838	0.43053, 0.79128

Phytoplankton Absorption Coefficients (aph_giop)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	Aqua Range	In situ Range
aph_giop412	209	0.82185^*	1.66731^*	0.00233, 0.82421	0.00141, 0.47654
aph_giop443	211	0.85452^*	1.59815^*	0.00311, 0.93034	0.00193, 0.54002
aph_giop488	211	0.91611^*	1.55841^*	0.00221, 0.63829	0.00102, 0.28555
aph_giop531	210	0.90931^*	1.69064^*	0.00050, 0.36753	0.00016, 0.15537
aph_giop547	206	1.03440^*	1.76493^*	0.00030, 0.30275	0.00006, 0.11235
aph_giop555	205	0.99191^*	1.77567^*	0.00020, 0.26505	0.00004, 0.09717
aph_giop667	209	1.13731^*	1.74573^*	0.00080, 0.52916	0.00038, 0.33381

Absorption Coefficient of Non-algal Material plus CDOM (adg_giop)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	Aqua Range	In situ Range
adg_giop412	197	0.72827^*	1.96579^*	0.00698, 1.72101	0.00674, 1.18787
adg_giop443	197	0.65554^*	2.08122^*	0.00376, 0.98503	0.00418, 0.71333
adg_giop488	197	0.55519^*	2.32123^*	0.00167, 0.43820	0.00218, 0.34561
adg_giop531	197	0.44876^*	2.69219^*	0.00082, 0.20208	0.00126, 0.20564
adg_giop547	197	0.44269^*	2.78596^*	0.00056, 0.15152	0.00094, 0.16032
adg_giop555	197	0.40969^*	2.92806^*	0.00050, 0.13119	0.00085, 0.15071
adg_giop667	197	0.22572^*	4.82567^*	0.00009, 0.01745	0.00011, 0.04039

backscattering coefficient of particles (bbp_giop)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	MODIS-Terra Range	In situ Range
bbp_giop412	159	0.87578^*	1.40350^*	0.00093, 0.00960	0.00063, 0.01606
bbp_giop443	159	0.87632^*	1.39425^*	0.00085, 0.00886	0.00056, 0.01555
bbp_giop488	159	0.87343^*	1.38610^*	0.00075, 0.00797	0.00049, 0.01489
bbp_giop531	159	0.86715^*	1.38735^*	0.00067, 0.00742	0.00044, 0.01438
bbp_giop547	159	0.87239^*	1.38679^*	0.00064, 0.00733	0.00042, 0.01415
bbp_giop555	159	0.86794^*	1.38912^*	0.00063, 0.00729	0.00041, 0.01410
bbp_giop667	159	0.85862^*	1.41248^*	0.00046, 0.00676	0.00031, 0.01303
bbp_giop678	3	0.78852^*	1.26820^*	0.00058, 0.00368	0.00076, 0.00480

Photosynthetically Available Radiation (par)

Product Name	#	Mean Bias	Mean Absolute Error (MAE)	Aqua Range	In situ Range
par	362	5.53442	5.53936	21.72650, 64.50283	8.89095, 61.75359

Particulate inorganic Carbon (pic)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	Aqua Range	In situ Range
pic	70	0.20271^*	5.48211^*	0.00001, 0.00318	0.00002, 0.02753

Particulate Organic Carbon (poc)

^* statistical calculations based on log10 (implies ignoring values equal to or less than zero)
Product Name	#	Mean Bias	Mean Absolute Error (MAE)	MODIS-Terra Range	In situ Range
poc	310	1.04007^*	1.41515^*	20.72490, 3464.74996	21.30000, 796.91229

References

Ahmad, Ziauddin, Bryan A. Franz, Charles R. McClain, Ewa J. Kwiatkowska, Jeremy Werdell, Eric P. Shettle, and Brent N. Holben. 2010. “New Aerosol Models for the Retrieval of Aerosol Optical Thickness and Normalized Water-Leaving Radiances from the SeaWiFS and MODIS Sensors over Coastal Regions and Open Oceans.” Applied Optics 49 (29): 5545. https://doi.org/10.1364/AO.49.005545.

Ahmad, Ziauddin, and Bryan A. Franz (2018), Uncertainty in aerosol model characterization and its impact on ocean color retrievals, in PACE Technical Report Series, Volume 6: Data Product Requirements and Error Budgets (NASA/TM-2018 – 2018-219027/ Vol. 6), edited by I. Cetinić, C. R. McClain and P. J. Werdell, NASA Goddard Space Flight Space Center Greenbelt, MD.

Balch, W. M., et al. (2005). "Calcium carbonate budgets in the surface global ocean based on MODIS data." Journal of Geophysical Research Oceans 110(C7): C07001, doi:07010.01029/02004JC002560.

GEBCO Compilation Group (2020) GEBCO 2020 Grid (doi:10.5285/a29c5465-b138-234d-e053-6c86abc040b9)

Gordon, H. R., et al. (2001). "Retrieval of coccolithophore calcite concentration from SeaWiFS imagery." Geochemical Research Letters 28(8): 1587-1590.

Hu, C., Feng, L., Lee, Z., Franz, B. A., Bailey, S. W., Werdell, P. J., & Proctor, C. W. (2019). Improving satellite global chlorophyll a data products through algorithm refinement and data recovery. Journal of Geophysical Research: Oceans, 124.

McKinna, L.I.W., Werdell, P.J. and C.W. Proctor. 2016. "Implementation of an analytical Raman scattering correction for satellite ocean-color processing." Optics Express, 24(14), A1123-A1137, doi: 10.1364/OE.24.0A1123

O'Reilly, J.E., and Werdell, P. J. (2019). Chlorophyll algorithms for ocean color sensors - OC4, OC5 & OC6, Remote Sensing of Environment 229: 32-47, https://doi.org/10.1016/j.rse.2019.04.021.

Voss, K. J., & Flora, S. (2017). Spectral dependence of the seawater-air radiance transmission coefficient. Journal of Atmospheric and Oceanic Technology, 34(6), 1203-1205. https://doi.org/10.1175/JTECH-D-17-0040.1

Westberry et al. (2013) Influence if Raman scattering on ocean color inversion models, Applied Optics, 52(22), 5552-5561, https://doi.org/10.1364/AO.52.005552

Zibordi, G., Voss, K. J., Johnson, B. C., & Mueller, J. L. (2019). Protocols for Satellite Ocean Colour Data Validation: In Situ Optical Radiometry. IOCCG Ocean Optics and Biogeochemistry Protocols for Satellite Ocean Colour Sensor Validation, Volume 3.0, IOCCG, Dartmouth, NS, Canada. http://dx.doi.org/10.25607/OBP-691