Open Access
Add to BibSonomy
Abstract
Remote sensing reflectance obtained from space-borne ocean color sensors is of great importance to carbon cycle and ocean-atmospheric interactions by providing biogeochemical parameters on the global scale using specific algorithms. Vicarious calibration is necessary for obtaining accurate remote sensing reflectance that meets the application demands of atmospheric correction algorithms. For ocean color sensors, vicarious calibration must be done prior to atmospheric correction. The third Chinese Ocean Color and Temperature Scanner (COCTS) aboard the HY1C satellite was launched on September 7, 2018, and it will provide essential ocean color data that will complement those of existing missions. We used field measurements from the Marine Optical Buoy (MOBY) and aerosol information provided by the MODerate Imaging Spectroradiometer (MODIS) aboard the Terra satellite to calculate vicarious calibration coefficients, and we further evaluated the applicability of the established vicarious calibration approach by cross-calibration using MODIS data on the global scale. Finally, the established vicarious calibration coefficients were used to retrieve the aerosol optical depth and remote sensing reflectance, which were compared to Aerosol Robotic Network-Ocean Color (AERONET-OC) data and MODIS-Terra and Ocean and Land Color Instrument (OLCI)-Sentinel-3A operational products. The results show that the vicarious calibration coefficients are relatively stable and reliable for all bands ranging from visible to near-infrared and can be used to obtain accurate high-quality data.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Ocean color satellite sensors can provide data on global distributions of chlorophyll, particulate organic carbon, primary production and other ocean-related biogeochemical parameters [1–6]. Therefore, ocean color data are of great importance to global carbon cycling, ocean-atmospheric interactions, and climate change [7,8]. However, for ocean color data to be used, it must satisfy strict calibration accuracy requirements [9]. For example, if we wish to obtain a good chlorophyll data, the leaving-water radiance used to retrieve chlorophyll concentrations must have an uncertainty value of less than 5% according to a band ratio algorithm [10]. Further, the radiance at the top of the atmosphere must have an uncertainty value of less than 0.5% [9,11]. However, it is difficult to meet uncertainty requirements strictly via prelaunch and on-orbit calibration mainly due to the influence of the atmosphere and due to effects of sensor degradation after launch. To eliminate errors brought about by atmospheric correction, vicarious calibration must be employed. When this cannot be executed, cross-calibration can serve as an alternative approach [12].
As the central drawback of ocean color satellite sensors, data generated are easily affected by cloud cover and sunlight obstructions, making it difficult to achieve good coverage at the global scale over one or a few days when addressing specific scientific problems. Thus, multiple sensors must be combined to improve data coverage. For example, to demonstrate large-scale ecological processes at the global scale, at least 3 sun-synchronous ocean color satellite sensors must be simultaneously used in space [13].
Currently, on-orbit operational sun-synchronous ocean color satellite missions use 6 sensors, including MODerate Imaging Spectroradiometer (MODIS)-Terra, MODIS-Aqua, Visible Infrared Imaging Radiometer Suite (VIIRS)-National Polar-Orbiting Partnership (NPP), VIIRS-NOAA20, Ocean and Land Color Instrument (OLCI)-Sentinel-3A, and OLCI-Sentinel-3B. MODIS sensors may stop being used at any time due to their long-term and extended service periods relative to the specified lifetime. To maintain or enhance good data coverage, more sun-synchronous ocean color satellite sensors will be needed. China has made an effort to provide an important supplement with the launch of HY-series ocean color satellite sensors since 2002. In 2018, HY1C (the third Hai Yang series satellite) was launched. The Chinese Ocean Color and Temperature Scanner (COCTS) is fitted to the HY1C. As an important supplementary data source, it is necessary to evaluate the stability of COCTS-HY1C. The objective of this paper is to provide an in-depth evaluation of performance of vicarious calibration coefficients using high accuracy in situ data generated by a marine optical buoy (MOBY) positioned adjacent to Lanai Island [14] and well-calibrated MODIS data.
This paper is organized as follows. First, background information on COCTS-HY1C and the vicarious calibration method is provided. We then describe variations in vicarious coefficients observed over days of operation. We then use cross-calibration results of the global scale generated by MODIS to evaluate their applicability. Finally, retrieval and validation results for production based on the established calibration coefficients are given.
2. Background information on COCTS-HY1C
HY1C is a sun-synchronous orbit satellite that was launched on September 7, 2018. Its orbit spans 782 km, its tilting angle is set to 98.522 degrees, and its overpass time is set to 10:30 AM ± 30 min local time. The COCTS sensor aboard the satellite is a moderate resolution imaging scanner with a nadir spatial resolution of 1.1 km and with central bands at 412, 443, 490, 520, 565, 670, 750 and 865 nm, respectively. Unlike MODIS and other sensors, it does not have its own on-orbit calibration system and it has two thermal infrared bands. Corresponding band set and performance parameters are listed in Table 1.
Table 1. Technical parameters for COCTS-HY1C. Measurement parameters are set to the typical input spectral radiance value (mW·cm-2·μm-1·sr-1)
3. Vicarious calibration method
Vicarious calibration involves computing the gain coefficient included in Eq. 1, where the theoretical signal $\textrm{L}_{\textrm{TOA}}^{\textrm{theory}}$ uses sea-truth water-leaving radiance. The gain coefficient is used to multiply the radiance value measured at the top of the atmosphere to have the atmospheric-corrected water-leaving radiance retrieve this true value. A flowchart for this computation method is outlined in a previous work [9,11,15]. The vicarious calibration flowchart used for the present study is shown in Fig. 1. It’s noted that the present method is different from the Refs. [9,11,15], because COCTS-HY1C does not have its own on-orbit calibration system, the radiance at the longest near-infrared band can’t be directly used to determine the aerosol optical depth. We use well-calibrated MODIS-Terra data to collect information on aerosols, including aerosol types and aerosol optical depths (AODs), as the local time at the descending MODIS-Terra intersection is approximately 10:30 AM, and thus the change that occurs due to water and aerosols can be neglected. We further use the viewing geometry and National Centers for Environmental Prediction (NCEP) data to compute the contributions of Rayleigh and aerosol radiance and diffuse transmittance of air molecules and aerosols for sensor viewing path, and transmittance due to gas absorption. Water radiance contributions are determined using in situ data measured by MOBY. It’s noted that the aerosol look-up tables used in this research for COCTS-HY1C is generated in the same way as the operational atmospheric correction process from National Aeronautics and Space Administration (NASA) based on Ahmad et al. [16]. Therefore, in this research we used the MODIS-TERRA to select which aerosol type from the atmospheric correction aerosol look-up tables should be used. Only high-quality MOBY data were used, and the following selection criteria for match-up points between MOBY and satellite data were used. First, a ± 1-hour time window from the satellite overpass time was assigned. Second, when a MOBY measurement fell within the time window, COCTS-HY1C pixels with a 5 × 5 box centered at the position of the MOBY measurement were extracted. When all of the extracted pixels were valid and when the coefficient of variation was less than 0.15 (defined in [17]), the value of the pixel closest to the MOBY location and corresponding MOBY data were used.
After the above physical quantities were determined, we could apply them to obtain the value simulated at the top of the atmosphere and to further derive the vicarious calibration coefficient using Eq. 1. The spectral response function for COCTS-HY1C was applied when using MOBY data. For the NIR bands, as we did not consider water contributions, the same strategy as that shown in Fig. 1 was used without measuring water contributions.
4. Vicarious calibration results for COCTS-HY1C
Figures 2(a)-2(h) show variations in the vicarious coefficient at bands 412, 443, 490, 520, 565, 670, 750 and 865 nm, respectively, for days of operation after launch. Vicarious calibration gains observed over the first 110 days after launch are not computed because the commissioning phase involved preprocessing. As is shown in Fig. 2, 18 points are computed. Gain coefficients at central bands 412, 443, 490, 520, 565, 670, 750 and 865 nm are very stable and oscillate around values of 0.84016, 0.90678, 0.89926, 0.93862, 0.92849, 0.91111 and 0.85642, respectively. The standard deviations are relatively small compared to the amplitudes of the gains, and are 0.02012, 0.02310, 0.02509, 0.02168, 0.02260, 0.02327, 0.02988, and 0.04744 for these bands. Specific values for each vicarious calibration and corresponding statistical parameters are listed in Table 2. Vicarious coefficients derived for COCTS-HY1C do not oscillate around 1 mainly because the vicarious coefficients were derived via prelaunch absolute radiometric calibration. As COCTS-HY1C does not have its own on-orbit calibration system as stated above, the prelaunch absolute radiometric calibration coefficients should decrease due to significant environment changes.
Table 2. Vicarious calibration gain coefficients and statistical parameters
5. Evaluation of the vicarious calibration gain based on MODIS cross-calibration results
To further evaluate applicability at the global scale, we use results (water-leaving radiance and aerosol optical depths and types) obtained from MODIS-Terra as input and compute the radiance at the top of the atmosphere using the viewing geometry taken from COCTS-HY1C. The normalized Lw drawn from MODIS was used and then converted into COCTS geometry via BRDF correction. The locations of the selected images are shown in Fig. 3, red polygons were obtained from COCTS-HY1C, and white ones were obtained from MODIS-Terra. Time periods and cross regions used for cross-calibration are shown in Table 3.
Table 3. Regions and time periods used for cross-calibration
Figures 4(a)-4(h) show the cross-calibration results. To determine the degree of consistency between the cross-calibration and vicarious calibration results, the x-axis represents the theory-computed total radiance measured at the top of the atmosphere using MODIS data, and the y-axis shows radiance values measured at the top of the atmosphere by COCTS multiplied by the vicarious coefficient established in Table 2. As is shown in Fig. 4, for different regions, the radiance value at the top of the atmosphere presents different amplitude values due to different aerosol load and water constituents, but it is always valued around the 1:1 line. These results indicate that vicarious calibration coefficients for the MOBY site are stable and applicable at the global scale. Cross-calibration coefficients established by the cross points for each image and relevant statistical parameters are listed in Table 4. Compared with the vicarious calibration gain coefficients, the relative differences between the average values for all bands are 0.6%, 1.2%, 1.1%, 3.2%, 2.4%, 1.9%, 1.9%, and 0.7%, respectively. So, in the present context of a sensor lacking an on-board calibration system and relatively large uncertainty made by cross-calibration, the differences are acceptable.
Fig. 4. The cross-calibration results derived from MODIS-Terra data. Note that the y-axis presents COCTS radiometry values determined after the application of average vicarious gains listed in Table 2.
Table 4. Cross-calibration coefficients and statistical parameters. The results reflect values determined before the application of vicarious gains
6. Application for atmospheric correction
To further validate the effectiveness of vicarious calibration coefficients, the near-infrared iterative atmospheric correction algorithm [18] was used. The retrieved results were compared with in situ measurements taken from the AERONET-OC site using Version3 Level 1.5 Lwn data. A spectral comparison of four temporal and spatial match-up points positioned at the Venise AERONET-OC site on January 10, 13, and 16, 2019 and on February 4, 2019 is shown in Figs. 5(a)–5(d). Scatterplot comparisons of retrieved remote sensing reflectance (Rrs) values at 443, 490, 520 and 565 nm for January to April, 2019 for specific AERONET-OC sites (Galata_platform, Gloria, Irbe_Lighthouse, Socheongcho, USC_SEAPRISM_2, WaveCIS_Site_CSI_6, Zeebrugge-MOW1 and Venise) are shown in Fig. 6. It can be clearly observed that despite the presence of discreteness mainly due to the change in water bodies influenced by the use of temporal and spatial match-up windows and uncertain measurements and algorithms, the results are reasonable and oscillate approximately 1:1 line. Additionally, the spatial distribution of the ratio of Rrs values retrieved at 443 to 565 nm, which are often used to derive chlorophyll, is outlined in Fig. 7(a) with the corresponding ratio for relevant adjacent MODIS bands shown in Fig. 7(b). It is evident that these data have the same spatial distribution and amplitude. Mesoscale eddies can be effectively depicted.
Fig. 5. Comparison of retrieved and in situ measured Rrs values for the Venise AERONET-OC site for January 10 (a), 13 (b), and 16, 2019 (c) and for February 4, 2019 (d).
Fig. 6. Scatterplot comparisons of COCTS and AERONET-OC Rrs values measured at 443 nm (a), 490 nm (b), 520 nm (c) and 565 nm (d).
Fig. 7. Comparison of distributions of COCTS-HY1C and Terra-MODIS chlorophyll a concentrations for January 13, 2019.
7. Discussion
Another issue concerns the use of the MODIS-Terra satellite sensor for data on aerosols. As this sensor is currently quite old, the MODIS-Terra calibration process is complex. To illustrate the reliability of the vicarious calibration coefficient established above, we use OLCI-Sentinel-3A to apply a cross-check. As is shown in Figs. 8(a)–8(d), we compare Rrs values taken at 412, 443, 490, 520, 565 and 670nm between COCTS and OLCI-Sentinel-3A with a match-up area at 59°W∼56°W, 46°S∼43°S for January 13, 2019. It can be clearly seen that, the scatter points are almost located around 1:1 line, especially for the points with high point density. Of course, we can also see clearly that for the bands at 412 and 443nm, there still exist a systematic overestimation of COCTS. It’s hard to conclude that which is right, because of the different data processing scheme for COCTS and OLCI and this research is to get a quick calibration after post-launch of COCTS-HY1C, without an on-board calibration system, and further analysis and evaluation will be made in future. Overall, the results further confirm the preliminary approach.
Fig. 8. Comparisons of COCTS and OLCI-Sentinel-3A Rrs values measured at 412 nm (a), 443 nm (b), 490 nm (c), 520 nm (d), 565 nm (e), and 670 nm (f) (For OLCI data, the relevant adjacent bands were used) on January 13, 2019.
8. Conclusions
Field measurements drawn from MOBY and aerosol data provided by MODIS-Terra were used to calculate vicarious calibration coefficients, and we further evaluated the applicability of the established vicarious calibration method by cross-calibration using MODIS data of the global scale. Finally, the established vicarious calibration coefficients were used to retrieve the remote sensing reflectance and were compared to AERONET-OC data and the MODIS-Terra and OLCI-Sentinel-3A operational products. The results show that vicarious calibration coefficients are relatively stable and reliable for all bands ranging from visible to near-infrared and can be used to obtain accurate, high-quality data.
However, the vicarious calibration gains obtained through this study were only derived over a short period spanning from January to March 2019. Thus, this work represents an attempt to calibrate COCTS-HY1C data quickly after launch, and the stability of vicarious calibration coefficients occurring over this short period may not have been accounted for. The stability of vicarious calibration coefficients must be evaluated at length in the future in addition to the uncertainty of products used. Further, as COCTS-HY1C does not adopt its own in-orbit calibration system, it is difficult to monitor the temporal degradation of its sensor, and thus, vicarious calibration methods must be updated regularly.
Funding
National Key R&D Program of China (2016YFC1400906); National Natural Science Foundation of China (41431176, 61705211); National Natural Science Foundation of China-Shandong Joint Fund for Marine Science Research Centers (U1406404).
Acknowledgments
We thank Yuebin Yang, Bin Wu, Kaiqi Shi, Jiheng Zhang, Yuxiao Zhang, and Guohui Zhang from our team who worked hard to process the data. We also express our gratitude to the NASA Ocean Biology Processing Group for affording us access to MODIS-Terra data and SeaDAS software, to the Goddard Space Flight Center for the providing AERONET-OC data, to the NOAA for providing NCEP data and to the MOBY team for the maintaining the MOBY system and for providing MOBY data. We gratefully acknowledge Principle Investigators Giuseppe Zibordi, Nima Pahlevan, Burton Jones, Curtiss Davis, Young-Je Park, Dimitry Van der Zande, Bill Gibson, Robert Arnone and Sherwin Ladner from AERONET-OC for maintaining and distributing data for the scientific community. Additional gratitude should be given to Dr. Giuseppe Zibordi from Joint Research Centre of the European Commission and Professor Zhongping Lee from University of Massachusetts, Boston for their valuable comments and discussion. Moreover, valuable comments provided by the three anonymous reviewers are much appreciated.
References
1. C. Hu, Z. Lee, and B. Franz, “Chlorophyll aalgorithms for oligotrophic oceans: A novel approach based on three-band reflectance difference,” J. Geophys. Res.: Oceans 117(C1), C01011 (2012). [CrossRef]
2. M. Stramska, “Particulate organic carbon in the global ocean derived from SeaWiFS ocean color,” Deep Sea Res., Part I 56(9), 1459–1470 (2009). [CrossRef]
3. D. Antoine, J. M. André, and A. Morel, “Oceanic primary production: 2. Estimation at global scale from satellite (coastal zone color scanner) chlorophyll,” Global Biogeochem. Cycles 10(1), 57–69 (1996). [CrossRef]
4. Z. Chen, F. E. Muller-Karger, and C. Hu, “Remote sensing of water clarity in Tampa Bay,” Remote Sens. Environ. 109(2), 249–259 (2007). [CrossRef]
5. Z. Lee, K. L. Carder, and R. A. Arnone, “Deriving inherent optical properties from water color: a multiband quasi-analytical algorithm for optically deep waters,” Appl. Opt. 41(27), 5755–5772 (2002). [CrossRef]
6. M. J. Behrenfeld, E. Boss, D. A. Siegel, and D. M. Shea, “Carbon-based ocean productivity and phytoplankton physiology from space,” Global Biogeochem. Cycles 19(1), GB1006 (2005). [CrossRef]
7. M. J. Behrenfeld, R. T. O’Malley, D. A. Siegel, C. R. McClain, J. L. Sarmiento, G. C. Feldman, and E. Boss, “Climate-driven trends in contemporary ocean productivity,” Nature 444(7120), 752–755 (2006). [CrossRef]
8. W. W. Gregg, M. E. Conkright, P. Ginoux, J. E. O’Reilly, and N. W. Casey, “Ocean primary production and climate: Global decadal changes,” Geophys. Res. Lett. 30(15), 1809 (2003). [CrossRef]
9. H. R. Gordon, “In-orbit calibration strategy for ocean color sensors,” Remote Sens. Environ. 63(3), 265–278 (1998). [CrossRef]
10. H. R. Gordon, “Atmospheric correction of ocean color imagery in the Earth Observing System era,” J. Geophys. Res.: Atmos. 102(D14), 17081–17106 (1997). [CrossRef]
11. B. A. Franz, S. W. Bailey, P. J. Werdell, and C. R. McClain, “Sensor-independent approach to the vicarious calibration of satellite ocean color radiometry,” Appl. Opt. 46(22), 5068–5082 (2007). [CrossRef]
12. E. J. Kwiatkowska, B. A. Franz, G. Meister, C. R. McClain, and X. Xiong, “Cross calibration of ocean-color bands from Moderate Resolution Imaging Spectroradiometer on Terra platform,” Appl. Opt. 47(36), 6796–6810 (2008). [CrossRef]
13. IOCCG Working Group, “Minimum requirements for an operational ocean-colour sensor for the open ocean,” Report of an IOCCG working group held in Villefranche-sur-Mer, France (1998).
14. D. K. Clark, M. A. Yarbrough, M. Feinholz, S. Flora, W. Broenkow, Y. S. Kim, B. C. Johnson, S. W. Brown, M. Yuen, and J. L. Mueller, “MOBY, A Radiometric Buoy for Performance Monitoring and Vicarious Calibration of Satellite Ocean Color Sensors: Measurement and Data Analysis Protocols,” Chapter 2. in Ocean Optics Protocols for Satellite Ocean Color Sensor Validation, Revision 4, Volume VI: Special Topics in Ocean Optics Protocols and Appendices. J.L. Mueller, G.S. Fargion, and C.R. McClain, eds. NASA Goddard Space Flight Center, Greenbelt, MD. NASA/TM2003-211621/Rev4-Vol.VI:3-34, (2003).
15. M. Wang, W. Shi, L. Jiang, and K. Voss, “NIR-and SWIR-based on-orbit vicarious calibrations for satellite ocean color sensors,” Opt. Express 24(18), 20437–20453 (2016). [CrossRef]
16. Z. Ahmad, B. A. Franz, C. R. Mcclain, E. J. Kwiatkowska, J. Werdell, E. P. Shettle, and B. N. Holben, “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,” Appl. Opt. 49(29), 5545–5560 (2010). [CrossRef]
17. S. W. Bailey and P. J. Werdell, “A multi-sensor approach for the on-orbit validation of ocean color satellite data products,” Remote Sens. Environ. 102(1-2), 12–23 (2006). [CrossRef]
18. S. W. Bailey, B. A. Franz, and P. J. Werdell, “Estimation of near-infrared water-leaving reflectance for satellite ocean color data processing,” Opt. Express 18(7), 7521–7527 (2010). [CrossRef]
