In situ validation of MERIS marine reflectance off the southwest Iberian Peninsula: assessment of vicarious adjustment and corrections for near-land adjacency

Figures & data

Figure 1. Satellite image of MERIS algal pigment index 1 (API 1) for the southwest coast of Portugal with the inset showing locations of Stations A (37° 00′ 39″ N and 8° 53′ 58″ W), B (36° 56′ 06″ N and 8° 52′ 48″ W), and C (36° 51′ 33″ N and 8° 50′ 16″ W) off the peninsula at Sagres.

Figure 2. Gain spectra at 95% confidence limits for MERIS third reprocessing used for VIC (adapted from Lerebourg et al. (2011)).

Table 1. Statistics for the bio-optical properties at Stations A, B, and C. Spectral diffuse attenuation coefficient at 490 nm (K_d(490)); algal pigment index 1 – total chlorophyll a and its degradation products (API 1); total suspended matter (TSM); absorption coefficient at 443 nm for phytoplankton (a_ph); absorption coefficient at 443 nm for non-algal particles (a_nap); and absorption coefficient at 443 nm for yellow substance (a_ys).

		Station A	Station B	Station C
Depth	Average ± SD	41.1 ± 7.0	104.4 ± 3.3	160.3 ± 40.5
(m)	Min. – Max.	32.3–57.0	96.0–109.0	103.0–242.0
Transparency	Average ± SD	11.0 ± 2.8	14.6 ± 5.7	16.5 ± 6.1
(m^−1)	Min. – Max.	7.0–16.5	7.0–29.0	7.0–31.0
Kd(490)	Average ± SD	0.13 ± 0.05	0.09 ± 0.05	0.08 ± 0.04
(m^−1)	Min. – Max.	0.06–0.25	0.04–0.22	0.03–0.18
API 1	Average ± SD	0.91 ± 0.63	0.71 ± 0.54	0.50 ± 0.34
(mg m^−3)	Min. – Max.	0.10–2.27	0.09–2.10	0.08–1.17
TSM	Average ± SD	1.87 ± 0.61	1.79 ± 0.61	1.76 ± 0.63
(ug.^–^l)	Min. – Max.	0.80–3.00	0.70–2.90	0.45–2.80
aph (443)	Average ± SD	0.06 ± 0.03	0.05 ± 0.04	0.04 ± 0.03
(m^−1)	Min. – Max.	0.01–0.12	0.01–0.15	0.01–0.13
	Total % (Average ± SD)	48.4 ± 16.5	45.4 ± 20.3	46.8 ± 21.5
anap (443)	Average ± SD	0.01 ± 0.01	0.01 ± 0.01	0.00 ± 0.00
(m^−1)	Min. – Max.	0.00–0.03	0.00–0.04	0.00–0.02
	Total % (Average ± SD)	8.8 ± 6.6	8.8 ± 9.5	6.4 ± 7.4
ays (443)	Average ± SD	0.05 ± 0.03	0.06 ± 0.04	0.04 ± 0.03
(m^−1)	Min. – Max.	0.01–0.12	0.01–0.16	0.00–0.11
	Total % (Average ± SD)	42.8 ± 18.4	45.8 ± 21.1	46.8 ± 22.4

Figure 3. Ternary diagram for Stations A, B, and C showing the relative contribution of absorption coefficients at 443 nm for the absorption by phytoplankton (a_ph), the absorption by non-algal particles (a_nap), and the absorption by yellow substance (a_ys). Each measurement is represented by (+).

Table 2. Summary of the uncertainty budget for water-leaving reflectance (ρ_w) determined for the TACCS at Station A (SA), Station B (SB), and Station C (SC) for wavelengths 443, 490, and 560 nm.

	443 nm			490 nm			560 nm
Uncertainty source	SA	SB	SC	SA	SB	SC	SA	SB	SC
Absolute calibration of Lu(λ)	2.8	2.8	2.8	2.8	2.8	2.8	2.8	2.8	2.8
Self-shading corrections	0.7	0.7	0.7	0.7	0.7	0.7	0.7	0.7	0.7
Absolute calibration of Ed(λ)	3.1	3.1	3.1	3.1	3.1	3.1	3.1	3.1	3.1
Bio-optical assumptions	2.2	2.2	2.2	1.2	1.2	1.2	2.0	2.0	2.0
Geometrical effects	4.5	4.5	4.5	4.5	4.5	4.5	4.0	4.0	4.0
Environmental perturbations	2.8	3.0	3.8	3.1	3.2	5.0	4.0	3.5	4.6
Quadrature sum (%)	7.1	7.2	7.6	7.0	7.1	8.1	7.4	7.1	7.7

Table 3. Number of matchups (N) with potential for MERIS validation in Sagres (Initial). Number of matchups after processing procedures, without vicarious adjustment (NoVIC); with vicarious adjustment (VIC); with ICOL processing (ICOL); with vicarious adjustment and ICOL processing (VIC+ ICOL) at Station A (SA), Station B (SB), and Station C (SC).

N	SA	SB	SC
Initial	134	131	128
NoVIC	90 (67%)	108 (82%)	115 (90%)
VIC	103 (77%)	97 (74%)	115 (90%)
ICOL	51 (38%)	108 (82%)	113 (88%)
VIC+ ICOL	85 (63%)	113 (86%)	115 (90%)

Figure 4. Scatter plots of MERIS water-leaving reflectance (ρ_w) versus in situ ρ_w at 443, 490, and 560 nm for Stations A, B, and C. The 1:1 relationship is represented by the solid diagonal line, whereas the linear regression for the data set is represented by the dashed lines. For each matchup point, the vertical bar indicates one standard deviation within the 3 × 3 pixel box used for the matchup and the horizontal bar represents the uncertainty budget for in situ ρ_w. Colours represent the different processing combinations with NoVIC (red), VIC (black), ICOL (green), and VIC + ICOL (blue).

Figure 5. Histograms showing percentage differences for water-leaving reflectance (ρ_w) and curves for normal fit at 443, 490, and 560 nm for Stations A, B, and C. Colours represent the different processing combinations with NoVIC (red), VIC (black), ICOL (green), and VIC + ICOL (blue).

Table 4. Matchup statistics for MERIS water-leaving reflectance (ρ_w) for Stations A, B, and C with NoVIC.

	Station A (N = 90)						Station B (N = 108)						Station C (N = 115)
λ (nm)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)
412	0.177	0.007	0.540	1.166	16.6	27.5	0.705	0.003	0.925	1.203	20.3	22.7	0.714	0.005	0.854	1.242	24.2	26.3
443	0.441	0.004	0.662	0.998	−0.2	20.5	0.816	0.002	0.936	1.084	8.4	13.5	0.835	0.002	0.927	1.133	13.3	16.7
490	0.690	0.002	0.757	0.895	−10.5	16.4	0.800	0.002	0.819	0.972	−2.8	10.0	0.806	0.002	0.877	1.037	3.7	9.4
510	0.691	0.002	0.709	0.866	−13.4	17.5	0.721	0.003	0.661	0.950	−5.0	10.0	0.675	0.002	0.811	1.014	1.4	9.0
560	0.586	0.002	0.588	0.826	−17.4	20.6	0.660	0.003	0.506	0.899	−10.1	11.6	0.811	0.000	0.904	0.912	−8.8	14.4
665	0.224	0.000	0.584	0.162	−83.8	84.4	0.358	0.000	0.498	0.458	−54.2	54.2	0.482	0.000	0.823	0.730	−27.0	36.5

Table 5. Matchup statistics for MERIS water-leaving reflectance (ρ_w) for Stations A, B, and C with VIC.

	Station A (N = 103)						Station B (N = 97)						Station C (N = 115)
λ (nm)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)
412	0.01	0.008	−0.101	1.359	35.9	47.9	0.422	0.000	0.687	0.699	−30.1	32.5	0.770	0.000	0.755	0.772	−22.8	24.7
443	0.118	0.006	0.226	1.204	20.4	31.9	0.586	0.000	0.746	0.782	−21.8	24.1	0.853	0.000	0.820	0.847	−15.3	17.6
490	0.644	0.004	0.532	1.067	6.7	17.2	0.518	0.004	0.546	0.857	−14.3	16.8	0.795	0.002	0.754	0.916	−8.4	11.8
510	0.667	0.003	0.529	1.041	4.1	16.2	0.362	0.006	0.322	0.838	−16.2	17.3	0.590	0.003	0.574	0.905	−9.5	12.8
560	0.605	0.003	0.557	0.990	−1.0	14.6	0.373	0.004	0.339	0.826	−17.4	18.4	0.742	0.001	0.779	0.912	−8.8	14.4
665	0.094	0.000	0.379	0.951	−4.9	37.6	0.185	0.000	0.484	0.381	−61.9	62.9	0.321	0.000	0.663	0.667	−33.3	41.3

Table 6. Matchup statistics for MERIS water-leaving reflectance (ρ_w) for Stations A, B, and C with ICOL.

	Station A (N = 51)						Station B (N = 108)						Station C (N = 113)
λ (nm)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)
412	0.121	0.007	0.503	1.283	28.3	31.8	0.696	0.004	0.851	1.185	18.5	20.9	0.797	0.004	0.912	1.207	20.7	23.1
443	0.336	0.004	0.638	1.125	12.5	18.8	0.796	0.002	0.886	1.079	7.9	13.0	0.857	0.002	0.955	1.106	10.6	14.5
490	0.612	0.004	0.673	1.003	0.3	9.0	0.756	0.002	0.780	0.979	–2.1	10.2	0.793	0.002	0.871	1.025	2.5	9.0
510	0.497	0.005	0.544	0.970	–3.0	10.4	0.660	0.004	0.617	0.963	–3.7	10.0	0.664	0.002	0.806	1.006	0.6	9.0
560	0.476	0.005	0.335	0.917	–8.3	13.4	0.611	0.003	0.495	0.918	–8.2	11.3	0.798	0.000	0.926	0.666	–33.4	11.6
665	0.216	0.001	0.300	0.758	–24.2	29.1	0.372	0.000	0.498	0.645	–35.5	38.4	0.483	0.000	0.855	0.754	–24.6	35.0

Table 7. Matchup statistics for MERIS water-leaving reflectance (ρ_w) for Stations A, B, and C with VIC + ICOL.

	Station A (N = 85)						Station B (N = 113)						Station C (N = 115)
λ (nm)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)	R^2	Intercept	Slope	MR	RPD (%)	APD (%)
412	0.362	0.003	0.448	0.785	–20.8	23.8	0.459	0.001	0.640	0.727	–27.3	30.2	0.785	–0.001	0.813	0.734	–26.6	28.1
443	0.620	0.002	0.611	0.864	–13.0	18.4	0.640	0.001	0.724	0.809	–19.1	23.2	0.844	0.000	0.857	0.819	–18.1	19.9
490	0.757	0.002	0.744	0.926	–6.8	13.6	0.631	0.003	0.638	0.876	–12.4	15.1	0.779	0.001	0.783	0.902	–9.8	12.6
510	0.706	0.002	0.711	0.927	–6.7	13.0	0.523	0.004	0.459	0.861	–13.9	15.4	0.567	0.003	0.604	0.898	–10.2	12.9
560	0.593	0.003	0.585	0.918	–7.6	14.2	0.433	0.003	0.389	0.850	–15.0	15.9	0.738	0.001	0.816	0.907	–9.3	14.7
665	0.134	0.001	0.236	0.821	–17.0	24.0	0.247	0.000	0.544	0.571	–42.9	44.6	0.328	0.000	0.715	0.696	–30.4	38.2

Figure 6. Transects of water-leaving reflectance (ρ_w) at 490 and 560 nm for 17 November 2008, 27 May 2009, 18 March 2011, and 24 November 2011 extending perpendicular from the coast up to approximately 24 km offshore. Each trace has undergone different processing combinations with NoVIC in red, VIC in black, ICOL in green, and VIC + ICOL in blue. The dashed vertical lines indicate the approximate locations of Stations A, B, and C from the coast with the value for in situ ρ_w shown as full black circles.

Figure 7. Water-leaving reflectance (ρ_w) at 560 nm from MERIS full-resolution satellite images from 17 November 2008, 27 May 2009, 18 March 2011, and 24 November 2011 at similar coordinates for the stations shown in Figure 1. Each date shows images with different processing combinations NoVIC, VIC, ICOL, and VIC + ICOL.

Table 8. Comparison of published statistical data for MERIS matchups with those at Sagres for wavelengths 443, 490, and 560 nm, where N = number of matchups, R² = coefficient of determination, MR = mean ratio, RPD = relative percentage difference, and APD = absolute percentage difference.

	AAOT(Zibordi, Mélin, and Berthon 2006)			BOUSSOLE(Antoine, Guevel, et al. 2008)			Bohai Sea(Cui et al. 2010)			SagresStation A			SagresStation B			SagresStation C
	443	490	560	443	490	560	443	490	560	443	490	560	443	490	560	443	490	560
N	41	41	41	61	64	63	13	13	13	103	103	103	97	97	97	115	115	115
R^2	0.79	0.92	0.95	0.38	0.44	0.34	0.55	0.41	-0.01	0.12	0.64	0.61	0.59	0.52	0.37	0.85	0.80	0.74
MR	–	–	–	1.32	1.16	1.21	–	–	–	1.20	1.07	0.99	0.78	0.86	0.83	0.85	0.92	0.91
RPD (%)	42	21	15	31.6	15.8	21.3	–10	1	5	20.4	6.7	–1.0	–21.8	–14.3	–17.4	–15.3	–8.4	–8.8
APD (%)	43	22	18	35.8	18.6	25.3	18	15	16	31.9	17.2	14.6	24.1	16.8	18.4	17.6	11.8	14.4

Table 9. Matchup statistics for MERIS water-leaving reflectance (ρ_w) comparing BOUSSOLE (Antoine, Guevel, et al. 2008) with Station C at Sagres. The sample numbers (N) for BOUSSOLE are 20, 61, 64, 63, 64, and 64 for the wavelengths 412, 443, 490, 510, 560, and 665 nm, respectively. At Sagres, N = 115 for all the wavelengths listed in the Table.

	BOUSSOLE					Station C
λ (nm)	R^2	Intercept	Slope	RPD (%)	APD (%)	R^2	Intercept	Slope	RPD (%)	APD (%)
412	0.43	0.009	0.93	60.2	62.7	0.71	0.005	0.854	24.2	26.3
443	0.38	0.008	0.70	31.6	35.8	0.83	0.002	0.927	13.3	16.7
490	0.44	0.006	0.69	15.8	18.6	0.81	0.002	0.877	3.7	9.4
510	0.24	0.007	0.52	21.5	23.5	0.67	0.002	0.811	1.4	9
560	0.34	0.003	0.64	21.3	25.3	0.81	0.000	0.904	−8.8	14.4
665	0.16	0.001	0.89	59	69.1	0.48	0.000	0.823	−27	36.5

Table 10. Comparison of MERIS matchups from AAOT site (*Zibordi et al. 2013; ^†Kajiyama, D’Alimonte, and Zibordi 2014) with Sagres ocean Station C with MERIS water-leaving reflectance (ρ_w) after NoVIC in bold italics and after VIC in normal text for wavelengths 413, 443, 490, 560, and 665 nm, where N = number of matchups, R² = coefficient of determination, RPD = relative percentage difference, and APD = absolute percentage difference.

	AAOT (N = 160)* AAOT (N = 86)*						AAOT (N = 100)† Sagres (N = 115)						Sagres (N = 115)
λ (nm)	R^2	RPD (%)	APD (%)	R^2	RPD (%)	APD (%)	R^2	RPD (%)	APD (%)	R^2	RPD (%)	APD (%)	R^2	RPD (%)	APD (%)
412	0.19	55	69	0.43	–40	48	0.39	–43	44	0.71	24.2	26.3	0.77	–23	25
443	0.46	17	36	0.71	–23	29	0.67	–26	26	0.83	13.3	16.7	0.85	–15	18
490	0.80	11	17	0.88	–8	14	0.81	–12	14	0.81	3.7	9.4	0.80	–8	12
560	0.90	7	15	0.91	–2	12	0.88	–8	12	0.81	–8.8	14.4	0.74	–9	14
665	0.82	–4	44	0.80	–16	33	0.77	28	36	0.48	–27	36.5	0.32	–23	41

McClain, C. R., J. R. Christian, S. R. Signorini, M. R. Lewis, I. Asanuma, D. Turk, and C. Dupouy-Douchement. 2002. “Satellite Ocean-Color Observations of the Tropical Pacific Ocean.” Deep Sea Research Part II: Topical Studies in Oceanography 49: 2533–2560. doi:10.1016/S0967-0645(02)00047-4

 Web of Science ®Google Scholar

Zibordi, G., F. Mélin, and J.-F. Berthon. 2006. “Comparison of SeaWiFS, MODIS and MERIS Radiometric Products at a Coastal Site.” Geophysical Research Letters 33: L06617. doi:10.1029/2006GL025778.

 Web of Science ®Google Scholar

Antoine, D., F. d’Ortenzio, S. B. Hooker, G. Bécu, B. Gentili, D. Tailliez, and A. J. Scott. 2008. “Assessment of Uncertainty in the Ocean Reflectance Determined by Three Satellite Ocean Color Sensors (MERIS, SeaWiFS and MODIS-A) at an Offshore Site in the Mediterranean Sea (BOUSSOLE Project).” Journal of Geophysical Research 113: C07013. doi:10.1029/2007JC004472.

 Web of Science ®Google Scholar

Cui, T., J. Zhang, S. Groom, L. Sun, T. Smyth, and S. Sathyendranath. 2010. “Validation of MERIS Ocean-Color Products in the Bohai Sea: A Case Study for Turbid Coastal Waters.” Remote Sensing of Environment 114: 2326–2336. doi:10.1016/j.rse.2010.05.009.

 Web of Science ®Google Scholar

Zibordi, G., F. Mélin, J.-F. Berthon, and E. Canuti. 2013. “Assessment of MERIS Ocean Color Data Products for European Seas.” Ocean Science 9: 521–533. doi:10.5194/os-9-521-2013.

 Web of Science ®Google Scholar

Kratzer, S., C. Brockmann, and G. Moore. 2008. “Using MERIS Full Resolution Data to Monitor Coastal Waters – A Case Study from Himmerfjärden, a Fjord-Like Bay in the Northwestern Baltic Sea.” Remote Sensing of Environment 112: 2284–2300. doi:10.1016/j.rse.2007.10.006.

 Web of Science ®Google Scholar

Kajiyama, T., D. D’Alimonte, and G. Zibordi. 2014. “Match-Up Analysis of MERIS Radiometric Data in the Northern Adriatic Sea.” IEEE Geoscience and Remote Sensing Letters 11 (1): 19–23. doi:10.1109/LGRS.2013.2244844.

 Web of Science ®Google Scholar