@prefix this: . @prefix sub: . @prefix np: . @prefix rdf: . @prefix prov: . @prefix npx: . @prefix dc: . @prefix xsd: . a np:Nanopublication; np:hasAssertion sub:assertion; np:hasProvenance sub:provenance; np:hasPublicationInfo sub:pubinfo; dc:created "2026-07-25T16:48:15.385Z"^^xsd:dateTime; dc:creator ; dc:license ; npx:introduces sub:turbidity-strong-limb-open-source-westerschelde; npx:wasCreatedAt ; "Open-source Sentinel-2 turbidity retrieval in the Westerschelde confirms the Sado strong-turbidity finding"; . sub:turbidity-strong-limb-open-source-westerschelde a ; "2026-07-25"^^xsd:date; "Open-source Sentinel-2 turbidity retrieval in the Westerschelde confirms the Sado strong-turbidity finding"; """This replication confirms the original claim. Sent et al. (2021) found that Sentinel-2 MSI retrieves water-quality parameters with parameter-dependent accuracy — strong for turbidity, weak for chlorophyll-a — and selected a C2RCC + Nechad (cN783) chain that retrieved turbidity in strong agreement with in situ (R² = 0.84). Testing that turbidity limb in an independent, more turbid estuary (the Westerschelde, 2016–2026) with a fully open-source atmospheric-correction chain (Acolite + Nechad et al. 2010, the aN783 chain), we obtained strong agreement with in situ turbidity: R² = 0.92 with a near-unity slope (0.86), meeting and slightly exceeding the paper's proprietary result. This is the strong side of the parameter asymmetry the paper describes, and it replicates even open-source: an operationally reproducible chain that uses no proprietary processor recovers estuarine turbidity at, and slightly beyond, the accuracy of the paper's selected cN783 chain, in a different optical regime."""; ; """From results/turbidity_satellite_acolite.parquet, agreement computed in linear space (scripts/matchup_stats.py), matching the paper's linear treatment of turbidity (turbidity is reported in FNU on both sides — no unit conversion): N = 11 match-ups R² = 0.92 (ours, aN783) vs R² = 0.84 (original, selected cN783) slope = 0.86 (near unity) RMSE = 10.13 FNU BIAS = +6.91 FNU (slight overestimate) The open-source Acolite + Nechad (aN783) chain therefore meets and slightly exceeds the original study's selected C2RCC + Nechad (cN783) R² of 0.84, with a near-unity regression slope, in a different and more turbid estuary. The small positive bias (+6.91 FNU) indicates a mild systematic overestimate but does not degrade the correlation. Turbidity is the strong side of the parameter asymmetry seen across the three retrieved parameters. Match-up figure: figures/asymmetry.png."""; """Four factors bound this conclusion. 1. Sample size. N = 11 turbidity match-ups is small — fewer valid pairs than the Chl-a limb (N = 33), and comparable to the original (N = 19–21). The R² = 0.92 therefore carries a wide confidence interval. The direction of the result (strong agreement, meeting/exceeding the original) is robust; its precise magnitude is not. 2. Processor rating. The paper reports Acolite as its worst-performing atmospheric-correction processor, yet here the open-source Acolite + Nechad chain matches and slightly exceeds the paper's selected C2RCC + Nechad result. That is a notable finding, but a single-site one — it should not be read as a general claim that Acolite outperforms C2RCC for turbidity everywhere. 3. Site and period change. This is a Replication, not a Reproduction: the Sado in situ data are not public (paper Data Availability Statement), so we validated in the Westerschelde (2016–2026), a different and more turbid mesotidal estuary. A different optical regime tests generalisability, not the correctness of the original analysis. Agreement was computed in linear space, matching the paper's treatment of turbidity. 4. Systematic bias. The retrieval carries a slight positive bias (+6.91 FNU), a mild overestimate that does not degrade the correlation but should be noted for any quantitative use. A grid-sensitivity check on a HEALPix / DGGS (GRID4EARTH) representation of the corrected fields is deliberately out of scope here and deferred to a separate follow-up chain that will extend this one, so the primary validation stays on the native Sentinel-2 grid."""; ; ; . sub:assertion prov:wasAttributedTo . "Anne Fouilloux" . sub:sig npx:hasAlgorithm "RSA"; npx:hasPublicKey "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"; npx:hasSignature "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"; npx:hasSignatureTarget ; npx:signedBy .