Corrigendum: Optical water characterization and atmospheric correction assessment of estuarine and coastal waters around the AERONET-OC Bahia Blanca

A Corrigendum on
Optical water characterization and atmospheric correction assessment of estuarine and coastal waters around the AERONET-OC Bahia Blanca

by Arena M, Pratolongo P, Loisel H, Tran MD, Jorge DSF and Delgado AL (2024). Front. Remote Sens. 5:1305787. doi: 10.3389/frsen.2024.1305787

In the published article, there was an error in Figure 5 as published. The conversion to L_wn(λ) of the “ACOLITE-DSF” products was miscalculated due to an additional division by a factor of pi. Consequently, scatterplots of the “ACOLITE-DSF” match-up exercise exhibited significantly underestimated values with errors of large magnitude. The corrected Figure 5 and its caption appear below.

Figure 5

Figure 5. Scatterplots of the match-up exercise of the atmospheric correction algorithm (AC) retrieval and the in situ L_wn (443, 490, 510, 560 nm). Colors indicate the OWT: OWT 3 in green, OWT 4 in yellow.

A correction has been made to 3. Results, 3.3 Validation of OLCI (Sentinel 3A-B) L_wn(λ), paragraph 2. This paragraph previously stated:

“At 412 nm, the “Standard NN” presented the lowest RSME (0.29 mW cm² sr µm)⁻¹ and NMAE (12%) (Supplementary Figure S2, Supplementary Material). Similar results were obtained with the l2gen processor whose performance was relatively intermediate (RSME: 0.51 and 0.73, and NMDE: 24 and 38 at 412 nm for “SeaDAS” and “SeaDAS-ALT”, respectively). Results from “SeaDAS” and “SeaDAS-ALT” were similar, with a slight tendency to overestimation in both cases. In the case of blue-green bands, overestimation resulted in large biases, NMDEs and offsets (Figure 5). POLYMER and ACOLITE-DSF consistently underestimated L_wn (λ) with negative biases and lower coefficients of determination. It has to be noted that POLYMER performance improved at 510, 560, and 665 nm, but large RSMEs and NMAEs, as well as larger dispersion were observed when compared to the operational algorithm. In the study area, ACOLITE-DSF produced large underestimations with errors of large magnitude in the visible bands (Figure 5).”

The corrected paragraph appears below:

“At 412 nm, the “Standard NN” presented the lowest RSME (0.29 mW cm² sr µm)⁻¹ and NMAE (12%) (Supplementary Figure S2, Supplementary Material). Similar results were obtained with the l2gen processor whose performance was relatively intermediate (RSME: 0.51 and 0.73, and NMDE: 24 and 38 at 412 nm for “SeaDAS” and “SeaDAS-ALT”, respectively). Results from “SeaDAS”, “SeaDAS-ALT” and “ACOLITE-DSF” were similar, with a slight tendency to overestimation in all cases. In the case of blue-green bands, overestimation resulted in large biases, NMDEs and offsets (Figure 5). POLYMER consistently underestimated Lwn (λ) with negative biases and lower coefficients of determination. It has to be noted that POLYMER performance improved at 510, 560, and 665 nm, but large RSMEs and NMAEs, as well as larger dispersion were observed when compared to the operational algorithm (Figure 5).”

The authors apologize for these errors and state that these do not change the scientific conclusions of the article in any way. The original article has been updated.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.