General comment:

The North Atlantic Ocean is an important CO2 sink (Takahashi et al. 2009) and this region contains high concentrations of anthropogenic CO2 in the water column (Khatiwala et al., 2013; Steinfeldt et al, 2024). Thanks to numerous observations in this basin since the seventies, it is now well known that simulated carbonates systems properties, including CO2 fluxes, OA and Cant, present significant bias. Quoting Perez et al., 2024: “The largest disagreement in the CO2 flux between GOBMs and pCO2 products is found north of 50°N”. Bias between models and data-based estimates are also observed for the Cant inventories in the North Atlantic (Perez et al, 2024, their figure 7). This calls for new analysis based on series of cruises to investigate the Cant variability, from seasonal to multi-decadal, such as recently presented by Bajon et al (2025). In this context, it is also important to extend and/or revise the data when observational bias are suggested or identified (e.g. Wang et al , 2025 for oxygen). Here the authors suggest that a correction of their pH data obtained along the OVIDE-BOCATS sections should be applied (by 0.011 on average). As an example, the comparison of data presented in the NEADW is convincing. Although this has apparently no impact on the pH trends, they show that the correction leads to a shift for the ASD. This new dataset (including 2021 and 2023 cruises) represents an important product, not only to re-investigate the drivers of ASD (Lauvset et al, 2020) but also for comparisons of pH data from BGC-Argo floats in this region (Wimart-Rousseau et al, 2024) as well as for model validation. The dataset includes 23500 corrected pH data from 11 cruises that will be probably revisited in the next GLODAP version. I wondered why authors did not include their AT data in the file that would help to calculate CT and Cant as well.

The document is well structured, figures and tables adapted. I recommend publication after minor revision. Below are listed specific comments.

;;;;;;;      Specific comments:

C-01: Title: “Two decades of pHT measurements along the GO-SHIP A25 section”. For readers not familiar with GO-SHIP and cruises numbers, maybe specify this is in the North Atlantic: Two decades of pHT measurements along the GO-SHIP A25 section in the North Atlantic.

C-02: Line 59: Not sure that Ishii et al (2025) is a correct reference for OA and BGC-Argo data.

C-03: Line 139: “as well as the ocean's capacity to absorb, store, and transport CO2 (Pérez et al., 2013; Zunino et al., 2015).” You can add reference to Bajon et al, (2025) when published.

C-04: Line 141:  “and understanding the SPNA's response to climate change (Rodgers et al., 2023).” Is it the correct reference for the response to climate change ? DeVries et al, (2023) would be more appropriate.

C-05: Figure 1: I guess one of the cruise in 2014 (GEOVIDE) extended to the west off Greenland (stations south of Labrador Sea). Are these stations included in the new dataset ? If yes, this should be shown in figure 1.

C-06: Line 176: “…should be adjusted by +0.0047 pHT units (Lee et al., 2000).” I think the reference should be DelValls and Dickson (1998). Lee et al used this correction to revise the dissociation constants.

C-07: Line 188:  “(see Fig. S1 in Álvarez et al., submitted).” The paper by Álvarez et al. is not available at that stage (but would be happy to read it).

C-08: Line 387: “This fit, based on 6,910 samples from the 2018, 2021, and 2023 cruises (Supporting Information Fig. 3)”. In Figure S3, N= 2673. Is the fit based on 6910 or 2673 samples ?

C-09: Line 530: “While corrections related to the mCP dye addition effect were included in the data published in GLODAPv2.2023 (Lauvset et al., 2024), the 488A-based correction described in Sect. 2.2.3 had not yet been incorporated”. This is an important information for those who used and will use GLODAPv2.2023. Maybe also indicate here if the most recent cruises (2021 and 2023) have been submitted to GLODAP for the next version ?

C-10: Line 536: “We present a new database comprising 23,535 seawater samples with spectrophotometric pHT values,…”. I wondered why authors did not include their AT data.

C-11: Line 597: “To assess this impact, aragonite saturation horizons were recalculated using in situ temperature, salinity, AT, and pHT values,…”. Interesting sensitivity test, but the AT data are not in the files (correct ?).

C-12: Line 600: “This reevaluation reveals a more pronounced reduction in aragonite saturation at the surface (from -0.040 to -0.065),…”. I suspect this is reduction from pre-industrial period. Please clarify.

C-13: Line 622: “Notably, a persistent pHT minimum appears in the Iceland Basin between 500 m and 1,000 m, associated with intermediate waters with high Apparent Oxygen Utilization”. On this topic, the impact of biological processes on pH distribution was quantified by Lauvset et al (2020).

C-14: Line 651: “The highest OA rates (< -0.002 pHT yr-1) are observed in the surface layers (0—500 m) due to direct air-sea CO2 exchange.” Interestingly, such rate was also deduced in the NASPG from other surface data (-0.0021 pHT yr-1, Reverdin et al, 2018).

C-15: Line 652: “These upper layers also exhibit high interannual pHT variability (Fig. 10), which correlates negatively with AOU (Supporting Information Fig. 6b).” Figure 10 does not show the interannual variability. Maybe refer to Figure 9 here or present another figure for surface layer (in Supp Mat ?).

C-16: Line 673: “NEADW originates from Antarctic Bottom Water, formed in the Vema Fracture Zone, and is largely devoid of Cant (Steinfeldt et al., 2024).” See also Mercier and Morin (1997) who first investigate the AABW in the Atlantic Fracture Zones.

C-17: “6. Data availability” I wanted to explore the files but unfortunately, no access. On “zenodo” the message is: The record is publicly accessible, but files are restricted to users with access.

;;;;;;;;;;;;;;;;;;;;; Reference added in this review, not listed in the MS:

Bajon, R., Carracedo, L. I., Mercier, H., Asselot, R., and Pérez, F. F.: Seasonal to long-term variability of natural and anthropogenic carbon concentrations and transports in the subpolar North Atlantic Ocean, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2025-4425, 2025.

DeVries, T., Yamamoto, K., Wanninkhof, R., Gruber, N., Hauck, J., Müller, J. D., et al. (2023). Magnitude, trends, and variability of the global ocean carbon sink from 1985 to 2018. Global Biogeochemical Cycles, 37, e2023GB007780. https://doi.org/10.1029/2023GB007780

Khatiwala, S., Tanhua, T., Mikaloff Fletcher, S., Gerber, M., Doney, S. C., Graven, H. D., et al. (2013). Global ocean storage of anthropogenic carbon. Biogeosciences, 10(4), 2169–2191. https://doi.org/10.5194/bg‐10‐2169‐2013

Lauvset, S. K., Carter, B. R., Perez, F. F., Jiang, L.‐Q., Feely, R. A., Velo, A., & Olsen, A. (2020). Processes driving global interior ocean pH distribution. Global Biogeochemical Cycles, 34, e2019GB006229. https://doi.org/ 10.1029/2019GB006229

Mercier, H and Morin, P: Hydrography of the Romanche and Chain Fracture Zones, 1997 JOURNAL OF GEOPHYSICAL RESEARCH-OCEANS, VL 102, 10373, DI 10.1029/97JC00229

Reverdin, G., Metzl, N., Olafsdottir, S., Racapé, V., Takahashi, T., Benetti, M., Valdimarsson, H., Benoit-Cattin, A., Danielsen, M., Fin, J., Naamar, A., Pierrot, D., Sullivan, K., Bringas, F., and Goni, G.: SURATLANT: a 1993–2017 surface sampling in the central part of the North Atlantic subpolar gyre, Earth Syst. Sci. Data, 10, 1901-1924, https://doi.org/10.5194/essd-10-1901-2018, 2018.

Takahashi, T., et al, 2009. Climatological Mean and Decadal Change in Surface Ocean pCO2, and Net Sea-air CO2 Flux over the Global Oceans. Deep-Sea Res II, doi:10.1016/j.dsr2.2008.12.009

Wang, Z., et al, 2025. Bias Evaluation for Sensor-Based Dissolved Oxygen from CTD and Profiling Floats in the World Ocean Database JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY, 42, DOI: 10.1175/JTECH-D-25-0027.1

Wimart-Rousseau, C., Steinhoff, T., Klein, B., Bittig, H., and Körtzinger, A.: Technical note: Assessment of float pH data quality control methods – a case study in the subpolar northwest Atlantic Ocean, Biogeosciences, 21, 1191–1211, https://doi.org/10.5194/bg-21-1191-2024, 2024.

;;;;;;;;;; end review

General Comments

I really enjoyed reading this paper. It is well written, thoughtfully laid out, and has incorporated many rigorous experiments explained in significant detail. I liked the background information provided for the A25 section and felt it gave good context to the remainder of the manuscript. This large pH_T dataset will be a great resource for researchers. I particularly commend the authors for including the absorbance values in their pH_T dataset, which the authors correctly point out will be useful in the future if pH_T characterizations are updated. I also appreciate that the authors put these pH_T revisions in the context of the larger scale OA trends. I recommend acceptance of the manuscript after some minor revisions.

Specific Comments

Lines 74-75: pH_T calculated from A_T and C_T is known to have potential issues, such as the pH-dependent pH offset and larger uncertainty. I would argue that this sentence should be rephrased to convey that both (1) pH_T reported with no additional internal consistency corrections AND (2) pH_T calculated from A_T and C_T could both affect the reliability of OA analyses. The key is having directly measured pH_T that can be “trusted”.

Lines 78-79: The pH method was detailed previous to the 1990s (Robert-Baldo et al. 1985; Byrne and Breland, 1989), using mCP as the indicator dye began in the 1990s.

Line 100: Was this the same lot of dye used for all 11 cruises? Or were different batches of dye made over time, just with repeat purchases from the same brand?

Eq. 1: Need to define what pK₂ is for this equation. You should also make it clear that this pK₂ is different from the pK₂ defined in Clayton and Byrne (1993). They list their K₁ and K₂ as formation constants, whereas the K₂ in this manuscript is a dissociation constant. Their K₂ describes the formation of H₂I, whereas the K₂ in this manuscript describes the dissociation to form H⁺ and I₂^-.

Lines 171-172: Again, make clear that the pK₂ shown here is actually labeled as log K₁ (so –pK₁) in the CB’93 paper. I understand why you are showing yours as dissociation constants, but since you are referencing CB’93, you need to make it clear that they are not the same.

Lines 171-173: This sentence is almost identical to a sentence in Lee et al. (2000) and should be properly cited, not just citing the last sentence of the paragraph.

Line 178: Would be useful to mention here that the mCP in CB’93 was also made in DI water. Which is different than what is done in this study (and described in detail later).

Line 185: I think this value should be 0.018 from L’11.

Line 205: Since you’re referencing eq. (11) of DB’17, you should note that your “R_unpur” is noted as “R_obs” in their paper.

Fig. 2: It needs to be made clear that these ₄₃₄A_imp values are specific to one lot from each of these vendors, which are listed in the CB’93 paper. These ₄₃₄A_imp values aren’t generic for all lots of impure mCP with the same vendor name.

Fig. 2: What is the purified line showing? Is this using the same R_unpur and comparing the CB’93 and L’11 parameterizations? It’s difficult to follow the text in lines 220-222 describing this.

Line 219: List the concentration for the mCP dyes determined in this study.

Line 240: Clarify that the blank measurement were recorded at the three target wavelengths (434, 578, 730 nm) like you do for the dye-addition measurements (line 246).

Lines 249-252: It should be clarified that measurements of R are insensitive to slight temperature changes (which is what Byrne and Breland 1989 noted). Also, Byrne and Breland were using cresol red, not mCP. However, pH is quite sensitive to temperature so accurate T measurements are necessary.

Line 264: Specify that an R of 1 corresponds to a pH_T of 7.67 at what salinity and temperature?

Line 279: Can authors justify why their double-dye additions used 50 uL increments instead of 75 uL like their experiments? The authors account for the difference correctly in their equation 6, but I think readers will wonder the reasoning.

Line 297: Might be useful to add a clarifying sentence here to explain that at pH_T y=0, the R (pH) of the original sample and the R (pH) of the indicator are the same, which is why no change is observed.

Line 307: For a couple of the cruises, normalizing the delta R with the isosbestic point reduced the R². Can the authors explain why that would be?

Line 339: R=1 (or pH_T=7.65) at what temperature and salinity?

Fig. 3: Also need to specify which temperature condition

Lines 368-369: How did the researchers determine when the volume deviated by more than 20%?

Lines 395-396: Authors should note that some uncertainty is introduced by collecting duplicates on 2 different Niskin bottles, rather than 2 samples from the same Niskin. (i.e., small leaking in one bottle, biological activity, delay in closing between Niskins, etc.)

Fig. 4: The grey shading is useful but hard to see. Can the authors make it slightly darker? Min and max values are also very small and hard to read.

Line 425: Would be useful to include the final dye concentration in the sample cell for the 2 dye types for a direct comparison, since the experiments used different volumes and stock indicator concentrations.

Fig. 5: Might be useful to reorganize the figure so that the purified dye column is on the right-most side. The unpurified dye columns could then all be shown together as increasingly improving the results to most closely match the purified dye results.

Fig. 6: The “N=176” within the figure seems unnecessary, since it’s listed in the figure caption. Its placement is confusing since it makes it appear that the “N=176” is specifically referring to the middle column of the figure, rather than all of the data.

Lines 514-516: Mention this finding is reasonable since you are working with two different lots of dye. The phrasing now makes it seem like you are recommending a change to the DB’17 value, rather than that you have a different ₄₃₄A_imp value because you have a different lot of dye.

Line 543: No pH values were flagged as 4? If you’re going to highlight the number of flag 3, you should include the number of other flags as well.

Lines 545-547: So, were the pH_T data from 2002-2018 not corrected using the DB’17 adjustment and L’11 equation? Or this was what was in GLODAPv2 before and will be updated using DB’17 and L’11 for v3? It needs to be very clear which data is in each version. Same for the 2021 and 2023 cruises – which procedure was used and what is the final product? Spell it out here for readers who skim, since this is a very important point.

Line 603: What’s the uncertainty in these saturation depth changes? I would be curious what is the range in the depth that saturation = 1 based on the uncertainty in pH measurements themselves?

Fig. 9: Difficult to read pH scale numbers and numbers in each panel, also difficult to read the x and y axes labels and numbers.

Lines 649-651: What are the uncertainties on these pH rate changes per year?

Technical Corrections

Line 14: Provide the acronym meanings for GO-SHIP and OVIDE-BOCATS.

Line 97: Provide acronym meaning for CLIVAR.

Line 163: Should be CB’93.

In my opinion the submitted manuscript “Two decades of pH_T measurements along the GO-SHIP A25 section” is excellent and I suggest should be accepted after minor revisions.

Measurements of any parameter to high accuracy and precision over a long timescale are challenging especially where no certified reference material is available. Further confounding the challenge is the evolving consensus in best practice. The attractive simplicity of spectrophotometric seawater pH measurements has been long established, however the various improvements and refinements potentially thwart production of a consistent high quality long-term data set.

This paper summarises the fundamentals and the issues with dye impurity well then proceeds to discuss various replication and quality assurance protocols used to eventually ascribe a justifiable offset with uncertainty between pH_T data calculated from absorbances determined with unpurified and purified dye stocks. It is difficult to imagine anything further that could have been done ensure the data quality.

Comments

There are a lot of acronyms used and not all are explained on their first use.

Line 163 CB’13 should be CB’93

Line 192 expansion of DB’17.

Line 193 language could be improved

Line 337 orange squares in Fig. 3

Line 410 Font sizes, point colour makes chart difficult to when printed.

Line 447 Colours faint when printed. Could authors comment on why the σ on the purified dye differences (0.0014) seems notably smaller than the unpurified dye (0.0024) for Tris but not modified seawater (both 0.0017).

Line 591 upper-right insert.

Couldn’t see the dataset due to restricted access in Zenodo