Comment on os-2020-120

Short summary: This technical note introduces the use of gas tight bags for the storage of pH buffer solutions, as an alternative to the storage in bottles. The authors demonstrate through comparison pH measurements with the spectrophotometric method that the decrease of the buffer pH can be limited to <0.01 over a period of one year and under laboratory conditions. This is an important finding indicating the potential suitability of such storage bags for repeated calibration measurements of pH instruments during field deployments. The observed pH decrease is attributed to the accumulation of CO2 and likely sources of CO2 are discussed.

bag.As this happened even under laboratory conditions, it appears likely that the structural integrity of the bags will represent a major challenge for successful field deployments of the bags, in particular in the sometimes harsh conditions of the coastal zone.This aspect is mentioned in the discussion, but I think it could be emphasized a bit more, in order to avoid that the approach is considered "ready-to-go" without further testing.
(3) Although the essential information is already covered throughout the manuscript, I would like to encourage the authors to summarize specific instructions for potential users of this approach in a dedicated "Recommendations" section.This section could be guided along three questions: What are the core requirements (material, fittings, etc) of gas tight bags to be used for pH buffer storage?What needs to be considered for proper handling of the bags (cleaning, filling, storage, etc)?Which measures could be taken to ensure that the bags perform well under in situ conditions?

Specific comments
L. 11: Usually, the buffers are prepared as "equimolal" not "equimolar" solutions (DelValls and Dickson, 1998).Please check and -if applicable -correct throughout ms.L. 13: I wouldn't consider the tris batches "experimental conditions", but rather replicates of the same experiment.Consider removing it from the list here.
L. 15: It appears equally (or even more) important to inform the reader about the range rather than the mean of pH changes that the buffers are likely to experience.This information is indicated by the CI (± 0.0008 yr-1), but surprisingly this CI seems to disagree with the range given in the caption of Figure 3 where the upper and lower bounds are given as -0.0028 and -0.0091 pH per year.Can you clarify how the CI of the slope relates to the upper and lower bounds, and include the upper and lower bounds in the abstract?L. 33: Uncertainty thresholds are listed in the wrong order.It should be 0.02 for the weather and 0.003 for the climate goal (Newton et al., 2015).
L. 37: The expression "roughly once per decade throughout most of the ocean" sounds to me as if only one major ocean cruise can be conducted per decade.Maybe change to "decadal reoccupations of a few major sections per ocean basin".L. 81ff.: When I understood correctly, bag type 1 was custom made and type 2 is commercially available.Is this correct and can it be clarified early in chapter 2? L. 99: When I understand it correctly, the unit error affected only the HCl concentration but not the TRIS concentration.If this is the case, then the TRIS/TRISH+ ratio was not exactly 1:1.Please clarify this, and also revise the use of the term "equimolal" accordingly.
L. 100: Please include the degree of purity of the reagents, where possible.(See also comment on l. 223 below).
L. 107-122: The information in text and Table 1 appear redundant.Maybe the text could be restricted to general explanations of procedures, whereas the table could cover the specific routines for each test.
L. 118: The description of the CT measurements would profit from more technical details.Did you need to adjust the method to measure the comparably low CT concentrations in the buffer solutions?Were CRM measured along with the buffer samples?If yes, please indicate the batch.
L. 133: The temperature dependence of the TRIS buffer applied here strictly refers to the equimolal buffer composition, whereas some deviations need to be expected for the buffer composition used in this study.Can you estimate how large this difference might be?Does the dependence of pHspec,Tc on temperature agree with the expected temperature dependence of the buffer solutions?L. 136: The time series of CRM measurements does not show a systematic drift, but a spread (~ 0.03 pH units) that is about an order of magnitude higher than the precision that can be achieved with the instrument used, and also larger than the good agreement between measured and calculated TRIS pH in this study would suggest (compare l. 185 -192).Measured pH values appear clustered by CRM batch and/or measurement time.Would you have an explanation for this and an idea how this would impact your findings?L. 137: I'm a bit sceptical about the approach to correct for dye impurities.First, I'm wondering if this correction is required at all, as the aim is to track pH changes (or better pH stability) over time and therefore the pH-dependent impact of dye impurities should be almost identical for all buffer measurements.More importantly, the comparison measurement of pure vs impure dye made on seawater solutions should also be affected by the pH perturbation of the dye addition.This pH perturbation is related to the pH of the stock solution and can be different for the two stock solutions used here.Did you minimize this pH perturbation by adjusting the stock solution pH to the sample pH, or correct for it by extrapolating your measurements to zero dye concentration?If not, I'm afraid your correction term in Eq. ( 2) might be impacted.Please revise this approach.
L. 145: What do you mean with "normal practices"?Is this a standard operating procedure, or a threshold defined in your lab? Can you provide a reference?L. 198: The estimates of the upper and lower bounds should be given more weight.From a users perspective and for the application of this storage solution without regular pH test measurements, the likely range of pH changes seems even more important the average rate of change!Please explain in the main text, how these bounds must be interpreted in contrast to the 95% CI of the slope.
L. 204: The consequence of bag damage deserves more attention, in particular with respect to the use of bags under in situ conditions.How can this damage happen even under laboratory conditions, and more importantly, how can it be avoided?L. 214: Could your interpretation "that the drift in tris pH was primarily driven by an increase in CO2" also be supported by the change of CT over time?Did you see a consistent increase of CT?I assume this should be the case, due to the fact that pH is decreasing over time and CT and pH seem to be correlated.However, an explicit statement about this would not hurt.L. 216: Can you please describe in more detail what you mean with an "ad hoc acid-base equilibrium model of seawater including tris in addition to the CO2 and other minor acid-base systems"?I've an idea what you mean but it is not entirely clear to me.L. 223: Respiration of organic matter is proposed as one potential source for the accumulation of CT.Could you try to relate the amount of accumulated CT to the size of potential sources?Would it be possible to give a conservative estimate of how much organic material could cover the inner wall of the bags?Which quantity of organic matter must be expected to be contained in the reagents used to produce the buffer solutions?Is tris itself -which is also routinely used in biological experiments to stabilize pH -likely to be respired?I think a bit more detailed discussion to this end would help to identify how the accumulation of CT can be prevented in the future.L. 280: I was not able to access the data at UC San Diego Library Digital Collections through the doi, nor through a keyword search.Please make sure that the data are correctly uploaded and accessible.__________