We thank Reviewer #1 for the positive and helpful comments. Our response is provided below.

“First, it should be clear where in the artificial sediment the littleneck clams (L. staminea) live. Pore water chemistry was determined at a depth of 8 cm (Line 172) and compared in the results section with samples taken from the overlying water. Since (carbonate chemistry) parameters are different for these two samples, the sediment likely contains a gradient of conditions. It should be clear where the clams exactly lived (even if it varies between specimens, between conditions or in time) to indicate what conditions they experienced. Now it is implied that they all lived at a depth of 8 cm and that those pore water conditions were the ones experienced by the clams.”

Response: This feedback shows that we need to do a better job describing the experiment. The clams' precise vertical position of the clams was not measured, though they were limited by their small size to the top of the sediment (generally above 8cm). The 8cm depth for porewater measurement was chosen based on the amount of porewater needed to run the chemistry analysis, which was ~250 mL per sample (one 125mL sample for pH and one 125 mL sample for alkalinity). Because of the large volume of porewater collected, our measurements represent an average of porewater chemistry, as opposed to a measurement of porewater chemistry at 8cm specifically. In our revised manuscript, we will clarify these points in greater detail.

“More importantly, the Discussion lacks the necessary depth. I suggest the authors extend the discussion by comparing their results with earlier experiments with bivalves under OA-conditions. Is reduced growth under acidified conditions reported before and were responses comparable? Or do they seem to vary between species? What does this mean for natural environments in which the sediment is rich/ poor in carbonates? Is the applied level of acidification representative for a near-future scenario? How fast does the shell hash dissolve? And how does this fit on the growing literature on alkalinity enhancement to mitigate acidification?”

Response: This is a good suggestion; in the revised manuscript we will do more to place our findings within a broader ecological and applied context. Our applied level of acidification was indeed chosen to mimic near-future scenarios, but much more could be done to compare our work to others who have done the same. 

Regarding comparisons between species–previous studies have shown marked differences in the response of marine calcifiers to ocean acidification, with molluscs generally being more resilient than groups such as corals or dinoflagellates (Bednaršek et al. 2025). A recent metaanalysis on bivalves (Hu et al. 2026) suggests that calcification rates generally decrease under moderate ocean acidification conditions, but that these results vary widely by species and lifestage. For example, juvenile clams such as Tridacna squamosa and Chamelea gallina show decreased calcification under OA conditions (Armstrong et al. 2022; Sordo et al. 2021) while the scallop Argopecten purpuratus and the mussel Choromytilus chorus did not (Lagos et al. 2016; Benitez et al. 2018).

Field work on shell hash introduction similarly suggests that the results are mixed, and depend on location as well as time since treatment (Doyle et al. 2022; McGarrigle et al. 2024). That means carbonate-rich versus carbonate-poor sediments cannot be assumed to be “good” or “bad: for bivalve health.. A 2024 additionality study argues that natural alkalinity release occurs in many coastal settings, and that organic-matter supply and hydrodynamics can matter more than carbonate content alone because added alkalinity can suppress natural CaCO₃ dissolution.

Several studies have looked at the dissolution rate of biogenic calcite and aragonite, which places the kinetic constant for both at ~5 yr⁻¹ , producing a dissolution rate of ~0.05-0.2 mol alkalinity g^-1 d⁻¹ depending on the saturation (Biçe et al. 2025). This puts the first-order half life at about 50 days, which is consistent with the loss of shell hash efficacy reported after ~2 months in McGarrigle et al. (2024). 

In conclusion, our results fit squarely into the growing yet complex literature on ocean alkalinity enhancement: shell hash can raise pH, total alkalinity, and saturation state with respect to carbonate minerals (e.g. aragonite and calcite) and is being explored as a mitigation strategy for marine calcifiers, but ecological responses are mixed. We will do a more thorough job in the revised manuscript to incorporate these ideas and expand upon them.

References: Armstrong, Eric J., et al. "Elevated temperature and carbon dioxide levels alter growth rates and shell composition in the fluted giant clam, Tridacna squamosa." Scientific Reports 12.1 (2022): 11034.

Bednaršek, Nina, et al. "Assessment framework to predict sensitivity of marine calcifiers to ocean alkalinity enhancement–identification of biological thresholds and importance of precautionary principle." Biogeosciences 22.2 (2025): 473-498.

Benítez, Samanta, et al. "High pCO2 levels affect metabolic rate, but not feeding behavior and fitness, of farmed giant mussel Choromytilus chorus." Aquaculture Environment Interactions 10 (2018): 267-278.

Biçe, K., Myers Stewart, T., Waldbusser, G. G., and Meile, C.: The effect of carbonate mineral additions on biogeochemical conditions in surface sediments and benthic–pelagic exchange fluxes, Biogeosciences, 22, 641–657, https://doi.org/10.5194/bg-22-641-2025, 2025.

Doyle, Bridget, and Leah I. Bendell. "An evaluation of the efficacy of shell hash for the mitigation of intertidal sediment acidification." Ecosphere 13.3 (2022): e4003.

Hu, Nan, Zhuonan Wang, and Jiamin Sun. "A global meta-analysis reveals consistently negative effects of ocean acidification on marine cultured bivalves: implications for future bivalve aquaculture." Reviews in Fish Biology and Fisheries36.1 (2026): 10.

Lagos, Nelson A., et al. "Effects of temperature and ocean acidification on shell characteristics of Argopecten purpuratus: implications for scallop aquaculture in an upwelling-influenced area." Aquaculture Environment Interactions 8 (2016): 357-370.

Luff, R. and Wallmann, K.: Fluid flow, methane fluxes, carbonate precipitation and biogeochemical turnover in gas hydrate-bearing sediments at Hydrate Ridge, Cascadia Margin: numerical modeling and mass balances, Geochim. Cosmochim. Ac., 67, 3403–3421, 2003. 

McGarrigle, Samantha A., Mia C. Francis, and Heather L. Hunt. "Effects of experimental addition of algae and shell hash on an infaunal mudflat community." Estuaries and Coasts 47.6 (2024): 1617-1636.

Sordo, Laura, et al. "Long-term effects of high CO2 on growth and survival of juveniles of the striped venus clam Chamelea gallina: implications of seawater carbonate chemistry." Marine Biology 168.8 (2021): 123.

We thank Reviewer #2 for the detailed feedback. In short, we agree that there are better statistical approaches to the ones used in this manuscript. We have onboarded a new co-author, Dr. Gabriel Ng, who is helping us reanalyze the data using GLMMs as suggested. Below we provide point-by-point responses to the Reviewer’s concerns and our plans to reanalyze the data:

“The problem here is that pH and header tank are completely confounded. Any difference between ambient and acidified treatments could be due to pH itself, or some other/unmeasured characteristic of the header tank (microbial community, chemistry, etc.). Because there is only one header tank per pH level, there is no way to separate these effects statistically. This is particularly concerning because the seawater in the header tanks was replaced every three weeks (line 153), which may have allowed water quality to evolve differently between tanks and deteriorate independently of pH, thereby confounding the interpretation of treatment effects.”

Response: We agree that the experimental design does not allow us to separate pH from header tank as independent experimental factors. The revised manuscript will explicitly acknowledge this limitation and will temper the language throughout so that the results are presented as tank-level treatment patterns rather than formally replicated pH effects. Still, we note that many aspects of seawater chemistry were measured in water coming from the header tanks (dissolved oxygen, pCO2, DIC, alkalinity, ΩAragonite, ΩCalcite, salinity, temperature…) and pH appears to be the most relevant difference. The water chemistry data therefore indicates a consistent difference between the two header-tanks, and the biological responses track those differences. That said, we agree that this does not eliminate the confounding issue, and we will revise the text to make clear that our interpretation is limited by the lack of independent replication at the header-tank level.

“Lack of replication for the pH treatment. Strictly speaking, substrate effect is testable, but pH effect is not formally testable because pH has only one experimental unit (header tank) per level. The pH × substrate interaction is also problematic because pH lacks replication at the level at which it was applied. Overall, shell ash was replicated among buckets and its effect can be evaluated. In contrast, pH was manipulated at the header-tank level with only one header tank per treatment, so pH effects and pH × substrate interactions should be interpreted as treatment-level patterns rather than formally replicated tests.”

We agree. In the revision, we will use GLMMs to test for a header tank effect and a “header x substrate” interaction. Rather than a population, we will ask whether the substrate differs in response between one header tank vs the other.

“2. Low replication of the shell ash treatment. Another important limitation of the experiment is that there are only 2 buckets per treatment combination. So the effective replication for testing pH and substrate effects is low. The confidence intervals and p-values should be interpreted cautiously because they are based on only 8 experimental units.”

Response: We agree that the bucket-level replication is limited. In the revision, we will emphasize effect sizes, confidence intervals, and the limited number of experimental units, rather than relying on p-values alone. We will also make clear that clams within the same bucket are not independent replicates. The revised analysis will therefore be built around the bucket as the experimental unit, with individual clams nested within each bucket. 

“3. The statistical models used are unclear and probably unappropriated. The statistical approach does not appear to reflect the hierarchical structure of the experimental design. The authors compared several ANOVA models, including additive and interaction models with tank ID treated as a blocking factor, and selected the preferred model based on AIC. However, tank is not a blocking factor but rather the experimental unit to which treatments were applied. Individual clams within a tank are therefore not independent observations.

Given the nested structure of the data (clams within tanks), the analysis should be based on a general linear mixed-effects model (GLMM), with pH and shell-ash treatment specified as fixed effects and tank included as a random effect to account for the non-independence of observations within tanks. Model selection among different ANOVA formulations does not address this issue and may lead to pseudoreplication. Line 260: For whatever reason, the variable 'new growth' was analyzed using GLMM, which seems appropriate.”

Response: We agree that GLMMs will better represent the hierarchical structure of the experiment. In the revised manuscript, we will use buckets as a random effect to account for the fact that shell hash was randomized between buckets rather than for individual clams. The models will be structured based on the initial hypotheses, so that we don’t need to use AIC to determine which model to go with. We will also revise the text so that the statistical model is described consistently across response variables. The current manuscript already uses a mixed-effects approach for percent new growth, and we will extend that logic to the rest of the analyses so that the full statistical treatment is internally consistent. 

“More broadly, the statistical analyses are not presented consistently across response variables, making it difficult to evaluate the strength of the evidence and compare results among parameters. In addition, the manuscript does not provide statistical analyses for the carbonate chemistry parameters measured in the overlying water, a summary table for carbonate chemistry parameters in pore water, or the full test statistics associated with the growth responses.

I recommend that the authors provide a comprehensive summary table reporting the statistical results for all measured variables. At a minimum, this table should include the model used, sources of variation, numerator and denominator degrees of freedom, effect estimates and test statistics (F-values and p-values). Such a table would greatly improve transparency, facilitate interpretation of the results, and allow readers to assess the consistency of treatment effects across response variables.”

Response: We will restructure the Results so that each response variable is analyzed and reported in a consistent format. We will add a summary table that reports, for each response, the model used, the relevant test statistics, degrees of freedom, estimates, and p-values. This will make it easier for readers to evaluate the strength of the evidence across carbonate chemistry and growth responses.

“In the same vein, the statistical treatment of the carbonate chemistry parameters could be improved. Rather than conducting separate analyses for pore water and overlying water, the authors could consider fitting a single model including water type as an additional factor. Such an approach would allow formal testing of differences between water compartments while accounting for the hierarchical structure of the experiment. Depending on the sampling design, water type could be treated as a repeated measure or as a factor nested within tank.”

Response: In the revised manuscript, we will re-evaluate the pore-water and overlying-water chemistry using a model that explicitly accounts for the repeated sampling structure of the experiment. This should allow formal assessment of differences between pore water and overlying water while respecting the nested and repeated-measures nature of the data. 

“In addition, it is unclear how temporal replication was handled in the analyses. If measurements collected at different sampling dates were treated as independent replicates, this would constitute pseudoreplication because observations through time within the same tank are not independent. Instead, sampling date should be incorporated explicitly into the model, for example as a repeated-measures factor (or fixed effect), with an appropriate covariance structure to account for temporal dependence. The authors should clarify how temporal replication was handled and revise the analyses if repeated observations were treated as independent replicates.”

Response: We agree and will address this directly in the revision. Repeated chemistry measurements through time within the same experimental units are not independent, and we will analyze them accordingly. We will incorporate sampling date explicitly into the water-chemistry models and use an appropriate repeated-measures or mixed-effects structure to account for temporal dependence. We will also clarify this in the Methods so that the handling of time is transparent to readers.

“Finally, I do not understand how to interpret Table 4. Also, I cannot figure out which treatments differed from which. Could you please add different letters to indicate significant differences directly in Figure 4?”

Response: We will revise Table 4 so that it is easier to interpret and will add clear significance groupings in the figure panels. We will also ensure that the figure captions explain the comparisons clearly.

“Growth parameters and husbandry conditions are difficult to appreciate. As a bivalve biologist, I was struck by the small increments in shell size observed after 90 days at a temperature of 15°C. Is this normal for this species? I would appreciate a discussion about this specific point. How does this growth compare to what happens under natural field conditions?”

Response: There is not much work on growth rates in Leukoma staminea. Our previous work suggests they reach ~7-15 mm after 200 days (Kempf et al. 2023), and specimens from British Columbia reach 22 to 35 mm long between 1.5 and 2.5 years old (https://www.dfo-mpo.gc.ca/species-especes/profiles-profils/littleneck-clam-palourde-pac-eng.html). So these do not appear to be fast-growing animals. The juveniles in this study were already 200 days old when the 90-day experiment began, so the measured growth reflects post-settlement laboratory growth under controlled conditions rather than early larval development. We will discuss these points in the revised manuscript.

Reference: Kempf, Hannah L., David A. Gold, and Sandra J. Carlson. "Investigating the Relationship between Growth Rate, Shell Morphology, and Trace Element Composition of the Pacific Littleneck Clam (Leukoma staminea): Implications for Paleoclimate Reconstructions." Minerals 13.6 (2023): 814.

“Discussion needs further development. The discussion is far too short. I would very much appreciate some thoughts on the potential of using shell ash to cope with ocean acidification. Do you think this is feasible? Is it scalable? Would it be acceptable to growers and fishermen? Furthermore, given the limitations of the experimental design, I would be grateful for a statement on this in the discussion.”

Response: Please see our response to Reviewer #1 for a detailed response on how we plan to improve the Discussion. Our expanded and revised Discussion section will also consider the potential of shell hash in applied settings. In short, there is some data suggesting this approach is scalable, but the effects of shell hash mitigation appear to vary widely by location and species. There is also great heterogeneity in how bivalves respond to ocean acidification more broadly. And of course, we will explicitly discuss the limitations of our experimental design so that the applied implications are not overstated.

“Please respect the following variable order in the manuscript text, figures and tables: pH, TA and omega”

Response: We will standardize the order of variables to pH, TA, and Ω throughout the text, figures, and tables. 

“Figure 3 plots pHtot in pore waters as a function of time for each treatment. This is the only variable presented with such detail. Why is that? Please consider doing the same for other carbonate chemistry parameters, at least in a supplementary file.If not, please provide a justification. It is likely that the TA will increase over time in the shell ash treatment, particularly under OA conditions. In any case, time should be investigated and considered a repeated measure in the statistical model (see point 3).”

Response: pH_tot had an interesting pattern, which is why we highlighted it in Figure 3. As part of our general approach to treat temporal chemistry patterns more transparently in the revised manuscript, we will add time-series plots for the other carbonate chemistry variables, at least in the supplementary materials. As discussed earlier, we will also consider time in our revised statistical analysis, which will bring this time component to the fore.

“Add some description of the shell ashes composition, grain size, …”

Response: Much of this is in the methods (for example, we note that shell material was crushed and size-graded to 0.25–0.5 cm), but we will prepare a specific subheading for shell hash preparation to make this clearer and more explicit.

“F-values in the text are not well reported. The author reported as F(numerator df) while it should be F(numerator df, denominator df).”

Response: We will fix this in the revised manuscript.