elktoeChemistry: Data and code for Elktoe Chemistry project

NM comments {-}

What is current status of A. Rav?
Perhaps specify the research hypotheses investigated in this study.
1. Assess the compositional differences in shell composition as a function of specimen health/morbidity and intra-watershed location.
2. Among these populations: a. compare inner (nacreous) to outer (prismatic layers) shell compositions during the experimental period. b. Compare inner to outer shell compositions preceding the experimental period, as sampled by replicate whole-valve transects away from the umbo, to assess whether innermost and/or outermost valve layers, biomineralized prior to the experiment, were affected by environmental exposure associated with relocation.
Need to say how old they were; even if retroactive from the annuli counting. How does the experimental period duration compare to their ages? Where do we expect to see growth during the experiment? Since seasonality if a factor for shell growth, when did the experiment take place in this regard? How many specimens had traverses that spanned nacreous AND prismatic calcite deposition during the experiment? These are helpful stats for assessing the population from which inferences are made.
On figure 2: possibly include predominant land use and location of any industrial sites on this map? What are grey areas? What is lake? Could bedrock geology also be shown?
On LOD: Vague; For laser traverses, there is little control over baseline values during the traverse. Accordingly for travers, the Limits of detection are inferred from the analysis of standards analyzed ~hourly over the analytical session. These are typically short traverses with a duration of 60 sec, run in replicate. The short durations, bracketed by similar duration background measurements (when laser is not firing; aka gas blanks) makes it practical to model the predicted baseline variance when the standard replicates were actually measured. Thus, LODs are essentially based on the standard deviation of the baseline measurements bracketing the standard iterations, and then applied to model what baselines were like during the standard analysis. http://iolite-software.com/support/IoliteManualv3.09.pdf
On GAM fit: I think it would be helpful to see an example of this fit. We often do moving median followed by moving average smooths (using same boxcar length) to minimize effect of high-frequency outliers. Reader would like to know how representative these fits are, so think an example would be beneficial to consider.
How do we specifically identify biomineralization within the experiment period? Do we only consider valve intervals since the last annuli was deposited? Maybe I missed this, but we should document this. Are we comparing compositions for the entire valve interval (prismatic, nacre) as opposed to evaluating compositional variations within these intervals? For example, if we see an increase in prismatic layer Pb toward the valve exterior, it might be muted if we present the average Pb for the entire interval as opposed to the last year of growth or last 100µm of growth.
This section selectively documents if we could verify a change in composition (nacreous and/or prismatic layer) related to the experiment? But, it also seems like there should be other foundational results presented, such as:
Results of specimen characterization: How old were they? Annuli counting results
How did specimen health vary over the experiment interval? Was it aberrant?
How did specimen health relate to intra-watershed location? This might be an important factor in interpreting results.
How do elemental compositions vary for the same layers (regardless of the experiment). How much within-layer variance; How does it change as a function of age?
We have a major opportunity to document chemical composition variations associated with these Unionid bivalves – that will make a nice reference database for future studies. Why not make this one of the project goals, in addition to the others regarding population decline? By not tabulating this in the manuscript, we do not show the actual data very effectively (which is a challenge for such a large study). This is also an opportunity to document the natural variability among specimens, which is an interpretation variable otherwise.
What about presenting graphical results showing detail for all transect for a few select individuals? Similar to Fig 2.
Seeing is believing and I think the reader might appreciate such a presentation.
I’m finding it very difficult to understand the significance of A through D in the figure caption. I think I understand what each figure generally shows with respect to the null hypothesis (although is D for prismatic layer during experiment only?). It’s a lot of work to understand this – can we make it easier to follow for the reader?
But we haven’t really evaluated the possibility that metal concentrations might change following biomineralization. If outer shell regions, closest to ambient water, are most susceptible to such changes, the outermost portions of the prismatic calcite layer might have variations related to the change in environment associated with the experiment? Another way to ask this questions might be do we see chemical gradients in the prismatic calcite layer? If so, is it a natural result of the biomineralization process over time or could it be superimposed by the environment?
In what? Mn is a very redox sensitive element. Changes in pH could affect Mn solubility. Large runoff events (e.g., fertilizer) that increased nutrient contents in rivers might reduce dissolved oxygen and increase Mn solubility in the benthic realm?
Again, we need to document specimen health early in the results section. But. . . . . this potential makes it all the more important to evaluate the portions of the shell that may have experienced the effects of the ambient environment, even if they were not actively growing.