the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Aug 2026
Otolith trace element composition reveals spatial and ontogenetic structuring in the glacier lanternfish Benthosema glaciale (Reinhardt, 1837)
Abstract. Detecting habitat use in mesopelagic fishes is hampered by a lack of detailed surveys. Otolith chemistry can fill this gap by providing lifetime chronologies of the environmental conditions experienced by individuals. We analysed otolith transects of six trace elements with known environmental or physiological sensitivity in the dominant North Atlantic myctophid, the glacier lanternfish Benthosema glaciale. Fish were sampled across a complex fjord–shelf system where topographic and hydrographic conditions create vertical and horizontal structuring. Element:Ca ratios changed with age in all six elements. Overall, Sr increased through life, while Ba declined during juvenile stages and rose again in adults. These paired trajectories align with a shift from early exposure to surface-influenced waters toward longer residence in deeper water masses as fish grow. Mg and Mn were highest early in life and declined thereafter, suggesting ontogenetic control linked to growth and physiology. P increased from around the age of maturation, while B showed a weaker age structure. Spatial patterns were dominated by a contrast between the coastal site and the fjords, with additional element-specific differences among fjords. Individual fish also differed persistently in their overall elemental levels and in how their chemical trajectories changed with age. Several elements covaried after accounting for age, site and individual structure, most strongly Sr with Ba, suggesting shared incorporation processes or exposure histories beyond their common ontogenetic and spatial trends. Our results indicate that otolith chemistry in B. glaciale records life-stage-specific information on water-mass exposure, providing a retrospective window on habitat use where direct observation is impractical. More broadly, they highlight the value of this approach for linking biological and biogeochemical processes in deep and vertically structured systems.
- Preprint
(1798 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-4521', Anonymous Referee #1, 24 Aug 2026
-
RC2: 'Comment on egusphere-2026-4521', Anonymous Referee #2, 22 Sep 2026
Comments to the Author
Saltalamacchia et al. present a study on how otolith chemistry can be used to study the spatial structure of a common mesopelagic fish, taking into account the ontogenetic influence on elemental incorporation. The findings are interesting, and the study offers valuable insights into elements that are less commonly examined in otolith chemistry research (P and B).
The use of GAMMs is an interesting, well-described, and well-suited approach for this dataset. The results match the stated aims, and the claims stay within the scope of the study. The complexity of separating environmental from physiological drivers of elemental incorporation in otoliths is well integrated and thoughtfully discussed throughout the manuscript.
While the specific research questions are well addressed, I feel the authors fall somewhat short of fully achieving their broader aim of clarifying "how otolith chemistry can be used to infer habitat use in deep-water species," due to the lack of an in-depth multi-elemental analysis. Aside from the autocorrelation analysis, elements are treated separately throughout. I think this gap should at least be acknowledged as an avenue for future exploration, or the authors should justify why a multi-elemental approach was not pursued.
Some minor methodological details could be added for greater precision and clarity (see specific comments below).
Given how frequently environmental conditions are discussed throughout the manuscript, I think it would be worth moving a figure from the supplementary material into the main text, so readers can more easily grasp the environmental variation across sampling sites. Figure A4 could be a good candidate for this.
Overall, this is a well-written and well-balanced manuscript that makes a valuable contribution to the fields of mesopelagic fish ecology and otolith chemistry. Please see the comments below for suggested minor corrections.
Detailed comments
Abstract
L17: Consider listing the six trace elements in parentheses after first mention (e.g., Sr, Ba, Mg, Mn, P, B).
Introduction
L110: Previous paragraphs on Sr and Ba explained how they were expected to react. Why was this not done for B as well? Are changes in pH expected among habitats?
L121-122: “In otoliths, Mn is influenced not only by ambient availability but also by growth and metabolism-related processes, which complicates its interpretation as a purely environmental tracer (Cavole et al., 2023; Turner and Limburg, 2015).” This sentence makes it sound like it’s a particular case for Mn, while it is also true for most elements used in microchemistry.
Material and methods
L158-161: Could it be worth mentioning why the slow growth period is happening in summer and the fast growth in winter, as it is usually the opposite?
L186: Is there a grit size associated with the grinding paper? It might be worth mentioning.
L195: I am confused by the mention of the ablation points accompanying the ablation transect here: “we ran edge-to-edge transects (ablation points) along the major otolith growth axis and passing through the core”, as you mention a constant travel speed of the laser beam. Is it a series of points or a continuous ablation? Or are the ablation points referring to the output data points? Consider clarifying.
L198: How much time does it take, approximately, to run 7–10 otoliths? Is it roughly every hour? It would be helpful to add that information.
Results
L309-310: I am wondering if this is the clearest way to write it: “The average number of laser ablation points per translucent and opaque increments was, respectively, 4-9 at age 0; 6-23 at age 1; 6-10 at age 2; 4-6 at age 3; 4-4 310 at age 4; and 3-4 at age 5.” You could consider a small table instead.
L329: I am unsure of the added value of setting a reference age of 2. When it is mentioned again at L358, I would prefer the age to be mentioned directly.
Citation: https://doi.org/10.5194/egusphere-2026-4521-RC2 -
AC1: 'Reply on RC2', Francesco Saltalamacchia, 24 Sep 2026
We thank both reviewers for their thoughtful and constructive comments. We are working through each point carefully and appreciate the opportunity to seek clarification through the interactive discussion platform.
Regarding the final point raised by RC2:
“L329: I am unsure of the added value of setting a reference age of 2. When it is mentioned again at L358, I would prefer the age to be mentioned directly.”
Could RC2 please clarify whether the concern relates to the centring of the age variable before fitting the GAMMs, or only to the wording in the Results, where “age 2” could be used directly instead of “the reference age?
Citation: https://doi.org/10.5194/egusphere-2026-4521-AC1 -
RC3: 'Reply on AC1', Anonymous Referee #2, 24 Sep 2026
Thank you for your consideration. The comment only referred to the wording.
Citation: https://doi.org/10.5194/egusphere-2026-4521-RC3
-
RC3: 'Reply on AC1', Anonymous Referee #2, 24 Sep 2026
-
AC1: 'Reply on RC2', Francesco Saltalamacchia, 24 Sep 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 223 | 75 | 36 | 334 | 31 | 24 |
- HTML: 223
- PDF: 75
- XML: 36
- Total: 334
- BibTeX: 31
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Overall assessment
Saltalamacchia et al. present a paper that uses otolith trace elemental chemistry to understand habitat use of glacier lanternfish (Benthosema glaciale), an important myctophid, in Norway. This paper is generally well written; the statistical approach is sound and the conclusions drawn are appropriate for the data.
They examine temporal changes (within in individuals’ lifespan) and spatial differences (across 5 sampling sites) in six different trace elements within otolith. They primarily sought to address three research questions (1) how otolith elemental composition in B. glaciale is structured across age; (2) whether age related patterns in elements (particularly Sr and Ba), vary with expected shifts in depth-related habitat use and water-mass exposure; and (3) whether elements covary within and among individuals.
To address these questions the authors use two primary approaches to analyse elemental data generated though laser ablation – ICP-MS. 1) Generalised Additive Mixed Models (GAMMs) to examine how elementals in otolith varied with age. This was their primary approach to reconstructing ontogenetic changes habitat usage. 2) a Bayesian multivariate mixed-effects model to evaluate covariation within and between individuals. This statistical approach is generally sound and appropriate for the data.
They found that all six elements showed significant age-related structure, with Sr:Ca increasing progressively through life and Ba following a U-shaped trajectory that declined through the juvenile period before rising again in adults a pattern consistent with fish moving from surface-influenced waters into deeper, more stratified habitat as they mature. Mg and Mn declined with age, while P increased from around the age of maturation and B showed only weak age structure. Spatial differences were dominated by a consistent contrast between the coastal site and the fjords, with additional site-specific variation superimposed on this pattern, and individual fish differed persistently in both their overall elemental levels and how those levels changed with age.
The multivariate analysis further showed that several elements covaried once age, site, and individual trends were accounted for, most notably a strong positive association between Sr and Ba at both the within- and among-individual level, alongside a physiologically consistent positive coupling between Mg and Mn and a growth-related negative association between Sr and the metabolic tracers Mg and Mn.
This leads the authors to conclude that otolith chemistry can be used to reconstruct habitat usage and the ontogenetic changes in vertical habitat. The covariance structure in elements observed between elements within and between individuals lead them to conclude that categorisation of otolith elements as either environmental or physiological is too simplistic and that age, site, and individual-level trends need to be partitioned out.
Although overall this manuscript is sound, there are some aspects, particularly around methodological detail and framing of a few interpretive claims, that would benefit from clarification, and these are outlined in the comments below.
General comments
The potential for ontogenetic changes in growth rate to affect the temporal resolution of the LA-ICP-MS record deserves discussion. Because ablation proceeds at a constant travel speed (line 196), a single laser spot integrates over a longer period of otolith formation in slower-growing, older fish than in faster-growing juveniles (Fig. 3). This raises the possibility that finer-scale chemical variation is smoothed out in older age classes, which could be important when interpreting the flattening of trajectories in later life (Fig. 4).
Average increment widths are shown pooled across all sites in Fig. 3, and Fig. A3 gives site-specific cumulative otolith distance at each age landmark. It would be useful to report site-specific increment widths by age class, likely in the supplement given the probable size of the table. This would clarify whether the differences already visible in Fig. A3 translate into differences in age-specific increment width, and whether such differences could themselves contribute to the spatial patterns in E:Ca.
AIC values are not reported in Table A1, only the qualitative rationale for each candidate formulation. Since model selection by ΔAIC < 2 is described in the text (line 260), the AIC or ΔAIC values should be added as a column so readers can evaluate model support directly.
The models do not appear to explicitly account for autocorrelation along the ablation transect. Adjacent points on the same transect are likely more similar to each other than the mean structure (shared smooth plus random slope) predicts, beyond what the per-fish random intercept and random age slope can capture. The authors could clarify whether this was tested, for example via a residual autocorrelation diagnostic per fish, and if structure remains, whether an AR1 term would be appropriate, or explain why this is unlikely to affect the reported trends.
The rationale for interpolating rather than removing below-LOD values is only briefly addressed in the Limitations section (lines 542 to 544), where it is described as conservative and unlikely to bias results. This reasoning should be expanded and justified.
As some sites there is a small sample size (e.g. n is between 6 and 8 at three sites), while I understand the logistics and the problems that can arise with obtaining appropriate sample sizes for this type of study I think that there should at least be some discussion of the implications of this and the impacts that it could have on the overall results.
Otolith weighing is mentioned as part of the extraction method, but no otolith size or weight data are reported in the Results. Was this compared among sites? This is distinct from the increment-width data in Fig. 3 and Fig. A3 and could add useful context for interpreting site differences.
I appreciate figure 3 showing the differences in mean otolith size. What is driving the larger size of the opaque section in year 1 – I see that the SD is also much larger here, are there one or two individuals that are driving this? How might this much larger mean increment width impact results and what the LA-ICP-MS results show?
Are there site-specific differences in vertical habitat use, and if so, what are the ecological and methodological implications of this?
Could the authors use the full multi-elemental suite together, rather than each element on its own, to better parameterise movement across water masses? This might dampen the impact of physiological effects on any single element and give a more robust proxy of habitat use. This is different from the covariation analysis in section 3.2, which looks at whether pairs of elements move together, not whether the elements can be combined into a single index. While I am not suggesting that the data is reanalysed it might be worth touching on this in the discussion.
Specific comments
Line 19: It would be good to include the specific elements measured here, since only those with results are mentioned in the abstract.
Line 50: Is it just size, or also ontogenetic variation in the intensity or direction of DVM? The two are closely linked, but it might be worth mentioning the ontogenetic aspect here too, given how heavily this manuscript leans on age-based differences.
Line 53: Is having quotation marks around "voyage recorders" necessary? Worth checking with the copy editor.
Line 58-59: Minor rewording could make this clearer. It's not immediately clear whether they mean the amount of new material deposited is proportional to activity, or that the elements incorporated are proportional to activity.
Line 112: The evidence for pH impacting boron levels is a bit mixed. While Limburg et al. (2023) found that the overall decrease in pH over time has been associated with changes in B, this might not always hold at the individual or population level under controlled experimental conditions, particularly over the course of an individual's life. This pattern could be weaker within individuals.
Line 127: P is far less frequently used as part of the elemental suite in otolith studies; it would be great to have a sentence or two here just providing a little more reasoning for its inclusion.
Figure 1: The labelling of Sognefjord twice on the map confused me. It took me a while to understand that one label refers to the fjord itself while the second is pointing to the sampling site. None of the other fjords have this dual labelling, so it might be worth removing the general fjord label or making it visually distinct from the sampling site labels.
Line ~205: While most of the LA-ICP-MS detail is reported in Saltalamacchia et al. (2026), I think the authors still need to provide at least the precision and accuracy here from the standards.
Line 216: Were measurements taken across the otolith before sectioning? This might help with understanding how accurately the sections were cut.
Figure 2: It might be worth changing the colours used to highlight the sectional plane and the laser transect. It took me a little while to realise the figure annotations were representing different things.
Line 290: The abbreviation MCMC hasn't been defined in full yet, so it should be added here.
Line 392: What if sexual maturation itself is driving the change in elemental composition, rather than, or in addition to, the habitat shift?
Line 407: Is this a habitat thing or an incorporation thing?
Line 425: I think this section is well structured and relates well to the previous paragraph discussing the physiological modulation of Sr. It might be worth adding a caveat around the fact that depth use stabilises but Sr levels still change, and flagging that physiological effects may be contributing to this later-life pattern.
Line 440-443: While reading this I made a note in the margins that said "this is good stuff."
Line 485: The authors discuss age a bit here, but what about the potential for fish size to impact these relationships between elements and physiological/environmental conditions? I know size and age are pretty confounded, but size independent of age has been shown to have an impact elsewhere.
Line 498: The fish sampled here are from mid-water rather than near the benthos. How far from the benthos could redox-influenced sediment Mn still be affecting E:Ca chemistry in these individuals?
Line 520: I would be interested in the potential future use of P:Ca as a tracer of metabolism.