Use of Telemetry Data to Quantify Life History Diversity in Migrating Juvenile Chinook Salmon (Oncorhynchus tshawytscha)
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsGeneral comments:
The paper describes an analysis of telemetry data collected over ten years on the out-migration of stocked chinook salmon in a California River. The paper is interesting, but can be improved. In general, I think the intro can be made to better focus the attention of the reader. I have made several specific comments about this. I also think the paper would benefit from a deeper discussion/description of the implications of using standard deviation of migration time as a response variable; but also to more clearer explain what exactly the SD values represent: are they SD of individual fish movements per day or the variation in the tagged group of fish? I wasn't 100% sure.
I also suggest either being more cautious or being very clear about your use of the word 'diversity' - only because in biological papers it's typically reserved for describing communities. However, I can also see the link to the benefits of diversity in ecological contexts (e.g., genetic, community composition, habitat) as a buffer against perturbation. And maybe this is already well-established in the paper - that you are arguing that diversity of life history traits like out-migration of salmon is another one of these protective mechanisms. Maybe heterogeneity is a good word, but now I'm probably drifting into preference word-smithing rather than making a meaningful contribution.
Specific comments:
Lines 32-59: This paragraph seems to be communicating a complex set of inter-related ideas. One seems to be the generic explanation for behaviours like anadromy in salmonids. Another seems to be that the benefits to salmon of anadromy are based on overlapping (and typically synchronized) schedules of things like temperature. A third seems to be that the required (relative) synchrony makes such a system fragile. Maybe I’m wrong about the components (other such as management strategies also make an appearance in lines 54-59), but the point of this comments is that this paragraph is very complex and can probably either be simplified or split up.
Lines 60-69: I get the general point, but how you want to connect this material to the previous paragraph is not explicit. Arguably, you are missing a topic sentence here – and you could potentially pull lines 54-59 into this paragraph and make the topic sentence something about management strategies and what types of measurements people use to determine if those strategies are successful?
Lines 95-97: why is there need to expand this toolbox? Presumably there are gaps/problems with the existing approaches? What are some of these problems?
Lines 103-105: You can perhaps be a little more specific here without going overboard. You could potentially even pose a hypothetical question that you are actually deeply interested in addressing in this paper such as ‘Researchers and managers can use advances in telemetry to address questions related to the influence of land use change or damming on life history characteristics.’…but honed to this specific manuscript.
Lines 106: it’s too bad it’s probably considered uncouth to mention - one of the major advantages of studies like this one is leveraging existing data that (probably) required millions of dollars to collect not to mention the blood, sweat, and tears.
Line 108-111: this may come up later, but I may make a comment about standard deviation as a smoother; what I mean is that SD tends to strip out detail and other metrics, such as total distance covered along with something to account for stream distance covered (to separate a fish swimming 30 km to the ocean vs one swimming 30 km in circles) might also be helpful. Although in fairness I also sympathize with the analyst since so many things are easier said than done. But regardless, adding more justification of the value of using SD as a response variable - and the sorts of information which can be gleaned from this measurement will help – maybe part of this is in lines 91-94?
Lines 125-128: maybe this addresses my previous point?
Lines 161-162: nice
Table 1: do you have any data on release numbers? It might be helpful. Also, I don’t know exactly how to do it, but you might also want to consider finding a way to convert this table into a figure; it’s relatively large as a table and is ‘difficult’ to comprehend – depends on what you are trying to communicate though.
Line 187: what’s up with detections prior to a fish’s release? And were there any signal patterns indicating a deceased fish?
Line 202-204: Are you assuming the remaining 9,379 undetected fish died? Arguably these are really censored data since you don’t actually know what happened to them – except for those individuals which may have been predated (those acoustic tags migrating upstream). I’m not sure what, if any implications this has, but I know that ‘losing’ participants (e.g., someone just stops coming for check-ups, or the study ends) is common in human trials. I’m not suggesting you need to re-do your analysis to account for this phenomenon, but there may be some value in follow-up work.
Line 209-211: probably worth highlighting these locations specifically in Figure 1 with maybe asterisks on the appropriate label; makes comprehension faster.
Line 248: I’m not quite sure what ‘access across the landscape’ means.
Lines 269-271: It seems like this analysis is asking about migration from different release points?
Line 272: what are B1 and B2?
Table2: same comment as table 1; this is a big table but doesn’t communicate very much information.
Line 317: why not? Is it an ephemeral channel?
Line 335: might be worth putting these data into a figure like a boxplot (even though I typically advise against boxplots as an explanatory tool).
Lines 357-359: I’m not really sure what this means. I’m also curious about the implications of focusing on variability of measurements (what are the movement scenarios which increase variability?) And it’s variability of individual fish right? So this means that the ‘wetter’ the year, the more likely the migratory pattern is to be patchy? (some days with little movement and some days with a lot?) Or maybe the data says that in wet warm years, is when there tends to be more variability in migration? (which I think makes sense especially if the fish seek out and remain in thermal (or other) refugia?)
Figure 2: I think it bears repeating throughout the paper, or at least constantly reminding the readers why you are interested in variability of migration time rather than some other metric. Also, I typically recommend people add dots to the boxplots to show the actual observations; both together are typically more informative than each alone.
Lines 362-365: I don’t quite understand what this means
Line 370: is diversity and variability the same thing? Also on line 123.
Lines 372: ok fair enough; I’m not really sure how Figure 3 helps me interpret this result. Also, units are missing from Figure 3. And I have a nit-picky formatting tip to save space in figure axes: if all axis labels have the same number of placeholder zeros, it’s better to strip many, most, or all of those zeros from the axis label than to repeat the values multiple times in the axis label and include a multiplier in the axis title. In the case of Figure 3A, 20, 40, 60, and 80 (along with outflow x 1000) might be less cluttered. Although you don’t really have a cluttering problem, but still. Might be useful for future work.
Figure 4: is the line in the plot the global smoother? And presumably it is a continuous line? Seems kinda crazy wild, but I don’t have a ton of experience with GAMs and none with HGAMs. It might be worth stating in plain language what the black line is indicating. And what are the grey pixels? Are those NAs?
Lines 421-422: and presumably these categories were developed from a continuous variable? And these categories are derived from the material described in lines 216-219? I think it may be worth adding some units back into this interpretation. If I’m reading figure 4 correctly, it looks like when flows are above ~ 60,000 cfs and temperatures are between 14 and 17 is when the variability in the migration time is the greatest? Also, I’m curious about the effect of outliers on SD. I know you say in line 360 that no outliers were determined, but does this refer to the distribution of SDs or of the individual movements – also see comment above about lines 357-359. I think my point here is that clarifying what’s driving the differences in the wet years – meaning all of the migrators took longer or a few took much longer is important. From Figure 2, and absence of whickers on the wet years suggests the former, but still. It’d be re-assuring for you to say so specifically.
Line 457-460: the relevance of this to lines 454-457 may need some clarification
Line 463-464: was time of release not included in the model?
Lines 504-506: I think it’s very normal for measurements like temperature used in field studies to be complex indices of many predictors.
Lines 512-515: do you have a sense of how closely the measured water temperatures reflect other parts of the study basin?
Line 518-519: over what time period?
Lines 554-577: arguably this could (or should) be labelled as a ‘conclusion’
Comments on the Quality of English Languagenothing major. There are a few complex statements I didn't understand but I've highlighted some of these in the review comments.
Author Response
Thank you for a thoughtful review, please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThis manuscript requests minor revisions as outlined in the comments belows:
Comment 1 (Abstract, Line 20): the Sacramento River, California, between 2007 and 2017. Only the Upper Sacramento River have the data by release group from 2007 and 2017 (Table 1).
Comment 2 (Meterials and Methods): What the mean for the Rkm (Line 188) and RKM (Lines 243, 245 and 246)?
Author Response
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions in track changes in the re-submitted files.
Comment 1 (Abstract, Line 20): the Sacramento River, California, between 2007 and 2017. Only the Upper Sacramento River have the data by release group from 2007 and 2017 (Table 1).
Response 1: Yes, that is correct. As a synthesis of many studies, not all years include all release groups. We have data for fish released in the Middle Sacramento River starting in 2008 and Tidal Delta starting in 2012. We have made this point more clear in the abstract by adding this sentence after line 20: "In this synthesis we examined a wide variety of landscape and demographic drivers at high resolution by incorporating many individual telemetry studies, with variability in release location by year, environmental conditions and all runs of salmon that are present in the watershed."
Comment 2 (Materials and Methods): What the mean for the Rkm (Line 188) and RKM (Lines 243, 245 and 246)?
Response 2: Thank you for identifying our undefined acronym. Rkm or RKM is river kilometer (measure of distance along a river from its mouth). We have replaced Rkm or RKM with river kilometer throughout the manuscript.
Round 2
Reviewer 1 Report
Comments and Suggestions for Authorsno additional comments