This study is an important contribution to our understanding of waterfowl conservation and population ecology in Europe. Recovery of marked birds, typically through harvest by waterfowl hunters, is an important means of obtaining data to assess survival and harvest probabilities in waterfowl, but the ability to differentiate between natural and harvest mortality requires a better understanding of reporting probabilities (the proportion of banded/ringed birds that are harvested by hunters that are also reported to banding authorities). In North America we have had numerous studies using reward bands to estimate this “band reporting rate”, but comparable studies have not been conducted elsewhere, until this study. I thoroughly reviewed this preprint and my overall assessment is strongly supportive. I have only a few suggestions for potential improvement.

It might be nice to bound the reporting rate estimates between 0 and 1 by formally including reporting rate in the model likelihoods rather than estimating it as a derived parameter. I can’t use the link to your code, so I’m unable to see exactly how you modeled this, but you could seemingly model reporting probability directly by including it in the likelihood anywhere that Brownie’s f or Seber’s r appears for birds marked with reward rings.

Lines 328-330: You conclude this paragraph with a statement about your results supporting additive mortality from hunting, but the rationale for this isn’t explained (I’m not disputing your claim, but you haven’t clearly articulated why you believe your results support partially additive mortality). The stark difference in estimated harvest probabilities between newly ringed and previously ringed (i.e., direct vs. indirect in North American terminology) suggests that heterogeneity in vulnerability to harvest might (also) be very important in these populations and thereby contribute to compensation of harvest. Coauthor Emilienne Grzegorczyk presented intriguing results on survival heterogeneity at the latest EURING conference and it might be worthy of a little bit of discussion here.

Minor edits: Line 77 or thereabouts: Because there is an extensive literature on reporting probabilities from North America, but quite different terminology, it might be nice to include a Methods paragraph clarifying ring/band/tag recovery as identical, young vs. adult and hatch-year vs. after-hatch-year, and define the terms direct vs. indirect recovery in terms of time since marking.

Line 154: In addition to the inscription included on reward rings, it would be helpful to indicate the exact inscription provided on standard rings. In North America we observed a pronounced increase in band reporting probabilities when band inscriptions were modified to include toll-free phone numbers and later, web addresses.

When do (most) of your recoveries occur? It would be helpful to include information on timing of harvest in France. Given that you include season of banding as a covariate on survival, subsequent estimates of survival beyond the first year will be hunting season to hunting season. It might be nice to more formally address timing of banding by including a “partial year survival” term in the first diagonal of your m-arrays. This could be a shared annual survival term, but partitioned into portions based on how much of the year an average bird would have to survive (e.g. S^(5/12) if 5 months or S^(9/12) if 9 months).

In North American ducks, we would expect to see pronounced differences in seasonal survival between sexes due to breeding risks incurred by females. For example, spring releases of female mallards would be expected to have lower survival to the first hunting season than spring releases of males. It might be nice to indicate in the methods that you ignored sex in your analysis given small sample sizes (given interactions with species, age, and timing, it might require 6-12 df to properly address), but future analyses based on additional data might wish to investigate sex differences in both survival and recovery probabilities.

You have a nice literature review, but there are a few additional papers that would be worth including: Lines 64-65: Either of the two Riecke et al. 2022 Journal of Animal Ecology papers would be good to cite for an example of how reporting probabilities can help partition annual survival into harvest and natural mortality. Koons et al. (2014, Wildfowl) would be a nice paper to cite here for life-history differences in relation to body size. The results from nasal-marked teal are intriguing, and I suspect that nasal markers might influence survival, vulnerability to harvest, and reporting probability. Arnold et al. 2016 J. Wildl. Manage., Szymanski et al. 2020 Wildl. Soc. Bull., Reinecke et al. 1992, J. Wildl. Manage., Caswell et al. 2012, J. Wildl. Manage.).

Minor changes to wording: Abstract, line 49: I think you mean “subjected” rather than “submitted”. Intro, line 57: “elaborate” rather than “elaborated”. Intro, line 83: use “to” instead of “on”. Intro, line 89: use “of” instead of “or”. Methods, line 126: use “drop-door” instead of “door-falling” Line 161: “departmental”. Line 196: “parameter” (not plural). Line 298: “a heavy predator-control program was in place”. Line 344-345: Curiosity effect has been hinted at in some other research.