Habitat Impacts on the Golden Eagle’s Foraging Ecology and Nest Site Selection in Poland

Comments

Answer

Reviewer 1

This is an interesting and useful paper which considers interactions between habitat and diet in a breeding group of Polish golden eagles. The relatively frequent records of several smaller meso-predators in the diet are particularly interesting. In the main it is well-written and employs a logical set of text, even if in a few places some terms and use of English could be better. I would expect, however, that English is not the first language for the authors and so in that circumstance the paper is admirable. (In a revision you will probably have to change species names to lower case, as per the journal requirement.)

Line 13. Replace ‘after that’ with ‘hereafter’. You should also consider removing ‘the’ before using GE, on several occasions (even in the Abstract) and be aware that on line 33 you stop using GE and revert to golden eagle for some reason. Check line 31 too (‘Eagle’), for consistency.

It was corrected

Line 16. Use ‘On average’ not ‘in the average’

It was corrected

Line 25. a not an

Line 27. ‘Suggesting’ would be better than ‘indicating’ when no prey/food surveys were done

It was corrected

Line 32. ‘Identified’, on spelling

It was corrected

Line 75. Delete ‘built’

It was corrected

Line 76. Cliffs, not rocks

It was corrected

Line 76-79. Please check the English here, as there are several problems

It was corrected

Line 83. Suggest replacing ‘occasionally’ with ‘also’ for more accuracy

It was corrected

Line 91. Least

It was corrected

Line 109. Latitude, and altitude also?

It was completed

Line 124. Check text “Central The...”

It was corrected

Line 127. Human would be better than population

Line 152-154. There is a problematic issue with the use of ‘random points’ as an alternative for eagle nest sites when nest sites are not random in the environment. As the authors (sort of) note in line 199 and more explicitly in line 274-276 (and many other studies have shown) golden eagles require either particular large trees or cliffs for nest sites, and these are not random features of the environment. The authors could have used some stratification in their selection of random points to at least partially account for this (like there is no way that eagles would nest in farmland) but haven’t. Such an additional analysis would take some time and effort but would probably still not account for all the potential biases and the false equivalency of random points. At the very least the authors should note that this ‘random point’ selection procedure could probably have flaws and may affect results and conclusions. Personally, I do not like such “random” procedures to evaluate non-random data, when to my mind they often just weaken rather than strengthen any results and conclusions. When usually unnecessary and flawed I don’t see the point of them or any value they bring

We have clarified the description of the method of drawing random points. Random points were added to the map. 

The drawing procedure created a layer consisting of forested areas and non-forested mountain tops. Urbanized and agricultural areas and water bodies were excluded.

We generally agree that random points are not a remedy for everything. Hence, we believe it is worth indicating how the points chosen by the birds differed from potential points drawn in the environment.

Line 156-158. How did you treat results from pellets and food remains from nest? Apparently in combination, but not separately? The biases associated with these two different collection/analytical collection methods are different, and so can be difficult to combine (Whitfield et al. 2009 [cited in the submission]) although see Sidiropolous et al Diversity 14 (2022) on this. There is a lack of clarity in this important part of the methods which should be rectified

It was explain in text.

Line 159-161. It is also not clear how the Minimum Number of Individuals (MNI) analytical method was employed (e.g. Sidiropolous et al Diversity 14 2022, and references therein). Some clarification again is needed here. There is a paucity of relevant references on this critical method

It was explain in text

Line 161-162. Using the MNI method (not noted) is not quite this simple. More information is needed on what was done, and if MNI principles were actually used. Especially when combining results from pellets and nest remains collections

It was explain in text

Line 168-171. Sorry I’m confused here. You note a GLM but then say GLMs. You say that you used a GLM using data for each nest. Does that mean you ran 28 GLMs (as there seems to have been 28 nests with sufficient sample) or did you use nest site as the dependent and included as independent predictors (not dependent) the lat and long? Or did you use lat and long as the dependent (based on grid references for each of the 28 nest sites) and used the prey proportions as independents? I think that it’s the latter, but it’s not immediately clear in the description

We have clarified the terms GLM and the dependent and independent variables described.

Line 170. What was your cut-off on numbers of food items? Ah, I see later (line 174) that it was 15 items, but this seems to be across several years per nest site/territory (line 178). Others have used 10 items

We decide to put the threshold on 15 level, for better data quality. With 10 preys modeling was much more difficult.

Line 1710-173. Again I’m slightly confused. You say that independent variables were proportions of the six most frequent bird food categories and four/five most frequent mammal food categories. That adds up to 10 or 11 (really should not be “or”: what was it, 10 or 11?!). Soon after (line 175-178) you give a list of these independent variables: pigeons etc etc. which adds up to 10 or 11 (if “other predatory mammals” is number 11). Additionally, I think that at this point you have to acknowledge that in choosing these independent food proportion variables that this was after preliminary scrutiny of the data, when your analyses were not based on the lowest taxonomic level, and you grouped some into categories. That shouldn’t be a problem for your study but could be worth pointing out in a revision

It was corrected in text. We used 6 birds and 5 mammals categories.

Line 178. Good to see that you used the nest site/territory as the analytical datum basis, rather than pooling data across all sites. I know it will probably be difficult, but did you have sufficient samples to consider year as a variable, or even look at the possibility of annual variation? This is simply a query and not any recommendation for additional analyses/radical change to the manuscript

We really wanted to achieve this. Unfortunately, the data were not as accurate for each revir and year of the study to be able to construct reliable models.

Line 179-180. Once more, you need to be clearer: what is an “a priori” test…? A priori is defined as “relating to or denoting reasoning or knowledge which proceeds from theoretical deduction rather than from observation or experience”

The a priori test was used to make a preliminary assessment of the significance of the influence of individual variables in the GLM. Only the post-hoc test, in turn, indicated the direction and quality of the influence on the dependent variable. In this case, we can agree that it is not necessary and was removed from the text.

Line 181. English again, sorry to say. Drafted?

It was corrected.

Line 190-191. The “respective years” doesn’t make much sense since no years are given to which “respective” could refer to

It was corrected

Line 194. ‘value of’ could be deleted. ‘Was’ not ‘were’

It was corrected

Line 205-206. You say “and from potential feeding grounds” but later go on to say that they could often take “forest” species as prey? Some contradiction perhaps? Why not simply say “open ground” or similar?

It was corrected

Line 228 Table 3. These are pooled results, yes? Be good if so, to make this clear

It was completed

Line 294-295. Sentence doesn’t make sense

It was corrected

Line 315-322. The high proportions of mesopredators in the diet is fascinating. For the northern goshawk Accipiter gentilis, another apex predator (if smaller and in its forest environment where studied) it has been shown that when the subpopulation increased, breeding success declined and presumed availability of more ‘preferred’ prey declined there were more such ‘mesopredators’ (owls, other birds of prey) through “superpredation” (Hoy et al 2017 https://nsojournals.onlinelibrary.wiley.com/doi/abs/10.1111/jav.01387). I am not suggesting that the present study should look at dietary shifts through changes in density nor consider breeding success at particular nest sites (although I would presume the study recorded such data). However, it may be interesting if the authors could perhaps speculate on such a possibility in the Discussion (given their wider knowledge of their study system) when they may be able to pass some comment on this, when they have recorded relatively high proportions of smaller predators being killed through “superpredation”. Just a thought

This issue was included in the discussion

Comments on the Quality of English Language

I have made many comments where the English could be improved. These comments are not exhaustive, however, since to be an editor in this regard is not my role. I would recommend that the English be examined thoroughly and revised where necessary. This does not simply relate to issues like grammar etc, and to some obvious glaring textual errors but to some of the scientific 'terms' used inappropriately or without clarity.

The English language was corrected

Whilst methods are not novel and results not of high international impact, this is an easy to comprehend study with useful information for a national conservation effort. I have a few relatively minor comments on the study design and analysis

1. I would suggest that the nest and random points buffers should not overlap, but this is not specified in the methods. It would be helpful to include the random buffers and the study area boundary in Fig.1

It was not overlapped.

2. Was the age of nest trees estimated from trunk dimensions?

The age of nest trees was given from Forest Data Bank. It was explain.

3. Line 205-208. You says you are reporting averages so don't need to use the word 'about'. I would stick with reporting percentages throughout, so 2/3  would become 67%

It was corrected

4. It would be helpful to give a clearer rational for being interested in latitude and longitude in relation to diet - do you anticipate any patterns? Also, why no interest in altitude? My impression is that you are interested in relationships between diet composition and landscape composition. Perhaps ordination is a statistical approach you could consider. Have you screened your dependent and independent variables for their distribution types - transformations may be requited for any that depart from normality

When analyzing the data, we noticed regional variation. We wanted to see if there were correlations to explain the local variation in prey share. We know that the Golden Eagle can forage farther than the analyzed buffers surrounding the nest. We took into account that analysis of the nest surroundings may not explain the differences. During further analysis, it turned out that the studied nests did not differ significantly in the characteristics of the surrounding landscape, so they could not be explained by general landscape parameters.

Hence, an attempt to explain the perceived differences in the food composition of different pairs of anteaters by a general spatial distribution relationship. Latitude and longitude potentially carried information about local differences in prey availability, types of human activity, and landscape characteristics that were more difficult to describe.

We checked the types of distributions and they were close to normal.

5. You don't given a reference for how prey species were assigned to forest, open and anthropogenic habitats. Can you add a sentence explaining how this was done?

The necessary references have been added

Comments on the Quality of English Language

The English is generally very good, but with occasional odd word usage and sentence construction that detracts a little from it's flow but ultimately has little impact on understanding. I think this could be resolved by an edit from a fluent English speaker

English was corrected by native speaker