Managing Moose from Home: Determining Landscape Carrying Capacity for Alces alces Using Remote Sensing

Round 1

Reviewer 1 Report

Notes:

 

General notes:

 

Overall this was a well-written paper. I noticed very few grammar issues, and they are identified by line number below. I recognize that this work would have required an extensive amount of effort, and I like the general idea of being able to use remote sensing data to estimate population carrying capacity. At the same time, I do have a few more questions after reading the manuscript several times.

 

1. Why Allegheny area with no moose?

 

It seems like the Pennsylvania study area was put in as an afterthought, or perhaps because it was processed in the same way as the NY study area. If the objective is to estimate moose NCC, they why include a study area that has no moose?

 

2. How handle high elevation mountain areas?

 

I am not familiar with NY (AP), but it seems like there should be some high-elevation areas that have no moose habitat value. It would be good to describe how those were handled, or make it clear that moose can live on mountain-tops.

 

3. Not clear how landscape cover data evaluated

 

I am a little concerned with how the landscape cover data were evaluated. As described, it appears to me that data were taken to be correct, but there does not appear to be documentation of who data was obtained from, specific layers, how it was merged. E.g., what timber companies provided harvest/age class data, how much of the landscape was this, etc.

 

4. NCC as a term. I originally thought Mautz was the first to use the term, now I am not so sure. He did the fat cycle in deer paper in WSB, but didn’t use NCC. Hanley or Hobbs?

 

5. AUD calculations. I have not seen the Peterson M.S. thesis that was used to provide the NCC values for each cover type. Given that it is an unpublished M.S. thesis, it might be worthwhile to expand on why the NCC of so many forest types was identified as 0, especially in an aging forest you would expect to see canopy gap formation, for example.

 

Overall, I think this paper represents a good approach to data synthesis, I guess I would just like to see a bit more supporting information for the conclusions that are drawn. I can’t comment on the random forest statistical approach.

 

Specific line number comments:

 

35        This has been previously called nutritional carrying capacity – Mautz? First used in WSB. Then Hanley, Hobbs, others.

 

62        includes (not include)

 

74        I’m not so sure I would apply the term “rapid” to this technique. It would have taken hours to identify and process landsat scenes across decades.

 

85        Also earlier, but it is not clear to me why an area with no moose is included as a study area.

 

106       Dependent on

 

166/7    Awkward phrasing

 

193       as _a_ predictor variable

 

196       Need more details on the training data – how acquired, how verified, etc.

 

223       How much did the assumption that lands in public ownership > 20 years were mature affect results? Did they conflict with the NDVI approach or enhance (or reduce variability by forcing about half of landscape to have an NCC of 0?)

 

253       Figure 3 doesn’t seem to add to much in terms of explaining

 

259       Estimate of moose density based on Reference 39. No AUD’s in conifer, lowland / mixed, or wetlands.

 

539       Same as 550 – check for others.

550       Should have book citation with editors, no indication this is just a chapter.

552       Same as 550

 

Author Response

Reviewer 1:

I would like to thank you for reviewing the manuscript. I found your comments to be constructive and helpful. I have addressed the majority of your concerns. Please see a more detailed response below to your non-structural comments (i.e. misspellings, punctuation). See my comments in red:

1. Why Allegheny area with no moose?

 It seems like the Pennsylvania study area was put in as an afterthought, or perhaps because it was processed in the same way as the NY study area. If the objective is to estimate moose NCC, they why include a study area that has no moose?

 The Pennsylvania study site was used to pilot the remote sensing method of identifying timber harvest. I thought it appropriate to include in that it applies the methodology to two different locations, one northern hardwood forest and one mixed hardwood/balsam conifer forest. It allows us to show that the methods are applicable across different forest types, rather than just hardwoods. We initially calculated K for deer in the Allegheny, but the values used for landcover estimates were not as thoroughly vetted as there were for moose in the ANF, therefore we thought it inappropriate to include in a peer-reviewed publication. I added some more info to address this.

1. How handle high elevation mountain areas?

I am not familiar with NY (AP), but it seems like there should be some high-elevation areas that have no moose habitat value. It would be good to describe how those were handled, or make it clear that moose can live on mountain-tops.

Those areas were classified by NLCD as barren lands. Additionally, all of the areas of high elevation are public lands, so we know that they have not been harvested in 20+ years so we excluded those from the classification. I can add in this caveat in the methods.

1. Not clear how landscape cover data evaluated

I am a little concerned with how the landscape cover data were evaluated. As described, it appears to me that data were taken to be correct, but there does not appear to be documentation of who data was obtained from, specific layers, how it was merged. E.g., what timber companies provided harvest/age class data, how much of the landscape was this, etc.

I can clarify this in the methods that timber harvest data was provided by private companies. That data was then vetted using aerial imagery and ground trothing. 

1. NCC as a term. I originally thought Mautz was the first to use the term, now I am not so sure. He did the fat cycle in deer paper in WSB, but didn’t use NCC. Hanley or Hobbs?

 I added a more appropriate citation of Hobbs and Schwartz

1. AUD calculations. I have not seen the Peterson M.S. thesis that was used to provide the NCC values for each cover type. Given that it is an unpublished M.S. thesis, it might be worthwhile to expand on why the NCC of so many forest types was identified as 0, especially in an aging forest you would expect to see canopy gap formation, for example.

The Peterson paper is current in review. Ideally, I would provide a more adequate citation but I don’t want to re-invent the wheel and put the methodology out there twice. The methods themselves are rather complicated and I don’t want to water them down here.

 

Thank you again!!

Author Response File: Author Response.docx

Reviewer 2 Report

Overall, I enjoyed this paper very much, but was repeatedly confused as to why the PA study area was included in this project.  There are places where PA and Adirondack Park are called case studies, but I never saw a compelling reason to the inclusion of the PA study area.  There is some mention in the Discussion (~lines 340-342) that may explain the inclusion of PA, but it needs to be much more explicit.  Is PA historic moose range?  Are they proposing the Allegheny National Forest in PA could be used for moose habitat?  Was the PA study are just used for training data?

Specific comments:

Introduction

Lines 44-47: great and applicable!

There should be some mention of the other effects timber harvest may have on moose...increased deer abundance? Increased or decreased gastropod abundance?  Impacts on winter ticks?  Also, why are PA and NY study sites both included?

Study sites

Minor: use superscript in the square kilometer notation.

Sections 2.2.1 and 2.2.2

Please explain why 2 different methods were used.  Was it due to data availability?  And not to belabor the point, but why is the Allegheny site being used?

2.3 Remote Sensing Classification

Is this for PA only?  Equation for NDVI is missing a parenthesis.

Throughout the paper, Adirondack Park is variously referred to being in New York, NY, NE NY, northcentral New York, etc.  Please standardize.

Table 4.  This is the first place I am seeing that the 2 study areas are case studies. 

Table 5. Instead of labeling the 2 groups as simply A and B, please name them as AP and ANF.

In the discussion, I think there needs to be more care in addressing how timber management can impact moose...not simply increasing forage.

In the conclusion, there is a sentence (“It is even possible...”) that is intriguing, but does not seem to be supported by the results (or even mentioned in the results or discussion).  Please elaborate.

Author Response

Reviewer 2:

I would like to thank you for reviewing the manuscript. I found your comments to be constructive and helpful. I have addressed the majority of your concerns. Please see a more detailed response below to your non-structural comments (i.e. misspellings, punctuation). See my comments in red:

There should be some mention of the other effects timber harvest may have on moose...increased deer abundance? Increased or decreased gastropod abundance?  Impacts on winter ticks?  Also, why are PA and NY study sites both included?

 The Pennsylvania study site was used to pilot the remote sensing method of identifying timber harvest. I thought it appropriate to include in that it applies the methodology to two different locations, one northern hardwood forest and one mixed hardwood/balsam conifer forest. It allows us to show that the methods are applicable across different forest types, rather than just hardwoods. We initially calculated K for deer in the Allegheny, but the values used for landcover estimates were not as thoroughly vetted as there were for moose in the ANF, therefore we thought it inappropriate to include in a peer-reviewed publication. I added some more info to address this.

Is this for PA only?  Equation for NDVI is missing a parenthesis.

Fixed the equation and added a statement that this applies to both.

Table 4.  This is the first place I am seeing that the 2 study areas are case studies. 

Table 5. Instead of labeling the 2 groups as simply A and B, please name them as AP and ANF.

Fixed both tables.

In the discussion, I think there needs to be more care in addressing how timber management can impact moose...not simply increasing forage.

We decided not to add much more in that this quickly delves into issues of winter tick, etc. These issues are not currently prevalent in NY given the low moose densities and I don’t want to take away from the focus, which is more about the method itself and less about the outcome.

 

Thank you!!

Author Response File: Author Response.docx