Next Article in Journal
Economic Modelling of Poplar Short Rotation Coppice Plantations in Hungary
Previous Article in Journal
Stratification, Scarification and Application of Phytohormones Promote Dormancy Breaking and Germination of Pelleted Scots Pine (Pinus sylvestris L.) Seeds
 
 
Article
Peer-Review Record

Mid-Scale Drivers of Variability in Dry Mixed-Conifer Forests of the Mogollon Rim, Arizona

Forests 2021, 12(5), 622; https://doi.org/10.3390/f12050622
by Matthew Jaquette 1, Andrew J. Sánchez Meador 1,*, David W. Huffman 2 and Matthew A. Bowker 3
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Forests 2021, 12(5), 622; https://doi.org/10.3390/f12050622
Submission received: 6 April 2021 / Revised: 28 April 2021 / Accepted: 12 May 2021 / Published: 14 May 2021
(This article belongs to the Section Forest Ecology and Management)

Round 1

Reviewer 1 Report

The manuscript „Drivers of variability in dry mixed-conifer forests of the 2 Mogollon Rim, Arizona” reports the results of investigating historical variability and drivers of variability in dry mixed-conifer forests on the Mogollon Rim in northern Arizona. Excellent work on a highly relevant topic. The paper can consider to be of general and international interest. The authors clearly state the aim of the paper and present data, approach and methods in great detail and with relevant literature cited. Results are presented interesting and in detail.

Nevertheless, I would suggest to the author to improve the manuscript quality:

  • The manuscript is too long. It is 36 pages long, with two annexes.
  • The pictures should be better quality. In some places it is difficult to see notes.

Author Response

Comments and Suggestions for Authors

The manuscript „Drivers of variability in dry mixed-conifer forests of the 2 Mogollon Rim, Arizona” reports the results of investigating historical variability and drivers of variability in dry mixed-conifer forests on the Mogollon Rim in northern Arizona. Excellent work on a highly relevant topic. The paper can consider to be of general and international interest. The authors clearly state the aim of the paper and present data, approach and methods in great detail and with relevant literature cited. Results are presented interesting and in detail.

Nevertheless, I would suggest to the author to improve the manuscript quality:

The manuscript is too long. It is 36 pages long, with two annexes.

Response/Action: We have further edited the manuscript to ensure the text is concise yet comprehensive, with the majority of our edits shortening the introduction and discussion. While some edits were made to the methods and results, we choose to retain as much experimental detail as needed so that the methods and results can be reproduced. Where possible, sections have been shortened and material not vital to the study or necessary for context or discussion have been removed.

The pictures should be better quality. In some places it is difficult to see notes.

Response/Action: Figured were edited and enlarged for visual quality. Specifically, 8 and 9 had their annotations redone and were slightly enlarged (given page width limitations) for clarity.

Reviewer 2 Report

general comments:

In this paper, authors used a large data set of forest survey and reconstructed historical forest conditions. Contemporary and historical forest conditions were compared in terms of average size of trees, basal area, tree density, species composition, spatial autocorrelations, direct and indirect effects of topographic, climatic, and edaphic conditions on forest composition and structure. The data used in this study were large enough to describe the change in forest conditions at the regional level, and the methods used in the study were appropriate. The conclusions that the forest conditions have changed much from 1879 to 2014, and the changes were caused by fire regime disruption are acceptable.

One drawback on the method of reconstruction of historical forest conditions is that "this method of reconstructing forest structure may fail to detect some small trees that died and decomposed prior to the contemporary surveys (L. 169-170)”. Indeed, the reconstructed forest structures did include less small-sized trees than did contemporary forests. This may lead readers to a concern that the change in forest structure reported in this study is an artefact. I think that some discussion of this concern may be added.

 

 

specific comments:

L.120: Please express the length as a value in km unit.

 

Table A1 title: What does "potential" mean?

Is the unit for ppt correct?

 

L.193-194: "complete spatial randomness" is misleading for a term for not-autocorrelated spatial distributions of the values.

 

L.334-336: Please rephrase a sentence beginning with "We found...".

 

Figure 5: It is mentioned that EIV was calculated for each species in each PLOT (L. 174-176), but there is only one set of EIVs in Figure 5. Is this one example showing the result in one plot? Were EIVs in Figure 5 calculated for entire plots?

 

L.397: "random" may not be appropriate.

 

L.539: "indicate" -> "indicated"?

Author Response

Comments and Suggestions for Authors

General comments:

In this paper, authors used a large data set of forest survey and reconstructed historical forest conditions. Contemporary and historical forest conditions were compared in terms of average size of trees, basal area, tree density, species composition, spatial autocorrelations, direct and indirect effects of topographic, climatic, and edaphic conditions on forest composition and structure. The data used in this study were large enough to describe the change in forest conditions at the regional level, and the methods used in the study were appropriate. The conclusions that the forest conditions have changed much from 1879 to 2014, and the changes were caused by fire regime disruption are acceptable.

One drawback on the method of reconstruction of historical forest conditions is that "this method of reconstructing forest structure may fail to detect some small trees that died and decomposed prior to the contemporary surveys (L. 169-170)”. Indeed, the reconstructed forest structures did include less small-sized trees than did contemporary forests. This may lead readers to a concern that the change in forest structure reported in this study is an artefact. I think that some discussion of this concern may be added.

Response/Action: The reviewer is correct; however, we state “While this method of reconstructing forest structure may fail to detect some small trees that died and decomposed prior to the contemporary surveys, comparisons to historical surveys indicate that 91 to 94% of pre-settlement trees can be identified by contemporary surveys in areas lacking recent disturbance [48,53].” These methods have been used in numerous previous studies across a variety of forest conditions and are known to produce acceptable results with respectable detection rates. The reconstruction techniques are largely considered an “accepted” approach. In the Discussion, we bring the reviewer’s attention to the 300% increase in mean tree density (mean trees ha-1 increased from 165 to 657 during the observation period) which is too large to be the result of omission errors in our reconstruction model, which have had detection rates documented at or near 90%.

specific comments:

L.120: Please express the length as a value in km unit.

Response/Action: Converted.

Table A1 title: What does "potential" mean?

Response/Action: In the methods we explain that each latent or composite variable could have been represented by many measures, so we compiled a large pool of potential explanatory variables to select from (see Table A1 for a complete, detailed list of explanatory variables).” Rewritten for clarity and the caption for Table A1 now reads “Summary of all environmental variables considered for inclusion in models.”.

Is the unit for ppt correct?

Response/Action: It is in millimeters, but we can see how this was confusing and have removed the preceding “0.1” designation form each unit

L.193-194: "complete spatial randomness" is misleading for a term for not-autocorrelated spatial distributions of the values.

Response/Action: Rewritten to read “simulations of expected values given no significant auto-correlation to evaluate significance.”

L.334-336: Please rephrase a sentence beginning with "We found...".

Response/Action: Rewritten for clarity. This sentence is now two sentences which read: “We found multicollinearity between the explanatory variables of ‘temperature’ and ‘water’ which may have caused path coefficient inflation in the model. This issue was resolved by combining these two factors into a single ‘climate’ composite.”

Figure 5: It is mentioned that EIV was calculated for each species in each PLOT (L. 174-176), but there is only one set of EIVs in Figure 5. Is this one example showing the result in one plot? Were EIVs in Figure 5 calculated for entire plots?

Response/Action: Figure caption edited for clarity and it now specifies “across all plots combined.”

L.397: "random" may not be appropriate.

Response/Action: Rewritten for clarity. The sentence now reads “Interestingly, density was significantly negatively autocorrelated at 810m, but was other-wise not significantly autocorrelated at all other lags.

L.539: "indicate" -> "indicated"?

Response/Action: Corrected.

Reviewer 3 Report

  1. If I understood clearly, the previous article (titled “Reference conditions and historical fine-scale spatial dynamics in a dry mixed-conifer forest, Arizona, USA”) have shown most of the forest structure or composition in the two specific years 1879 and 2014 in the study region. The drivers of such variability should be the new insights from the current one. On the one hand, I want to know or recommend the authors to add detailed text for the updated knowledge of forest structure and composition in the two years, compared to the previous published article. On the other hand, the analyzed topography, climate and soil may do not significantly change in the study region. I am therefore wondering why such drivers are involved in the investigation. Actually, the most different factor is the fire disruption after 1879 as low-severity fires burned frequently with a mean fire interval of 2 to 8.5 years before that. The forests are currently the target of restoration efforts to increase ecological resilience and protect important municipal water supplies. So, the influence of forest management is the added factor. Have the authors ever considered the two most important different drivers? Definitely, any coefficients can happen or exist between the selected climate data (or topography or soil) and forest structure or composition included in the historical and contemporary models (as shown in Figs.7-9). How about the uninvolved factors? I mean, are there some criteria for the selection?
  2. From the title, I could expect that clear information can be obtained for the drivers of variability in dry mixed-conifer forests of the Mogollon Rim, Arizona. However, after I read it, the drivers are still not directly demonstrated. The generalized statements are useless. Actually, about half of the manuscript is referring characteristics of forest structure and composition in the two years 1879 and 2014. So, maybe it’s better to change the title as something like “Forest structure and composition in two specific years and their possible drivers in the Mogollon Rim, Arizona”.
  3. Information in the lines 27-28 “Managers can utilize this increased understanding of variation to tailor silvicultural prescriptions to environmental templates” is not clearly. Please re-edit it. I could not gain the asserted “increased understanding” probably because I am not expertized in the forest management field.  
  4. Meteorological climate data for the historical period 1895-1924 is unavailable, but, I think that for the contemporary period 1981-2010 is possible or at least available from the nearby sites. Considering climate data were acquired from the Parameter-elevation Regressions on Independent Slopes Model (PRISM), I am curious how about the confidence of the modeled data. Here, I am also care about the effect of some possible extreme events, e.g., severe drought, to the structure or composition during the two specific periods.  
  5. I always felt that the involved content of this manuscript is much. The Materials and Methods section is much longer than the other articles, which can reduce the overall interest or visual fatigue. Some parts are redundancy. I therefore suggest deleting the unnecessary sentences or reorganizing them.
  6. What’s the “ctly” in line 237 mean?
  7. Some of the coefficients embedded in Figs.8 and 9 are hard for reading.
  8. It’s of course not so comfortable or readable to separate one word into two lines, please change the format of the paragraph into justify throughout the full text.
  9. The figure number of the contemporary structural equation model is Fig.9, but not Figure 4 as shown in line 486.
  10. Line 725 should be changed into “As noted in our methods”.
  11. The reference list is much more for one research article, normally 60 is enough. Please delete some of the unnecessary ones.

Author Response

Comments and Suggestions for Authors

  1. If I understood clearly, the previous article (titled “Reference conditions and historical fine-scale spatial dynamics in a dry mixed-conifer forest, Arizona, USA”) have shown most of the forest structure or composition in the two specific years 1879 and 2014 in the study region. The drivers of such variability should be the new insights from the current one. On the one hand, I want to know or recommend the authors to add detailed text for the updated knowledge of forest structure and composition in the two years, compared to the previous published article. On the other hand, the analyzed topography, climate and soil may do not significantly change in the study region. I am therefore wondering why such drivers are involved in the investigation. Actually, the most different factor is the fire disruption after 1879 as low-severity fires burned frequently with a mean fire interval of 2 to 8.5 years before that. The forests are currently the target of restoration efforts to increase ecological resilience and protect important municipal water supplies. So, the influence of forest management is the added factor. Have the authors ever considered the two most important different drivers? Definitely, any coefficients can happen or exist between the selected climate data (or topography or soil) and forest structure or composition included in the historical and contemporary models (as shown in Figs.7-9). How about the uninvolved factors? I mean, are there some criteria for the selection?

Response/Action: We understand that this is a complex issue and we appreciate the reviewer’s perspective. The referenced study (Rodman et al. 2016) was based on four 1-ha stem-mapped plots, and while these plots were located near our study site, they do not provide a robust sample for describing mid-scale (a phrase that as added to the title for clarity) attributes, and they cannot fully describe the variety of historical and contemporary conditions in mixed-conifer forest of the Mogollon Rim. Additionally, the specific dates were selected to represent the historical conditions prior to degradation (i.e., a period immediately prior to fire exclusion) and contemporarily (i.e., when our data were collected). The authors feel that Rodman et al’s work (n=4) provided only limited information on variability in condition, composition, and structure present at the two time periods, in contrast to our more comprehensive work (n=270). As the reviewer points out, topography and soils have likely changed little and we used time-period specific climate information; all within a constrained yet plausible model form. Paired with this knowledge and the documented lack of recent active forest management in these ecosystems, the most plausible cause of observed differences between the historical and contemporary models was fire exclusion. We discuss this thoroughly throughout the manuscript. Additionally, while it may be possible to explore unrecorded or uninvolved factors (e.g., latent variables in the SEM framework) our focus was on investigating the influence of important and measurable environmental factors commonly used by managers. We do this in a framework that allows one to quantify strengths of relationships. We feel the manuscript provides adequate clarity on these issues.  

  1. From the title, I could expect that clear information can be obtained for the drivers of variability in dry mixed-conifer forests of the Mogollon Rim, Arizona. However, after I read it, the drivers are still not directly demonstrated. The generalized statements are useless. Actually, about half of the manuscript is referring characteristics of forest structure and composition in the two years 1879 and 2014. So, maybe it’s better to change the title as something like “Forest structure and composition in two specific years and their possible drivers in the Mogollon Rim, Arizona”.

Response/Action: Our objectives were “(1) What were the historical structural conditions warm/dry mixed-conifer forests on the Mogollon Rim? (2) How did historical structural conditions vary spatially, and has spatial variation changed since fire exclusion? and (3) What were the drivers of variability in historical dry mixed-conifer forests of the Southwest, and how have they changed con-temporarily in relative importance?” It should be clear to the reader that before one can explore drivers of variability, one must develop clear baselines for comparison. Therefore, extensive description of conditions is necessary for context. Furthermore, SEM is a framework that allows one to quantify strengths (i.e., the relative importance) of relationships in a model framework. We disagree that the generalized statements are useless and argue that they are the primary insights given the study design and the methods used. The manuscript describes forest conditions (and their variability) at two points in time (i.e., one that had developed over at least several centuries with frequent fire, and one reflecting a century of fire absence), and then explores the relative importance of a fixed set of key environmental factors important for making management decision and designing restoration treatments. Lastly, the observed path coefficients do indicate drivers of spatial, structural and compositional variability. We disagree with the reviewer’s suggestion to change the manuscript title as suggested and choose to keep it (with a slight modification – adding “mid-scale” for clarity on a previous comment) as written.

  1. Information in the lines 27-28 “Managers can utilize this increased understanding of variation to tailor silvicultural prescriptions to environmental templates” is not clearly. Please re-edit it. I could not gain the asserted “increased understanding” probably because I am not expertized in the forest management field.

Response/Action: Thank you for this clarification. Managers often need site specific information when designing restoration treatments, and they commonly use information like topography, soils, and climate to vary prescriptions and to develop targets with respect to desired structure and composition. With an increased understanding of the relative influence environmental factors both historically and contemporarily, managers will be better suited to develop treatments that ensure forest are resilience to disturbance. We clarified and edited the Discussion to better explain this issue.

  1. Meteorological climate data for the historical period 1895-1924 is unavailable, but, I think that for the contemporary period 1981-2010 is possible or at least available from the nearby sites. Considering climate data were acquired from the Parameter-elevation Regressions on Independent Slopes Model (PRISM), I am curious how about the confidence of the modeled data. Here, I am also care about the effect of some possible extreme events, e.g., severe drought, to the structure or composition during the two specific periods.

Response/Action: The PRISM data utilized were modeled using climatologically-aided interpolations, which assume the long-term average pattern (i.e., the 30-year normals) as first-guess of the spatial pattern of climatic conditions for a given month or day. These interpolations are robust to wide variations in station data density - a necessary aspect when modeling long time series. To further eliminate the influence of extreme events (as indicated on page 7 lines 267-270) we used 30-year normal for the two periods (1895-1924 to represent historical climate in 1879 and 1981-2010 to represent contemporary climate in 2014). While evaluating the accuracy of these data is beyond the scope of our work, more information on the subject (as cited) can be found here https://prism.oregonstate.edu/documents/PRISM_datasets.pdf

I always felt that the involved content of this manuscript is much. The Materials and Methods section is much longer than the other articles, which can reduce the overall interest or visual fatigue. Some parts are redundancy. I therefore suggest deleting the unnecessary sentences or reorganizing them.

Response/Action: We have further edited the manuscript to ensure the text is concise yet comprehensive, with the majority of our edits shortening the introduction and discussion. While some edits were made to the methods and results, we choose to retain as much experimental detail as needed so that the methods and results can be reproduced. Where possible, sections have been shortened and material not vital to the study or necessary for context or discussion have been removed. Please see response(s) above.

  1. What’s the “ctly” in line 237 mean?

Response/Action: This is an issue with the rendered pdf. The words “directly” and “indirectly” are correct as written.

  1. Some of the coefficients embedded in Figs.8 and 9 are hard for reading.

Response/Action: Figure edited for clarity.

  1. It’s of course not so comfortable or readable to separate one word into two lines, please change the format of the paragraph into justify throughout the full text.

This is a function of MDPI’s template and their formatting standard. Therefore, this is beyond our control.

  1. The figure number of the contemporary structural equation model is Fig.9, but not Figure 4 as shown in line 486.

Response/Action: Thank you for catching this important typo. Corrected.

  1. Line 725 should be changed into “As noted in our methods”.

Response/Action: Corrected.

  1. The reference list is much more for one research article, normally 60 is enough. Please delete some of the unnecessary ones.

Response/Action: Where possible, unnecessary references were removed.

Back to TopTop