GENERAL COMMENT

In this systematic review protocol, the authors present how they plan to conduct a systematic review on the effects of two different forest management practices on biodiversity: even-aged (clear-cutting) and uneven-aged harvesting of forest. The overarching aim is well defined and justified by scientific and societal needs. The authors provide a clear and concise background, further convincing the reader that they target an important policy- and management relevant topic. The manuscript is well written and in general follows the journal's guidelines (but see my point below about the stakeholder engagement section). However, there are a few major concerns with the present version of the manuscript, mostly related to how the sub-questions are framed. The question framing determines a substantial part of the coming review work and the authors need to revise this before starting with the review.

It is not clear for a reader what the aim of subdividing the primary question is. In S1 you will compare clear-cuts with non-intervention, but in S2 you will compare uneven-aged harvesting with both non-intervention and clear-cutting. I do not understand why these are divided into two questions. Please consider reframing these, either into one single question or with a clearer difference (not overlapping) between them. I would consider using "management" as the intervention, compared to non-management or alternative forms of intervention (see comment below). In addition, question S3 is rather unclear and need substantial clarifications. See my detailed comments below for more feedback on this.

There is also a seemingly misunderstanding of the use of Population in the PICO-framework and some vague statements concerning screening that need clarification. I would also like to see some elaboration on the study validity assessment description, and potential challenges in the analytical approach (eg. comparsion of BA and CI studies with the authors' approach) need consideration. I have given some recommendations for the search process, and hope that the authors will make use of those, including a reduced search string. I recommend the authors to have a look at previously published systematic reviews in the journal for inspiration on this, but the authors may also justify their final choice of method by pilot testing, e.g. of the search string. Although several parts needs improvements or clarifications, I am convinced that the authors will manage to revise the protocol according to my comments and believe an improved protocol will lead to a systematic review with valuable contribution to the forest management arena.

BACKGROUND

The background is in general well written and clearly captures the problem and the logic behind the study. However, according to the EEJ's guidelines for systematic review protocols the role of commissioners and stakeholders in the formulation of the question should be described and explained. Hence, the "Stakeholder engagement" section under Methods should be incorporated into the background. Some revisions are needed to not be too specific regarding details, e.g. PICO-terms. Please see other previously published protocols in EEJ for inspiration.

Paragraph 1 p.4 L1-4: First few sentences are very anthropocentric. I suggest you start with a line stating the threat to biodiversity in general. Then have a second sentence stating that although biodiversity is vital for human well-being human actions threaten biodiversity worldwide.

Paragraph 2 p.4 L13: "In the boreal…". Also, please add a proper reference to the first sentence. p.4 L16: Remove "the" from "the long term", and please specify what "declines" (not just with the example, ie. "biodiversity declines" or "declines of biological communities"). Paragraph 3 p.4 L20 - p.1 L4: Why no examples from Norway? Please add or justify why not. Paragraph 4 p.5 L5: regime => regime¬s.

OBJECTIVE

Paragraph 1 Why is Norway not mentioned in the Background, when all other countries are?

QUESTION

Paragraph 2 P.6 L1-2: Please clarify what you mean that you will include data from either intervention or comparator or both. I do not follow the logic behind this. You will need to have a comparator to estimate the effect.

Questions (p. 6-7)

Why are S1 and S2 separated? To me it seems that both questions target an intervention (even- or uneven-aged forestry) and comparing it to non-intervention or alternative forms of interventions. I think you can combine these into one question, otherwise you need to separate them more clearly. From an analytical point of view I can see the point of conceptualising like this, but combining these two into one question does not stop you from analysing the data in line with how you have framed these two questions.

First I thought your main question would be what the difference is between the two management regimes, but as you frame it now it seems more like you are interested in effects of management and how that depends on management regime. In that case, why not just use "forest management" as the intervention, comparing it to non-intervention OR alternative forms of forest management? That is possible, both conceptually and analytically. I would recommend you to consider such an approach, otherwise justify your question framing more clearly.

The before/after description belongs to study validity assessment (which needs some elaboration, see below). This is what you do in S3, although you only specify clear-cutting as the intervention and uneven-aged management is included in the comparator. This needs clarification and/or justification. As it is now it is rather confusing.

In addition, the outcomes in S3 needs specification. What scale are you targeting? And what is the difference from S1-2? Maybe this actually needs specification also in S1 and S2, e.g. that you refer to diversity within vs. outside the forest stands?

METHODS

Move the Stakeholder engagement to background. Might need some revisions though to suit the Background section.

SEARCHING FOR ARTICLES

p. 8: The "population" does not refer primary to the biological communities found within the forests. It is the forests that form the population, of course including the biological communities found within them, but they rather belong to the outcomes. This comment is also valid for the Questions section.

p. 8-9: Including the comparator in the search string seems unnecessary. That means you will find articles on for example bird communities in national parks without any relevance for the study due to lack of intervention in terms of forest management. I suggest you save yourself some work and exclude the comparator terms from the search string.

p. 9: The search string seems to contain a lot of redundant search terms. Why do you not just define the population (boreal forests; including all terms that define that), the interventions and the outcomes? I doubt you will fail to find articles concerning effects on biodiversity if you exclude the organism group specification.

p. 9: I recommend you to use [*] to define open starts or ends of terms in the search string. In addition, it might be worth considering to include "vegetation" in the outcome part of the search string.
Languages

p. 9: Why not conduct searches also in Norwegian? I acknowledge you base the selection on your language skills, but translating the search terms from Swedish to Norwegian seems feasible, as well as understanding the outcome of such a search.

In the last sentence in this paragraph you mention that you will adjust search language based on database content. I guess you mean you restrict this adjustment to your selected languages?

PUBLICATION DATABASES

You can use boolean operators also for google/google scholar. Please consider doing so, and indicate how many of the hits you will screen for relevance? I find it odd to base it on "no longer relevant", which is not a replicable approach.

ARTICLE SCREENING AND STUDY INCLUSION CRITERIA

Please specify what you mean with "screening in agreement". There are various measures of such agreement, I strongly recommend you to use one of those.

Please clarify what you mean with "If there are multiple studies from one study site, they will be appraised as a group to avoid inclusion of duplicate data". I do understand that you want to identify such overlapping studies, but I find that irrelevant for the eligibility criteria section. There might, for example, be one study which qualifies for the review and one not - so why appraise them as a group?

TABLE 2

Please revise your use of "Population".

Does "abundance" in the Outcomes include abundance of individual species? Please clarify, and if so, please motivate the use of individual species as a measure of biodiversity. In addition, defining the outcome as "changes" is not appropriate, as the outcome also applies to the comparator or the state before harvesting. Consider remove "changes".

Please clarify the last sentence of the Study design criteria. I do not understand.

STUDY VALIDITY ASSESSMENT

Please specify more in detail what you refer to with these categories. In addition, "Sample size", "Sampling" and "Replication" are quite overlapping. I suggest you provide more elaborated bullet points with some descriptions on what you will assess as low, medium or high quality. It is rather unclear now. Also - I suggest you include something like "methodological description insufficient", which you then can use for those studies lacking appropriate method description (can apply to several of the other criteria).

Will you use other weighting approaches, or just base it on the general validity assessment?

DATA EXTRACTION STRATEGY

Will you set a limit in terms of years since publication for contacting authors?

Potential effect modifiers and sources of heterogeneity

I suggest adding time since intervention started as an effect modifier in case you will compare regularly harvested forest plots, i.e. how long a patch has been under a certain management regime. I guess that in many cases there are not that many forest generations to consider, but still.

DATA SYNTHESIS AND PRESENTATION

You state that you will consider different types of meta-analyses. However, in case you have no data to conduct proper meta-analysis you might aim for alternative analytical approaches that do not count as a meta-analysis. I see no fundamental problem in this statement for the protocol though, but keep that in mind.

I do not understand why you will not consider meta-analytical approaches for question S3. Please justify and clarify.

In addition, I think your review would benefit from defining, at this stage, how you aim to solve the potential problem with combining BA, CI and BACI designs. It might not be appropriate to combine these measures if a CI study compares pristine forest with a clear-cut and a BA study compares the biodiversity state before and after harvesting in a regularly harvested forest. I think this needs some consideration before conducting these kind of analyses.

REFERENCES

Please check year of publication and other information in the reference list. For example, there is no publication year for reference 1-4, 7, 15, 17-21, and missing full names in some references (eg. 13, 15, 16). In general, check format of the references. I do not think "April" is necessary for reference 5.

ANNEX 2 Please check reference format, first name initial(s) should come after family name.