Species Diversity Associated with Foundation Species in Temperate and Tropical Forests

Round 1

Reviewer 1 Report

This paper proposes a method for finding foundation tree species, using five large mapped plots in temperate and tropical forests, and starting with a known foundation species on a very familiar plot at Harvard Forest. Overall, this paper is well written, the analyses were nicely done (and a lot of work), and they present a good advance for an important concept in forest ecology that is inherently hard to deal with.  

The validity of the statistical fingerprints are incredibly tricky to evaluate.  I was skeptical at first, but then came to the conclusion that they probably do work and that the ideas presented in the paper will provoke discussion and further evaluation of the foundation species concept and its relationship with forest dynamics concepts and species traits. Following are some ideas for the authors to think about while they refine the paper.  

Statistical fingerprints of foundation species.  I agree that hemlock is a foundation species, however, under conditions of low deer abundance and multi-aged forests (conditions which were more common given the disturbance regime and deer abundances prior to European settlement), it would fall within the reverse J pattern among species, since there would be more sapling-sized trees (1-5 cm dbh) given its high tolerance for shade. The fact that it does not fit that pattern today seems likely to be an artifact of the state of landscape conditions at this point in time. In other words, problems with reproduction on the contemporary landscape could make hemlock more likely to fall outside the reverse-J pattern among species. It could also be that the fingerprint includes shade-tolerant species that fall in the high-density region of the plot, whether or not they are outliers, as well as intermediate-tolerant species that are more likely to be outliers of the reverse-J distribution.  I did appreciate the discussion on lines 352-360 that partly acknowledge the complexities of using this particular fingerprint.

The second statistical fingerprint for a candidate species is having a broader distribution of subplots on the basal area versus number of individuals plot, essentially showing a number of subplots with both high numbers of individuals and basal area, than the other candidate species.  However, given their intermediate shade tolerance, red oak, red maple and white pine, are very likely to display either a lot of small trees (on subplots that had a recent disturbance) or a few large ones (on subplots that had a disturbance a long time ago) on a given subplot most of the time due to self-thinning regimes and their inability to regenerate in their own shade in the absence of continued disturbance. At first I questioned whether this fingerprint metric was really just separating high shade tolerance versus intermediate shade tolerance. On second thought, however, I noted that with intermediate disturbance regimes (moderately frequent disturbances of moderate size and severity, e.g. small understory fires or multi-tree blowdowns), these species could easily also fall within the reverse-J pattern for the first statistical fingerprint and have subplots with both high basal area and high numbers of individuals (second fingerprint) as well. Therefore, whether species fall within the reverse J pattern seems more likely to be an indicator of whether the disturbance regime is aligned with the disturbance adaptation traits of the species, than whether they are foundation species.  As a third thought, I noted that the way disturbance regimes work in a landscape where hemlock dominance is supported by the climate and soils, such an intermediate disturbance regime is unlikely to occur without knocking hemlock out of the dominant position, and therefore the statistical fingerprint does work, but only in the context of this landscape (with some nuances regarding the factors affecting hemlock reproduction noted above), so the fingerprints are context dependent. Hemlock partly controls the disturbance regime so that its characteristics work in concert with the species’ traits, which I guess makes for even stronger evidence in favor of hemlock as a foundation species.

The codispersion analyses seem well founded. I appreciate the authors explanations of how this works in the discussion section-essentially it seems that the foundation species create alternate local environments that support different communities of species smaller species that depend on the foundation species, while at the same time the foundation species also influence the environment of the entire ecosystem. However, the fact that foundation species are clumped, means that non-foundation species are also clumped and the areas they dominate probably also support some species that would otherwise not occur in the forest. This is one of the ambiguities in the foundation species concept.

Author Response

Please see attached pdf

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors investigated the species diversity associated with „foundation species” in three temperate forests and in other three tropical forests. The analyzed plots are six very large forest, established continually since 1980 and re-surveyed at around 5 years intervals. The Ms is very well written, very well documented and well organized and the statistical analysis seems to be done properly. It is a very interesting paper and suitable for the scope of FORESTS journal.
I have just several specific comments/suggestions that could clarify more some aspects I found not very understable for the general reader.

P6. L.147. I don’t understand what means “the site PIs”. Please add more information/detail to do it clearer.

P6. L.149. Not clear what means this “outlier”. Please explain more this term (is there any test to classify one value as an outlier?)

P7. Figure 4. It is not clear for me why several points were not considered as part of the “reverse J” distribution (e.g. in TY site the point having a total number >10000 and mean diameter around 3-4m, or 5000 individuals vs. 5 m dbh, or in LFDP the point from upper left part – 11000 individuals vs. 5-6 m dbh, or in BCI the point 3800 individuals or 2800 for 6-10 cm dbh).

P7. Figure 4. Legend. What means “important” species? Please either detail or replace it by other more term word.

P8. Figure 5. Legend. Again “important”. Please be clearer.

P9. Table 1. Caption of the table. Again what is the site “PIs”.

P9. L.160. Please add a reference for the Bray-Curtis dissimilarity.

FOR all Figures please use a larger font for the text of diversity measures, or sites….

Author Response

please see attached pdf

Author Response File: Author Response.pdf