Multi-Century Reconstruction of Pandora Moth Outbreaks at the Warmest/Driest Edge of a Wide-Ranging Pinus Species

Round 1

Reviewer 1 Report

 

This is an excellent paper that brings in many new techniques to examine insect outbreaks. It provides an important analysis of pandora moth outbreaks on the Kaibab Plateau.

Lines 48-52: I appreciate the focus on the interaction of multiple disturbances and especially the interaction between pandora moth and fire. This was proposed by Speer 1997 with a graphic from this finally published in Speer and Kulakowski 2017. This works proposes the hypothesis that pandora moth is controlling the fire regime by removing fine needle fall for a ten-year period resulting in less fire spread. 

Line 53: Large and small scale are oppositely defined in geography and ecology, leading to confusion in biogeography that intersects these two disciplines. It is more clear to use broad and fine scale.

Lines 53-67: I agree that this is the technique that we always use, but it may mean that we are unable to detect pandora moth outbreaks during droughts. Possibly if the outbreak is extreme, it might be the most limiting factor and show a more negative change than drought does. But if drought is the dominant limiting factor and a pandora moth outbreak occurs, we may not detect it. I don't have a solution to this, but it has been bothering me for a while.

Lines 151-154: I am not asking you to change your standardization, but I would now go with a longer standardization curve especially with older trees to maintain as much signal as possible at the 40 year time period since we have documented return intervals in this range. We are potentially removing half of the 40-year signal with this standardization. Currently, I would use an Age-Dependent Spline that Ed Cook recommends by making it default in the current versions of ARSTAN.

Lines 180-191: I appreciate the complication of other insect outbreaks in PSME and like your solution of using a multi species non-host control. Future work in North American could draw from Ed Cook's NADA and even select all other species rather than the host. This broad scale analysis is likely to be a better measure of climate than any one species non-host chronology.

Figure 2: I appreciate this analysis comparing the host and non-host climate responses at the 90% confidence intervals. Too often in the past, we would just show their responses and declare that they are similar.

Figure 3: What do you think happened with the most recent outbreak where it is not recorded well with the precip or the PSME data. It is too bad that your PIED chronology does not extend through the modern outbreak. This is probably part of your hypothesis of lower pandora moth damage during the fire suppression period.

Also, can you speak to the discrepancy between the PSME non-host and the PIED non-host records in the 1700s. Do you think that Spruce Budworm outbreaks back in time, might have complicated the reconstruction?

Figure 4: It is interesting that in each of your images, you have a large ring the year after what you declare is the start of your outbreak. My observations in Oregon always had the first year half the size of normal, followed by the two smallest rings in the signature. If 1748, 1877, and 1977 are the true start to your outbreaks, I wonder if your trees are releasing the year after first major defoliation as an induced response to try to outgrow the insect. Is it also possible that the start should be 1979, 1879, and 1750? Your documented start of the outbreak is 1978 for the modern outbreak, and there is often a growth in the population necessary before the trees seem to record the outbreak. The start of my signatures seemed to coincide more to the peak of the outbreak rather than the start of the outbreak.

Lines 327-328: I don't understand how using precipitation as a non-host could result in severe droughts being mislabeled as outbreaks. Severe droughts should be removed as possible pandora moth events because it is being used as a non-host.

Lines 413-424: Speer and Kulakowski 2017 argue that pandora moth probably shares a simbiotic relationship with poderosa pine forests because pandora moth can suppress fire for ten years during which ponderosa pine can establish and grow large enough to survive the next fire. I appreciate your more ecologically balanced stance rather than viewing insects as pests.

With Kristen De Graauw’s masters thesis, I am surprised that we did not reconstruct pandora moth outbreaks on the Kaibab Plateau. I remember that we talked about this and we were using a standard technique to reconstruct outbreaks from ITRDB chronologies without the original wood. Because of that we took a fairly conservative approach to the data, which I believe is why Kristen did not document outbreaks there despite historical documentation of outbreaks in this area.

Specific Comments

Line 61: …attributed TO other factors…

Line 227: There is an extra period at the end of Figure 2 caption.

Line 412: …pandora moths WILL affect forests under future conditions.

References

Speer, J.H., 1997. A dendrochronological record of pandora moth (Coloradia pandora, Blake) outbreaks in central Oregon (Doctoral dissertation, The University of Arizona).

Speer, J.H. and Kulakowski, D., 2017. Creating a buzz: insect outbreaks and disturbance interactions. Dendroecology: tree-ring analyses applied to ecological studies, pp.231-255.

Nice work!

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear colleagues,

I have read your manuscript with great pleasure. It is absolutely classical, methodically correct and well-elaborated research. However, I should point out some aspects, which can make your reasoning better. 

First, the parameter values of outbreak identification (4 years of duration, -1.28 GSI decline and 25% proportion of defoliated trees) looks arbitrary. At any rate the reasonableness of these values have not justified in the text. R language allow us examine the set of values combinations for searching better one (e.g., https://doi.org/10.3390/insects12020090) like in machine learning. If you change the values of parameters, you possibly can achieve reconstruction that is more precise.

Second, you have used the term ‘outbreak duration’ (for example, in the lines 267–268). However, in fact, you deal with a length of narrow-ring parts of core or chronology, not with outbreak duration indeed. The defoliation described by J.M. Schmidt and D.D. Bennett, for example, continued no longer than five years (1979–1983), but not nine as in Table 2. The late portion of such parts had appeared after outbreak cessation due to prolonged effect of defoliation (lack of assimilates). In addition, the duration of periods with reduced radial growth could be affected by additional stresses (drought, competition etc.), which impact on rings width is difficult to be assessed.  

At last but not least, the Discussion could be more ecologically-oriented in context of population dynamics of forest defoliators. In fact, you have shown that pandora moth population on the Kaibab Plateau have got strong cyclycity and influenced by weather. The (quasi)cyclic oscillations of nisects number and the weather impact are very important both for theoretical and for applied forest entomology. The useful concepts of outbreak cyclicity you can find, for example, in (https://doi.org/10.1016/0169-5347(96)81066-4, http://dx.doi.org/10.1016/B978-0-12-078148-5.50005-1 and http://dx.doi.org/10.1002/9781119407508), about weather influence on forest insects outbreaks in (http://dx.doi.org/10.1002/9781119407508 and in some parts of https://doi.org/10.1007/978-1-4899-0789-9).

 

Best wishes,

Your reviewer. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx