Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Reconstructing spruce budworm outbreak severity: a comparison of paleoecological and tree-ring signals

  • Marc-Antoine Leclerc ,

    Contributed equally to this work with: Marc-Antoine Leclerc, Martin Simard, Hubert Morin

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

    leclercmarcantoine@gmail.com

    Affiliation Department of Fundamental Sciences, Université du Québec à Chicoutimi, Chicoutimi, Québec, Canada

    ⨯
  • Martin Simard ,

    Contributed equally to this work with: Marc-Antoine Leclerc, Martin Simard, Hubert Morin

    Roles Conceptualization, Methodology, Supervision, Validation, Writing – review & editing

    Affiliation Department of Geography, Université Laval, Québec City, Québec, Canada

    ⨯
  • Hubert Morin

    Contributed equally to this work with: Marc-Antoine Leclerc, Martin Simard, Hubert Morin

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing

    Affiliation Department of Fundamental Sciences, Université du Québec à Chicoutimi, Chicoutimi, Québec, Canada

    ⨯

Abstract

Applying a centennial or millennial perspective to disturbance regimes permits an understanding of how these events have varied in the past in relation to climate change. Correctly interpreting this variability is crucial when preparing sustainable forest management practices for future warming. The eastern spruce budworm (Lepidoptera) is the most important biotic disturbance in the eastern Canadian boreal forest. Adult moths are covered by chitinous scales, and lepidopteran scale records in lake sediments have been analyzed to reconstruct Holocene spruce budworm populations. However, the magnitude of these scale accumulations has yet to be calibrated using an independent proxy. Here, we determine whether the impacts of spruce budworm defoliation are recorded by both sedimentary lepidopteran scale accumulations and tree-ring widths. Agreement between proxies was found at five of nine sites and strongest between the proportion of affected trees and scale accumulations while agreement in signal synchronicity was found at six of nine sites and strongest when comparing scale accumulations to a growth suppression index. A species-based composite chronology relying on white spruce produced the clearest outbreak record for both proxy records. Peak scale accumulations correlated well with smaller tree-ring widths, demonstrating that larger scale accumulations correspond to more severe defoliation events. Therefore, lepidopteran scales provide reliable records of spruce budworm abundance serving as a proxy record ameliorating our understanding of how budworm impacts have fluctuated at centennial and millennial time scales in the context of past climate change.

Citation: Leclerc M-A, Simard M, Morin H (2025) Reconstructing spruce budworm outbreak severity: a comparison of paleoecological and tree-ring signals. PLoS One 20(8): e0329406. https://doi.org/10.1371/journal.pone.0329406

Editor: Michal Bosela,, Technical University in Zvolen, SLOVAKIA

Received: April 12, 2025; Accepted: July 15, 2025; Published: August 12, 2025

Copyright: © 2025 Leclerc et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information.

Funding: The author(s) received funding from the ‘NSERC Industrial Research Chair on black spruce growth and the effect of the spruce budworm on landscape heterogeneity in the boreal forest’ acquired by Dr. Hubert Morin. HM received the award Grant number: 499381-15 NSERC Industrial Research Chair on the growth of black spruce and the effect of the spruce budworm on landscape heterogeneity in the boreal forest Natural Sciences and Engineering Research Council of Canada URL: https://www.nserc-crsng.gc.ca/index_eng.asp Funders did not play a role in the study design, data collection, analysis, decision to publish or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Climate change is expected to favour more frequent and severe disturbance events [1,2]. In the boreal biome, potentially affected disturbance regimes include insect outbreaks, as insect survival, growth, and development are influenced directly or indirectly by a changing climate. The projected climate-related consequences include the direct effects of temperature change [3–5], modified insect and host-tree distributions [6,7], feeding on novel host-tree species [8–10], and an altered insect-host tree phenology, possibly rendering less vulnerable hosts more vulnerable [11–13]. Indirect climate effects, via drought and variable precipitation regimes, could result in greater host-tree mortality in conjunction with insect outbreaks [14–18], or a loss in outbreak regularity [19].

In the eastern Canadian boreal forest, the spruce budworm [Choristoneura fumiferana Clemens], a native lepidopteran defoliator experiencing episodic outbreaks, is the main biotic disturbance [20,21]. The spruce budworm completes its life cycle in a single year beginning as an egg, molting through six larval stages, forming a pupa, and emerging as an adult moth that can then reproduce and disperse [21–24]. Defoliation by larvae occurs preferentially on the current year needles of mature balsam fir [Abies balsamea (L.) Mill.], with older needles targeted at high populations densities [21,25,26]. Meanwhile, white [Picea glauca (Moench) Voss] and black [Picea mariana (Mill.) Britton, Sterns & Poggennburg] spruce serve as alternate budworm hosts [27]. Defoliation by the larvae often results in growth reductions [28–31], and severe defoliation will result in tree mortality, particularly in the preferred host [32,33] as opposed to black spruce which has historically experienced lower levels of defoliation [34] but this may no longer be the case given climate change [5,11–13]. Twentieth-century outbreaks of this insect in eastern Canada have occurred roughly every 30–40 years [31,35–37] varying in duration [38] and severity [32,39–41] producing important consequences for timber supply, and carbon cycling [42].

Spruce budworm outbreaks over the last 300–400 years have been reconstructed using dendrochronology with the detection and severity of such events identified by narrow tree-ring widths [29–31,35,36,43–45]. Narrower ring widths reflect a lagged signal [46–48] of inhibited growth due to the reduction in a tree’s photosynthetic capacity [46,49], acting as a proxy of abundant spruce budworm and its feeding on needles, i.e., defoliation, once the background climate signal is removed [50]. However, the accuracy of dendrochronological records of outbreaks can depend on the tree species cored [29,30,51–53], surviving individuals of past disturbance events and their ability to record outbreaks [54] i.e., survivor bias, and is limited by a tree’s lifespan [55] resulting in a fading record by biasing reconstructions to more recent events [53,56]. As such, dendrochronological reconstructions of spruce budworm outbreaks at millennial or multi-millennial scales are limited [57].

Therefore, the use of other proxy records is required to extend spruce budworm outbreak reconstructions over thousands of years. Previous paleoecological approaches to identify spruce budworm outbreaks have included sediment records of larval head capsules, and their feces but these proxies exhibited low abundance and degradation issues, respectively [58,59]. More recently, chitinous lepidopteran scales, which cover the myriads of adult moths present during an outbreak, have proven to be an excellent proxy of spruce budworm populations [60,61]. The abundance of scales in the sedimentary record—thousands cover a single adult budworm moth—and their long-term preservation in the lake sediment record related to their chitinous composition [62–64] improve on the former insect outbreak proxies. Moreover, the abundance of scales deposited into small lakes mirrors the local insect biomass, and lepidopteran scales are incorporated into the lacustrine sediment within the year [65]. Finally, there is a high agreement between the photographic aerial survey record (1967-present) and the abundance peaks in the lepidopteran scale sedimentary record [66].

Despite the potential of scales to track long-term spruce budworm abundance, further calibration remains. Here, we compare local dendrochronological and corresponding lepidopteran scale sedimentary records from a series of lakes. We aim to establish whether both proxy records record similar impacts of spruce budworm defoliation and, if so, assess signal synchronicity. We posit two hypotheses: 1) larger accumulations of lepidopteran scales in the sedimentary records should match tree-rings records showing growth suppression/greater percentage of affected trees because greater scale accumulations in the sediment record should track larger adult moth populations around the lake and more severe defoliation by larvae; and 2) the signals from the sedimentary and tree-ring records should be relatively synchronous because lepidopteran scale incorporation into the sediment is relatively quick [65], and the effects of defoliation as recorded by tree-ring widths become evident two to four years following a defoliation event [47,48].

Methods

Ethics statement

This study was conducted in Québec’s unprotected public lands and permission was obtained from Québec’s Ministry of Forest, Fauna and Parks, the governing body managing these lands, and field work did not require a permit. No protected or endangered species were used or sampled in the course of this research.

Lepidopteran scales]

Field sampling, subsample preparation, and identification.

The same sites, surface sediment samples, age-depth models, and methods described in Leclerc et al. [66] were used to obtain sedimentary lepidopteran scale records of spruce budworm outbreaks. Briefly, a gravity corer [67,68] sampled the water-sediment interface of nine lakes located at sites that have recently sustained variable levels of defoliation [66] (Table 1). Sediment cores were sampled at a 1 cm resolution and scale recovery and analysis followed the protocol outlined in Leclerc et al. [66]. Finally, the age-depth models using 210Pb activity derived from the constant rate of supply model [69,70] were the same as those published in Leclerc et al. [66].

thumbnail
Download:
Table 1. Characteristics of the sampled lakes and of their surrounding forest.

https://doi.org/10.1371/journal.pone.0329406.t001

Tree-ring record for the 1900–2019 period

Field sampling, preparation, and identification.

We produced local chronologies at each site by coring multiple trees (20–32 of the largest and oldest individuals) within a 200 m radius around each lake (from herein ‘site’ will refer to the sediment collected from each lake and the respective cored trees) but up to an 800 m radius to sample enough individuals. We extracted one to two cores from each tree at a height of approximately ≤30 cm above the ground surface, or in certain cases ≤70 cm above the ground surface depending on the presence of rot in the stem. Cores were prepared and mounted following standard dendrochronological procedures [73]. Cores were scanned, and ring widths were measured using WinDendro software [74,75]. Samples were visually cross-dated using PAST5 software [76] guided by the Pearson correlation coefficient, threshold values of 60%–70% for the ‘Gleichlauf test’, and ≥4 for the Baillie/Pilcher and Hollstein t-tests to ensure correct dating [77]. We subsequently verified cross-dated sample placement via COFECHA v.6.06 [78] to confirm visual cross-dating. We then used chronology stripping [79] and the expressed population signal (EPS) [80,81] to select individual trees that would make up the most representative chronology for the 1900–2019 period for each site in the dplR package [82]. These selected individual trees were detrended and standardized with a 60-year cubic spline in dplR [82] to remove any low frequency signals [36,83–86], and highlight higher frequency changes that would result from spruce budworm defoliation to obtain ring-width indices, and finally produce the most representative site chronologies. The reliability of these chronologies over the study period (1900–2019) was assessed using Subsample Signal Strength with a threshold value of 0.85 [80]. These and all subsequent analyses were performed in the R statistical environment (v.4.0.5) [87].

Defoliation events described by the mean growth suppression index (GSI), and proportion of defoliated trees (‘percent affected’) for each site were obtained using the R package dfoliatR [88] using the methods and criteria developed by Swetnam et al. [50]. Defoliation events were first identified using the defoliate_trees function with default parameters [50,88] but adjusting the bridging events and series-end events to capture the current outbreak and avoid detecting multiple defoliation events that were likely part of a single event [88–90]. We calculated each site’s mean GSI and percent affected at an annual resolution with the outbreak function using default settings [88–90]. The tree-level GSI is usually calculated as [50,88,89]:

(1)

where GSIi is the growth suppression index in year i, H and NH are the standardized host-tree and non-host ring-width series, respectively, NHm is the mean non-host series’ ring-width, and is the ratio between the standard deviation of the standardized host and non-host series [50,88]. In this study, lack of local non-host individuals precluded the creation of a non-host series and use of a non-host correction such that the GSIi simply became the standardized host-tree ring-width series:

(2)

In lieu of a non-host correction, the Standardized Precipitation Evapotranspiration Index (SPEI) [91,92] from 1901–2011 [93] was used to confirm that growth reductions recorded by the tree-rings were in fact due to defoliation by the spruce budworm (S1 Appendix).

Correlation and synchronicity between scale and tree-ring records

The correlation and synchronicity between lepidopteran scale and tree-ring records (mean GSI and percent affected) were assessed using wavelet analysis. First, lepidopteran scale accumulations over the entire sediment chronology for each site were interpolated to an annual resolution by determining the relative proportion that the sediment subsample(s) contributed to each respective year and then weighing the subsample(s) based on these respective proportions to obtain an accumulation for each year [94] using the pretreatment function from the paleofire R package [95] to retain the variability in these accumulations. The interpolation permitted directly comparing scale accumulation rate, mean GSI, and percent affected. Further, interpolation to a constant time step attempted to reduce potential effects of sediment compaction that could have occurred in the scale records. The signal of each variable for each site was modelled using generalized additive models (GAMs) with a penalized cubic regression spline in the mgcv package [96]. Model fit was evaluated with the gam_check function by assessing the distribution of the residuals, normality, and the relationship between the measured and predicted variables [96].

Lake 8 required removing and imputing a point (CE 1952) in the mean GSI data set to better model the mean GSI signal and ensure the constant time step required for wavelet analysis. The comparison between the original and altered data set is presented in S2 Appendix.

The correlation and synchronicity between the lepidopteran scale and mean GSI signals and between the scale and the percent affected signals were assessed using wavelets with the biwavelet R package [97] over the time period (1900–2019) common to both records as the trees cored were not old enough to go further back in time. The use of wavelets is particularly powerful for analyzing and comparing non-stationary signals through time [98–102]. We applied a continuous Morlet wavelet transform, using default settings [97], to obtain the temporal trends for each variable because this particular continuous wavelet transform is robust to noise, can extract phase interaction information between a pair of time series [99,100,103], offers a relatively good balance between time and frequency [99], and has a high frequency resolution [100] allowing for precise and accurate identification of spruce budworm impact signals. For each site, we tested the strength of each variable’s signal (significance in the power spectrum) against a background signal generated by an autoregressive process with a lag of 1, using a test and an value of 0.05 [97,98].

Subsequently, we assessed any correlations between the scale and tree-ring records in the biwavelet R package via cross-wavelet analysis and wavelet coherence running 1000 Monte Carlo randomizations for the latter [97]. Cross-wavelet analysis reveals points in time and return intervals (i.e., periodicities) where strong signals between the variables overlap, and which signal leads or lags [99,104]. In the time–frequency dimension, wavelet coherence identifies correlations between a pair of measured series as a function of frequency, with the correlation ideally displaying a consistent relationship between signals, i.e., phased-locked behaviour [98–100,105]. The significance for both wavelet coherence and cross-wavelet analysis was tested against a background signal generated by an autoregressive process with a lag of 1, using a test, and an value of 0.05.

Finally, we determined the nature of the interaction between the two proxies by determining the phasing of the signal pairs, that is, which signal was leading or lagging relative to the other [99,100]. Phasing was key in testing the posited hypotheses by quantifying the agreement/disagreement between the proxy record signals (see below) ultimately determining whether lepidopteran scales were a reliable recorder of spruce budworm outbreaks. Phasing was represented graphically by arrows overlaid on wavelet power spectra with the arrowhead direction indicating the nature of the interaction. Arrows pointing to the right indicate that the two signals are in-phase when the peaks and troughs of both signals are synchronous. An arrow pointing to the left indicates an anti-phase relationship when the peak of a signal lines up with the trough of the other signal [99]. An upward arrow indicates that the scale record lags the dendrochronological record by [99] i.e., a peak/trough in lepidopteran scales occurs slightly after a peak/trough in either the mean GSI or percent affected. A downward arrow indicates that the scale record leads the tree-ring record by , i.e., a peak/trough in lepidopteran scales occurs slightly before a peak/trough in either the mean GSI or percent affected [99]. We calculated the circular mean and standard deviation using the circular R package [106] for the entire zone of significant high overlapping power and in zones of high correlation (R2 ≥ 0.80) over the entire study period (1900–2019) and for each known outbreak period (1912–1929, O3; 1946–1959, O2; and 1975–1992, O1) [35,36,43] for each site.

Using the circular mean values and standard deviation, we described the signals as being in agreement (lepidopteran scale and mean GSI signals showing an anti-phase behaviour) or disagreement (in-phase behaviour represents disagreement). We expect large spruce budworm impacts, related to large adult moth populations (abundant scale accumulations), will be reflected by a lower mean GSI (narrower tree-ring widths) because of greater larval defoliation. Conversely, lepidopteran scale and percent affected signals agree when these records exhibit in-phase behaviour and disagree when exhibiting anti-phase behaviour.

Composite chronologies

We created composite lepidopteran scale and tree-ring chronologies to determine whether a clearer signal could be obtained by combining sites and basing these composite records on tree species (three white spruce and six black spruce sites). We combined standardized individual tree-level ring widths from the three white spruce sites to obtain the white spruce composite. In contrast, the black spruce composite was built using the standardized individual tree-ring widths from the six black spruce sites. The respective ring-width and percent affected composite records were developed using the dplR [82], and dfoliatR [88] packages respectively. We combined the annually interpolated lepidopteran scale accumulations, obtained via the pretreatment function in the R paleofire package [95], from the respective white and black spruce sites to obtain a composite mean lepidopteran scale accumulation for each tree taxa.

Results

Age-depths models, and lepidopteran scale and tree-ring proxy records

The age-depth models derived from 210Pb measurements used to determine lepidopteran scale accumulations displayed similar shapes and low variability in date estimates over the study period (1900–2019; supplementary material in [66]). Lepidopteran scale accumulations, ring-width indices, and percent trees affected varied from site to site (S3 Appendix).

Wavelet analyses

The individual continuous Morlet wavelet transforms revealed strong significant return intervals (periodicities) in the scale and tree-ring records (Fig 1). Areas of significant high power in lepidopteran scale abundance had a periodicity of 8–32 years in seven out of nine sites from 1950–2019 (Fig 1). Areas of significant high power in the mean GSI, and percent affected also had a similar periodicity (8–32 years) for most of the study period (1900–2019) in seven of the nine sites (Fig 1). We observed areas of high power for mean GSI over the entire study period at all sites except Site 2 and high power for percent affected over the entire study period at all sites except Sites 1, and 2 (Fig 1).

thumbnail
Download:
Fig 1. The power spectra of the paleo and tree-ring variables from 1900 to 2019.

Warmer colours (red, orange, yellow) indicate high power relative to the background signal generated by an autoregressive process with a lag of 1, whereas cooler colours represent weaker power [98]. Statistically significant zones of power were determined using a test (p < 0.05) and are delineated by a thick black line [98]. The light grey shading delineates the recommended zone of interpretation, i.e., the cone of influence [100]. Squares in the left-hand corners of the spectra indicate the tree species cored; white, white spruce; black, black spruce.

https://doi.org/10.1371/journal.pone.0329406.g001

The cross-wavelet analysis between both proxies revealed a stronger signal between lepidopteran scales and the percent trees affected than between lepidopteran scales and the mean GSI at an 8-to-32-year periodicity over the entire study period for all sites (Fig 2; Table 2). Agreement at five of nine sites was observed in both lepidopteran scale-mean GSI, and lepidopteran scales-percent affected comparisons. Further consistent agreement between records was observed at Sites 2, 3, 4, and Bélanger for both comparisons (Fig 2; Table 2), noting however, that phases were variable within the zones of significant overlapping power, and also among sites.

thumbnail
Download:
Table 2. Characteristics of tree-ring chronologies and wavelet analysis results comparing spruce budworm outbreak proxies.

https://doi.org/10.1371/journal.pone.0329406.t002

thumbnail
Download:
Fig 2. The power spectra resulting from the cross-wavelet analysis applied to the paleo and dendrochronological records.

Warmer colours (red, orange, yellow) indicate high common power relative to a background signal generated by an autoregressive process with a lag of 1, whereas cooler colours represent weaker power [98]. Statistically significant zones of common power were determined using a test (p < 0.05) and are delineated by a thick black line [98]. The arrows in the areas of statistical significance help specify the type of association. Arrows pointing left indicate that the signals are anti-phase, where a peak of one signal lines up with a trough of the other signal. Arrows pointing right suggest that signals are in-phase, where peaks and troughs of both signals line up. Downward arrows indicate that the scale record leads the tree-ring record by , whereas upward arrows indicate the opposite. The light grey shading delineates the recommended zone of interpretation, i.e., the cone of influence [100]. Squares in the left-hand corners of the spectra indicate the tree species cored; white, white spruce; black, black spruce.

https://doi.org/10.1371/journal.pone.0329406.g002

We observed a highly significant wavelet coherence between the scale and tree-ring records at a one-to-four-year periodicity over the study period (Fig 3). Agreement between records was observed at six, and five of nine sites for scale-mean GSI, and scale-percent trees affected comparisons in the zones of high correlation (R2 ≥ 0.80) respectively (Fig 3; Table 2). Further, consistent agreement was observed at Sites 1, 2, 3, 5, and Bélanger (Table 2) suggesting that greater scale accumulations coincided with narrower ring widths or a greater proportion of affected trees. Significant correlations observed at higher periodicities were noteworthy at Sites 2, 3, and 8 (Fig 3).

thumbnail
Download:
Fig 3. The power spectra of the wavelet coherence analysis applied to the paleo and dendrochronological records.

Comparison between the lepidopteran scale and mean GSI time series and the lepidopteran scale and percent affected time series. Warmer colours (red, orange, yellow) indicate high correlation relative to a background signal generated by an autoregressive process of a lag of 1, whereas cooler colours represent weaker correlation [98]. Statistically significant zones were determined using a test (p < 0.05) and are delineated by a thick black line [98]. The arrows in the areas of statistical significance help specify the type of association. Arrows pointing left indicate that the signals are anti-phase, where a peak of one signal lines up with a trough of the other signal. Arrows pointing right suggest signals are in-phase, where peaks and troughs of both signals line up. Downward arrows indicate that the scale record leads the tree-ring record by , whereas upward arrows suggest the opposite. The light grey shading delineates the recommended zone of interpretation, i.e., the cone of influence [100]. Squares in the left-hand corners of the spectra indicate the tree species cored; white, white spruce; black, black spruce.

https://doi.org/10.1371/journal.pone.0329406.g003

Finally, agreement between both proxies at the individual outbreak level varied between sites and depended on the applied analysis (Table 2). We observed a more consistent outcome for the wavelet coherence analysis, although this consistency was site dependent (Table 2). For example, we found consistent agreement and disagreement in wavelet coherence between both proxies for all outbreaks at three sites and one site respectively. Agreement for cross-wavelet analysis at the outbreak level was more variable, and only Lake Bois Joli displayed a constant disagreement between both proxies for all outbreaks (Table 2).

Composite chronologies

By creating species-specific composite chronologies, we obtained a stronger, and clearer signal of spruce budworm impact for both proxies in the white spruce composite relative to the black spruce composite. The white spruce composite of lepidopteran scale accumulation revealed three relatively distinct peak scale accumulations corresponding to the three 20th century outbreaks (Fig 4). The ring-width index revealed three clear areas of growth suppression with three corresponding peaks in the proportion of affected trees. We did observe, however, a consistent 10-year lag between the scale and tree-ring records. We only observed a single large peak scale accumulation in the black spruce composite from ~2000 AD (Fig 4). Similarly, the dendrochronological black spruce composite revealed a major growth suppression and an increased proportion of affected trees from ~2000 AD to 2019 (Fig 4). The variability in the black spruce composite proxy records was such that the other outbreak signals were difficult to identify (Fig 4) suggesting more variable levels of defoliation and/or a more variable ability to record defoliation by this species.

thumbnail
Download:
Fig 4. The white and black spruce composite chronologies of the scale and tree-ring records.

The white (left column) and black (right column) spruce records comprise (A) the mean lepidopteran scale accumulation, (B) the mean ring-width index (RWI), and (C) the proportion of affected trees (percent affected). The white spruce composite incorporates data from three sites. The black spruce composite incorporates data from six sites.

https://doi.org/10.1371/journal.pone.0329406.g004

Discussion

Proxy record strengths, and limitations

Both the scale and tree-ring records record the impact of spruce budworm defoliation, and the signals are relatively synchronous, validating the use of lepidopteran scales for reconstructing centennial to multi-millennial spruce budworm population fluctuations. Lepidopteran scales are assumed to be a direct measure of the size of the adult moth spruce budworm population as their count in the sediment represents the changing insect biomass at a site over time (a stand-level phenomenon), with a greater count at a particular depth reflecting a larger budworm population. This assumption is based on prior knowledge of the insect’s biology, including life cycle and development, feeding behaviour and impacts [21,23,107–109], and the epidemiology of past outbreaks [31,36,44,54,110]. In addition to representing changes in adult spruce budworm populations, lepidopteran scale accumulations are correlated with severe annual defoliation [65], indicating that greater accumulations may reflect more severe events. However, despite a relatively quick incorporation into the sediment record [63,65], the lepidopteran scale signal is affected by the rate of sediment deposition and mixing both highly variable among lakes, and even within a given lake [111]. There is also error associated with the dating of the sediment record [112] producing artificial lags or asynchrony between the sediment record and the tree-ring record.

The tree-ring record, from which the mean GSI and percent affected variables are derived, is an indirect indicator of the spruce budworm presence. Specifically, mean GSI requires the growth records of several individual trees to obtain a mean stand-level growth pattern. An individual tree’s reaction, i.e., the production of narrower growth rings, to a defoliation event is nonetheless unique in that multiple contributing factors, such as site conditions [113], species composition [39,40,114,115], genetics [116,117], age [33,118], and tree height [119,120] can influence the extent, and speed that a tree exhibits growth suppression [41,46,121]. The reaction and response time can thus introduce variation in the mean stand growth to obscure the spruce budworm signal.

The percent affected metric, also based on tree ring widths, reflects the proportion of trees affected by the spruce budworm. More specifically, it is binary; depending on the applied defoliation threshold, a tree is counted as having been defoliated or not [88]. Contrary to the mean GSI, which quantifies the change in the mean stand-level growth over time, percent affected is a coarser measure of the spruce budworm’s impact on the stand because it testifies to changes in the proportion of trees considered as defoliated over time [88]. However, both are based on the survivors of past outbreaks and other disturbances. Therefore, these survivors of past defoliation events may not necessarily reflect the true severity of the event [53,54,83]. Moreover, harvesting and wildfires may remove trees that likely recorded a spruce budworm signal. Those remaining trees tend to be younger and typically less affected by defoliation [32,33,118]. Finally, the tree species may influence the recorded spruce budworm signal. Defoliation on white spruce is reliably recorded [5,29,30,51], whereas the defoliation signal in black spruce may not be as clear [28,30] despite their use in reconstructing outbreaks [48,52]. Therefore, it is important to note that although the tree-ring record is more temporally precise than the paleo proxy, recorded annual changes in growth depend on survivors, and incorporates large amounts of variability such that annual changes in local defoliator populations cannot be directly inferred from this record [122].

Proxy record signal comparison, and factors affecting proxy signals

Both the scale and tree-ring records highlighted strong significant return intervals (periodicities) in spruce budworm populations. The strong return interval signal identified by both proxies typically ranged from 16 to 32 years over the study period likely reflects the 20th century return interval of spruce budworm outbreaks that occurred every 30–40 years at the landscape scale [31,35,43,54,110]. The similar intervals identified indicate that both proxies detected spruce budworm impacts. A strong lepidopteran scale signal occurred in the latter half of the 20th century but was not observed in the first half. This pattern was evident at most sites, and may be due to little variability in scale accumulations from 1900–1950 which may have been amplified by the interpolation process making it difficult to distinguish peaks from background accumulations, in combination with the shortening of the scale record to match the dendrochronological record (S3 Appendix). Conversely, we observed a strong signal in the dendrochronological data set over the entire study period (1900–2019) for both mean GSI and percent affected. However, the strength, return intervals, and time over which the signal was observed in both the scale and tree-ring records were site dependent. For example, at Site 4 a strong signal is observed in mean GSI and percent affected prior to 1950 but not after 1950, meanwhile we observed a strong signal in the scale record after 1950. Conversely, Site Bois Joli displayed a strong signal in both proxy records throughout the study period.

The cross-wavelet analysis also identified zones of strong significant overlapping signals common to both records at intervals of 16–32 years over the study period (1900–2019), suggesting that both proxy records, relative to one another, detect periods of high spruce budworm impact. However, we observed a difference in the strength of overlapping power when comparing lepidopteran scales to the mean GSI instead of percent affected. Although significant, the weak overlap in power in the mean GSI signal relative to the percent affected signal may be because the accumulation of lepidopteran scales reflects the stand-level phenomenon of insect biomass at a site. Mean GSI on the other hand, using the mean annual ring width of the stand [88], may introduce more tree-level variability, and wavelet analysis may be sensitive to this variability. If this tree-level variability exceeds the background signal generated by the analysis, it will be considered as signal, and although a general pattern may emerge, a fuzzier signal occurs in years when a single tree expresses good growth, and the majority do not. It is also worth acknowledging that lepidopteran scale accumulation interpolation to an annual resolution could have also affected the detected scale-mean GSI signal pair comparison.

Two reasons may explain the stronger overlap in the lepidopteran scales and percent trees affected pair. The first relates to the nature of the paleo proxy. As mentioned above, lepidopteran scales accumulations reflect the insect population in a stand, and not on an individual tree and moreover, scale accumulations reflect moderate and severe defoliation events [65]. The second reason is methodological relating to how the percent affected metric is calculated. As stated earlier, the percent affected metric does not assess the exact magnitude of growth reduction expressed by each tree, it relies on predetermined thresholds to classify individual trees as being defoliated or not and based on this classification determines the proportion of defoliated trees in a stand [50,88]. Additionally, the applied thresholds most likely reflect moderate and severe defoliation. As such, the coarser representation of the spruce budworm’s impact represented by the proportion of affected trees compares similar defoliation impacts and better corresponds to the stand-level phenomenon measured by the scale accumulations resulting in the stronger overlap of high common power. In contrast, the mean GSI measures individual tree-level changes in growth obtaining an average outcome for the stand noting that in this study a non-host correction could not be applied. However, site-level SPEIs from 1901−2011 did not reveal drought conditions (values exceeding −2) during the known outbreak periods (S1 Appendix). Further, correlations and cross-correlations between site-level SPEI and tree-ring chronologies were insignificant (weakly correlated at site Bois Joli) suggesting that growth reductions recorded were due to the defoliation by the spruce budworm (S1 Appendix).

Wavelet coherence revealed a significant correlation over a short time period between the scale and tree-ring records, suggesting that the signals were synchronous. This correlation was typically observed at a periodicity of four years or less, although the periodicity ranged from 0 to 16 years. Here, the return intervals (periodicities), can be interpreted as the time required for the proxies to record the spruce budworm signal. The zero-to-four-year periodicity between the signals likely results from a combination of the time required for scale incorporation into the sediment, potential errors associated with the dating of the sediment core that carry over into the age–depth models for each core [112,123], and the lag of two to four years after the onset of an outbreak for defoliation to become apparent in individual trees [47,48]. However, we also recorded a significant correlation at longer intervals, and the implications of this remain unclear. A significant correlation at an interval ≥16 years was present in the latter half of the 20th century for sites 2, 3, 4, and 8. These zones of high correlation may result from chance, producing a spurious correlation between the proxy records at longer intervals.

Site history, particularly changes to the proportion of host-trees, can influence the spruce budworm signal recorded by the lepidopteran scales and tree-rings producing variable areas of high power and dampen one or both proxy record signals affecting detected correlations. For example, timber harvesting, as observed around lakes 2, 3, and 8 in the latter half of the 20th century (Table 1 and S3 Appendix), has the potential of altering host-tree proportions and influence which individual trees remain on site. A reduction in host-tree proportion limits spruce budworm abundance likely resulting in small populations causing the paleo-proxy to record accumulations that are indistinguishable from background accumulations. Interestingly, even given a low abundance of spruce budworm, the few surviving host-trees left behind from harvesting may record low defoliation or possibly more severe defoliation that does not reflect the budworm population due to a greater density of larvae on a smaller food source [124–126]. In this situation, the proxy records could produce more monotonous or even opposing signals potentially resulting in spurious correlations. For example, the effect of past and current harvesting is well demonstrated within Site 2, as neither individual proxy exhibits a strong signal (zones of significant power). Curiously, the proxies at Site 2 still exhibit a significantly strong correlation, most likely spurious. Analogously, previous fires may also produce spurious correlations at longer return intervals as observed at Site 4 modifying the proportion host-trees in a stand via the stochastic nature of ignitions, burn area, and tree mortality. Finally noting that strong signals at other return intervals may also be a result of mathematical artifacts such that multiples of a particular return interval may be highlighted. In this case, the biologically relevant 32-year interval would be identified along with the mathematically, but not biologically relevant intervals of 8 or 16 years.

Wavelet analysis revealed that sedimentary lepidopteran scale and tree-ring records identify spruce budworm outbreaks (both individually and when compared with each other), and their signals are synchronous. The early part of the statement is supported by the presence of areas of high power (red zones; Fig 1) that may be significant (thick black line; Fig 1) in the individual and cross-wavelet spectra coinciding with the approximately 30-year spruce budworm outbreak cycle (30-year period along the y-axis; Figs 1 and 2). The latter part of the statement is supported by the presence of areas of high significant power within one to four years in the wavelet coherence spectra (Fig 3). The final piece of interpretation involves assessing the phases (arrows in the cross-wavelet and coherence figures; Figs 2 and 3) to characterize the relationship between the two proxies. The average directionality of the arrows indicates that both proxy records agree when large lepidopteran scale accumulations coincide with narrow tree rings or when large lepidopteran scale accumulations coincide with a large proportion of affected trees (Table 2). However, this agreement between proxies at the outbreak level was highly variable suggesting a complex relationship (Table 2).

Composite chronology comparison

The signal obtained from the composite chronologies was species-dependent. We observed a strong, clear signal in the white spruce composite with distinct periods of high lepidopteran scale accumulations, growth suppression, and large proportions of defoliated trees. This finding agrees with previous studies that demonstrated white spruce to be a good recorder of spruce budworm outbreak events [27,29,30,51], and our recorded periods coincide very well with known outbreak periods of 1912–1929, 1949–1959, 1975–1992 [35,36,44,45]. However, there was a consistent 10-year lag between the lepidopteran scale accumulations and the dendrochronological variables, which could result from the paleo-proxy’s inability to differentiate between periods of no or light defoliation [65] corresponding to low population levels. Scales may not record the entire progression of each outbreak, but record the point at which the insect populations surpass a certain threshold corresponding to periods of moderate or severe defoliation [65] whereas tree-rings may be more sensitive and able to record defoliation at lower insect abundances. Alternatively, the lag may be an artifact of sediment dating errors related to the estimation of background 210Pb levels that can result from natural variation in precipitation [127,128], sediment compaction and/or remobilization, and floods events for example [70,129–132]. In this case, the errors would be associated with sediment layers closest to the water-sediment interface that constrain the maximum age (and errors) at the top of the chronology [112]. Additionally, the taphonomical processes that govern incorporation of sedimentary charcoal into lake sediments including bioturbation, erosion re-deposition, and subaqueous slumping [123] also likely affect the incorporation and distribution of lepidopteran scales into the sediment column.

In contrast, the black spruce composite record was not as clear as the white spruce record. We recorded periods of growth suppression around 1915, 1950, and 1980, although the latter two low-growth periods were not very evident, whereas the current outbreak was clear as evidenced by peak scale accumulations ca. 2000–2019 and low-growth. The noisier signal in the black spruce composite record may result from the more variable proportion of balsam fir in the surrounding forest of each site affecting defoliation severity, where defoliation severity in black spruce is greater with an increasing proportion of balsam fir [126,133,134]. Additionally, the historical phenological mismatch between larval emergence and budburst along with cooler temperatures may have limited population build-ups and attenuated defoliation on black spruce [5,11–13,34], ultimately weakening past outbreak signals.

Future research avenues

Finally, a multi-proxy comparison between tree-rings with other paleo-proxies and between different paleo-proxies could also prove fruitful. For example, a comparison using known, recent outbreaks (approximately last 150 years) between tree-rings and spruce budworm head capsules and frass would help in ameliorating the interpretation of their respective accumulations in the sediment [58,59]. Such a comparison could even be extended over longer time periods using subfossil trees [57]. Similarly, a comparison between tree-rings and other paleo-proxies such as isotopes, and pollen ratios [135] could be investigated. Additionally, utilising a combination of lepidopteran scales, pollen, and macrofossils (both plant and insect) could provide more complete and robust long-term reconstructions. The use of ancient DNA may be a promising avenue of research [136,137] as it has been used to reconstruct faunal [138] and plant diversity [139,140] during the Holocene, but would again need to be compared with other spruce budworm proxies.

Conclusion

Our study demonstrated that lepidopteran scales and tree ring-derived variables recorded the impact of 20th-century spruce budworm outbreaks and that the signals obtained from these proxies were (relatively) synchronous. A key takeaway from this study is that using a multi-proxy approach is essential for reconstructing past disturbances. Comparisons of proxy records can determine whether events are recorded in time and help better interpret the magnitude of a given event. More importantly, multiple lines of evidence provide a more robust interpretation of past disturbance events because of differential recording abilities of the archives and the varying degree of influence by site history. Both lepidopteran scales and tree-ring widths performed well in recording the impacts of the spruce budworm, indicating that both proxies are reliable for longer-term reconstructions, and although scales are a valid long-term proxy, care must be taken in site selection, and sediment sampling to ensure accurate reconstructions. As more lakes are sampled and long-term spruce budworm chronologies are obtained from lepidopteran scales, regional composite scale-tree-ring chronologies will offer the best means of assessing how spruce budworm impacts varied at centennial and millennial scales in the context of past climate change because they reduce variability associated with site history, lake-level taphonomical processes, and help circumvent the inability of every lake to record every outbreak [54,83]. This is especially important as there does not appear to be a particular suite of site properties, based on the current data used in this study, that govern precise and accurate spruce budworm event reconstructions requiring further investigation.

Acknowledgments

We thank first and foremost Dr. Olivier Blarquez who provided methodological and analytical advice and feedback to this project prior to an abrupt career change. We also thank 2 anonymous reviewers, and Dr. Murray Hay for constructive feedback on earlier versions of the manuscript. François Gionest for help in the field, and Marika Tremblay and Guillaume Vigneault for help in the laboratory.

References

  1. 1. Millar CI, Stephenson NL. Temperate forest health in an era of emerging megadisturbance. Science. 2015;349(6250):823–6.
  2. 2. McDowell NG, Allen CD, Anderson-Teixeira K, Aukema BH, Bond-Lamberty B, Chini L, et al. Pervasive shifts in forest dynamics in a changing world. Science. 2020;368(6494):eaaz9463. pmid:32467364
  3. 3. Berg EE, Henry DH, Fastie CL, De Volder AD, Matsuoka SM. Spruce beetle outbreaks on the Kenai Peninsula, Alaska, and Kluane National Park and Reserve, Yukon Territory: Relationship to summer temperatures and regional differences in disturbance regimes. For Ecol Manage. 2006;227:219–32.
  4. 4. Bentz BJ, Régnière J, Fettig CJ, Hansen EM, Hayes JL, Hicke JA. Climate change and bark beetles of the western United States and Canada: Direct and indirect effects. BioScience. 2010;60(8):602–13.
  5. 5. Régnière J, St-Amant R, Duval P. Predicting insect distributions under climate change from physiological responses: spruce budworm as an example. Biol Invasions. 2012;14:1571–86.
  6. 6. Jepsen JU, Hagen SB, Ims RA, Yoccoz NG. Climate change and outbreaks of the geometrids Operophtera brumata and Epirrita autumnata in subarctic birch forest: evidence of a recent outbreak range expansion. J Anim Ecol. 2008;77(2):257–64.
  7. 7. Jepsen JU, Kapari L, Hagen SB, Schott T, Vinstad OPL, Nilssen AC. Rapid northwards expansion of a forest insect pest attributed to spring phenology matching with sub-arctic birch. Global Change Biology. 2011;17(6):2071–83.
  8. 8. Cullingham CI, Cooke JEK, Dang S, Davis CS, Cooke BJ, Coltman DW. Mountain pine beetle host-range expansion threatens the boreal forest. Molecular Ecology. 2011;20:2157–71.
  9. 9. Erbilgin N, Ma C, Whitehouse C, Shan B, Najar A, Evenden M. Chemical similarity between historical and novel host plants promotes range and host expansion of the mountain pine beetle in a naïve host ecosystem. New Phytol. 2014;201(3):940–50. pmid:24400902
  10. 10. Rosenberger DW, Venette RC, Maddox MP, Aukema BH. Colonization behaviors of mountain pine beetle on novel hosts: implications for range expansion into northeastern North America. PLoS One. 2017;12(5):e0176269.
  11. 11. Pureswaran DS, De Grandpré L, Paré D, Taylor A, Barrette M, Morin H. Climate-induced changes in host tree-insect phenology may drive ecological state-shift in boreal forests. Ecology. 2015;96(6):1480–91.
  12. 12. Pureswaran DS, Neau M, Marchand M, De Grandpré L, Kneeshaw D. Phenological synchrony between eastern spruce budworm and its host trees increases with warmer temperatures in the boreal forest. Ecol Evol. 2018;9(1):576–86. pmid:30680138
  13. 13. Fuentealba A, Pureswaran D, Bauce É, Despland E. How does synchrony with host plant affect the performance of an outbreaking insect defoliator?. Oecologia. 2017;184(4):847–57. pmid:28756489
  14. 14. Allen CD, Macalady AK, Chenchouni H, Bachelet D, McDowell N, Vennetier M. A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. For Ecol Manage. 2010;259:660–84.
  15. 15. Allen CD, Breshears DD, McDowell NG. On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene. Ecosphere. 2015;6(8):129.
  16. 16. Hart SJ, Veblen TT, Eisenhart KS, Jarvis D, Kulakowski D. Drought induces spruce beetle (Dendroctonus rufipennis) outbreaks across northwestern Colorado. Ecology. 2014;95(4):930–9. pmid:24933812
  17. 17. Hart SJ, Veblen TT, Schneider D, Molotch NP. Summer and winter drought drive the initiation and spread of spruce beetle outbreak. Ecology. 2017;98(10):2698–707. pmid:28752623
  18. 18. DeGrandpré L, Kneeshaw DD, Perigon S, Boucher D, Marchand M, Pureswaran D. Adverse climatic periods precede and amplify defoliator-induced tree mortality in eastern boreal North America. Journal of Ecology. 2019;107(1):452–67.
  19. 19. Esper J, Büntgen U, Frank DC, Nievergelt D, Liebhold A. 1200 years of regular outbreaks in alpine insects. Proc Biol Sci. 2007;274(1610):671–9. pmid:17254991
  20. 20. MacLean DA. Impacts of insect outbreaks on tree mortality, productivity, and stand development. Can Entomol. 2015;148(S1):S138–59.
  21. 21. Nealis VG. Comparative ecology of conifer-feeding spruce budworms (Lepidoptera: Tortricidae). Canadian Entomologist. 2016;148:S33–57.
  22. 22. Royama T. Population Dynamics of the Spruce Budworm Choristoneura Fumiferana. Ecological Monographs. 1984;54(4):429–62.
  23. 23. Royama T, MacKinnon WE, Kettela EG, Carter NE, Hartling LK. Analysis of spruce budworm outbreak cycles in New Brunswick, Canada, since 1952. Ecology. 2005;86(5):1212–24.
  24. 24. Régnière J, Nealis VG. Ecological mechanisms of population change during outbreaks of the spruce budworm. Ecol Entomol. 2007;32:461–77.
  25. 25. Piene H. Spruce budworm defoliation and growth loss in young balsam fir: Defoliation in space and unspaced stands and individual tree survival. Can J For Res. 1989;19:1211–7.
  26. 26. MacLean DA, Hunt TL, Eveleigh ES, Morgan MG. The relation of balsam fir volume increment to cumulative spruce budworm defoliation. For Chron. 1996;72(5):533–40.
  27. 27. Hennigar CR, MacLean DA, Quiring DT, Kershaw JA. Differences in spruce budworm defoliation among balsam fir, and white, red, and black spruce. Forest Science. 2008;54(2):158–66.
  28. 28. Blais JR. Some relationships of the spruce budworm, Choristoneura fumiferana (Clem.) to black spruce Picea mariana (Moench) Voss. For Chron. 1957;33(4):364–72.
  29. 29. Blais JR. Effects of defoliation by spruce budworm (Choristoneura fumiferana Clem.) on radial growth at breast height of balsam fir (Abies balsamea (L.) Mill.) and white spruce (Picea glauca (Moench) Voss.). For Chron. 1958;34(1):39–47.
  30. 30. Blais JR. Collection and analysis of radial-growth data from trees for evidence of past spruce budworm outbreaks. For Chron. 1962;38(4):474–84.
  31. 31. Blais JR. Trends in the frequency, extent, and severity of spruce budworm outbreaks in eastern Canada. Can J For Res. 1983;13(4):539–47.
  32. 32. MacLean DA. Vulnerability of fir-spruce stands during uncontrolled spruce budworm outbreaks- a review and discussion. For Chron. 1980;56(5):213–21.
  33. 33. MacLean DA. Effects of spruce budworm outbreaks on the productivity and stability of balsam fir forests. For Chron. 1984;60(5):273–9.
  34. 34. Nealis VG, Régnière J. Insect-host relationships influencing disturbance by the spruce budworm in a boreal mixedwood forest. Can J For Res. 2004;34:1870–82.
  35. 35. Boulanger Y, Arseneault D. Spruce budworm outbreaks in eastern Québec over the last 450 years. Can J For Res. 2004;34(5):1035–43.
  36. 36. Boulanger Y, Arseneault D, Morin H, Jardon Y, Bertrand P, Dagneau C. Dendrochronological reconstruction of spruce budworm (Choristoneura fumiferana) outbreaks in southern Quebec for the last 400 years. Can J For Res. 2012;42:1264–76.
  37. 37. Volney WJA, Fleming RA. Spruce budworm (Choristoneura spp.) biotype reactions to forest and climate characteristics. Global Change Biology. 2007;13:1630–43.
  38. 38. Gray DR, Régnière J, Boulet B. Analysis and use of historical patterns of spruce budworm defoliation to forecast outbreak patterns in Québec. For Ecol Manage. 2000;127(1–3):217–31.
  39. 39. Su Q, Needham TD, MacLean DA. The influence of hardwood content on balsam fir defoliation by spruce budworm. Can J For Res. 1996;26:1620–8.
  40. 40. Colford-Gilks AK, MacLean DA, Kershaw JA, Béland M. Growth and mortality of balsam fir- and spruce-tolerant hardwood stands as influenced by stand characteristics and spruce budworm defoliation. For Ecol Manage. 2012;280:82–92.
  41. 41. Houndode DJ, Krause C, Morin H. Predicting balsam fir mortality in boreal stands affected by spruce budworm. For Ecol Manage. 2021;496:119408.
  42. 42. Dymond CC, Neilson ET, Stinson G, Porter K, MacLean DA, Gray DR, et al. Future Spruce Budworm Outbreak May Create a Carbon Source in Eastern Canadian Forests. Ecosystems. 2010;13(6):917–31.
  43. 43. Morin H, Laprise D. Histoire récente des épidémies de la tordeuse des bourgeons de l’épinette au nord du lac-saint-jean (québec): une analyse dendrochronologique. Can J For Res. 1990;20(1):1–8.
  44. 44. Morin H. Dynamics of balsam fir forests in relation to spruce budworm outbreaks in the boreal zone of Québec. Can J For Res. 1994;24(4):730–41.
  45. 45. Blais JR. Regional Variation in Susceptibility of Eastern North American Forests to Budworm Attack Based on History of Outbreaks. The Forestry Chronicle. 1968;44(3):17–23.
  46. 46. Ericsson A, Larsson S, Tenow O. Effects of early and late season defoliation on growth and carbohydrate dynamics in Scots pine. J Appl Ecol. 1980;17(3):747–69.
  47. 47. Krause C, Morin H. Changes in radial increment in stems and roots of balsam fir [Abies balsamea (L.) Mill.] after defoliation by spruce budworm. For Chron. 1995;71(6):747–54.
  48. 48. Krause C, Luszczynski B, Morin H, Rossi S, Plourde PY. Timing of growth reductions in black spruce stem and branches during the 1970s spruce budworm outbreak. Can J For Res. 2012;42:1220–7.
  49. 49. Deslauriers A, Caron L, Rossi S. Carbon allocation during defoliation: testing a defense-growth trade-off in balsam fir. Front Plant Sci. 2015;6:338. pmid:26029235
  50. 50. Swetnam TW, Thompson MA, Sutherland EK. Using dendrochronology to measure radial growth of defoliated trees. 639. United States Department of Agriculture. 1985.
    • 51. Bouchard M, Pothier D. Spatiotemporal variability in tree and stand mortality caused by spruce budworm outbreaks in eastern Quebec. Can J For Res. 2010;40:86–94.
    • 52. Tremblay M-J, Rossi S, Morin H. Growth dynamics of black spruce in stands located between the 51st and 52nd parallels in the boreal forest of Quebec, Canada. Can J For Res. 2011;41(9):1769–78.
    • 53. Pothier D, Elie J-G, Auger I, Mailly D, Gaudreault M. Spruce budworm-caused mortality to balsam fir and black spruce in pure and mixed conifer stands. Forest Science. 2012;58(1):24–33.
    • 54. Jardon Y, Morin H, Dutilleul P. Périodicité et synchronisme des épidémies de la tordeuse des bourgeons de l’épinette au Québec. Can J For Res. 2003;33:1947–61.
    • 55. Burns RM, Honkala BH. Silvics of North America volume 1: conifers and volume 2: hardwoods. United States Department of Agriculture Forest Service Agriculture Handbook 654; 1990.
    • 56. Duchesne L, Houle D, Ouimet R, Caldwell L, Gloor M, Brienen R. Large apparent growth increases in boreal forests inferred from tree-rings are an artefact of sampling biases. Sci Rep. 2019;9(1):6832. pmid:31048703
    • 57. Simard S, Morin H, Krause C. Long-term spruce budworm outbreak dynamics reconstructed from subfossil trees. J Quat Sci. 2011;26(7):734–8.
    • 58. Simard I, Morin H, Potelle B. A new paleoecological approach to reconstruct long-term history of spruce budworm outbreaks. Can J For Res. 2002;32(3):428–38.
    • 59. Simard I, Morin H, Lavoie C. A millennial-scale reconstruction of spruce budworm abundance in Saguenay, Québec, Canada. Holocene. 2006;16(1):31–7.
    • 60. Navarro L, Harvey A-É, Morin H. Lepidoptera wing scales: a new paleoecological indicator for reconstructing spruce budworm abundance. Can J For Res. 2018a;48(3):302–8.
    • 61. Navarro L, Harvey A-É, Ali A, Bergeron Y, Morin H. A Holocene landscape dynamic multiproxy reconstruction: How do interactions between fire and insect outbreaks shape an ecosystem over long time scales?. PLoS One. 2018b;13(10):e0204316. pmid:30278052
    • 62. Richards AG. Studies on Arthropod Cuticle. I. The Distribution of Chitin in Lepidopterous Scales, and its Bearing on the Interpretation of Arthropod Cuticle1,2. Annals of the Entomological Society of America. 1947;40(2):227–40.
    • 63. Sturz H, Robinson J. Anaerobic decomposition of chitin in fresh-water sediments. In: Muzzarelli R, Jeuniaux C, Gooday GW, editors. Chitin and nature and technology. Boston: Springer. 1986. p. 531–8.
      • 64. Montoro Girona M, Navarro L, Morin H. A secret hidden in the sediments: Lepidoptera scales. Front Ecol Evol. 2018;6(2):1–5.
      • 65. Tremblay E. Quantification de la relation entre la défoliation des peuplements forestiers causée par la tordeuse des bourgeons de l’épinette (Choristoneura fumiferana Clem.) et les écailles du papillon de la tordeuse des bourgeons de l’épinette retrouvées dans les sédiments lacustres annuels boréaux. M. Sc. Thesis, Université du Québec à Chicoutimi. 2022. Available from: https://constellation.uqac.ca/id/eprint/8874//.
        • 66. Leclerc MA, Simard M, Blarquez O, Morin H. Lepidopteran scales in lake sediments as a reliable proxy for spruce budworm outbreak events in the boreal forest of Eastern Canada. Holocene. 2024;34(7):978–86.
        • 67. Renberg I. The HON-Kajak sediment corer. J Paleolimnol. 1991;6(2):167–70.
        • 68. Renberg I, Hansson H. The HTH sediment corer. J Paleolimnol. 2008;40:655–9.
        • 69. Appleby PG, Oldfield F. The assessment of 210Pb data from sites with varying sediment accumulation rates. Hydrobiologia. 1983;103:29–35.
        • 70. Binford MW. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake sediment cores. J Paleolimnol. 1990;3:253–67.
        • 71. MFFP (Ministère des forêts, de la faune et des parcs). Shapefiles of the forest inventory data including forest composition and previous disturbances; 2021b [cited May 15 2020. ]. Database: Shapefiles [Internet].Available from: https://www.donneesquebec.ca/recherche/dataset/carte-ecoforestiere-avec-perturbations
          • 72. MFFP (Ministère des forêts, de la faune et des parcs). Shapefiles of the aerial surveys done to assess the damage caused by the spruce budworm from 1967-present; 2021a [cited May 15 2020. ]. Database: Shapefiles [Internet]. Available from: https://www.donneesquebec.ca/recherche/fr/dataset/donnees-sur-les-perturbations-naturelles-insecte-tordeuse-des-bourgeons-de-lepinette
            • 73. Stokes MA, Smiley TL. An introduction to tree-ring dating. Chicago: University of Chicago Press. 1968.
              • 74. Regent Instruments Inc. WinDendro: Tree ring and wood density image analysis system. 2017.
                • 75. Guay R, Gagnon R, Morin H. A new automatic and interactive tree ring measurement system based on a line scan camera. The Forestry Chronicle. 1992;68(1):138–41.
                • 76. Sciem. Past5 (Personal analysis system for tree ring research) version 5.0.576 software. 2015.
                  • 77. Sciem. Past5 (Personal Analysis System for Tree Ring Research) Software Instruction Manual. Vienna. 2015.
                    • 78. Holmes RL. Computer-assisted quality control in tree-ring dating and measurement. Tree Ring Res. 1983;43:69–78.
                    • 79. Fowler A, Boswijk G. Chronology stripping as a tool for enhancing the statistical quality of tree-ring chronologies. Tree Ring Research. 2003;59(2):53–62.
                    • 80. Wigley TML, Briffa KR, Jones PD. On the average value of correlated time series, with applications in dendroclimatology and hydrometeorology. J Appl Meteorol Climatol. 1984;23(2):201–13.
                    • 81. Buras A. A comment on the expressed population signal. Dendrochronologia. 2017;44:130–2.
                    • 82. Bunn A. Statistical and visual crossdating in R using the dplR library. Dendrochronologia. 2010;28(4):251–8.
                    • 83. Bouchard M, Kneeshaw D, Bergeron Y. Forest dynamics after successive spruce budworm outbreaks in mixedwood forests. Ecology. 2006;87(9):2319–29. pmid:16995632
                    • 84. Bouchard M, Kneeshaw D, Messier C. Forest dynamics following spruce budworm outbreaks in the northern and southern mixedwoods of central Quebec. Can J For Res. 2007;37:763–72.
                    • 85. Robert LE, Kneeshaw D, Sturtevant BR. Effects of forest management legacies on spryce budworm (Choristoneura fumiferana) outbreaks. Can J For Res. 2012;42:463–75.
                    • 86. Robert LE, Sturtevant BR, Cooke BJ, James PMA, Fortin MJ, Townsend PA. Landscape host abundance and configuration regulate periodic outbreak behavior in spruce budworm Choristoneura fumiferana. Ecography. 2018;41:1556–71.
                    • 87. R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. 2021.
                      • 88. Guiterman CH, Lynch AM, Axelson JN. DfoliatR: An R package for detection and analysis of insect defoliation signals in tree rings. Dendrochronologia. 2020;63:125750.
                      • 89. Swetnam TW, Lynch AM. A Tree-Ring Reconstruction of Western Spruce Budworm History in the Southern Rocky Mountains. Forest Science. 1989;35(4):962–86.
                      • 90. Lynch AM. What tree-ring reconstruction tells us about conifer defoliator outbreaks. In: Barbosa P, Letourneau DK, Agrawal AA, editors. Insect outbreaks revisited. Blackwell Publishing Ltd. 2012. p. 126–54.
                        • 91. Vicente-Serrano SM, Beguería S, López-Moreno JI. A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index. J Clim. 2010;23(7):1696–718.
                        • 92. Vicente-Serrano SM, Beguería S, López-Moreno JI, Angulo M, El Kenawy A. A new global 0.5° gridded dataset (1901-2006) of a multiscalar drought index: comparison with current drought index datasets based on the Palmer Drought Severity Index. J Clim. 2010;11(4):1033–43.
                        • 93. Environment and Climate Change Canada. Twelve-month historical Standardized Precipitation Evapotranspiration Index values for Canada from 1900-2011; 2015 [accessed June 17 2024]. Database: CSV file [Internet]. Available from: https://open.canada.ca/data/dataset/55fa8f11-aa8c-4232-86da-2e370239e096
                          • 94. Higuera PE, Brubaker LB, Anderson PM, Hu FS, Brown TA. Vegetation mediated the impacts of postglacial climate change on fire regimes in the south‐central Brooks Range, Alaska. Ecological Monographs. 2009;79(2):201–19.
                          • 95. Blarquez O, Vannière B, Marlon JR, Daniau A-L, Power MJ, Brewer S. Paleofire: An R package to analyse sedimentary charcoal records from the Global Charcoal Database to reconstruct past biomass burning. Comput Geosci. 2014;72:255–61.
                          • 96. Wood SN. Generalized Additive Models: An Introduction with R. 2nd ed. Chapman and Hall/CRC. 2017.
                            • 97. Gouhier TC, Grinsted A, Simk V. R package biwavelet: Conduct univariate and bivariate wavelet analyses (Version 0.20.21); 2021. Available from: https://github.com/tgouhier/biwavelet
                              • 98. Torrence C, Compo GP. A practical guide to wavelet analysis. Bull Am Meteorol Soc. 1998;79(1):61–78.
                              • 99. Grinsted A, Moore JC, Jevrejeva S. Application of the cross wavelet transform and wavelet coherence to geophysical time series. Nonlinear Process Geophys. 2004;11:561–6.
                              • 100. Cazelles B, Chavez M, Berteaux D, Ménard F, Vik JO, Jenouvrier S, et al. Wavelet analysis of ecological time series. Oecologia. 2008;156(2):287–304. pmid:18322705
                              • 101. Fernández-Macho J. Wavelet multiple correlation and cross-correlation: A multiscale analysis of Eurozone stock markets. Physica A: Statistical Mechanics and its Applications. 2012;391(4):1097–104.
                              • 102. Fernández-Macho J. Time-localized wavelet multiple regression and correlation. Physica A. 2018;493:1226–38.
                              • 103. Mallat S. A wavelet tour of signal processing. San Diego: Academic Press. 1998.
                                • 104. Labat D. Recent advances in wavelet analyses: Part 1. A review of concepts. J Hydrol (Amst). 2005;314:275–88.
                                • 105. Chatfield C. The analysis of time series: An introduction. 4th Ed. Chapman and Hall; 1989.
                                  • 106. Agostinelli C, Lund U. R package ‘circular’: Circular Statistics (version 0.4-93); 2017. Available from: https://r-forge.r-project.org/projects/circular/
                                    • 107. Baskerville GL. Spruce Budworm: Super Silviculturist. The Forestry Chronicle. 1975;51(4):138–40.
                                    • 108. Cooke BJ, Nealis VG, Régnière J. Insect defoliators as periodic disturbances in northern forest ecosystems. In: Johnson EA, Miyanishi K, editors. Plant disturbance ecology: the process and the response. Amsterdam, the Netherlands: Elsevier. 2007. p. 487–525.
                                      • 109. Royama T, Eveleigh ES, Morin JRB, Pollock SJ, McCarthy PC, McDougall GA. Mechanisms underlying spruce budworm outbreak processes as elucidated by a 14-year study in New Brunswick, Canada. Ecol Monogr. 2017;87(4):600–31.
                                      • 110. Blais JR. Epidemiology of the spruce budworm in western Ontario: A discussion. The Forestry Chronicle. 1985;61(6):494–8.
                                      • 111. Bennett KD, Buck CE. Interpretation of lake sediment accumulation rates. Holocene. 2016;26(7):1092–102.
                                      • 112. Higuera PE, Sprugel DG, Brubake LB. Reconstructing fire regimes with charcoal from small-hollow sediments: A calibration with tree-ring records of fire. Holocene. 2005;15(2):238–51.
                                      • 113. MacKinnon WE, MacLean DA. The influence of forest and stand conditions on spruce budworm defoliation in New Brunswick, Canada. Forest Science. 2003;49(5):657–67.
                                      • 114. Campbell E, MacLean DA, Bergeron Y. The severity of budworm-caused growth reductions in balsam fir/spruce stands varies with the hardwood content of surrounding forest landscapes. Forest Science. 2008;54(2):195–205.
                                      • 115. Chen C, Weiskittel A, Bataineh M, MacLean DA. Modelling variation and temporal dynamics of individual tree defoliation caused by spruce budworm in Maine, US and New Brunswick, Canada. Forestry. 2019;92:133–45.
                                      • 116. Mageroy MH, Parent G, Germanos G, Giguère I, Delvas N, Maaroufi H, et al. Expression of the β-glucosidase gene Pgβglu-1 underpins natural resistance of white spruce against spruce budworm. Plant J. 2015;81(1):68–80. pmid:25302566
                                      • 117. Méndez-Espinoza C, Parent GJ, Lenz P, Rainville A, Tremblay L, Adams G, et al. Genetic control and evolutionary potential of a constitutive resistance mechanism against the spruce budworm (Choristoneura fumiferana) in white spruce (Picea glauca). Heredity (Edinb). 2018;121(2):142–54. pmid:29453424
                                      • 118. Krause C, Gionest F, Morin H, MacLean DA. Temporal relations between defoliation caused by spruce budworm (Choristoneura fumiferana Clem.) and growth of balsam fir (Abies balsamea (L.) Mill.). Dendrochronologia. 2003;21(1):23–31.
                                      • 119. Reams GA, Brann TB, Halteman WA. A nonparametric survival model for balsam fir during a spruce budworm outbreak. Can J For Res. 1988;18:789–95.
                                      • 120. Lavoie J, Montoro Girona M, Morin H. Vulnerability of conifer regeneration to spruce budworm outbreaks in the eastern Canadian boreal forest. Forests. 2019;10:850.
                                      • 121. Bauce É, Crépin M, Carisey N. Spruce budworm growth, development and food utilization on young and old balsam fir trees. Oecologia. 1994;97(4):499–507. pmid:28313739
                                      • 122. Swetnam TW, Lynch AM. Multicentury, regional-scale patterns of western spruce budworm outbreaks. Ecol Monogr. 1993;63(4):399–424.
                                      • 123. Conedera M, Tinner W, Neff C, Meurer M, Dickens AF, Krebs P. Reconstructing past fire regimes: methods, applications, and relevance to fire management and conservation. Quat Sci Rev. 2009;28:555–76.
                                      • 124. Yamamura K. Biodiversity and stability of herbivore populations: influences of the spatial sparseness of food plants. Popul Ecol. 2002;44:33–40.
                                      • 125. Otway SJ, Hector A, Lawton JH. Resource dilution effects on specialist insect herbivores in a grassland biodiversity experiment. J Anim Ecol. 2005;74:234–40.
                                      • 126. Bognounou F, De Grandpré L, Pureswaran DS, Kneeshaw D. Temporal variation in plant neighborhood effects on the defoliation of primary and secondary hosts by an insect pest. Ecosphere. 2017;8(3):e01759.
                                      • 127. Turekian KK, Nozaki Y, Benninger LK. Geochemistry of atmospheric radon and radon products. Annu Rev Earth Planet Sci. 1977;5:227–55.
                                      • 128. Rangarajan C, Madhavan R, Gopalakrishnan SS. Spatial and temporal distribution of Lead-210 in the surface layers of the atmosphere. J Environ Radioact. 1986;3:23–33.
                                      • 129. Appleby PG. Three decades of dating recent sediments by fallout radionuclides: a review. Holocene. 2008;18(1):83–93.
                                      • 130. Kirchner G. 210Pb as a tool for establishing sediment chronologies: examples of potentials and limitations of conventional dating models. J Environ Radioact. 2011;102(5):490–4. pmid:21145144
                                      • 131. Baskaran M, Nix J, Kuyper C, Karunakara N. Problems with the dating of sediment core using excess (210)Pb in a freshwater system impacted by large scale watershed changes. J Environ Radioact. 2014;138:355–63. pmid:25085208
                                      • 132. Delbono I, Barsanti M, Schirone A, Conte F, Delfanti R. 210Pb mass accumulation rates in the depositional area of the Magra River (Mediterranean Sea, Italy). Continental Shelf Research. 2016;124:35–48.
                                      • 133. Atsatt PR, O’dowd DJ. Plant defense guilds. Science. 1976;193(4247):24–9. pmid:17793989
                                      • 134. Cotton-Gagnon A, Simard M, De Grandpré L, Kneeshaw D. Salvage logging during spruce budworm outbreaks increases defoliation of black spruce regeneration. For Ecol Manage. 2018;430:421–30.
                                      • 135. Morris JL, le Roux PC, Macharia AN, Brunelle A, Hebertson EG, Lundeen ZJ. Organic, elemental, and geochemical contributions to lake sediment deposits during severe spruce beetle (Dendroctonus rufipennis) disturbances. For Ecol Manage. 2013;289:78–89.
                                      • 136. Parducci L, Bennett KD, Ficetola GF, Alsos IG, Suyama Y, Wood JR, et al. Ancient plant DNA in lake sediments. New Phytol. 2017;214(3):924–42. pmid:28370025
                                      • 137. Rijal DP, Heintzman PD, Lammers Y, Yoccoz NG, Lorberau KE, Pitelkova I, et al. Sedimentary ancient DNA shows terrestrial plant richness continuously increased over the Holocene in northern Fennoscandia. Sci Adv. 2021;7(31):eabf9557. pmid:34330702
                                      • 138. de Bruyn M, Hoelzel AR, Carvalho GR, Hofreiter M. Faunal histories from Holocene ancient DNA. Trends Ecol Evol. 2011;26(8):405–13. pmid:21529992
                                      • 139. Clarke CL, Edwards ME, Brown AG, Gielly L, Lammers Y, Heintzman PD, et al. Holocene floristic diversity and richness in northeast Norway revealed by sedimentary ancient DNA (sedaDNA) and pollen. Boreas. 2018;48(2):299–316.
                                      • 140. Liu S, Stoof-Leichsenring KR, Kruse S, Pestryakova LA, Herzschuh U. Holocene vegetation and plant diversity changes in the north-eastern Siberian treeline. Frontiers in Ecology and Evolution. 2020;8:560243.