Return of the moth: rethinking the effect of climate on insect outbreaks
The sudden interruption of recurring larch budmoth (LBM; Zeiraphera diniana or griseana Gn.) outbreaks across the European Alps after 1982 was surprising, because populations had regularly oscillated every 8–9 years for the past 1200 years or more. Although ecophysiological evidence was limited and underlying processes remained uncertain, climate change has been indicated as a possible driver of this disruption. An unexpected, recent return of LBM population peaks in 2017 and 2018 provides insight into this insect’s climate sensitivity. Here, we combine meteorological and dendrochronological data to explore the influence of temperature variation and atmospheric circulation on cyclic LBM outbreaks since the early 1950s. Anomalous cold European winters, associated with a persistent negative phase of the North Atlantic Oscillation, coincide with four consecutive epidemics between 1953 and 1982, and any of three warming-induced mechanisms could explain the system’s failure thereafter: (1) high egg mortality, (2) asynchrony between egg hatch and foliage growth, and (3) upward shifts of outbreak epicentres. In demonstrating that LBM populations continued to oscillate every 8–9 years at sub-outbreak levels, this study emphasizes the relevance of winter temperatures on trophic interactions between insects and their host trees, as well as the importance of separating natural from anthropogenic climate forcing on population behaviour.
KeywordsEuropean Alps Dendroecology Insect outbreaks North Atlantic Oscillation Population cycles Zeiraphera diniana or griseana
Periodic larch budmoth (LBM; Zeiraphera diniana or griseana Gn.) outbreaks are a classic example of population cycles (Baltensweiler and Rubli 1999), with densities regularly oscillating every 8–9 years, ranging from ~ 1 to 20,000 larvae per host tree (Larix decidua Mill.). Recurring LBM epidemics, within an altitudinal range of ~ 1700 to 2000 m asl (Johnson et al. 2010), have been reconstructed without clear interruption over several centuries across the Alpine arc (Büntgen et al. 2009; Hartl-Meier et al. 2017; Kress et al. 2009; Nola et al. 2006; Rolland et al. 2001; Saulnier et al. 2017), and back to mediaeval times in the Swiss Lötschental (Esper et al. 2007). Periodic outbreaks of this foliage-feeding Lepidoptera species can affect the functioning and productivity of larch forest ecosystems (Berryman 2002), with an estimated reduction in aboveground biomass of > 1100 kg ha−1 in the first 3–4 years after forest defoliation (Peters et al. 2017). Geographical patterns of twentieth century LBM outbreaks across the Alpine arc indicate recurring outbreaks likely appear as eastward travelling waves from western epicentres in the southern French Alps to subalpine regions in central and eastern Austria (Bjørnstad et al. 2002; Johnson et al. 2004). Despite extensive demographic studies of LBM oscillations since the mid-twentieth century (Baltensweiler et al. 1977), and comprehensive tree ring-based evidence of historical population peaks during the past three centuries (Büntgen et al. 2009), the relationship between biotic and abiotic factors responsible for the system’s stable periodicity is not fully understood (Baltensweiler and Rubli 1999; Turchin et al. 2003).
The unprecedented disappearance of widespread LBM forest defoliation since the late-1980s was unexpected (Wermelinger et al. 2018). While the hiatus of three Alpine-wide outbreaks from circa 1990 to 2010 may have afforded aesthetic benefits to residents and possibly even benefitted tourism (Wermelinger et al. 2018), the system’s failure has been interpreted as an example of an extraordinary breakdown of ecological behaviour in response to global warming (Esper et al. 2007; Ims et al. 2008). However, it remains unclear how, if at all, climate affects the intensity, frequency, and persistence of cyclic LBM population outbreaks. Though still debatable, previous work suggests increased winter temperatures lead to decreased egg survival during diapause (Baltensweiler 1993), and the temporal offset between larval and foliar development may dampen LBM population growth (Benz 1974; Wermelinger et al. 2018). Moreover, it has been argued that rising temperatures can shift the insect’s outbreak epicentres to higher elevations (Johnson et al. 2010). As climates warm, conditions for optimal LBM growth may move upwards, above the current distribution of dense and widespread subalpine larch forests, where sufficient foliage resources are lacking (Hartl-Meier et al. 2017). Despite numerous studies on a variety of biotic and abiotic factors that potentially influence the LBM system (Büntgen et al. 2009; Hartl-Meier et al. 2017; Johnson et al. 2010; Kress et al. 2009; Nola et al. 2006; Rolland et al. 2001; Saulnier et al. 2017), the identity of climatic controls on the sudden ‘collapse’ of widespread LBM outbreaks after 1982 remains a mystery.
Here, we combine high-resolution meteorological and dendrochronological measurements to report on the unexpected recent return of LBM outbreaks in 2017 and 2018 at different sites of the Alpine arc. We argue that even though widespread insect defoliation of subalpine larch trees across the European Alps has been absent for nearly 4 decades, LBM populations in this region most likely continued to oscillate every 8–9 years at sub-outbreak levels. Moreover, we suggest the current LBM outbreak conceivably occurred at higher elevations, and that the period, during which previous LBM outbreaks were lacking since 1982, was characterised by a climatic regime shift over the North Atlantic/European sector that resulted in particularly warm winters. Our study not only tests and supports the three interrelated theories of warming-induced elevational range shifts, phenological mismatch and trophic disruption, but also implicates the ecophysiological importance of winter temperatures for forest insect population outbreaks.
Materials and methods
Insect outbreak data
To investigate the timing and magnitude of past LBM defoliation impacts on larch growth, incremental measurements of wood production were obtained using a combination of dendrochronological and wood anatomical methods (Büntgen 2019). After the 2017 growing season (20-Sep-2017), 23 individual larch trees from around 1900 m asl, on the north-exposed slope of the Swiss inner-alpine Lötschental were sampled (Esper et al. 2007; Peters et al. 2017). From one core (tree 04b), a continuous thin section of 10–25 µm was cut with a sliding microtome (Gärtner and Nievergelt 2010). A combination of water and pure alcohol was used to moisten the cross-sectional surface, and a drop of glycerol was applied to prevent drying. The sample was bleached for 5 min with Eau de Javel (sodium hypochlorite and potassium hypochlorite), then double stained with a 1:1 Safranin and Astra-Blue solution that colours lignified structures red and non-lignified material blue (Gärtner and Schweingruber 2013). Surplus stain was removed with ethanol to further dehydrate the thin sections, which we finally embedded in Canada balsam under a permanent cover glass and dried for 12–24 h at 60 °C. All remaining larch samples were mounted and polished with sand paper of progressively finer grain size up to 800 grit. Ring width of these samples were measured on a Velmex Tree Ring Measuring System with a resolution of 0.001 mm (Velmex Inc., Bloomfield, NY, USA).
Climate response analysis
Monthly mean temperatures (as well as the corresponding maximum and minimum values) and monthly precipitation totals were compiled from a high-resolution, gridded (0.25° × 0.25°) dataset of surface measurements (E-OBS v17.0) (Haylock et al. 2008), which were averaged over the greater study area (6–12° East and 46–47° North). Measurements from local meteorological stations in the eastern Swiss Alps were used for validation of the gridded product (Büntgen et al. 2014). Monthly mean temperatures, recorded at high-elevation stations in Grisons, were strongly correlated with the temperature indices from the nearest E-OBS grid cell (r > 0.9). A tipping- or breakpoint timeseries analysis that identifies periods at which statistical properties and patterns of a timeseries change (Rodionov 2004), was applied on the individual meteorological data for detecting possible regime shifts. The breakpoint test is a sequential version of the cumulative deviations of the means test (Rebstock 2002), combined with the t test (Rodionov 2004). One can think of it more simply as a sliding t test able to detect multiple inflection points in a timeseries. The gridded temperature datasets, as well as the monthly resolved indices of the North Atlantic Oscillation (NAO; Hurrell 1995) were used for comparison with LBM population dynamics since 1950. The NAO is a major source of seasonal to interdecadal variability in the Earth’s climate system (Hurrell 1995), and its climate effect is most pronounced in winter. A positive winter index (NAO +) produces increased westerlies and, consequently, mild and wet November–February conditions in central Europe, including the Alpine arc. In contrast, when the index is low (NAO−), westerlies are suppressed, and much of central Europe is dominated by cold and dry winters. However, a strong high-pressure influence from the east can result in a persistent blocking of westerly zonal flow. The NAO indices were extracted from NOAA’s Climate Prediction Centre (CPC; Chen and van den Dool 2003). Phenological observations of larch foliage expansion were extracted from four high-elevation sites in the Swiss Alps between 1500 and 1800 m asl (Davos, Lenzerheide, Pontresina and St. Moritz). It should be noted that historical long-term observations of larch and LBM phenology originated from the host dominated subalpine forest belt in the eastern Swiss Alps, which has been identified previously as a key biogeographic hotspot for cyclic LBM outbreaks (Baltensweiler and Rubli 1999). This region has further been the geographical focus of considerable LBM population research in the twentieth century (see Baltensweiler et al. 2008 and references therein). All continuous timeseries were expressed as day of the year (DOY). Though encompassing periods both with and without LBM outbreaks, the temporal availability of precise LBM and meteorological data restricts our analyses to post-1950.
LBM outbreaks during colder winter climates
Mean Fig. 3
LBM cycles and winter climates
In addition to distinct year-to-year winter temperature variability since 1950, the most obvious feature is the regime shift from generally cooler winters before 1987 to mostly warmer conditions afterwards (Fig. 3b). With a temperature range from − 0.46 to 0.21 °C (Fig. S1b), all monthly winter combinations reveal substantial temperature differences between the cooler high-peak and warmer low-peak outbreak period. The recent Alpine winter cooling originates from a large-scale mode change over the North Atlantic/European sector. Known as the ‘Beast from the East’, exceptionally cold Arctic airmasses with an anticyclonic structure were stretching from Siberia to the British Isles, covering large parts of Asia and almost all of Europe from 24-Feb-2018 to the second half of March 2018. Apart from the remarkable agreement between cold winters and the occurrence of LBM outbreaks, no other season (e.g. summer), or meteorological parameter (e.g. precipitation), was found to be statistically related. Differences between cold November–February temperatures and relatively warm April conditions, however, reveal specifically large variances before and during the four LBM outbreaks that were recorded between 1953 and 1982, and again in 2018 (Fig. S1a). The three LBM cycles that were not associated with extensive population outbreaks between 1998 and 2009 were characterised by a preponderance of mild winters and overall small seasonal temperature differences.
Our study describes a remarkable, recent, eco-climatic phenomenon: nearly four decades ago, population and forest ecologists were aghast when the regular, clockwork-like cycles of landscape-scale defoliation caused by the LBM in the subalpine larch belt of the European Alps suddenly stopped. Based on the varying widths of precisely dated tree rings and their mean site chronologies, it was concluded that LBM outbreaks had persisted for more than 1200 years (Esper et al. 2007), so it was rather unexpected that intensive population peaks, and their subsequent forest defoliation and discoloration impact, failed to materialize after 1981. Despite much speculation about the apparent ‘collapse’ of LBM outbreaks (Büntgen et al. 2009; Esper et al. 2007; Hartl-Meier et al. 2017; Ims et al. 2008; Johnson et al. 2010; Kress et al. 2009; Nola et al. 2006; Rolland et al. 2001; Saulnier et al. 2017), the underlying ecological processes causing this change remained unresolved.
Based on yearly larval counts and visual observation of defoliation in the eastern Swiss Alps (Grisons) (Wermelinger et al. 2018), as well as yearly counts of larval density in a permanent plot at 1800 m asl in the southern French Alps (Les Combes, Briançon) (Roques and Goussard 1982), our study—though limited in space and time—reveals that LBM population oscillations most likely did not stop during the last decades. This detracts from the postulated ‘collapse’ of cyclic population dynamics in response to climate change (Ims et al. 2008). Since some of the contemporary defoliation reports are from relatively high-altitudes between 1600 and 2000 m asl, they are indicative of outbreaks near the upper distributional limit of extensive larch stands. This finding corroborates predictions of temperature-mediated elevational shifts in LBM outbreak epicentres (Johnson et al. 2010). From the Little Ice Age until the mid-1980s, the optimal elevational range for LBM outbreaks was likely below the upper limit of extensive subalpine larch forests. Since the 1990s, the temperature sensitive LBM range though possibly exceeded the host resource distribution by moving upwards into the predominantly treeless Alpine zone. The recent return of LBM outbreaks does not exclude a downward shift in epicentres from above the ecotone back into the subalpine forest zone. Dendrochronological fingerprints of the four LBM outbreak peaks between 1953 and 1982 (Fig. 2j–k) suggest the current population boost will also cause abrupt reductions in the ring width and/or latewood density of its subalpine larch hosts. In line with dendrochronological findings from the French Alps (Saulnier et al. 2017), our tree-ring evidence from the Swiss Lötschental confirms the absence of strong outbreaks after 1982 (Fig. 2l), the period for which systematically collected, quantitative LBM data are limited for most of the Alpine arc, including Switzerland.
Our study argues for the often-neglected relevance of cold season climate for organismal responses (Williams et al. 2015; Büntgen and Krusic 2018), most pronounced at higher latitudes and elevations, where the impacts of cold season temperature and precipitation persist through most of the year. Trends and extremes in winter climate, which usually exceed those during other months (Renner and Zohner 2018), impact species-specific chilling requirements, the risk of frost injury, demands on energy and water balance, as well as phenological synchrony and the composition and interactions within communities (Williams et al. 2015; Renner and Zohner 2018). In addition to describing uneven rates of change in the phenology of interacting species, our results provide evidence of the persistence of LBM cycles during recent decades. While this finding contradicts arguments for a strong vulnerability of population cycles to climate (Ims and Fuglei 2005; Ims et al. 2008), it does not imply that transient population dynamics remain unaffected (i.e. although the system’s cyclic behaviour continued during the last decades, the amplitude of oscillations remained low; Fig. 2l). Instead, we argue for the importance of shifts in the NOA that can, at least sporadically, dampen or even reverse the ecological effects of increasing temperatures. Since internal climate oscillations play a critical role for cold season ecosystem functioning in central Europe (Stenseth et al. 2002), consideration of its teleconnection patterns is important for disentangling the effects of naturally and anthropogenically forced climate change (Rosenzweig and Neofotis 2013). In terms of ‘detection and attribution’—a modern principle in climate change research—improved understanding of the sensitivity of trophic interaction in biological and ecological systems to abiotic forcing factors will better inform and guide local to regional management strategies, national and international conservation programmes, and more generally, global environmental policy. Nevertheless, our study suggests the complexity of scale-dependent interactions between different, overlapping climate forcing agents. While a pause in recurring, geographically extensive LBM outbreaks after 1981 describes a response to anthropogenic global warming (Esper et al. 2007; Johnson et al. 2010; Saulnier et al. 2017), the recent return of a high LBM populations points towards the importance of synoptic-scale, internal climate variability versus externally forced, large-scale climate change (Stenseth et al. 2002). Unfortunately, our limited understanding of the spatiotemporal characteristics of LBM outbreaks prior to the mid-twentieth century hinders similar analyses of LBM dynamics further back in time.
In conclusion, tree-ring evidence from the Swiss Alps suggests that larch budmoth outbreaks have occurred every 8–9 years over the past 1200 years. The occurrence of Alpine-wide LBM outbreaks, however, ceased after 1982 but the underlying causes of this change have been uncertain. The sudden return of LBM outbreaks in 2017 and 2018 might be related to anomalous cold European winters due to a persistent negative phase of the NAO, which almost resembles conditions of the four consecutive insect outbreaks between 1953 and 1982. In contrast, three warming-induced mechanisms (1–3 or a combination thereof) help explain the disruption of LBM outbreaks: (1) high egg mortality, (2) asynchrony between egg hatch and foliage growth, and (3) upward shifts of outbreak epicentres. Although we demonstrate that the LBM cycle did not stop during the past decades, our findings stress the climate sensitivity of transient population dynamics. By providing a strong negative relationship between winter temperatures and LBM outbreaks, this study not only overcomes previous speculations, but also underlines the ecophysiological importance of separating natural from anthropogenic climate forcing of trophic interactions.
We dedicate this study to the late Werner Baltensweiler (1926–2008). C. Körner (Basel) and M. Carrer (Padua) kindly provided information on local LBM outbreaks that occurred in 2018 in different parts of the Italian Alps. We are particularly thankful to J.-B. Daubrée, T. Bélouard and staff of the French Forest Health Department for delivering LBM defoliation data from the French Alps. J. Giraud, P. Vigne and the municipality of L’Argentière-la-Bessée kindly offered logistic assistance during intensive fieldwork.
Author contribution statement
BW and DN initiated the study and UB designed and performed the analyses. BW and AR compiled recent LBM evidence. UB, AL, PJK and JE wrote the paper with input from all others. Each author provided discussion and approved submission.
UB received funding from “SustES - Adaptation strategies for sustainable ecosystem services and food security under adverse environmental conditions” (CZ.02.1.01/0.0/0.0/16_019/0000797), whereas AL was supported by OP RDE grant EVA4.0 (CZ.02.1.01/0.0/0.0/16_019/0000803).
Compliance with ethical standards
Conflict of interest
Authors declare that they have no conflict of interests.
- Baltensweiler W, Rubli D (1999) Dispersal – an important driving force of the cyclic population dynamics of the larch bud moth. Forest Snow Landsc Res 74:3–153Google Scholar
- Benz G (1974) Negative Rückkoppelungen durch Raum- und Nahrungskonkurrenz sowie zyklische Veränderung der Nahrungsgrundlage als Regelprinzip in der Populationsdynamik des Grauen Lärchenwicklers, Zeiraphera diniana (Guenée) (Lep., Tortricidae). Zeitschrift für Angewandte Entomologie 76:196–228CrossRefGoogle Scholar
- Berryman A (2002) Population cycles: the case for trophic interactions. Oxford University Press, OxfordGoogle Scholar
- Büntgen U, Krusic PJ (2018) Non-traditional data and innovative methods for autumn climate change ecology. Climate Res 75:215–220Google Scholar
- EUFORGEN (2018) http://www.euforgen.org/distribution_maps.html. Accessed 6 Sep 2018
- Gärtner H, Schweingruber FH (2013) Microscopic preparation techniques for plant stem analysis. Kessel, RemagenGoogle Scholar
- Roques A, Goussard F (1982) Amélioration de la technique de prévision des pullulations de la tordeuse du Mélèze, Zeiraphera diniana Guénée (Lép. Tortricidae). Acta Oecologica Oecologia Applicata 3:35–45Google Scholar
- Wermelinger B, Forster B, Nievergelt D (2018) Cycles and importance of the larch budmoth. WSL Fact Sheet 61:1–12Google Scholar
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.