Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Climate change and invasions by non-native species each constitute major threats to forest ecosystems worldwide (Dale et al. 2001), yet attempts to synthesize broadly their interacting impacts are limited (Engel et al. 2011; Hellmann et al. 2008; Walther et al. 2009). The inherent complexity of each and the potential confounding and compounding effects when considered jointly have hampered analysis of actual and potential effects on forest ecosystems. Another consideration is that the paramount stage of biological invasions, the arrival stage, is predominately a result not of climate change but of increasing global trade through which new species ‘hitchhike’ on products, nursery stock, packing materials, and in ship hulls and ballast water (Hulme et al. 2008; Lockwood et al. 2007). Consumer demand for many foreign manufactured goods is also not generally dependent upon or influenced by climate, at least not in the short-term. In fact, one could argue that due to changing climates, there could be a greater emphasis on reducing carbon footprints associated with foreign imports with a consequent increase in the desire to purchase locally available products that in turn could conceivably reduce the volume of imports and species that are introduced with these imports.

In the absence of an environmental revolution and major shifts in economic patterns and in governmental policy, however, the arrival of new species will likely continue to increase as long as global trade increases. Once a new species arrives to a novel habitat, the other stages of the invasion process, establishment, spread, and impact, could be directly and indirectly influenced positively or negatively by climate and changes in climate. In this chapter, we first briefly describe the causes and dynamics of biological invasions and climate change independently. We then focus on how these forces interact to affect forest ecosystems by mediating range shifts, altering population outbreak frequency and intensity, changing seasonality and voltinism, and decoupling interactions between and among trophic levels. In some cases, we rely on examples from native species in native ranges, and then extrapolate these observations to non-native species in non-native habitats. We lastly consider the economic ramifications of the interaction between biological invasions and climate change.

2 Dynamics of Biological Invasions

Biological invasions consist of four distinct stages: arrival, establishment, spread, and impacts (Lockwood et al. 2007). The arrival stage is defined as the movement of a species from an area in which it is established to a novel habitat. Although the interchange of biota among biogeographic regions is integral to the history of life (Crosby 1986; di Castri 1989), arrival rates of new species have increased dramatically in recent decades due to increases in global trade and travel (Aukema et al. 2010). Establishment after arrival is a critical transition, and it is believed that most arriving species fail to establish (Simberloff and Gibbons 2004). One of the most important factors influencing the establishment of newly arrived species is propagule pressure: the number of individuals in the arriving founder population, the number of independent introductions, or a combination of both (Lockwood et al. 2007). For example, low-density populations tend to be subject to environmental and demographic stochasticity, and Allee effects (positive density-dependence, Stephens et al. 1999). Allee effects can arise when individuals in sparse populations encounter difficulties finding suitable mates, satiating natural enemies, and overcoming host defense mechanisms, all of which can exacerbate the challenges that small founder populations already face during the establishment phase (Liebhold and Tobin 2008; Taylor and Hastings 2005; Tobin et al. 2011). Other factors influencing establishment success include the level of genetic diversity in founder populations, availability of resources, degree of habitat disturbance, and presence or absence of competitors, mutualists, and natural enemies (Lockwood et al. 2007). Another factor that is important to consider in the context of this chapter is climate suitability (Beaumont et al. 2009; Hayes and Barry 2008; Thuiller et al. 2005), and thus shifting climatic patterns could have critical consequences for the establishment success of invaders.

Following successful establishment, a species often begins to expand its geographic range, often through a process known as stratified dispersal in which local growth and diffusive spread is coupled with long-distance population ‘jumps’ (Hengeveld 1989). These population jumps are often facilitated through anthropogenic (e.g., Gilbert et al. 2004), hydrological (e.g., Davidson et al. 2005), and atmospheric transport mechanisms (e.g., Isard et al. 2005), and can greatly accelerate the rates of spread of invading species (Shigesada and Kawasaki 1997). Long-range dispersal can also serially initialize new invasions, which are effectively subject to the same constraints that affect the establishment success of any newly-arrived species in a non-native environment (Liebhold and Tobin 2008; Taylor et al. 2004).

The fourth stage of the invasion process is the ecological and economic impacts of non-native species (Lockwood et al. 2007). Impacts due to invasive species can vary dramatically among species, and within a species depending upon the region being invaded. A recent analysis on non-native insects in forests within the continental United States has suggested that only a minority of introduced and established insects (≈14 %) since 1860 have caused major damage (Aukema et al. 2010), yet this minority of species can still account for several billion USD in costs (Aukema et al. 2011; Holmes et al. 2009; Pimentel et al. 2005). The costs to the United States of the non-native emerald ash borer, Agrilus planipennis Fairmaire, which are primarily associated with the treatment, removal, and replacement of ash, Fraxinus spp., is predicted to be 10.7 billion USD alone (Kovacs et al. 2010). It is also likely that the impacts and consequent costs due to non-native species will be affected by climate and changes in those regimes.

3 Climate Change

Recent increases in the concentrations of atmospheric greenhouse gases, most notably carbon dioxide, methane, and nitrous oxide, have led to a corresponding change in local and global climates. Global mean surface temperatures have increased by 0.3-0.6 °C over the last century (Mann et al. 1998), and global temperatures are projected to continue to increase by the end of the next century (Intergovernmental Panel on Climate Change 2007). The projected increase in global surface temperatures is thought to range from 1 °C under a low (B1) greenhouse gas emission scenario, which assumes substantial mitigation and reductions in greenhouse gas emissions, to 6 °C under a high (A1fi) greenhouse gas emission scenario, which assumes temperatures under the current conditions and without any mitigating strategies (Hayhoe et al. 2007; Kunkel et al. 2008).

The increase in global surface temperatures can also result in a number of cascading effects. For example, as temperatures warm, there could be changes in hydrological cycles leading to increases or decreases in precipitation patterns, such that some places could be more prone to drought while others more prone to flooding. Changes in climatic variability, including increases in storm event frequency and intensity, could also be the result of recent warming trends (Rosenzweig et al. 2001). More subtle changes, such as the diminishment of winter snow pack, could also have important ecological ramifications for both insects and trees. Consequently, climate change has the potential to destabilize many ecosystem functions, and cause major changes to the dynamics of individual species and to those communities in which they reside.

Insect species are particularly sensitive to climate change because many of their physiological processes are temperature-dependent. Insects could respond to changing climates by going extinct, moving into areas with a more tolerable weather regime, or adapting in situ. Climate change can affect insect species differentially depending upon the latitude at which they live; for example, temperate insects that evolved under strong seasonality could inherently have greater phenotypic plasticity for withstanding the pressures of climate change relative to tropical species already near their thermal tolerances, and thus could be more capable of adapting to warming trends (Deutsch et al. 2008). Such changes could include accelerated developmental rates with a consequent change in the timing of year-to-year phenological events, and/or in the number of generations per year. Species abundance can be directly affected by temperature, resulting in population density extremes, from extinction to outbreaks. The distributional ranges of insects can also be affected by the removal of biogeographic boundaries formed by climatic factors.

4 Ecological Interactions Between Forest Insect Invasions and Climate Change

We now turn our attention to the effects, both documented and those projected to occur, of climate change on insect invasions in forest ecosystems. Although non-native species across a diversity of taxa have been introduced to novel areas (Pimentel 2002; Simberloff and Rejmánek 2011), forest insects, such as xylophagous species, represent a particularly important group of invaders because of their propensity for importation in solid wood packaging material, wood dunnage, and wood pallets (Bartell and Nair 2003; Brockerhoff et al. 2006a; Liebhold et al. 1995; Wingfield et al. 2010), all of which are important components in global trade. Forest insects can also be introduced through imports of infested timber and forest products, as well as on infested live plants (Brasier 2008; Liebhold et al. 2012; Reichard and White 2001). In recent decades, a number of economically damaging, non-native forest insect species have been or were likely introduced through global trade routes, such as A. planipennis (Poland and McCullough 2006) and Anoplophora glabripennis (Motschulsky) (Smith et al. 2009) in North America, Dendroctonus valens LeConte in China (Yan et al. 2005), and Sirex noctilio F. in Australasia, South Africa, South America and most recently, North America (Slippers et al. 2012a).

4.1 Changes in Geographic Ranges

There exists a rich literature around bioclimatic envelope modeling documenting how the distribution and abundance of forest insects are restricted to (or released from) suitable habitat(s) by various climatic factors in time and space (e.g., Hlasny and Turcani 2009; Jonsson et al. 2011; Logan et al. 2003; Robinet et al. 2007; Williams and Liebhold 1995a). Changes in climate compared to historic norms have relaxed these boundaries, allowing for rapid range expansion in some species, and range retraction in others. In this section, we will discuss three types of changes in the geographic ranges of forest insects: expansions in latitudinal range, expansions in geographic range, and range retractions. We restrict our treatment to natural dispersal, even though human-assisted transport may be a confounding factor in many range expansions (Brockerhoff et al. 2006b) that could be detected post hoc within the genetic structure of subpopulations (Kerdelhué et al. 2009).

In montane ecosystems, range expansions frequently become first apparent in altitudinal changes rather than latitudinal changes. Simulation studies have shown that the mountain pine beetle, Dendroctonus ponderosae Hopkins, which is native to North America, responds to increases in temperature first by increasing its altitudinal range before becoming more apparent at more northern locations in western Canada (Sambaraju et al. 2012). In the Greater Yellowstone Ecosystem of the western United States, D. ponderosae has had a long association with lodgepole pine forests with infrequent outbreaks at climatically-inhospitable high elevations. Recently, however, it has exhibited behavior outside of its observed range of natural variability, and threatens to decimate high-elevation five needle pines (Logan et al. 2010; Logan and Powell 2001), a host species that does not appear to have a co-evolved relationship with this insect. These pronounced shifts in elevation are not restricted to bark beetles. In recent years, for example, the winter moth, Operophtera brumata (L.), has defoliated stands of mountain birch at elevations up to the tree line in coastal areas of northern Norway (Hagen et al. 2007). Historically, O. brumata outbreaks have occurred at lower elevations due to bioclimatic and/or competitive effects with a sympatric species, the autumn moth Epirrita autumnata (Borkhausen) (Ammunet et al. 2010). The gypsy moth, Lymantria dispar (L.), also a defoliator, is expected to continue a progression into regions of higher elevations in Europe as it doubles its range compared to 50 years ago (Hlasny and Turcani 2009).

Latitudinal range shifts are perhaps more dramatic than altitudinal shifts, as humans may notice defoliated or killed trees more quickly in regions where such activity was historically unapparent. There exist a number of examples of lepidopteran defoliators in Europe where this phenomenon has occurred over the past decade. While O. brumata has expanded into higher elevations and moved northeast across Norway, the more cold-hardy E. autumnata has exhibited a concomitant increase into colder, continental areas (Jepsen et al. 2008). Recent phenological changes in budburst have facilitated a rapid northward range expansion of another geometrid, the scarce umber moth, Agriopis aurantiaria Hübner, in the same region (Jepsen et al. 2011). In France and Portugal, higher elevation and northward latitudinal incursions have been documented for the pine processionary moth, Thaumetopoea pityocampa (Denis and Schiffermüller), a frequent defoliator of pine and cedar (Arnaldo et al. 2011; Battisti et al. 2005). In North America, a warming climate could facilitate the expansion of L. dispar into areas where current overwintering temperatures are too cold to permit survival (Régnière et al. 2009). However, a constraint in the southern geographic range of L. dispar is the lack of sufficient cooling periods to terminate diapause (Gray 2004); thus, although warming temperatures could allow for northern expansion of L. dispar, there could a concomitant restriction in its potential and realized southern range.

Latitudinal range expansion has not been limited to Coleoptera and Lepidoptera. For example, damage from the poplar woolly adelgid, Phloemyzus passerinii (Signoret), was recently recorded for the first time in northern France (Rouault et al. 2006). The European pine sawfly, Neodiprion sertifer (Geoffroy), is currently limited in its range by the degree of freeze tolerance of its eggs, is also expected to move northwards over the next decades (Veteli et al. 2005). Lethal mortality to overwintering life stages is frequently cited as a key delimiter for range margins (Bale and Hayward 2010), which, once ameliorated, could permit rapid expansion in a number of species.

A notable example of climatic release permitting range expansion has been the recent outbreak of D. ponderosae in western Canada (Aukema et al. 2006). Mountain pine beetle has been historically limited in range by a −40 °C thermal isocline, restricting its Canadian range to a line primarily west of the Rocky Mountains (Safranyik et al. 1975). Recently, this insect breached the historic geoclimatic divide and is now reproducing in lodgepole pine forests of northwestern Alberta (de la Giroday et al. 2011, 2012; Robertson et al. 2009). This is of grave concern because range expansions of this magnitude could permit access to new hosts and new habitat corridors. For example, D. ponderosae is now reproducing in an area where lodgepole pine hybridizes with jack pine (Cerezke 1995; Cullingham et al. 2011), which could permit further range expansion eastward through the boreal forest of North America. This insect seems well adapted to jack pine, and exhibits elevated reproductive rates in evolutionarily naïve hosts (Cudmore et al. 2010).

Of course, climatic suitability across the extent of a new range is never assured, and different models could have contradicting outcomes. Projections of habitat suitability for D. ponderosae across the boreal forest of Canada through global simulation models show a wide range of projections, from high to low climatic suitability, depending upon the scenario applied (Safranyik et al. 2010). Projections of climatic suitability for the same insect in the Rocky Mountain region of the western United States, however, are more uniform, predicting decreases in suitable habitat by up to 50 % by the year 2050. In contrast, projections for the same regions for a different insect, the western pine beetle, Dendroctonus brevicomis LeConte, indicate increases or decreases in suitable habitat, depending on the scenario (Evangelista et al. 2011). Potential range retractions are not restricted to bark beetles. Lymantria dispar and the nun moth, Lymantria monacha (L.), are projected to lose up to 900 km from their southern European ranges as temperatures warm (Vanhanen et al. 2007). Many projections of climate warming demonstrate either rapidly altered ranges of the host trees (Refehldt et al. 2006) or maladaptive seasonal phenology that disrupts insect development and host procurement (Williams and Liebhold 1995a).

Climatically-mediated range shifts could also lead to changes in physiological and morphological traits of insects. Indeed, latitudinal clines in physiological, reproductive, and morphological traits of insects are well known (Blanckenhorn and Demont 2004). For example, wing size tends to be larger at higher latitudes in Drosophila subobscura Collin (Gilchrist et al. 2001) and genetic change in this species appears to be tracking climate change (Balanya et al. 2006). Furthermore, invasive populations of this species formed a latitudinal gradient in wing size very similar to that in native populations in only 20 years, which suggests that some invasive species could evolve rapidly in response to elevated temperatures. A number of widespread native forest Lepidoptera exhibit similar adult size across their latitudinal range but express reductions in fecundity and (or) larger offspring with increasing latitude (Ayres and Scriber 1994; Harvey 1983; Parry et al. 2001). A latitudinal shift in a number of fitness parameters is evident in introduced populations of the fall webworm, Hyphantrea cunea (Drury), in Japan (Gomi 2007; see Sect. 4.3). Because such latitudinal clines appear to be common in many species, selection could be expected to form clines in invasive species as their distributional range expands.

4.2 Changes in Outbreak Frequency and Intensity

The study of forest insect outbreaks has a long and storied history, and many of the species that have been the focal point of these studies have broadly served as model systems for studying the conceptual and mechanistic processes of insect population dynamics (e.g., Barbosa and Schultz 1987). Several forest insect species undergo regular cycles in population density, from innocuous to outbreak and back. Apart from being inherently fascinating, forest insect outbreaks often cause considerable economic and ecological damage by affecting nutrient cycles, animal and plant populations (Frost and Hunter 2004; Payette et al. 2000; Work and McCullough 2000), and human uses of forests for timber and recreation (Coyle et al. 2005). Another aspect of many forest insect outbreaks is that they can often be spatially synchronized (Haynes et al. 2009; Johnson et al. 2005; Peltonen et al. 2002), in which the congruence in the temporal variation of abundance among geographically distinct populations results in outbreaks over a large spatial scale (Bjørnstad et al. 1999). Geographically widespread outbreaks can be particularly important for a number of reasons. First, spatially synchronous outbreaks could dilute regulating effects of natural enemies that could otherwise provide local control (Royama 1984). Second, they can reduce the ecological landscape’s ability for buffering because most areas within an ecosystem experience simultaneous disturbance (Lovett et al. 2002). Lastly, they can overwhelm the budgetary and logistical efforts available for protection of economic assets or ecological functions through suppression programs (Tobin et al. 2012).

Many non-native species are innocuous in their native ranges where they evolved in concert with host tree defensives and the regulatory effects of natural enemies, but can be problematic when introduced into naïve environment and/or to naïve host species (Liebhold et al. 1995; Hu et al. 2009). Such a contrast is apparent when comparing the insect borers A. planipennis and Agrilus anxius Gory. The latter species is native to North America where it is relatively innocuous to North American birch species unless coupled with severe stress such as drought. In contrast, A. anxius causes significant mortality in non-native birch species planted in North America, whether trees were stressed or not (Nielsen 1989). Similarly, A. planipennis rarely causes mortality in ash in its native range in Asia, while all North American ash species appear to be extremely susceptible (Poland and McCullough 2006). Non-native species often lack natural enemies in a new area, and consequently, the ‘enemy release’ hypothesis has been suggested to play a role in the ability of species to invade novel habitats and reach outbreaking densities (Keane and Crawley 2002; Torchin et al. 2003).

The effect of climate change could result in changes to both the outbreak intensity and the periodicity of forest insect outbreaks (Logan et al. 2003; Volney and Fleming 2000). Trends in climate warming are thought to have had direct effect on the development, intensity, and geographic extent of outbreaks of D. ponderosae Hopkins, in North America (Kurz et al. 2008). Although this species is native to western North America, it is now achieving outbreak densities in northern British Columbian forests where it had never previously been found, at least not over the last few centuries of direct observation (see Sect. 4.1). Similarly, outbreaks of other eruptive bark beetle species such as the southern pine beetle, Dendroctonus frontalis Zimmermann, the Mexican pine beetle, Dendroctonus mexicanus (Hopkins), and the European spruce bark beetle, Ips typographus (L.), are expected to exhibit outbreaks of increased magnitude as temperature and precipitation regimes change (Kausrud et al. 2012; Waring et al. 2009).

In Fennoscandia, climate warming is thought to have shifted the geographic distribution of outbreaks of O. brumata (Hagen et al. 2007). Although this species is native to Fennoscandia, it, like D. ponderosae in North America, is believed to have crossed altitudinal barriers that were previously impassable due to climate that was historically unfavorable to its winter survival. In addition to the expanse of areas experiencing outbreaks, the duration of outbreaks is thought to have increased due to climate warming (Jepsen et al. 2008). Examples of the interplay between climate change and outbreaks by non-native species are less documented, likely due to the fact that invasions by non-native forest insects are a more recent phenomenon (Aukema et al. 2010), relative to the time scale that native species have existed in their native ranges. Indeed, many important invasive species are still in the active range expansion phase of their colonization, making it impossible to partition climatic influences separately from other processes driving spread. However, the change in outbreak dynamics in native species is likely not unique and will likely result in related changes in the outbreak dynamics of non-native species.

Not all forest insects necessarily benefit from recent trends in climate change, adding complexity to our efforts to understand expected patterns. For example, regular outbreaks have been recorded for the larch budmoth, Zeiraphera diniana Guénée, in the European Alps (Bjørnstad et al. 2002). A recent dendrological reconstruction of these outbreaks has suggested that this cyclical behavior had been occurring for at least 1,173 years and during previous climatic events, such as periods of warming during the Middle Ages and cooling during the Little Ice Age (Esper et al. 2007). However, since 1981, Z. diniana outbreaks have been conspicuously absent with the supposition that recent trends in climatic warming have upset the balance of a system that previously had exhibited remarkable stability (Esper et al. 2007). However, because the absence of Z. diniana outbreaks is thought to be due to climate-mediated disruption of the stability of this system, non-native species, which are agents of disturbance in themselves, may or may not be less prone to collapse in forest ecosystems that are also experiencing disturbance due to changing climates.

4.3 Changes in Seasonality and Voltinism

Increasing temperatures will have direct consequences for insects including alterations to life cycle duration (developmental rate) and changes in voltinism (the number of generations per year). While most insects are capable of increased growth rates at elevated temperatures, a key factor is during which part of a particular insect’s life history that the temperature change occurs. Thus, generalizations concerning the response of insect growth rate and development to global climate change must be tempered with knowledge that a species may behave idiosyncratically with respect to temperature.

At the level of individual insect species, a major determinant of the response to climatic shifts is the type of life-cycle and the developmental strategy employed. Danks (2006) suggested that insect development could be viewed as either an active default, where it proceeds until some reliable environmental cue signals it to stop, or a passive default, where development stops at a preset point irrespective of current environmental conditions and does not resume again until an appropriate cue is received. A good example of this dichotomy would be a multivoltine species that produces additional generations as long as diapause-inducing cues are absent, while an univoltine species would develop faster in its single generation, but would still be constrained by an obligate diapause to one generation annually. Insects using an active developmental default will likely receive greater benefits from warming temperatures than those with passive default development systems. Because many multivoltine insect species use photoperiodic cues to initiate diapause, which do not change in response to changing climates (Tauber and Tauber 1976), sufficient increases in temperature prior to the onset of diapause-inducing photoperiods could be a key determinant to the number of generations possible per year under future climate scenarios (Chen et al. 2011; Tobin et al. 2008).

Many geographically widespread insects exhibit latitudinal gradients in voltinism (Wolda 1988); thus, shifting temperatures should slide the boundaries between voltinism states in predictable directions. Voltinism may be relatively plastic, and in some species governed by photoperiod, temperature, and host plants, whereas in others it is fixed (Tauber and Tauber 1976). For those species with flexible voltinism, warming may be advantageous, permitting faster growth and additional generations annually (Bale et al. 2002; Tobin et al. 2008). In Europe, extensive data sets encompassing hundreds or even thousands of species of Lepidoptera have allowed comparisons among different time periods. For butterflies and moths with the capacity for multivoltinism, there have been significant increases in the frequency of species exhibiting bi- or multivoltine life cycles, with much of this increase occurring in the last two decades (Altermatt 2010; Pöyry et al. 2011).

Increased voltinism could promote faster population growth because more offspring are being produced per seasonal time period, thus increasing the likelihood of outbreaks of pest species or elevating non-pests or minor pests to a more economically important stature (Steinbauer et al. 2004; van Asch and Visser 2007). In forested ecosystems, changes in the voltinism of Lepidoptera and of Coleoptera (particularly scolytid bark beetles), are of concern as these groups contain some of the most economically damaging forest pests. A number of native bark beetle species in both Europe and North America have shifted lifecycles by adding annual generations, in an apparent response to moderating temperatures at higher latitudes or altitudes (Berg et al. 2006; Jonsson et al. 2009; Werner et al. 2006). With respect to the spruce beetle, Dendroctonus rufipennis (Kirby), this change was associated with devastating outbreaks in Alaska’s Kenai Peninsula (Sherriff et al. 2011).

Despite the apparent high frequency of shifts in voltinism in native insects in many temperate zones globally, documentation of such shifts in invasive forest insects has thus far been rare. It is not known if the apparent rarity in forest insects is an artifact or represents real patterns. There are several possibilities for this absence. First, some invasive species could have a fixed voltinism, such as L. dispar, which is exclusively univoltine. Another possibility is that simply too little is known about the biology of many invasive species. Additionally, invasive species could lack sufficient genetic diversity or plasticity to respond to climate shifts, owing to small founder population size and the relatively short interval of observation. Successful species could have also been introduced with, and indeed may owe their success to, the expression of an appropriate voltinism for a given region.

One well-known case of shifting voltinism concerns H. cunea, a relatively benign defoliating lepidopteran accidentally introduced from North America to Europe and Asia where it has become a major pest. In Japan, the founding population was bivoltine, but within 50 years of introduction and coupled with a southward spread, trivoltine populations became the norm in warmer areas (Gomi and Takeda 1996). The shift to a trivoltine life-cycle was associated with a subtle but biologically significant change in sensitivity to photoperiod (Gomi 2007).

Multivoltinism appears to be rare in eruptive folivores, as most appear to be constrained to the nutritionally superior, but inherently risky (see Sect. 4.4) early season foliage of woody plants (Hunter 1991, 1995). For invasive folivorous species with obligate diapause, such as many univoltine spring-feeders, increasing temperatures could provide respite from natural enemies because development will accelerate through the vulnerable larval period, a function of escape from the trap of slow-growth and high mortality (Benrey and Denno 1997; Zalucki et al. 2002). This, however, makes the assumption that natural enemies will not respond similarly to elevated temperatures, or that they will not quickly adapt to a seasonal shift in prey abundance. Some, but not all, studies have suggested that L. dispar outbreaks are correlated with warmer spring temperatures in the year of, and the year prior to, defoliation (Elkinton and Liebhold 1990), although the mechanism underlying the pattern is not known. Communities of forest Lepidoptera irrespective of taxonomic affinity, especially those that are spring-feeders, exhibit concordant population dynamics, suggesting commonality of either positive or negative responses to a significant environmental driver like meteorological conditions (Raimondo et al. 2004; Stange et al. 2011).

A major concern from an invasive species perspective could be the response of wood borers. Unlike folivorous insects, many wood and cambium feeders have considerable plasticity with respect to voltinism and are constrained mainly by the combination of wood as a nutritionally poor resource and the relatively low temperatures in temperate and boreal forests. Increased voltinism observed in variety of native bark beetles (e.g., Faccoli 2009) could be a harbinger of what to expect in this particular guild. Voltinism in A. glabripennis is a function of latitude in China with southern populations requiring only a single year to complete development (Hu et al. 2009), while populations introduced into North American are variable, with both semivoltine or univoltine emergence recorded. Another invasive cerambycid, the brown spruce long-horned beetle, Tetropium fuscum (Fabricius), currently has a univoltine lifecycle but has been recorded as bivoltine in parts of its native range. In China, D. valens has already devastated vast tracts of Chinese red pine. In its native range in the southern United States, this species has up to three generations per year, but only one and perhaps a partial second generation has been recorded in China (Sun et al. 2004), suggesting that this insect could become an even greater threat in its introduced range under warming temperatures. These examples highlight the flexible voltinism apparent in many wood feeding insects and thus a high propensity to benefit from climatic warming.

4.4 Decoupling Species Interactions

One of the first and probably best-documented effects of anthropogenic driven climate change has been a phenological shift in the seasonal occurrence of a diverse array of organisms. Phenology, the seasonally influenced timing of developmental processes (e.g., Visser et al. 2010), is strongly correlated with temperature regime for many organisms including plants, insects, and vertebrates (Parmesan 2006; Root et al. 2003). In temperate regions, a large number of species have shifted seasonal biological activities such as onset of bud break, flowering time, emergence, or migrating earlier or maintaining activity later in the season as a response to recent changes to the onset of spring and the increasing length of the growing seasons, respectively. For example, the spring phenology of European Lepidoptera has advanced significantly over the past four decades (Altermatt 2010; Roy and Sparks 2000; Stefanescu et al. 2003) as it has or will for other insects (Hassall et al. 2007; Logan et al. 2003; Masters et al. 1998); these changes are apparently correlated with an increase in degree-day availability early in the season (Parmesan 2006).

Although changes in the phenology of individual species are well-described (Menzel et al. 2006; Robinet and Roques 2010), less attention has been paid to climatically driven mismatches to the trophic relationships of interacting species, despite predictions about the important negative consequences of asynchrony and its resultant decoupling (Donnelly et al. 2011; Singer and Parmesan 2010). Climatically driven decoupling is expected when synchrony between species is disrupted in time or space (Stenseth and Mysterud 2002). Decoupling can be viewed from either a temporal or spatial perspective. Spatially, rapid range expansion by a species could decouple relationships between predator and prey (Menendez et al. 2008; see Sect. 4.1), whereas temporally, differential response to shifting temperatures could lead to a phenological decoupling of a species relationship, be it plant-herbivore, predator-prey, or tritrophic interactions.

Whether or not a system will become phenologically decoupled depends on the response of the participant species to climatic drivers. For example, no net change could occur if the interacting species respond similarly to the same environmental cues or to different environmental cues in a way that is highly correlated. However, decoupling might be expected where species are responding to specific cues that become less correlated as temperatures and/or seasonality change. For example, a photoperiodic response by one species could lead to a divergent phenology if an interacting species responds primarily to degree-day accumulations. In tri-trophic relationships, elucidating the effects of climatic shift will be difficult and the relative changes in responses by an herbivorous insect, its host plant, and its natural enemies could be neutral, negative, or positive depending on the degree of decoupling and the nature of the decoupled mechanism(s).

The potential for climatically driven phenological decoupling of herbivorous insects and their host plants has long been recognized (Buse and Good 1996; Dewar and Watt 1992; Harrington et al. 1999), but has been investigated extensively in only a few systems. The importance of phenological synchrony of insect herbivores with host plants varies between and among species, functional feeding guild, and the seasonal activity period of a species with effects likely to be neutral or negative, as a positive effect seems implausible. The same mechanisms that drive asynchrony and decoupling, however, could allow insects to utilize hosts that were previously outside of their phenological range as climatic change differentially alters seasonal timing of tree and herbivore (e.g., Jepsen et al. 2011).

Sensitivity to phenological change is likely to be greatest for spring feeding species (Forkner et al. 2008), but could also affect other seasonal guilds depending on the nature and magnitude of change. Increased voltinism (see Sect. 4.3) may push some phenologically insensitive species into more vulnerable early or late season envelopes. Species whose activity (i.e., egg hatch, larval emergence) is timed to bud burst of host trees may be susceptible to even relatively small alterations to synchrony. These species often have a narrow window of opportunity to maximize growth because they are constrained by starvation if they emerge too early, and by declining nutritional value and increasing secondary phytochemical concentration of maturing leaves should development be delayed until later in spring (Ayres and MacLean 1987; Feeny 1970; Hunter 1993; Jones and Despland 2006; Martel and Kause 2002; Parry et al. 1998). While the effects of phenological asynchrony are best known for Lepidoptera, negative consequences have been shown in many other insect herbivores including Homoptera, Diptera, Coleoptera, and Hymenoptera (Dixon 1976; Fox et al. 1997; Martel et al. 2001; Yukawa and Akimoto 2006).

We know of no study that has specifically addressed phenological decoupling of an insect-plant interaction in the context of biological invasions, but it seems unlikely that introduced species would differ substantially from native species. Although direct research is lacking, extrapolation is possible from a few well-studied native insect-plant interactions. One exemplar insect, O. brumata, could be particularly instructive in elucidating consequences of climate change on phenological decoupling as it is well-studied in its native range, and is currently invading North America.

The winter moth is a univoltine spring feeder native to Great Britain and Europe, but was accidentally introduced to Nova Scotia (1940s), British Columbia (1960s), and more recently Massachusetts (2000s) and other New England states (Elkinton et al. 2010; Roland and Embree 1995). Flightless females ascend host trees such as oaks in the fall, and oviposit on branches and twigs in the canopy. Eggs hatch in the following spring in close proximity to bud break. The fitness consequences of synchrony with bud break are significant (van Asch and Visser 2007) as the newly hatched larvae have a limited ability to survive starvation if emergence is early but suffer from the declining value of maturing foliage if emergence is late.

Winter moth has been extensively studied, especially in Great Britain (e.g., Varley et al. 1973) and the existence of several long term data sets allows insight to the effects of climatic change on O. brumata with pendunculate oak. In the Netherlands, the onset of winter moth egg hatch and bud break of this oak species have advanced considerably over a quarter century (Visser and Holleman 2001). Egg hatch, however, has advanced more than bud break, decreasing synchrony by 2–14 days depending on the year. While late hatch decreases fitness through reductions in fecundity, a shift of 5 days too early can result in mortality of 90 % or more, suggesting that an increasingly premature hatch relative to bud break is non-trivial. Winter moth has sufficient genetic variability that selection should act to push hatch time closer to bud break (van Asch et al. 2010), although this does not appear to have happened naturally thus far. Although the effects of climatic change on O. brumata phenology and synchronicity with host plants have not been studied in North America where it is invasive, its extensive use of multiple tree genera in the northeastern United States could buffer it from any deleterious consequences of climatic change.

Considerably less is known about the effects of climate change on phenology in other invasive forest insects, even for those that have been extensively studied. For example, while various models (Régnière et al. 2009; Williams and Liebhold 1995b; see Sect. 4.1) have suggested that the geographic range of L. dispar in North America will expand northward under various warming projections, the potential for asynchrony with host tree species has not been explored. The responses of trees to warming at higher latitudes may differ from the temperature response of L. dispar egg hatch, thus increasing the risk of phenological mismatch. Although L. dispar is sensitive to tree phenology (Hunter 1993; Hunter and Elkinton 2000; Stoyenoff et al. 1994), it has life-history attributes that could mitigate many of the most deleterious effects of asynchrony. Similar to O. brumata, L. dispar larvae feed on a wide-variety of woody plants, which ensures that at least spatially, some hosts will be available to neonates. Furthermore, the temporal distribution of egg hatch, both within a single egg mass and among egg masses in a population, spans extended periods (Gray et al. 2007; Hunter 1993), which also increases the likelihood that the highly mobile larvae will encounter phenologically suitable hosts.

The phenological relationship between insects and plants do not occur in a vacuum; rather, it is a template upon which other environmental factors also enhance or attenuate the effects of asynchrony. Temperature and climatic shifts also occur in concert with rising levels of CO2, which could increase or decrease quality of plant tissue for herbivores depending on species and functional feeding guild (Cornellisen 2011; Stiling and Cornelissen 2007). Other environmental feedbacks and covariates associated with climate change could further confound any analysis. Based on limited studies to date, it seems unlikely that phenological asynchrony will be of significant long-term consequence for many native insect herbivores. Even less confidence can be attached to predictions about non-native species. As many forest insect invasions are initiated from genetically-limited founder populations from only portions of their native range, it is unclear if responses to climatic shifts will differ from that seen in populations of native species. However, many successful invasive species are habitat or host generalists and may express considerable phenotypic plasticity while other species, despite apparently limited genetic diversity, have nonetheless rapidly adapted to climatic variability in recipient regions (Gomi et al. 2009).

The potential for climatic disruption or alteration of coupled species relationships also applies to higher trophic levels, which often exert considerable top-down regulation on herbivore populations. The relative synchrony between natural enemies and their prey could be maintained under climatic change if the organisms respond similarly to the same variable or to variables that remain highly correlated. The decoupling of such relationships could occur as divergent responses to the same or to correlated variables, although few predator-prey interactions and even fewer multi-trophic systems have been examined in detail. A recent study documented considerable species turnover in samples of subarctic parasitoid communities when compared to historical data sets from the same localities, with patterns suggesting a link to climate warming (Fernández-Triana et al. 2011). In a different study, a meta-analytical approach suggested that an amplification of climatic variability was negatively correlated with parasitism of tree-feeding Lepidoptera, particularly for specialist hymenopteran parasitoids, which were disproportionately affected relative to tachinid parasitoids with broader host ranges (Stireman et al. 2005). Thus, at least in the short-term, the influence of parasitoids on the population dynamics of their prey could be reduced, which has important ramifications for outbreak species whether native or introduced. However, selection has favored herbivore life-history strategies that maximize temporal enemy-free or enemy-reduced space, and differential responses to climate variables could also force greater overlap between some herbivores and their natural enemies (Hance et al. 2007).

There is evidence of decoupled predator-prey relationships due to climatic shifts in a number of insectivorous birds in Europe. The fitness of these birds is greatest when the timing of reproduction corresponds with a peak in biomass of primarily lepidopteran caterpillars (Both et al. 2006; Visser et al. 1998, 2006). Warmer springs have shifted this peak earlier, and higher temperatures have compressed the period of abundance as larvae complete their development more rapidly. Although reproductive activity of birds has also advanced, it has not done so at the same rate as caterpillar biomass peak, and this relationship appears to becoming increasingly asynchronous. The implications for invasive forest insects are unclear, but since birds are important predators of many herbivorous insects, especially in low-density insect populations (Holmes et al. 1979; Marquis and Whelan 1994; Parry et al. 1997), a diminution of their capacities would likely benefit both native and non-native insects alike. However, there remains much uncertainty in the degree and importance of divergence of bird and insect prey phenology due to climate change; for example, a long time series data set in England suggests that the great tit, an important predator of O. brumata larvae, has been able to track the shift in spring phenology of its prey item over time (Charmantier et al. 2008).

The success of many invasive insects in forests owes at least in part to an incomplete or missing natural enemy component, and thus the potential effects of climate change on this trophic level could be largely moot for some species. However, for invasive species held in check by classical biological control introductions, climatic shifts potentially could alter these important interactions, leading to a resurgence of previously suppressed populations or hamper efforts to develop new biological control programs. Indeed, it has long been recognized that classic biological control could be vulnerable to climate change (Cannon 1998) because the interactions between introduced enemies and their non-native prey could be inherently more susceptible to decoupling than those interactions involving native species. Many introductions of control agents are initiated with relatively low genetic diversity, which potentially limits the adaptive response to changing climate. Second, specific biotypes of natural enemies, particularly parasitoids, are often selected to match current climatic conditions in a given region (Robertson et al. 2009); these may or may not be suitable for future climatic envelopes. Conversely, some biological control organisms that are currently climatically limited in parts of an invader’s range may become more effective under warming scenarios (Siegert et al. 2009; Zalucki and van Klinken 2006).

Many of the predictions concerning the decoupling of insect herbivore-host or predator-prey interactions are overly general or simplistic because we lack the necessary knowledge to make these predictions in all but a few systems. The effects of climate change on decoupling interactions involving non-native species are even more difficult to generalize as the relationships are often novel and are occurring in environments different from those where the species evolved. For example, in North America, the non-native tachinid Cyzenis albicans (Fallén) is regarded as the most effective regulator of invasive O. brumata populations, but this parasitoid is a trivial source of mortality in native populations (Roland and Embree 1995; Varley et al. 1973). Thus, it may be difficult to generalize the effects of climate from donor to recipient regions, even for well-studied systems.

5 Economic Ramifications of Invasions in the Face of Climate Change

Despite the challenges associated with predicting the ecological consequences of climate-mediated effects on biological invasions, it is arguably even more difficult to quantify the economic costs due to all of these interacting forces. After all, reliable estimates of the economic costs due to specifically non-native forest insects alone are largely lacking (Aukema et al. 2011). Even though these costs are challenging to estimate, they are not always difficult to envision. For example, increases in the availability of suitable habitat due to changing climates facilitating invasions into new areas could in turn increase the costs associated with its management (Cannon 1998; Kiritani 2006). Similarly, increases in abundance, and outbreak intensity and frequency due to climate warming is likely to lead to increased management costs (Hellmann et al. 2008; Rosenzweig et al. 2001; Waring et al. 2009). Costs could also include the increase in the energy footprint of food and fiber production systems due to this increased need for pest control measures (Gandhi and Herms 2010; Pimentel 2002).

Other potential consequences, however, can be complex and involve a cascading array of effects across one or more trophic levels. One such effect of climate change, and specifically the role of increased concentrations of carbon dioxide and ozone in the atmosphere, is the potential change in host plant nutritional quality. For example, plants grown under high levels of carbon dioxide can cause changes in the carbon-to-nitrogen ratio of plant tissues (Hamilton et al. 2005); consequently, herbivores feeding on such plants could eat more leaf matter to compensate for the reduced nutritional quality of their host plants (Coviella and Trumble 1999; Dermody et al. 2008; but see Kopper and Lindroth 2003). Increases in herbivory due to changes in concentrations of atmospheric gases, coupled with increases in herbivore abundance, insect developmental rates, and voltinism owing to increases in surface temperatures (Bale et al. 2002; Chen et al. 2011; Tobin et al. 2008; Yamamura and Kiritani 1998), could have dramatic implications to pest management practices and the costs required to achieve pest control. Furthermore, a need to increase pest control tactics, specifically the use of chemical insecticides, could also intensify the inimical effects to non-target species (Pimentel et al. 1980) as well as select for resistance in the target species (Roush and McKenzie 1987).

Because of the potential for climate change to decouple interactions between natural enemies and their prey (Simon et al. 2002; Stireman et al. 2005), the use of biological control as a management tactic against non-native forest pests could be rendered less effective. In particular, classical biological control has received renewed interest in combating non-native insect pests (Hajek et al. 2007; Hajek and Tobin 2010; Hoddle 2004), and increased scrutiny is given to the specificity of introduced agents to avoid the historical blunders from the import of generalist natural enemies (Elkinton et al. 2006; Simberloff and Stiling 1996; Strong and Pemberton 2000). Because of the need for specificity in selecting a natural enemy for introduction, changes in climate – even if subtle – could influence aspects of these interspecific interactions, and the suitable range of one species could be affected by climate differently than the other. For example, the parasitic nematode Deladenus siricidicola is an effective biological control agent of the wood wasp S. noctilio in Argentina (Corley et al. 2007) and Australia (Neumann and Minko 1981). As a nematode, it is likely more sensitive to changes in moisture conditions, which are predicted to be affected by changes in climate (Rosenzweig et al. 2001), than its insect host. Indeed, the observed geographic variation in the effectiveness of D. siricidicola as a biological control agent of S. noctilio could be due to variation in climate among regions (Slippers et al. 2012b). Although additional and specific forest insect examples are still rare, climate and projections in climate will likely need to be considered when evaluating the short- and long-term efficacy of an introduced natural enemy (Zalucki and van Klinken 2006).

6 Conclusions

In addition to many of the “known unknowns” described above, a final consideration in the context of climate change and its effect on forest insect invasions is the proverbial “unknown unknowns”. The dynamics of forest insects and their interactions with associated pathogens and natural enemies, together with interactions with host species, can be difficult to predict when species are introduced to a new area. Indeed, although many biological and ecological aspects are often highlighted as important when considering the invasion potential of a species and in formulating risk assessments (Liebhold and Tobin 2008; Lockwood et al. 2007; Worner and Gevrey 2006), developing a general paradigm of species invasiveness with broad application has proved challenging (Hulme 2003; Lonsdale 1999; Rejmánek and Richardson 1996). Coupling the uncertainty of biological invasions with the complexity of climate change and its variable effect on individual species and to those communities in which they interact complicates this challenge even further. Innocuous species today could quite possibly become quite invasive under future climatic conditions, whether in their native range, an introduced habitat, or both. Greater attention should be given to this complexity through examinations of landscape-level climatic changes and its combined effect on ecosystem inhabitants.