Extreme climatic events affect populations of Asian chestnut gall wasps, Dryocosmus kuriphilus, but do not stop the spread

Agricultural and Forest Entomology (2021), DOI: 10.1111/afe.12448 Extreme climatic events affect populations of Asian chestnut gall wasps, Dryocosmus kuriphilus, but do not stop the spread Maria J. Lombardero∗, Fernando Castedo-Dorado†and Matthew P. Ayres‡ ∗Unidade de Xestión Ambiental e Forestal Sostible. Departamento de Produción Vexetal e Proxectos de Enxeñaría, Universidade de Santiago de Compostela, Lugo, Rúa Benigno Ledo, S/N, 27002 Lugo, Spain, †Departamento de Ingeniería y Ciencias Agrarias. Escuela de Ingeniería Agraria y Forestal. Campus de Ponferrada, Universidad de León, Avda. de Astorga s/n. 24401 Ponferrada, León, Spain and ‡Department of Biological Sciences, Dartmouth College, Hanover, NH 03755, U.S.A. Abstract 1 Global climate change affects the frequency of extreme weather events that can influence plant–insect interactions. 2 We evaluated how the late-spring frost and severe drought that occurred in Spain in 2017 affected interactions between the invasive gall insect, Dryocosmus kuriphilus, and the native tree, Castanea sativa. We assessed effects on insect survival, fertility, population growth, and effects through changes in tree palatability and in other pests and pathogens. 3 Late-spring frost reduced D. kuriphilus to 25–40% of previous abundance. Wasp populations recovered rapidly (>7-fold in 3 years), consistent with density-dependence in population dynamics. 4 Larvae affected by freeze or drought were smaller. Female fecundity was affected by the freeze 1 year later. 5 Late-spring frosts and severe drought affected leaf size and physiology. Water content was higher within galls, but nitrogen was higher within galls in non-freeze plots after weather conditions improved. 6 Freezing also influenced the secondary chemistry of leaves. Phenol concentrations were lower, and terpenes higher, in frozen plots, while condensed tannins remained the same. Condensed tannins were reduced to half in the drought year. 7 Freezing had limited effects on damage from other pests and pathogens. 8 Our work expands understanding of how climate and weather affects forest pests. Keywords Community interactions, density-dependence, drought, invasive pest, late-spring frost, secondary metabolites. Introduction There is a strong agreement that global climate change has important effects on forests and forest communities. Furthermore, pest and disease outbreaks are themselves important as sources of natural disturbance (Kirilenko & Sedjo, 2007; Ayres & Lombardero, 2018). Global climate change can influence forest disturbance caused by insects and pathogens through changes in temperature and precipitation that can affect their survival, reproduction, dispersal, and geographic distribution (Ayres & Lombardero, 2000; Bale et al., 2002). However, climate change can involve not only shifts in average climate parameters, but also changes in the variability and predictability of weather patterns Correspondence: María J. Lombardero, Tel.: 34 982 823 150; fax: 34 982 823 001; e-mail: [email protected] (e.g., more extreme weather events and reduced winter snow cover; Harris et al., 2019; Wagner, 2020). The impact on insect populations may be even more dramatic when a growing season includes multiple climatic events. For example, the coincidence of spring frost and summer drought has been associated with general declines in insect abundance within a community of oak folivores (Marquis et al., 2019). One consequence of global climate change, at least in Mediterranean countries, is increased aridity and a higher frequency of extreme climatic events such as late-spring frosts and heat waves (Diffenbaugh et al., 2005; Christensen et al., 2007; Vitasse & Rebetez, 2018). Warming of temperatures in late winter and early spring tends to cause a general advancement of phenology (Menzel & Fabian, 1999; Parmesan & Yohe, 2003) including insects; shifts towards earlier seasonal © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 2M. J. Lombardero et al. activity are widespread (Forrest, 2016). This can trigger the so called ‘false spring’ and induce earlier plant growth, which increases the risk of vulnerable plant tissues to subsequent spring frosts (Inouye, 2000; Bigler & Bugmann, 2018). Consequently, late-spring frosts, i.e., below-freezing temperatures in late spring, are among the most critical extreme events in temperate and boreal regions (Zohner et al., 2020). Late-spring frosts affect growth and reproduction of plants and may cause considerable economic losses. For example, the economic loss for agricultural crops after the 2007 late-spring frost in the U.S. was up to $112 million, with fruit crop loss of $86 million (Warmund et al., 2008). Further examples of late-spring frost damage include 1995 in France (Ningre & Colin, 2007); 2010 in theNEofU.S.A.(Hufkenset al., 2012); and 2011 and 2016 in Switzerland, south Germany, and northeastern France (Kreyling et al., 2012; Vitasse et al., 2018). Late-spring frosts can kill foliage and therefore reduce photosynthesis and tree growth (Dittmar et al., 2006; Bräuning et al., 2016). Trees can partially compensate by producing new leaves (Kramer & Kozlowski, 1979; Neilson & Wullstein, 1983). However, refoliation is costly in resources and the length of the photosynthetic season is shorter nonetheless (Zohner et al., 2019). Late-spring frost may affect insect populations directly by killing eggs or early larvae exposed to the low temperatures, as well as by starvation when the tree foliage is killed (Meurisse et al., 2012). Late-spring frost can also exert substantial impacts on trees, including increased mortality, altered morphology, decreased growth, changes in primary nutrients for insects (e.g., N), and changes in leaf chemical defences (St. Clair et al., 2009; Hufkens et al., 2012; Man et al., 2013). Consequently, late-spring frost may also influence the susceptibility of plants to insects or pathogens, although the number of studies that have examined the connection between freeze damage and insect susceptibility in trees is limited. Some authors suggest freeze as a predisposing factor to secondary insect pests and fungi (Wargo, 1996; Wolken et al., 2009; La Spina et al., 2013). Increased growth has been reported for insects feeding in tissue reflushed after freeze damage (Thomson et al., 2001; Rubert-Nason et al., 2017). The few studies of freeze-induced changes in phytochemistry and plant–insect interactions have reported an increase in phenolic glycosides and condensed tannins (St. Clair et al., 2009) or mixed effects depending on the genotypes (Rubert-Nason et al., 2017). Rapid climate change tends to increase co-occurrence of extreme events as heat stress and drought (Mittler, 2006). The impact on insect communities of rising temperature has been broadly studied during the last 25 years (e.g., Bale et al., 2002; Estay et al., 2014). Direct effects of warmer temperatures on insect physiology include rapid insect development and increased overwinter survival (Bale et al., 2002). Nevertheless, warmer summer temperatures do not always reduce generation time (Forrest, 2016), and can result in increased mortality, smaller size, and lower dispersal capacity (Pineau et al., 2017; Mech et al., 2018). Insects experience the effects of drought and heat directly but also indirectly through changes in host suitability (e.g., phytochemistry) and changes in the broader biotic community (e.g., competitors, enemies, and species that share common enemies) (Rouault et al., 2006). Insect herbivores may be affected by drought because it can affect nutritional quality of plant tissue (Holopainen et al., 2018) and reduced water content may interfere with nitrogen acquisition by insects (Huberty & Denno, 2004). However, the impact of drought on susceptibility of trees to insect infestations and mortality is unclear (McDowell et al., 2011; Gely et al., 2020). It has been suggested that plant nitrogen content can increase with moderate drought (Mattson & Haack, 1987; Kolb et al., 2016) and decrease with severe drought (He & Dijkstra, 2014; but see Gely et al., 2020). The effects of drought on tree resistance to pests may vary depending on the feeding guild of insect herbivores (Larsson, 1989; Jactel et al., 2012). Moderate water-stress can limit woodborer performance (Lieutier et al., 2004), but improve performance of defoliators due to increased concentration of soluble nitrogen in foliage (Mattson & Haack, 1987; Larsson, 1989). Nevertheless, experimental studies have found mixed responses of leaf-feeders to plant water stress (Rouault et al., 2006; Castagneyrol et al., 2018). This could be because the different sub-guilds within defoliators (e.g., leaf chewers, gall formers, and leaf miners) may have different responses to plant water stress (Huberty & Denno, 2004; Jactel et al., 2012). For leaf chewers and gall makers, survival, density, and overall performance have been found to decrease with drought (Huberty & Denno, 2004; Jactel et al., 2012). Generalizations are further complicated because plant stress can have opposite effects on different life-stages of some galling insects (Bjorkman, 1998). Galling insects have an intimate relationship with their host plants (Price et al., 1987; Stone et al., 2002). The nutrition and defences of galling insects can depend upon their success in manipulating host plant morphology and physiology (Tooker et al., 2008; Giron et al., 2016; Oliveira et al., 2016). There are three major hypotheses for the adaptive significance of gall induction and the evolution of gall morphology. The Nutritional Hypothesis states that galls provide a high-quality nutrient source with less plant defensive compounds than other feeding modes (Price et al., 1987; Crespi et al., 1997). The Microenvironment Hypothesis states that galls protect the insects from unfavourable abiotic conditions as temperature changes, desiccation, or ultraviolet radiation (Price et al., 1987; Crespi et al., 1997; Miller et al., 2009). The Enemy Hypothesis argues that morphology and chemistry of the gall tissue protect the insect from predators, parasitoids and pathogens (Stone & Schonrogge, 2003). The aim of this study was to assess how extreme climatic events, such as the late-spring frost of 2017, affected interactions between the native chestnut tree, Castanea sativa Mill., with its most important pest, the invasive gall insect Dryocosmus kuriphilus Yasumatsu, in Northern Spain. The year 2017 was also exceptionally warm (1.1 ∘C above the annual mean of the period 1981–2010; the warmest year since 1965) and was the year with the second most severe drought since 1965 (AEMET, 2017, 2018). Therefore, we also analysed how the warm temperatures and the severe drought may affect the plant–insect in that interaction. We evaluated direct effects of weather on insect performance by quantifying insect survival, fitness, and reproduction capacity. We also evaluated indirect effects of weather on D. kuriphilus population growth due to (i) © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 Extreme climatic events and D. kuriphilus 3 Figure 1 Locations visited after a late-spring frost in May 2017 in Galicia (NW of Iberian Peninsula). Fourteen of the 16 study plots are visible; the other two are located in the Northeast (Castroverde and Masoucos) and were not included in the figure to improve visualization of the survey transect. [Colour figure can be viewed at wileyonlinelibrary.com] changes in trees palatability, and (ii) changes in the associated communities of other insects and pathogens living in the same trees. Material and methods Study area The study was carried out in Galicia (NW of Spain). From January to March of 2017, we established 16 study plots to follow the damage caused by D. kuriphilus on Castanea sativa in the study area. These plots were in the central area of Galicia (Fig. 1) with similar climatic conditions. All 16 study plots were private chestnut plantations with a double purpose: nut and wood production. The plots were discrete within a landscape matrix of agricultural and forest land. The plots had different owners and different agronomic histories which can influence the site index, even for plots located very close. In addition, spatial variability in local site conditions, involving microtopography and soil, can lead to large spatial variability in site quality (Skovsgaard & Vanclay, 2013). The quality of each plot for chestnut growth was assessed through the site index (the expected dominant height of the plot at a reference age of 45 years), according to the site index curves developed by Patricio and Nunes (2017). For our analyses, we recognized eight high-quality sites and eight low-quality sites based on whether the height of dominant trees was greater than or less than 25 m at 45 years. Study species Sweet chestnut (Castanea sativa, Fagaceae) is a tree native to Southern Europe and Asia Minor. Sweet chestnut is a forest species of great interest in northwestern Spain for their economic value and for the area occupied, more than 100 000 ha, of which about 45 000 ha are in Galicia (MARM, 2011). The species is widely cultivated because it produces high-quality wood and edible seeds, the chestnuts, which have been used as human and animal food since ancient times (Conedera et al., 2004). Dryocosmus kuriphilus (Hymenoptera: Cynipidae) is originally from China and it is considered the most important insect pest of chestnuts worldwide (EPPO, 2005). It was first reported in Japan in 1941 damaging local chestnut orchards (Aebi et al., 2006). The species was detected in Europe, in Italy in 2002 (Brussino et al., 2002). Since then, it has spread throughout Europe reaching Spain in 2012 and Galicia in 2014 (Nieves-Aldrey et al., 2019). Currently, it has spread throughout the territory (Gil-Tapetado et al., 2021a). © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 4M. J. Lombardero et al. Climatic events April of 2017 in Galicia was characterized by two climatic events. There was a very warm period during early April due to the entrance of warm air from the North of Africa, raising temperatures above 30 ∘C in many areas and reaching a maximum of 34.3 ∘C on 9 April at Arnoia (Ourense). The exceptionally warm period promoted early growth in chestnuts as well as other plants. Then, a cold front penetrated Europe on 17 April, followed by a large incursion of dry polar air. Consequently, most of Europe experienced a series of nights that reached freezing temperatures. In Galicia, the coldest nights were from 27–29 April when temperature dropped to as low as −8.2 ∘C in Baltar (Ourense) (Meteogalicia, 2018). The rest of 2017 was hot and dry in Galicia; the average temperature for 2017 was 0.8 ∘C above the mean for 1981–2010 while precipitation was 24% lower (966 vs. 1299 L/m2). Hydric balance was negative since April (−104.1 L/m2), and reached the lowest value in July (−148.9 L/m2) in the weather station of Campus de Lugo, in the vicinity of one of the study plots (data available at https://www.meteogalicia.gal/observacion/ estacionshistorico/historico.action?idEst=10053). In contrast, 2018 had relatively normal temperatures and precipitation that was 11% higher than the mean for 1981–2010 (Meteogalicia, 2018, 2019). We did some additional measurements in 2020 (see Attack level and population growth section), which was also a warm year in Galicia but with near-average precipitation (5% lower than normal; Meteogalicia, 2021). Effect of frost on Dryocosmus kuriphilus Survival. To assess freeze damage directly to the insect pest, 3 days after the late-spring frost (2 May), we sampled along a transect of about 250 km through the study area. Along the way, we examined 39 chestnut plantations (including the 16 main study plots), 20 affected by the late-spring frost and 19 non affected (Fig. 1). In each plantation, we recorded the geographic coordinates (WGS 84, World Geodetic System 1984) with a GPS device and whether or not the plot suffered freeze damage. Given the variable topography, the study plots, frozen or not, were interspersed across landscape (Fig. 1). In 23 of these chestnut plantations (13 affected by late-spring frost and 10 non-affected), we collected 20 galls chosen haphazardly from within each of five different trees (total of 100 galls per plot). Galls were moved to the lab, stored in individual boxes within the refrigerator and dissected within 3 days to assess larval survival. Individual cell chambers containing ectoparasitoids were excluded from the estimations. We repeated this sampling in each of the next 3weeks. Attack level and population growth. In summer 2017, after freeze-damaged trees had flushed new shoots and leaves, we assessed gall abundance in the 16 main study plots. We selected two branches at random in each of 12–27 trees per plot. The number of trees varied among plots due to the availability of trees. On each branch, we located and examined the portion of the shoot that grew in the previous summer (2016). Within that length of shoot, we counted all the galls produced in the current year (2017). As a covariate representing variation in the size of shoots, we also counted the number of buds that were present at the end of previous year’s growing season (2016). The branches were labelled, and thus we were able to repeat measurements on the same trees and branches in 2018 and 2020 (again using the previous year’s buds as covariates). The resulting data allowed us to estimate galls/shoot for each plot in each year and per capita population growth rate between years. Larval size and adult fecundity. To assess effects of the freeze event on D. kuriphilus, we weighed 145 late-instar larvae in 2017 from seven of the study plots (10–41 larvae/plot) and in 2018 we weighed 166 larvae from the same plots (21–27 larvae/plot). We also assessed female fecundity by counting the number of eggs produced by emerging female adults from galls collected in the same study plots (n=57 and 107 females in 2017 and 2018, respectively). Galls were collected in the study plots and stored in individual boxes in the lab until females emerge. Eggs were counted by observing the dissected abdomen of the emerging females under the microscope. We compared these data with the only previous measurements of female eggs collected in the area. These data are from Lugo, the plot where the insect was detected in Galicia for the first time. The Lugo plot was among the 39 sites in the present study and was not affected by the frost. Thus, we had measurements of eggs per female in that location since 2015. Gall characteristics. In 2018, at each of 16 main study sites, we collected and measured 51–160 galls of each of three types of galls based on the tissue where the gall is formed (shoot, leaf, or stipule). Variation in sample size among plots and gall types was due to the availability of trees and galls of the scarcest type (stipule galls). In the laboratory, we measured three perpendicular axes of each gall with digital callipers and averaged the three axes to estimate gall diameter. All of galls were subsequently dissected to measure gall wall thickness (with a calliper) and count the number of feeding chambers (each representing one gall wasp). Effect of frost on tree nutritional quality and palatability In June of 2017, after the leaves had re-flushed, we were able to collect and analyse 10 samples of ungalled leaves (each from a different tree) from each of six study plots (three plots that experienced freezing and three that did not). We avoided occasional leaves in the unfrozen plots that exhibited any signs of freezing damage. In the lab, we measured leaf size, nitrogen, and water content as well as total phenols, condensed tannins, and terpenes. In the same sampling, we also collected another sample of leaves with D. kuriphilus galls (one sample from each of five trees per plot). For each galled leaf, we analysed water and N content separately in the gall itself and in the leaf tissue surrounding the gall. We repeated this sampling in 2018 in the same six study plots. To analyse water and N content, leaves were weighed fresh then oven-dried at 60 ∘C for 48 h. The dried samples were weighed again to estimate water content, milled to a fine © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 Extreme climatic events and D. kuriphilus 5 powder, and submitted to instant oxidation (as 0.1 g tissue samples); the gases released were quantified with an Elemental Analyser (LECO-TruSpec). Analyses were performed by the analytical unit of the University of Santiago de Compostela (RIAIDT). To analyse concentrations of total terpenes, we followed Wainhouse et al. (1998). One gram of leaf or gall from each sample was cut in very small sections and terpene compounds were quantitatively extracted twice with n-hexane (with each extraction including 25min in an ultrasonic bath). Later, the plant material was recovered by filtration using quantitative filter paper. Then, the solvent was evaporated at room temperature inside an extraction chamber, and the mass of the non-volatile terpene residue was measured with a precision scale. Total leaf phenolics were determined following Sampedro et al. (2011). Phenolics were extracted from 0.5 g of leaf tissue with aqueous methanol (1:1 vol:vol) in an ultrasonic bath for 15 min, followed by centrifugation and subsequent dilution of the methanolic extract. Total phenolic content was determined colorimetrically using Folin–Ciocalteu in a BioTek Elx 850 microplate reader at 740 nm, quantified with a standard curve of tannic acid and expressed as mg tannic acid equivalents g−1dry mass of plant tissue. To analyse condensed tannins, the same extract was assayed with butanol – hydrochloric acid reagent (0.7 g ferrous sulphate heptahydrate in 50mL concentrated HCl and n-butanol added to make 1 L), and absorbance was measured at 550 nm (Waterman & Mole, 1994) with the same microplate reader, using as standard purified condensed tannins of quebracho (Schinopsis balansae Engl., Unitan Saica, Buenos Aires). Phytosanitary inventory In summer 2018, we carried out an inventory of insect and fungi present in the same 266 study trees of the 16 main study plots (12–27 trees/plot). Identification was done to species level for the common organisms and to feeding guild for the unknown ones. For each pest or pathogen on each study tree, one of us (MJL) assessed severity of the damage on a scale from 0 to 5. Statistical analysis Population abundance, measured as galls per shoot, was analysed with a general linear model that included site quality, freezing in 2017 or not, and their interaction as fixed effects, site nested within site quality and freezing as a fixed effect, tree nested within site as a random effect, the number of buds per shoot at the end of previous year’s growing season (potential sites for galls) as a continuous variable, and no intercept. Larval size and adult fecundity were analysed with an analysis of variance (anova) that included year, freezing, year x freezing, and plot nested within freezing. Leaf size, terpenes, phenols, and condensed tannins were analysed with an anova that included year, freezing, and their interaction as fixed effects. Each replicate sample represented a different tree. Water and nitrogen content were analysed with an anova that included year, tissue type (ungalled leaf, ungalled portion of galled leaf, or gall), freezing or not in 2017, their 2and 3-way interactions, and plot nested within freeze and year; tree was a random effect nested within plot and year. The size and wall thickness of galls were analysed with an anova that included freezing, site quality, gall type, and their interactions as fixed effects, and plot nested within freezing and site quality as a random effect. Cells per gall were log-transformed prior to analysis to improve normality and homoscedasticity. Damage levels from common pests and pathogens were analysed with an anova that included freezing, site quality, and their interaction as fixed effects, and plot nested within freezing x site quality as a random effect. Regression models were used to compare attack level of D. kuriphilus with the severity of other insects and pathogens. Statistical analyses were performed with the package JMP (SAS Institute Inc.). Results Impact of extreme climatic events on Dryocosmus kuriphilus At the time of the freeze, chestnut trees in all plots were growing shoots and expanding leaves. Trees from plots affected by the freeze lost their crowns completely (Fig. 2), while the non-affected plots either suffered no visible damage or only had some leaves that were partially frozen. Due to local topography, the freeze had variable effects across our study plots: seven were affected and nine were not (Table 1). Freezing effects on D. kuriphilus varied with gall type. Frozen galls in leaves dropped with the leaves within a few days and larvae of D. kuriphilus presumably died inside (Fig. 2b). Frozen galls in shoots showed less initial damage (Fig. 2c) and the larvae remained alive longer. Shoot galls sampled 3 days after the frost contained 96% of the larvae alive. A week after the frost, survival was still 87% (Fig. 2d), but gall tissue deteriorated progressively, and larval survival declined accordingly. Larval survival dropped to 68% and then only 25% over the next 2 weeks as the shoot galls rotted (Fig. 2e,f). We did not observe larval mortality in plots not affected by the late-spring frost. When new leaves grew back in the plots affected by the late-spring frost some contained galls. By the end of larval development in 2017, surviving galls per shoot in frozen plots was only 25–50% of that in unfrozen sites, with stronger reductions in plots that were best for chestnut growth (Fig. 3). In 2018, galls per shoot were still lower in plots that had been frozen, but less so than in 2017. By 2020, gall abundance had fully recovered in frozen plots on high-quality sites and was even higher than unfrozen plots on low-quality sites where wasp densities averaged almost 7 galls per chestnut shoot (Fig. 3). In 2017 and 2020, average gall densities were significantly higher in low-quality sites for chestnut (F1295 =14.03, P<0.001, and F1,67 =4.28, P=0.042, respectively). The rapid recovery of D. kuriphilus populations in frozen plots was attributable to strong density-dependence, with per capita growth rate of D. kuriphilus being high when they were relatively rare and decreasing as abundance increased (Fig. 4). In plots with low abundance, growth rates corresponded to population doubling times of just over 3years (R≈2.25 galls/gall/year). The © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 6M. J. Lombardero et al. (a) (b) (c) (d) (e) (f) Figure 2 (a) Freeze damage on chestnut leaves and shoots. (b) Dryocosmus kuriphilus larvae inside the frozen gall. (c) Bud gall apparently not affected by frost. (d) Gall inner tissue showing necrosis while the larvae was still alive. (e, f) Larvae alive inside the gall 3 weeks after the frost. [Colour figure can be viewed at wileyonlinelibrary.com]. apparent equilibrium abundance (K) was about 5 galls/shoot; plots with <5 galls/shoot tended to increase and those with more tended to decrease (Fig. 4). The deterministic density-dependent function in Fig. 4 predicts recovery from low abundance (0.1 galls/shoot) to 90% of equilibrium in ∼5years. The mass of late instar larvae differed conspicuously between years; larvae of the same developmental stage were about 30% larger in 2018 than in 2017 (Fig. 5, upper; F1,297 =183.03, P<0.0001). There was also an effect of frost but only in 2017 when larvae from frozen plots were about 13% smaller than © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 Extreme climatic events and D. kuriphilus 7 Table 1 Characteristics of the 16 main study plots. Site index (mean ±SE) for lowand high-quality sites were 23.3 ±0.3 and 26.2 ±0.4 m, respectively Locality Latitude (∘N) Longitude (∘E) Elevation (m.a.s.l.) Frost damage Minimum temp. (∘C) Site quality Site index (m) Carballal 42.871 7.883 524 No 1.4 High 25.6 Castroverde 43.101 7.338 534 No −1.0 High 27.2 O Corgo 42.912 7.504 518 No −1.9 High 28.2 Rumín 42.854 7.870 549 No 1.4 High 25.8 Verín II 42.760 7.868 574 No 1.4 High 25.4 Gaioso 43.126 7.669 403 Yes −3.3 High 25.4 Uriz I 43.117 7.649 414 Yes −3.3 High 26.2 Uriz II 43.114 7.651 407 Yes −3.3 High 25.5 Buratai 42.994 7.508 448 No −1.9 Low 23.7 Calde 42.953 7.611 571 No −1.9 Low 23.0 Masoucos 42.996 7.276 683 No −0.6 Low 24.6 Torible 43.023 7.681 544 No −1.9 Low 24.2 Verín I 42.760 7.869 580 No 1.4 Low 23.0 Casa Laranxa 43.118 7.660 411 Yes −2.8 Low 22.6 Esperante 42.964 7.598 520 Yes −2.8 Low 23.6 Robra 43.089 7.613 407 Yes −3.3 Low 22.2 those from unfrozen plots (Fig. 5, upper; main effect of freeze: F1,297 =8.69, P=0.003; year x freeze interaction: F1,297 =7.48, P=0.007). In the year of the frost, fecundity did not differ among females, but in the next year, adult females in frozen plots had barely half the fecundity as those in unfrozen plots or the previous year (Fig. 5, lower; year x freeze interaction: F1,155 =9.36, P=0.0026). The low fecundity of females in frozen plots 1 year after the freeze was a notable pattern in our study population. It contrasts with the stability of fecundity over 3 years in our oldest study site (Lugo, not affected by the freeze), which is where the gall wasp has been established the longest (first detected in 2014): mean ±SE =125 ±6, 116 ±6, and 138 ±8 in 2015, 2017 and 2018, respectively; F2,102 =2.40; P=0.09. Impact of extreme climatic events on chestnut leaves and galls There was severe damage to chestnut trees in the plots that froze in 2017. All leaves and shoots that were expanding at that time were destroyed and dropped within few days (Fig. 2a). Trees responded with refoliation but the length of leaves in 2017 was reduced in frozen plots: mean ±SE =9.7 ±0.9 vs. 14.3 ±0.9 cm, respectively (F1,38 =10.34; P=0.003). Leaf size in frozen plots was still 26% smaller in 2018: 14.3±1.0 vs. 19.5 ±0.8 cm, respectively (F1,38 =20.14; P<0.0001). Leaf size may also have been affected by the drought because leaf length was less in 2017 than in 2018: 12.0 ±0.6 vs. 17.3 ±0.6 cm, respectively (F1,77 =27.76; P<0.0001 for a year effect). Analyses of leaves without galls showed variable effects from freezing on phytochemistry (Fig. 6). In frozen plots compared to unfrozen plots, total phenols were significantly lower (≈23%), and total terpenes were significantly higher (≈26%) (F1,63 =4.32 and 4.70, respectively; P<0.04). Condensed tannins were similar between frozen and unfrozen plots but were more than twice as high in 2018 as in 2017 (F1,63 =4.70, P=0.03). anovas of the data in Fig. 6 revealed no other significant main effects or interactions. Galls always had higher water content than the surrounding leaf tissue in galled leaves or than ungalled leaves (Fig. 7, upper; least square means ±SE =82.1 ±0.4, 63.3 ±0.4, and 64.4 ±0.4%; F2,167 =647.07, P<0.0001, Table 2). Water content of leaves and galls was higher in 2018 compared with the drought year of 2017 (least square means ±SE =72.9 ±0.4 vs. 66.9 ±0.4%; F1,75 =113.38; P<0.0001); this was especially true in the ungalled leaves and galled leaves. In 2017, water content of ungalled leaves and galled leaves was markedly higher in frozen plots (which had reflushed leaves) than in unfrozen plots (63–65% vs. 55–59%, Fig. 7) but this was not true in 2018 (66–68%; freeze x year interaction: F1,75 =24.86, P<0.0001). There was also a frost x tissue type interaction (F2,167 =8.01; P=0.0005) because ungalled leaves and the ungalled part of galled leaves had particularly low water content in unfrozen plots in 2017 (Fig. 7, upper). Years were similar with respect to average nitrogen content of leaves and galls, but there were patterns with respect to tissue types and exposure to freezing (Fig. 7, lower; Table 2). The highest N concentrations in 2017 were in the relatively small reflushed leaves (with and without galls) in plots that froze (2.96–3.16%). However, N content of the galls themselves was relatively low in these samples. In 2018, N content tended to be lower in the frozen plots than the unfrozen plots, especially in galls and the ungalled portion of leaves with galls. In 2018, galls were of similar size in plots that froze or not in 2017: mean diameter ±SE =9.16 ±0.15 vs. 9.16 ±0.13 mm, respectively (F1,12 =2.82, P=0.12). Shoot galls were larger than leaf galls, which were larger than stipule galls: 11.19 vs. 9.44 vs. 7.44 mm (SE =0.13 for all; F2,1887 =526, P<0.0001); this range was greater in good quality sites than poor quality sites: 7.24 to 11.33 mm vs. 7.64 to 11.04mm (site quality x gall type: F2,1887 =4.42, P=0.012). The number of chambers per gall was positively related to gall size and showed similar patterns as gall size: no effect from the freeze (F1,12 =0.00, P=0.95), and stronger differences between gall types on good quality sites (1.28 to 2.62 vs. 1.28 to 2.06 cells/gall for stipule galls to bud galls; site quality x gall type: F2,1887 =8.66, P=0.0002). The average thickness of galls walls was about 2.45±0.06 mm © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 8M. J. Lombardero et al. Figure 3 The abundance of Dryocosmus kuriphilus galls in plots that were and were not frozen in the year of the late-spring frost (2017), the following year, and 2 years later. Data are grouped by sites that were low vs. high quality for growth of chestnut trees. irrespective of freezing, site quality, or any of their interactions (P>0.12). Impact of weather and gall wasps on other pests and pathogens Our survey of 266 trees carried out in 2018 and distributed across 16 plots revealed frequent occurrences of three types of folivorous insects (defoliators, leaf miners, and leaf skeletonizers) and four types of fungi (Cryphonectria parasitica,Mycosphaerella maculiformis,Gnomoniopsis sp., and Phomopsis sp.) Damage from one or more of these agents were evident in 17–96% of the trees. Most trees showed damage from at least one of the insect groups and one of the fungal groups (97% and 87%, respectively). There were no significant correlations among the seven Figure 4 Per capita change in the abundance of Dryocosmus kuriphilus from 2017 to 2018 as a function of density in 2017. Each point represents a plot. Some plots were frozen in spring of 2017 and some were not. There was additional variation among plots in initial density due to time since invasion by D. kuriphilus. The relationship was log-linear and suggests an equilibrium (K) of about 5 galls/shoot (with an average annual shoot being 25 cm long and containing 5 buds). (a) (b) Figure 5 Larval mass (a) and fecundity (b) of Dryocosmus kuriphilus in 2 years in plots that were and were not frozen in spring of 2017. agents in their damage levels (|r|<0.26 for 22 pairwise comparisons, n=266). There was only limited evidence for effects of freezing in 2017 on damage from pests and pathogens in 2018. Leaf miner damage was higher in previously frozen plots: mean damage score ±SE =0.30 ±0.03 vs. 0.08 ±0.03, respectively (F1,7.8 =26.47, P<0.0001), and skeletonizer damage was lower © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 Extreme climatic events and D. kuriphilus 9 (a) (b) (c) Figure 6 Foliar concentrations of (a) Total phenols, (b) Condensed tannins, and (c) Terpenes in 2 years in plots of chestnut that were and were not frozen in spring of 2017. in previously frozen plots: damage score =0.15 ±0.06 vs. 0.48 ±0.05, respectively (F1,12.6 =15.89, P=0.002). There were no significant effects of the 2017 freeze on any of the other five agents (P>0.18). There were no significant effects for any of the seven agents from site quality or freeze x site quality. There was a tendency for variation in attack severity by fungi among replicate plots (Wald’s Pvalue =0.03, 0.03, 0.06, and 0.07 for Cryphonectria,Gnomoniopsis,Mycosphaerella,andPhomopsis, respectively). None of the insect groups showed variation in attack levels among replicate plots. Damage from skeletonizers tended to decrease with increasing damage from gall wasps (P=0.002, n=260; regression of residuals from above anova vs. % crown damage from gall wasps). However, damage from Gnomoniopsis and Mycosphaerella tended to decrease with increasing damage from gall wasps (P=0.0002 and P=0.02, respectively, n =260). The other four agents of damage showed no relationship to abundance of gall wasps. Figure 7 Foliar concentrations of water (upper) and nitrogen (lower) in 2 years in plots of chestnut that were and were not frozen in spring of 2017. Measurements included control leaves, leaf tissue around galls in galled leaves, and the galls themselves. Discussion The late-spring frost of April 2017 caused severe freezing damage to both natural and cultivated vegetation though Europe (Vitasse & Rebetez, 2018). There were catastrophic consequences for fruit growing and viticulture. Economic losses were estimated at €3.3 bn mainly for fruit and wine growers – especially in Italy, France, Germany, Poland, Spain and Switzerland (Munich Re, 2017). The major damage in Galicia occurred in wine and chestnut plantations (DOGA, 2017). Altitude, even just a few meters, made big differences among sites; plots located at lower altitude suffered more damage. In clear nights with temperature inversions, colder near-ground temperatures bring greater frost damage at lower topographic areas (Clarke, 1946). The late-spring frost strongly affected chestnut trees, but also populations of Dryocosmus kuriphilus. Abundance after the frost was reduced to 25–50% of the previous abundance in plots that froze (Fig. 3). Presumably, most larvae died by starvation because they were unable to leave their deteriorating gall tissue. Surprisingly, in plots where trees lost completely the crown, trees flushed back with galls in leaves and shoots, most likely because some buds were not yet developed by the end of April. We cannot reject the possibility that D. kuriphilus lays some eggs in dormant or adventitious buds, as is known in other gall makers when abundance is high and oviposition sites are limited (Hails & Crawley, 1991). However, plots with low attack levels, and presumably limited competition, also showed galls after refoliation. It could be adaptive for females so spread © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448 16 M. J. Lombardero et al. Tooker, J.F., Rohr, J.R., Abrahamson, W.G. & De Moraes, C.M. (2008) Gall insects can avoid and alter indirect plant defenses. New Phytologist,178, 657–671. https://doi.org/10.1111/j.1469-8137 .2008.02392.x. Vannini, A., Vettraino, A., Martignoni, D. et al. (2016) Does Gnomoniopsis castanea contribute to the natural biological control of chestnut gall wasp? Fungal Biology,121, 44–52. https://doi.org/10.1016/j.funbio .2016.08.013. Villalpando, S.N., Williams, R.S. & Norby, R.J. (2009) Elevated air temperature alters an old-field insect community in a multifactor climate change experiment. Global Change Biology,15, 930–942. https://doi.org/10.1111/j.1365-2486.2008.01721.x. Vitasse, Y. & Rebetez, M. (2018) Unprecedented risk of spring frost damage in Switzerland and Germany in 2017. Climatic Change,149, 233–246. https://doi.org/10.1007/s10584-018-2234-y. Vitasse, Y., Schneider, L., Rixen, C., Christen, D. & Rebetez, M. (2018) Increase in the risk of exposure of forest and fruit trees to spring frosts at higher elevations in Switzerland over the last four decades. Agricultural and Forest Meteorology,248, 60–69. https://doi.org/10 .1016/j.agrformet.2017.09.005. Wagner, D.L. (2020) Insect declines in the anthropocene. Annual Review of Entomology,65, 457–480. https://doi.org/10.1146/annurev-ento011019-025151. Wainhouse, D., Ashburner, R., Ward, E. & Rose, J. (1998) The effect of variation in light and nitrogen on growth and defence in young Sitka spruce. Functional Ecology,12, 561–572. https://doi.org/10.1046/j .1365-2435.1998.00232.x. Walther, G. (2010) Community and ecosystem responses to recent climate change. Philosophical Transactions of the Royal Society B – Biological Sciences,365, 2019–2024. https://doi.org/10.1098/ rstb.2010.0021. Wang, C., He, J., Zhao, T. et al. (2019) The smaller the leaf is, the faster the leaf water loses in a temperate forest. Frontiers in Plant Science, 10, 58. https://doi.org/10.3389/fpls.2019.00058. Wargo, P.M. (1996) Consequences of environmental stress on oak: predisposition to pathogens. Annales Des Sciences Forestieres,53, 359–368. https://doi.org/10.1051/forest:19960218. Warmund, M.R., Guinan, P. & Fernandez, G. (2008) Temperatures and cold damage to small fruit crops across the eastern United States associated with the April 2007 freeze. HortScience,43, 1643–1647. https://doi.org/10.21273/HORTSCI.43.6.1643. Waterman, P.G. & Mole, S. (1994) Analysis of Phenolic Plant Metabolites. Blackwell Scientific, U.K. Weis, A.E., Price, P.W. & Lynch, M. (1983) Selective pressures on clutch size in the gall maker Asteromyia carbonifera.Ecology,64, 688–695. https://doi.org/10.2307/1937190. Wills, A.J. & Farr, J.D. (2017) Gumleaf skeletoniser Uraba lugens (Lepidoptera: Nolidae) larval outbreaks occur in high rainfall Western Australian jarrah (Eucalyptus marginata) forest after drought. Austral Entomology,56, 424–432. https://doi.org/10.1111/aen.12255. Wolken, J.M., Lieffers, V.J., Landhaeusser, S.M. & Mulak, T. (2009) Spring frost and decay fungi are implicated in suppressing aspen re-growth following partial cleaning in juvenile stands. Annals of Forest Science,66, 805. https://doi.org/10.1051/forest/2009072. Wool, D., Aloni, R., Ben-Zvi, O. & Wollberg, M. (1999) A galling aphid furnishes its home with a built-in pipeline to the host food supply. Entomologia Experimentalis et Applicata,91, 183–186. https://doi .org/10.1046/j.1570-7458.1999.00482.x. Yarnes, C.T., Boecklen, W.J. & Salminen, J. (2008) No simple sum: seasonal variation in tannin phenotypes and leaf-miners in hybrid oaks. Chemoecology,18, 39–51. https://doi.org/10.1007/s00049-0070391-y. Zohner, C.M., Rockinger, A. & Renner, S.S. (2019) Increased autumn productivity permits temperate trees to compensate for spring frost damage. New Phytologist,221, 789–795. https://doi.org/10.1111/nph .15445. Zohner, C.M., Moa, L., Renner, S.S. et al. (2020) Late-spring frost risk between 1959 and 2017 decreased in North America but increased in Europe and Asia. Proceedings of the National Academy of Sciences, 117, 12192–12200. https://doi.org/10.1073/pnas.1920816117. Accepted 26 April 2021 © 2021 The Authors. Agricultural and Forest Entomology published by John Wiley & Sons Ltd on behalf of Royal Entomological Society. Agricultural and Forest Entomology, doi: 10.1111/afe.12448