Polar Biol (2015) 38:967–974 DOI 10.1007/s00300-015-1655-6

ORIGINAL PAPER

Abrupt changes in invertebrate herbivory on woody at the forest–tundra ecotone

Mikhail V. Kozlov • Boris Yu. Filippov • Natalia A. Zubrij • Vitali Zverev

Received: 20 September 2014 / Revised: 26 January 2015 / Accepted: 27 January 2015 / Published online: 4 February 2015 Ó Springer-Verlag Berlin Heidelberg 2015

Abstract Invertebrate herbivores, insects in particular, due to climate-driven range expansion and increased play important functional roles in terrestrial ecosystems. At abundances of -feeding insects. background (non-outbreak) densities, they consume 5–15 % of woody plant foliage in forests across the planet. Keywords Background herbivory Climatic changes Á Á At the same time, almost nothing is known about the levels Defoliators Gallers Miners Northern of insect herbivory in Arctic tundra. To test the hypothesis Á Á Á that the amount of plant biomass lost to insects in tundra is substantially less than in subarctic forests, in 2013, we Introduction explored foliar herbivory in woody plants at three sites in the Arctic tundra and four sites in the subarctic forests of Exploration of species-poor and fragile Arctic ecosystems European Russia. A vast majority of foliar damage was is crucial for understanding the general principles of eco- imposed by externally feeding defoliators. In forests, system functioning (Post et al. 2009; Wookey et al. 2009; defoliators damaged three times more and consumed Link et al. 2013). These studies have recently become eight times more area than in the tundra. No miners particularly important because climate change appears were found in the tundra, and gallers affected five times disproportionately rapid towards the poles (Walther et al. less leaf area in the tundra compared with forests. An 2002; Doney et al. 2012), and we are in serious danger of abrupt decrease in loss of woody plant foliage to insects losing knowledge of the present state of biotic interactions between subarctic forests and tundra (from 4.34 to 0.56 %) in the Arctic. On the other hand, ongoing climatic change supports the existence of a latitudinal gradient in herbivory motivates the need to rapidly assess the consequences of in terrestrial ecosystems. More studies are needed to pre- the warming process on the functioning of ecosystems at dict how tundra plants, which have been historically polar latitudes (Link et al. 2013). Rapid assessment may be exposed to low levels of insect herbivory, will cope with critical for processes mediated by arthropods, a group that the increased levels of damage that are expected to occur is expected to exhibit a particularly strong response to climatic changes in the high Arctic (Hodkinson and Bird 1998; Strathdee and Bale 1998). M. V. Kozlov (&) V. Zverev Section of Ecology,Á University of Turku, 20014 Turku, Finland Insect herbivores play important functional roles in e-mail: mikoz@utu.fi terrestrial ecosystems, from direct effects on plant growth, survival and reproduction to indirect regulation of evapo- B. Yu. Filippov transpiration and nutrient cycling processes (Mattson and Department of Zoology and Ecology, Northern (Arctic) Federal University, Severnaya Dvina Emb. 17, 163000 Arkhangelsk, Addy 1975; Seastedt and Crossley 1984; Hunter et al. Russia 2012). At background (non-outbreak) densities, insect herbivores consume 5–15 % of woody plant foliage in N. A. Zubrij forests across the planet (Coley and Aide 1991; Coley and Institute of Ecological Problems of the North, Ural Branch of the Russian Academy of Sciences, Severnaya Dvina Emb. 23, Barone 1996). At the same time, almost nothing is known 163000 Arkhangelsk, Russia about the levels of plant damage by insects in tundra. A

123 968 Polar Biol (2015) 38:967–974 vast majority of studies addressing herbivory in the Arctic (e.g. Olofsson et al. 2009; Wookey et al. 2009; Legagneux et al. 2012; Stien et al. 2012) focus on vertebrate herbi- vores, implicitly assuming a negligible role of insect her- bivory in high-latitude ecosystems. Moreover, researchers exploring plant-feeding insects in Arctic ecosystems (e.g. MacLean and Jensen 1985; Kukal and Dawson 1989; Ro- ininen et al. 2002; Lundbye et al. 2012; Roslin et al. 2013) usually do not collect information on the level of plant damage by insect herbivores. We are aware of only two Fig. 1 Study area and sampling sites quantitative studies of leaf area lost by woody plants from insect herbivory in the tundra biome (Olofsson et al. 2007; Torp et al. 2010), and these data originated from the rela- (Fig. 1; Table 1). All sites were selected in natural habitats tively southern shrubby tundra of Fennoscandia. with negligible levels of human-induced disturbance. The This shortage of information is especially critical selection of sampling localities and of the time of sampling because of the current debate on the existence of latitudinal was driven by factors other than the goals of the present patterns in insect herbivory. A recent meta-analysis (Moles study. In particular, all tundra sites were visited during the et al. 2011a) did not find the support for the hypothesis expedition of research vessel ‘Professor Molchanov’; (Coley and Barone 1996; Grime 2001) that herbivory therefore, our selection was presumed to be random in decreases with latitude (or the associated increase in relation to the existing level of herbivory which is variable environmental harshness with latitude). However, Kozlov in both space and time. et al. (2013) argued that conclusions by Moles et al. All tundra sites were located relatively close to the (2011a) may be valid only for regions with temperate cli- seashore. The maximum height of woody plants within mate and relatively smooth environmental gradients, 20–50 km from the sampling sites in the tundra was less between the 30th and 50th parallels in both hemispheres, than 1 m (Walker et al. 2005, and pers. obs.). The sampled whereas at high latitudes, herbivory decreases substantially communities on Kolguev Island and on Belyi Nos Penin- towards the poles. Latitudinal changes in herbivory may be sula were classified as low- tundra and on Vaygach especially pronounced within the transition zone between Island as graminoid prostrate dwarf-shrub tundra (after forest and tundra, which is one of the world’s most Walker et al. 2005). In surroundings of Murmansk, samples prominent ecotones and is associated with a strong climatic were collected in sparse low-stature woodland (5–7 m in gradient and an abrupt change in the structure and com- height) formed by mountain birch (Betula pubescens ssp. position of plant communities (Sveinbjo¨rnsson et al. 2002). czerepanovii (Orlova) Ha¨met-Ahti). This locality lies Along with climate, the differences in herbivory between beyond the northern distribution limit of Norway spruce these biomes may be caused by latitudinal changes in leaf (Picea abies (L.) Karst.) and forms a transition from north mechanical properties (Onoda et al. 2011), in plant taiga forest to low-shrub tundra. In Naryan-Mar, samples defensive chemistry (Moles et al. 2011a, b) and in predator were collected from a patch of sparse Norway spruce forest pressure (Bjo¨rkman et al. 2011). (12–15 m) within the forest–tundra zone. Study sites near In the present study, we explored foliar herbivory in Lovozero and Arkhangelsk were selected in dense north common species of woody plants at several sites in Arctic taiga forests formed by Norway spruce (14–18 m), birches tundra, forest–tundra and north taiga forests of European (mountain birch in Lovozero and white birch, Betula Russia. We aimed at quantifying community-wide losses in pendula Roth in Arkhangelsk) and European aspen (Pop- foliar biomass of woody plants to insects from three major ulus tremula L.). feeding guilds (defoliators, miners and gallers) to test the The summer of 2013 was significantly warmer relative hypothesis that invertebrate herbivory in tundra is sub- to long-term records (Table 1). However, the climatic stantially lower than in subarctic forests. gradient between tundra and forest sites was not affected, as can be concluded from significant correlations between long-term averages and actual temperatures of July across Materials and methods our study sites (r = 0.92, n = 7, P = 0.0038).

Study sites Sampling and processing

Leaves of woody plants were collected in 2013 from three In forests, we selected 3–6 species of woody plants that localities in tundra and four localities in forested habitats were most common at our study sites (except for conifers),

123 Polar Biol (2015) 38:967–974 969

Table 1 Characteristics of study sites

Locality Latitude, N Longitude, E Altitude, m a.s.l. Temperature of July, °C Vegetation zone and Sampling date community type Long-term average 2013

Vaygach Island 70°2501600 59°0303400 7 5.0 8.9 Low-shrub tundra 15–20 August 2013

Belyi Nos Peninsula 69°3601400 60°1204100 5 7.3 11.8 Low-shrub tundra 6 July 2013

Kolguev Island 68°4700600 49°1904600 7 7.5 12.5 Dwarf-shrub tundra 4 July 2013

Murmansk 68°5305900 33°4001100 195 12.6 14.8 Forest–tundra 18 August 2013

Naryan-Mar 67°3801200 53°0303700 10 13.3 17.5 Forest–tundra 11 July 2013

Lovozero 67°5301400 34°2501700 305 13.6 14.4 North taiga forest 15 August 2013

Arkhangelsk 64°2504900 40°5802700 20 16.3 16.6 North taiga forest 28 September 2013 while in tundra, we sampled all available woody species. number of leaves in the sample (including undamaged To avoid a bias, plant species to be sampled in forested leaves). Then the number of leaves in each damage class habitats were selected before visiting our sites, i.e. without was multiplied by the median value of the damaged leaf a knowledge on their current levels of damage by insects. area (i.e. 0.5 % for the damage class 0.01–1 %), and the This protocol allowed for comparison of community-wide obtained values were summed for all damage classes levels of herbivory between two biomes, forest and tundra, within a sample separately for each feeding guild. The on the basis of site-specific values. We collected one second response variable, average proportion of leaf area branch (with 100–200 leaves) from each of 2–5 individuals lost to (or damaged by) insects from each of these feeding (depending on plant abundance and the available time for guilds, was calculated by dividing the obtained values by sampling) of each species (for sample sizes, consult the total number of leaves in the sample (including Table 2). Plant individuals were selected by pointing at undamaged leaves). For defoliators, we also calculated the them from the distance of 10–15 m, from which the level third response variable, proportion of leaf area lost from of herbivory cannot be evaluated; this protocol minimized the damaged leaf, by dividing the obtained value by the the possibility of confirmation bias, i.e. the tendency of number of damaged leaves. humans to seek out evidence and interpret it in a manner The proportion of leaves damaged by external defoliators that confirms their existing ideas and hypotheses (Wilg- and the leaf area consumed by these insects (both total loss enburg and Elgar 2013). For both Vaccinium vitis-idaea L. and loss per damaged leaf) were log-transformed prior to the and Salix rotundifolia Trautv., we chose individual stems analysis to meet the normality assumption. Because each that had 10–40 leaves and aggregated several stems site was sampled only once, it was impossible to include the growing next to each other. All leaves from these branches/ sampling period (mid- or late summer/autumn) and the stems (including petioles of completely consumed leaves) study site in the same statistical model. Therefore, we used were collected and preserved between the sheets of paper two different models of ANOVA (SAS GLM procedure, as ordinary herbarium specimens. SAS Institute 2009) to test the hypotheses that (1) foliar Following widely used methodology (Southwood et al. damage does not differ between two sampling periods and 1982; Fox and Morrow 1983; Alliende 1989), each leaf (2) foliar damage differs between tundra and forest biomes. was attributed to one of the following damage classes Distributions of data on foliar damage by miners and gallers according to the proportion of the leaf area that was con- were greatly skewed (due to multiple zero values) and were sumed or damaged (galled, mined or skeletonized): intact therefore analysed by a nonparametric Kruskal–Wallis test. leaves, 0.01–1, 1.1–5, 6–25, 26–50, 51–75 and 76–100 %. Additionally, for defoliators, we correlated response vari- The numbers of leaves in each damage class were recorded ables to mean temperatures of July, which were previously separately for each feeding guild. All samples were pro- identified as the best predictor of plant losses to defoliators, cessed by the same person (MVK). leafminers and sap-feeders in north taiga forests (Kozlov 2008; Kozlov et al. 2013, 2015). Data analysis

The first response variable, the proportion of leaves dam- Results aged by each of three feeding guilds (defoliators, miners, gallers), was calculated by dividing the number of leaves The vast majority of damage recorded on leaves of woody damaged by insects from the respective guild by the total plants was imposed by externally feeding defoliators. Only

123 970 Polar Biol (2015) 38:967–974

Table 2 Levels of foliar damage of woody plants by insect herbivores Localitya Plant species Sample size Foliar damage (%), mean ± SEb (numbers of plants/ leaves) Defoliators Miners Gallers DL LA DL LA DL LA

Vaygach Salix lanata L. 3/372 35.5 ± 10.0 1.41 ± 0.66 0 0 0 0 Island Salix myrtilloides L. 3/388 9.4 ± 3.0 0.67 ± 0.29 0 0 0 0 Salix reticulata L. 3/414 4.6 ± 1.9 0.81 ± 0.41 0 0 1.0 ± 0.6 0.14 ± 0.10 Salix rotundifolia Trautv. 3/556 7.7 ± 0.9 0.30 ± 0.01 0 0 0.6 ± 0.5 0.09 ± 0.07 Mean values 14.3 0.80 0 0 0.4 0.6 Kolguev Salix lanata L. 2/222 40.2 ± 28.6 1.69 ± 0.70 0 0 0 0 Island Salix lapponum L. 4/516 5.0 ± 4.4 0.19 ± 0.11 0 0 0 0 Salix myrtilloides L. 2/262 4.0 ± 0.9 0.29 ± 0.08 0 0 0 0 Salix phylicifolia L. 3/339 8.5 ± 0.7 0.40 ± 0.25 0 0 0 0 Mean values 14.4 0.64 0 0 0 0 Belyi Nos Salix arctica Pall. 1/167 1.2 0.04 0 0 0 0 Peninsula Salix lanata L. 4/550 2.7 ± 0.8 0.01 ± 0.00 0 0 0 0 Salix myrtilloides L. 4/551 6.3 ± 3.0 0.95 ± 0.57 0 0 0.2 ± 0.2 0.01 ± 0.01 Mean values 3.4 0.33 0 0 0.07 0.003 Naryan-Mar Alnus viridis ssp. fruticosa 2/200 69.7 ± 11.7 11.33 ± 0.56 0 0 0 0 (Ruprecht) Nyman Betula nana L. 2/200 3.5 ± 1.5 0.20 ± 0.16 0 0 0 0 Betula pubescens Ehrh. 2/200 27.5 ± 5.5 0.89 ± 0.42 0 0 0 0 Salix lanata L. 2/200 82.5 ± 15.5 9.80 ± 5.27 0 0 26.6 ± 26.6 1.22 ± 1.22 Salix phylicifolia L. 2/200 73.5 ± 10.5 11.26 ± 4.60 0 0 0 0 L. 2/200 0 0 0 0 0 0 Mean values 42.8 3.69 0 0 4.4 0.2 Murmansk Betula nana L. 5/500 2.6 ± 1.9 0.84 ± 0.58 0 0 0 0 Betula pubescens ssp. 5/505 58.6 ± 7.1 4.13 ± 1.75 0.8 ± 0.5 0.05 ± 0.03 0 0 czerepanovii (Orlova) Ha¨met-Ahti Salix phylicifolia L. 5/500 26.2 ± 7.5 2.89 ± 0.86 0 0 0 0 Vaccinium myrtillus L. 5/500 3.1 ± 0.8 0.68 ± 0.42 0 0 0 0 Mean values 22.6 2.14 0.2 0.01 0 0 Lovozero Betula nana L. 3/300 11.3 ± 4.9 2.81 ± 1.85 0 0 0 0 Betula pubescens ssp. 5/499 42.9 ± 9.2 3.35 ± 1.53 1.2 ± 1.0 0.06 ± 0.05 0 0 czerepanovii (Orlova) Ha¨met-Ahti Salix phylicifolia L. 5/500 77.8 ± 4.3 19.01 ± 3.08 0 0 0 0 Vaccinium myrtillus L. 5/500 22.2 ± 13.1 10.77 ± 7.68 0 0 0 0 Vaccinium uliginosum L. 5/500 22.4 ± 9.2 8.61 ± 5.95 0 0 0 0 Vaccinium vitis-idaea L. 4/367 2.4 ± 0.7 0.39 ± 0.21 0 0 0 0 Mean values 29.8 7.49 0.2 0.01 0 0 Arkhangelsk Betula pendula Roth 3/394 71.6 ± 12.3 3.69 ± 1.54 1.2 ± 0.4 0.01 ± 0.01 17.1 ± 7.6 0.49 ± 0.30 Populus tremula L. 3/334 47.0 ± 15.2 1.03 ± 0.53 0 0 17.5 ± 4.7 0.18 ± 0.09 Vaccinium myrtillus L. 3/327 9.1 ± 2.3 0.34 ± 0.10 0 0 0 0 Mean values 42.6 1.69 0.4 0.003 11.5 0.22 a For characteristics of localities, consult Table 1 b Measures of foliar damage: DL, proportion of leaves damaged by insects from the respective feeding guild; LA, average proportion of leaf area lost to (or damaged by) insects from the respective feeding guild. Standard errors (SE) reflect variation among plant individuals

15 of 11263 examined leaves were mined by moth larvae Neither proportion of leaves damaged by external defoli- (from families Stigmellidae and Gracillariidae), and only ators (mean ± SE based on site-specific values: mid-summer, 186 leaves bore galls formed by insects or mites. 23.8 ± 8.5 %; late summer/autumn: 26.7 ± 6.0 %;

123 Polar Biol (2015) 38:967–974 971

Table 3 Sources of variation in characteristics of plant damage by defoliating insects (ANOVA, type III sum of squares). For definition of response variables, consult text Source df Proportion of damaged Proportion of leaf area lost from Proportion of leaf area lost leaves the damaged leaf to defoliating insects Mean square FPMean square FPMean square FP

Biome (tundra vs. forest) 1 6.62 4.93 0.04 2.21 6.36 0.02 12.35 7.54 0.01 Site (Biome) 5 0.74 0.55 0.74 1.05 3.01 0.03 1.09 0.66 0.65 Error 23 1.34 0.35 1.64

F1,28 = 0.91, P = 0.35) nor the total loss of leaf area caused sampled in both biomes (S. lanata and S. phylicifolia) was by these insects (mid-summer, 2.80 ± 1.27 %; late summer/ higher in forest sites than in tundra sites (Table 2), and the autumn: 3.51 ± 1.20 %; F1,28 = 1.81, P = 0.91) differed average proportion of leaf area lost by these two between the two sampling periods. This conclusion was valid species from the damaged leaf in forests (14.98 ± 3.12 %) also for miners (v2 = 2.49, df = 1, P = 0.11) and gallers was nearly five times as high as in tundra (3.29 ± 1.98 %; 2 (v = 0.51, df = 1, P = 0.58). Therefore, samples collected F1,6 = 12.8, P = 0.01). in mid-summer were combined with samples collected in late summer or autumn for the analysis of differences in herbivory between forest and tundra biomes. Discussion Plant species collected from the same site generally differed in the level of damage by defoliators at forest sites Our study demonstrated that insect herbivores in tundra but not at tundra sites (Table 2). Leaf damage by defolia- ecosystems consumed only 0.56 % of the foliar biomass of tors at forest sites significantly exceeded the damage at woody plants, which was much lower than in forest eco- tundra sites (Table 3). The proportion of leaf damage and systems close to the northern tree limit (4.34 %). This the total loss of leaf area were three times and eight times result is in line with the conclusion by McNaughton et al. higher, respectively, in the forest than in the tundra (1989) that highly productive ecosystems sustain a larger (Fig. 2). Consistently, the average proportion of leaf area level of herbivory per unit of primary production than less lost to defoliators from the damaged leaf in forests productive ecosystems. Furthermore, the detected differ- (11.64 ± 3.74 %) was twice as high as in tundra ences between losses of woody plant foliage to insects in (6.90 ± 1.58 %). The proportion of leaves damaged by northern forests and tundra habitats, that were separated by defoliators significantly (r = 0.78, n = 7 sites, P = 0.04) (on average) 2.5° latitude (Table 1), were of the same increased with long-term mean temperature of July, and magnitude as the differences between northern and south- average proportion of leaf area lost to these insects also ern forests in Fennoscandia, separated by 10° latitude tended to increase with temperature (r = 0.71, n = 7, (Kozlov 2008 and unpublished data). This result supports P = 0.07). Although no miners were found in tundra sites the hypothesis (Kozlov et al. 2013) that gradients in and gallers affected five times less leaf area in tundra than background herbivory are stronger at high latitudes than in in forests (Fig. 2), the difference between biomes in plant regions with temperate climates, in particular because the damage by these two feeding guilds did not reach the level slope of the temperature gradient increases towards the of statistical significance (miners: v2 = 2.19, df = 1, poles (Terborgh 1973). This discovery emphasizes the need P = 0.14; gallers: v2 = 0.06, df = 1, P = 0.80). for further study of both the levels of the background (Salix spp.) were the only group of woody herbivory and associated consequences for the fitness of plants sampled in both forest and tundra biomes (although plants in the Arctic. In particular, systematic measurements the species composition of willows differed between for- of invertebrate herbivory north of the Arctic circle, from ests and tundra: Table 2). Within willows, both the mag- north taiga forests to polar deserts, are of special impor- nitude and statistical significance of the differences tance for revealing the general pattern in the relationship between the biomes in the amount of chewing herbivory between latitude, woody plant diversity, plant life form and were higher than in the community-wide comparisons losses of foliage to insects. (average proportion of damaged leaves, forest: 60.7 ± The absolute levels of foliar herbivory in our tundra sites

17.2 %, tundra: 9.5 ± 3.2 %, F1,9 = 30.2, P = 0.0004; are similar to earlier data from tundra sites in Iceland, average proportion of leaf area lost to insects, forest: where defoliating insects damaged 1.0–13.1 % (mean 10.81 ± 4.66 %, tundra: 0.54 ± 0.14 %, F1,9 = 74.5, P \ 5.2 %) of leaves in Salix herbacea L. and 1.0–26.2 % 0.0001). The damage of two willow species that were (mean 6.4 %) of leaves in Vaccinium uliginosum L.

123 972 Polar Biol (2015) 38:967–974

in one of the subarctic forest sites during the last 30 years (a) 40 Forest (Hunter et al. 2014), and herbivory on birches at the Tundra northern tree limit during the years with warmer-than- average summer temperatures has increased to the level 30 typical for habitats located 500–800 km to the south (Kozlov et al. 2013). Thus, climate-driven range expansion 20 and increased abundances of species with generally more southern distributions may add to the species pool and * abundance of herbivores in tundra ecosystems, thereby 10 Damaged leaves (%) increasing herbivory pressure on arctic plants, because woody plants are likely to move northwards much slower than insects. 0 Chewers Miners Gallers The consequences of increased insect herbivory for Arctic ecosystems are difficult to predict due to acute (b) shortage of information on insect–plant relationships in the tundra. Although this knowledge gap was identified long 5 ago (Haukioja 1981; Danks 1986), little progress has been made in this area of research (but see Lundbye et al. 2012). 4 In northern taiga forests, a relatively small increase in background herbivory due to climate warming can poten- 3 tially cause severe negative impacts on tree growth (Zvereva et al. 2012). However, the question arises of whether these 2 predictions can be similarly applied to the tundra biome. Moles et al. (2011a) surprisingly found that chemical 1 * defences in plants from higher latitudes were significantly higher than in plants from lower latitudes. Our data indi- Consumed / damaged leaf area (%) 0 Chewers Miners Gallers rectly support this conclusion because in tundra, defoliating Fig. 2 Proportion of damaged leaves (a) and proportion of leaf area insects consumed a significantly smaller part of any dam- lost to (or damaged by) insects (b) from three feeding guilds in aged leaf than in forests, which can indicate that they are subarctic forests (n = 4) and tundra sites (n = 3). Bars indicate SE changing their feeding sites more frequently than in forests. based on site-specific means; an asterisk indicates significant Restriction of the comparison to two willow species that difference between biomes (for statistics, consult Table 3 and text) were sampled in both biomes further confirmed this con- clusion. Changing of feeding sites allows herbivores to (Kozlov et al. 2009). In contrast, all published data from avoid local defensive responses of their host plants tundras of northern Fennoscandia (44.3 % of leaves dam- (Edwards and Wratten 1983; Bergelson et al. 1986; Zver- aged in S. lanata: Olofsson and Strengbom 2000; 2.56 % of eva and Kozlov 2000), and changing the feeding site at leaf area consumed in S. glauca: Olofsson et al. 2007; and lower levels of leaf damage can indicate that woody plants 10.0 % of leaf area consumed in B. nana: Torp et al. 2010) in tundra have a stronger decrease in leaf palatability in reported foliar damage levels in the same range as we response to damage than do woody plants in forests. If our found in the northern forests. However, Fennoscandia is a interpretation is correct, then plant defence may hamper the region that does not include ‘true’ tundra according to the use of tundra plants by more southern herbivores, which classification by Walker et al. (2005), may need some time to adapt to the high level of plant Low levels of foliar herbivory in the tundra are consis- defences. On the other hand, the predicted increase in tent with the low diversity and relatively low abundances herbivory is likely to be valid for that have recently of herbivorous insects in this biome. According to Danks invaded the tundra (Sturm et al. 2001; Tape et al. 2006) (1986), the ratio between the number of insect species because most of these shrubs are growing in subarctic feeding on plants and the number of plant species forests and therefore are familiar to more southern herbi- decreased from 3 to 4 in temperate and northern forests to vores. In this case, the increase in insect herbivory, along 0.1–1.2 in tundra. However, the situation is likely to with mammalian herbivory (Olofsson et al. 2009), has a change in the near future, as many insect species have been potential to slow down the rate of shrub encroachment in reported to expand their distribution towards the North tundra ecosystems. (Parmesan et al. 1999; Warren et al. 2001). Furthermore, To conclude, our findings on the sharp decrease in foliar the abundances of many herbivorous moths have increased damage of woody plants by insects between subarctic

123 Polar Biol (2015) 38:967–974 973 forests and arctic tundra support the existence of a latitu- Hunter MD, Kozlov MV, Ita¨mies J, Pulliainen E, Ba¨ck J, Kyro¨ E-M, dinal gradient in herbivory in terrestrial ecosystems. More Niemela¨ P (2014) Current temporal trends in moth abundance are counter to predicted effects of climate change in an studies are needed to predict how tundra plants, which have assemblage of subarctic forest moths. Glob Change Biol been historically exposed to low levels of folivory, will 20:1723–1737 cope with increased levels of damage resulting from cli- Kozlov MV (2008) Losses of birch foliage along geographical mate-driven range expansion and the increased abundance gradients in Northern and Central Europe: a climate-driven pattern? Clim Change 87:107–117 of insect herbivores. Kozlov MV, Zvereva EL, Zverev VE (2009) Impacts of point polluters on terrestrial biota: comparative analysis of 18 Acknowledgments We thank E. Yu. Churakova for help in iden- contaminated areas. Springer, Dordrecht tifying plants and E. L. Zvereva and three anonymous reviewers for Kozlov MV, van Nieukerken EJ, Zverev V, Zvereva EL (2013) commenting on an earlier draft of the manuscript. Research visits by Abundance and diversity of birch-feeding leafminers along MVK and VZ to the study sites were supported by INTERACT (Grant latitudinal gradients in Northern Europe. Ecography agreement no. 262693 under the EC 7th Framework Programme) and 36:1138–1149 by the Otto Malm´ ’s Foundation. Fieldwork by BYF and NAZ was Kozlov MV, Stekolshchikov A, So¨derman G, Labina E, Zverev V, supported by the Ministry of Education and Science of the Russian Zvereva EL (2015) Sap-feeding insects on forest trees along Federation (Grant 5.4615.2011) and the Russian Foundation for the latitudinal gradients in northern Europe: a climate-driven pattern. Basic Research (11-04-98814-north). Glob Change Biol 21:106–116 Kukal O, Dawson TE (1989) Temperature and food quality influences feeding behavior, assimilation efficiency and growth rate of arctic woolly-bear caterpillars. Oecologia 79:156–532 References Legagneux P, Gauthier G, Berteaux D, Beˆty J, Cadieux M-C, Bilodeau F, Bolduc E, McKinnon L, Tarroux A, Therrien JF, Alliende MC (1989) Demographic studies of a dioecious tree. II. The Morissette L, Krebs CJ (2012) Disentangling trophic relation- distribution of leaf predation within and between trees. J Ecol ships in a High Arctic tundra ecosystem through food web 77:1048–1058 modelling. Ecology 93:1707–1716 Bergelson J, Fowler S, Hartley S (1986) The effects of foliage Link H, Piepenburg D, Archambault P (2013) Are hotspots always damage on casebearing moth larvae, Coleophora serratella, hotspots? The relationship between diversity, resource and feeding on birch. Ecol Entomol 11:241–250 ecosystem functions in the Arctic. PLoS ONE 8:e74077 Bjo¨rkman C, Berggren A, Bylund H (2011) Causes behind insect Lundbye H, Johansson DK, Andersen MR, Philipp M (2012) The folivory patterns in latitudinal gradients. J Ecol 99:367–369 effect of a seed-sucking bug on seed germination of an arctic Coley PD, Aide TM (1991) Comparison of herbivory and plant cushion plant. Ecoscience 19:209–212 defenses in temperate and tropical broad-leaved forests. In: Price MacLean SF Jr, Jensen TS (1985) Food plant selection by insect PW, Lewinsohn TM, Fernandes GW, Benson WW (eds) Plant– herbivores in Alaskan arctic tundra: the role of plant life form. animal interactions: evolutionary ecology in tropical and tem- Oikos 44:211–221 perate regions. Wiley, New York, pp 25–49 Mattson WJ, Addy ND (1975) Phytophagous insects as regulators of Coley PD, Barone JA (1996) Herbivory and plant defenses in tropical forest primary production. Science 190:515 forests. Ann Rev Ecol Syst 27:305–335 McNaughton SJ, Oesterheld M, Frank DA, Williams KJ (1989) Danks HV (1986) Insect plant interactions in arctic regions. Rev Ecosystem-level patterns of primary productivity and herbivory Entomol Que´ 31:52–75 in terrestrial habitats. Nature 341:142–144 Doney SC, Ruckelshaus M, Duffy JE, Barry JP, Chan F, English CA, Moles AT, Bonser SP, Poore AGB, Wallis JR, Foley WJ (2011a) Galindo HM, Grebmeier JM, Hollowed AB, Knowlton N, Assessing the evidence for latitudinal gradients in plant defence Polovina J, Rabalais NN, Sydeman WJ, Talley LD (2012) and herbivory. Funct Ecol 25:380–388 Climate change impacts on marine ecosystems. Annu Rev Mar Moles AT, Wallis IR, Foley WJ, Warton DI, Stegen JC, Bisigato AJ, Sci 4:11–37 Cella-Pizarro L, Clark CJ, Cohen PS, Cornwell WK, Edwards Edwards PJ, Wratten SD (1983) Wound induced defenses in plants W, Ejrnæs R, Gonzales-Ojeda T, Graae BJ, Hay G, Lumbwe FC, and their consequences for patterns of insect grazing. Oecologia Magan˜a-Rodrı´guez B, Moore BD, Peri PL, Poulsen JR, Veldt- 59:88–93 man R, von Zeipel H, Andrew NR, Boulter SL, Borer ET, Fox LR, Morrow PA (1983) Estimates of damage by herbivorous Campo´n FF, Coll M, Farji-Brener AG, De Gabriel J, Jurado E, insects on Eucalyptus trees. Aust J Ecol 8:139–147 Kyhn LA, Low B, Mulder CPH, Reardon-Smith K, Rodrı´guez- Grime JP (2001) Plant strategies, vegetation processes, and ecosystem Vela´zquez J, Seabloom EW, Vesk PA, van Cauter A, Waldram properties. Wiley, Chichester MS, Zheng Z, Blendinger PG, Enquist BJ, Facelli JM, Knight T, Haukioja E (1981) Invertebrate herbivory at tundra sites. In: Bliss LC, Majer JD, Martinez-Ramos M, McQuillan P, Prior LD (2011b) Cragg JB, Heal DW, Moore JJ (eds) Tundra ecosystems: a Putting plant resistance traits on the map: a test of the idea that comparative analysis. Cambridge Univ Press, Cambridge, plants are better defended at lower latitudes. New Phytol pp 547–555 191:777–788 Hodkinson ID, Bird J (1998) Host-specific herbivores as sensors of Olofsson J, Strengbom J (2000) Response of galling invertebrates on climate change in Arctic and alpine environments. Arctic Alpine Salix lanata to reindeer herbivory. Oikos 91:493–498 Res 30:78–83 Olofsson J, Dahlgren J, Witzell J (2007) Grey-sided voles increase the Hunter MD, Reynolds BC, Hall MC, Frost CJ (2012) Effects of susceptibility of Northern willow, Salix glauca, to invertebrate herbivores on ecosystem processes: the role of trait-mediated herbivory. Ecoscience 14:48–54 indirect effects. In: Ohgushi T, Schmitz OJ, Holt R (eds) Trait- Olofsson J, Oksanen L, Callaghan T, Hulme PE, Oksanen T, mediated indirect interactions: ecological and evolutionary Suominen O (2009) Herbivores inhibit climate-driven shrub perspectives. Cambridge Univ Press, Cambridge, pp 339–370 expansion on the tundra. Glob Change Biol 15:2681–2693

123 974 Polar Biol (2015) 38:967–974

Onoda Y, Westoby M, Adler PB, Choong AMF, Clissold FJ, Sveinbjo¨rnsson J, Hofgaard A, Lloyd A (2002) Natural causes of the Cornelissen JHC, Diaz S, Dominy NJ, Elgart A, Enrico L, Fine tundra-taiga boundary. Ambio Spec Rep 12:23–29 PVA, Howard JJ, Jalili A, Kitajima K, Kurokawa H, McArthur Tape K, Sturm M, Racine C (2006) The evidence for shrub expansion C, Lucas PW, Markesteijn L, Pe´rez-Harguindeguy N, Poorter L, in Northern Alaska and the Pan-Arctic. Glob Change Biol Richards L, Santiago LS, Sosinski EE, Van Bael SA, Warton DI, 12:686–702 Wright IJ, Wright SJ, Yamashita N (2011) Global patterns of leaf Terborgh J (1973) On the notion of favorableness in plant ecology. mechanical properties. Ecol Lett 14:301–312 Am Nat 107:481–501 Parmesan C, Ryrholm N, Stefanescu C, Hill JK, Thomas CD, Torp M, Olofsson J, Witzell J, Baxter R (2010) Snow-induced Descimon H, Huntley B, Kaila L, Kullberg J, Tammaru T, changes in dwarf birch chemistry increase moth larval growth Tennent WJ, Thomas JA, Warren M (1999) Poleward shifts in rate and level of herbivory. Polar Biol 10:693–702 geographical ranges of butterfly species associated with regional Walker DA, Raynolds MK, Danie¨ls FJA, Einarsson E, Elvebakk A, warming. Nature 399:579–583 Gould WA, Katenin AE, Kholod SS, Markon CJ, Melnikov ES, Post E, Forchhammer MC, Bret-Harte MS, Callaghan TV, Christen- Moskalenko NG, Talbot SS, Yurtsev BA, the other members of sen TR, Elberling B, Fox AD, Gilg O, Hik DS, Høye TT, Ims the CAVM Team (2005) The circumpolar arctic vegetation map. RA, Jeppesen E, Klein DR, Madsen J, McGuire AD, Rysgaard S, J Veget Sci 16:267–282 Schindler DE, Stirling I, Tamstorf MP, Tyler NJC, van der Wal Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, R, Welker J, Wookey PA, Schmidt NM, Aastrup P (2009) Fromentin JM, Hoegh-Guldberg O, Bairlein F (2002) Ecological Ecological dynamics across the Arctic associated with recent responses to recent climate change. Nature 416:389–395 climate change. Science 325:1355–1358 Warren MS, Hill JK, Thomas JA, Asher J, Fox R, Huntley B, Roy Roininen H, Danell K, Zinovjev A, Vikberg V, Virtanen R (2002) DB, Telfer MG, Jeffcoate S, Harding P, Jeffcoate G, Willis SG, Community structure, survival and mortality factors in arctic Greatorex-Davies JN, Moss D, Thomas CD (2001) Rapid populations of Eupontania leaf gallers. Polar Biol 25:605–611 responses of British butterflies to opposing forces of climate Roslin T, Wirta H, Hopkins T, Hardwick B, Va´rkonyi G (2013) and habitat change. Nature 414:65–69 Indirect interactions in the high Arctic. PLoS ONE 8:e67367 Wilgenburg E van, Elgar MA (2013) Confirmation bias in studies of SAS Institute (2009) SAS version 9.2 for Windows. SAS Institute, nestmate recognition: a cautionary note for research into the Cary behaviour of animals. PLoS ONE 8(1):e53548 Seastedt TR, Crossley OA Jr (1984) The influence of arthropods on Wookey PA, Aerts R, Bardgett RD, Baptist F, Brathen KA, ecosystems. Bioscience 34:157–161 Cornelissen JHC, Gough L, Hartley IP, Hopkins DW, Lavorel Southwood TRE, Moran VC, Kennedy CEJ (1982) The richness, S, Shaver GR (2009) Ecosystem feedbacks and cascade abundance and biomass of the arthropod communities on trees. processes: understanding their role in the responses of Arctic J Anim Ecol 51:635–649 and alpine ecosystems to environmental change. Glob Change Stien A, Ims RA, Albon SD, Fuglei E, Irvine RJ, Ropstad E, Biol 15:1153–1172 Halvorsen O, Langvatn R, Loe LE, Veiberg V, Yoccoz NG Zvereva EL, Kozlov MV (2000) Pollution suppresses delayed (2012) Congruent responses to weather variability in high arctic inducible resistance in boreal willow Salix borealis. Ecol Lett herbivores. Biol Lett 8:1002–1005 3:85–89 Strathdee AT, Bale JS (1998) Life on the edge: insect ecology in Zvereva EL, Zverev VE, Kozlov MV (2012) Little strokes fell great arctic environments. Annu Rev Entomol 43:85–106 oaks: minor but chronic herbivory substantially reduces birch Sturm M, Racine C, Tape K (2001) Climate change—increasing growth. Oikos 121:2036–2043 shrub abundance in the Arctic. Nature 411:546–547

123