Both I. cembrae and T. gabrieli are native to the Alps in Central Europe where European larch (Larix decidua, Mill.) is the host tree. Although Siberian larch (Larix sibirica, Ledeb.) occurred naturally in Scandinavia shortly after the last glacial period (Kullman, 1998), it disappeared from the Scandinavian peninsula probably due to changes in the climatic conditions (Larsson-Stern, 2003). In Sweden, the first recorded plantation of European larch dates from 1763 (Schotte, 1917 in Larsson-Stern, 2003). Today, Larix x eurolepis, (A Henry), a hybrid between the European and the Japanese larch, (L. kaempferi, Lamb. Carr), is the most planted larch variety in Sweden (Johansson, 2012; Larsson-Stern, 2003). In southern Sweden (Götaland), larch constitutes 0.2% of the total growing stock including all tree species (Swedish University of Agricultural Sciences, 2022) while in Denmark larch constitutes 3.4% of total growing stock (Nord-Larsen et al., 2020). The increased use of larch in forest plantations is thought to have enabled I. cembrae and T. gabrieli to spread and establish in new areas in Europe (Ravn, 2012; Lindelöw et al., 2015; Lynikiene et al., 2021; Grégoire et al., 2024).

The current known distribution of I. cembrae outside its native area ranges from UK and the Netherlands in the west (Luitjes, 1974; Alexander, 2007), to Denmark and Sweden in the north (Ravn, 2012; Lindelöw et al., 2015; Grégoire et al., 2024). Ips cembrae generally colonizes weakened trees. In the native range outbreaks of I. cembrae have occurred after storm feelings and snow-breakages (Grégoire and Evans, 2004; Grodzki, 2008). The flight period in central Europe occurs between the end of April to the middle of September (Grucmanová et al., 2014; Holuša et al., 2014). It is not known if I. cembrae is attracted to host volatiles although it is known that I. cembrae is not attracted to α-pinene (Stoakley et al., 1978). The male initiates colonization of breeding substrate by boring into the bark. After successful colonization, an aggregation pheromone is released by the males that is strongly attractive to both males and females. The males create a mating chamber and is joined by three to five females, which create maternal galleries where they oviposit. After producing a first brood the parent beetles may emerge to establish a second brood (sister brood) in another tree. The species can have one to two generations per year, depending on environmental conditions.

The current distribution of T. gabrieli outside its native area ranges from the UK in the west to Denmark and southern Sweden, and recently Lithuania, towards the northeast (Crawshay, 1907; Ericson, 2010; Kahanpää, 2017; Lynikiene et al., 2021). Like I. cembrae, T. gabrieli is considered a secondary pest, but when larch trees are weakened due to drought it may cause considerable tree mortality (Evans et al., 2004). The flight period starts in May and extends to early August in southern Sweden (Schroeder et al., 2020). Mating may take place right after the emergence of the new generation, possibly even before dispersal, as well as later in the summer (Hanks, 1999). Thus, in contrast to I. cembrae, mated females can colonize (i.e., lay their eggs) host trees independent of males, and aggregation with conspecifics to overwhelm tree defence is not required. It is not known if T. gabrieli uses host tree volatiles to locate breeding substrate, but it is shown that the species was not attracted to α-pinene and ethanol (Schroeder et al., 2020). However, as it has been demonstrated that a closely related species, T. fuscum, responds to a blend of five monoterpenes that are released from stressed spruce trees (Sweeney et al., 2004), it is possible that also T. gabrieli is attracted to host volatiles. Males of T. gabrieli are thought to release a pheromone (fuscumol) as it was found to attract both males and females (Schroeder et al., 2020). Three other Tetropium species are attracted to fuscumol and the pheromone has been demonstrated to be released by the males of two of these species (Silk et al., 2007; Sweeney et al., 2010).

The study was conducted in Denmark (in the regions Zealand, Funen, South Denmark, Mid and North Jutland), and Southern Sweden (in the regions Skåne, Småland, Blekinge, Kalmar and Halland), during 2020. Both I. cembrae and T. gabrieli had previously been detected in many of the regions included in the current study (Isacsson, unpublished data; Ravn and Harding, 1995; Ravn, 2012; Lindelöw et al., 2015; Schroeder et al., 2020). With the above mentioned regions, we refer to the administrative entities in Denmark and southern Sweden, except for the island of Funen, which, although administratively belongs to Southern Jutland, we defined as its own entity due to its isolated location. The number of study locations differed between regions spanning from one (Funen) to 14 (Skåne). The first records of T. gabrieli in Denmark dates from 1890, with a few subsequent sightings, but with a limited distribution (Hansen, 1996). The first report of I. cembrae in Denmark dates from 1995 (Ravn and Harding, 1995) and it is currently established over large parts of the country (Ravn, 2012). The first record of T. gabrieli in Sweden occurred in 2007 in the province of Blekinge (Ericson, 2010). Ips cembrae was first reported in Sweden in 2011 in the province of Skåne (Lindelöw et al., 2015).

Forty-two stands of hybrid larch were selected, ranging from 0.79 to 9.74 ha in size. All stands were monocultures except for five stands, which had undergrowth trees, consisting mainly of spruce. A subset of the stands had previously been included in a monitoring program for I. cembrae conducted by Swedish Forest Agency (Lindelöw et al., 2015) while others were selected from larch stand data provided by forest owners. The minimum distance between selected stands was 1.8 km. For each stand the diameter at 1.3 m height of 40 randomly selected trees (10 trees around each trap position, see below) were measured. Only one standing tree, successfully attacked by T. gabrieli in the previous year (2019), was found in two of our focal stands (both in Sweden, one in Blekinge and one in Kalmar region). In the remaining stands, there were no signs of any local hibernating populations of neither I. cembrae nor T. gabrieli. In the 42 stands, the relative population density of I. cembrae and T. gabrieli were estimated with pheromone-baited traps. In addition, in 34 stands colonization success was assessed by felling larch trees in early spring that subsequently were inspected for attacks.

To analyze the relationship between amount of larch in the adjacent landscape and the relative population density of I. cembrae and T. gabrieli, we used two buffer zones with a radius of 500 and 1,000 m respectively, from the center of each focal stand. We chose 1,000 m as the maximum buffer radius not to violate the independence of our observations (i.e., overlapping buffers). In addition to the focal stand, the quantity of larch (in hectares) in the created buffer zones was calculated. We used the quantity of larch as a proxy for presence of suitable breeding material like weakened standing trees, wind-felled trees and logging residues. The mean, standard deviations and range of stand and landscape variables are provided in Table 1. For all stands in Denmark, and a subset of ten stands in Sweden, the forest owners provided detailed maps with all the larch stands around the focal stands. For the remaining stands in Sweden maps were lacking and we estimated the amount of larch based on aerial images Google maps and maps from the Swedish Forest Agency website.¹ The estimates from aerial images were accurate as demonstrated by a check against the detailed maps provided by the ten forest owners in Sweden.

To estimate the relative population density of I. cembrae and T. gabrieli, for each of the selected stands, we used two pheromone-baited traps (ECONEX MULTIFUNNEL-12^®, Econex SL, Murcia, Spain) for each species. The trapping was standardized by placing the traps inside larch stands, at least 10 m from the stand edge with a minimum distance of 50 m between them. The traps were hung from two 1.8 m wooden poles. A 1:1 mixture of propylene glycol and water was used in the trap jars. Cemsan dispensers (Flügel Gmbh, Germany; active substance 3-Methyl-3-buten-1-ol), were used for attracting I. cembrae and aPhinity BSLB dispensers (racemic E-fuscumol, Andermatt Biocontrol, Switzerland) for attracting T. gabrieli. The lures were not changed during the experiment since the attraction of the lures should last for the duration of an entire trapping season based on the pheromone data sheet provided by the manufacturer. All traps were baited between 13th and 19th of April 2020 and emptied four times between 12th of May and 27th of July 2020, at approximately 3 weeks intervals. This period includes the major part of the flight activity of T. gabrieli as shown in an earlier trapping study conducted in the study area (Schroeder et al., 2020). There is no study on seasonal flight activity of I. cembrae from Scandinavia but the considerably slower development rate, and higher temperature requirements, compared with I. typographus (Schebeck and Schopf, 2017) suggests that most of the flight has occurred during the trapping period (c.f. Fritscher and Schroeder, 2022). As the important predator of I. cembrae, Thanasimus spp., is attracted to bark beetle pheromone, the funnel traps also provided an estimate of their relative population density.

Ips cembrae adults were determined based on morphological characteristics using the identification key from Douglas et al. (2019), and sexed by inspecting the genitalia through dissection. Tetropium gabrieli adults were determined based on Pfeffer (1995), and Ehnström and Holmer (2007), and sexed by inspection of the tip of the abdomen (Yanega, 1996). As in many Cerambycidae, a portion of the sixth abdominal tergite is visible beyond the pygidium of males.

Both I. cembrae and T. gabrieli are known to colonize felled trees (Lindelöw et al., 2015; Resnerová et al., 2020; Schroeder et al., 2020). Five larch trees were felled in early spring (2–8 March in Denmark and 16–22 March 2020, in Sweden) in each of the 34 selected stands. Stands with felled trees had a minimum size of 1.8 ha. The minimum distance between the felled trees within stands was 30 m and with a minimum distance of 50 m from the pheromone-baited traps. The trees to be felled were randomly selected (although the distance to the previous felled tree was kept to a minimum of 30 meters). At each trap emptying, the whole length of the trunk, and branches with a diameter exceeding 3 cm of the felled trees, were inspected for the presence of boring dust and entrance holes, indicating possible attacks of I. cembrae. If boring dust was found, the bark was removed and species present were determined. At the final inspection (15–25 September), on every half meter of the trunk, pieces of bark were peeled off with an axe, to check for possible larval galleries of Tetropium (not possible to determine to species level from galleries) and I. cembrae attacks, as well as other non-target bark- and wood-boring beetles.

From each tree with I. cembrae or Tetropium attacks, one 1.2 m log was cut for rearing of adult beetles and enemies. As few trees were colonized by I. cembrae no data from rearing of this species is presented. All collected logs were placed in insect nets and stored outside the laboratory in Uppsala during winter. On 30 and 31 March 2021, one 60 cm stem section was cut from each log and waxed at both ends to prevent desiccation. Each stem section was hung under a translucent roof outside, in a cotton sack, with a plastic collection funnel attached to a collection jar located at the bottom of the sack. The collection jars were emptied throughout the summer. The data collection was completed by the end of August when all the logs were debarked, and dead adults and larvae were recorded. The only Tetropium species emerging was T. gabrieli. Emerging parasitoids belonged to either Braconidae or Ichneumonidae family based on morphological characteristics (see Schroeder et al., 2020 for more details). Parasitism rates of T. gabrieli larvae was estimated for each stem section by dividing the number of emerging parasitoids by the sum of emerging T. gabrieli adults and parasitoids [we assume that each emerged parasitoid killed one larva, based on the life history of the species found to parasitize T. gabrieli previously (see Schroeder et al., 2020)].

All statistical analyses were performed using R ver. 4.3.3 (R Core Team, 2024). For the analyses, locations without any catches or lacking colonization by either species were excluded. These locations were assumed to be outside the distribution area of the species based on the results from this and previous studies (Ravn and Harding, 1995; Ravn, 2012; Lindelöw et al., 2015). Thus, for the analyses of trap catches, 36 locations remained for I. cembrae and 32 locations for T. gabrieli. For the analyses of tree colonization, 30 locations remained for I. cembrae and 26 for T. gabrieli.

For I. cembrae, no significant correlation was found between any of the explanatory variables and the size of the local population in the model including the 500 m buffer zone. However, a positive significant relationship was found between relative population density and the amount of larch within the 1,000 m buffer (Table 2). Region explained some of the variance in both the 500 and 1,000 m model (between 0.88 and 9.06%, see Table 2). In both models for T. gabrieli, the relative population density showed a significant positive relationship with the size of the focal stand (Table 2), but not with the quantity of larch in the surrounding landscape.

Ips cembrae colonized four felled larch trees, distributed among two of the 30 stands included in the analyses (Figure 2A). The two locations with colonized trees were those with the highest trap catches (Figure 2B) and, there was a significant difference in the mean (±SE) number of trapped beetles between the two locations with colonized trees and the 17 locations without colonized trees (287.5 ± 59.7 versus 35.4 ± 5.8, t-test, p < 0.001). No colonization occurred in stands with mean trap catches below 80 beetles. No standing trees were attacked by either of the species in the focal stands in the year of the experiment.

Tetropium gabrieli colonized 46 felled larch trees, distributed over 12 of the 26 stands included in the analyses (Figure 3A). The number of colonized trees varied from one to five per location. All five trees were colonized in 30% of the locations. Colonization occurred at mean catches as low as 0.5 beetles per stand, and took place in all stands with mean trap catches of two beetles or more (Figure 3B).

The number of felled trees colonized by T. gabrieli was significantly and positively related to the mean number of trapped individuals, and the mean number of trapped females and males (Table 3). We found a negative significant relationship between mean diameter of the cut trees and the number of colonized trees in the models including the mean number of trapped individuals and the mean number of trapped males.

A total of 676 T. gabrieli emerged from 42 stem sections out of the total 46 used in the rearing experiment. The mean (±SE) rate of emergence per stem section was 16.1 (±3.2) beetles. There was no significant difference between the number of emerged males and females (p = 0.16, chi-square test).

A total of 7684 Thanasimus beetles (7649 T. formicarius and 35 T. femoralis), were caught in the I. cembrae traps at the 36 locations. There was no significant relationship between the mean numbers of trapped I. cembrae and Thanasimus (Figure 4). The ratio between trapped Thanasimus/I. cembrae was 6.6 to 1. One location was removed from the model as it was considered an outlier; however it is included in the figure.

A total of 184 parasitoids emerged from the stem sections colonized by T. gabrieli. The average rate of parasitism was 30.1% (SE = 5.0%). There was no significant relationship between the proportion of parasitized T. gabrieli larvae and the density of larvae (Figure 5A) or the proportion of parasitized larvae and the mean number of trapped T. gabrieli (Figure 5B).

Three other bark- and wood-boring species emerged from the T. gabrieli colonized stem sections, Serropalpus barbatus (Schaller), (Coleoptera, Melandryidae), Urocerus gigas (L), (Hymenoptera, Siricidae) and Cryphalus sp. (Coleoptera, Scolytinae). A total of 121 S. barbatus adults emerged from 13 of the 47 stem sections, 28 U. gigas emerged from seven of the sections and two adults of Cryphalus sp. were found after debarking one stem section at the end of the rearing experiment.

Our results show that relative population density of both species was positively related to habitat availability (see also Schroeder, 2013; Kautz et al., 2016; Gohli et al., 2024). However, the species differed in the spatial scale of their response to habitat quantity: Tetropium gabrieli showed a positive response to the size of the focal stand, in both the 500 m and 1,000 m buffer models. Whereas I. cembrae showed a positive response to the quantity of larch at the 1,000 m buffer. The result for I. cembrae might be explained by the presence of local populations in larch stands within the 1,000 m radius resulting from the previous year. The positive relationship of T. gabrieli with the size of the focal stand was not only due to the local hibernating populations in two of the stands. Tetropium gabrieli still showed a positive response to the size of the focal stand even after removing the only two localities with hibernating populations. Research has shown that responses to habitat availability at different spatial scales is often species specific due to different life history traits (Holland et al., 2004).

Both species showed a positive relationship between the relative population density and the probability of colonization of the cut trees. At too low population sizes no colonization occurred. This indicates the presence of the component Allee effects for host finding (i.e., finding of suitable breeding material) and/or mate finding for both species. For the bark beetle I. cembrae, a much higher relative population density was required for colonization than for the longhorn beetle T. gabrieli. At the same time, the mean population size of I. cembrae in the colonized localities was significantly higher than in localities without colonization. The large difference in minimum relative population density (i.e., trap catch) required for colonization events strongly indicates that lower population density is required for T. gabrieli than for I. cembrae. Reproductive strategy could be an explanation for this difference. Tetropium gabrieli females can, after mating, oviposit in many substrates independent of males. Whereas, for I. cembrae, a male must first find the object and then successfully attract a female (i.e., the female cannot by herself colonize an object). In addition, each female colonizes potentially two objects. However, when comparing the trap catches of the two species, we need to be aware that it is hard to know if both species are equally attracted to the pheromone bait used. Regardless of the magnitude of the trap catch colonization of the felled logs is, however, a good indication of reproductive activity for both species.

Based on the trap catches, there was no relationship between the relative population density of I. cembrae and the predator Thanasimus spp. Thanasimus formicarius, the dominant species preys on over 20 species, and it is considered one of the most important predator for many bark beetle species, including I. cembrae (Schroeder, 2001; Kenis et al., 2004; Warzée and Grégoire, 2006). Being generalists, Thanasimus spp. densities most likely depend on the presence of other potential prey in the area. At low population densities of the prey generalist predators are thought to be important predators to maintain the population at low densities. In the present study, the average catch of Thanasimus was almost five times higher than that of I. cembrae. When present Thanasimus spp. adults feed on adult bark beetles and their larvae feed on bark beetle larvae under the bark. Thus, one Thanasimus beetle could kill more than one bark beetle and therefore, predation pressure at low density of I. cembrae could be high. In our study, this could mean that the predator may have prevented some colonization in localities with small populations of I. cembrae by being attracted to the felled trees by host volatiles and preying on I. cembrae males prior to boring into the bark. Following the start of pheromone production the predator may also prey on the I. cembrae females before they have managed to enter the mating chamber. Our findings indicate that the potential presence of a component Allee effect of predation pressure could influence successful colonization and establishment of I. cembrae negatively. Using trap catches enables us only to look at the presented relationship, other natural enemies, like parasitoids and woodpeckers, could be important predators as well. Unfortunately, we could not measure this because of the minimum colonization of our felled logs.

For T. gabrieli there was no significant relationship between neither the relative population density and the proportion of larvae parasitized, nor the number of larvae in the stem sections and the proportion of them that were parasitized. Hence, there does not seem to be a component Allee effect of enemy pressure on reproductive success after colonization. Although the rate of parasitism in our study was higher compared to Schroeder et al. (2020), it was still lower than the rate of parasitism reported from its native range (Kenis and Hilszczanski, 2004). In Schroeder et al. (2020) only three parasitoids species, classified as native, were found to parasitize the T. gabrieli larvae. All these species are also associated with the native T. fuscum and T. castaneum. The parasitoids of the present study were not determined to the species level but belonged to the same families as the ones in the study by Schroeder et al. (2020). The limited presence of other species under the bark suggests that T. gabrieli larvae faced no competition for food (phloem) from other species, as S. barbatus and U. gigas larvae feed in the wood instead.

To show a direct effect on population size (the demographic Allee effect) one often needs to investigate those factors that affect population size (component Allee effects) (Lewis and Kareiva, 1993; Berec et al., 2007). In this study, we investigated host location, mate finding and predation pressure as components of the demographic Allee effect. Our study species exhibit distinct ecological differences, and we find that enemy pressure, as well as the mechanisms for host and mate finding, are specific to each species. Based on the natural enemies included in our study, we suggest that component Allee effect involving enemy pressure is more important for I. cembrae than for T. gabrieli in the way it affects their success in population establishment and thus invasion of a new previously uncolonized area with fragmented habitat.

Species establishment and spread are processes that happen over time. Time series collected by the Swedish Forest Agency (Issacson unpublished data) show that trap catches of I. cembrae vary per year and it is general knowledge that based on weather conditions, insect populations can show large variation in densities between years. In that respect, our study is a snapshot in time. To circumvent replication over time, we designed the experiment at a large spatial scale including areas where the species were fully established (Denmark) and areas where evidence of establishment was more tentative (southern Sweden). Based on the geographical scale and the extensive spatial replication, we believe that our results represent those ecological processes influencing population establishment we are aiming to measure.