We compiled world-wide records of insect species recorded on Pinus radiata that incorporated the original lists of Ohmart (1980, 1981, 1982a, b) and additional records from New Zealand and other countries which had been continuously added to and curated by John Bain (Scion (New Zealand Forest Research Institute)). Beginning in 2018, this list was thoroughly reviewed and updated with a literature search using the Scopus database (see below) as well as forward and backward searches in relevant publications. In 2020, a list independently compiled by Helen Nahrung (University of the Sunshine Coast, Queensland, Australia) with records from Australia was incorporated. Additional records were added between 2020 and 2022 by further interrogating the literature, Scion’s Forest Health Database and other available databases. The main criterion for inclusion in our pine pest list was that species feed on any tissue of P. radiata (see below for more information on the impact classification). The full species list with references is available as Suppl. material 1: table S1 and at the online repository Zenodo (Brockerhoff et al. 2023).

Using the scientific name of each insect as the search term, the current taxonomy, synonyms and distribution in native and introduced ranges were retrieved for all species, initially by systematic searches using Scopus (https://www.scopus.com), Google Scholar (https://scholar.google.com), the CABI Invasive Species Compendium (https://www.cabi.org/ISC), the Global Biodiversity Information Facility (GBIF, https://www.gbif.org), NZOR – New Zealand Organisms Register (https://www.nzor.org.nz), the Atlas of Living Australia (https://www.ala.org.au), the Australian Faunal Directory (https://biodiversity.org.au/afd/home), and the Australian Plant Pest Database (https://www.appd.net.au) as well as Google (https://www.google.com) and Wikipedia (https://en.wikipedia.org). Other databases (some taxon-specific) and literature records were accessed as required, and in some cases, experts were consulted directly (see references in Suppl. material 1: table S1). Establishment data were cross-checked against the ‘International non-native insect establishment data’ database (Turner et al. 2021b).

The species list was standardised taxonomically using the GBIF taxonomic database (GBIF Secretariat 2021) and the “taxize” package in R (Chamberlain and Szöcs 2013). For any names not recognised by GBIF, standardisation was performed manually via searches of other databases and literature. Coleoptera family names were based on the framework in Bouchard et al. (2011), and Lepidoptera families as per Mally et al. (2022).

For each species, native and non-native occurrences were grouped by biogeographic regions defined as shown below. Our biogeographic regions are mostly aligned with those of Udvardy (1975) but not strictly because our information sources were specific to countries of occurrence, whereas the borders of Udvardy’s biogeographic realms often pass through countries (i.e., one country can be in more than one region).

Our regions are defined as follows:

• Western Palearctic (“W Palearctic”): Europe, North Africa and Near East;
• Eastern Palearctic (“E Palearctic”): Northern and eastern Asia and including the Indo-Malayan region;
• South West Pacific (“SW Pacific”): Australasia and Pacific Islands (excluding Hawaii);
• Afrotropic: Sub-Saharan Africa;
• Nearctic: North America including all of Mexico and Hawaii;
• Neotropic: South and Central America (excluding all of Mexico) and the Caribbean.

Using the information on occurrences of native species and establishments of non-native species, we compiled for each biogeographic region (i) the number of native species feeding on P. radiata, (ii) the number of established non-native species feeding on P. radiata, and (iii) the number of species originating from each region that became established in another region or in another country in the same region.

Three datasets with border interceptions were analysed to determine which of the species on our list have been intercepted during border inspections of imports, vessels and containers, and in some cases international mail and passenger baggage. Post-border interceptions were not considered.

Unless otherwise stated, analyses with border interception data were conducted using an international dataset. This recent dataset is a collection of international border interceptions between 1995 and 2021 in New Zealand, Australia, South Africa, South Korea, Japan, Canada, the United States, the United Kingdom and the European and Mediterranean (EPPO) region. The international interception dataset is comprised of the border interceptions described in Turner et al. (2021a). In addition, we queried South African border interceptions from Saccaggi et al. (2021), additional border interceptions from Japan between 1996–2019 extracted from http://www.pps.go.jp/TokeiWWW/Pages/report/index.xhtml (Plant Protection Station, The Ministry of Agriculture, Forestry and Fisheries of Japan), and updated EPPO border interceptions for the 2011–2021 period from the Europhyt annual interception reports. Included in the international dataset was the New Zealand data subset which spans the period from 2000–2017 (Turner et al. 2021a) which was used for a country-specific analysis.

Additional statistics were drawn from two older border interception databases. Firstly, the Scion BUGS database for New Zealand 1950–2000 which contains border interceptions of species relevant for trees, and secondly the USDA 1949–2008 interceptions of Scolytinae and Cerambycidae (Brockerhoff et al. 2014).

Each species on the list was assigned one of three impact ratings relating to evidence of pest status on P. radiata: ‘negligible impact’ – species where no interventions, management or damage records were found; ‘low-medium impact’ – species with evidence of damage, management or control but this was either short-term, localised or minor; and ‘high impact’ – species that required ongoing management and/or had significant economic effects, such as severe damage to forest or amenity trees and/or are important vectors of highly damaging pathogens of P. radiata. Species causing severe impacts on human or veterinary health (e.g. from urticating hairs of caterpillars) were also considered ‘high impact’. In some cases, we combined species in the low-medium and high impact categories as species of ‘non-negligible impact’. Impacts related to market access were excluded in our study because these are often associated with species that do not damage live trees or cause no damage at all. Likewise, impacts of species whose recorded damage was exclusive to timber in service, such as borers in dry deadwood, were excluded because the focus of our assessment was on insects feeding on living trees. Consequently, species exclusively affecting market access or causing only damage to timber in service were classified as having negligible impact.

Our impact classification differs from the now widely used EICAT classification (IUCN 2020) because our impacts relate mainly to damage to Pinus radiata planted for commercial purposes outside their native range and in some cases also to trees in their native range, whereas EICAT focuses only on “impact to native taxa” (IUCN 2020, p. 8). However, our categories can be translated to approximately corresponding EICAT categories (‘negligible impact’ = ‘minimal concern’; ‘low-medium’ = ‘minor’; ‘high impact’ = ‘moderate’). None of the insects considered in our list have a ‘major’ or ‘massive’ impact according to EICAT as both these involve at least local extinction of the affected species.

The final dataset containing all insect species feeding on Pinus radiata was analysed and visualised in R version 4.1.2 (2022-05-20). When analysing by biogeographic region, we excluded seven cosmopolitan species with a widespread distribution across multiple biogeographic regions where it could not be determined which regions were part of the native or non-native range. When analysing non-native species, we included species which were successfully eradicated as these represent the establishment potential in the absence of a post-border biosecurity response. For example, four of the species invasive to New Zealand fell into this category (Coptotermes acinaciformis, Coptotermes frenchi, Cryptotermes brevis and Teia anartoides).

Comparisons were made among all insects on the pine pest list (i.e., any species feeding on Pinus radiata) as well as specifically among the “non-negligible” impact species (those in the combined low-medium or high impact categories).

To investigate relationships between border interceptions and establishments, the number of species was compared by taxon groups which were defined at the level of insect orders with the exception of four particularly species-rich and important families/subfamilies (Cerambycidae, Scolytinae, Geometridae and Tortricidae) which were analysed separately. If relationships between interceptions and establishments were independent of taxon group, we would expect the number of established species in each group to be relative to the number of intercepted species in each group and proportional to the ratio of established insect species per intercepted insect species (i.e. expected number of establishments in taxa group = (total number of established insects)/(total number of intercepted insects)*(number of intercepted insects in taxa group). We assume that the number of established species per group can then be described by a Poisson distribution and calculate a prediction interval for each of our taxa groups. The prediction interval bounds were calculated to show the region within which all 11 taxa groups would be expected to fall 95% of the time. When calculating the interval quantiles, a Bonferroni correction was used for multiple comparisons.

The relationship between the number of native and non-native insects per biogeographic region was visualised on a scatter plot. The effect of feeding guilds (i.e., borers, defoliators, sap-feeders and others) was visualised by adding ellipses showing the 95% confidence intervals for a multivariate t-distribution (Fox and Weisberg 2011).

Pearson’s chi-square test was used to test for evidence of differences in proportions between groups (i.e., negligible vs non-negligible, intercepted vs not intercepted, feeding guilds), followed by pairwise comparisons of proportions using the Holm (1979) method of adjustment for multiple comparisons. In situations where expected counts were fewer than 5, Fisher’s exact test was used instead.

We found records of 649 insect species (in 438 genera, 83 families and nine orders) feeding on P. radiata (Table 1, Suppl. material 1: table S1). An additional 11 records were named at the genus level only; these were all of negligible or low impact, and as their identity could not be confirmed, they were excluded from the analyses (but are listed in Suppl. material 1: table S1). Coleoptera is the most represented order (299 species or nearly 50% of all species), followed by Lepidoptera (224 species), Hemiptera (65 species), Blattodea (i.e., termites), Hymenoptera and other orders. Twenty-eight species were categorised as ‘high impact’ and 168 species as ‘low-medium impact’ (Table 1). The remaining 453 species (nearly 70% of the species total) were considered to have negligible impacts on P. radiata as no records of damage were found for these species (Table 1). Of the 49 insects on our list that are known to vector diseases, evidence of detrimental impact exists for 37 species. In terms of feeding guilds, most species are either borers or defoliators while sap-feeders and other guilds such as root feeders and cone insects are less represented (Table 2).

The 28 species classified as high-impact comprised 17 Coleoptera (12 of which are true bark beetles (Scolytinae)), eight Lepidoptera, two Hymenoptera and one Hemiptera (Tables 1, 3). Twenty of these high-impact species are borers, seven are defoliators, and one is a sap-feeder (Table 3), with significant differences in proportions between groups (Table 2, Fisher’s Exact Test, P=0.016). Seventeen of the 28 high-impact species are known vectors of serious pathogens affecting P. radiata, especially the pitch canker fungus Fusarium circinatum (Table 3). Other species are high-impact pests in their own right such as the European six-toothed bark beetle Ips sexdentatus which can occasionally cause substantial tree mortality.

Seven cosmopolitan species which occur in multiple regions and for which the native range could not be determined were excluded from the analysis of native or invaded ranges except for the specific analysis for New Zealand and Australia (see below). Most native species feeding on P. radiata were recorded in the SW Pacific region (42% of all non-cosmopolitan species, with 167 species being native to Australia and 107 species native to New Zealand), followed by the Nearctic (20%), the Afrotropic (16%), the W Palearctic (12%) and the Neotropic region (12%) (Fig. 1A). The fewest native species feeding on P. radiata were recorded in the E Palearctic (6%). Despite the large number of species recorded for the SW Pacific, this region has just one native high-impact species (the Australian psychid moth Hyalarcta huebneri (Table 3)). The three southern hemisphere regions have the highest proportions of species with negligible impact and an average proportion of low-medium impact species (Fig. 1A). The W Palearctic has a high proportion and the largest number of high-impact species recorded on P. radiata (15 species: 8 Coleoptera, 5 Lepidoptera and 2 Hymenoptera), followed by the Nearctic region (11 species: mainly Coleoptera) and the E Palearctic region (8 species: 3 Coleoptera, 3 Lepidoptera and 2 Hymenoptera). However, there is considerable overlap in the native regions of these species. For example, eight high-impact species native to W Palearctic are also native to E Palearctic. The Neotropic has two high-impact species, the bark beetle Ips mexicanus in the northern part of this region (in the native range of pines) and Ormiscodes cinnamomea, a polyphagous saturniid in Chile.

Figure 1.

Impact levels of insect species feeding on Pinus radiata and their biogeographic ranges, excluding cosmopolitan species. (A) Species native to each biogeographic region. (B) Species non-native to each biogeographic region. (C) Species native to a biogeographic region (x-axis) which have established somewhere outside their native range (could be in the same biogeographic region e.g. from Australia to New Zealand). Note that the East Palearctic includes records from the Indo-Malayan region.

Twelve of the 28 high-impact species have already become established somewhere in the world, and six of these have become established in more than one biogeographic region (Table 3). The biogeographic regions with the most invasions of high-impact species (6) are the Nearctic or Neotropic (Fig. 1B); all but one of these species are native to Europe (and adjacent parts of the W Palearctic region), with the remainder being a native species from the Nearctic which invaded the Neotropic. Other regions with several establishments of high-impact species are the SW Pacific (four species, two native to the W Palearctic and two native to the Nearctic), and the Afrotropic (four species with three of these being native to Europe), while the Western and the E Palearctic had only one established high-impact species each (Fig. 1B).

The SW Pacific is the region with the most native species that became established somewhere outside their native range (both beyond and especially in other countries within their native biogeographic region), followed by the W Palearctic and the E Palearctic (Fig. 1C). However, the W Palearctic contributed by far the most high-impact species that became established somewhere, followed by the E Palearctic and the Nearctic. Although the E Palearctic ranks second in terms of high-impact species that established somewhere outside their native region, these are all species with a native range that extends from Europe across northern Asia, and it is difficult to ascertain the actual part of the region from which the invasion occurred.

Considering the source regions and invaded regions together, a clear picture of invasion routes emerges (Fig. 3). The W Palearctic is the main source region of invaders that colonised mainly the Nearctic, the Neotropic and the SW Pacific regions for all species (Fig. 3A) and species of non-negligible impact (Fig. 3B). Furthermore, the SW Pacific region has by far the most species that invaded other parts of the same region (Figs 3A, B, 4). However, these concern only species of negligible or low-medium impact as there are no high-impact species native to this region which established anywhere.

Figure 3.

Global movement of all insects feeding on Pinus radiata (A), and those with non-negligible impact (B). The thickness of each arrow is relative to the number of species native to the source biogeographic region established in the destination biogeographic region. Some species had native ranges spanning multiple biogeographic ranges, and in general it is not known if regions were used as bridgeheads, so the arrows represent all possible movements. Note that the East Palearctic includes records from the Indo-Malayan region.

Figure 4.

The number of species feeding on Pinus radiata that are native to each region and established (or not) outside their native range for non-negligible impact. Cosmopolitan species are excluded. Note, many of the native species from the SW Pacific are native to Australia but established in New Zealand – this is an example of a “Within” region establishment. Also note that some species are native to more than one biogeographic range, e.g., Palearctic species native to Europe and Asia, but this is not shown here. Note that the East Palearctic includes records from the Indo-Malayan region.

Of all the species in the pine pest list, 185 (29%) were intercepted during border import inspections at least once internationally between 1995–2021 (Table 1). Of these, 83 species (13% of the pine pest list) were intercepted specifically at New Zealand’s border between 2000–2017. An additional eight species were intercepted earlier (i.e., between 1950 and 2000 and recorded in New Zealand’s BUGS database), and a further two species were intercepted and recorded in United States interception records from 1949–2008. Therefore, a total of 195 species were intercepted at least once at a border. More than 60% of species of Scolytinae (bark and ambrosia beetles) on the list were intercepted (at least once, Table 1), a significantly greater percentage than other beetle groups and several other taxa (Fig. 2B). In terms of feeding guilds, the percentage of intercepted species was greatest for sap-feeders and borers, and least for defoliators, while differences from ‘other’ guild members were not significant (Table 2). The most frequently intercepted species were mainly sap-feeders (including Thrips tabaci, Thysanoptera: Thripidae, 42,302 interceptions; Aonidiella aurantii, Hemiptera: Diaspididae, 8,782 interceptions; Pseudococcus longispinus, Hemiptera: Pseudococcidae, 3,341 interceptions; and several other Hemiptera), as well as the defoliators Helicoverpa armigera and Helicoverpa punctigera (Lepidoptera: Noctuidae, 8,668 and 1,793 interceptions, respectively) and a borer, the bark beetle Hylurgus ligniperda (Coleoptera: Curculionidae, 1,766 interceptions) (see Suppl. material 1: table S1 for a complete list of interceptions). The proportion of species that were intercepted was significantly higher among the non-negligible species than for negligible species (One sided, 2-sample test for equality of proportions without continuity correction on log-transformed data, Chi-squared = 8.210, p-value=0.002).

Among the species feeding on P. radiata, the number of intercepted species in a taxonomic group was strongly positively correlated with the number of established species in a family (Pearson’s correlation coefficient, using log-transformed data: 0.92, P<0.001, Fig. 5). Of the 185 intercepted insect species (considering the international interception dataset from 1995–2021), 104 (56%) have already established somewhere (including cosmopolitan species), and 71% of the 146 species which have already invaded somewhere were intercepted (Fig. 6). Conversely, only 9% of the 464 species that were not intercepted have already invaded somewhere (Fig. 6). This indicates that species that are often intercepted also have a considerably higher likelihood of becoming established. Taxa with particularly high percentages of interceptions include the Scolytinae, Hemiptera and ‘other’ orders (mainly Orthoptera and Thysanoptera) (Table 1, Fig. 2B). Scolytinae and Hemiptera also have a high percentage of species that became established (Table 1, Fig. 2A).

Figure 5.

Number of species in the complete all-species pine pest list (649 species total) per taxonomic group that were intercepted and/or established, shown on a log-log scale. The black line represents where the taxa would fall on average if the number of established species was proportional to the number of intercepted species. The dashed lines show the prediction interval within which the taxa groups are expected to fall if establishments occurred at proportionally similar rates to interceptions, based on a Poisson model, alpha = 0.05, with Bonferroni correction accounting for 11 comparisons between taxa.

Figure 6.

Mosaic plot of the number (and percentages) of species according to their intercepted, establishment, or impact status. Established species are those established in a region outside their native range and are inclusive of cosmopolitan species and species that were subsequently eradicated. Interceptions are based on the international interceptions dataset covering the period 1995–2021. Species with negligible impact on Pinus radiata in light grey, those with non-negligible impact (i.e., low-medium and high impact) in dark grey.

Of the 28 high-impact species, 15 (54%) have been intercepted internationally (Table 3), and of the 168 species of low-medium impact, 56 (33%) have been intercepted (Suppl. material 1: table S1).

Of the 196 species with non-negligible impact, 19 species have been intercepted internationally more than 100 times (in decreasing order: Thrips tabaci, Helicoverpa armigera, Helicoverpa punctigera, Hylurgus ligniperda, Hylastes ater, Arhopalus ferus, Lymantria dispar, Ips sexdentatus, Heliothrips haemorrhoidalis, Epiphyas postvittana, Gnathotrichus sulcatus, Dendroctonus valens, Bradysia impatiens, Gnathotrichus retusus, Agrotis infusa, Nysius vinitor, Orthotomicus erosus, Arhopalus rusticus and Leptoglossus occidentalis) (Suppl. material 1: table S1). All but four of these 19 species have become established outside their native range (i.e., only Ips sexdentatus, Gnathotrichus sulcatus, Gnathotrichus retusus and Nysius vinitor have not yet invaded anywhere, to our knowledge).

A significantly greater percentage (36%) of species with non-negligible impact were intercepted than species with negligible impact (25%¸P=0.002, see details above; Fig. 6), but the difference in impacts between species that had become established (or not) was marginally non-significant (P=0.053, see details above; Fig. 6, Suppl. material 3: table S3). For some taxa, differences were observed for both parameters. For example, among Cerambycidae with non-negligible impact, a higher percentage has been intercepted (78%) (compared with only 29% of all Cerambycidae feeding on Pinus radiata), and a higher percentage (44%) have established outside their native range (compared with 20% of all Cerambycidae). For Scolytinae feeding on Pinus radiata, nearly two thirds (64%) were intercepted, 51% were of non-negligible impact, and 36% are already established (Table 1), significantly higher proportions than for Cerambycidae (Chi-squared tests with Yates’ continuity correction, impacts: Chi-squared = 19.19, df = 1, P<0.001; interceptions: Chi-squared = 13.52, df = 1, P< 0.001). This suggests that Scolytinae feeding on Pinus radiata are more likely to be intercepted, become established and have negative impacts than Cerambycidae. By contrast, only nine (22%) of the 40 Geometridae on the list have non-negligible impacts (and none fall into the high impact class), and few have been intercepted (5%) or become established (8%) (Table 1).

With a total of 649 insect species, our compilation of world-wide records of insects feeding on Pinus radiata represents a considerable increase over the last such comprehensive effort by Ohmart about 40 years ago (Ohmart 1980, 1981, 1982a, b). Although many of the most damaging insects of P. radiata were recognised then, several new threats have emerged. For example, the spongy moth Lymantria dispar was known as an occasional defoliator of P. radiata but it was considered “of little consequence” (Ohmart 1980). However, major outbreaks of L. dispar causing considerable defoliation have been reported recently from Spain (Castedo-Dorado et al. 2016) and we now classify this defoliator as a high-impact species. Another species that has only been recognised in this century as a potentially serious pest of P. radiata is the nun moth, Lymantria monacha (Withers and Keena 2001). Although other pine species have long been known to suffer sometimes severe defoliation by L. monacha in Europe, it was established through laboratory feeding trials that P. radiata is a highly suitable host for this defoliator (Withers and Keena 2001). Insects acting as vectors of the pitch canker disease, caused by the fungus Fusarium circinatum, are also of particular concern. The severe impacts of this disease on P. radiata have been known for some time (Wingfield et al. 2008b), and this is one of the main reasons why P. radiata is considered by the IUCN to be ‘endangered’ in its native range in California (Farjon 2013). However, the important role of insects such as the cone beetle Conophthorus radiatae as critical vectors in the transmission of the pathogen has only been appreciated in the last 25 years (Hoover et al. 1996; Brockerhoff et al. 2016). This is the main reason why insects capable of acting as vectors of F. circinatum are listed by us as high impact.

Altogether, we rated 28 insect species as high impact. Most of these species are native to the Palearctic or Nearctic where pines are native, while only three species originate from parts of the southern hemisphere where P. radiata and other pines are planted as non-native species. This is consistent with observations on insects feeding on northern hemisphere plants in southern hemisphere regions such as New Zealand and Australia; these insects originate mainly from the northern hemisphere where their host plants or close relatives are native while comparatively few insects native to the southern hemisphere have colonised these plants which have few or no relatives in the native southern hemisphere flora (Brockerhoff et al. 2010; Harvey et al. 2012). Likewise, few native insects in Europe damage non-native trees without close relatives (i.e., no congeneric species) in the European flora while those with close relatives are colonised by a larger suite of native plant-feeding insects (Branco et al. 2015; Padovani et al. 2020). This applies particularly to insects with a higher degree of host specificity but less so to polyphagous species.

More than half of the high-impact species are from the W Palearctic where they are normally found on European pine species. This means there are more high-impact species that have jumped from other pines to P. radiata (with which they have not co-evolved) than high-impact species with long associations with P. radiata in its native range. Such new associations between plant-feeding insects and new host plants often cause more severe damage than on their natural hosts. This is well illustrated by the pine processionary moth, Thaumetopoea pityocampa, which is considerably more damaging on P. radiata planted in Europe than on native European pines (Cobos-Suarez and Ruiz-Urrestarazu 1990), probably because P. radiata has not had the opportunity to evolve adaptations against this defoliator, in contrast to southern European pines which have co-evolved with T. pityocampa. The European six-toothed bark beetle Ips sexdentatus is mainly known as a secondary pest with relatively minor impacts (such as vectoring and facilitating blue-stain fungus infections) but during outbreaks in its native range, it can attack and kill live trees, albeit mainly those that are already weakened by other factors (Cobos-Suarez and Ruiz-Urrestarazu 1990). Other Palearctic species causing high impacts are more problematic in southern hemisphere regions where P. radiata has been planted than on P. radiata or other pines in Europe. Most notable among these are the Sirex woodwasp, Sirex noctilio (Slippers et al. 2015), and the European pine shoot moth, Rhyacionia buoliana (Alvarez and Ramirez 1989) which probably benefited from a combination of release from natural enemies (Mitchell and Power 2003; Colautti et al. 2004; Lombardero et al. 2008) and a highly susceptible tree species which has not co-evolved with these insects. High susceptibility in such cases may occur as novel host trees tend to have more limited resistance against non-native insects that are naturally associated with closely related trees, especially when the novel host has no experience with a congeneric native insect (Mech et al. 2019).

Even among the high-impact species native to the Nearctic, several are new associations where P. radiata represents a novel host. This includes, most notably, Ips grandicollis, the eastern five-spined engraver or five-spined bark beetle, which is native to eastern North America, with its range not sympatric with the natural distribution of P. radiata. Ips grandicollis invaded Australia where it can be highly damaging in P. radiata plantations and sometimes causes tree mortality by itself or in combination with attack by Sirex noctilio (Neumann 1987). Another species in this category is Pissodes nemorensis, an eastern North American weevil that can damage small trees and also acts as a vector of the pitch canker fungus, both of which have invaded South Africa (Gebeyehu and Wingfield 2003; Brockerhoff et al. 2016).

It is important to note that many of the high-impact species cause more substantial damage on P. radiata outside their native range. This applies, for example, to Sirex noctilio, Ips grandicollis, Essigella californica and Rhyacionia buoliana. These species probably benefit from freedom of natural enemies compared with the situation in their native regions (Mitchell and Power 2003; Colautti et al. 2004). An equivalent situation may occur in regions where the insect is native but the tree is non-native and probably colonised less by natural enemies as in the case of T. pityocampa in P. radiata plantations in Europe. In addition, the simplified monoculture environment typical especially of southern hemisphere plantation forests probably has a lower abundance and diversity of natural enemies than more diverse forests which tend to be more common in the native region of P. radiata and other pines (Stemmelen et al. 2022).

When considering all insects (not only those with high impact), the SW Pacific (i.e., in Australia or New Zealand) was the region with the greatest number of native species feeding on P. radiata (42% of all non-cosmopolitan species). This is rather surprising as there are no native pines or other Pinaceae in that region, and consequently, one would not expect a large number of species feeding on P. radiata. There are indeed a few native SW Pacific species that have caused noticeable damage in P. radiata plantations such as Pseudocoremia suavis (Lepidoptera: Geometridae) during outbreaks in New Zealand in the 1950s and 60s (White 1974). However, no outbreaks of this species have been recorded for nearly 50 years, and it is now relatively rare (Berndt et al. 2004), suggesting that these outbreaks were unusual occurrences. Consequently, we have rated P. suavis as low-medium impact. The majority of SW Pacific species (about 80%) have no or negligible impacts and the remainder are almost entirely in the low-medium impact category. The reason for the large number of records of species of little relevance is the existence of rigorous forest health surveillance systems in Australia and New Zealand where trees in plantation forests, urban areas, plant nurseries and high-risk sites near ports, airports and transitional facilities (where imports arrive and are cleared by biosecurity officials) are inspected regularly and any insects found are submitted for diagnostic identification (Bulman 2008; Carnegie and Nahrung 2019). These surveillance programmes are designed to detect incursions of non-native insect pests and pathogens as well as damage from known pests and pathogens but they also yield records of native species found on P. radiata even though most of these are not damaging. In other regions where P. radiata occurs as a native or non-native species, such non-damaging species are not recorded and published to the same extent and publications focus more on species causing more severe damage. Otherwise, the large number of species native to the Nearctic is consistent with this being the region where P. radiata and many other pines are native and as a result, there is a large fauna of insects feeding on pines. By contrast, the small number of species native to the E Palearctic may seem somewhat surprising given that there are many native pines and other Pinaceae in that region. However, P. radiata is not planted on a large scale in that region, and we are only aware of experimental plantings in China on an area covering hundreds of hectares (Bi et al. 2003, 2008). As the number of species colonising non-native trees is positively correlated with the area planted (Branco et al. 2015), it is plausible that there are comparatively few records of insects feeding on P. radiata from China and the E Palearctic (Bi et al. 2008). In addition, our list is probably not entirely complete because sources in languages other than English, especially in the grey literature, may have been missed. This potential bias may have affected especially our records from the E Palearctic and Neotropic with a higher proportion of non-English literature. Furthermore, some regions are under-studied regarding biological invasions, especially in the E Palearctic and Afrotropic regions (Pyšek et al. 2008).

With 146 established non-native insects feeding on P. radiata, 22% of all species in our database have already successfully invaded other regions. This large number of invasions is likely to be related to the substantial international trade in pine logs, timber, wood packaging material and propagation material used for the establishment of P. radiata plantations in non-native regions. International trade in logs, timber and goods shipped with wood packaging materials such as pallets are important pathways facilitating invasions especially of bark beetles, longhorn beetles and other wood borers (Brockerhoff et al. 2006, 2014; Meurisse et al. 2019; Vilardo et al. 2022), the groups most represented among established non-native species. Trade in live plants used for propagation is another important invasion pathway which is particularly relevant for sap-feeders in the order Hemiptera and defoliating and other Lepidoptera (Liebhold et al. 2012; Meurisse et al. 2019), the second- and third most numerous groups of non-native species feeding on P. radiata.

Nearly half of the 28 high-impact species we identified already occur somewhere as established non-native species. However, only six are established in more than one non-native region, indicating a large potential for additional invasions. Also, there are differences between regions in the number of established species. For example, there are only four established high-impact species in the SW Pacific while the remaining 86% are not yet present, which suggests there is considerable benefit in continuing and enhancing biosecurity measures aimed at preventing the arrival and establishment of these species (Sequeira and Griffin 2014; Ormsby and Brenton-Rule 2017).

Nearly a third of the species on our list (29%) were intercepted at least once in the countries for which we could access border interception data. For bark beetles, the percentage of intercepted species was even higher and exceeded 60%. Fifteen of the 28 high-impact species were intercepted, in some cases hundreds of times (e.g., Ips sexdentatus, Lymantria dispar and Hylastes ater). This highlights that pathways exist by which many of these species are transported with international trade and that there is a high potential for additional invasions to occur. Positive relationships between the number of interceptions of species and the probability of invasions have been documented, especially for groups such as bark beetles and longhorn beetles which are often well-identified and are less affected by insufficient identification or omission in interception data (e.g., Brockerhoff et al. 2014; Turner et al. 2020; Nahrung and Carnegie 2021). Our analyses specific to insects feeding on P. radiata were consistent with these trends as we found a positive correlation between the number of intercepted species within a taxonomic group and established species in that group (Nahrung and Carnegie 2021). Although some key pathways, such as the use of wood packaging materials, have been mitigated with some effect (Haack et al. 2014), the sheer volume of international trade means that some risk of introduction remains.