We obtained all AQIM records from October 2003 (when WPM was first inspected as part of AQIM) through 2020 where WPM was associated with the inspected consignments. This period included data from different phases of ISPM 15 as follows.

Therefore, the US regulations related to ISPM 15 were broadly consistent within each of these four phases. We analyzed the data in three separate ways: all cargo-related WPM records as a whole; separately for the three major survey programs (perishables, Italian tiles, and GWPM); and separately for three major US trading partners (China, Italy, and Mexico).

For those AQIM records with WPM where insects were detected, we first categorized each insect taxon as a likely bark- or wood-infesting insect or not. Some of the intercepted insects were agricultural pests and likely associated with the imported perishable cargo, while several others were considered hitchhikers that inadvertently contaminated the cargo or WPM. Classifying insects as wood borers was straightforward when they were identified to genus or species, based on life history descriptions available in the literature. Insects identified to only family or subfamily level were classified as likely wood borers if they were in the beetle (Coleoptera) families Buprestidae or Cerambycidae, or the weevil (Curculionidae) subfamilies Platypodinae or Scolytinae (bark and ambrosia beetles). Similarly, records of carpenterworms (Lepidoptera: Cossidae) and woodwasps (Hymenoptera: Siricidae) were classified as likely wood borers. Weevils (Curculionidae) in the subfamily Cossoninae were also classified as wood borers when the beetles were associated with WPM, because many Cossoninae species are true wood borers (Jordal et al., 2011). However, we did not classify the few records of wood-infesting insects identified as Bostrichidae (powderpost beetles) or Isoptera (termites) as borers because these insects can infest WPM after treatment and therefore cannot be used to judge the effectiveness of ISPM 15 (Haack et al., 2014). When more than one species of wood borer was found in the same consignment, all taxa were recorded, but the individual consignment was simply considered infested regardless of the number of distinct borer species found.

Prior to analysis, we excluded all Canadian records because the US did not require Canadian WPM to meet ISPM 15 standards during the sampling period, given that the vast majority of bark- and wood-infesting insects native to Canada also occur in the US (Bright, 1976, 1987, 2021; Bousquet et al., 2017). In addition, for the period 2006–2020, we only included those records where the WPM had the ISPM 15 mark, indicating that the WPM was apparently treated to ISPM 15 standards. Note that most of the WPM inspected since 2006 has been marked with the ISPM 15 mark (Figure 2). We will refer to this reduced dataset as the “final dataset” used in the analyses below.

We separately tested the borer detection rate in WPM for 2003–2004 against the rates in the 2005–2006, 2007–2009, and 2010–2020 periods. We constructed 2 × 2 contingency tables for each comparison and analyzed each using Fisher’s exact test (one-sided probability, PROC FREQ, SAS Institute, Cary, NC). We used a significance level of α = 0.1 and did not correct for multiple comparisons because infestation rates of WPM are usually low and we wanted to reduce the likelihood of committing a Type II error (i.e., a false negative). The above procedure was used to examine borer detection rates for all WPM records overall, for the three main survey programs (perishables, Italian tiles, and GWPM), and for the three major US trading partners (China, Italy, and Mexico). We also conducted chi-squared tests (PROC FREQ, SAS Institute), and calculated running averages for certain categories and inspected them for long-term trends.

The AQIM database contained 109,709 inspection records with WPM from 2003 to 2020. These records represented consignments from 172 countries, with the top 15 countries being Mexico (30.4% of all records), Canada (18.2%), Italy (12.7%), China (4.6%), Costa Rica (3.6%), Guatemala (2.7%), Netherlands (2.5%), Dominican Republic (2.4%), Brazil (2.1%), India (2.0%), Ecuador (1.6%), Honduras (1.5%), Turkey (1.4%), Spain (1.4%), and Germany (1.4%). The number of countries from which WPM originated increased during each period, and more than doubled from 2003–2004 to 2010–2020 (Table 1), although annual numbers of countries with wood borer detections have dropped since 2018 (Figure 3). Note that when data are expressed on an annual basis, 2003 only contains data for 3 months (October-December), given that AQIM first started recording WPM data in October 2003.

For the entire 2003–2020 dataset, 88.6% of the inspected WPM items were recorded as pallets, 7.5% as crating, 2.4% as dunnage, and 1.4% as other (e.g., wood blocks and spools; Figure 4). These four types of WPM were encountered in each of the three main survey programs during 2003–2020. For the GWPM program, the WPM types were 64.7% pallets, 23.1% crating, 8.2% dunnage, and 4.0% other. For perishables the WPM types were 98.5% pallets, 1.2% crating, 0.1% dunnage, and 0.2% other. For Italian tiles the WPM types were 92.0% pallets, 5.9% crating, 1.1% dunnage, and 1.0% other.

The final dataset used for analysis had 87,571 records after removal of all consignments from Canada, as well as those records with unmarked WPM during the years 2006–2020. As an aside, no borers were detected in any of the Canadian WPM records that were excluded.

Wood borers were detected in 180 of the 87,571 consignments with WPM (0.21%). Of these detection records, 67.2% were in pallets, 29.4% were in crating, 2.2% were in dunnage, and 1.1% were classified as other (Figure 4). When expressed as borer detection rates, wood pests were detected significantly more often on crating (0.64%) than pallets (0.12%), dunnage (0.15%), or other (0.13%) (χ² = 124.1, P < 0.0001). The 180 infested consignments originated from 30 countries, with 10 from Europe (including Russia), 8 from Asia (including Turkey), 6 from South America, and 5 from both Central America and Mexico (Table 1). The number of countries from which infested WPM originated generally increased over time (Table 1 and Figure 3). During the 2003–2020 period, borers were detected in WPM in eight or more years for only five countries (Figure 5), which also were the countries with the most borer detections: China (40), Italy (27), Mexico (22), Costa Rica (18), and Turkey (15) (Table 1). Consignments with infested WPM were detected in all months of the year, showing no strong seasonal interception pattern (Figure 6).

The AQIM data showed that WPM was associated with several hundred different types of commodities. The top five perishable imports with WPM were (decreasing order) broccoli, pineapples, tomatoes, bananas, and peppers. The top five non-perishable imports with WPM were (decreasing order) tiles, auto parts, stone slabs, machinery, and metal. The most common commodities with infested WPM followed these same trends (Table 2 and Supplementary Table 1).

The infested consignments contained seven families or subfamilies of wood borers (with all seven being referred to as families for simplicity): five families of Coleoptera (Buprestidae, Cerambycidae, Platypodinae, Scolytinae, and other Curculionidae), plus Siricidae and Cossidae (Tables 2, 3). Five of the 180 infested consignments had borers from two different families and one consignment had borers from three families, and therefore when organized by borer family they represented 187 borer-consignment combinations (Tables 2, 3). Cerambycidae were the most commonly intercepted borers overall (93 of 187, or 49.7%), with Scolytinae being the second most frequent (39.0%) (Table 3). The diversity of wood borers found in WPM increased over time, with the Cerambycidae and Scolytinae consistently being the two most common borer families intercepted during each of the four time periods (Table 3). The 3-year running average for the mean annual number of detections for the four most intercepted beetle families indicated that cerambycid interceptions were consistently high, scolytine interceptions were the next most frequent and surpassed the cerambycid detections in the 2010s for a few years, and both buprestid and platypodine interceptions remained low but peaked in the mid-2010s (Figure 7). Most of the wood borer families recorded in this study were found in crating and pallets (Table 3). The most common genera of wood borers of those identified to genus or species are listed in Table 2. For the Scolytinae that were identified to genus or species (45% of 73 interceptions), 100% were true bark beetles (which develop primarily under bark) during 2003–2004 and 2005–2006, but only 40% in 2007–2009, and 23% in 2010–2020, with the others being ambrosia beetles (which develop inside wood) (Supplementary Table 1).

The presence of bark on WPM was recorded in nearly all the original AQIM records from 2003 to 2020, although patch size of any residual bark was not recorded (N = 109,707). The percentage of WPM consignments with bark was 3.4% in 2003–2004, 2.6% in 2005–2006, 2.3% in 2007–2009, and 1.2% in 2010–2020. When compared to 2003–2004, each of the later three periods had significantly reduced percentages of WPM with bark (Fisher’s exact test P = 0.007 for 2005–2006; P < 0.0001 for both 2007–2009 and 2010–2020). Similarly, the 46% decrease in the percentage of WPM with bark from 2007–2009 to 2010–2020 was significant (Fisher’s exact test P < 0.0001).

Overall, bark was present in 22.5% of the 180 wood borer detections in WPM. By period, bark was present on 40% of detections in 2003–2004, 46% in 2005–2006, 39% in 2007–2009 records, and 15% in 2010–2020, demonstrating a significant decline for the last period (Fisher’s exact test P = 0.004). Unfortunately, AQIM data do not indicate if the detected borers were found under bark (if present) or in the wood, but simply that bark was present somewhere on the WPM.

The presence of the official ISPM 15 mark on WPM was first recorded in the AQIM database in October 2005. Overall, 72% of the WPM inspected during October-December 2005 was marked. The percentage of marked WPM entering the US increased to 89% in 2006 and 95–99% was marked during 2007–2020 (Figure 2).

Within the three AQIM survey programs, the overall detection rate of borer-infested WPM entering the US immediately before ISPM 15 implementation (2003–2004) was 0.34% (Table 4). Rates were significantly lower (P < 0.1) in two subsequent periods: 0.13% during 2005–2006 (61% reduction compared to 2003–2004, P = 0.017), and 0.18% during 2007–2009 (47% reduction, P = 0.054). The 2010–2020 rate of 0.22% (36% reduction) was also lower, and nearly significant (P = 0.102). The overall infestation rate of WPM at 2-year intervals also indicated a sharp drop during 2005–2006 and a temporary rise in infested WPM during the mid-2010s (compare the “Total” bars in Figure 8).

Wood borer detection rates were greatest for the GWPM program (Table 4 and Figure 9), and cerambycids were the most common borers associated with infested GWPM (66% of detections; Supplementary Table 1). We found no significant differences in detection rates by period, which ranged from 0.33% during 2005–2006 (P = 0.486) to 0.41% during 2007–2009 (P = 0.603) (Table 4). In fact, when viewed at 2-year intervals, borer detection rates of GWPM were rather consistent over time (Figures 8, 9). Of all inspections classified as GWPM, most originated from China (18.4%), with the next five highest countries of origin being (decreasing order) India, Brazil, Germany, Italy, and Turkey.

For the Italian tiles program, wood borer detection rates in WPM declined by almost two thirds from 2003–2004 (0.62%) to 2010–2020 (0.23%) and was near zero during much of 2005–2009 (Table 4 and Figures 8, 9). When compared to the 2003–2004 detection rate, these reductions were significant during the 2005–2006 (P = 0.020) and 2007–2009 (P = 0.018) periods, but not the 2010–2020 period (P = 0.127; Table 4). Wood borer detection rates on Italian tiles spiked from 2015 to 2018 for unknown reasons, but never exceeded 0.1% (Figure 8). As expected, 99.8% of all consignments recorded as Italian tiles originated from Italy. Most detections in this survey program were cerambycids (68%; Supplementary Table 1).

For the perishable goods program, wood borer detection rates declined overall from 0.20% in 2003–2004, to 0.06–0.07% during 2005–2006 and 2007–2009 (∼70% reduction), and to 0.12% during 2010–2020 (39% reduction, Table 4 and Figure 9). These reductions were only significant for the period of 2007–2009 (P = 0.077, Table 4). When viewed at 2-year intervals, detection rates in WPM associated with perishables decreased at first and then increased again during 2009–2018, peaking during 2013 (Figure 8). Overall, 66.9% of all consignments of perishables with WPM originated from Mexico. The next five countries with the highest number of perishable records were (decreasing order) Costa Rica, Guatemala, Netherlands, Ecuador, and the Dominican Republic. Scolytines were the most frequent borers detected in WPM associated with perishables (77%; Supplementary Table 1).

We found no significant reductions for wood borer detection rates in WPM from China during any post-ISPM 15 period (Table 5). The great majority of the detections for China appeared in the GWPM program (91.0%). For China, the WPM infestation rate was 1.26% during 2003–2004, and ranged from 0.58 to 1.11% during the next three periods (Table 5).

We did, however, find significant rate reductions for borer detections in WPM from Italy and Mexico (Table 5). For Italy, borer detection rates were significantly lower during each period compared to 2003–2004, although they lessened over time (Table 5). As expected, most wood borer detections in WPM from Italy were on Italian tiles (86.7%). Borer detection rates in WPM from Mexico continually declined over time, from a 64% reduction during 2005–2006, to an 85% reduction during 2010–2020 (Table 4). Nearly all borer detections from Mexico were on perishables (98.9%). Because detections of infested WPM for each of the above three countries were primarily associated with a single, distinct survey program, these three countries greatly influenced detection rates within the three survey programs listed in Table 4.

Costa Rica and Turkey ranked fourth and fifth overall for wood borer detections in WPM, with borer detection rates generally increasing over time for both countries. The borer detection rate for Costa Rica was 0.072% (1 of 1,382 consignments) during all of 2003–2009 but increased to 0.665% (17 of 2,555) during 2010–2020. Similarly, for Turkey, the borer detection rate was zero during 2003–2004 (0 of 59 consignments), 0.83% during 2005–2006 (1 of 120), 1.12% during 2007–2009 (4 of 356), and 1.05% during 2010–2020 (10 of 952).

The most common borers detected in WPM from the above five countries varied by country. Cerambycids were the most frequently intercepted wood borers from China (78%), Italy (67%), and Turkey (66%), and usually associated with imports of tiles, stone slabs, metal, and machinery (Supplementary Table 1). By contrast, scolytines were the most commonly intercepted borers on goods from Costa Rica (78%) and Mexico (91%) and were mostly associated with imports of perishables goods (Supplementary Table 1). Moreover, 88% of the scolytines identified to genus or species from Costa Rica were ambrosia beetles (7 of 8, with 4 others identified to Scolytinae only), whereas 100% were bark beetles in the case of Mexico (6 of 6, with 13 identified to Scolytinae only) (Supplementary Table 1).

The current Agriculture Quarantine Inspection Monitoring (AQIM) dataset consists of thousands of records that have been collected at US ports since the late 1990s, from consignments that were randomly selected within various cargo survey categories and uniformly inspected. These qualities give us confidence that AQIM data can be used to compare infestation rates of WPM in several major cargo pathways both before and after US implementation of ISPM 15 during 2005–2006.

We estimated that overall borer detection rates for WPM entering the US in these cargo pathways declined by 36–61% during and after implementation of ISPM 15 (Table 4). However, percentage rate reductions by period have lessened over time. In annual terms, detection rates have been relatively constant since 2005, except for a spike in rates associated with Italian tiles from 2015 to 2018 (Figure 8). That happened despite the modifications made to ISPM 15, such as the 2009 requirements for debarking, limiting the size of residual bark patches, requiring debarking prior to fumigation, and lengthening the fumigation period.

One feature of our analysis is that AQIM data only started in 2003. Bulman (1992) reported a 4.3% borer detection rate in WPM for randomly selected maritime containerized consignments that entered New Zealand during 1989–1991. Compared to that rate, the three later time periods listed in Table 4 reflect detection rate reductions of greater than 95%, which would all be highly significant (Fisher’s exact test P < 0.0001). Many countries were probably already improving the phytosanitary quality of their WPM in the 1990s and early 2000s. For example, beginning in 1992, the European Commission required that pine WPM from countries with pinewood nematode (Bursaphelenchus xylophilus) be either fumigated or heat treated before export to Europe (Allen et al., 2017). In 1998, the US required WPM from China to be heat treated, fumigated, or treated with a preservative prior to arrival in the US (US Department of Agriculture Animal and Plant Health Inspection Service [USDA APHIS], 1998). Then, the North American Plant Protection Organization [NAPPO] (2001) drafted and approved a regional standard entitled “Import Requirements for Wood Dunnage and Other Wood Packing Materials into a NAPPO Member Country” (North American Plant Protection Organization [NAPPO], 2001). This regional standard was never implemented because it soon served as the basis for ISPM 15, which was adopted in 2002 (Allen et al., 2017). In addition, several countries implemented ISPM 15 before the US (see Introduction). Thus, it is likely that the pre-ISPM 15 WPM infestation rate would have been significantly greater in the 1990s, if data were available, than the pre-ISPM 15 rate used in the present study for 2003–2004.

It is worth noting that few non-native wood borers have apparently become established in the United States in the last decade. For example, the Eurasian Quercus-infesting ambrosia beetle Xyleborus monographus was found in California beginning in 2017 (Rabaglia et al., 2020), and the Asian hardwood-infesting cerambycid Dere thoracica was found in Georgia starting in 2020 (Traylor et al., 2021). Xyleborus monographus has been intercepted seven times during 2009–2021 at US ports, originating from multiple countries and usually in association with WPM. Similarly, D. thoracica has been collected multiple times in WPM from China at US ports (Eyre and Haack, 2017; Traylor et al., 2021). Even though these two species were collected only recently, they likely became established many years earlier given the typical lag time between initial establishment and first detection (Brockerhoff and Liebhold, 2017). Another example is the discovery of the Anoplophora glabripennis infestation, an Asian cerambycid, in South Carolina in 2020, which was at least seven years old when discovered (Coyle et al., 2021). However, DNA analyses indicated that the South Carolina population matched the current A. glabripennis infestation in Ohio that was discovered in 2011, suggesting domestic spread of A. glabripennis possibly through movement of infested firewood (Coyle et al., 2021). Alternatively, both A. glabripennis infestations could have been two separate introductions that originated from the same source population in Asia. Given that annual volume of imports to the US has increased 68% from 2003 to 2020, while borer detection rates have remained rather steady, an increase in introductions and establishments might be expected. However, the apparent lower establishment rates for wood borers in recent years, compared to the late 1990s and early 2000s (Haack, 2006), suggests that ISPM 15 is helping to mitigate risks in the US and remains an important component of the global phytosanitary system.

Although we anticipated that borer detection rates in WPM would consistently decline with implementation of each ISPM 15 rule change, there are several possible explanations for why infestation rates have remained fairly stable since 2005. The six explanations reviewed in detail by Haack et al. (2014) remain relevant to the current study. Briefly, they were (1) wood borer tolerance to the approved treatments, especially heat treatment at 56°C for 30 min, (2) unintentional non-compliance due to factors like poorly calibrated equipment, (3) outright fraud where WPM is stamped with an ISPM 15 mark without treatment, (4) colonization of WPM by borers after proper treatment, (5) data issues with AQIM such as pest detections being recorded on a consignment basis, and (6) artificially low pre-ISPM 15 detection rates because some shippers started using treated WPM when exporting to the US in anticipation of its implementation in 2005. To this list, we add the possibilities that (7) borer detection has increased over time by inspectors, (8) changes in trading partners that may vary in ISPM 15 enforcement and implementation, as well as the diversity of local borers, (9) variability in the occurrence of borer outbreaks among trading partners, and (10) limits on the compliance levels that can be achieved under an international standard such as this one.

With respect to thermotolerance by wood borers, it is important to recognize that the primary heat treatment schedule adopted in ISPM 15 (56°C for 30 min throughout the profile of the wood including the core) was initially developed to kill pinewood nematode (Allen et al., 2017), with the recognition that some wood pests were able to tolerate those conditions (Haack et al., 2014). For example, the Asian buprestid beetle known as emerald ash borer, Agrilus planipennis, was the focus of many heat-treatment studies after its initial discovery in the US in 2002 (see discussion in Haack et al., 2014). Some authors reported various levels of A. planipennis survival after heat treatment, but none precisely tested 56°C for 30 min at the core of the wood, which is the ISPM 15 requirement (Allen et al., 2017; Food and Agriculture Organization [FAO], 2019). Based on studies by Myers et al. (2009), the heat treatment schedule for interstate movement of ash (Fraxinus) firewood within the US was set at 60°C for 60 min, but the data supporting this new schedule was not considered sufficient to change import regulations by the European Commission (European Food Safety Authority Panel on Plant Health [EFSA PLA], 2012).

More recently, MacQuarrie et al. (2020) and Haack and Petrice (2022) reported complete mortality of A. planipennis when heat-treating infested wood to the current standard. However, Haack and Petrice (2022) demonstrated that some borers did survive in both hardwood and conifer bolts when treated at 56°C for 30 min. Moreover, they showed that borer mortality increased as the air temperature in the heating chamber was increased from 60° to 75°C (testing at 60°, 65°, 70°, and 75°C). Given these recently published findings, and that ISPM 15 does not stipulate a minimum chamber temperature during heating cycles, a proportion of the live borers found in heat-treated WPM by the AQIM program from 2005 to 2020 likely did survive heat treatment. Studies are currently underway to test in vitro lethal temperatures of various borers using a precision water bath apparatus that should help reduce the confounding effects of the thermal properties of wood on heat treatment testing of borers (Noseworthy et al., 2022).

Unintentional non-compliance can occur when managers attempt but fail to apply the appropriate heat treatment or fumigant dose to all WPM during a treatment cycle (Haack et al., 2014). Factors such as poor equipment calibration, cold pockets, and improper sensor placement or stacking of WPM in kilns, among others, can all lead to uneven treatment, with some WPM pieces not reaching the minimum dose. Therefore, the occurrence of some live borers could reflect such situations. Many recommendations on how WPM is to be handled and tested during heat treatment and fumigation sessions have been published by the IPPC (Food and Agriculture Organization [FAO], 2017). Additionally, fraud can occur when non-treated WPM is stamped with the ISPM 15 mark and used to export cargo (Pallet Enterprise, 2015; Eyre et al., 2018; Papyrakis and Tasciotti, 2019). The incidence of this type of fraud at a global scale is unknown, given that each trading partner’s National Plant Protection Organization is responsible for their own respective audits and ISPM 15 compliance. Unfortunately, no effective diagnostic tools exist that can independently test WPM for treatment compliance, although methods have been investigated (Iline et al., 2014; Kim et al., 2019). While over 98% of the WPM entering the US in the past decade has been stamped as ISPM 15 compliant (Figure 2), it is also true that all live insects evaluated in this paper were within stamped and apparently compliant WPM.

Post-treatment colonization of WPM can occur, especially when large patches of bark are retained on WPM (Evans, 2007; Haack and Petrice, 2009). A few borers can infest bark-free wood, including some ambrosia beetles and cerambycids, but most require bark to be present during colonization, oviposition, and early larval development (Duffy, 1953; Haack and Petrice, 2009; Haack et al., 2017). This fact was the rationale for requiring WPM to be made from debarked wood in 2009, which significantly (Fisher’s exact test P < 0.0001) decreased the percentage of marked WPM entering the US with bark from 2.3% in 2007–2009 (481 of 20,635 records) to 1.2% in 2010–2020 (905 of 71,887). The requirement for debarking WPM is likely the reason why fewer true bark beetles (that develop under bark) have been detected in WPM in recent years compared with ambrosia beetles (that develop in wood) (see discussion above and Supplementary Table 1). Nevertheless, it is possible that a few of the wood borers found in WPM in the present study, such as some ambrosia beetles, colonized the WPM after treatment.

With respect to AQIM data collection protocols, detections of wood borers are reported for an entire consignment, giving taxonomic information for the wood pests found but not the total number of individuals found, nor the number of infested pieces of WPM. Consequently, borer infestation evaluated at the per-WPM-piece level could have fallen dramatically over time, but analyses at the consignment level cannot uncover such trends. For example, Schortemeyer et al. (2011) presented worst-case infestation levels of Agrilus planipennis for a typical pallet where every piece of wood used to construct the pallet was considered infested at typical field levels. They calculated that such a pallet could contain as many as 38 A. planipennis larvae. If double stacked, a 40-foot-long shipping container with 44 pallets infested at worst-case levels could yield 1,672 A. planipennis. Alternatively, if a consignment had just one A. planipennis infested pallet at the worst-case level, and further if that pallet’s ISPM 15 treatment had killed 95% of the borers present, then there would be only about two A. planipennis individuals present per container. In each situation, the two containers would be recorded as infested in AQIM, but the worst-case founder population of 1,672 A. planipennis individuals clearly poses a much greater risk of establishment than the ISPM 15-mitigated risk founder population of 2 individuals, especially when considering Allee effects (Liebhold and Tobin, 2008; Brockerhoff et al., 2014; Ormsby, 2022). Moreover, pest establishment is less likely if the entering imported cargo goes to disparate geographic locations, because surviving insects need to find both suitable host plants and potential mates (if sexually reproducing).

All the above analyses assumed that the detection abilities of port inspectors were uniform over time. However, Agricultural Specialists for the US Customs and Border Protection might be detecting wood borers more reliably than before because of improvements in training, inspection techniques, or technology. If so, the apparent impact of ISPM 15 would be lessened in recent years. Estimating the effectiveness of inspections is not currently possible, which typically requires a “leakage survey” conducted on inspected goods to see how often target pests are missed (Robinson et al., 2009).

Changes in trading partners could also influence the apparent effectiveness of ISPM 15, especially if countries vary widely in the quality of enforcement and implementation of ISPM 15 (Papyrakis and Tasciotti, 2019). For example, based on recent borer detection trends (Table 5) the phytosanitary quality of WPM appears to have improved over time in Mexico, stayed rather steady in China, and declined in Costa Rica and Turkey (see discussion above). Infestation rates on WPM from Italy have generally improved over time as well (Table 5), except for a recent rate spike (see above). In addition, given that species of trees and their associated borers often vary by country and world region, changes in trading partners can affect both the type of wood used to construct WPM and the composition of associated borers.

Even when trading partners do not change, local outbreaks of both native and non-native wood borers can influence which tree species are used to construct WPM at the origin, and their potential pest load. For example, borer outbreaks often follow periods of environmental stress such as drought, severe and repeated defoliation, and widespread wind damage (Haack and Petrice, 2019; Pureswaran et al., 2022). Wood from such trees could be heavily infested, with high pest loads prior to any ISPM 15 treatment. If the treatments are not 100% effective, then live borers would likely be more commonly encountered in WPM during outbreak periods. In addition, planting non-native tree species can lead to outbreaks of native borers. For example, in China, outbreaks of the native buprestid Agrilus planipennis occurred in areas planted to North American ash (Fraxinus) (Dang et al., 2022). Similarly, outbreaks in China of the native cerambycid A. glabripennis occurred in areas planted to non-native poplars (Populus) (Haack et al., 2010). In both cases, widespread ash and poplar mortality occurred, especially in the 1990s. Some of these dead and dying trees were likely used to construct WPM, which then could have resulted in these two Asian borers being introduced to the US and other countries (Haack et al., 2010; Dang et al., 2022).

Still, the influence of changes to ISPM 15 treatment requirements alone may not always lead to reduced wood borer infestation rates. This is because in an open international commercial system such as this, program integrity depends on voluntary compliance (Fletcher et al., 2017; Yoe et al., 2020; Meyerson et al., 2022). Reducing infestation levels to near zero principally via application of the ISPM 15 standard would require nearly universal compliance by industry, using highly effective treatments. However, that may be difficult to achieve without either very strong incentives (Rossiter and Hester, 2017) or intensive oversight and significant penalties for non-compliant exporters (Hood et al., 2019). Improved ISPM 15 education and outreach by IPPC, along with greater cooperation and information sharing amongst importing and exporting countries, could create global feedback that facilitates a very high proportion of WPM in the system becoming compliant.

The diversity of borers reported in WPM in the present study increased from two major borer groups in 2003–2004 to seven in 2010–2020 (Table 1). In part, this outcome reflects the longer 2010–2020 sampling period, along with many more countries being sampled in 2010–2020 (159 vs 75) and more countries being the source of infested WPM in 2010–2020 than in 2003–2004 (25 vs. 5). Moreover, as the number of trading partners increases for any country it is also likely that the number of tree species used to construct WPM will increase (Krishnankutty et al., 2020), along with borer diversity, thus increasing the species pool of potential invasive wood borers. Nevertheless, although borer diversity increased over time in the present study (Table 3), all the major families of borers intercepted during the 2010–2020 period have been collected in WPM at US ports in earlier decades (Haack, 2001, 2006) as well as in other countries (Bain, 1977).