Discussion
We aimed to document the introduction of the Eucalyptus snout beetle in Ecuador, determine its taxonomic placement, and model potentially suitable habitat throughout South America. With this framework, we provide the first genetic evidence for the presence ofGonipterus platensis in Ecuador. Furthermore, our ecological niche model analysis suggests areas of suitable habitat throughout a broad range of climatic and elevational regimes that may have played a role in the establishment of populations of Eucalyptus snout beetles in South America and could promote its future expansion into new areas.
The phylogenetic analysis we showed herein highlighted large genetic variation across the Gonipterus scutellatus species complex. However, most of the samples analyzed in this study came from Australia and Tasmania, with few samples coming from Portugal, Spain and South Africa (Mapondera et al., 2012), and only the nine samples we collected coming from South America. DNA barcoding data (mitochondrial COI sequences) clearly resolved the identity of the beetles found in Ecuador as G. platensis , which has also been documented in other South American countries (i.e., Argentina, Brazil, Colombia, Chile, and Uruguay) (Garcia et al., 2019; Mapondera et al., 2012; Schröder et al., 2020). The genetic diversity of G. platensis was one of the lowest among the species analyzed in this study (see values of π for each clade in Fig. 2), which coincides with the results found by Mapondera et al. (2012) for this species in Western Australia. In fact, the Ecuadorian samples share the same haplotypes with individuals from Tasmania (Garcia et al., 2019), which suggests that the Ecuadorian population could have originated from a Tasmanian source. Notwithstanding, an analysis including genetic samples from Colombia and other South American countries is needed to thoroughly examine whether the Ecuadorian samples were a product of a direct introduction from the native range or if they were secondarily introduced from another South American country. An introduction from Colombia seems plausible given geographical proximity and the continuity of suitable climatic habitat predicted by our model (Fig. 3). Moreover, our phylogeny suggested that there are at least three different geographical origins of the G. platensis specimens found outside Australia. Indeed, two different sequences of this species from Spain and Portugal (JN391479_ESP and JN391480_POR) are the most different compared to the sequences of South America and Tasmania, as evidenced by their long branches in the phylogeny (see Gonipterus platensis clade in Fig. 2). Additional sampling and genetic analysis are necessary from Eucalyptus snout beetle populations from Southeastern Australia and from other South American countries to truly pinpoint the origin of invasive populations.
Two main challenges have been identified when producing ecological niche models (ENM) for invasive species (Lake et al., 2020). First, invasive species are often in disequilibrium with the novel environment they occupy, and second, generating ENM projections from native to invaded ranges may be problematic for correlative model approaches (Elith et al., 2010). We followed an ENM approach at the species level, including all unique localities of G. platensis in South America available in GBIF (GBIF.org, 2022; www.gbif.org; https://doi.org/10.15468/dl.8mwpht). This approach is justifiable because few records were available in open-source databases from the native range of Australia (N = 9). Such a low number of occurrences would bias ENM predictions, particularly when projected from native ranges into novel environments where the species was introduced. Furthermore, using only the native range to estimate the potential suitable habitat of a species may result in misrepresentation of predictions, particularly if the colonization into the invaded range is characterized by niche expansion (Broennimann et al., 2007). To mitigate the spatial data bias, we rarefied records and used a background extent that included only potentially reachable areas (following Elith et al., 2010). This choice of background would reduce the degree of model extrapolation (Elith et al., 2010). Despite the uneven sampling, our approach allowed us to produce estimates of the areas where G. platensis occurs, but also assess regions of suitable habitat where these beetles could potentially be found or invade (Fig. 3).
The ENM predictions presented herein attained a high true positive fraction for the currently known observations (Appendix, Fig. A1). Notably, the westernmost locality in the Buenos Aires province of Argentina showed the lowest predicted distribution values and was excluded by the LPT95 threshold. This inland locality of Argentina was characterized by a narrow range of high temperatures (i.e., little variation in temperature seasonality) and generally dry conditions (i.e., lowest precipitation of coldest quarter) that were unique compared to all other occurrences. Despite the lack of observation records from GBIF, model predictions showed that suitable habitat forG. platensis exists in southeastern Brazil in the city of Curitiba, PR, where the beetle was first documented as invasive for the country (Fig. 3; op. cit. G. scutellatus in Freitas, 1979). This region of suitability in Brazil spans from the state of Espírito Santo, south through Rio Grande do Sul, and west into Mato Grosso do Sul, in which G. platensis is widely recognized as a pest ofEucalyptus forests (de Souza, 2016; Wilcken & Oliveira, 2015).
In the northwestern parts of South America, ENM predicts suitable habitat across the Andes from Colombia to Bolivia at high elevations (Fig. 3). In Ecuador, G. platensis was predicted to have significantly higher elevations than elsewhere in the continent (Fig. 4; Table 1), with highly suitable habitat existing in and around the cities of Quito, Ambato, and Cuenca and with at least moderately suitable habitat occurring all along the Ecuadorian Andes (Fig. 3). Suitable habitat at high elevations was also predicted throughout Perú, which may represent potential areas for invasion (Fig. 3). In the southern countries, the model resulted in high to moderate values of suitability at lower elevations in Chile, Argentina, and Uruguay, respectively (Fig. 3). Examination of the climate envelopes revealed that G. platensis in Ecuador occupies significantly less seasonal and drier conditions than in other parts of the invaded range at similar elevations (Fig. 4; Table 1). Nonetheless, when considering the entire distribution, this beetle occurs across a wide elevational range and a variety of suitable environmental conditions where it could potentially sustain long term populations. The sampling bias, which likely excluded areas occupied by G. platensis but not included in GBIF or not yet documented across South America, renders our ENM predictions somewhat exploratory. Nonetheless, these results combined with climate envelope analysis highlights the potential of G. platensis to occupy a broad range of environmental conditions, which may allow it to become highly invasive in different parts of South America.
Finally, we make a call to study the ecological and economic impacts of the invasion of Eucalyptus snout beetle and its hosts (e.g.,Eucalyptus globulus ) in South America, particularly in Colombia and Ecuador. In northern South America little research has been conducted on forestry of Eucalyptus species, whereas in Brazil and Chile, forestry research is quite active. In the particular case of Ecuador, Eucalyptus trees were planted in the late 19th century, mainly in volcanic soils, known as ‘cangagua’, in the inter-Andean valleys (FAO, 1981). These valleys have had human influence for millennia (Bush et al., 2022; Young, 2009), so perhaps these degraded environments have lost natural enemies that could exploit these insects as new hosts. We thus recommend extensive sampling of Eucalyptussnout beetles and potential controllers in Ecuador to test this hypothesis. Also, as far as we know and according to records in GBIF,Gonipterus platensis remains concentrated in a couple of areas of the Ecuadorian Andes (around the city of Quito in the North and in the city of Loja in the South, Appendix, Fig. A1), but our model predicts the whole range as climatically suitable for invasion. In fact, climatic suitability expands across the border to the Peruvian Andes, within the area of distribution of Eucalyptus spp. plantations (Castillo Vera et al., 2019; Luzar, 2007), where no reports of the insects have been made so far. All this highlights the urgency of conducting more studies on the invasion of this pest to design effective control and/or prevention measures.