Changes in propagule pressure from single or multiple regions directly contribute to the success or failure of nonnative species establishment6,8,23,24. In this study, we collected and measured the quantity and diversity of seeds, over time obtained from the air-intake grilles of refrigerated containers, with two seasons for comparison. Targeting the trans-oceanic transport of a single commodity in this industrial trade system that serves as a transport vector for hitchhiking seeds provided reduced variation in which to quantify propagule pressure, including propagule size (Fig. 1) and propagule number (Supplementary Fig. 2) of plant species considered to be of high-risk to agriculture in the USA19. Our key finding is that influx is sufficient and reproduction of these species is high enough to represent a risk of population(s) establishment in and around the shipping port, even with the bottlenecks of escape from the shipping container, subsequent germination, and seedling survival (Figs. 2, 3).
Over 20,000 shipping containers are moved as import or export daily on the GCT 14, providing ample volumes for passive hitchhikers to establish at the GCT and surrounding areas. In fact, we found steady arrival of shipping containers over the approximately 32-week shipping season (Supplementary Fig. 2). Conversely, we found strong seasonal variation in propagule number (i.e., number of seeds per refrigerated shipping container; Fig. 1). We estimated for the FNW, S. spontaneum, that over 40,000 seeds entered GCT during the two shipping seasons (Table 1). This level of propagule pressure is clearly sufficient to represent introduction and establishment risk of a clonal, perennial, fecund species that likely does not require a large initial propagule size, even if the escape rate from the shipping containers is exceedingly small (Table 1; Fig. 2). In this study, the four focal monocotyledonous taxa all had similar seed sizes, and no other larger-sized propagative material of these species (e.g., rhizomatous material or cuttings) were encountered during our study at the GCT.
The theoretical literature postulates that increased numbers of propagules (i.e., propagule size6,8,9,10,11) and pressure (which includes propagule size and frequency as a rate) increases the likelihood of nascent population establishment and population size and diversity6,10; however, among our four focal taxa, a nascent population may establish from a single seed during arrival at a suitable terrestrial substrate, such as the GCT’s greenspaces19. Persistence of an extremely small population can, and is likely to, be facilitated by asexual propagation and spatial spread of these particular plant taxa. Theoretical population biology intrinsically includes propagule pressure within the invasion process6,10, and empirical studies measuring propagule pressure have demonstrated its importance as the most important and generalizable predictor of nonnative invasion success24. Propagule pressure in itself is also the factor most influenced by human activity8,9. Therefore, our study adds additional support to the importance of propagule pressure (see Figs. 2 and 3), and in this system, there is sufficient propagule pressure (i.e., influx from Fig. 3) for invasion success, even if escape rates from shipping containers, germination, and survival are low.
Though S. spontaneum is the only FNW we encountered, we were also able to identify Arundo donax L., a species that is listed as noxious by 46 of the USA’s 50 states25. We were not able to identify seeds to a taxonomic level sufficient to determine origin status for the other 28 taxa encountered (Table S4), but for our three additional focal species, we suggest that two species are native and one is likely introduced. We found Typha domingensis (Pers.) Steud. (native), Andropogon glomeratus (Walter) Britton, Sterns & Poggenb. and A. virginicus L. (both native), and Phragmites australis (Cav.) Steud. (nonnative), already established on-port at the GCT in a previous study that demonstrated that the Port of Savannah is a hub of nonnative species richness19.
For any of the species collected on the shipping containers, the propagules have the potential of being picked up en route to the GCT or, with the exception of S. spontaneum since it is not established there, at the GCT. Most of the taxa already have cosmopolitan distributions, and actual escape rates from the shipping containers are not yet known, meaning that the seeds could make multiple journeys on cargo ships across oceans before being released from a container. Also, the seasonality of seed dispersal coincides with dispersal time in the northern hemisphere, which may apply to seed sources in Panama, the Caribbean, or the USA. Andropogon glomeratus and A. virginicus occur throughout North and Central America (including Panama and the Caribbean)26. As native species to the southeastern USA, propagules escaping refrigerated shipping containers are not of significant concern, although they could be homogenizing genetic composition if genotypes from other portions of the parental established ranges are introduced here. Additionally, Andropogon propagules may result in introductions of nonnative species to South America if the seeds remain on the containers and are viable for return trips. Typha domingensis has nearly a global distribution, and though it is native to the southeastern USA, its presence at the Port of Savannah could also indicate the presence of admixed genotypes. Moreover, our morphological identification of the seeds could not distinguish to the species level, and T. angustifolia L., a nonnative species, could have been represented in our samples, though this species is well established and widely distributed already in the USA26. Phragmites australis, a noxious weed in 6 USA states25, is already found worldwide26. The genus Phragmites contains 4 species, of which only Phragmites australis is native to portions of North America; however, intra- and inter-specific hybridization among genotypes has resulted in the influx of nonnative lineages from Europe and Asia, which have spread to areas of the continent where it is not native27,28.
The most interesting case is S. spontaneum, the FNW. This species is established only in Florida on the USA mainland, where it was introduced for historical and extant breeding programs with sugarcane29. This recent report29 showed that it was naturalized in only three counties, but we have documented it growing in six counties and in cultivation in one additional county (Supplementary Fig. 4). We did not find it growing at the GCT at the Port of Savannah. Yet, it is known that S. spontaneum, which is native to the Indian subcontinent, is well established in the Panama Canal region29,30 along the shipping route of interest. The number of propagules we intercepted and estimated, along with nontrivial germination rates and high survivorship of seedlings, indicate that this species represents a real threat of establishment outside of Florida. Combined with other modeled estimates that S. spontaneum can establish throughout the majority of the USA29, we suggest that this species represents a significant risk of negative invasive species impact, earning its FNW listing in the 1980s18,29.
All four of our focal taxa share common life history features that have been suggested to be characteristic of invasive plant species: asexual reproduction through rhizomes, persistence in a wide range of environmental conditions, prolific seed production (Table 1 and citations within), wind pollination, and wind dispersal31,32,33. These traits have the potential to enhance geographic spread into new ranges and rapidly lead to single-species domination of local plant communities. All of these taxa have a life history and ecology similar to the very successful southeastern USA invasive species cogongrass (Imperata cylindrica (L.) P. Beauv.) that has been demonstrated to benefit from intraspecific heterosis and multiple introductions34,35,36.
A previous study used molecular barcoding of seedlings germinated from seed collected from Season 1 in this study, and they identified some seedlings as: S. spontaneum, Typha sp(p)., Phragmites sp(p)., and Andropogon sp(p).37, as identified here. Seeds that were grouped as S. spontaneum in this study resulted in seedlings that returned haplotypes for the genus Phragmites (rbcL haplotype 1 and matK haplotypes 3 and 437) and Saccharum along with other genera37. There are two interesting and opposing forces at play here. First, in sorting seeds morphologically, there is the potential to group similar looking seeds of different species. The molecular barcode result that shows Phragmites haplotypes in seeds morphologically identified as S. spontaneum is evidence of misidentification and inaccurate sorting of seed. Second, some haplotypes showed equally correct molecular identification across multiple genera of grasses, indicating that these standard molecular barcode sequences for plants may not have the species-level resolution necessary for molecular identification of some of the highest threat invasive grass species.
There are two key approaches to mitigating the risk that propagules of nonnative taxa will become established: 1) prevent propagules from hitchhiking on transoceanic cargo ships, in this case, becoming attached to shipping containers at their point-of-origin or stops along the way (that result from “trans-shipping”), and 2) prevent viable propagules from entering and establishing in the USA, via inspection and interception by the “gatekeepers” of biosecurity at international points-of-entry. These agricultural inspectors are tasked with the interception of propagules of insects, fungi, and all other nonnative or “actionable” taxa, in addition to the seeds of plants. One potential solution to reduce invasion risk by vascular plant seed is to employ a scaled-up version of the research approach we implemented here of backpack vacuuming air-intake grilles of refrigerated shipping containers. Another possibility in lieu of labour-intensive vacuuming of intake grilles is to conduct research on efficacy of liquid pre-emergent herbicide application to the air-intake grilles. For either approach, our data support that these interventions may not be needed year-round for important species like S. spontaneum, which have a clear import seasonality on this particular commodity. For example, based on our data, seed removal measures may only be needed in October, November, and early-mid December.
In the face of poorly resourced capacity for inspection and the potential of diminishing fiscal resources and human capital, consequences include acceleration of biodiversity loss, economic and environmental impacts, and on-going biotic homogenization. The interception efforts to prevent the entry of nonnative propagules of all nonnative taxa worldwide will ultimately conserve local endemism, biodiversity, economic output, and ecosystem services that are interrupted or extirpated by biological invasions1,3. This research aimed to identify key risks and highlights the need for improved strategies for efficacious prevention and interception of nonnative, particularly plant, propagules prior to establishment, though such prevention approaches can be designed and applied for many taxa. Enhancing the capacity, speed, and frequency of successful prevention programs will be required to minimize or eliminate the real risks posed by viable hitchhiking propagules associated with economic trade and sea/air transportation of commodities and people.