Introduction

The rate of detection of non-native species in aquatic ecosystems worldwide has increased substantially over recent decades, with seaweeds (marine macroalgae) accounting for over 14% of detected introductions (Bailey et al. 2020). Although the number of known non-native seaweed species has increased in recent years (Williams and Smith 2007; Thomsen et al. 2016a; Costello et al. 2021), this probably represents an underestimate, owing to undetected introductions and a lack of taxonomic information (Bailey et al. 2020). Non-native seaweeds have potentially pervasive impacts on marine biodiversity and ecosystem functioning, typically exerting negative effects on the abundance and diversity of native biota (Williams and Smith 2007; Thomsen et al. 2016a), yet their effects are context dependent, with the potential to be exacerbated by climate change and other environmental stressors (Bax et al. 2002; Thomsen et al. 2016b; Bennett et al. 2021). Similarly, the “invasion success” of a particular non-native seaweed is difficult to predict, depending on the attributes of both the non-native species and the invaded ecosystem and, perhaps less importantly, the lack of coevolution with native species (Thomsen et al. 2016a).

In order to detect and monitor the arrival, distribution and impact of non-native seaweeds effectively, it is essential that species are reliably identified. Certain groups of seaweeds are notoriously difficult to distinguish based on morphology alone and require molecular-assisted taxonomy for accurate identification (Saunders and Kucera 2010; Tran et al. 2022). DNA-barcoding studies of morphologically indistinct and phenotypically variable taxa, such as foliose green seaweeds of the cosmopolitan genus Ulva, have revealed cryptic diversity and confirmed the presence of non-native species in many parts of the world (e.g. Heesch et al. 2009; Krupnik et al. 2018; Wei et al. 2022). In addition, the family Ulvaceae, to which this genus belongs, includes a disproportionately high number of known non-native species, and is regarded as a group with a great proclivity for invasion (Williams and Smith 2007). The use of a robust taxonomic approach for enhancing our knowledge of seaweed biodiversity is particularly important in remote areas where invasive species pose a significant threat to marine ecosystems comprising a relatively high proportion of rare and/or endemic species (Clubbe et al. 2020; Dawson et al. 2022), especially given the greater tendency for non-native seaweeds to exert negative effects on native biota in relatively pristine environments (Thomsen et al. 2016a).

The remote sub-Antarctic island of South Georgia supports a rich seaweed flora (John et al. 1994; Wells et al. 2011; Mrowicki and Brodie, pers. obs.), although its shallow subtidal ecosystems remain poorly studied (Barnes et al. 2006; Rogers et al. 2015), particularly with regard to seaweeds (Clubbe et al. 2020). Existing seaweed species lists for South Georgia are based on morphological identification and outdated species concepts, making it difficult to make biogeographical comparisons; however, given the dominance of endemic and range-edge species in South Georgia’s benthic invertebrate fauna (Hogg et al. 2011), its seaweed flora may be expected to have similar characteristics. The unique inshore marine biodiversity of South Georgia faces threats predominantly from climate change and invasive species (Hogg et al. 2011; Rogers et al. 2015), including seaweeds, which could be introduced to the island via international shipping (Dawson et al. 2022) involving tourist, research and fishing vessels. Many of these vessels arrive from the Falkland Islands, located approximately 1450 km west of South Georgia, which, therefore, represent a potential source of marine non-natives. While it is estimated that non-native species account for 70% of the island’s terrestrial flora (Clubbe et al. 2020), there have been no confirmed reports of non-native seaweeds in South Georgia, despite a number of species having been identified as high risk in terms of their likely arrival, establishment and ecological impact (Dawson et al. 2022). Here, we provide the first record of a non-native seaweed in South Georgia, Ulva fenestrata, and evidence of its establishment in the Falkland Islands, supported by molecular-assisted taxonomy.

Methods

Specimen collection

Seaweed specimens were collected from the shallow sublittoral zone at Grytviken, South Georgia (14th November 2021) and from the intertidal zone at four sites in the Falkland Islands (December 2013–February 2018; Figs. 1 and 2; Online Resource 1). Tissue samples for DNA extraction (~ 0.1 cm2 with no visible epiphytes) were either dissected from fresh material and preserved in silica gel in the field (South Georgia) or removed from dried herbarium specimens (Falkland Islands). All material is housed in the algal herbarium at the Natural History Museum, London, UK (BM).

Fig. 1
figure 1

a The locations of the Falkland Islands and South Georgia in the southwest Atlantic Ocean (dashed lines represent the respective Exclusive Economic Zone boundaries), and collecting sites for Ulva fenestrata in (b) the Falkland Islands and c South Georgia

Fig. 2
figure 2

Ulva fenestrata specimens collected during this study. Clockwise from left: specimen SG-21-347 from South Georgia; specimens 021213-01iv, FLK503, FLK851 and FLK1058 from the Falkland Islands. Scalebars represent 5 cm. Specimen collection details are given in Online Resource 1

DNA sequencing

DNA extraction, amplification and sequencing methods are described in Brodie et al. (2021). Two plastid markers were sequenced: for all specimens, (1) an 855 bp region of the translation elongation factor Tu (tufA), using primers tufGF4 (Saunders and Kucera 2010) and tufAR (Famà et al. 2002); and, for the South Georgia specimen only, (2) the 742 bp 5′ end of the rubisco large subunit (rbcL-3P), using primers GrbcLFi (Saunders and Kucera 2010) and 1385R (Manhart 1994). PCR thermal profiles for tufA and rbcL-3P followed Saunders and Kucera (2010). Sanger sequencing was performed at the Natural History Museum using a 3730xl DNA Analyzer (Applied Biosystems, USA).

Data analyses

Forward and reverse sequences were aligned, edited and trimmed using Unipro UGENE (v40.0; Okonechnikov et al. 2012). Additional sequences were obtained from GenBank and added to tufA and rbcL-3P datasets, including three outgroup sequences per dataset (Online Resource 2). Resulting tufA (n = 24) and rbcL-3P (n = 26) were aligned separately using MAFFT (v7.505; Katoh and Standley 2013) via the L-INS-I algorithm. Prior to phylogenetic analyses, regions of poor alignment and sites containing gaps were removed from each alignment using Gblocks (v0.91b; Castresana 2000), resulting in 674 and 621 sites for tufA and rbcL-3P, respectively. Analyses were performed on both single-marker alignments separately, in addition to a concatenated alignment (tufA + rbcL-3P) including 18 sequences (total 1295 sites). Best-fit partitioning schemes and corresponding nucleotide substitution models, based on the corrected Akaike Information Criterion (AICc), were determined for all three alignments using PartitionFinder (v2.1.1; Lanfear et al. 2016; Online Resource 3).

Phylogenetic relationships were inferred based on Bayesian and maximum likelihood optimality criteria, using MrBayes (v3.2.6; Ronquist et al. 2012) and RAxML-NG (v1.1.0; Kozlov et al. 2019), respectively. Bayesian analysis was performed using two runs of four Markov chains over 5 million generations, sampling every 100 generations, with the first 25% of trees discarded as “burnin”. Convergence was diagnosed using the average standard deviation of split frequencies (ASDSF; ≪ 0.01) in addition to potential scale reduction factors (PSRF; ≈1) and estimated sample sizes (ESS; > 200) for all parameters, and by examining parameter log-likelihood traces using Tracer (v1.7.1; Rambaut et al. 2018). Maximum likelihood analysis involved 100 tree searches using 50 random and 50 parsimony-based starting trees, followed by non-parametric bootstrapping with 5000 replicates to generate Felsenstein (FBP) branch support values for the best scoring tree, with convergence assessed post hoc using the autoMRE criterion.

Results

Molecular phylogenetic analyses confirmed the presence of U. fenestrata in South Georgia (tufA and rbcL-3P; Fig. 3, Online Resources 4, 5) and the Falkland Islands (tufA only).

Fig. 3
figure 3

Bayesian phylogenetic tree inferred for tufA sequence data. Node values indicate maximum likelihood bootstrap support values/Bayesian posterior probabilities (‘–’ denotes < 50% support). Clades with low overall support are collapsed for easier interpretation. The Ulva fenestrata clade is highlighted in grey, with samples from South Georgia and the Falkland Islands emphasised in bold. Scalebar represents number of substitutions per site

The tufA single-marker analysis produced two subclades within the U. fenestrata clade, one containing samples from the North Pacific, including the holotype from Kamchatka in Eastern Russia, and the other containing samples from South Georgia, the Falkland Islands and the North Atlantic, although these subclades were not well supported (Fig. 3). While tufA sequences were identical within each subclade, they differed by a single base pair (0.0015%) between subclades. For rbcL-3P, South Georgia U. fenestrata was resolved within a well-supported clade including the holotype and samples from Japan, New Zealand and both sides of the North Atlantic (Online Resource 4). While the South Georgia specimen was identical to the holotype in terms of rbcL-3P sequence, the maximum pairwise difference within the clade was 2 bp (0.0032%). The concatenated tufA + rbcL-3P analysis supported the single-marker results and did not provide any further resolution (Online Resource 5). While U. fenestrata appears to be widespread in the Falkland Islands, distributed from the southwest coast of West Falkland to the east coast of East Falkland (Fig. 1), the species was only collected from Grytviken Jetty in King Edward Cove on the north coast of South Georgia (Online Resource 1).

Discussion

This study provides the first evidence of a non-native seaweed, U. fenestrata, in South Georgia and confirms its establishment in the Falkland Islands, where it is now distributed widely. There is only one previous record of a non-native marine species in South Georgia, the mussel Mytilus edulis, a live individual of which was found at King Edward Point in 1974 (Ralph et al. 1976). Although this species appears not to have become established, it is still considered a high-risk potential invader (Dawson et al. 2022).

The genus Ulva is a good example with which to illustrate the problems in establishing whether a species is non-native and how long it has been in a particular area. High phenotypic variability combined with a lack of distinctive morphological characters means that molecular data are necessary to confirm species identities (Maggs et al. 2007; Saunders and Kucera 2010; Tran et al. 2022). As a result, the misapplication of species names is widespread within Ulva, and this has led to a misunderstanding of species distributions. Ulva fenestrata is a case in point. For about 250 years, the name U. lactuca was generally used for cold temperate Ulva species until Hughey et al. (2019) sequenced the holotype of U. lactuca, demonstrating that it was almost identical to the epitype of U. fasciata Delile, a species found in warm temperate to tropical seas. Having considered that the U. lactuca holotype was from the Indo-Pacific, the authors applied the earlier name U. fenestrata to the cold temperate Northern Hemisphere species (Hughey et al. 2019). Establishing the species concept of U. fenestrata demonstrates the importance of reliable taxonomy based on robust scientific and historical evidence for compiling accurate species lists and distinguishing taxa.

The length of time U. fenestrata has been present in South Georgia and the Falkland Islands is unknown. Based on the collecting date of specimen 021213-01iv in the current study, the species has been in the Falkland Islands since at least 2013. Other historical specimens in the Natural History Museum (BM) algal herbarium may represent this species based on morphological examination but require molecular confirmation. This includes 14 specimens identified as U. lactuca collected between 1842 and 2003 from a range of localities in West and East Falkland, raising the possibility that U. fenestrata has been present in the Falkland Islands for a long time. In the case of South Georgia, for which even fewer historical data exist, there are two specimens identified as U. latissima collected from Cumberland Bay and King Edward Cove in 1913. In a list of South Georgia seaweed species compiled from published records and BM herbarium specimens (John et al. 1994), the only species of Ulva listed was U. lactuca Linnaeus var. macrogyna [sic] Reinsch with the caveat “Status of this species requires confirmation”. Ulva lactuca var. macrogonya Reinsch, which presumably is what John et al. (1994) referred to, is not an Ulva species. Reinsch (1890), who based his description on material collected from South Georgia in 1883, described this species as comprising a single layer of cells, whereas Ulva species are composed of two cell layers (Maggs et al. 2007). More recently, Wells et al. (2011) collected U.lactuca” from King Edward Cove in 2010, which is a more likely candidate for U. fenestrata, awaiting identification via molecular analysis. Another specimen named U. lactuca was collected from Husvik Harbour during the same expedition, but this has since been identified as a species belonging to the genus Protomonostroma (Mrowicki, unpublished).

Until now, the only records of U. fenestrata in the Southern Hemisphere were from the South Island of New Zealand, relating to specimens collected from three disparate localities in 2003 and 2004 (Heesch et al. 2007, 2009), which were initially identified as U. lactuca but later confirmed as U. fenestrata (Hughey et al. 2019), now recognised as an introduced species in this region (Nelson et al. 2021). All three specimens were collected from artificial structures in harbours (Heesch et al. 2007), indicating human-mediated dispersal between sites. In South Georgia, the only confirmed (i.e. this study) and potential records of U. fenestrata are from King Edward Cove, the main port of entry, where the main settlement of King Edward Point and former whaling station of Grytviken are located, and which is frequented by tourist, research and fishing vessels of international origin. Further, two recent expeditions have surveyed and collected seaweeds extensively along the coast of South Georgia (Wells et al. 2011; Mrowicki and Brodie, unpublished) and found no evidence of U. fenestrata outside of this locality. This suggests that U. fenestrata was introduced to South Georgia via shipping, despite not necessarily during recent years, given that the island was a centre for sealing and whaling activity from the late 18th to mid-twentieth centuries (Hoffman et al. 2011; Calderan et al. 2020). It is also possible that this species arrived by ‘hitchhiking’ on drifting anthropogenic debris (Barnes 2002) or on seaweed rafts (Avila et al. 2020), which are known to travel thousands of kilometres around the Southern Ocean and act as vectors for species dispersal between distant landmasses, including South Georgia and New Zealand (Fraser et al. 2022). Determining the timeframe and pathway of introduction of U. fenestrata in South Georgia requires further investigation involving additional field sampling of remote areas in the Southern Hemisphere and detailed molecular analyses of historical and recent herbarium specimens. In particular, population genetics and genomic data may be used to track the geographical origin of U. fenestrata and even predict its invasion success (Estoup and Guillemaud 2010; Matheson and McGaughran 2022), which would shed more light on its population status in South Georgia and other regions of the Southern Hemisphere where it is considered a non-native (Nelson et al. 2021).

Ulva species tend to be successful invaders compared to other seaweed taxa owing to their opportunistic life history and rapid growth rate, and this genus is highlighted as a target for monitoring and management (Williams and Smith 2007). DNA-barcoding studies have been instrumental in detecting non-native Ulva species in many parts of the world (e.g. Heesch et al. 2009; Krupnik et al. 2018; Steinhagen et al. 2019; Wei et al. 2022), including regions such as China where their rapid spread is causing severe ecological and economic impacts (Xie et al. 2020). For South Georgia, a recent horizon scanning study (Dawson et al. 2022) listed five seaweed species (none of which were Ulva species) as potential invasive non-natives in South Georgia, of which two were identified as “high-risk”. In addition to these other seaweeds, our findings provide a baseline for monitoring U. fenestrata in South Georgia and emphasise the need for strict biosecurity measures to minimise the risk of introducing non-native species into potentially fragile ecosystems. This study also paves the way for more detailed molecular analyses of herbarium specimens to establish the likely timeframe, source and mode of arrival of U. fenestrata in South Georgia, providing further evidence to guide biosecurity policy.