Introduction

Approximately, 30% of the ~200 known instances of establishment of non-native organisms in Antarctic and sub-Antarctic terrestrial ecosystems involve insects (Frenot et al. 2005), the large majority of these being on the sub-Antarctic islands. However, very few of these instances are well documented (but see Slabber and Chown 2002; Lee et al. 2007, 2009) or the subject of ongoing directed monitoring or research programmes. Thus, while there is an increasing awareness of the threat posed by non-native organisms to these ecosystems (Frenot et al. 2005, 2008; Convey et al. 2006), assessment of their occurrence, distribution, impacts and current status remains largely reliant on the synthesis of anecdotal and opportunistic observations.

Sub-Antarctic and also Antarctic terrestrial ecosystems, in effect highly isolated islands of habitat, are thought to be particularly vulnerable to the establishment of non-native species (D’Antonio and Dudley 1995; Frenot et al. 2005; Convey et al. 2006). These ecosystems appear to contain vacant niches allowing the establishment of previously unrepresented trophic functions (Cushman 1995; Chown et al. 2008), and the typically adversity-selected ecological characteristics of the native biota provide them with limited resistance against invading species with greater competitive abilities (Kennedy 1995; Convey 1996).

The examples of two flightless and predatory carabid ground beetles that have been introduced to South Georgia, one of which also to Îles Kerguelen, over the last century are particularly pertinent (Vogel 1985; Ottesen 1990; Ernsting 1993; Ernsting et al. 1995, 1999; Todd 1996; Chevrier et al. 1997; Brandjes et al. 1999). These beetles, Merizodus soledadinus (Guerin-Ménéville) (previously known as Oopterus soledadinus) and Trechisibus antarcticus (Dejean), provide two of the few known examples of the establishment of South American rather than European/boreal non-native species in the sub-Antarctic. Both beetles are voracious predators of small insects and other invertebrates. Where they occur on these two sub-Antarctic islands, their presence has significantly modified the local food webs, which contain no other comparable indigenous predators. The presence of M. soledadinus on Îles Kerguelen has been found to lead to almost all other large invertebrates becoming locally extinct, including the endemic flightless dipteran Anatalanta aptera Eaton, a species that plays a major role in organic matter decomposition here (see Frenot et al. 2008; Laparie et al. 2010). On South Georgia, the presence of T. antarcticus leads to significant reductions in populations of the endemic herbivorous perimylopid beetle, Hydromedion sparsutum (Müller), through the predation of younger instar larvae and has thereby been implicated in a selective alteration of life cycle characteristics towards more rapid larval development in these prey species (Ernsting et al. 1999; see also Chown and Smith 1993). Other endemic species on South Georgia that may be at risk from these predators include two ground-dwelling and flightless dipteran flies Eretmoptera murphyi Schaeffer (Chironomidae) and Antrops truncipennis Enderlein (Sphaeroceridae), although there has been no investigation of any interactions with these species.

These beetles are thought to have been accidentally transported from source locations within their natural range in the Falkland Islands or southern (cool temperate) South America to South Georgia and Îles Kerguelen, although the precise date and detail of introduction events are unconfirmed, as is typically the case in this region (Convey and Lebouvier 2009). Merizodus soledadinus was most likely introduced to Îles Kerguelen in the early 20th Century (Jeannel 1940; Chevrier et al. 1997). The introduction event in South Georgia is thought to have been more recent, as M. soledadinus was first recorded there near Grytviken in 1963 (Darlington 1970) (see Fig. 1 for locations on South Georgia mentioned in the text). There may have been at least two separate introductions on this island, as M. soledadinus is known from two distinct locations (centred on the abandoned whaling stations at Grytviken and Husvik, while the second non-native carabid (T. antarcticus) was first recorded from the vicinity of the Husvik station in 1982 (Vogel 1985).

Fig. 1
figure 1

Locations on South Georgia mentioned in the text. Top left: position of South Georgia relative to South America, the Antarctic Peninsula and Scotia arc achipelagoes; bottom left: South Georgia, with (inset) top right East and West Cumberland Bays; bottom right: previous range extent of T. antarcticus in 1995/6 in West Cumberland Bay as reported by Brandjes et al. (1999); at this time, M. soledadinus was restricted to the station vicinity at Husvik and between Grytviken and the settlement at King Edward Point, along with a newly located small colony at Hut Point, Jason Harbour, on the Busen Peninsula

Full size image

The most recent survey data available from South Georgia relating to the distribution of these carabid beetles date from the 1995/6 austral summer (Brandjes et al. 1999). At that time, M. soledadinus largely remained restricted to the immediate vicinity of the two abandoned whaling stations at Husvik and Grytviken, although with a new population restricted to Hut Point, Jason Harbour (a location frequently visited by personnel then stationed at King Edward Point), discovered in that season. By 1995/6, T. antarcticus had spread several kilometres along the coast from its single introduction site at Husvik, both along the Busen and Tönsberg Peninsulas, and with separate populations established near to Leith Station and on Grass Island (Fig. 1 indicates the extent of the distribution of this species in 1995/6, based on information taken from Brandjes et al. (1999: Fig. 5)). The current study integrates recent opportunistic records and observations from organised surveys of specific locations on South Georgia, obtained during several recent austral summer seasons, in order to provide an update on the current distribution of the two established carabids and document changes since the mid-1990s.

Methods

Collections

Collections of carabid beetles were made on an opportunistic basis during November 2002 (Maiviken) and late February 2007 (Stromness, Husvik, Grytviken, Corral Bay), by examining habitats in which they are typically common at their already known locations, particularly under debris (e.g. wood, whalebone) amongst vegetation at low altitude. The environs of the abandoned whaling station at Prince Olav Harbour were also examined during the latter season, but no carabids were encountered. As part of a more general survey of terrestrial arthropod diversity between 21 January and 7 February 2006, 10 locations around South Georgia were visited, 440 samples of various vegetation substrata collected for Tullgren-style invertebrate extractions and visual inspection made under debris (timber, whalebone, etc.). Finally, during the 2008/9 austral summer, further collections were made at 21 locations using a range of standard invertebrate sampling techniques (pitfall, yellow water pan and Malaise trapping, vacuum sampling, chemical gradient Berlese extraction using naphthalene of vacuum samples and plant litter and direct searching under stones, litter, moss etc.) as part of the South Atlantic Invasive Species Project (SAISP) expedition (Key and Key 2009). The 2006 and 2008/9 collections were aimed at maximising the species diversity information obtained. Given the range of substrata extracted and techniques used, combined with necessarily short (1–3 h) visits to most locations, we have not attempted any analyses of species abundances, either within the current data sets or relative to previous reports. Locations (GPS) of all sites examined are given in Supplementary Table 1.

The two species are morphologically similar. Differences that are apparent on close examination are provided for guidance in Table 1. As part of a separate study examining the relatedness of different introduced populations of M. soledadinus in the sub-Antarctic using molecular biological techniques (L. Lalouette, L. Konecny and C. Douady pers. comm.), the species identity of individual beetles obtained in November 2002 and February 2007 was confirmed.

Table 1 Characters used to separate Trechisibus antarcticus and Merizodus soledadinus
Full size table

Results

Records of the two carabid species are integrated in Supplementary Table 1. The previous distributional extents are illustrated in Fig. 1. To the west of the previously recorded distribution, the range of T. antarcticus is extended to include Fortuna Bay, where the beetle was present in samples obtained near the main tourist landing location, although not from other sampling locations elsewhere in this bay. Access to Leith Station is currently prohibited, due to presence of asbestos and other risks, by the Government of South Georgia and the South Sandwich Islands, so no changes to the species’ previous limit near to this station could be assessed. Fortuna Bay is linked with the area where this species has previously been recorded via a mountain pass (>300 m a.s.l.) to the west of Leith Station, which is occasionally used by visitors to the island, and the Bay is itself a frequent landing place for visitors from cruise ships and private yachts. There was no opportunity to visit locations on the Busen Peninsula beyond the previous distributional limit of T. antarcticus, or the Jason Harbour location for M. soledadinus, so no assessment of changes in this area can be made, except that neither species was noted at Carlita Bay in January 2006 (Supplementary Table 1).

The range of M. soledadinus in East Cumberland Bay is extended eastwards to the Greene Peninsula and Corral Bay. In the vicinity of Grytviken, its distribution has expanded to include Maiviken. Both Corral Bay and the Greene Peninsula are isolated from the beetles’ previously documented distributions, and each other, by large tidewater glaciers.

Discussion

The data obtained in the current study, drawn from several different observational sources, indicate important changes in the distributions of both the non-native carabid species known to be established on South Georgia. The continued expansion of the range occupied by T. antarcticus is as predicted by Ernsting (1993), Ernsting et al. (1995) and Brandjes et al. (1999), who found the species to be expanding rapidly during the early and mid-1990s. Likewise, although at that time M. soledadinus showed little sign of expanding beyond the two abandoned whaling station boundaries, its expansion from Grytviken to Maiviken (connected by a low-altitude valley with a regularly used path and also by boat traffic) is consistent with the macroclimatic warming trends affecting South Georgia (see Cook et al. 2010), as is the collection in 2009 of a single specimen at an altitude of 275 m on Mount Hodges above Grytviken station. However, this is based on the untested assumption that this warming will act to relax thermal limits on activity or distribution of this species.

The expansion of the area occupied by M. soledadinus to Corral Bay, on the Barff Peninsula, and to the Greene Peninsula, is of particular significance as both these locations are isolated from the previous distribution by tidewater glaciers (Fig. 1), and hence, this cannot simply have taken place by the (flightless) insects moving overland. An analogous challenge is presented by the 1996 discovery of an isolated population of this species at Hut Point, Jason Harbour. Corral Bay is a frequently used landing and access point for boat traffic and personnel moving from the research station at King Edward Point to sites on the Barff Peninsula and beyond. Boat traffic also links King Edward Point with locations throughout East and West Cumberland Bays, and Jason Harbour is a regular landing location. The Greene Peninsula, however, is a protected area with restrictions on landing without permission. Similarly, the lack of intermediate records of T. antarcticus between the newly discovered location at Fortuna Bay and its previously recorded limit near the boundary of Leith whaling station is suggestive of a further direct human role in its expansion to this site. Thus, these new distributional records are all suggestive of an inadvertent human involvement in the underlying dispersal events.

Neither species was apparent in general collections from locations on the eastern coast of the Barff Peninsula or further east in 2006 or 2009 (Ocean Harbour, Godthul, Hound Bay, Moltke Harbour, Cooper Bay, Larsen Harbour). However, there is now no significant geographical obstruction to the expansion of the species eastwards as far as the tidewater glaciers in Royal Bay. Cook et al. (2010) have recently emphasised the significance of the retreat of several glaciers on South Georgia to expose beaches, across which large vertebrates (reindeer, rats) have already successfully passed. Brandjes et al. (1999) similarly note that T. antarcticus had by 1996 successfully crossed a non-vegetated pebble isthmus to reach an otherwise isolated part of the Tönsberg Peninsula between Husvik and Stromness. As with these vertebrates, the ecosystem impacts of invasive predatory invertebrates are now potentially faced across a much larger proportion of the biologically rich and more productive north-east coastline of South Georgia than was previously the case.

In the absence of published ecophysiological studies, it is not currently known whether the beetles face environmental constraints on the expansion of their distributions, as appears to be the case for the non-native flesh fly Calliphora vicina on Îles Kerguelen (Frenot et al. 2008). There is also no detailed macro- or microclimate monitoring programme currently in place on South Georgia, either in the context of documenting and identifying longer term trends or in that of allowing comparison of environmental conditions between different locations on the island.

The strongly negative impacts of these predatory carabids on endemic perimylopid beetle life cycle and abundance have been noted earlier. In that the native species continue to co-occur with the carabids, albeit at much reduced population density, there is no suggestion at present that they are at risk of local extinction, although it is not clear whether an equilibrium state has been reached (see Ernsting et al. 1995). Suggestively, in the SAISS survey and collections, of 42 samples containing the indigenous rove beetle Halmaeusa atriceps (C.O. Waterhouse) (286 individuals) and 50 samples containing one or both species of carabid (total 392 individuals), there were only three samples containing both staphylinds and carabids, possibly suggesting a negative interaction. It would also be of interest to investigate the co-occurrence of the endemic flightless chironomid midge Eretmoptera murphyi with these predatory beetles, as endemic flightless dipterans on Îles Kerguelen have been shown to be very vulnerable to predation by M. soledadinus (Frenot et al. 2008). The SAISS recorded E. murphyi only at Prince Olav Harbour, where it was extremely abundant particularly in swards of the non-native grass Poa annua. Recent studies of this chironomid (Allegrucci et al. 2006) have relied on material collected from Bird Island, with the species not being encountered at locations (near Grytviken and Husvik) that have previously provided source material in the 1980s (Ring et al. 1990; Block et al. 1992) or are assumed to have been the source for the species’ anthropogenic introduction to maritime Antarctic Signy Island in the late 1960s (Block et al. 1984; Convey and Block 1996). Although only an anecdotal observation, this may indicate that this species has become less widespread or common in areas where carabids now occur.

These records highlight the likely continued importance of human assistance in the introduction and local redistribution of invasive non-native predators to the exceptional terrestrial ecosystems of the sub-Antarctic islands. No remediation (eradication) measures are likely to be practicable or effective. Hence, reliance on the effective application of biosecurity measures to minimise the risk of further anthropogenically assisted range expansion, or introduction events, is now the only practicable approach, particularly to ensure that areas of the island currently isolated from those already colonised by features such as tidewater glaciers remain unaffected.