Sources of contamination

Recent studies have further defined the nature of local chemical contamination in Antarctica and the main sources or types of chemical contamination are now well established: fuel spills, heavy metals/metalloids (typically copper, lead, zinc, cadmium, mercury, arsenic) and polychlorinated biphenyl (PCB) contamination derived from abandoned waste disposal sites, chemicals disposed of through the sewage system including metals and polybrominated diphenyl ether (PBDE) flame retardants, as well as the release of other persistent contaminants, such as polycyclic aromatic hydrocarbons (PAH) and polychlorinated dibenzodioxins (PCDDs), from combustion processes (e.g. Snape et al. Reference Snape, Stark, Cole, Gore, Duquesne and Riddle2001b, Santos et al. Reference Santos, Silva, Schaefer, Alburquerque and Campos2005, Negri et al. Reference Negri, Burns, Boyle, Brinkman and Webster2006, Hale et al. Reference Hale, Kim, Harvey, La Guardia, Mainor, Bush and Jacobs2008). A significant amount of persistent atmospheric contaminants is also transported to Antarctica from other continents, especially in the Southern Hemisphere. The import of trace gases such as carbon dioxide (from the burning of biomass and fossil fuels) and chlorofluorocarbons (CFCs, used as flame-retardants and refrigerants) has significantly changed the Antarctic atmosphere in recent decades. Through sea ice, persistent atmospheric contaminants are also transferred to water and organisms and can accumulate in tissues and biomagnify in food chains. For example, dichlorodiphenyldichloroethylene (DDE), a derivative of dichlorodiphenyltrichloroethane (DDT), a pesticide that was used globally, has been found to be widespread in Antarctic organisms (Bargagli Reference Bargagli2005) despite restrictions on its use under the Stockholm Convention on Persistent Organic Pollutants.

Chemical contamination from abandoned waste disposal sites and past fuel spills are a legacy from when environmental management was less stringent prior to the entry into force of the Environmental Protocol in 1998 (Fig. 1). There is no comprehensive inventory of contamination in Antarctica, although the amount of contaminated soil and waste has been estimated to be of the order 1–10 million m³ (Snape et al. Reference Snape, Stark, Cole, Gore, Duquesne and Riddle2001b). In global terms, this is a very small volume of contaminated material, and Antarctica's overall contamination legacy is not a large footprint in absolute terms. To put the amount of contamination into context, however, requires consideration of the proportion of habitat that is affected and the geographic scale of assessment. Only about 0.34% of the Antarctic continental area is ice-free (BAS 2004), mostly in the Peninsula and remote mountain regions. The total surface area of rocky outcrops within 5 km of the coast has been estimated to be around 5970 km² (D. Smith, Australian Antarctic Data Centre, personal communication 2008). This area encompasses all Antarctic terrestrial maritime land with rock and soil, and, combined with the shallow nearshore areas, is a very rare and unique type of habitat on a continent that is mostly ice-covered. It is often these isolated Antarctic oases where most wildlife congregates, and coincidentally this is also where most of the research stations are located and where tourists visit. Viewed this way, it becomes practicable to visualize an area about six times the size of King George Island, representing the total coastal Antarctic polar habitat with rock/soil and shallow water, being occupied by 53 active research stations as well as a number of disused stations and infrastructure, many with one or more abandoned waste disposal areas and/or fuel contaminated soil, and a contaminated footprint in the adjacent marine ecosystem. Whilst in some cases it appears that the footprint can extend for up to several hundred metres offshore, the marine effect has yet to be investigated for most coastal stations with the extent of outward ripple effects unknown. Thus, Antarctic chemical contamination may be limited in its absolute areal extent, but the significance of its impact is magnified as it affects a large proportion of an extremely rare habitat. Using the language of the Environmental Protocol, the impacts of such chemicals in the Antarctic environment are neither “minor” nor “transitory” (see Pineschi Reference Pineschi, Caroli, Cescon and Walton2001), and without active remediation and further changes in management practices, localized chemical contamination will continue to have a significant and long-term impact both on the environment and on scientific research.

Fig. 1. Waste disposal site at the abandoned Wilkes Station. (Photo by Ian Snape)

Chemical contamination associated with on-going sewage disposal is of growing concern. Hale et al. (Reference Hale, Kim, Harvey, La Guardia, Mainor, Bush and Jacobs2008) found that contemporary sewage management practices at McMurdo Station and Scott Base are insufficient to prevent the local dispersal and accumulation of PBDE. These types of contaminants, like the more commonly known globally-dispersed persistent organic pollutants, such as dichlorodiphenyltrichloroethane (DDT) and hexachlorobenzene, are important contaminants in the Antarctic because they are highly persistent, bioaccumulate and are toxic (Weber & Goerke Reference Weber and Goerke2003, Goerke et al. Reference Goerke, Weber, Bornemann, Ramdohr and Plotz2004). PBDE derived from materials on station that were treated with flame retardants were found in dust, wastewater treatment sludge, and sediment and fish immediately adjacent to the sewage outfall. Hale et al. (Reference Hale, Kim, Harvey, La Guardia, Mainor, Bush and Jacobs2008) found that levels of PBDE in these fish are comparable to urbanized areas in North America, and concluded that wastewater maceration alone, as allowed under the Environmental Protocol, is permitting significant PBDE and other chemical contamination.

Other minor sources of local contamination include blown cement dust (Adamson et al. Reference Adamson, Adamson and Seppelt1994), airborne pollution from exhaust emissions and incineration (O'Brien et al. Reference O’Brien, Todd and Kriwoken2004), and isolated contamination events associated with our everyday presence in Antarctica. Near-shore sediments in Winter Quarters Bay, McMurdo Sound have also been found to be contaminated with butyltin, probably as a result of abrasion of antifouling paint from the hulls of icebreakers or from ship groundings (Negri et al. Reference Negri, Hales, Battershill, Wolff and Webster2004).

Several detailed chemical studies have attempted to fingerprint the source of contaminants or to use chemical signatures to delineate the timing of contamination or rates and mechanisms of degradation. For example, Townsend & Snape (Reference Townsend and Snape2002, Reference Townsend and Snape2008) used stable lead isotopes to document the distribution of Australian-derived lead contamination adjacent to Casey Station (constructed and operated by Australia since 1969), and to differentiate Australian- and American-derived lead sources adjacent to the nearby abandoned Wilkes Station (constructed by the US in 1957 and operated by Australia between 1961 and 1969). For fuel spills, Snape et al. (Reference Snape, Harvey, Ferguson, Rayner and Revill2005) and Rayner et al. (Reference Rayner, Snape, Walworth, Harvey and Ferguson2007) used chemical fingerprints to differentiate between multiple spill sources where liability (and subsequent payment) was ultimately determined based on the extent of specific spill events. The fingerprinting approach has also been extended to monitoring natural attenuation rates and differentiating between various processes of degradation, such as evaporation, biodegradation and water washing, and for determining the efficacy of various remediation treatments (e.g. Gore et al. Reference Gore, Revill and Guille1999, McIntyre et al. Reference McIntyre, Harvey, Ferguson, Wressnig, Snape and George2007, Revill et al. Reference Revill, Snape, Lucieer and Guille2007).

Longevity of chemicals and their ecological impacts

Apart from fuel spills that occur in high-energy marine environments, where dispersal is rapid, research findings show that chemical contamination is long-lived in Antarctica. Gore et al. (Reference Gore, Revill and Guille1999) described fuel spills from the Bunger Hills that were virtually unaltered after more than 10 years in the ground. Snape et al. (Reference Snape, Ferguson, Harvey and Riddle2006) and Revill et al. (Reference Revill, Snape, Lucieer and Guille2007) found that spills at Casey Station were variably weathered and that evaporation caused substantial losses immediately following the event but, after that, the fuel half-life was on the order of decades. Similarly, Aislabie et al. (Reference Aislabie, Fraser, Duncan and Farrell2001) described enhanced populations of hydrocarbon-degrading microbes more than 30 years after hydrocarbon contamination. Several studies have used recently developed genetic tools to describe the impacts of fuel spills on microbial communities and relate the impacts to contamination (e.g. Evans et al. Reference Evans, Hills and Dickson2000, Aislabie et al. Reference Aislabie, Fraser, Duncan and Farrell2001, Delille et al. Reference Delille, Coulon and Pelletier2004). The unpublished results of a long-term oil-spill experiment undertaken at Casey Station by Stark and co-workers indicates that near-shore communities at several trophic levels remain significantly affected five years after the pollution event (J. Stark, Australian Antarctic Division, personal communication 2008). The longevity of contamination is also probably to be influenced by the volatility of the chemical and the substrate on which it is deposited.

Recent studies have demonstrated that chemical contamination has caused ecological impacts at many different trophic levels, from bacteria to vertebrates (e.g. Evans et al. Reference Evans, Hills and Dickson2000, Stark et al. Reference Stark, Snape and Riddle2003c, Negri et al. Reference Negri, Burns, Boyle, Brinkman and Webster2006). For example, increased numbers of hydrocarbon-degrading bacteria (e.g. Rhodococcus, Acinetobacter, Pseudomonas, and Sphingomonas) and decreased microbial community diversity in soils have been observed, as have higher occurrences of pathological anomalies in fish (Evans et al. Reference Evans, Hills and Dickson2000, Aislabie et al. Reference Aislabie, Fraser, Duncan and Farrell2001). Marine sediments contaminated with hydrocarbons and heavy metals display greater abundance of polychaetes and gastropods compared to uncontaminated sediments, although the response is location dependent (e.g. Stark et al. Reference Stark, Snape and Riddle2003c).

Several studies have found that scales of patchiness, of both physical-chemical properties and biological diversity, are high in the Antarctic (Stark et al. Reference Stark, Riddle and Simpson2003a, Reference Stark, Snape, Riddle and Stark2005, Gasparon et al. Reference Gasparon, Ehrler, Matschullat and Melles2007). Manipulative experiments have been particularly useful in demonstrating cause and effect of chemical contamination (Stark et al. Reference Stark, Snape and Riddle2003c, Thompson et al. Reference Thompson, Goldsworthy, Riddle, Snape and Stark2007). However, one of the major difficulties for highly contaminated sites affected by multiple contaminant sources, such as Winter Quarters Bay or the sites around Casey Station, is that determination of the precise cause and effect of contamination is difficult. For example, Evans et al. (Reference Evans, Hills and Dickson2000) studied the effects of contaminants on fish in and around Winter Quarters Bay. They concluded that metal contamination had only a limited effect on fish, but that observed changes in the pathologic condition of fish were more likely to be due to organic contaminants such as polycyclic aromatic hydrocarbons (PAH) and PCBs. Crockett & White (Reference Crockett and White2003) advocate that the uncertainty associated with attributing effects to specific pollutants should be a focus of future research because it could help refine management or remediation to focus on the most potent chemical sources.

On-site remediation

The importance of managing contaminated sites was emphasized in the Environmental Protocol through its Annex III, Waste Disposal and Waste Management. Article 1 of Annex III requires that past and present waste disposal sites on land and abandoned work sites of Antarctic activities shall be cleaned up by the generator(s) of such wastes and the user(s) of such sites. This obligation does not require the removal of any structure or waste material in circumstances where the removal by any practical option would result in greater adverse environmental impacts than leaving the structure or waste material in its existing location.

The high costs, logistical difficulties, environmental risks and political sensitivities make bulk earth extraction, transport and disposal (dig-and-haul) an unattractive proposition. The challenge for research scientists and engineers is to develop robust low-cost alternatives that can be applied on site. A range of techniques has recently been investigated, including multistage water treatment systems, permeable reactive barriers, and bioremediation of petroleum hydrocarbons (see Snape et al. Reference Snape, Morris and Cole2001a, Northcott et al. Reference Northcott, Snape, Scales and Stevens2005, Filler et al. Reference Filler, Snape and Barnes2008 and references therein). Bioremediation could have the greatest potential cost saving over dig-and-haul if it can be developed to work as well on site as in the laboratory. New Zealand, USA, Argentina, Australia, UK and France have all investigated the bioremediation potential of indigenous microorganisms under a range of engineering conditions (see Aislabie et al. Reference Aislabie, Balks, Foght and Waterhouse2004, Walworth et al. Reference Walworth, Reynolds, Rutter, Snape, Filler, Snape and Barnes2008 and references therein). Treatment by such means will probably take longer, but provided off-site dispersal can be controlled, bioremediation offers a low-cost alternative strategy.

Quantitative ecological risk assessment

The need for objective risk assessment information that is specifically relevant to the Antarctic environment is well illustrated by the difficulties encountered when designing remediation programmes with an absence of clean-up criteria (Stark et al. Reference Stark, Snape and Riddle2006, Snape et al. Reference Snape, Acomb, Barnes, Bainbridge, Eno, Filler, Plato, Pol, Raymond, Rayner, Riddle, Rike, Rutter, Schafer, Siciliano, Walworth, Filler, Snape and Barnes2008). Others have also noted the paucity of quantitative environmental risk data for the Antarctic (Chapman & Riddle Reference Chapman and Riddle2005). Following methodologies developed in Canada and Australia, Snape et al. (Reference Snape, Acomb, Barnes, Bainbridge, Eno, Filler, Plato, Pol, Raymond, Rayner, Riddle, Rike, Rutter, Schafer, Siciliano, Walworth, Filler, Snape and Barnes2008) proposed a weight-of-evidence approach to derive trigger levels and set remediation targets for soils. The approach uses data from chemical modelling, ecotoxicology and ecological community impacts to derive a conservative table of trigger values (minimum values at which reaction or response is required by managers) for a particular contaminant. The idea is that ecosystem health will be protected from contamination if chemical levels stay below the trigger value. The physical characteristics of Antarctica, such as low temperature, seasonal snow or sea ice cover, and limited soil development, influence how contaminants behave and interact with biota in ways that can influence ecological risk. Trigger values derived from some temperate organisms may be too high, i.e. not conservative enough to protect Antarctic or sub-Antarctic soil ecosystems (e.g. Schafer et al. Reference Schafer, Snape and Siciliano2007, Snape et al. Reference Snape, Acomb, Barnes, Bainbridge, Eno, Filler, Plato, Pol, Raymond, Rayner, Riddle, Rike, Rutter, Schafer, Siciliano, Walworth, Filler, Snape and Barnes2008). Antarctic organisms take much longer to develop than similar temperate species. Sensitive development stages last longer, and organisms are especially vulnerable to chemical contamination over a longer period of time. King & Riddle (Reference King and Riddle2001) found that, when the Antarctic sea-urchin (Sterechinus neumayeri (Meissner)) is exposed to copper and cadmium during the sensitive larval stage of reproduction, its EC50 values (the concentration of a contaminant that creates an effect in 50% of the test organisms) are approximately 1/5 to 1/10th of the values found in similar temperate species if the test is allowed to run to the same stage of development. However, it remains the case that very few comprehensive ecological studies have been undertaken that can be used directly to feedback into a trigger value.

Without doubt, Antarctica lags far behind other global regions regarding the existence of environmentally relevant data for risk-based decision-making (Chapman & Riddle Reference Chapman and Riddle2005, Chapman et al. Reference Chapman, McDonald, Kickham and McKinnon2006). In accord, Crockett & White (Reference Crockett and White2003) called for focussed research to quantify contaminated sediment deposition rates, improved delineation of sources and fluxes into and out of Winter Quarters Bay, and to identify specific contaminants that are responsible for apparent toxic impacts. These research needs are common to sites adjacent to all other coastal stations, and such information could help manage contaminated sites that are difficult and extremely expensive to remediate.

Ecological impacts

In the last decade, most sewage impact studies have focussed on benthic invertebrates, particularly near McMurdo and Casey Stations (populations of approximately 1000 and 50, respectively). Before sewage treatment started at McMurdo in 2003, long-term release of untreated sewage led to significantly reduced benthic community abundance around the location of the outfall. However, at all but the most contaminated sites, biodiversity was equal or greater than at control sites (Conlan et al. Reference Conlan, Kim, Lenihan and Oliver2004). Reduced assimilation of organic sewage material near the outfall was caused by the formation of an anaerobic microbial biofilm that caused avoidance by megafauna scavengers (Kim et al. Reference Kim, Thurber, Hammerstrom and Conlan2007). Using a technique that examined carbon- and nitrogen-isotope ratios in sewage, sediments and invertebrates, Conlan et al. (Reference Conlan, Rau and Kvitek2006) suggested that generalist benthic feeders could be used as biomonitors for sewage contamination at sites across Antarctica. Studies at Casey Station showed that even comparatively low volumes of sewage could affect the near-shore marine environment, with impacted sites generally having lower species richness, biodiversity and variability compared with control sites (Stark et al. Reference Stark, Riddle and Simpson2003a). Evidence suggests that sewage-derived heavy metals may impact upon soft-sediment assemblages (Casey Station, Stark et al. Reference Stark, Riddle, Snape and Scouller2003b) and the clam Laternula elliptica (King & Broderip) (Rothera Research Station, Lohan et al. Reference Lohan, Statham and Peck2001).

Comparatively little work has been done on the effects of sewage on other biological groups. In experiments, it has been shown that sewage may cause genotoxic effects and pathological anomalies in Antarctic fish (Van Ngan et al. Reference Van Ngan, Gomes, Passos, Ussami, Campos and Rocha2007). Whilst a comparison of heavy metal effects between Winter Quarters Bay and an almost pristine site suggests that levels of heavy metals found in sewage may have little direct effect on fish (Evans et al. Reference Evans, Hills and Dickson2000). A recent report of E. coli from two fur seal pups suggest that pathogenic microbes can be ingested by marine mammals but it is not yet clear what the pathway might be (Hernandez et al. Reference Hernandez, Prado, Torres, Waldemstrom, Haemig and Olsen2007). Effects upon indigenous marine microorganisms are little understood, although George (Reference George2002) showed that Antarctic marine microorganisms could break down detergents commonly found in sewage but at a lower rate than in temperate locations.

Dispersal

Sewage contamination levels around outfalls depend upon the biological and physical environmental characteristics, the volume of sewage released and the degree of treatment. The extents of several Antarctic station sewage plumes were published before the implementation of the Environmental Protocol (Hughes Reference Hughes2004 and references therein). However, since 1998, sewage plume extents, measured using faecal microorganisms, have been published for McMurdo (Edwards et al. Reference Edwards, McFeters and Venkatsean1998), Casey (Morris et al. Reference Morris, George, Tate, Cathers, Hughson and Ruckstuhl2000), Dumont d'Urville (Delille & Delille Reference Delille and Delille2000) and Rothera Stations (Hughes Reference Hughes2004, Hughes & Thompson Reference Hughes and Thompson2004). Both chemical (faecal sterols and hydrocarbons) and microbial (faecal coliforms, Clostridium perfringens and viruses) sewage indicators showed that contamination was generally localized around the outfalls (see Table I). Indeed, Santos et al. (Reference Santos, Silva, Schaefer, Alburquerque and Campos2005) found elevated metal concentrations in marine sediments only in the immediate vicinity of Comandante Ferraz Station sewage outfall. Ensuring optimal operational performance of wastewater facilities is notoriously difficult, particularly if sewage treatment plants are installed, making on-going sewage impact monitoring essential (Hughes & Blenkharn Reference Hughes and Blenkharn2003, Gröndahl et al. Reference Gröndahl, Sidenmark and Thomsen2008). Ecological impacts are probably modulated by the mixing regime of the discharge area. They are likely to be more marked if discharge is in shallow waters or in an enclosed bay, while strong currents and rapid mixing are likely to dilute impacts.

Table I. Extent of pollution from Antarctic research station sewage outfalls, measured using chemical and microbial indicators.

STP = sewage treatment plant

Environmental factors affecting numbers of viable faecal coliform bacteria in sewage plumes include: the level of microbial input by the station population and transient local wildlife, seawater temperature and salinity flux and dilution effects by local currents and solar ultraviolet radiation (UV) (Hughes Reference Hughes2003a). The negative effects of UV on microbial viability are increased by high seawater oxygen concentrations, but are reduced by UV attenuation caused by increasing water depth, sea ice presence and algal blooms. Antarctic solar radiation also naturally reduces bacterial cell viability with increasing exposure time and decreasing radiation wavelength, ozone column depth, cloud cover and solar zenith angle (Hughes Reference Hughes2005).

Not all human activities are located at the coast. Inland, sewage is generally disposed of either in “sewage bulbs” in ice beneath stations (e.g. Amundsen-Scott Base, Halley Research Station), while field parties may bury faecal waste in shallow snow pits, crevasses, or dispose of it directly into the sea when in coastal locations. The legacy of sewage disposal can be long-term. Using molecular techniques, bacterial genetic material from sewage organisms was detected many years after deposition (Sjoling & Cowan Reference Sjoling and Cowan2000). Spore-forming strains such as Bacillus and Clostridium sp. were shown to survive up to 40 years (Hughes & Nobbs Reference Hughes and Nobbs2004), but faecal coliforms were vulnerable to ultraviolet radiation and desiccation (Hughes Reference Hughes2003b). On the Antarctic Peninsula, regional warming has reduced snow cover around some nunataks and coastal areas, resulting in previously buried (and long forgotten) faecal material melting out of the ice (see examples in Hughes & Nobbs Reference Hughes and Nobbs2004). Some Antarctic operators do not permit dumping of faecal waste in the field, and return it to research stations for disposal, although this itself raises other potential environmental and safety concerns, such as the risk of spills and further contamination during storage in the field, and health and safety concerns for field and logistics staff who are handling and transporting the waste. Antarctic Specially Protected Area (ASPA) management plans generally stipulate that sewage disposal is forbidden within the protected area. While the Environmental Protocol encourages the removal of waste from field camps, the disposal of sewage within the vast majority of Antarctica is still possible, without breaching the terms of Environmental Protocol. The long-term potential impact of sewage is not known largely because it has not been systematically investigated.

Impacts in the vicinity of research stations

In view of the scarcity of reports of vegetation impacts in the scientific literature since 1998, Antarctic Treaty Inspection Reports - reporting the inspections of Antarctic research stations - provide one of the few sources of information documenting recent vegetation impacts in the vicinity of stations. Seven Inspection Reports available since 2001, four of which include inspections of stations where vegetation issues are mentioned, indicate to some extent that impacts may have reduced since the entry into force of the Environmental Protocol, although their limitations (Harris Reference Harris1991) mean that consistency of reporting cannot be assumed. Two of the Inspection Reports cover stations on the western side of the Antarctic Peninsula (United States 2001, United Kingdom et al. 2005), one on Ross Island (Australia 2005) and the other in Dronning Maud Land (Norway 2001). In general they indicate a widespread and high level of awareness of the measures and requirements of the Environmental Protocol amongst the stations inspected, including the need to protect vegetation, although this was not always seen to be translated into practice. There is also some inconsistency in reporting detail - for instance damage reported to moss beds on King George Island in 2005 (United Kingdom et al. 2005) was not mentioned in the previous inspection in 2001 (United States 2001), but it is not possible to state whether the damage had not been observed in 2001, or that it had occurred in the interim.

Impacts at popular visitation sites

Occasional anecdotal comments in the scientific literature relating to tourist visits to Antarctica prior to 1998 indicate that a certain level of trampling of vegetation probably occurred from tourists during that time (Chen & Blume Reference Chen and Blume1997 and references therein, Hansom & Gordon Reference Hansom and Gordon1998). Rigorous scientific data obtained either in this period or more recently are largely lacking. Naveen (Reference Naveen2004) reported results from a ten-year study monitoring impacts at tourist sites on the Antarctic Peninsula since 1994; the report measured a number of faunal and floral indicators, but left unclear was whether any impacts were observed.

The tourism industry organization - The International Association of Antarctic Tour Operators (IAATO) - was formed in 1991 and currently represents over 90% of tour companies operating in Antarctica (IAATO 2008). IAATO has collaborated with the Antarctic Treaty Consultative Parties (ATCPs) to produce the document Guidance to Visitors (ATS 1994), its members have demonstrated an apparently high level of compliance to both its own and Environmental Protocol requirements (IAATO 2007), and it continues to collaborate with the ATCPs to produce Site Guidelines for Visitors for specific sites. Site guidelines have been adopted by the ATCPs for 18 of the most-visited sites in the Antarctic Peninsula area (ATS 2008a). Of these 18 sites, only Barrientos Island and Half Moon Island, both in the South Shetland Islands, have “known impacts” to vegetation, while 11 have “potential impacts” to vegetation (including the sites with known impacts). Five sites have “closed areas” partly due to sensitive vegetation, and for two sites, guided walking routes are recommended to avoid vegetation trampling. Sixteen of the 18 sites support significant vegetation, and these guidelines include clear instructions to avoid walking on any vegetation susceptible to disturbance. At Barrientos Island, successive years of visits by tourist groups had resulted in “the erosion of multiple footpaths through vegetation between the eastern and western end of the island” (ATS 2008a). As a solution, the site guideline describes a new guided walking route which avoids vegetation trampling. Along with the other 17 Site Guidelines, it provides a good example of collaborative practical problem solving between the ATCPs and the tourist industry to minimize vegetation impacts.

Pedestrian approach

When approached by humans, Antarctic wildlife may alter their behaviour. Studies to determine minimum appropriate approach distances in seabirds began in the early 1990s and have continued (e.g. Pfeiffer & Peter Reference Pfeiffer and Peter2003, Martin et al. Reference Martin, de Neve, Fargallo, Polo and Soler2004, de Villiers et al. Reference de Villiers, Bause, Giese and Fourie2006, Burger & Gochfeld Reference Burger and Gochfeld2007, Holmes et al. Reference Holmes, Giese and Kriwoken2008). Altered behaviour in birds may result in energetic and/or time costs (Burger & Gochfeld Reference Burger and Gochfeld2007). Some birds display elevation of heart rate (e.g. Weimerskirch et al. Reference Weimerskirch, Shaffer, Mabille, Martin, Boutard and Rouanet2002, de Villiers et al. Reference de Villiers, Bause, Giese and Fourie2006) or increased expression of stress hormones (e.g. Fowler Reference Fowler1999), but demonstration of direct fitness consequences of these responses remain elusive. Gentoo penguins (Jouventin et al. Reference Jouventin, Stahl, Weimerskirch, Mougin, Croxall, Evans and Schreiber1984, Holmes Reference Holmes2007) and southern giant petrels (Jouventin et al. Reference Jouventin, Stahl, Weimerskirch, Mougin, Croxall, Evans and Schreiber1984, Woehler et al. Reference Woehler, Riddle, Ribic, Huiskes, Gieskes, Rozema, Schorno, van der Vies and Wolff2003, Pfeiffer & Peter Reference Pfeiffer and Peter2004) are especially sensitive to human disturbance and repeated approaches to breeding wandering albatrosses (Diomedea exulans Linn.) influenced their chick survival (Wheeler et al. Reference Wheeler, de Villiers and Majiedt2008), but species-specific sensitivity varies according to site [consider gentoo penguins at Macquarie Island (Holmes Reference Holmes2007) and at Port Lockroy (Muller-Schwarze Reference Muller-Schwarze1984, Cobley & Shears Reference Cobley and Shears1999, cited in de Villiers Reference de Villiers2008)].

Whether visitor impacts lead to reduced recruitment and population abundances of seabirds over longer time scales is difficult to judge because habituation may occur (Cobley & Shears Reference Cobley and Shears1999, Otley Reference Otley2005) and individuals may relocate to other sites. It has, for instance, become evident that the behaviour and the physiology of individual animals can be affected strongly by parameters such as visit duration, visit frequency and intensity and duration of contact by researchers (e.g. Salwicka & Stonehouse Reference Salwicka and Stonehouse2000, Weimerskirch et al. Reference Weimerskirch, Shaffer, Mabille, Martin, Boutard and Rouanet2002, Pfeiffer Reference Pfeiffer2005). This can have severe consequences, potentially lengthening foraging trips and sometimes resulting in abandonment of offspring (e.g. Wilson et al. Reference Wilson, Coria, Spairani, Adelung and Culik1989). Species, however, vary in their susceptibility to human disturbance, although habituation typically only takes place if disturbance is regular and predictable (de Villiers Reference de Villiers2008). Van Polanen Petel et al. (Reference van Polanen Petel, Giese, Wotherspoon and Hindell2007) reported that, among Weddell seals (Leptonychotes weddellii Lesson) and their pups on Windmill Islands in East Antarctica, habituation took place over ten approaches within two hours, while irregular visits did not result in any sign of habituation. At Dumont d'Urville Station, Adélie penguins have been observed to nest underneath station buildings, sheltered from harsh winds, despite high human frequentation (Y. Ropert-Coudert, personal communication 2008), as do gentoo penguins at Port Lockroy. Pfeiffer (Reference Pfeiffer2005) studied effects of human visits of between 5 and 100 m to breeding and resting southern giant petrels and skuas. Scientists triggered the strongest reactions when they checked nests. However, unguided station members and tourists who were walking off usual paths also caused an increase in flight and defensive behaviour, and the reactions of birds varied according to each individual's degree of habituation to people.

Handling of animals

Some scientific research requires catching animals, handling them and fitting external equipment, such as leg bands, flipper tags, tracking devices or ‘crittercams’. In some cases more invasive procedures are involved, including sampling of stomach contents by water offloading and the implantation of various types of biological and environmental loggers. Devices attached to the body of an aquatic animal compromise the animal's streamlining and thereby induce a modification to the animal's foraging performances (e.g. Ropert-Coudert et al. Reference Ropert-Coudert, Wilson, Yoda and Kato2007). This, in turn, may lead either to a decrease in the food load captured or, if the load is kept constant, to an increase in the energy expenditure of the animal. Similarly, putting bands on penguin flippers for identification has led to increased energy expenditure during swimming in the laboratory, physical injury, and mortality during moult (e.g. Jackson & Wilson Reference Jackson and Wilson2002, Gauthier-Clerc et al. Reference Gauthier-Clerc, Gendner, Ribic, Fraser, Woehler, Descamps, Gilly, Bohec and Maho2004). On the other hand, it has been shown in the field that while foraging trips increase slightly in length, food loads brought back to the chick also increase among banded compared to unbanded Adélie penguins, indicating the complexity of the issue (Dugger et al. Reference Dugger, Ballard, Ainley and Barton2006, see also Ballard et al. Reference Ballard, Ainley, Ribic and Barton2001). In some recent studies, flipper bands have been replaced with subcutaneous transponders, though the information gained is not usually equivalent. However, transponders, if not applied judiciously, could also compromise survival in some individuals by introducing bacteria and causing infection (Clarke & Kerry Reference Clarke and Kerry1998). Tracking devices are often attached to the bodies of flying seabirds, such as albatrosses and petrels, and subsequent nest abandonment and extended foraging periods by the birds in question have been observed. However, by limiting the weight of the device to less than 3% of the body mass of the birds, the likelihood of these negative impacts has been reduced (Phillips et al. Reference Phillips, Xavier and Croxall2003).

Construction, noise and light

The construction of infrastructure could potentially result in significant disturbance to Antarctic wildlife. Peter et al. (Reference Peter, Buesser, Mustafa and Pfeiffer2008) recorded a reduced breeding activity of Wilson's storm petrels (Oceanites oceanicus Kuhl) during and after the recent extension of the Chilean airstrip on Fildes Peninsula on King George Island. In the case of the construction of an airstrip at Pointe Géologie in East Antarctica between 1984 and 1992, two islands were destroyed, together with 10–35% of the nesting sites of Adélie penguins, cape petrels (Daption capense Linn.) and snow petrels. Breeding pairs were removed from their nesting areas by station personnel or left the area on their own. Monitoring over a period of 14 years showed that, by 1999, the populations most heavily affected by the destruction of their habitats have re-established themselves to levels prior to disturbance (Micol & Jouventin Reference Micol and Jouventin2001).

Noise and visual intrusion arising from aircraft operations can also disturb Antarctic wildlife (Hughes et al. Reference Hughes, Waluda, Stone, Ridout and Shears2008, Harris Reference Harris2005). According to de Villiers (Reference de Villiers2008), the impacts of aircraft operations on wildlife range from insignificant (e.g. Burton & van den Hoff Reference Burton and van den Hoff2002) or minor behavioural changes (Giese & Riddle Reference Giese and Riddle1999, Southwell Reference Southwell2005), to increases in heart rate and temporary nest desertions resulting in some egg or chick mortality (Wilson et al. Reference Wilson, Culik and Adelung1991), multiple nest desertions (e.g. Sladen & Leresche Reference Sladen and Leresche1970) and mass panic and the resulting death of thousands of birds (Rounsevell & Binns Reference Rounsevell and Binns1991). For example, Giese & Riddle (Reference Giese and Riddle1999) observed behavioural reactions of emperor penguin (Aptenodytes forsteri Gray) chicks to air traffic. All chicks demonstrated increased vigilance and the majority of the animals started to walk or run away at the approach of a helicopter at an altitude of 1000 m. Hughes et al. (Reference Hughes, Waluda, Stone, Ridout and Shears2008) also found a significant change in the behaviour of breeding king penguins (Aptenodytes patagonicus (Miller)) during helicopter overflights at altitudes between 230 and 1768 m. Generally, as the helicopter approached, many non-incubating adults and juveniles walked away from the approaching helicopter while all incubating birds stayed on their nests. Pfeiffer (Reference Pfeiffer2005) showed a stronger response of southern giant petrels and skuas to aircraft over-passes following seldom-used routes compared to main routes. The decrease in the southern giant petrel population at Point Géologie has been higher than that in other populations in East Antarctica, and has been attributed, at least in part, to the higher level of human disturbance there, notably the helicopter landing zone, which is only 40 m from the traditional main breeding area (Micol & Jouventin Reference Micol and Jouventin2001). However, in some cases habituation to air traffic may occur (Cobley & Shears Reference Cobley and Shears1999, Otley Reference Otley2005, Hughes et al. Reference Hughes, Waluda, Stone, Ridout and Shears2008).

Underwater, marine vertebrates are particularly sensitive to anthropogenic noise. The noise associated with all forms of transport is likely to be audible to animals under the sea surface, and underwater explosions can in some cases cause mortality of penguins (Brown & Adams Reference Brown and Adams1983, van Polanen Petel et al. Reference van Polanen Petel, Terhune, Hindell and Giese2006, cited in de Villiers Reference de Villiers2008). Whale watching from ships or small boats can also disturb marine mammals (COMNAP 1999). It has been suggested that high source-level hydroacoustic equipment poses some risk to cetacean hearing and could potentially lead to auditory damage. Equipment of the strength and frequency implicated elsewhere in the world is not used in the Antarctic, although lower intensities may generate avoidance behaviour (SCAR 2006a, Kremser et al. Reference Kremser, Klemm and Kötz2005).

Lights on ships at night can disorient birds, and bird strikes on vessels operating in the Southern Ocean are not infrequent (Black Reference Black2005). The level of mortality is generally, but not always, low. The use of deck lights under conditions of reduced visibility in the vicinity of major breeding sites of burrow-nesting petrels, such as the blue petrel (Halobaena caerulea (Gmelin)) and the common diving petrel (Garrodia nereis (Gould)), have led to occasional incidents where hundreds of birds collided with ships overnight, and were found dead or dying on the deck at dawn (Black Reference Black2005).

Natural and anthropogenically-assisted colonization of Antarctica

Even with Antarctica's geographical and environmental isolation, it is clear that natural transfer and establishment events take place over evolutionary timescales, both into and out of the region (Clarke et al. Reference Clarke, Barnes and Hodgson2005, Barnes et al. Reference Barnes, Hodgson, Convey, Allen and Clarke2006), via several potential dispersal routes (reviewed by Hughes et al. Reference Hughes, Ott, Bölter, Convey, Bergstrom, Convey and Huiskes2006). At temperate Gough Island in the South Atlantic and sub-Antarctic Marion Island in the Indian Ocean, it is estimated that natural colonization processes have been outweighed by anthropogenic introductions by at least two orders of magnitude over the few hundred years since the islands’ discovery (Gaston et al. Reference Gaston, Jones, Hänel and Chown2003, Gremmen & Smith Reference Gremmen and Smith2004). It is not clear that the same magnification can be applied to colonization of the much more severe Antarctic continental environment, as fewer potential colonists will posseses appropriate pre-adapted biological features (Convey Reference Convey and Rogan-Finnemore2008). In the sub-Antarctic, the last two centuries of human contact have seen major changes in the structure and functioning of terrestrial ecosystems (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005, Reference Frenot, Convey, Lebouvier, Chown, Whinam, Selkirk, Skotnicki, Bergstrom and Rogan-Finnemore2008, Convey et al. Reference Convey, Frenot, Gremmen, Bergstrom, Bergstrom, Convey and Huiskes2006). These changes resulted from the anthropogenic introduction of many non-indigenous plants and animals, both deliberately and inadvertently. The sub-Antarctic provides a clear warning of the consequences of the establishment of invasive non-indigenous species (sensu Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005) within the Antarctic continent, which is expected to increase in likelihood under climate change. Although very few non-indigenous species are yet established in the latter region, it is clear that many analogous transfers of biota do occur, with the major vectors including cargo, vehicles, food, clothing and people themselves (Sjoling & Cowan Reference Sjoling and Cowan2000, Whinam et al. Reference Whinam, Chilcott and Bergstrom2004, Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005, Lewis et al. Reference Lewis, Riddle and Smith2005, Reference Lewis, Bergstrom and Whinam2006, Hughes et al. Reference Hughes, Ott, Bölter, Convey, Bergstrom, Convey and Huiskes2006).

The early years of human activity in Antarctica included many examples of the deliberate import of a range of vertebrates for transport, food and recreational purposes. Such imports are no longer permitted under the Antarctic Treaty. However, on many sub-Antarctic islands (primarily Kerguelen, Macquarie, Marion, South Georgia), the human-mediated introduction of non-native vertebrates (rats, cats, rabbits, sheep, reindeer, mouflon) has led to major negative impacts for nesting birds and their chicks (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005). The 1960s and 1970s also saw a number of botanical transplant experiments into Antarctic field locations, particularly at the South Orkney Islands (Smith Reference Smith1996), assessing the characteristics required of potential colonizing species. Again, such transplants are no longer permitted, and the ATS provides formal mechanisms for controlling the import into Antarctica, handling and subsequent destruction of non-indigenous biota for scientific purposes. However, despite the clear strictures provided under the ATS, there continue to be new instances of angiosperms being deliberately imported into the Antarctic and planted amongst native vegetation. The most clear-cut example occurred around 1997 near Great Wall Station, King George Island, when several grasses, other graminoids and Cerastium sp., were planted in fellfield beside a vehicle track (Smith Reference Smith, Huiskes, Gieskes, Rozema, Schorno, van der Vies and Wolff2003). These persisted over several years and at least one species (Cerastium sp.) produced several new plants nearby. All plants that could be located were removed in 2005 (H.-U. Peter, personal communication 2008).

Known established non-indigenous species - invertebrates and plants

The number of instances of non-indigenous species establishing in Antarctica is much lower than that for the sub-Antarctic (Table II), with a total of up to eight species (five proven) known from the former, compared with approximately 200 from the latter. There is no confirmed evidence of any of these persistent non-indigenous species becoming invasive. Additional to the small number of species that have become established in Antarctica, there are many anecdotal and a few published instances of non-indigenous biota existing synanthropically (i.e. their survival is only possible in direct association with human activity) within active station buildings over periods of at least several years (Hughes et al. Reference Hughes, Walsh, Convey, Richards and Bergstrom2005, Greenslade Reference Greenslade2006). There are also anecdotal reports of vertebrates kept as “pets” (dogs, birds, fish) and plants on several stations and ships, despite this being prohibited under the Antarctic Treaty.

Table II. The occurrence of established non-indigenous species of terrestrial biota across Antarctic biogeographical zones (updated from Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005).

The five proven examples, established for at least several years, are all in the immediate vicinity of research stations. Three of these are grasses - Poa annua Linn. near Arctowski Station, King George Island (Olech Reference Olech1996), P. pratensis Linn. at Cierva Point near Primavera Station, northern Antarctic Peninsula (Smith Reference Smith1996) and P. trivialis at Syowa Station, on the continental coast (Japan 1996; this plant was removed in 2007). There are no published studies of population or distribution trends, although there are anecdotal reports that the area occupied by the P. annua population at Arctowski Station has expanded in recent years, and there have been some partial attempts at removal. The presence of these species is significant, as both P. annua and P. pratensis are alien species that are invasive on several sub-Antarctic islands (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005, Reference Frenot, Convey, Lebouvier, Chown, Whinam, Selkirk, Skotnicki, Bergstrom and Rogan-Finnemore2008).

The remaining known persistent non-indigenous species are invertebrates accidentally introduced to the immediate vicinity of Signy Station, South Orkney Islands. They were almost certainly introduced during the plant transplant experiments that took place in the 1960s as mentioned above (e.g. Convey & Block Reference Convey and Block1996, Dozsa-Farkas & Convey Reference Dózsa-Farkas and Convey1997), although their presence was not recognized until the early 1980s. They are an enchytraeid worm, Christensenidrilus blocki Dozsa-Farka & Convey, and a brachypterous chironomid midge, Eretmoptera murphyi Schaeffer, and both are otherwise known only from sub-Antarctic South Georgia. Given the climatic similarity between South Georgia and the South Orkney Islands, these species illustrate the likelihood of successful establishment by species already possessing appropriate preadaptations; the fact that the midge is parthenogenetic further increases this probability (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005).

Three further records of non-indigenous Collembola (springtails) have been published from maritime Antarctic sites, but no subsequent assessment has been made of their persistence or status. The cosmopolitan springtail Hypogastrura viatica (Tullberg) has been recorded from the South Shetland Islands (where there are many research stations and much human activity) and Léonie Island in Marguerite Bay (several kilometres from Rothera Station, Adelaide Island) (Greenslade Reference Greenslade1995). This species is a particular concern, as it or its congeners have become aggressive species in their invasions of some sub-Antarctic islands, including South Georgia and Macquarie Island (e.g. Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005, Greenslade Reference Greenslade2006). Finally, Folsomia candida (Willem) and Protaphorura sp. have been reported from Deception Island, South Shetland Islands (Greenslade & Wise Reference Greenslade and Wise1984). However, their association with human activity is unproven as, although this island is the location of several research stations, it is also geothermally active with areas of heated ground. Such habitats encourage the natural establishment of lower-latitude taxa otherwise unable to survive in the Antarctic terrestrial environment (e.g. Convey et al. Reference Convey, Greenslade and Pugh2000a, Smith Reference Smith2005).

Microbiota

The potential significance of microbial introductions to, and subsequent movement within, the Antarctic continent has only recently received any prominence, and very few data are available (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005, Convey Reference Convey and Rogan-Finnemore2008), other than those relating to the risks associated with sewage (where non-indigenous faecal microbes may infect local bird and marine mammal populations). This is compounded by the difficulty of separating a “newly recorded” microbial taxon that has been introduced anthropogenically, from one that is present naturally and is either cosmopolitan in distribution or has dispersed to the region by natural means. Further, as with some of the invertebrate examples given above, even where evidence of introduction has been proposed, no follow-up studies have assessed status or change (Convey Reference Convey and Rogan-Finnemore2008 and references therein). Circumstantial evidence of microbial introductions is provided by studies that compare diversity at human impacted sites (e.g. close to research stations) and pristine areas, assuming that taxa present only at the former are likely to be associated with human activity (e.g. Upton et al. Reference Upton, Pennington and Haston1997, Azmi & Seppelt Reference Azmi and Seppelt1998). A particular risk arising from anthropogenic microbial introduction is that of disease in wildlife (Kerry et al. Reference Kerry, Riddle and Clarke1999). Proof of disease source is problematic (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005) as, even where evidence of disease exposure exists, alternative sources exist, as many Antarctic birds and mammals encounter humans beyond the continent itself during at least part of the year while some, particularly skuas and gulls, regularly forage around lower latitude ports and refuse sites as well as following ships at sea.

Advances in the application of molecular tools promise to improve our ability to identify non-indigenous microbes - for instance, while the classically described Antarctic algal flora is thought to be largely cosmopolitan (Broady Reference Broady1996), molecular studies of various microbial groups are now showing them to be more distinct and by implication indicative of more ancient evolutionary isolation (Lawley et al. Reference Lawley, Ripley, Bridge and Convey2004, Boenigk et al. Reference Boenigk, Pfandl, Garstecki, Harms, Novarino and Chatzinotas2006). Molecular methods that ease the detection of specific microbes as “indicators” of human activity are also now available (Baker et al. Reference Baker, Tow and Cowan2003).

Intra-continental transfer

While controlling the risk of establishment of species not already native to the Antarctic continent is clearly important, a separate but fundamentally important risk remains. The Antarctic continent is not a single biogeographical unit and shows considerable intra-continental regionalization (Chown & Convey Reference Chown and Convey2007, Convey & Stevens Reference Convey and Stevens2007, Convey et al. Reference Convey, Gibson, Hillenbrand, Hodgson, Pugh, Smellie and Stevens2008). Therefore, there are clear threats to regional biodiversity from transfers occurring within the continent. In this context, the strength of the biogeographical boundary across the southern Antarctic Peninsula (the “Gressitt Line” - Chown & Convey Reference Chown and Convey2007) is significant, as several major higher taxonomic groups (in particular the Acari, Collembola and Nematoda) share no or virtually no species across it.

Molecular biological studies now indicate that many elements of the Antarctic terrestrial biota are considerably more ancient or evolutionarily distinct than previously suspected (reviewed by Convey et al. Reference Convey, Gibson, Hillenbrand, Hodgson, Pugh, Smellie and Stevens2008), highlighting an often unappreciated but important element of biodiversity - that of local evolutionary intraspecific differentiation (see Chown & Convey Reference Chown and Convey2007). The highly fragmented and isolated nature of Antarctic terrestrial ecosystems provides an ideal environment in which differentiation can occur. Therefore, these ecosystems are clearly vulnerable to the inadvertent transfer of native biota from one part of Antarctica to a different region where they are alien. The risks are further inflated as it is nearly always the case that any such biota transferred are likely already to be well adapted to the environmental challenges of their introduction site.

Even within distinct Antarctic regions, analogous risks exist, as there are examples of isolated terrestrial communities lacking elements otherwise ubiquitous for that region. For instance, Convey et al. (Reference Convey, Smith, Peat and Pugh2000b) identified that terrestrial ecosystems of isolated nunataks on the northern coast of maritime Antarctic Charcot Island lack any representatives of the Collembola, a group known from every other community examined in the maritime Antarctic. This lack of an otherwise ubiquitous biological and functional group provides an almost unique research opportunity to address fundamental questions in ecological theory relating to species and functional redundancy and the control of ecosystem structure. Any human contact with the ecosystems of Charcot Island, inevitably originating from a maritime Antarctic location where Collembola are native, creates a serious risk of the transfer of these arthropods. In recognition of this risk, the ice free regions of Charcot Island have recently been proposed and accepted as a new Antarctic Specially Protected Area (ASPA), imposing stringent entry requirements on any researchers proposing to visit this location (ATS 2008b).

A second and final example relates to the exceptional inland terrestrial ecosystems of Ellsworth Land, southern Palmer Land, at the base of the Antarctic Peninsula) (Convey & McInnes Reference Convey and McInnes2005). These ecosystems are currently unique worldwide in being dominated by Tardigrada and lacking any representatives of the Nematoda, a group previously thought to be ubiquitous in faunal communities (Freckman & Virginia Reference Freckman and Virginia1997). These ecosystems face at least two serious risks. First, as above, they lack many of the major taxonomic groups that are generally common in most other Antarctic terrestrial ecosystems, hence the probability of inadvertent transfer of pre-adapted Antarctic biota is again magnified. Second, this region hosts “forward support” facilities and field camps supporting, primarily, operations of the British Antarctic Survey (Sky Hi Nunataks) and tourism activities of Antarctic Logistics and Expeditions (Patriot Hills), as well as being visited by the aircraft and staff of other national operators. Lying south of the biogeographical boundary formed by the Gressitt Line, this region is clearly at enhanced risk of intracontinental transfer of species native only to one part of the Antarctic continent.

Marine non-indigenous species

The potential for introduction of alien marine taxa to the Antarctic region has been recognized in recent years (Frenot et al. Reference Frenot, Chown, Whinam, Selkirk, Convey, Skotnicki and Bergstrom2005, Lewis et al. Reference Lewis, Riddle and Smith2005, Reference Lewis, Bergstrom and Whinam2006), but has largely not formed a subject of study to date. Accordingly, very few data or records of occurrence are available. The non-indigenous green alga Enteromorpha intestinalis (Linn.) is well established in the intertidal zone at Half Moon Island in the South Shetland Islands, possibly introduced via the hulls of visiting vessels (Clayton et al. Reference Clayton, Wiencke and Klöser1997). Lewis et al. (Reference Lewis, Hewitt, Riddle and McMinn2003) investigated the possibility of transport of non-indigenous marine biota between Tasmania, Macquarie Island and the Antarctic continent by national operator and tourist ships, identifying three pathways relating to the transport of i) planktonic organisms from the Southern Ocean and Tasmanian waters, ii) epibenthic organisms from Tasmania, and iii) fouling assemblages. Lee & Chown (Reference Lee and Chown2007) and Lewis et al. (Reference Lewis, Riddle and Smith2005) report instances of transport of marine biota into the Antarctic. Both anthropogenic and natural marine debris provide a vehicle for the transport of adhering biota into the Antarctic region (see Barnes et al. Reference Barnes, Hodgson, Convey, Allen and Clarke2006), and may also have a direct and negative impact on indigenous biota through ingestion or entanglement. While this risk would appear to apply primarily to the marine biota, many coastal sites in the sub- and maritime Antarctic have been documented to accumulate substantial quantities of marine debris annually (Gregory & Ryan Reference Gregory, Ryan, Coe and Rogers1997), including some that have been polluted before any human has set foot on them (Convey et al. Reference Convey, Barnes and Morton2002). This, thereby, extends the risk of introduction to include intertidal, supralittoral and even terrestrial biota capable of surviving exposure to the marine environment during transfer. Again, while the potential of these routes have been identified, no instances of establishment of non-indigenous biota through its utilization have been identified. Somewhat inaccurate publicity has been accorded to the recent record of a potentially invasive North Atlantic spider crab species being recorded from a trawl off the Antarctic Peninsula (Tavares & de Melo Reference Tavares and de Melo2004) and the potential for deep sea crabs to invade from lower latitudes in the Southern Ocean (Thatje et al. Reference Thatje, Anger, Calcagno, Lovrich, Portner and Arntz2005). However, the former relates to a single record, with no subsequent records or evidence of establishment, while the latter relates to potential distributional shifts rather than actual evidence of anthropogenically mediated invasion.

Impacts on target populations

Fish

Comprehensive summaries and detailed reviews of the first 15–20 years of finfishing in the Southern Ocean are available (e.g. Gon & Heemstra Reference Gon and Heemstra1990, Kock Reference Kock1992, Agnew Reference Agnew2004, Duhamel et al. Reference Duhamel, Gasco and Davaine2005, Kock et al. Reference Kock, Reid, Croxall and Nicol2007). Finfishing in the Southern Ocean began in 1961 with the operation of Soviet distant-water fleets in the south-west Atlantic and then spread rapidly into the Scotia Sea. The fishery in this sector began on the marbled rockcod at South Georgia. The stock was depleted within two seasons after about 500 000 tonnes was taken (CCAMLR 1990). After some years of low fishing effort, the fishery switched to the mackerel icefish (Fig. 4). It is possible to conclude that the Soviet fleet significantly reduced rockcod and icefish stocks at South Georgia over three years before moving to the Antarctic Peninsula area (around Elephant Island and the South Shetland Islands) and the Kerguelen Plateau, where, again, the stocks of the same species were overexploited (Kock Reference Kock1992). Other fish stocks that declined substantially after only a few years of fishing were the green notothenia (Gobionotothen gibberifrons (Lönnberg)), the Scotia Sea icefish (Chaenocephalus aceratus (Lönnberg)) and the South Georgia icefish (Pseudochaenichtys georgianus Norman; Gon & Heemstra Reference Gon and Heemstra1990, Kock Reference Kock1992).

Fig. 4. Catch statistics for the marbled rock cod (Notothenia rossii) and mackerel icefish (Champsocephalus gunnari), in the statistical area 48 (South Georgia and Antarctic Peninsula). (Data from CCAMLR database)

In total, 12 species of finfish have been commercially fished in the Southern Ocean. Data on stock estimates are not available for all species or for all locations and, indeed, very little information exists on the pre-exploitation biomass of the exploited species. Published studies continue to demonstrate little recovery of many of the stocks that were exploited before the creation of CCAMLR. For example, Kock et al. (Reference Kock, Belchier and Jones2004) reported that no substantial recovery of the marbled rockcod has been observed at South Georgia and off Elephant Island, more than two decades after the closure of the fishery. Further, no recovery of the stocks of marbled rockcod or humped rockcod (Gobionotothen gibberifrons) has been observed in the particularly well-researched coastal zones of the South Shetland Islands (Barrera-Oro & Marshoff Reference Barrera-Oro and Marschoff2007). Barrera-Oro & Marshoff (Reference Barrera-Oro and Marschoff2007) concluded that “despite an overall increasing trend of N. rossii (marbled rockcod) catches from 1991 to 2006, the current level was half of that found in the early 1980s (before the fishery), while those of G. gibberifrons (humped rockcod) had further declined and remained close to zero”. Despite the fact that surviving individuals of the species have increased in size, they have been unable to reproduce to rebuild the populations, resulting in exceedingly low fish populations. Ainley & Blight (Reference Ainley and Blight2008) have hypothesized that a climate oscillation may be responsible for lack of recovery. In the vicinity of Elephant Island, trawl surveys showed that 30 years of fisheries have had considerable impacts on the stocks of mackerel icefish, marbled rockcod, blackfin icefish (Chaenocephalus aceratus) and humped rockcod (Kock Reference Kock1998). Jones et al. (Reference Jones, Kock and Balguerías2000) concluded that overall levels of biomass around the South Orkney Islands are so low that most fish stocks are probably not in a state that could withstand even limited exploitation.

The fishery for the deep sea species Patagonian toothfish (Dissostichus eleginoides) began in the late 1970s in South Georgian waters and the Scotia Arc, though the degree to which it was fished prior to record keeping is unknown (Kock Reference Kock1992). This was the first fishery that CCAMLR had the opportunity to manage, other than by closure. The Scientific Committee of CCAMLR, in the absence of ecological data, has opted for what it believes to be a precautionary approach to this fishery. It introduced an international scheme of scientific observors, and a consensus agreement to establish a Total Allowable Catch (TAC) within small-scale management units. Due to its high commercial value, the toothfish fishery has suffered continuing and pervasive threats from illegal, unreported and unregulated (IUU) fishing, with IUU contributing significantly to the total global catch of toothfish (Agnew et al. Reference Agnew, Pearce, Peatman, Pitcher and Pramod2008). By the early 2000s, the total IUU catch for toothfish was estimated to be at least double the legal one, well in excess of the aggregate global limit recommended for the regulated fisheries for all CCAMLR waters. Some evidence indicates that IUU activities are under control in areas subject to national jurisdiction, e.g. Kerguelen Shelf, South Georgian Shelf (Croxall & Nicol Reference Croxall and Nicol2004), but may be intensifying in others, especially over high-sea and oceanic banks. This is particularly true in the Indian Ocean sector where extensive exploitation of large adult toothfish by IUU vessels could have already caused substantial, long-term damage to the stock (CCAMLR 2007). Over the years, the toothfish fishery has expanded farther south and included a second species, the Antarctic toothfish (D. mawsoni). The fishery has now reached as far as it possibly can, to the Antarctic's southernmost shores, those of the Ross Sea.

Incidental mortality and bycatch

Seabirds

Longline fishing operations for toothfish began under CCAMLR supervision in 1986. It quickly became clear that this fishery carried considerable risk of fatality for seabirds like albatrosses and petrels, which are scavengers and thus readily feed on longline baits, get hooked and drown (Fig. 5). Decreases in albatross populations on the sub-Antarctic islands became evident in the mid-1990s, particularly at South Georgia and Iles Crozet where the longest series of annual population counts were available (Fig. 6; Croxall et al. Reference Croxall, Prince, Rothery, Wood, Robertson and Gales1998, Jouventin & Weimerskirch Reference Jouventin, Weimerskirch, Kerry and Hempel1990). Observations and census data linked population decline to incidental mortality associated with longline fisheries (Croxall et al. Reference Croxall, Rivera, Moreno and Kennelly2007). Delord et al. (Reference Delord, Gasco, Weimerskirch and Barbraud2005) reported that 26 668 seabirds were killed between 2001 and 2003 by legal toothfish longline fishery boats operating around the Iles Kerguelen and Iles Crozet. Species included (in decreasing order by number caught) white-chinned petrel (Procellaria aequinoctialis Linn.), grey petrel (Procellaria cinerea Gmelin), giant petrels (Macronectes sp.), black-browed albatross (Diomedea melanophrys Temminck) and grey-headed albatross (D. chrysostoma Forster). The highest mortality of white-chinned petrels, grey petrels, black-browed albatrosses and grey-headed albatrosses occurred during the chick-rearing period. Longline vessels using automatic baited lines caught many more birds than those using manually baited lines (Delord et al. Reference Delord, Gasco, Weimerskirch and Barbraud2005). Mortality levels from the longline fisheries in CCAMLR waters were unsustainable for the albatross populations involved (CCAMLR 2002). These long-lived birds delay breeding until ten years old, do not necessarily breed every year, and only produce one fledgling at best when successful. They are thus highly susceptible to adult mortality. In view of the mortality rates from longline fisheries their decline could be potentially irreversible within two or three decades (Croxall & Nicol Reference Croxall and Nicol2004).

Fig. 5. Black browed albatross killed in an Antarctic longline fishery. (Photo by Carlos Moreno)

Fig. 6. Numbers of pairs of breeding albatrosses in study colonies at South Georgia, south of the Antarctic circumpolar front (unpublished data, courtesy of the British Antarctic Survey).

IUU fishing has compounded the problem. The CCAMLR Scientific Committee estimated that the number of seabirds killed in IUU fishing alone could have ranged between 151 000 and 543 000 during the ten years between 1996 and 2006 (CCAMLR 2007b). Around the Prince Edward Islands, as is true elsewhere, seabird mortalities from legal and IUU longline fishing have led to significant impacts on the breeding populations of several species of seabirds (Nel et al. Reference Nel, Ryan and Watkins2002).

Impacts on marine ecosystems

Antarctic marine systems have suffered major perturbation through hunting and fishing over the last two centuries and the consequential changes in these ecosystems (see reviews in Ballance et al. Reference Ballance, Pitman, Hewitt, Siniff, Trivelpiece, Clapham, Brownell, Estes, Demaster, Doak, Williams and Brownell2006, Ainley & Blight Reference Ainley and Blight2008) have lost attention in current research (Ainley et al. Reference Ainley, Ballard, Ackley, Blight, Eastman, Emslie, Lescroël, Olmastroni, Townsend, Tynan, Wilson and Woehler2007, see also Nicol et al. Reference Nicol, Croxall, Trathan, Gale and Murphy2007). The prevailing theory of the 1980s, the so-called “krill surplus” hypothesis, postulated that near extermination of large whales in the 20th century led to a surplus of krill as a key prey in the Southern Ocean ecosystem, with consequent increases in the numbers of other krill consumers (Ballance et al. Reference Ballance, Pitman, Hewitt, Siniff, Trivelpiece, Clapham, Brownell, Estes, Demaster, Doak, Williams and Brownell2006). While some studies have found evidence that supports this hypothesis, others argue that the loss of whales alone is not adequate to explain the population trends of some supposed key krill-eating species (e.g. gentoo, Adélie and macaroni penguin (Eudyptes chrysolophus Brandt); see Emslie & Patterson Reference Emslie and Patterson2007, Ainley & Blight Reference Ainley and Blight2008, McClintock et al. Reference McClintock, Ducklow and Fraser2008).

The impacts of fisheries on Antarctic marine ecosystems are poorly understood. An analysis by Ainley & Blight (Reference Ainley and Blight2008) indicates that removal of demersal finfish from insular and northern continental shelves during the 1960s–1980s had significant effects on populations of fish-eating predators, such as gentoo and macaroni penguins, Antarctic fur seals (Arctocephalus gazella Peters), southern elephant seals, and Antarctic shags (Phalacrocorax bransfieldensis Murphy; see also Barrera-Oro et al. Reference Barrera-Oro and Marschoff2007). The removal of finfish means that some of its predators may suffer food scarcity, some of its prey, such as salps (e.g. Salpa thompsoni (Foxton)), will experience less predation and in addition climate effects will become more prominent (e.g. Osterblöm et al. Reference Osterblöm, Hansson, Larsson, Hjerne, Wulff, Elmgren and Folke2007). This introduces the possibility of trophic cascades rippling through foodwebs at the community level, with changes in the medium-term biodiversity and relative abundance of other species in the demersal community (Worm et al. Reference Worm, Barbier, Beaumont, Duffy, Folke, Halpern, Jackson, Lotze, Micheli, Palumbi, Sala, Selkoe, Stachowicz and Watson2006). For example, preliminary information collected in waters off Ross Island indicates that fish-eating killer whales have become less frequent and Adélie penguins have begun to feed on silverfish (toothfish prey) to a greater degree during chick feeding since the advent of the toothfish fishery (DeVries et al. Reference DeVries, Ainley and Ballard2008). These first-order ecological effects in the Ross Sea remain under study but the potential for trophic cascades is high (Ainley et al. Reference Ainley, Ballard, Ackley, Blight, Eastman, Emslie, Lescroël, Olmastroni, Townsend, Tynan, Wilson and Woehler2007) in accordance with scenarios played out in other parts of the world's oceans with the removal of top predators, of which the toothfish is one (Pauly & Maclean Reference Pauly and Maclean2003, Osterblöm et al. Reference Osterblöm, Casini, Olsson and Bignert2006, Daskalov et al. Reference Daskalov, Grishin, Rodionov and Mihneva2007). Furthermore, the destruction that bottom trawling and longline gear can cause to benthic communities has become a topic of concern in the Southern Ocean as elsewhere in the world's oceans (CCAMLR 2007, Kock et al. Reference Kock, Reid, Croxall and Nicol2007). There is sufficient evidence globally that benthic habitats comprising slow-growing, habitat-forming, sessile species, such as deep sea corals and sponge communities could take much longer than three decades to recover from significant fisheries disturbances. However, little information exists on the extent to which damage might have already occurred in the Southern Ocean (CCAMLR 2007).