We visited 25 different sites on Antarctic islands and the Antarctic Peninsula during the Chilean Scientific Antarctic Expeditions of 2007/08 and 2009/10—the austral summer seasons—in order to search for populations of Poa annua (see Molina-Montenegro et al. 2012). In each site we visually surveyed the presence of P. annua around every building and point of landing (Fig. 1). Each plant individual was labelled in order to avoid recording the same individuals twice, and its distance to the closest buildings was recorded. Out of all sites visited we found individuals of P. annua at six of them. The plant was recorded at one point at each of five sites, whereas at the Polish Arctowski Antarctic Station it was recorded at two points.

In order to find potential drivers of P. annua establishment, we performed a regression analysis using the abundance of this species and (i) mean number of tourists landed in each point where P. annua has been recorded, (ii) mean number of scientists and operators landed in each point where P. annua has been recorded, (iii) distance to the main building of each station and (iv) level of soil disturbance. We used the time in years from the last time when any building was constructed or renovated at each station as a proxy for soil disturbance. Drawing from the database of the International Association of Antarctica Tour Operators, we considered the mean number of visits (tourists and non-tourists) landed during the 2007/08 and 2009/10 summer seasons—the growing seasons that coincided with our fieldwork.

To assess the effect of different soil-patch types on establishment, we performed a transplant experiment with adult individuals of P. annua collected close to the coastline (within 30–40 m to the sea). This experiment was conducted in a study area near Arctowski Station. All plants used ranged in height from 5 to 7 cm and appeared healthy. Each individual plant was excavated together with the soil around the roots (ca. 500 g) and kept well-watered in a plastic box under natural conditions of light and temperature for 12 h until the transplant. Plant status was visually assessed just before the transplant (plants showing foliar and/or root damage were excluded). We considered two soil-patch types for the experiment; the first type was in the bare ground between the resident vegetation (natural) and the second was in the disturbed soil surrounding the scientific stations. We selected three patches by each soil-type, with distances of 23, 41 and 56 m among each pair of sites (natural–disturbed). At each patch, 10 P. annua individuals were planted within each of three soil-patches of each type (n=30 individuals per soil-type). Individuals were planted at least 5 cm apart from each other. Transplants were carried out during the 2010/11 growing season and seedling survival was evaluated after six weeks. Additionally, we compared nitrogen availability and soil moisture between five non-disturbed and five disturbed patches. A soil sample of 10 cm depth (ca. 100 g) was taken in each patch. This depth was selected is this the average depth of the root zone of P. annua in the study. Soil samples were stored in sealed plastic bags and sent for analyses to determine the concentration of nitrate ( N O 3 - ; Robarge et al. 1983) and ammonium ( N H 4 + ; Longeri et al. 1979). Soil moisture (soil matric potential) was measured in five non-disturbed patches and five disturbed patches during the mid-growth season (January 2010). At each sampling point, a 2725 series Jet Fill Tensiometer (Soil Moisture Equipment Corp., Santa Barbara, CA, US) was dug into the soil to a depth of 10 cm. Tensiometers were placed at mid-day and, after a stabilization period of 2 h, the soil matric potential was recorded. At the end of experiment, all surviving individuals were removed from the site and destroyed.

Finally, we performed a manipulative experiment to assess the effect of soil disturbance on germination of Pou annua, Colobanthus quitensis and Deschampsia antarctica. We collected seeds from the inflorescence of the previous year's adult healthy-looking plants growing in tussocks in ice-free zones close to the coastline. At the commencement of this experiment three seeds of each species were tested to assess viability. This was done by cutting the seed in half on the equatorial axis with a razor blade after 24 h of immersion in tetrazolium chloride. All seeds were viable.

Each experimental unit was a plastic pot (50×40×30 cm) filled with soil extracted from the study site. Each experimental unit consisted 10 seeds of P. annua or C. quitensis or D. antarctica (n=6 pots per species; total n=18 pots). Seeds were buried 4–5 cm deep in a plastic pot, as this was the maximum depth at which P. annua seeds were found at the study site. Pots were subjected to a 20/4 h light/dark photoperiod at 4°C and were watered with 100 ml of water every four days. To determine the effect of soil disturbance on germination, half the pots of each species was subjected to manual disturbance in the top 4–5 cm of soil every three days until emergence of the first seedling (total n=9 pots). After the first germination, manual disturbance ceased in order to avoid seedling damage, but the recording of germination continued. The remaining pots—the controls—were undisturbed. Disturbance was deemed similar in intensity and effects to that of footfall and light vehicle traffic. At the conclusion of the experiment, which ran for three months, we recorded the number of emerged seedlings.

During the 2007/08 and 2009/10 growing season, the presence and abundance of P. annua at six different sites was recorded out of a total of 25 sites in the South Shetland Islands (one site at Deception Island and two sites at Arctowski Station) and the Antarctic Peninsula (Bernardo O'Higgins Station, Gabriel González Videla Station, and Almirante Brown Station). Regression analysis did not show a relation between P. annua abundance and the mean number of tourists or non-tourists landed at each scientific station (Fig. 2a, b). However, the distance of P. annua individuals to the main station building showed a positive but not significant correlation (Fig. 2c), and the abundance of P. annua was positively and significantly correlated with the number of years elapsed since any building was constructed or renovated (Fig. 2d).

In the transplant experiment, the survivorship of P. annua individuals was significantly higher (F _{1,  8}=6.11; P=0.039) in sites with soil disturbance than without disturbance (Fig. 3). Soil moisture was not different (F _{1,  8}=28.52; P=0.09) between disturbed and non-disturbed patches of soil (−25±4 and −30±6 MPa disturbed and non-disturbed, respectively). In contrast, the available nitrogen content of the disturbed soil was significantly higher (F _{1,  8}=231.21; P<0.01) than in non-disturbed conditions (7.56±1.2 and 3.98±1.6 mg·K⁻¹ disturbed and non-disturbed, respectively).

Overall, the percentage of germination was significantly increased under the disturbance treatment (F _{1,  24}=563.33; P<0.001), but this trend was not similar for all species ( F _{2,  24}=364.25; P<0.001). For P. annua the germination increased with disturbance, for C. quitensis germination decreased significantly under disturbance while for D. antarctica the germination percentage was not different between the disturbance and control treatments (Fig. 4).

In the 21st century, the number of visitors to the Antarctica and the scale of human influence have increased dramatically, mainly due to tourism (IAATO 2013). There has been concern about the potential impact of visitors on the Antarctic environment (Culik & Wilson 1995; Chown et al. 2012). Several studies have aimed to detect the effects of this visitor pressure (Chwedorzewska & Korczak 2010).

We did not find a positive correlation between Poa annua abundance and the mean number of tourists or scientists visiting the Antarctic. However, a positive relationship between the abundance of P. annua and the time since a base underwent construction suggests that this kind of human activity influences the establishment phase of this species. We believe that human activity associated with research stations enhances soil nutrients and modifies the soil matrix, and therefore facilitates the germination and establishment of P. annua. Several studies have shown that cargo and vehicles transported to Antarctic research stations contain propagules of many plants (Whinam et al. 2005; Lee & Chown 2009b; Chown et al. 2012). Hughes et al. (2010) highlighted that research station construction also facilitates the transport of non-native seeds. We found that time since construction influences the presence of P. annua, most likely because more time allows for the establishment and population expansion of the plant. In Antarctica, ice-free areas near the coast are rather scarce and are hotspots for local fauna and flora (Convey 1996) as well as humans (Chown et al. 2012). High traffic around research stations creates a high level of soil disturbance, which can facilitate the establishment of exotic plants (Chwedorzewska 2008; Cuba-Díaz et al. 2013), as our experiments support. In stressful environments such as those found in the Antarctic terrestrial ecosystems, it has been predicted that alien species are unlikely to colonize and extend their range of distribution unless some process relaxes the environmental constraints (Dullinger et al. 2003; Pauchard et al. 2009).

It is widely assumed that soil disturbance enhances resource availability (e.g., Burke & Grime 1996; Davis et al. 2000), and some have examined the effect of disturbances in cold environments and shown increased nutrients, or higher accessibility, following disturbance (Stanton et al. 1994; Grieve 2000). We have elsewhere shown that alien species take advantage of high resource availability (e.g., nutrient or water), improving their ecophysiological performance (Molina-Montenegro et al. 2011). In the study reported here, we found that disturbed soils had higher levels of nutrients, as well as germination and survival percentage of P. annua, suggesting that disturbance by humans may generate microhabitats more favourable to P. annua than to native plants. Alternatively, the enhanced establishment and survival of P. annua in disturbed areas could result from non-random spatial distribution due to wind. In cold environments some plant species show greater establishment and survival percentages when sheltered from the desiccating effect of wind (Resler et al. 2005). In Antarctica P. annua has been shown to become established at sites sheltered from wind, such as alongside buildings (Olech 1996). More surveys should be conducted to further explain the spatial distribution and performance of P. annua in Antarctica.

Since it was first recorded in the mid-1980s (Olech 1996), P. annua is the only introduced plant species to have established permanent scattered populations in the Shetland Islands and on the Antarctic Peninsula (Frenot et al. 2005; Chwedorzewska 2008, 2009; Molina-Montenegro et al. 2012). The species has shown a steady, constant increase in abundance and number of new populations (Molina-Montenegro et al. 2012). Having spread to alpine ecosystems (Johnston & Pickering 2001), northern Europe (Gederaas et al. 2012) and Sub-Antarctic islands (Frenot et al. 2005), P. annua is “generalist” invasive species that tolerates stressful environments and takes advantage of favourable conditions, e.g., enhanced nitrogen availability in disturbed soils.

Surprisingly, nowadays there is no international policy to control the introduction of non-native species to Antarctica. Our study provides evidence that Antarctic environments could be severely impacted in the near future by such introductions and in the long-term they may have unsuspected consequences on Antarctic diversity. Although, some efforts have been made to accomplish this goal, they are still insufficient to mitigate the negative impacts of human activities on Antarctic biodiversity (see Hughes & Convey 2014).

In summary, P. annua is positively correlated with human activities and its survival and germination is enhanced in soil disturbance conditions. We conclude that human impacts, mainly associated with research activities, may have negative effects on the native species but positively influence the abundance and spread of the invasive P. annua on the Antarctic continent.