The international pet trade is a major source of emerging invasive vertebrate species. We used online resources as a novel source of information for accidental bird escapes, and we investigated the factors that influence the frequency and distribution of bird escapes at a continental scale. We collected information on over 5,000 pet birds reported to be missing on animal websites during the last 15 years in Australia. We investigated whether variables linked to pet ownership successfully predicted bird escapes, and we assessed the potential distribution of these escapes. Most of the reported birds were parrots (> 90%), thus, we analysed factors associated with the frequency of parrot escapes. We found that bird escapes in Australia are much more frequent than previously acknowledged. Bird escapes were reported more frequently within, or around, large Australian capital cities. Socio-economic factors, such as the average personal income level of the community, and the level of human modification to the environment were the best predictors of bird escapes. Cheaper parrot species, Australian natives, and parrot species regarded as peaceful or playful were the most frequently reported escapees. Accidental introductions have been overlooked as an important source of animal incursions. Information on bird escapes is available online in many higher income countries and, in Australia, this is particularly apparent for parrot species. We believe that online resources may provide useful tools for passive surveillance for non-native pet species. Online surveillance will be particularly relevant for species that are highly reported, such as parrots, and species that are either valuable or highly commensal.
Citation: Vall-llosera M, Cassey P (2017) Leaky doors: Private captivity as a prominent source of bird introductions in Australia. PLoS ONE 12(2): e0172851. https://doi.org/10.1371/journal.pone.0172851
Editor: Christopher A. Lepczyk, Auburn University, UNITED STATES
Received: August 16, 2016; Accepted: February 10, 2017; Published: February 24, 2017
Copyright: © 2017 Vall-llosera, Cassey. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: MV was funded by the Invasive Animals Cooperative Research Centre (http://www.invasiveanimals.com/) (Project No. 1.L.4). PC was an Australian Research Council (http://www.arc.gov.au/) Future Fellow (FT0914420) and this work was supported by the Australian Research Council Discovery Grant (DP140102319). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: MV was funded by the Invasive Animals Cooperative Research Centre (http://www.invasiveanimals.com/). There are no patents, products in development or marketed products to declare. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
International wildlife trade moves millions of individual animals (of many thousands of species) around the world every year . The pet industry, including the popular practice of keeping cage birds, is a large component of wildlife trade , and a significant source of new invasive vertebrate species [3–5]. However, the factors that influence which and how pet animals are released into the wild remain largely unstudied. Unlike the introduction of other wildlife commodities (e.g., fish stocking or game animal acclimatisation), the release of pet animals is an unfortunate consequence of the trade . Pet species are either deliberately abandoned by their owners or can accidentally escape into the wild [6–8]. In some Asian countries, animals are also freed as part of religious practices [9,10]. Very few of these introduction events leave any identifiable records. This may be a substantial knowledge gap in contemporary invasion biology, which may be hindering biosecurity efforts for detecting new invasions and for preventing their impacts.
Former pets turned invasive species are capable of causing extinctions of native plants and animals and reducing biodiversity. They can threaten the organization and functioning of native communities through predation, competitive exclusion, disease transfer or hybridization [11–14]. However, not all released pets pose a high biosecurity risk. In order to become invasive, released pet birds must survive in the wild, establish a self-sustaining population and spread beyond the point of introduction [15,16], and indeed most introductions fail to do so . Introduced bird populations are thought to be highly vulnerable to extinction by stochastic factors, given their generally small population sizes [18,19]. In addition, captive bred birds may have reduced invasive ability, through loss of genetic diversity or key natural behaviours [20–22].
Australia boasts some of the strictest biosecurity and quarantine regulations in the world, particularly in regard to the trade and importation of wildlife [23,24]. However, detections of new non-native vertebrates at large (in the wild) have continued to increase . A recent review of vertebrate animal incursions in Australia reported over 227 bird incursion events detected by border and post-border biosecurity stakeholders during the 1990s . Birds are also the most frequently released taxa (escapes and thefts) from public zoos . Given that an estimated 13% of Australian households keep pet birds , private keeping may be the greatest existing source of bird introductions .
In the last decade, predominantly in higher income countries, different internet citizen initiatives have developed, particularly through social networks, with the aim of reuniting lost or stolen pets with their owners (e.g., https://www.parrotalert.com/; https://www.lostandfoundpetsaustralia.com). These resources provide relevant information on the identity and location of missing animals. We hypothesised that the distribution of bird escapes (i.e., unintentional pet bird losses) would be related to proxies of pet ownership; such as socio-economic characteristics of the community (Table 1). We used this information to predict the spatial distribution of the relative probability of bird escapes occurring across mainland Australia.
Materials and methods
We collated a unique dataset of accidental bird introductions in Australia using online reports of missing pet birds. We collected information on bird escapes from websites listed on the ‘Lost and Found’ section of the Australian Rescue and Rehoming Resource (ARRR) webpage (http://arrr.id.au/lostandfound.html; S1 Table), which contained reports for missing birds, for the period 1999–2013. From each report we extracted information on: (i) the identity of the species; (ii) the date and location of the report (including geographic coordinates); (iii) number of individuals missing (if known); and, (iv) the type of report (e.g., ‘Lost’, ‘Found’, or ‘Stolen’). We checked every report to assess the correct identification of the bird to the species level, or otherwise classified the bird as ‘Unknown’. We standardised the species scientific names according to Clements checklist of birds of the world Version 6.8 (http://www.birds.cornell.edu/clementschecklist/). Individual reports of bird escapes were cross-checked across all web pages to exclude duplicate reports of missing birds, and also in an endeavour to track the outcome of the escapees. On some occasions, birds reported initially as lost were later reported as found or even reunited with the owner. No personal data from the websites were recorded and all of the ‘Terms and Conditions’ of the individual sites were adhered to. The resulting data were analysed anonymously.
Economic data (average personal income level) and population age structure data were obtained from the Australian Bureau of Statistics. Previous research shows that non-native species richness is higher in areas of high human density . Therefore, we included, as covariates in our analysis, measures of human population distribution such as: (i) Human Influence Index (HII), a measure showing direct human impact on ecosystems using eight measures of human presence ; and (ii) land use type for each Australian territory. HII was extracted from the Global Human Dataset of the Last of the Wild Project . Land use data were obtained from the Australian Bureau of Agricultural and Resource Economics and Sciences website (see S1 Appendix for the complete details on data extraction, the list of references, and the corresponding links). All analyses were conducted in the R software environment for statistical computing and graphics version 3.03 . Geographic data for all the predictors were available at different resolutions, therefore all of the datasets were resampled to the lowest resolution available, 5km grid cells, using the nearest neighbour method with the R-packages ‘raster’  and ‘dismo’ . This resolution matches the scale to which missing bird report data were collected (i.e., suburb level) and captures accurate spatial heterogeneity of where birds are most likely to escape captivity.
Over 90% of all individual birds reported ‘missing’ were parrots (Order: Psittaciformes) (see Results). Consequently, we analysed factors that were predicted to be associated with the frequency of unintentional parrot losses (Table 2). We collected information of the traded parrots in Australia from the bird price guides published by the Avicultural Societies from: (i) South Australia (seven editions: 2003, 2005, 2006, 2008, 2009, 2012, 2013 ); (ii) Victoria (three editions: 2011, 2012, 2013 ); and (iii) the dataset completed from the classified advertisements section of the two most popular aviculturist magazines in Australia (Australian Birdkeeper and Australian Aviculture [45,46]). We obtained information for 126 traded parrot species.
We collected information on the following pet/keeper related variables for each parrot species: (i) Price (AU$), information on abundance of parrots in captivity is not readily available in Australia because there is no registry of pet ownership for bird species. For this reason we used species price as a proxy for abundance. We predicted cheaper species should be more frequently released, because we assumed they are more abundant in private captivity. In a recent paper, we demonstrated that price and abundance in private captivity are inversely correlated for bird species in the domestic Australian trade . Likewise, several lines of evidence point to the relationship between abundance in captivity and species price in the pet trade. Perception of rarity increases peoples’ willingness to pay high prices [58,59], therefore rarer species tend to be more expensive. The price for the escaped parrot species was obtained from the bird price guides. Price was estimated as the median, in Australian dollars, across all years and states, for a parent-reared pair comprising male and female of good quality and healthy condition. We considered only the price of the nominal subspecies because, for some species, there was considerable intraspecific variation in the price due to the presence in the trade of different subspecies or artificial breed colour mutations. Price was available for 107 species (84.9%); (ii) Body mass (g), body mass was estimated as the average body size of the adult male of the nominal subspecies. Body mass data was sourced from Dunning . This information was available for 125 species (99.2%); (iii) Longevity (years), for each species, longevity was measured as the maximum number of years an animal is known to have survived in captivity [61–64]. Longevity was available for 88 (69.8%) species; (iv) Docility, we measured whether the species possessed (or not) attractive behavioural traits for the keeper as follows . First, we reviewed information on parrot husbandry from the Birdcare webpage (http://birdcare.com.au/). The webpage extracts and summarizes information on bird husbandry from articles published in the Australian Aviculture and Australian Birdkeeper journals since 1947. For each of the parrot species we systematically checked the relevant information, searching for key words and terms that described the behavioural traits of the species in captivity related to docility. Second, we scored each species according to their descriptions as: (1) ‘Demanding’, if it was described as being harder to keep, for example, being shy, quiet, secretive, nervous, noisy, aggressive, prone to bite and/or requiring proper training or socialization; or (0) ‘Not demanding’, if described as active, playful, peaceful, and/or with the ability to learn to mimic and talk. This information was available for 116 species (92%); and (v) Native status of the species (native to Australia or not), the status of the species were defined according to the BirdLife Australia Working List of Australian Birds classification (http://birdlife.org.au/conservation/science/taxonomy; v1.1). The species were scored as ‘Non-native’ if their area of natural distribution did not include any part of the Australian territory, or ‘Native’ if otherwise. This information was available for all of the 126 parrot species. Price and longevity were loge transformed and body mass was log10 transformed for further analysis. See S2 Table for the complete parrot dataset.
Spatial distribution of bird escapes.
We used reports of missing birds for the period 2011–2013, including those in which the identity of the species was unknown. The majority of reports (> 85%) were from this period (see Results). We used generalised linear models (GLMs) to identify the factors associated with the spatial distribution of bird escapes in Australia. Escapes were recorded as a single presence data point. All reports provided information of the location, usually at the suburb level, and we used Google Maps to locate the geographic coordinates for each location. GLMs requires absence data but can be substituted by background data. Because no true-absence data are available for accidental escapes, a total of 10,000 random points were generated to be used as background data . Given that the chance of escape, or detecting an escapee, in uninhabited areas (e.g., desert) is arguably close to zero, we used HII to place the random pseudo-absences within a buffer around human inhabited areas (including towns, cities and major roads) . We used the lower scores of the HII to identify and exclude the areas of the Australian mainland without significant human impact (HII < 4; ). All areas outside the buffer were assumed to have a zero value, whereas all areas within the buffer were available for prediction. The buffer zone covered 61% of the surface of Australia. GLMs were fitted using the R package ‘dismo’. The occurrence of bird escapes (presence and pseudo-absence) were modelled with binomial variance and a logit link function. To avoid problems with model fitting, due to collinearity, we checked for correlated pairs of variables (Pearson’s r ≤ |0.7|). Collinearity amongst the variables was low (S3 Table). To balance model fit and predictive performance, the models were calculated using a 10-fold cross-validation procedure, which were run 50 times to ensure stable estimates of model evaluation statistics. A candidate set of models was built by including all possible combinations of explanatory variables to identify the models that provided the best support for the data, using the R-package ‘MuMIn’ . The relative support for each model structure was assessed by ranking models based on Akaike Information Criterion corrected for small sample size (AICc) and AICc weights (wAICc) . For the models with ΔAICc < 2 from the best model (lowest AICc) for each run, we calculated the model-averaged coefficients for the estimates and standard deviations (S4 Table). The distribution of the values of the estimates, and standard deviations for the GLMs, is presented as a coefficient plot.
Determinants of parrot escapes.
We used all reports of missing parrots, for the entire recorded period (1999–2013). Phylogenetic generalized least square models (PGLS, ) were fitted to analyse the relationship between the frequency of parrot escapes and the pet/keeper related variables using the ‘pgls’ function in the R package ‘caper’ [70,71]. We used PGLS models in order to account for the phylogenetic relatedness between the data points in the dependent variable, i.e. the species. This method addresses the issue of the ancestral relatedness by specifying a covariance matrix, which reflects the phylogenetic distances between the species. Each species is weighted according to the phylogenetic distance with the other species: the higher the correlation, the lower the weight given to that species. The PGLS model includes the estimated parameter Pagel’s lambda that controls the strength of the phylogenetic signal between the species . Although escaping captivity is not a biological trait itself, i.e. shaped by evolution, we used PGLS because the propensity for escape is likely to be correlated with species traits, such as longevity and body mass, that are known to have a strong phylogenetic signal .
We based our phylogenetic informed analyses on the phylogenetic tree for bird species proposed by Jetz et al. . The structure of this molecular-based tree is not known for certain, and so we incorporated uncertainty over the true phylogenetic relationship by repeating our analyses over a number of different phylogenetic trees for our species. One hundred avian phylogenetic trees were sampled at random online (http://birdtree.org), based on data from the complete avian phylogeny of Jetz et al.  and using the primary backbone tree of Hackett et al. . These trees were used as alternative phylogenetic hypotheses for the evolutionary relatedness of the parrot species. For the analysis we excluded two taxa (Barnardius barnadi and Pyrrhura roseifrons), which were considered subspecies by Jetz et al. .
Parrot escapes were measured as a binary variable where a parrot species reported to the websites was either scored as ‘Reported’ (N = 65), or ‘Unreported’ (N = 61). Frequency of parrot escapes was measured as the number of reports lodged in missing animal websites for each species. We analysed the relationship between the parrot pet/keeper related variables and the frequency of parrot escapes fitting independent PGLS models for each variable. The distribution of the values of the estimates and standard deviation for all generated models are presented as coefficient plots.
Bird escapes in Australia
We found reports for 5,139 bird escapes, corresponding to a total of 5,876 individuals from 91 bird species (six avian orders), for the period 1999–2013. For this period, the average number (± std err) of bird escapes was 120 ± 68 per month. Most reports of missing birds were from the period 2011–2013 (Fig 1). The majority of reports corresponded to lost birds (3,352 reports; 65.2%), followed by found birds (1,570 reports; 30.6%) and a small number of birds reported as stolen (217 reports; 4.2%). The average number of individuals reported per event was 1.14 ± 1.3 (N = 5,139). For stolen bird reports, the number of missing individuals reported was almost four times greater (3.8 ± 6; N = 217) than the average number reported for lost or found birds (1.04 ± 0.45; N = 5,103) (Fig 1). Missing birds were reported from urban or peri-urban areas, most of them concentrated within or around Australia’s seven major metropolitan regions, particularly along the east coast from Melbourne to Brisbane (Fig 2). This area accounted for 75% of all missing birds reported in Australia.
Presented as the number of missing bird reports per 60x60 km cell resolution. The Brisbane region accounted for 31.4% of all incursion events, followed by Sydney (29.9%) and Melbourne (16.1%).
Native and non-native bird species were represented in similar proportions. During the study period, 42 non-native (45.7%) and 49 Australian native species (53.3%) were reported. However, the number of reports for native species was more than twice (3,366 escapes; 65.5%) the number of reports for non-native species (1,548 escapes; 32.5%) (Table 3). Parrots were the most reported avian order (Psittaciformes); including 72% of reports and 93% of all individuals (Table 3). The most frequently reported bird species was the Cockatiel Nymphicus hollandicus, a native species, with 28.9% of all reports, 2.5 times more frequent than the second, the Rose-ringed Parakeet Psittacula krameri (11.1% of all reports), which was the most frequently reported non-native species (Table 4).
Spatial distribution of bird escapes
The performance of the model for predicting the spatial distribution of bird escapes was very high (average AIC [5th; 95th percentile] = 2,444.6[2,382.4; 2,501.7]; compared to the null model, AIC = 17,523). The distribution of bird escape reports was related to socio-economic factors of the community. The areas with higher average personal income level observed a greater frequency of reports. We found that bird escapes were more likely to occur in areas with higher levels of human modification to the environment. Bird escape reports were related to higher levels of HII and strongly positively related to intensive land uses (Fig 3 and S5 Table).
Determinants of parrot escapes
For the parrot species reported on missing animal websites (compared with all parrot species present in the Australian domestic bird trade) we found evidence for a relatively strong phylogenetic signal in the tendency for parrots to be reported as lost, found, or stolen (Estimated median lambda [5st, 95th percentiles] = 0.62 [0.55, 0.69]). We found a strong association between parrot escapes and the species price. Parrots reported on missing animal web sites were cheaper than species that have never been reported as escapees (Fig 4A). Similarly, parrot species that escaped most frequently were cheaper than species that less frequently escaped, or that did not escape at all (median t-value [5st, 95th percentiles] = -7.21 [-7.44, -7.02], df = 105, P < 0.001; Fig 4B). Parrot escapes were positively associated with the geographical origin and with the behaviour of the species as a pet (Fig 5). The parrot species escaping most frequently were mostly natives and had docile temperaments, i.e. species regarded as playful and peaceful (Fig 5 and S6 Table).
(a) Whether or not a parrot species has escaped captivity, and (b) the frequency of parrot escapes (non-natives and natives) in Australia.
Pet trade is the dominant transport pathway for non-native bird species globally , and it is estimated that four million live birds are traded annually . As a consequence, captivity is a major source of introduced bird species. In Australia, we have found that non-native bird introductions occur much more frequently than previously reported . The difference is not in the number of species (38 in Henderson & Bomford  c.f. 41 in this study), but in the number of introduction events and individuals: seven times more than previously recorded. When we included escaped natives, the numbers of reported species doubled and the number of introduction events and individuals increased by a factor of 22, with respect to previous estimates. The role of the pet keeping in the introduction pathway of vertebrates (particularly birds) has been largely overlooked [52,76]. The main reason is that animals reported to online public platforms are not generally recorded in the scientific literature or in the records of Australian biosecurity agencies. The difference in the proportion of reports for lost and found birds indicates that most birds that go missing are never found again, once in the wild. Indeed, detecting escaped birds appears to have a low probability. We suggest that only using reports of non-native birds seen at large might not be the best method for detecting emergent pests and quantifying incursions.
The expansion of the internet has changed how information is exchanged worldwide . These developments have produced new biosecurity risks [78,79], but also novel surveillance opportunities. In the case of escaped pet birds, the owners are able to communicate their losses with the local community of pet keepers in the hope of being reunited with their pets, and these records are available for tracking introduction events. The evidence collected from these online resources has revealed that hundreds of captive pet birds are unintentionally released from private households every month across the continent. We have no evidence to believe that the dramatic increase after 2010 is due to more birds escaping. Instead, we suggest that this is a likely consequence of the growing popularity, and participation, in social media and internet usage.
We found, as expected, that the distribution of bird escapes is strongly linked to the distribution of human population in Australia. Missing bird reports came from urban and peri-urban areas [37,38]. We found that the distribution of escaped birds is related to socio-economic factors of the community. Reported bird escapes were more frequent in areas with higher incomes. This supports our hypothesis that personal wealth could be related to higher pet ownership . Moreover, most areas with high average income correspond to highly inhabited areas. However, it appears that greater personal income is not related to the economic value of escaped species, because the species most frequently reported were, on average, less economically valued species.
The number of escaped bird species reported is low compared to the number of bird species known to be present in private captivity in Australia. Parrot species constitute only 35.9% of the traded bird species in Australia , yet lost parrots made up the majority of reports on the missing animal websites. We have no evidence to suggest that parrots escape more than other species. In public institutions (i.e. zoos), previous studies have shown that parrots were not more likely to escape, than other species . However, we suggest that parrots are more likely to be reported missing than other cagebird species. We do not know whether this is because parrots are more valued (sentimentally) than other bird groups, or because their relatively large body-size, bright colours and behaviour are considered more likely to be successfully found and returned. Alternately, finches and softbills (see S2 Appendix for definitions) are underrepresented in missing animal reports despite being abundant and widespread in private captivity in Australia . We suggest that missing individuals from these bird groups are probably less likely to be reported as lost, even for those with high economical value, because of lower chances of recovering the escaped birds. Regardless, if other bird species are escaping as frequently as parrots, and not being reported, then the pet trade in Australia has a very leaky aviary door .
Parrots species that were more frequently reported as escaped were cheaper, on average, than parrot species that were less frequently reported, or that were never reported as escaped. This result supports our hypothesis that less economically valued (more abundant) species are more likely to be deliberately or accidentally released . Previous research has found that popular species in the pet trade are introduced more frequently than rarer species [47,80,81]. Given that the complete confinement of birds in captivity can never be guaranteed, the more widely held a species is the greater the likelihood that an escape will occur . In addition, because abundance in private captivity for most bird species in Australia is unknown; we suggest that price can be used as a proxy to predict the captive pet bird species which have higher chances of being released, or escaping . Species price can also be related to abundance in the wild. Recent research on the illegal bird trade and parrot conservation showed that the species that are more abundant and accessible in the wild tend to be the cheaper, while rare species tend to be more expensive [83–85]. Perception of rarity is a factor driving extinctions of wild-caught traded species, particularly in countries where poaching for the pet trade is poorly controlled [83,86,87].
Australian native species were the most frequently escaped parrots. In a previous study we showed that native parrot species are almost as abundant in private captivity as non-natives . Recent studies also revealed that easily available species are more common in the bird markets [83,85], and Australia has a great diversity and abundance of native parrot species . This situation contrasts with bird trade in other Western countries like Europe , where most traded bird species are non-native, but it is more similar to Asian countries [9,10]. Native species are not always introduced in their natural range within the country , and this is particularly apparent for range restricted species. There is plenty of evidence of native bird species becoming pests and causing impacts just as significant as non-native species [90–92]. Some studies have suggested, that the so called ‘domestic exotics’ should be considered as biosecurity risks, alongside non-native species, outside their natural distribution .
The very high frequency of bird escapes suggests that the establishment of new non-native populations will eventually occur . However, despite the high rate of reported bird escapes in Australia, there is no evidence of the widespread risk of new ‘pet’ invaders. A small proportion of the escaped birds (30.6%) are found afterwards but most escapees disappear. Birds are generally kept in low numbers and, consequently, reported escapes involve a few individuals, and most often only a single individual (> 95% cases). It is possible that these escaped birds do not survive at large long enough to find mates and reproduce. In particular, captive birds may have lost some of the skills necessary to survive in nature [20,22]. For example, parrots intended to be sold as pets are usually socialized with humans in order for them to easily bond with their owners . Indeed, we found that more docile parrot species were more frequently reported on the missing animal websites as escapees. Owners may bond more easily with well-behaved species, therefore increasing their motivation for reporting the loss. Most recovered parrots are found in the vicinity of the release point, usually in bird feeders. The only circumstances in which the probability of establishment would be higher would be in the case of enough individuals released in one location, over a sufficient short period of time ; for example, repeated leaks from an unsecured aviary.
Our study has revealed that Australia has a genuine problem within domestic bird-keeping and onshore biosecurity reporting. The domestic pet trade constitutes an important pathway for the introduction of birds in Australia, and is of much greater risk than escapes from public holdings . However, unlike species kept in zoos and aquariums, measures for preventing the escapes and thefts from private holdings are non-existent or unenforced.
Banning the import of potential invasive wildlife has often been suggested as a preventive measure to mitigate the impact of invasive species [94,95]. Given that Australia already has a very reduced list of bird species that can be legally imported , we do not think further restrictions to the imports would affect the size of non-native species populations in captivity in Australia. Current domestic regulations on keeping non-native wildlife appear to be useful in preventing the introduction of some extreme risk species [96,97]. However, while regulating domestic breeding for these species (e.g. Rose-ringed Parakeet) may help to reduce population sizes available for escape, these kind of measures would certainly be very unpopular among the pet trade sector. International regulations to wildlife trade, such as CITES , have also been offered as a solution to stop over-harvesting of threatened species [99,100]. In a previous paper, we found that trade regulation at a national level had a greater effect on species price (and hence on abundance in captivity) in Australia than international bans . Finally, native Australian parrot wild populations are protected from international and domestic wildlife trade by biosecurity  and state and territory regulations . In general, native populations of Australian parrots are healthy and those species facing conservation issues have problems unrelated to harvesting for trade [102,103].
We suggest that bird reports from missing animal websites could be used as a passive surveillance system to monitor accidental bird introductions and to detect areas of high and frequent non-independent escapes. We advocate for intervention in situations that pose higher risk of establishment, such as when breeding activity in the wild is detected. We strongly recommend that introductions of the most frequently escaped non-native species should be monitored to prevent any future risk of invasion. For example, the Rose-ringed Parakeet is the most frequently escaping non-native bird species in Australia (260 releases only in 2013). The parakeet is considered an extreme risk species , and has previously been known to breed at large . The risk posed by the parakeet is clearly worthy of additional study .
Internet surveillance is a novel tool, based on public participation, but it also has its limitations. For example, there is great variability in individual people’s internet use, which we would expect to affect reporting. Globally, there is uneven access to the Internet, particularly in low-income countries, where many people lack access and access may be restricted to large cities (http://data.worldbank.org/indicator/IT.NET.USER.P2?view=map). Finally, we detected that there is a component of reporting bias regarding escaped species and there is still the likelihood that introductions will occur unnoticed.
Even with these limitations, we believe that reports to missing animal websites are a useful record of accidental pet bird introductions. Escaped parrots, in particular, may pose a greater risk to become new pest species. A better registry of bird keeping, and monitoring of escapes, is required to track incursions and to inform biosecurity agencies of the risk posed by the escapes of species present in the animal pet trade.
S1 Table. Complete list of consulted webpages.
List obtained from the Lost and Found section of the Australian Rescue and Rehoming Resource webpage. All webpages were last accessed in December 2013.
S2 Table. Complete parrot dataset.
NA means that the data was not available.
S3 Table. Pearson’s r correlation coefficient for predictors of bird escapes in Australia (N = 12,766).
Only continuous variables.
S4 Table. Summary of the components of the ΔAICc < 2 models.
Components were calculated using a 10-fold cross-validation and 50 runs. For each run, 32 models (combinations of the 5 variables) were compared.Variable codes are: 1. Human Influence Index; 2. Land uses; 3. Average personal income; 4. Elderly population; and 5. Children population.
S5 Table. Model coefficients (median estimates, [5th, 95th percentiles]) in GLMs for predictors of spatial distribution of bird escapes in Australia.
Coefficients were calculated using a 10-fold cross-validation and 50 runs.
S6 Table. Model coefficients (median estimates, [5th, 95th percentiles]) in PGLS models for determinants of parrot escapes.
Coefficients were calculated over 100 randomly selected likely phylogenetic trees. 'Estimate' means the 'lambda' value of the PGLS model.
S1 Appendix. Predictors of the spatial distribution of accidental bird escapes in Australia.
We thank the creators and contributors to Australian Rescue and Rehoming Resource for providing the data on missing pet birds in Australia, A. Woolnough and P. García-Díaz for conversations and advice on earlier drafts of the manuscript, S. Shan for advice on missing pets websites in South East Asian countries, T. Wittmann for help with copy editing and, finally, the academic editor and two anonymous reviewers for help with revising the manuscript.
- Conceptualization: MV PC.
- Data curation: MV.
- Formal analysis: MV.
- Funding acquisition: PC.
- Investigation: MV PC.
- Methodology: MV PC.
- Project administration: PC.
- Resources: PC.
- Supervision: PC.
- Visualization: MV PC.
- Writing – original draft: MV.
- Writing – review & editing: MV PC.
- 1. Broad S, Mulliken T, Roe D. The nature and extent of legal and illegal trade in wildlife. In: Oldfield S, editor. The trade in wildlife. Regulation for conservation. London and Sterling, VA: Earthscan Publications Ltd; 2003. p. 210.
- 2. Bush ER, Baker SE, Macdonald DW. Global trade in exotic pets 2006–2012. Conserv Biol. 2014;28(3):663–676. pmid:24661260
- 3. Romagosa CM. Live animal trade as vector for invasions. In: Canning-Clode J, editor. Biological Invasions in Changing Ecosystems. Vectors, Ecological Impacts, Management and Predictions. De Gruyter Open; 2015.
- 4. Hulme PE. Trade, transport and trouble: managing invasive species pathways in an era of globalization. J Appl Ecol. 2009;46(1):10–18.
- 5. Goss JR, Cumming GS. Networks of wildlife translocations in developing countries: an emerging conservation issue? Front Ecol Environ. 2013;11(5):243–250.
- 6. Robinson JM. The dynamics of avicultural markets. Environ Conserv. 2001;28(1):76–85.
- 7. Meenken D. Pet Biosecurity in New Zealand. Current state of the domestic pet trade system and options going forward. Wellington, New Zealand: Ministry for Primary Industries; 2012.
- 8. Padilla DK, Williams SL. Beyond ballast water: aquarium and ornamental trades as sources of invasive species in aquatic ecosystems. Front Ecol Environ. 2004;2(3):131–138.
- 9. Su S, Cassey P, Blackburn TM. Patterns of non-randomness in the composition and characteristics of the Taiwanese bird trade. Biol Invasions. 2014;16(12):2563–2575.
- 10. Su S, Cassey P, Vall-llosera M, Blackburn TM. Going cheap: Determinants of bird price in the Taiwanese pet market. PLoS One. 2015; 10(5): e0127482. pmid:26017386
- 11. Green SJ, Akins JL, Maljković A, Côté IM. Invasive lionfish drive Atlantic coral reef fish declines. PLoS One. 2012;7(3):e32596. pmid:22412895
- 12. Cadi A, Joly P. Impact of the introduction of the red-eared slider (Trachemys scripta elegans) on survival rates of the European pond turtle (Emys orbicularis). Biodivers Conserv. 2004;13(3):2511–2518.
- 13. Medina FM, Bonnaud E, Vidal E, Tershy BR, Zavaleta ES, Donlan CJ, et al. A global review of the impacts of invasive cats on island endangered vertebrates. Glob Chang Biol. 2011;17(11):3503–3510.
- 14. Baker J, Harvey KJ, French K. Threats from introduced birds to native birds. Emu. 2014;114(1):1–12.
- 15. Duncan RP, Blackburn TM, Sol D. The ecology of bird introductions. Annu Rev Ecol Evol Syst. 2003;34(1):71–98.
- 16. Kolar CS, Lodge DM. Progress in invasion biology: predicting invaders. Trends Ecol Evol. 2001;16(4):199–204. pmid:11245943
- 17. Williamson M, Fitter A. The varying success of invaders. Ecology. 1996;77(6):1661–1666.
- 18. Cassey P, Blackburn TM, Sol D, Duncan RP, Lockwood JL. Global patterns of introduction effort and establishment success in birds. Proc R Soc B-Biological Sci. 2004;271(S6):S405–S408.
- 19. Lockwood JL, Cassey P, Blackburn T. The role of propagule pressure in explaining species invasions. Trends Ecol Evol. 2005;20(5):223–228. pmid:16701373
- 20. Carrete M, Tella J. Wild-bird trade and exotic invasions: a new link of conservation concern? Front Ecol Environ. 2008;6(4):207–211.
- 21. Mason G, Burn CC, Ahloy J, Kroshko J, Mcdonald H, Jeschke JM. Plastic animals in cages: behavioural flexibility and responses to captivity. Anim Behav. Elsevier Ltd; 2013;85(5):1113–1126.
- 22. Cabezas S, Carrete M, Tella JL, Marchant TA, Bortolotti GR. Differences in acute stress responses between wild-caught and captive-bred birds: a physiological mechanism contributing to current avian invasions? Biol Invasions. 2013;15(3):521–527.
- 23. Alacs E, Georges A. Wildlife across our borders: a review of the illegal trade in Australia. Aust J Forensic Sci. 2008;40(2):147–160.
- 24. Department of Environment. List of Specimens taken to be Suitable for Live Import. Australia: Australian Government. Environment Protection and Biodiversity Conservation Act 1999; 2001.
- 25. Henderson W, Bomford M. Detecting and preventing new incursions of exotic animals in Australia. Canberra, Australia: Invasive Animals Cooperative Research Centre; 2011.
- 26. Henderson W, Bomford M, Cassey P. Managing the risk of exotic vertebrate incursions in Australia. Wildl Res. 2011;38(6):501–508.
- 27. Cassey P, Hogg C. Escaping captivity: the biological invasion risk from vertebrate species in zoos. Biol Conserv. 2015;181(1):18–26.
- 28. Animal Health Aliance. Pet Ownership in Australia. Canberra: The Animal Health Alliance (Australia) Ltd, Cox Inall Communications Pty Ltd; 2013. 92 p.
- 29. Chiron F, Shirley SM, Kark S. Behind the Iron Curtain: socio-economic and political factors shaped exotic bird introductions into Europe. Biol Conserv. Elsevier Ltd; 2010;143(2):351–356.
- 30. Jepson P, Ladle RJ. Bird-keeping in Indonesia: conservation impacts and the potential for substitution-based conservation responses. Oryx. 2005;39(4):1–6.
- 31. Trela JE. Age structure of voluntary associations and political self-interest among the aged. Sociol Q. 1972;13(2):244–252.
- 32. Sable P. Pets, attachment, and well-being across the life cycle. Soc Work. 1995;40(3):334–341. pmid:7761919
- 33. Albert A, Bulcroft K. Pets, families, and the life course. J Marriage Fam. 1988;50(2):543–552.
- 34. Vidović VV, Stetić VV, Bratko D. Pet ownership, type of pet and socio-emotional development of school children. Anthrozoos. 1999;12(4):211–217.
- 35. Williams JM, Muldoon J. Children and their pets: Exploring the relationships between pet ownership, pet attitudes, attachment to pets and empathy. Educ Heal. 2010;28(1):12–15.
- 36. Kidd AH, Kidd RM. Problem and benefits of bird ownership. Psychol Rep. 1998;83(1):131–138.
- 37. Chiron F, Shirley S, Kark S. Human-related processes drive the richness of exotic birds in Europe. Proc R Soc B Biol Sci. 2009;276(1654):47–53.
- 38. Strubbe D, Matthysen E. Establishment success of invasive ring-necked and monk parakeets in Europe. J Biogeogr. 2009;36(12):2264–2278.
- 39. Sanderson EW, Jaiteh M, Levy MA, Redford KH, Wannebo AV, Woolmer G. The Human Footprint and the Last of the Wild. Bioscience. 2002;52(10):891–904.
- 40. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. Vienna, Austria: R Foundation for Statistical Computing; 2015.
- 41. Hijmans RJ, Etten J van, Mattiuzzi M, Sumner M, Greenberg JA, Lamigueiro OP, et al. raster: Geographic data analysis and modeling. R package version 2.2–31. http://CRAN.R-project.org/package=raster. 2014.
- 42. Hijmans RJ, Phillips S, Leathwick J, Elith J. dismo: Species distribution modeling. R package version 0.9–3. http://CRAN.R-project.org/package=dismo. 2013.
- 43. The United Bird Societies of South Australia Inc. Bird Price Guide. 1st to 7th edition. 2003–2013.
- 44. The Avicultural Society of Australia Inc. Guide to Bird Prices. 2011–2012 to 2013–2014. Australian Aviculture. 2012–2014.
- 45. Australian Birdkeeper. Classifieds. July 2011 to November 2013. Australian Birdkeeper. 2011–2013.
- 46. The Avicultural Society of Australia Inc. Member’s Notices Birds for Sale & Wanted to Buy. April 2011 to November 2013. Australian Aviculture. 2011–2013.
- 47. Duggan IC, Rixon CAM, Macisaac HJ. Popularity and propagule pressure: determinants of introduction and establishment of aquarium fish. Biol Invasions. 2006;8(2):377–382.
- 48. McGill BJ. A test of the unified neutral theory of biodiversity. Nature. 2003;422(6934):881–885. pmid:12692564
- 49. Clubb SL. Captive management of birds for a lifetime. J Am Vet Med Assoc. 1998;212(8):1243–1245. pmid:9569162
- 50. Anderson PK. A bird in the house: An anthropological perspective on companion parrots. Soc Anim. 2003;11(4):393–418.
- 51. Engebretson M. The welfare and suitability of parrots as companion animals: a review. Anim Welf. 2006;15:263–276.
- 52. Cassey P, Vall-llosera M, Dyer E, Blackburn TM. The biogeography of avian invasions: history, accident and market trade. In: Canning-Clode J, editor. Biological Invasions in Changing Ecosystems Vectors, Ecological Impacts, Management and Predictions. Open: De Gruyter; 2015.
- 53. Ballouard JM, Brischoux F, Bonnet X. Children prioritize virtual exotic biodiversity over local biodiversity. PLoS One. 2011;6(8):e23152. pmid:21829710
- 54. Boehrer BT. Parrot Culture: Our 2500-Year-Long Fascination with the World’s Most Talkative Bird. Philadelphia, PA, USA: University of Pennsylvania Press; 2010.
- 55. Carrete M, Edelaar P, Blas J, Serrano D, Potti J, Dingemanse NJ, et al. Don’t neglect pre-establishment individual selection in deliberate introductions. Trends Ecol Evol Evol. 2012;27(2):67.
- 56. McCowan LSC, Rollins LA, Griffith SC. Personality in captivity: more exploratory males reproduce better in an aviary population. Behav Processes. 2014;107:150–157. pmid:25173618
- 57. Vall-llosera M, Cassey P. Physical attractiveness, constraints to the trade and handling requirements drive the variation in species availability in the Australian cagebird trade. Ecol Econ. 2017;131:407–413.
- 58. Angulo E, Courchamp F. Rare species are valued big time. PLoS One. 2009;4(4):e5215. pmid:19384416
- 59. Angulo E, Deves A-L, Saint Jalmes M, Courchamp F. Fatal attraction: rare species in the spotlight. Proc R Soc B-Biological Sci. 2009;276(1660):1331–1337.
- 60. Dunning JB. CRC Handbook of Avian Body Masses. Boca Raton, Florida: CRC Press; 2007.
- 61. Collar NJ. Family Psittacidae (Parrots). In: Del Hoyo J, Elliott A, Sargatal J, editors. Handbook of the Birds of the World (Volume 4): Sandgrouse to Cuckoos. Barcelona: Lynx Editions; 1997. p. 280–339.
- 62. Brouwer K, Jones ML, King CE, Schifter H. Longevity records for Psittaciformes in captivity. Int Zoo Yearb. 2000;37(1):299–316.
- 63. Juniper T, Parr M. Parrots: A Guide to the Parrots of the World. East Sussex, UK: Pica Press; 1998.
- 64. Higgins PJ. Handbook of Australian, New Zealand and Antarctic Birds. Volume 4: Parrots to Dollarbird. Melbourne, Australia: Oxford Univeristy Press; 1999.
- 65. Phillips SJ, Dudík M, Elith J, Graham CH, Lehmann A, Leathwick J, et al. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecol Appl. 2009;19(1):181–197. pmid:19323182
- 66. Vall-llosera M, Woolnough A, Anderson D, Cassey P. Improved surveillance for early detection of a potential invasive species: the alien Rose-ringed parakeet Psittacula krameri in Australia.
- 67. Barton K. MuMIn: Multi-model inference. R package version 1.9.13. http://CRAN.R-project.org/package=MuMIn. 2013.
- 68. Anderson DR, Burnham KP. Avoiding pitfalls when using information-theoretic methods. J Wildl Manage. 2002;66(3):912–918.
- 69. Ives AR, Garland TJ. Phylogenetic logistic regression for binary dependent variables. Syst Biol. 2010;59(1):9–26. pmid:20525617
- 70. Orme D, Freckleton R, Thomas G, Petzoldt T, Isaac N, Pearse W. caper: Comparative Analyses of Phylogenetics and Evolution in R. R package version 0.5.2. http://CRAN.R-project.org/package=caper. 2015.
- 71. Ho LST, Ane C. A linear-time algorithm for Gaussian and non-Gaussian trait evolution models. Syst Biol. 2014;63(3):1–38.
- 72. Pagel M. Inferring the historical patterns of biological evolution. Nature. 1999;401(6756):877–884. pmid:10553904
- 73. Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO. The global diversity of birds in space and time. Nature. 2012;491(7424):444–448. pmid:23123857
- 74. Hackett SJ, Kimball RT, Reddy S, Bowie RCK, Braun EL, Braun MJ, et al. A phylogenomic study of birds reveals their evolutionary history. Science. 2008;320(5884):1763–1768. pmid:18583609
- 75. Karesh WB, Cook RA, Bennett EL, Newcomb J. Wildlife trade and global disease emergence. Emerg Infect Dis. 2005;11(7):1000–1002. pmid:16022772
- 76. Blackburn TM, Lockwood JL, Cassey P. Avian Invasions. The Ecology & Evolution of Exotic Birds. Oxford Avian Biology Series. Oxford: Oxford University Press; 2009. 305 p.
- 77. Haythornthwaite C. Social networks and Internet connectivity effects. Information, Commun Soc. 2005;8(2):125–147.
- 78. Derraik JGB, Phillips S. Online trade poses a threat to biosecurity in New Zealand. Biol Invasions. 2009;12(6):1477–1480.
- 79. Kikillus KH, Hare K, Hartley S. Online trading tools as a method of estimating propagule pressure via the pet-release pathway. Biol Invasions. 2012;14(12):2657–2664.
- 80. García-Díaz P, Ross JV, Ayres C, Cassey P. Understanding the biological invasion risk posed by the global wildlife trade: propagule pressure drives the introduction and establishment of Nearctic turtles. Glob Chang Biol. 2015;21(3):1078–1091. pmid:25363272
- 81. Van Wilgen NJ, Wilson JRU, Elith J, Wintle BA, Richardson DM. Alien invaders and reptile traders: what drives the live animal trade in South Africa? Anim Conserv. 2010;13:24–32.
- 82. Hulme PE. Invasion pathways at a crossroad: policy & research challenges for managing alien species introductions. J Appl Ecol. 2015;52(6):1418–1424.
- 83. Pires SF. The illegal parrot trade: a literature review. Glob Crime. 2012;13(3):176–190.
- 84. Pires SF. The heterogeneity of illicit parrot markets: an analysis of seven Neo-tropical open-air markets. Eur J Crim Policy Res. 2015;21(1):151–166.
- 85. Olah G, Butchar SHM, Symes A, Guzmán IM, Cunningham R, Brightsmith DJ, et al. Ecological and socio-economic factors affecting extinction risk in parrots. Biodivers Conserv. 2016;25(2):205–223.
- 86. Courchamp F, Angulo E, Rivalan P, Hall RJ, Signoret L, Bull L, et al. Rarity value and species extinction: the anthropogenic Allee effect. PLoS Biol. 2006;4(12):e415. pmid:17132047
- 87. Marsden SJ, Royle K. Abundance and abundance change in the world’s parrots. Ibis. 2015;157(2):219–229.
- 88. Abellán P, Carrete M, Anadón JD, Cardador L, Tella JL. Non-random patterns and temporal trends (1912–2012) in the transport, introduction and establishment of exotic birds in Spain and Portugal. Divers Distrib. 2015.
- 89. Guo Q, Ricklefs RE. Domestic exotics and the perception of invasibility. Divers Distrib. 2010;16(6):1034–1039.
- 90. Chapman T, Massam M. Rainbow lorikeet. Note 200. Pest Note. Department of Agriculture and Food and Department of Environment and Conservation. State of Western Australia; 2007.
- 91. Tracey J, Bomford M, Hart Q, Saunders G, Sinclair R. Managing Bird Damage to Fruit and Other Horticultural Crops. Canberra: Bureau of Rural Sciences; 2007.
- 92. Bomford M, Sinclair R. Australian research on bird pests: impact, management and future directions. Emu. 2002;102(1):29–45.
- 93. Lockwood JL, Cassey P, Blackburn TM. The more you introduce the more you get: the role of colonization pressure and propagule pressure in invasion ecology. Divers Distrib. 2009;15(5):904–910.
- 94. Vall-llosera M, Sol D. A global risk assessment for the success of bird introductions. J Appl Ecol. 2009;46(4):787–795.
- 95. Keller RP, Lodge DM, Finnoff DC. Risk assessment for invasive species produces net bioeconomic benefits. Proc Natl Acad Sci USA. 2007;104(1):203–207. pmid:17190819
- 96. DPIPWE. Importing and Keeping Birds in Tasmania: List of Available Species and Import Guide. Department of Primary Industries, Parks, Water and Environment. Resource Management and Conservation Division; 2011.
- 97. Kirkpatrick W. Importing and keeping introduced mammals, birds, reptiles and amphibians in Western Australia. Perth; 2013.
- 98. Poole CM, Shepherd CR. Shades of grey: the legal trade in CITES-listed birds in Singapore, notably the globally threatened African grey parrot Psittacus erithacus. Oryx. 2016.
- 99. Cooney R, Jepson P. The international wild bird trade: what’s wrong with blanket bans? Oryx. 2005;40(1):18.
- 100. Pires SF, Moreto WD. Preventing wildlife crimes: solutions that can overcome the “Tragedy of the Commons.” Eur J Crim Policy Res. 2011;17(2):101–123.
- 101. Wilson A. Exotic Pest Birds. Current Management. State of Queensland, Department of Employment, Economic Development and Innovation; 2011. 60 p.
- 102. Drechsler M. Spatial conservation management of the orange-bellied parrot Neophema chrysogaster. Biol Conserv. 1998;84(3):283–292.
- 103. Snyder N, McGowan PJK, Gilardi J, Grajal A. Parrots: Status Survey and Conservation Action Plan 2000–2004. Gland, Switzerland and Cambridge, UK: International Union for Conservation of Nature and Natural Resources; 2000. 180 p.
- 104. DAFWA. Animal Pest Alert: Indian Ringneck Parakeet. Department of Agriculture and Food, Western Australia. Government of Western Australia; 2007.