Despite its social and economic benefits, the trade in ornamental species (henceforth, ‘ornamental trade') has become a major source of non-native fish introductions into freshwater ecosystems. However, the ornamental trade as a vector for introductions of non-native freshwater fishes is not well defined. We developed a framework incorporating elements of the biological invasion process and a typical ornamental fish trade supply chain to fill this gap. Records of non-native ornamental fishes introduced to freshwater environments of Australia, Belgium (Flanders), Canada (British Columbia), China (Guangdong), the Philippines, Poland, Singapore, the United Kingdom (England), and the United States of America (Florida) were reviewed to explore the pervasiveness of these introduced fishes in the wild. These regional case studies confirmed the prominence of the ornamental trade as a global vector for freshwater fish introductions beyond their natural range. Additionally, we examined freshwater fishes associated with the ornamental trade to identify ‘risky' species that could establish in recipient regions based on climate match. All regions assessed were at risk of new fish introductions via the ornamental trade, with the number of ‘risky’ species ranging from seven to 256. Further, there appears to be taxonomic bias in the freshwater ornamental fish trade, with 74% of the species belonging to just 10 families (of 67). Current prevention and management approaches and associated polices, regulations and legislation on aquatic non-native species within assessed regions fit five general categories: import controls, risk assessment, whitelist, blacklist, and release ban. However, these prevention/management efforts may not be sufficient to reduce the invasion risk associated with the ornamental fish trade. Recommendations including species- and vector-based risk assessments, better recording of species import consignments, increased public education and industry engagement, and early detection and rapid response are discussed in this review.
Human activities such as commercial shipping (Carlton, 1985), aquaculture (Naylor et al., 2001), and live animal trade (Weigle et al., 2005) are transforming natural ecosystems by transporting species across natural geographic barriers to habitats beyond their native ranges. Whereas some non-native (NN) species are considered beneficial (e.g. aquaculture species in some contexts), some others have detrimental effects on recipient environments, including hybridisation with resident populations, local extirpations and global extinctions of native species, disruption of ecosystem function, enhanced transmission of viruses and pathogens, and substantial damage to natural resources (e.g. Ricciardi and MacIsaac, 2011; Paolucci et al., 2013; Simberloff et al., 2013). Indeed, biological invasion is considered one of the leading threats to biodiversity (CBD, 2002), especially within freshwater ecosystems (Sala et al., 2000).
To understand biological invasions better, considerable research and management efforts have been directed to characterise major human-mediated vectors by which NN species are introduced (Hulme, 2015). Ship ballast water, which is a leading vector for introductions of aquatic NN species worldwide, has received extensive study, including biological surveys, efficacy of management strategies, new tools or methods for analysis of ballast water samples, monitoring in shipping ports, and risk assessments (Bailey, 2015). The International Maritime Organization (IMO) adopted the International Convention for the Control and Management of Ships' Ballast Water and Sediments to reduce the transfer of harmful aquatic organisms and pathogens, including invasive species (IMO, 2019). Many countries have also implemented regulations to manage the invasion risk associated with ballast water (e.g. Commonwealth of Australia, 2017; Government of Singapore, 2017; Government of Canada, 2019a). In contrast, the ornamental trade has received limited examination as a vector for aquatic species introductions (e.g. Copp et al. 2007; Chang et al., 2009). In fact, the industry remains largely unregulated or regulated but unenforced (Raghavan et al. 2013; Allen et al., 2017), with some exceptions (e.g. England and Wales for temperate species, Copp et al., 2005b; Florida and some other states in U.S.A., Hill, 2016, Tuckett et al., 2016; New Zealand, Duggan, 2011).
The trade in ornamental fishes, which is a multi-billion-dollar industry of generally strong and steady global growth in recent decades (Padilla and Williams, 2004; Monticini, 2010; Dey, 2016), includes captive and reared fishes, invertebrates, plants, accessories, aquaria, feeds, and drugs. The collection, breeding, import, and export of ornamental fishes involve at least 125 countries and ≈ 5,300 and 1,800 freshwater and marine fishes, respectively (Dey, 2016). Over two billion live ornamental fishes are moved worldwide each year (Monticini, 2010), with freshwater fishes from breeding facilities representing about 90% of the total trade (Evers et al., 2019). While this provides employment opportunities, raises public awareness on biodiversity and conservation issues, and even promotes human health through stress reduction (Helfman, 2007), there is increasing evidence that the ornamental fish trade is a major vector for NN fish introductions and translocations (e.g. Padilla and Williams, 2004; Gertzen et al., 2008; Chang et al., 2009). The accidental escape of cultivated species from fish farms, e.g. via the fish pond drainage and overflow during flood/spate events, is not uncommon (Courtenay and Stauffer, 1990; Naylor et al., 2001; Helfman, 2007) and intentional dumping of unwanted fishes can occur due to undesirable qualities and overproduction (Helfman, 2007). Ornamental fishes are often deliberately released into the environment from aquaria as a ‘humane’ disposal method of unwanted pets – a phenomenon observed in many regions of the world (e.g. Courtenay, 1999; Copp et al., 2005c; Gertzen et al., 2008). Some cultural and religious groups also release ornamental species, including freshwater fishes, as part of their rituals, i.e. mercy, prayer, or religious release (Crossman and Cudmore, 1999; Wasserman et al., 2019). Some species cause considerable negative impacts on the recipient ecosystems, with a number of freshwater ornamental fishes, including Walking Catfish (Clarias batrachus), common carp/koi (Cyprinus carpio), western mosquitofish (Gambusia affinis), and Largemouth (black) Bass (Micropterus salmoides) currently listed by the International Union for Conservation of Nature (IUCN) among the 100 world's worst invasive species (GISD, 2019), whereas others such as the topmouth gudgeon (Pseudorasbora parva) are listed as the worst invasive species in Europe (Nentwig et al., 2018).
The magnitude of the ornamental fish trade as a vector for species introductions is clearly of considerable concern (Copp et al., 2005c; Chang et al., 2009; Strecker et al., 2011). Here, we present a framework developed to gain a better understanding of the invasion process of ornamental fishes. We also provide a brief overview of NN ornamental fish introductions to freshwater ecosystems in selected regions of the world to demonstrate the importance of the ornamental trade as a global introduction vector. In addition, we identify which of the freshwater fish species currently in the ornamental trade pose a high risk of establishing NN populations in the assessed regions based on climate matching. Finally, we summarise current management efforts regarding introduced ornamental fishes by examining policies, regulations, and legislation regarding these species for each region.
Invasion process of non-native ornamental fishes
The global movement of ornamental fishes and their subsequent release into the environment may be viewed as a stage-based process, as with other biological invasions (see Blackburn et al., 2011). To help conceptualise the process, our framework comprises elements of the unified biological invasion framework of Blackburn et al. (2011) and a typical ornamental fish trade supply chain (Monticini, 2010) (Figure 1). The invasion process is broken into the transport, introduction, establishment, and spread stages. Barriers that may preclude a species from progressing to subsequent stages of the invasion process include: geography; domestication, cultivation, or captivity; survival and reproduction; and dispersal and environmental factors. The supply chain for ornamental fishes is complex and varies by region (Duggan, 2011). It typically involves harvesters, breeders, wholesalers, exporters, importers, retailers, and aquarists, moving fishes from their native habitats to new environments (Monticini, 2010). For simplicity, we consolidated exporters and importers under wholesalers to represent the movement of species from one region to another. A successful NN fish species must first circumvent the geographic barrier by being imported into a new region via the ornamental trade (transport stage) – this may represent the movement between the species' native habitat and the wholesaler if the species is collected or cultured in its native range. Alternatively, the species may be transferred from their native habitat to a culture facility outside of its native range. The latter scenario is more common (90% of the time) for freshwater species (Monticini, 2010). Entry to the recipient environment (introduction stage) can involve escape from the domestication/cultivation/captivity or deliberate release. Once introduced to the recipient environment, the species must survive and create a self-sustaining population (establishment stage), followed by dispersal of the species from the original introduction site(s) and the wider establishment of new populations (spread stage).
The likelihood of a taxon/population overcoming the barriers at each stage of the invasion process is determined by various factors (Colautti and MacIsaac, 2004), including propagule pressure, physical and chemical factors, community interactions, and/or species characteristics. Propagule pressure is a combined measure of the number of introduction events and the number of individuals released per event (Lockwood et al., 2005), with increasing numbers of events and individuals enhancing the probability of population establishment by decreasing environmental and demographic stochasticity, respectively (Simberloff, 2009). For freshwater ornamental NN fishes, a number of human demographic factors may be involved, including the intensity and diversity of fish imports (Copp et al., 2007), human population density (Copp et al., 2010a), popularity of fishes measured as the frequency of occurrence in pet stores (Duggan et al., 2006), and the distance of the recipient pond from the nearest road (Copp et al., 2005c). The physical and chemical properties of the transport process and the recipient environment may enhance or impede invasion success, depending on the physical and chemical requirements of the species at each life history stage (Colautti and MacIsaac, 2004; Briski et al., 2018). Transport conditions of ornamental species are optimised to minimise mortality because deaths decrease profits (Duggan, 2011). Community interactions such as competition, predation, and parasitism can hinder the transition of individuals from introduction to establishment, whereas facilitative interactions such as mutualism can increase the likelihood of invasion success (Simberloff and Von Holle, 1999; Colautti and MacIsaac, 2004).
Species biological traits such as morphology, behaviour, and ecophysiology can influence a species' propensity to transit to the next stage of the invasion process, and different sets of characteristics may be important for different stages (Kolar and Lodge, 2002; Marchetti et al., 2004; Van Kleunen et al., 2010). Ornamental species are often selectively bred for desired physical features such as colouration, body shape, and fin size and shape (Singh et al., 2010), enhancing their likelihood of moving through the supply chain to a new region (Duggan, 2011). Ornamental fish release into the environment, however, may be due to undesired characteristics, e.g. large body size, high reproductive output, and aggressive behaviour (e.g. Courtenay, 1999; Copp et al., 2005a; Gertzen et al., 2008). Greater longevity and cheaper retail price may also increase the probability of release as observed for pet reptiles and amphibians (Stringham and Lockwood, 2018). Once released into the wild, characteristics selected for ornamental purposes may affect the establishment, spread, and susceptibility to predation of released fishes (Duggan, 2011). For example, bright colour patterns of ornamental varieties are conspicuous and can be subjected to high predation risk (Copp et al., 2005c; Maan et al., 2008). Furthermore, the behaviour and physiology of ornamental fishes can be altered, often unintentionally, via the domestication and/or cultivation process, allowing them to cope better with stressors in novel environments (Copp et al., 2005a). For instance, domesticated fighting fishes (Betta spp.) are generally more aggressive than their wild counterparts (Verbeek et al., 2007), exhibiting lower cortisol levels in unfamiliar environments relative to wild types (Verbeek et al., 2008). This is consistent with human-induced adaptation to invade by which populations that are predisposed to human-altered habitats (e.g. strongly competitive due to confinement at high densities) are more likely to be successful in similarly human-altered recipient habitats (Hufbauer et al., 2012). In addition, characteristics including r-selected traits, broad physiological tolerance, high genetic variability, and high phenotypic plasticity are some common attributes of successful NN fishes (Kolar and Lodge, 2002; Moyle and Marchetti, 2006; Howeth et al., 2016), and are likely to be important for the invasion success of ornamental fishes.
Ornamental trade as a global vector for non-native freshwater fishes: how important is it really?
We reviewed the primary literature and reports to compile records of introduced fishes that could be attributed to the ornamental trade in fresh waters of Australia, Belgium (Flanders), Canada (British Columbia; B.C.), China (Guangdong), the Philippines, Poland, Singapore, the United Kingdom (U.K.; England), and the United States of America (U.S.A.; Florida). We selected these regions of the world based on knowledge of regional case studies on introduced freshwater ornamental fishes. While these nine regions are not exhaustive, they should be representative of the issue globally as they span multiple climate zones, countries (developed vs developing), jurisdictions, and socio-cultural values regarding ornamental fishes. The overview includes the statistics of the ornamental trade (when data were available) and examples of introduced ornamental fishes for each region. In some cases, we highlighted specific species to illustrate the impacts of NN ornamental fishes on the recipient ecosystems.
Over 1,100 freshwater fish species have been imported to Australia via the ornamental trade since the 1960s (McNee, 2002) of which more than 30 species have formed self-sustaining populations throughout the continent (Arthington and Mckenzie, 1997; García-Díaz et al., 2018). The number, distribution, and abundance of ornamental fishes in natural waterways in Australia has increased markedly, with ornamental fishes accounting for ≈ 80% of all NN fishes (Arthington and Mckenzie, 1997; Beatty and Morgan, 2013; Harris, 2013; García-Díaz et al., 2018). Indeed, the ornamental trade has become the leading vector for NN fish introductions since the 1970s (Harris, 2013; García-Díaz et al., 2018), with species established in Australian drainages including Oriental weatherfish (Misgurnus anguillicaudatus), goldfish (Carassius auratus), green swordtail (Xiphophorus hellerii), southern platyfish (Xiphophorus maculatus), blue acara (Andinoacara pulcher), and Siamese fighting fish (Betta splendens) (García-Díaz et al., 2018). However, the common carp remains the only species that has received major research and management attention, mainly because it has invaded the Murray-Darling River system, the largest and most economically important river system on the continent (Koehn and MacKenzie, 2004). The common carp abundance is so high in some places that commercial fisheries have been established for the species (Harris, 2013). While the high abundance no doubt causes increased competition for habitat and food resources, common carp has had major impacts on water quality of receiving environments, including increased turbidity as well as a concomitant reduction in light penetration and alteration of thermal regimes (Koehn et al., 2000; Vilizzi et al., 2014). It has also been thought to be a vector by which co-invading parasites have entered the continent (Lymbery et al., 2014).
Non-native fish introductions into Belgium began in Roman times with common carp (Van Neer and Ervynck, 1993), which was originally introduced as a food item, but the koi variety has become an important ornamental species. The next ornamental fish introduced to Belgium was goldfish (Raveret-Wattel, 1900), which is commonly kept in aquaria and garden ponds but occasionally found in open waters (Verreycken et al., 2007). In the late 19th century, a number of North American sport fishes, such as Pumpkinseed (Lepomis gibbosus) and Eastern Mudminnow (Umbra pygmaea), were introduced to Belgium and neighbouring countries (Louette et al., 2001). Pumpkinseed became a popular ornamental fish, and it is now widespread in Flanders' inland waters, occurring in all but a single river basin (Verreycken et al., 2007; Cucherousset et al., 2009), but most abundant (and invasive) in the eastern part of Flanders (Verreycken et al., 2007). Eastern mudminnow, also used as an ornamental fish, has established and become locally abundant in lakes and rivers of northeast Flanders (Verreycken et al., 2010). At least three sturgeon species, including sterlet (Acipenser ruthenus), Siberian sturgeon (A. baerii), and Russian sturgeon (A. guldenstaedtii), have been recorded in Flemish rivers since the 1990s, most likely escapees from garden ponds during floods or deliberately-released unwanted pets. Fathead minnow (Pimephales promelas), including the rosy-red variety, was first observed in Flemish open waters in 1995 (Verreycken et al., 2007). Belgium is the only introduced range in Europe where fathead minnow is reported to have self-sustaining populations in both river and pond systems (Anseeuw et al., 2005; H. Verreycken, personal observation).
Annual imports of ornamental fishes to Canada consist of ≈ 1,500 species of which 745 are freshwater (Mandrak et al., 2014), with at least 13 fish species found in open waters of B.C. being linked to the ornamental trade (Crossman, 1991; Ministry of Environment and Climate Change Strategy, 2017). Notable examples include goldfish, pumpkinseed, blotched snake-head (Channa maculata), common carp/koi, and Oriental weatherfish. Translocated native Canadian species, e.g. yellow perch (Perca flavescens) and northern pike (Esox lucius), which have been found occasionally in the ornamental trade, are expanding their range in parts of B.C. where they are not native. However, the establishment status for most of these species is unclear. Additionally, it is possible that some of these species were introduced to B.C. via other vectors, such as intentional stocking, illegal recreational fishing, and live food trade. The goldfish is perhaps one of the few true ornamental species introduced to, and established in, B.C. The first record of the species in the province was in 1935 from a large pond in Salmon Arm (Carl et al., 1967, cited in Troffe, 1999). Self-sustaining feral populations of goldfish are found mainly in the southern part of B.C., including the lower mainland, the southern interior, and southern Vancouver Island, though individuals of the species regularly appear in ponds, lakes, and streams near human population centres across the province (McPhail, 2007). Recently, an increasing number of reports of goldfish present in high numbers (hundreds to thousands) has raised concerns about the impact of the species on B.C.'s aquatic environments (e.g. CJFC Today, 2017; CBC News, 2018).
Over 500 freshwater ornamental fishes are bred in China (Luo et al., 2015), with ≈ 200 NN tropical fishes and some 300 varieties of goldfish and common carp/koi involved in the ornamental trade (Luo et al., 2015). Guangdong is particularly vulnerable to ornamental fish invasions because of its role as a trading hub of ornamental fishes for both domestic and international markets, its favourable climate for the establishment of sub-tropical fishes, and its connection to extensive water courses (Li et al., 2013). In addition, there are over 15,000 ha of ornamental fish farms, producing about 100 billion fishes per year in Guangdong (Ye et al., 2016). A recent risk screening, using the Aquatic Species Invasiveness Screening Kit (AS-ISK; Copp et al., 2016), identified eight ornamental fish species that pose a medium to high invasion risk to the middle and low reaches of the Pearl River, including Goldfish, Butterfly Peacock Bass (Cichla ocellaris), Common Pleco (Hypostomus plecostomus), Guppy (Poecilia reticulata), Amazon Sailfin Catfish (Pterygoplichthys pardalis), Zebra Tilapia (Heterotilapia buttikoferi), Green Swordtail and Southern Platyfish (H. Wei, R. Chaichana, D. Punyanuch, F. Liu, M. Nimtim, Y. Zhu, L. Vilizzi, G.H. Copp, S. Li and Y. Hu, unpublished data). The South American Armoured Catfishes (Pterygoplichthys spp.), most likely a species group including Amazon Sailfin Catfish, Vermiculated Sailfin Catfish (P. disjunctivus), and hybrids of the two species, are of special concern because they have established self-sustaining populations in the main drainages of Guangdong (Wei et al., 2017). The hybrids dominate in numbers, likely owing to an ability to adjust life-history traits and adapt to new environments (Wei et al., 2017). South American Armoured Catfish invasions have reportedly led to decreased catches of commercially important native species (e.g. Shrimp) in Guangdong (Wei et al., 2017).
Studies of ornamental fish species in wild systems have been generally overlooked in the Philippines. Whereas the Philippines is a major exporter of marine ornamental fishes collected from the wild, its ornamental fish farming industry is not sufficiently developed to meet local demands for freshwater species. Retail prices of ornamental fishes in the Philippines are higher than in neighbouring countries (e.g. Thailand and Malaysia), so intentional releases into the natural environment are unlikely but some expensive species have been reported from wild systems (e.g. Asian Arowana Scleropages formosus). One exception is ornamental fishes that also serve as a food source, and these species can escape from fish farms, ponds, and aquaria during floods and cyclones. These include South American Armoured Catfishes, cichlids such as Jaguar Cichlid (Parachromis managuensis) (Agasen et al., 2006; Rosana et al., 2006; Briones et al., 2016), Midas cichlids (Amphilophus citrinellum and A. labiatum), and Flowerhorn (a complex hybrid of several neotropical cichlid species; Briones et al., 2016), as well as Knifefishes (Chitala spp.). The latter species are amongst the most prominent ornamental ones introduced to the Philippines (Aquino et al., 2011; Guerrero, 2014). Originally farmed for food but found to be unpalatable, knifefishes were subsequently used in the ornamental industry because juveniles are an attractive aquarium fish. However, they can be highly predatory and may decimate native fish species. Originally confined to a limited number of lakes, knifefishes are spreading to other lakes via connected channels and gateways/floodways (Aquiono et al., 2011; Guerrero, 2014; Cuvin-Aralar, 2014).
At least 38 freshwater fish species have been introduced into Poland's inland waters over the past 800 years, with at least 10 of these species linked to the ornamental trade (Grabowska et al., 2010; Nowak et al., 2011; Witkowski and Grabowska, 2012). Common Carp was the first freshwater fish brought to Poland for aquaculture use, beginning in the 13th and 14th centuries (Witkowski, 2008), but as with elsewhere, the koi variety is a prominent ornamental fish (Nowak et al., 2011). Other ornamental fish introductions include Pumpkinseed, Goldfish, Amur (a.k.a Chinese) Sleeper (Perccottus glenii), and Eastern Mudminnow (Nowak et al., 2011; Witkowski and Grabowska, 2012), with most considered to be established (Grabowska et al., 2010; Witkowski and Grabowska, 2012). Thought to have been released into Polish waters by aquarists or as a contaminant of stocking consignments from the East (Antychowicz, 1994), Amur Sleeper has rapidly spread downstream (over 600 km) from its original detection in 1993 in the middle section of the River Vistula. Previously restricted to the Vistula basin (Grabowska et al., 2010), Amur Sleeper has expanded westward into the Oder River drainage in 2018 (G. ZieR ba, A. Kruk, J. Grabowska, unpublished data), likely due to its hardiness and popularity as a baitfish. The invasion by Amur Sleeper has led to the local extinction of the native Lake Minnow (Rhynchocypris percnurus) in a small bog pond near Sobibór (Wałowski and Wolnicki, 2010).
Over 120 NN freshwater fish species have been recorded in Singapore of which a large proportion (≈ 89%, as of 2010) have been linked to the ornamental trade (Ng and Tan, 2010; Yeo and Chia, 2010; H.H. Tan et al., unpublished data). This disproportionately high number of NN freshwater fishes, relative to the country's small land area of ≈ 720 km2, can be attributed at least in part to Singapore's prominent role in the global ornamental fish trade industry (Yeo and Chia, 2010). Detected species in Singapore's inland waters include Asian Arowana (Ng and Tan, 1997), Ocellate River Stingray (Potamotrygon motoro) (Ng et al., 2010), Threadfin Acara (Acarichthys heckelii) (Tan and Lim, 2008), South American Armoured Catfishes (Ng et al. 1993; Ng and Tan, 2010), and Peacock Basses (Cichla spp.) (Ng and Tan, 2010; Liew et al., 2012). First reported for a reservoir in Singapore in 2006, the Ocellate River Stingray was the first introduction of an elasmobranch outside the Neotropics. This species exemplifies the adverse impacts a NN ornamental fish could exert on Singapore's fresh waters due to its venomous caudal spine with retrorse serrations, which can pose a health hazard to humans (Ng et al., 2010; da Silva et al., 2015). Most likely released as an unwanted pet (Ng et al., 2010), the species has since been found in other reservoirs and connected streams, with at least two established reservoir populations (Ng et al., 2010; Ng and Tan, 2010; Tan and Zeng, 2015).
Over 30 fish species have invaded water ways in England since the 1800s, with the majority (≈ 45%) attributed to the ornamental vector (Lever, 1977; Copp et al., 2007). Ornamental fish introductions to England began in 1691 with the importation of goldfish (Lever, 1977), followed by releases to the wild prior to the beginning of the 20th century (West, 1910). The prevalence of abandoned pet goldfish in London ponds dates back at least to the 1950s (Wheeler, 1958), and its hybrids with its European Congener, Crucian Carp (Carassius carassius), were already being recorded in the 1960s (Marlborough, 1969). Widespread occurrence in the wild was rapid for some ornamental fishes (Copp et al., 2007), such as the rosy red variety of Fathead Minnow, though established populations are rare (ZieRba et al., 2010). Contaminated ornamental fish consignments were also responsible for the introduction of North American White Sucker (Catostomus commersonii), which came in shipments of live goldfish from the U.S.A. (Copp et al., 2006). Similarly, Topmouth Gudgeon, one of Europe's most invasive fish (Gozlan et al., 2010), and Sunbleak (Leucaspius delineatus) are believed to have been imported to England as a contaminant of consignments of Golden and/or Blue Orfe, ornamental varieties of Ide (Leuciscus idus), with subsequent dispersal linked to fish movements (Copp et al., 2010b). Species imported for ornamental purposes, but also used in aquaculture or for angling purposes, include Sturgeons (Britton and Davies, 2006a) and three ictalurid catfishes of North America – Black Bullhead (Ameiurus melas), Channel Catfish (Ictalurus punctatus), and White Catfish (Ameiurus catus) (Lever, 1977; Britton and Davies, 2006b; Copp et al., 2007). Tropical (pet) fishes have also been reported for open waters, but none of these have persisted in England's colder waters (ZieRba et al., 2010). For example, the North American temperate fish, Red Shiner (Cyprinella lutrensis), which was previously sold in England as an ornamental aquarium species, has yet to be reported in open waters.
Florida had accumulated few established ornamental fishes until the Blackchin Tilapia (Sarotherodon melanotheron) showed up in Tampa Bay in 1959 (Springer and Finucane, 1963). This euryhaline ornamental species has since established populations in brackish and freshwater habitats in west-central and east-central Florida (USGS, 2018). At least 107 more species of freshwater ornamental fishes have been introduced since this time (Tuckett et al., 2017; USGS, 2018; present study). With 31 species established, the ornamental trade has become the predominant vector for NN freshwater fishes in Florida (Nico and Fuller, 1999; J.E. Hill, unpublished data). Florida's tropical climate in the southeast portion and warm temperate climate throughout the remainder of the state makes the region vulnerable to invasion by species from tropical regions of the world. Southeast and west-central Florida are the main hotspots of introduction and establishment, both regions having historic ornamental aquaculture and large human populations in major cities, though east-central and southwest Florida are accumulating species over time. Cichlids make up over half of the established ornamental species in Florida (16 species), followed by catfish species of three families and four poeciliids. Relatively little is known about impacts of most established NN fishes in Florida (Schofield and Loftus, 2015). Most established ornamentals seemingly have minor and localised impacts; however, the loricariid catfishes of the genus Pterygoplichthys are among the most impactful. These species burrow for reproduction, increasing localised erosion (Nico et al., 2009a), create biogeochemical hotspots (Rubio et al. 2016), and graze on the epibiota of the imperilled Florida Manatee (Trichechus manatus latirostris) in at least one spring (Nico et al., 2009b).
Things can get worse: potentially ‘risky’ freshwater fishes in the ornamental trade
We consulted the primary literature, reports, and online databases to assemble a database of freshwater NN fishes in the ornamental trade for the nine recipient regions with which to carry out a climate match as a means of identifying species that could pose a risk of establishing invasive populations (see Supplementary Information document Appendix 1a (uaem_a_1685849_sm8827.pdf) for the database and 1b (uaem_a_1685849_sm8840.docx) for sources and methods used). The number of freshwater ornamental fishes with the potential to establish in the recipient regions on the basis of climate match differs by region, ranging from seven for Flanders and Poland to 256 for Australia (Table 1). The proportion of these species that have already formed self-sustaining populations in the recipient regions also varies widely, between 1% for Australia and 100% for the Philippines. The large number of ‘risky’ species for Australia may be due to the relatively large region size, which covers a range of climate zones, thereby providing suitable habitat for many species; however, the small proportion of established species could be attributed to the implementation of a national management strategy specific to ornamental fishes, which permits the importation of species deemed low risk only via a whitelist (see next section). There appears to be taxonomic bias in the freshwater ornamental fish trade, at least for the regions considered in this review. Of the freshwater ornamental fishes considered, 74% come from just 10 families (of 67): Cyprinidae (carps and minnows), Cichlidae (cichlids), Characidae (tetras), Poeciliidae (live bearers), Cobitidae (loaches), Osphronemidae (gouramis), Loricariidae (catfishes), Ambassidae (glassfishes), Acipenseridae (sturgeons), and Callichthyidae (catfishes) (Figure 2). Cyprinidae is the only family with member species documented in all nine studied regions. Only 22 families contain species that are considered established in the recipient habitats of which five families, including Cichlidae, Cyprinidae, Poeciliidae, Loricariidae, and Centrarchidae (sunfishes), contributed to 71% of the established fishes. The taxonomic patterns observed for freshwater ornamental fish are consistent with those reported by Lockwood (1999) for birds, wherein human influence increases the probability of purposeful transport (or import) of species with desired traits but not necessarily their likelihood of successful establishment in the recipient environments. There is a general mismatch between the numbers presented in this section and those in the regional case studies because this section enumerates species that are known to be distributed via the ornamental fish trade supply chains, with a fraction of them already introduced and established in the wild; whereas, numbers reported in the regional case studies represent past and current observations of ornamental fishes in the environment (boxes with solid and dotted outline in Figure 1, respectively).
Number of non-native freshwater ornamental fish species for which the native range matches the climate of the recipient area by continent and region. Numbers in parentheses are the number of species that have already formed self-sustaining populations in the recipient regions. The Köppen-Geiger (K-G) climate classes of each introduced region are also shown. A = equatorial, B = arid, C = temperate, D = cold, and E = polar.
Current legislation and risk analysis procedures pertaining to non-native ornamental fishes
We summarised current prevention and management strategies on ornamental fishes by examining legal frameworks that address aquatic NN species for each region. All studied regions have developed policies, regulations or legislation to control the importation, distribution, possession, and/or release of NN species, but most of them are not specific to ornamental fishes (but for England, see Copp et al., 2005b). The policy and management strategies implemented by the regions can be largely grouped into five non-mutually exclusive categories: import controls, risk screening and/or full assessment, whitelist (a list of NN species suitable for import), blacklist (a list of controlled or prohibited NN species), and release ban (prohibitions against the release of NN species into specified environments) (Table 2).
All regions have implemented import controls to regulate the importation of NN species. For example, the importation of live animals and plants into Australia is regulated by the Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) and the Quarantine Act 1908. In Canada, the importation of aquatic animals is regulated via the Health of Animals Regulations (C.R.C., c. 296) under the Health of Animals Act (S.C. 1990, c 21) and the Aquatic Invasive Species Regulations (SOR/2015-121) under the Fisheries Act (R.S.C., 1985, c. F-14). Canada's National Code on Introductions and Transfers of Aquatic Organisms also provides a decision-making framework and process for licencing the introduction and transfer of aquatic organisms; however, live aquatic organisms intended specifically for the aquarium and water garden trade are excluded from the code (DFO, 2017). In China, trans-continental and trans-provincial importation of aquatic animals is prohibited, unless approved under the Quarantine Regulation for Importation of Aquatic Animals (2016), Fishery Act (2013), and the Regulation on Aquaculture Seedling (2005). The importation of ornamental fish broodstock requires a quarantine certification and that the species be kept in closed facilities for domestication for at least six months according to the Quarantine Regulation for Exportation of Ornamental Fish (2000). The Keeping and Introduction of Fish (England) Regulations 2014, which replaced the Import of Live Fish Act 1980 and the Prohibition of Keeping or Release of Live Fish (Specified Species) Order 1998 (and related Orders), regulates the importation of temperate zone fishes into England. In Poland, the importation, breeding, selling, and keeping of NN species is regulated under the Nature Conservation Act (Dz.U. 2004 Nr 92). The import of all species into Singapore is regulated via four legislative acts: Animals and Birds Act (1965), Wild Animals and Birds Act (1965), Fisheries Act (1969), and the Wholesome Meat and Fish Act (1999) (Yeo and Chia, 2010).
Formal risk analysis, involving risk screening, full risk assessment, risk management and communication, of the species is part of the importing process for Australia, Canada, China, England, and the Philippines. Australia follows a risk analysis framework for the importation of products into the country. The process formally assesses the risks of proposed importations and enables either risk management measures to be proposed to reduce identified risks of importation or to prohibit trade should the risks be deemed unacceptable. In Canada, the National Code for Introductions and Transfers of Aquatic Species includes a risk assessment process to identify risks associated with the intentional introduction and transfer of aquatic organisms (Mandrak et al., 2011) and has been applied to screen some freshwater fishes in trade (Mandrak et al., 2014). In China, the Quarantine Act (1992), Agriculture Act (2013), Environmental Protection Act (2014), Wildlife Protection Act (2018), and directive documents related to environmental protection stipulate that quarantine or scientific assessments shall be conducted prior to the importation or introduction of NN species, and those used in aquaculture or for aquarium purposes should be kept in closed facilities for domestication. China's Aquatic Wildlife Protection Regulation (2013) requires the application for importing licences and the risk assessment report of the species to be submitted for approval. In Great Britain (England), NN species risk screening and analysis may involve either a screening phase only, using the Great Britain ‘Rapid Risk Assessment’ (GB NNSS, 2019), possibly complemented by the use of the Aquatic Species Invasiveness Screening Kit (Copp et al., 2016), or proceed directly to a full risk assessment. In the Philippines, a process to control the importation of fishes into the Philippines is implemented under the Fisheries Administrative (FAO) Order No. 221. The process involves a Letter of Request to Import and Application for a Permit to Import, an Import Risk Analysis for evaluation of risk factors regarding the environment, disease, and effects on native species, an inspection of the importer's facilities, and the insurance of the Sanitary and Phytosanitary Certificate if approved (Guerrero, 2014). Although not mandated in regulations, risk assessment is frequently used by state agencies in Florida to support decision-making on NN fishes, including ornamentals (e.g. Hill et al. 2014; Lawson et al. 2015).
A summary of key management strategies and legal frameworks relevant to aquatic non-native species, especially fishes, in the selected regions. I = import controls, RA = risk assessment, B = blacklist, W = whitelist or whitened blacklist (i.e. England), and RB = release ban. Policies and legislation with a focus on non-native ornamental fishes are denoted by an asterisk(*).
Some regions, including Canada, England, Flanders, Poland, and Singapore, use blacklists to control the movement of specific NN species into and within the regions. In Canada, the Aquatic Invasive Species Regulations (SOR/2015-121) impose restrictions around the importation, possession, transportation, and/or release of a list of prohibited and controlled species in specified geographic regions. At the provincial level, B.C. also regulates species via the Controlled Alien Species Regulation of the Wildlife Act (B.C. Reg. 94/2009), banning the possession, breeding, release, and transportation of a list of controlled species, including a few ornamental fishes, without a permit. In England, the Prohibition of Keeping or Release of Live Fish (Specified Species) Order 1998 includes a list of species subject to control. This Order's list, which was revised in 2003, extended controls to the keeping of NN species for commercial and private purposes (e.g. fish farmers and dealers, ornamental trade, and aquarists). Controls on the listed species is graduated, extending from a ‘general licence', which allows all persons in England to keep the species, to a ‘personal licence’, which limits fish keeping in quarantine-level holding facilities. The strength of the legislation was further enhanced by enactment of the Keeping and Introduction of Fish (England) Regulations 2014 in which the list of species subject to control was replaced by a list of taxonomic Orders (i.e. those containing temperate and semi-temperate species). This ‘whitening' of the blacklist effectively places regulatory controls on all species of the taxonomic Orders listed in one annex of the Regulation, except those listed in a second annex, which contains native and ‘tolerated' NN species. Members of the European Union (EU), including Flanders (Belgium), Poland and the U.K. in this case, have adopted the EU Regulation (No. 1143/2014) on the Prevention and Management of the Introduction and Spread of Invasive Alien Species, which includes a list of Invasive Alien Species of Union concern (the Union list). The principal risk scheme used for assessing these species (e.g. mosquitofishes; Copp and Verreycken, 2017) is a modified version of the Great Britain full risk assessment scheme, adapted to comply with the ‘minimum standards’ for risk assessment under the EU Regulation (Roy et al., 2018). The listed species are subject to restrictions on importing, selling, breeding, growing, and keeping. National lists also exist, such as Poland's list of species that may threaten native species and natural habitats if released into the natural environment under the Nature Conservation Act. Specific to the ornamental fish trade in Singapore, the Fisheries (Piranha) Rules (1971) prohibits the importation, rearing, and release of piranhas (Yeo and Chia, 2010).
Whitelists are also being used to manage NN species. Australia, the U.K., and the U.S.A. have both blacklists and whitelists. A Strategic Approach to the Management of Ornamental Fish in Australia specifies a list of species suitable for live import, dividing species into those that require or do not require a permit under the EPBC Act (Natural Resource Management Ministerial Council, 2006). There is also a national noxious species list, though this does not affect the estimated 1,000+ species in ornamental settings already in Australia (Harris, 2013). In England, the EU Regulation (No. 708/2007) concerning the use of NN and locally-absent species in aquaculture, which was implemented by way of a statutory instrument (the Alien and Locally Absent Species in Aquaculture Regulations for England and Wales, 2011), employs a whitelist approach, i.e. all NN species are subject to regulation except those on either the Regulation's exemption list (Annex IV) or by an exemption extended by an EU Member State for its territory. This contrasts the blacklist basis of the EU's Regulation (No. 1143/2014). In the U.S.A, a list of injurious wildlife species including fishes is maintained and their importation directly into Florida from outside of the continental U.S.A. is regulated under the Lacey Act of 1900. The wildlife trafficking provisions of the Act also provides for federal enforcement and penalties for interstate violations of fish and wildlife laws, including the illegal movement of NN species into Florida. At the state level, fishes are regulated under the Florida Administrative Code Chapter 68-5. A tiered, mixed list approach (i.e. prohibited, conditional species, and clean lists), with heavy reliance on dirty lists, is used to manage fishes imported into or cultured and possessed within the state (Hill, 2016). The Philippines has a List of Live Aquarium Fishes Allowed for Importation with a total of 91 species (Guerrero, 2014).
The introduction or release of NN species into the wild is prohibited in most regions. For example, there is a general prohibition in Canada against the introduction of any aquatic species into fish-bearing waters without a valid permit, licence, or authorisation under the Aquatic Invasive Species Regulations (Government of Canada, 2019b). In China, Regulation on Enhancement and Release of Aquatic Species (2009) prohibits the release of NN, hybrid, and transgenic aquatic species into natural water bodies. In England, the keeping or release of any new and listed species to open waters is presumed illegal under the Wildlife and Countryside Act (1981) and the Keeping and Introduction of Fish (England) Regulations 2014, though there are exceptions to some existing, tolerated species (i.e. common carp, goldfish, ide, crucian carp). In Flanders, the Species Act of 2009 (Besluit van de Vlaamse Regering met betrekking tot soortenbescherming en soortenbeheer) forbids the introduction of NN species into the wild without a permit. It is illegal to release any NN species into Florida's environment under the Florida Administrative Code Chapter 68-5. The Philippines Fisheries Code of 1998 on Introduction of Foreign Aquatic Species prohibits the introduction of aquatic organisms into Philippines waters without sound ecological justifications based on scientific studies. The Nature Conservation Act (Dz.U. 2004 Nr 92) in Poland prohibits the movement and introduction of species including fishes into the natural environment of Poland. In Singapore, the National Parks Board enforces regulations preventing the introduction of species into the nature reserves and parks, as well as waterways that connect to the reserves, under the Parks and Trees Act (2005, amended in 2017) (Yeo and Chia, 2010).
There are also policies, regulations, and legislation regulating genetically modified ornamental fishes. For instance, multiple genetic lines and species of transgenic fluorescent ornamental fish are cultured in Florida and are sold to aquarists in the U.S.A. (Hill et al., 2014). The commercial use of genetically modified ornamental fishes are regulated under the Federal Food, Drug, and Cosmetic Act. Each species and line combination is reviewed and approved at the state level as well as at the federal level under the New Animal Drug Application regulations. Aquaculture of NN species including ornamental fishes is regulated under the Florida Aquaculture Policy Act, Chapter 597, Florida Statutes. Florida aquaculture operations must maintain aquaculture certification since 2000. A major requirement of certification is mandatory compliance with Florida Aquaculture Best Management Practices (BMPs), including provisions for species culture and containment (FDACS, 2016).
While legal frameworks exist to address NN species, their effectiveness in mitigating invasion risk is generally unclear. There seems to be a patchwork of policies, regulations, and legislation as well as a lack of coordination among regulatory agencies and/or among provinces, states, and territories in managing NN species (e.g. Australia, Koehn and Mackenzie, 2004; Canada, Office of the Auditor General of Canada, 2019). A review of NN fish legislation in the early 2000s from many European and two North America countries (Canada and U.S.A.) found these legal instruments to be largely ineffective, except in England and Wales, due to a lack of precise regulations and limited enforcement of these regulations (Copp et al., 2005b). An exception in North America is the State of Florida, where mandatory BMPs have clearly reduced the probability of escape by cultured ornamentals (Tuckett et al. 2016). Furthermore, very few of these legal frameworks specifically address the ornamental fish trade or species.
Conclusions and recommendations
Our framework describes the invasion process of ornamental fishes and identifies determinants of invasion success pertaining to the ornamental trade. Characterisation of these components and factors contributing to invasion success (or failure) can inform future research and management efforts by identifying knowledge and data gaps, gaining a better understanding of potential control points, and developing monitoring and control programmes. The regional case studies of ornamental fishes observed in freshwater habitats demonstrate the magnitude of the ornamental trade as a global vector for introductions of freshwater fishes. The number of ‘risky’ freshwater species in the trade further highlights the potential invasion risk associated with the ornamental fish trade. These findings suggest that current efforts to control the transfer and release of NN species may be insufficient to reduce the invasion potential of freshwater ornamental fishes.
We recommend the use of a comprehensive risk analysis scheme as a means of identifying and assessing species, with the resulting assessments of compiled and evaluated information subjected to a peer-reviewed process to ensure that advice intended to inform current and future policy and management of ornamental fishes is evidence based. A number of risk screening approaches have been developed including scoring systems (e.g. FISK and AS-ISK, Copp et al., 2005a, 2016; Lawson et al., 2013; CMIST, Drolet et al., 2017), decision trees (e.g. Kolar and Lodge, 2002), and probabilistic models (e.g. Keller et al., 2007; Marcot et al., 2019). Full risk assessments can then be undertaken on those species identified in the screening process as posing a higher risk of becoming invasive in the risk assessment area (e.g. Cudmore et al., 2012; Copp and Verreycken, 2017), with vector assessment to identify components of the supply chain or stages of the invasion process where introduction risk is higher (Mandrak and Cudmore, 2015) and where management control points might exist. For instance, a vector assessment of live bait fisheries determined that the angler-release stage was a critical point for targeted management to reduce the overall invasion risk associated with the baitfish industry (Drake and Mandrak, 2014). Spatially explicit vector assessments can also be used to determine key entry points or release sites of ornamental fishes. For example, Chan et al. (2013) conducted a vector assessment to identify the major ports at greatest risk of invasions from ballast water discharge in the Canadian Arctic. A similar ‘hotspot' approach was undertaken for marine waters of Great Britain and Ireland (Tidbury et al., 2016). Risk assessments performed on individuals or groups of species can direct control efforts to those with the highest invasion potential, given limited resources (Mandrak and Cudmore, 2015). The ‘risky’ species identified in the present study should be prioritised for species risk assessments, especially for regions where risk assessments are not part of the existing introduced species management strategies.
Detailed, computer-based species import records would be useful for ornamental trade risk assessments, both at the vector- and species-levels, e.g. as exists for England (Copp et al., 2007; Tidbury et al., 2016). This information would also contribute to species monitoring, control, enforcement, and decision-making (Gerson et al., 2008; Chan et al., 2015b). While mechanisms to capture wildlife trade data exist, they often lack taxonomic resolution and are insufficient for tracking the movement of ornamental species into a country (Gerson et al., 2008; Allen et al., 2017; Rhyne et al., 2017) or the subsequent re-distribution post-boarder arrival. For example, the Harmonized Commodity Description and Coding System (HS) of the World Customs Organization, which is considered the most comprehensive dataset for global commodity trade, categorises wildlife commodities into broad taxonomic groups (Gerson et al., 2008; Chan et al., 2015b). Under this system, all live freshwater ornamental fishes are grouped under the HS code 030111, making it extremely difficult to detect specific species and quantify propagule pressure associated with the ornamental trade. We propose establishing a standardised mechanism for collecting detailed information on species imported into a region, including composition, species-level quantity, and origin.
Increased public education and broader industry engagement on the invasion risk associated with introduced fishes could discourage the intentional release of ornamental fishes into the natural environment (Chang et al., 2009). Programmes to raise awareness on NN species and to encourage the adoption of risk-lowering behaviour (e.g. re-homing unwanted pet fishes rather than releasing them into the wild) exist for some but not all regions, such as the Habitattitude™ campaigns partnered with the pet industry in the USA ( www.habitattitude.net/) and Canada ( www.habitattitude.ca/home-2/), as well as the ‘Be Plant Wise' education initiative in Great Britain ( www.nonnativespecies.org//beplantwise/index.cfm?), which focuses on ornamental plants but is also relevant to aquarium and pond fishes as contaminants of plant consignments (Copp et al., 2017). We are not aware of any formal evaluations on the effectiveness of these programmes, except perhaps for the U.K. where Sutcliffe et al. (2018, p.409) identified the following barriers to the uptake of biosecurity initiatives (e.g. Check Clean Dry and Be Plant Wise): “financial constraints linked to a lack of knowledge about invasion pathways and control measures, a focus on managing already established and visible INNS (invasive NN species), and collective action problems if others fail to undertake biosecurity”. Also, there is evidence that the Habitattitude™ campaign in the U.S.A. has raised awareness of aquarium and water garden owners on the NN species issue via social media platforms and pet surrender events (T. Campbell and D. Jensen, personal communication). Working with cultural and religious groups to identify low-impact release sites and species for mercy release has also been proposed as a management option to reduce the negative effects of ornamental species in natural environments while being sensitive to the groups’ sociocultural and religious value systems (Wasserman et al., 2019). We suggest the extension of awareness and education campaigns to all regions to prevent new introductions of ornamental fishes outside of their natural range.
Some ornamental fishes will inevitably escape or be released into the environment despite prevention efforts; thus, early detection and rapid response (EDRR) programmes are essential for minimising the invasion risk associated with the ornamental fish trade. Britton et al. (2011) provided a detailed review of various EDRR tools and options for managing NN fishes in the environment. In brief, these approaches include supressing new populations by removing individuals using techniques such as electrofishing and gill netting, which may not always be completely successful (Davison et al., 2017), containment of populations using an artificial barrier to prevent further spread, and when feasible, eradication using chemical treatments before populations become large and geographically widespread (Britton et al., 2011). Because it is often difficult to detect introduced fishes when they are occurring at low abundances, which is common for new or managed populations, we suggest using molecular techniques to aid in detection and monitoring efforts. A range of molecular-based detection methods has been developed and applied to detect the presence of fishes in aquatic ecosystems (e.g. Jerde et al., 2013; Davison et al., 2019). We also recommend prioritising EDDR programmes at potentially high-risk sites, such as more accessible water bodies near urban centres and culture facilities (Copp et al., 2005c, 2007).
The ornamental fish trade is a multi-billion-dollar industry that provides job opportunities and raises awareness about biodiversity. Fish keeping can also promote human health by relieving stress. At the same time, it is an emerging source of NN fishes in freshwater ecosystems around the world. Therefore, increased research and management efforts are needed to manage the invasion risk associated with the ornamental fish trade so that the socio-economic benefits of the industry can be fully realised without negatively impacting native ecosystems.
We thank M. Munawar, A. Zhan, MFIS-China Organizing and Scientific Committees, and the Aquatic Ecosystem Health & Management (AEHMS) Secretariat for their efforts in organising the Marine and Freshwater Invasive Species: Solutions for Water Security conference in Beijing, China, which led to the conception of this paper.
Supplementary data for this article can be accessed on the publisher's website.
We are grateful for the support from Fisheries and Oceans Canada, to F.T. Chan and T.W. Therriault; Cefas and the UK Department of Environment, Food & Rural Affairs, to G.H. Copp; the MFIS-China Organizing Committee to D.C.J. Yeo; the Central Public-interest Scientific Institution Basal Research Fund CAFS (No. 2018GH11 and 2019GH09), to H. Wei and D. Luo; the National Science Centre, Poland (decision No. DEC-2011/01/D/NZ8/01807), to G. ZieRba; the Florida Department of Agriculture and Consumer Services, the Florida Fish and Wildlife Conservation Commission, and the U.S. Department of Agriculture to J.E. Hill, and the PADI Foundation (No. 40551) to A.S. Gilles Jr.