David Almeida, 1,2 Filipe Ribeiro, 3,4 Pedro M. Leunda, 5,6 Lorenzo Vilizzi, 7 and Gordon H. Copp 1,2,6,8,

Transcription

1 Risk Analysis DOI: /risa Effectiveness of FISK, an Invasiveness Screening Tool for Non-Native Freshwater Fishes, to Perform Risk Identification Assessments in the Iberian Peninsula David Almeida, 1,2 Filipe Ribeiro, 3,4 Pedro M. Leunda, 5,6 Lorenzo Vilizzi, 7 and Gordon H. Copp 1,2,6,8, Risk assessments are crucial for identifying and mitigating impacts from biological invasions. The Fish Invasiveness Scoring Kit (FISK) is a risk identification (screening) tool for freshwater fishes consisting of two subject areas: biogeography/history and biology/ecology. According to the outcomes, species can be classified under particular risk categories. The aim of this study was to apply FISK to the Iberian Peninsula, a Mediterranean climate region highly important for freshwater fish conservation due to a high level of endemism. In total, 89 fish species were assessed by three independent assessors. Results from receiver operating characteristic analysis showed that FISK can discriminate reliably between noninvasive and invasive fishes for Iberia, with a threshold of 20.25, similar to those obtained in several regions around the world. Based on mean scores, no species was categorized as low risk, 50 species as medium risk, 17 as moderately high risk, 11 as high risk, and 11 as very high risk. The highest scoring species was goldfish Carassius auratus. Mean certainty in response was above the category mostly certain, ranging from tinfoil barb Barbonymus schwanenfeldii with the lowest certainty to eastern mosquitofish Gambusia holbrooki with the highest level. Pair-wise comparison showed significant differences between one assessor and the other two on mean certainty, with these two assessors showing a high coincidence rate for the species categorization. Overall, the results suggest that FISK is a useful and viable tool for assessing risks posed by non-native fish in the Iberian Peninsula and contributes to a watch list in this region. KEY WORDS: Alien species prescreening; aquatic conservation; biological invasions; hazard identification 1 Salmon & Freshwater Team, Cefas, Pakefield Road, Lowestoft, Suffolk, NR33 0HT, UK. 2 Centre for Conservation Ecology, Bournemouth University, Poole, Dorset, BH12 5BB, UK. 3 Centro de Oceanografia, Faculdade de Ciências da Universidade de Lisboa, Campo Grande Lisboa, Portugal. 4 Museu Nacional de História Natural e da Ciência, Universidade de Lisboa, Rua da Escola Politécnica 58, Lisboa, Portugal. 5 Gestión Ambiental de Navarra S.A., C/ Padre Adoain 219 Bajo, Pamplona/Iruña, Navarra, España. 6 University of Navarra, School of Sciences, Department of Environmental Biology,, c/irunlarrea 1, E-31008, Pamplona/Iruña, Navarra, Spain. 1. INTRODUCTION The Iberian Peninsula has a unique freshwater fish fauna with a high level of endemism, that is, >30% of native species, (1 4) and is increasingly threatened by human-mediated pressures such as 7 Murray-Darling Freshwater Research Centre, PO Box 991, Wodonga VIC 3689, Australia. 8 Environmental and Life Sciences Graduate Program, Trent University, Peterborough, Ontario K9J 7B8, Canada. Address correspondence to Gordon H. Copp, Salmon & Freshwater Team, Cefas, Pakefield Road, Lowestoft, Suffolk NR33 OHT, UK; gordon.copp@cefas.co.uk /13/ $22.00/1 C 2013 Society for Risk Analysis

2 2 Almeida et al. water extraction, pollution, discharge regulation, and non-native species. (5) In Europe, several countries show high ratios of non-native fishes, such as Greece (19%), (6) the United Kingdom (21%), (7) or the Czech Republic (23%). (8) Currently, the Iberian Peninsula is considered a non-native species hotspot for freshwater fishes, that is, >25% of total fish species are non-native, (9) where non-native fishes can exert a range of impacts on native ecosystems, including predation, competition, habitat alteration, hybridization, and disease transmission. (9,10) Yet, the number of non-native fishes in the Iberian Peninsula continues to increase despite concerns that these pose an increasing threat to biodiversity. (11,12) Thus, assessments of non-native fish risks are essential for controlling their numbers and thus mitigating their potential impacts. (13) To evaluate invasion risks in the Iberian Peninsula, most of the conducted research has focused on identifying key biological traits that facilitate successful fish invasions. (14,15) Recently, a risk assessment approach has been proposed for non-native fishes in the Iberian Peninsula (12) that is dependent on a profound understanding of invasion pathways and detailed knowledge of the key biological features that determine invasion success in this region. In aiming to aid in the prevention of future introductions, the Clavero (12) approach focused mainly on early invasion stages (i.e., arrival and establishment), and was less comprehensive regarding risks associated with the subsequent dispersal and impact stages of the invasion process. Furthermore, the Clavero approach (12) followed a region-specific procedure that is not easily applied outside the Iberian Peninsula, thus limiting its use to inform managers and policymakers in a broader context (e.g., European legislation). This contrasts the Fish Invasiveness Scoring Kit (FISK), which is a risk identification (screening) tool that has been successfully applied in different countries around the globe (16 20) and recently revised for wider use, including warm temperate and subtropical climatic zones, such as peninsular Florida, United States. (21) Adapted from the Weed Risk Assessment (22) by Copp et al. (23) for assessing non-native freshwater fishes in the United Kingdom, FISK consists of 49 questions that address eight topics (domestication/cultivation, climate and distribution, invasive elsewhere, undesirable traits, feeding guild, reproduction, dispersal mechanisms, and persistence attributes) under two main subject areas (biogeography/history, biology/ecology) that encompass all stages of the invasion process. (24) Details on the weighting of outcome scores, which range from 15 to 57, with which to rank the potential risk of a species being invasive as low, medium, or high are described in detail elsewhere. (25) The aim of this study was to undertake a risk identification assessment of non-native fishes relevant to Iberian drainage basins, and thus contribute to the broader geographic applicability of the FISK. Relevant species include existing translocated and non-native fishes, as well as possible future invaders of the peninsula. 2. METHODOLOGY A total of 89 fish species (Table I) was assessed with FISK v2. (21) The selection of species was based on the following criteria: (i) native species from the Iberian Peninsula that have been translocated across Iberian watersheds into drainage basins outside their native range; (ii) non-native species already present in the Iberian Peninsula that could spread into new Iberian areas; (iii) non-native species that are currently not present in the Iberian Peninsula but are located near the Perpignan corridor a recognized donor area of fish introductions to the Iberian Peninsula; (26) and (iv) non-native species that are not currently present either in the Iberian Peninsula or near the Perpignan corridor but are important for aquaculture or the aquarium trade. Of the 89 fish species, 16 (18.0%) were classified according to criterion (i), 25 (28.1%) as per criterion (ii), 34 (38.2%) as per criterion (iii), and the remaining 14 (15.7%) as per criterion (iv). Receiver operating characteristic (ROC) analysis (27) was used to assess the predictive ability of FISK to discriminate between invasive and noninvasive species. To this end, species were classified apriorias either invasive or noninvasive based on information available from the Invasive Species Specialist Group Database ( and from FishBase ( Statistically, an ROC curve is a graph of sensitivity versus 1 specificity (or alternatively, sensitivity vs. specificity), where in the present context sensitivity and specificity will be the proportion of invasive and noninvasive fish species, respectively, that are correctly identified by the FISK tool as such. A measure of the accuracy of the calibration analysis is the area under the ROC curve (AUC). If the AUC is equal to 1.0 (i.e., the ROC curve consists

3 Effectiveness of FISK for Assessing Fishes in Iberia 3 Table I. Fish Species Assessed with FISK v2 for the Iberian Peninsula Score CF Invasiveness/ Species Name Common Name Criterion Protection Status Mean Min Max SE Outcome Mean Min Max SE Abramis brama common bream 2 Noninvasive/LC MH Achondrostoma arcasii bermejuela 1 Noninvasive/VU M Acipenser baerii Siberian sturgeon 3 Noninvasive/NE M Acipenser naccarii Adriatic sturgeon 4 Noninvasive/CR M Alburnoides bipunctatus spirlin 3 Noninvasive/NE M Alburnus alburnus bleak 2 Noninvasive/LC H Ameiurus melas black bullhead 2 Invasive/NE VH Anguilla anguilla European eel 1 Noninvasive/CR M Aphanius fasciatus southern-european toothcarp 3 Noninvasive/LC M Astronotus ocellatus oscar 4 Noninvasive/NE M Australoheros facetus chameleon cichlid 2 Noninvasive/NE M Barbatula barbatula stone loach 3 Noninvasive/LC M Barbatula quignardi Lanquedoc stone loach 1 Noninvasive/LC M Barbonymus schwanenfeldii tinfoil barb 4 Noninvasive/NE M Barbus barbus European barbel 3 Noninvasive/LC M Blicca bjoerkna silver bream 2 Noninvasive/LC M Carassius auratus goldfish 2 Invasive/NE VH Carassius carassius crucian carp 3 Invasive/LC VH Carassius gibelio gibel carp 3 Invasive/NE VH Channa argus northern snakehead 4 Noninvasive/NE MH Channa marulius bullseye (great) snakehead 4 Noninvasive/LC MH Channa micropeltes giant snakehead 4 Noninvasive/VU M Chondrostoma nasus common nase 3 Noninvasive/LC M Cobitis bilineata Italian spined loach 2 Noninvasive/LC M Cobitis calderoni northern Iberian spined loach 1 Noninvasive/EN M Cobitis paludica southern Iberian spined loach 1 Noninvasive/VU M Ctenopharyngodon idella grass carp 3 Noninvasive/NE VH Cyprinus carpio carpio common carp 2 Invasive/VU VH Esox lucius northern pike 2 Invasive/LC H Fundulus heteroclitus mummichog 2 Noninvasive/NE MH Gambusia affinis mosquitofish 3 Invasive/NE H Gambusia holbrooki eastern mosquitofish 2 Invasive/NE MH Gobio alverniae Auvergne gudgeon 3 Noninvasive/LC M (Continued)

4 4 Almeida et al. Table I. (Continued) Score CF Invasiveness/ Species Name Common Name Criterion Protection Status Mean Min Max SE Outcome Mean Min Max SE Gobio gobio gudgeon 3 Noninvasive/LC M Gobio lozanoi Iberian gudgeon 1 Noninvasive/LC M Gobio occitaniae Languedoc gudgeon 3 Noninvasive/LC M Gymnocephalus cernua ruffe 3 Invasive/LC MH Hucho hucho huchen 3 Noninvasive/CR M Hypophthalmichthys molitrix silver carp 3 Invasive/NT MH Hypophthalmichthys nobilis bighead carp 3 Invasive/NE MH Ictalurus punctatus channel catfish 2 Invasive/NE VH Lates niloticus Nile perch 4 Invasive/LC MH Lepomis gibbosus pumpkinseed 2 Noninvasive/NE H Leucaspius delineatus sunbleak 3 Noninvasive/LC MH Leuciscus bearnensis Bearn beaked dace 3 Noninvasive/LC M Leuciscus burdigalensis beaked dace 3 Noninvasive/LC M Leuciscus idus ide 3 Noninvasive/LC M Leuciscus leuciscus common dace 3 Invasive/LC M Leuciscus oxyrrhis long-snout dace 3 Noninvasive/LC M Luciobarbus bocagei Iberian barbel 1 Noninvasive/LC M Luciobarbus graellsii Ebro barbel 1 Noninvasive/LC M Micropterus salmoides largemouth bass 2 Invasive/NE H Misgurnus anguillicaudatus Oriental weatherfish 2 Invasive/NE H Misgurnus fossilis European weatherfish 3 Noninvasive/LC M Neogobius melanostomus round goby 4 Noninvasive/NE H Oncorhynchus kisutch Coho salmon 3 Noninvasive/NE M Oncorhynchus mykiss rainbow trout 2 Invasive/NE MH Oreochromis aureus blue tilapia 4 Invasive/NE VH Oreochromis mossambicus Mozambique tilapia 3 Invasive/NT H Oreochromis niloticus Nile tilapia 3 Invasive/NE VH Pachychilon pictum Albanian roach 3 Noninvasive/LC M Parachondrostoma miegii Ebro nase 1 Noninvasive/LC M Parachondrostoma toxostoma sofie 3 Noninvasive/LC M Perca fluviatilis Eurasian perch 2 Invasive/LC MH Phoxinus bigerri Adour minnow 1 Noninvasive/LC M Phoxinus phoxinus European minnow 3 Invasive/LC M (Continued)

5 Effectiveness of FISK for Assessing Fishes in Iberia 5 Table I. (Continued) Score CF Invasiveness/ Species Name Common Name Criterion Protection Status Mean Min Max SE Outcome Mean Min Max SE Phoxinus septimaniae Languedoc minnow 3 Noninvasive/LC M Piaractus brachypomus pirapatinga 4 Noninvasive/NE M Pimephales promelas fathead minnow 3 Invasive/NE MH Poecilia reticulata guppy 2 Invasive/NE MH Pseudochondrostoma polylepis Iberian nase 1 Noninvasive/LC M Pseudorasbora parva topmouth gudgeon 2 Invasive/NE VH Pygocentrus nattereri red piranha 4 Noninvasive/NE M Rutilus rutilus roach 2 Invasive/LC H Salmo salar Atlantic salmon 1 Noninvasive/LC M Salmo trutta trutta sea trout 1 Invasive/NE MH Salvelinus fontinalis brook trout 2 Invasive/NE M Sander lucioperca pikeperch 2 Invasive/NE MH Scardinius erythrophthalmus rudd 2 Invasive/LC H Silurus glanis European catfish 2 Invasive/LC H Squalius alburnoides calandino 1 Noninvasive/VU M Squalius cephalus chub 1 Noninvasive/LC M Squalius pyrenaicus Iberian chub 1 Noninvasive/NE M Thymallus thymallus grayling 3 Noninvasive/LC M Tilapia zillii redbelly tilapia 4 Invasive/NE VH Tinca tinca tench 2 Noninvasive/LC MH Umbra pygmaea eastern mudminnow 3 Noninvasive/NE M Xiphophorus hellerii green swordtail 4 Invasive/NE M Xiphophorus maculatus southern platyfish 4 Invasive/NE M Notes: For each species, selection criterion, aprioriinvasiveness (as per and and protection status (categories initials as per along with summary statistics (SE = standard error) for corresponding FISK score, (risk) outcome, and certainty factor (CF), are reported. Criteria classified as: 1 = native species from the Iberian Peninsula that have been translocated to other Iberian areas; 2 = non-native species already present in the Iberian Peninsula that could spread into new Iberian areas; 3 = non-native species that are currently not present in the Iberian Peninsula but are located near the Perpignan corridor ; and 4 = non-native species that are currently not present either in the Iberian Peninsula or near the Perpignan corridor but are important for aquaculture or the aquarium trade. Outcome is based on a calibration threshold of between medium risk and high risk sensu lato species and classified as: medium (M) = [1, 20.25[; moderately high (MH) = [20.25, 25[; high (H) = [25, 30[; very high (VH) = [30, 57]. See text for computation of the CF. Species names after Leunda et al. (43) and

6 6 Almeida et al. of two straight lines, one vertical from 0,0 to 0,1 and the other horizontal from 0,1 to 1,1), then the test is 100% accurate because both sensitivity and specificity are 1.0 and there are neither false positives (i.e., noninvasive species categorized as invasive) nor false negatives (i.e., invasive species categorized as noninvasive). Conversely, if the AUC is equal to 0.5 (i.e., the ROC curve is a diagonal line from 0,0 to 1,1), then the test is 0% accurate as it cannot discriminate between true positives (i.e., actual invasive species) and true negatives (i.e., actual noninvasive species). Typically, the AUC will range between 0.5 and 1.0, and the closer the AUC to 1.0 the better the ability of FISK to differentiate between invasive and noninvasive species. Assessments were carried out separately and independently on all 89 species by three assessors (DA, FR, and PML) knowledgeable in the fish fauna of the Iberian Peninsula. Separate ROC curves were initially generated for each assessor and pair-wise differences between AUCs for the three resulting ROC curves were statistically tested (Mann-Whitney U- statistic, α = 0.05). Following between-curve comparisons, a mean ROC curve was computed on the mean scores from the three assessors. Based on the mean ROC curve, the best FISK threshold or cut-off value that maximizes the true positive rate (i.e., true invasive classified as invasive) and minimizes the false positive rate (i.e., true noninvasive classified as invasive) was determined using both Youden s J-statistic (28) and the point closest to the top-left part of the plot with perfect sensitivity or specificity. For the mean ROC curve, a smoothed mean curve was also generated along with bootstrapped confidence intervals of specificities along the entire range of sensitivity points (i.e., 0 to 1, at 0.1 intervals). (29) AUC comparisons were carried out using the online applet StAR (available at (30) whereas all other ROC analyses were done with package proc for Rx (31) using n = 2,000 bootstrap replicates. As each response of FISK for a given species is allocated a certainty score (1 = very uncertain; 2 = mostly uncertain; 3 = mostly certain; 4 = very certain), a certainty factor (CF) was computed as: (CQi )/(4 49)(i = 1,...,49), where CQ i is the certainty for question i, 4 is the maximum achievable value for certainty (i.e., very certain ), and 49 is the total number of questions comprising the FISK tool. The CF therefore ranges from a minimum of 0.25 (i.e., all 49 questions with certainty score equal to 1) to a maximum of 1 (i.e., all 49 questions with certainty score equal to 4). For further assessment of consistency between assessors, error (or confusion) matrices (32) were performed and coincidence rates determined for species categorization according to risk extent (i.e., low, medium, moderately high, high, very high). Also, the CFs for the 89 species were compared between assessors by a one-way permutational univariate analysis of variance (PERANOVA), using a Euclidean dissimilarity measure following normalization, n = 9,999 permutations of the raw data, and with a posteriori pair-wise comparisons at α = 0.05 (PERMANOVA+ for PRIMER v6). (33) 3. RESULTS There were no statistically significant differences (p > 0.05) between AUCs for the three assessorspecific ROC curves (Fig. 1 top). This justified computation of an overall ROC curve based on the mean scores, which showed an AUC of ( , 95% CI), hence well above 0.5 (Fig. 1 bottom). This indicated that FISK was able to discriminate reliably between invasive and noninvasive species for Iberia. Youden s J and closest point statistics provided the same best threshold of 20.25, which was therefore chosen as the calibration threshold of FISK risk outcomes for the Iberian Peninsula (Table I). Accordingly, the threshold was used to distinguish between medium risk species (i.e., species with FISK scores within the interval [1, 20.25[) and high risk sensu lato species (i.e., species with FISK scores within the interval [20.25, 57]), with the latter further categorized as per Britton et al. (34) into moderately high risk (interval [20.25, 25[), high risk (interval [25, 30[), and very high risk (interval [30, 57]), and with low risk species having a FISK score within the interval [ 15, 1[. Based on the above threshold and subsequent intervals, none of the 89 species assessed was categorized as low risk based on mean scores, 50 (56.2%) were categorized as medium risk, and the remaining 39 (43.8%) as high risk sensu lato, of which 17 (43.6%; 19.1% of total) were categorized as moderately high risk, 11 (28.2%; 12.4% of total) as high risk, and another 11 (28.2%; 12.4% of total) as very high risk (Table I). The three highest scoring species were goldfish Carassius auratus, gibel carp Carassius gibelio, and common carp Cyprinus carpio carpio,

7 Effectiveness of FISK for Assessing Fishes in Iberia 7 Fig. 2. Mean scores (±SE and n) for 89 fish species assessed by FISK for the Iberian Peninsula and ranked according to their invasiveness and protection status (cf. Table I). Thresholds are: <1 (low risk) and (high risk sensu lato), with medium-risk species in between. Risk categories are: L = low risk: [ 15, 1]; M = medium risk: [1, 20.25[; MH = moderately high risk: [20.25, 25[; H = high risk: [25, 30[; VH = very high risk: [30, 57]. Fig. 1. Upper graph: Receiver operating characteristic (ROC) curves for three assessors (initials of the first three authors) on 89 fish species assessed by FISK for the Iberian Peninsula. Lower graph: Mean ROC curve based on mean scores from the three assessors, with smoothing line and confidence intervals of specificities (see Table I). followed by the other eight very high-risk species, crucian carp Carassius carassius, black bullhead Ameiurus melas, Nile tilapia Oreochromis niloticus, topmouth gudgeon Pseudorasbora parva, blue tilapia Oreochromis aureus, grass carp Ctenopharyngodon idella, redbelly tilapia Tilapia zillii, and channel catfish Ictalurus punctatus (Table I). Among species classified apriorias noninvasive, but ranked here as medium risk, were those from categories of higher protection status, that is, critically endangered and endangered (Fig. 2). The remaining noninvasive fishes were classified apriorias least concern, vulnerable, and not evaluated. One species (common carp) was classified apriorias both vulnerable (the original wild form) and invasive (the domesticated form), which received the third highest score and was thus ranked as very high risk (Table I; Fig. 2). The remaining invasive species were ranked as moderately high risk (least concern) and high risk (near threatened and not evaluated). Mean certainty in response for all species was 3.3 ± 0.3 SE (i.e., above the category mostly certain ) and CF 0.83 ± 0.07 SE, and ranged from a minimum of 2.7 ± 0.4 SE (CF: 0.69 ± 0.09 SE) for tinfoil barb Barbonymus schwanenfeldii to a maximum of 3.8 ± 0.3 SE (CF: 0.95 ± 0.07 SE) for eastern mosquitofish Gambusia holbrooki (Table I). Species categorizations were more similar (i.e., higher coincidence rate) between FR and PML (69%) relative to DA (46% with FR and 60% with PML). There were significant differences in mean CF values between DA and FR (0.95 vs. 0.77) and between DA and PML (0.95 vs. 0.77), but not between FR and PML (Table II).

8 8 Almeida et al. Table II. Permutational Analysis of Variance (PERANOVA) in Certainty Factor (CF) Between Three Assessors (Initials of the First Three Authors) for 89 Fish Species Assessed with FISK for the Iberian Peninsula (cf. Table I) Source df MS F # /t P # Assessor <0.001 DA vs. FR 12.2 <0.001 DA vs. PML 22.5 <0.001 FR vs. PML Residual Note: Statistically significant effects and higher-order interactions are followed by a posteriori pair-wise comparisons (α = 0.05). F # = permutational F-value; t = t-test value; P # = permutational p-value (9,999 permutations). 4. DISCUSSION This study indicated that FISK assessments can discriminate reliably between invasive and noninvasive species for the Iberian Peninsula, irrespective of the assessor. The results obtained here were consistent with mean cut-off (i.e., threshold) values across different regions of the globe, (16,18,19,21,35) where they have been found to range from 17 (southeastern Australia) to 21 (peninsular Florida). The exception to this is the Balkans Region, (36) an area of Mediterranean-type climate where a threshold of 9.5 was identified using ROC analysis of FISK scores. This was due to a particularly high number of translocations (i.e., fish introductions across catchments within the same country) in this region, and thus many species were considered as invasive for analysis, although they do not display a strong invasive profile (i.e., low FISK scores) at the global scale. The establishment of a common threshold is essential to aid in the management of invasive species with regard to transborder legislation (e.g., across the European Union). Thus, policymakers can set limits in national and international legislations by using the scores from FISK assessments to rank species and thus decide on permissions for transporting and the usage of non-native fishes. Results also showed that the Iberian Peninsula is a region highly threatened by invasive freshwater fishes, with no species classified as low risk, >55% as medium risk, and >12% as very high risk. The highest scoring species (i.e., at category very high risk ) were those commonly introduced worldwide, including cyprinids (goldfish, gibel carp, common carp, topmouth gudgeon, grass carp), ictalurids (black bullhead, channel catfish), and cichlids (Nile tilapia, blue tilapia, redbelly tilapia). Also, high scoring was crucian carp, which is threatened in many parts of its native range but introduced in some southern European countries. (10) Many of these fishes are well known as valuable ornamental species for aquaculture and the aquarium trade. (37,38) Among them, five carp species appeared, which despite some of them being considered as naturalized (i.e., introduced and established many decades or centuries ago), were considered by the assessors to pose extremely high risks to native species and ecosystems, mainly through habitat alteration such as increased turbidity and uprooting of vegetation, for example, common carp in Banyoles Lake, Spain. (39) Eleven species ranked as high risk sensu lato were common within the top 20 species previously identified as highly invasive for Iberia. (12) Six other species were considered to be invasive in the Iberian Peninsula that in this study were classified as medium risk. Conversely, three species, specifically gibel carp, sunbleak Leucaspius delineatus, and snakehead Channa argus, were ranked as high risk sensu lato using FISK but were not by Clavero. (12) Variations in rankings between the two approaches (i.e., potential misclassifications) is due to methodological differences (e.g., regarding the regional focus) that require further investigation to develop stronger assessments in consensus. As a precautionary measure, these potentially misclassified species should be included in a watch list that includes fish species likely to become new invaders in the Iberian Peninsula until further investigations have been undertaken to determine their most appropriate classification. Certainty in responses was relatively high ( mostly certain ), which suggests a similar level of knowledge among the assessors regarding the species examined. Mechanisms of impact are well known for some species such as eastern mosquitofish, particularly in the Iberian Peninsula, with a wide range of literature on this species. (40 42) Conversely, other recently detected species with uncertain introduction status are poorly known; such is the case of tinfoil barb, with the lowest certainty. Coincidence rates between assessors for species rankings were of 69% at the most. It is difficult to obtain rates higher than this, as scores can be similar among assessors, but if they are slightly above or below the limit of a particular interval, then the category will change among assessors. Differences were also clear for certainties, and this is in relation to the more or less cautious criteria for each assessor. (16) These discrepancies between assessors show the importance of using several independent researchers

9 Effectiveness of FISK for Assessing Fishes in Iberia 9 to get accurate and representative assessments for a particular region. In conclusion, overall results suggest that like elsewhere, FISK has proved to be a useful and viable tool for assessing risks posed by non-native fish in the Iberian Peninsula. Furthermore, this article contributes to widen the existing watch list proposed by Clavero (12) for this region. Nevertheless, efforts should be made to expand this list, specifically to species commonly used by leading vectors of introduction, in particular game and aquarium fishes. ACKNOWLEDGMENTS D. Almeida was funded through Cefas by grants from the U.K. Environment Agency and the U.K. Department of Environment, Food & Rural Affairs (Defra), F. Ribeiro by the Portuguese Science Foundation (SFRH/BPD/46761/2008), and the participation of G. H. Copp was made possible through the support of Defra. REFERENCES 1. Almaça C. Freshwater fish and their conservation in Portugal. Biological Conservation, 1995; 72: Elvira B. Conservation status of endemic freshwater fish in Spain. Biological Conservation, 1995; 72: Smith KG, Darwall WRT (eds). The Status and Distribution of Freshwater Fish Endemic to the Mediterranean Basin. Gland, Switzerland and Cambridge, UK: IUCN, Reyjol Y, Hugueny B, Pont D, Bianco PG, Beier U, Caiola N, Casals F, Cowx I, Economou A, Ferreira T, Haidvogl G, Noble R, de Sostoa A, Vigneron T, Virbickas T. Patterns in species richness and endemism of European freshwater fish. Global Ecology and Biogeography, 2007; 16: Hermoso V, Clavero M. Threatening processes and conservation management of endemic freshwater fish in the Mediterranean Basin: A review. Marine and Freshwater Research, 2011; 62: Economidis PS, Dimitriou E, Pagoni R, Michaloudi E, Natsis L. Introduced and translocated fish species in the inland waters of Greece. Fisheries Management and Ecology, 2000; 7: Copp GH, Stakėnas S, Davison PI. The incidence of nonnative fishes in water courses: Example of the United Kingdom. Aquatic Invasions, 2006; 1: Lusk S, Lusková V, Hanel L. Alien fish species in the Czech Republic and their impact on the native fish fauna. Folia Zoologica, 2010; 59: Leunda PM. Impacts of non-native fishes on Iberian freshwater ichthyofauna: Current knowledge and gaps. Aquatic Invasions, 2010; 5: Ribeiro F, Leunda PM. Non-native fish impacts on Mediterranean freshwater ecosystems: Current knowledge and research needs. Fisheries Management and Ecology, 2012; 19: Ribeiro F, Collares-Pereira MJ, Moyle PB. Non-native fish in the fresh waters of Portugal, Azores and Madeira Islands: A growing threat to aquatic biodiversity. Fisheries Management and Ecology, 2009; 16: Clavero M. Assessing the risk of freshwater fish introductions into the Iberian Peninsula. Freshwater Biology, 2011; 56: Kolar CS, Lodge DM. Ecological predictions and risk assessment for alien fishes in North America. Science, 2002; 298: Vila-Gispert A, Alcaraz C, García-Berthou E. Life-history traits of invasive fish in small Mediterranean streams. Biological Invasions, 2005; 7: Ribeiro F, Elvira B, Collares-Pereira MJ, Moyle P. Lifehistory traits of non-native fishes in Iberian watersheds across several invasion stages: A first approach. Biological Invasions, 2008;10: Copp GH, Vilizzi L, Mumford J, Fenwick GV, Godard MJ, Gozlan RE. Calibration of FISK, an invasiveness screening tool for non-native freshwater fishes. Risk Analysis, 2009; 29: Mastitsky SE, Karatayev AY, Burlakova LE, Adamovich BV. Non-native fishes of Belarus: Diversity, distribution, and risk classification using the Fish Invasiveness Screening Kit (FISK). Aquatic Invasions, 2010; 5: Onikura N, Nakajima J, Inui R, Mizutani H, Kobayakawa M, Fukuda S, Mukai T. Evaluating the potential of invasion by non-native freshwater fishes in northern Kyushu Island, Japan, using the Fish Invasiveness Scoring Kit. Ichthyological Research, 2012; 58: Troca DFA, Vieira JP Potential invasive non-native fish farmed in the coastal region of Rio Grande Do Sul, Brazil. Boletim do Instituto de Pesca do Estado de Sao Paulo, 2012; 38: Verbrugge LNH, Van der Velde G, Hendriks AJ, Verreycken H, Leuven RSEW. Risk classifications of aquatic non-native species: Application of contemporary European assessment protocols in different biogeographical settings. Aquatic Invasions, 2012; 7: Lawson LL Jr, Hill JE, Hardin S, Vilizzi L, Copp GH. Revisions of the Fish Invasiveness Scoring Kit (FISK) for its application in warmer climatic zones, with particular reference to peninsular Florida. Risk Analysis, 2012; doi: /j x. 22. Pheloung PC, Williams PA, Halloy SR. A weed risk assessment model for use as a biosecurity tool evaluating plant introductions. Journal of Environmental Management, 1999; 57: Copp GH, Garthwaite R, Gozlan RE. Risk Identification and Assessment of Non-Native Freshwater Fishes: Concepts and Perspectives on Protocols for the UK. Cefas Science Technical Report. Lowestoft, UK, Available at: Accessed on January 30, Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC, McCauley DE, O Neil P, Parker IM, Thompson JN, Weller SG. The population biology of invasive species. Annual Review of Ecology and Systematics, 2001; 32: Vilizzi L, Lawson LL Jr, Hill JF, Hardin S, Copp GH. Evaluation of the FISK (Fish Invasiveness Scoring Kit) scoring system: A scenario-based comparison. Unpublished manuscript. 26. Clavero M, García-Berthou E. Homogenization dynamics and introduction routes of invasive freshwater fish in the Iberian Peninsula. Ecological Applications, 2006; 16: Bewick V, Cheek L, Ball J. Statistics review 13: Receiver operating characteristic curves. Critical Care, 2004; 8: Youden WJ. Index for rating diagnostic tests. Cancer, 1950; 3: DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics, 1988; 44:

10 10 Almeida et al. 30. Vergara IA, Norambuena T, Ferrada E, Slater AW, Melo F. StAR: A simple tool for the statistical comparison of ROC curves. BMC Bioinformatics, 2008; 9: Venables WN, Smith DM, the R Development Core Team. An Introduction to R: Notes on R, a Programming Environment for Data Analysis and Graphics. R Core Team, Renken H, Mumby PJ Modelling the dynamics of coral reef macroalgae using a Bayesian belief network approach. Ecological Modelling, 2009; 220: Anderson MJ, Gorley RN, Clarke KR. PERMANOVA for PRIMER: Guide to Software and Statistical Methods. Plymouth, UK: PRIMER E Ltd., Britton JR, Cucherousset J, Davies GD, Godard M, Copp GH. Non-native fishes and climate change: Predicting species responses to warming temperatures in a temperate region. Freshwater Biology, 2010; 55: Vilizzi L, Copp GH. Application of FISK, an invasiveness screening tool for non-native freshwater fishes, in the Murray- Darling Basin (south-eastern Australia). Risk Analysis, 2012; doi: /j x. 36. Simonović P, Tošić A, VassilevM, ApostolouA, MrdakD, Ristovska M, Kostov V, Nikolić V, Škraba D, Vilizzi L, Copp GH. Risk identification of non-native freshwater fishes in four countries of the Balkans Region using FISK (unpublished manuscript). 37. Welcomme RL. International Introductions of Inland Aquatic Species. Rome: Food and Agriculture Organization of the United Nations, Fisheries and Aquaculture Department, Savini D, Occhipinti Ambrogi A, Marchini A, Tricarico E, Gherardi F, Olenin S, Gollasch S. The top 27 animal alien species introduced into Europe for aquaculture and related activities. Journal of Applied Ichthyology, 2010; 26: García-Berthou E. Size- and depth-dependent variation in habitat and diet of the common carp (Cyprinus carpio). Aquatic Sciences, 2001; 63: Caiola N, de Sostoa A. Possible reasons for the decline of two native toothcarps in the Iberian Peninsula: Evidence of competition with the introduced eastern mosquitofish. Journal of Applied Ichthyology, 2005; 21: Alcaraz C, García-Berthou E. Life history variation of invasive mosquitofish (Gambusia holbrooki) along a salinity gradient. Biological Conservation, 2007; 139: Alcaraz C, Bisazza A, García-Berthou E. Salinity mediates the competitive interactions between invasive mosquitofish and an endangered fish. Oecologia, 2008; 155: Leunda PM, Elvira B, Ribeiro F, Miranda R, Oscoz J, Alves MJ, Collares-Pereira MJ. International standardization of common names for Iberian endemic freshwater fishes. Limnetica, 2009; 28: