This item is only available as the following downloads:
1 EVALUATION OF THE FISH INVASIVENESS SCREENING KIT (FISK V2) FOR IDENTIFYING THE INVASIVENESS RISK OF NON NATIVE FRESHWATER FISH ES IN PENINSULAR FLORIDA By LARRY LEON LAWSON JR. A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIE NCE UNIVERSITY OF FLORIDA 201 4
2 2014 Larry Leon Lawson Jr
3 To my family
4 ACKNOWLEDGMENTS I thank the members of my supervisory committee. Dr. Jeff Hill has provided continued support and guidance throughout my studies. His ability to objectively make conclusions based on the dat a presented and encouragement for me to do the same has been essential to my development as a scientist. Dr. Gordon Copp traveled halfway around the world to visit our lab and help advance my project. His insights into professional development and scientific writing will serve me well in future endeavors. While in Gainesville, Drs. Mike Allen and Kai Lorenzen helped guide my studies as a graduate student and improve d my performance in the classroom I also thank those who assisted with my research. Mr. Scott Hardin spent many long days, emails, and phone conversations explaining the various intricacies of non native species management and risk analysis. For someone of his professional stature to spend that amount of time mentoring a graduate student speaks volumes about his character Dr. Lorenzo Vilizzi did much of the stati stical a nalysis for my research and, e ven though we have never met, he did not once hesitate to assist me. The multiple survey participants, whom I will not name, were essential to the completion of this research. Mr. Craig Watson, Director of the Tropical Aquacul ture Laboratory, took a chance and hired me as his farm manager. We all work hard to achieve success in our family for their continued support as I pursue my career.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 LIST OF ABBREVIATIONS ................................ ................................ ............................. 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 Background ................................ ................................ ................................ ............. 11 FISK Evaluation in Peninsular Florida ................................ ................................ .... 15 2 METHODOLOGY ................................ ................................ ................................ ... 18 Species Selection ................................ ................................ ................................ ... 18 A Priori Classification and Regional Risk Survey ................................ .................... 18 FISK Assessments and Scoring ................................ ................................ ............. 21 Calibration of Score Thresholds ................................ ................................ .............. 22 Assessor Scoring Variation and Uncertainty ................................ ........................... 23 3 RESULTS ................................ ................................ ................................ ............... 29 4 CONCLUSION ................................ ................................ ................................ ........ 45 APPENDIX A REVISIONS OF THE FISH INVASIVENESS SCORING KIT (FISK) FOR ITS APPLICATION IN WARMER CLIMATIC ZONES, WITH PARTICULAR REFERENCE TO PENINSULAR FLORIDA ................................ ........................... 54 Materials and Methods ................................ ................................ ............................ 55 Findings ................................ ................................ ................................ .................. 58 D iscussion ................................ ................................ ................................ .............. 82 Acknowledgements ................................ ................................ ................................ 84 B EXPERT INVASIVENESS SURVEY QUESTIONNAIRE ................................ ........ 86 LIST OF REFERENCES ................................ ................................ ............................... 90 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 100
6 LIST OF TABLES Table page 2 1 List of fishes assessed using FISK v2 with abbreviations for use in tables and figures. ................................ ................................ ................................ ................ 25 3 1 Number of species classified by the three risk ranking methods ........................ 33 3 2 Comparison of risk classification between the mean FISK v2 scores and the a priori invasiveness ranking ................................ ................................ .............. 33 3 3 FISK v2 assessment results for peninsul ar Florida ................................ ............. 34 A 1 List of the 49 questions making up FISK v2 with highlighted (in italics) changes relative to FISK v1 (cf. Copp et al. 2005a). ................................ .......... 68 A 2 FISK v2 protocol with responses given for Barcoo Grunter Scortum barcoo as a real world application of FISK v2 in a management scenario. .................... 74
7 LIST OF FIGURES Figure page 3 1 Mean FISK v2 scores for the 95 freshwater fish species assessed using FISK v2 for peninsular Florida with error bars indicating minimum and maximum range in scores for multi assessed species. Thresholds are: axis labels are abbreviations for species names as appears in Table 2 1. ................ 38 3 2 Frequency distributions of mean FISK v2 scores, plotted separately for a priori non invasive (dashed line) and invasive (solid line) fishes. The vertical solid line represents the high risk threshold value of 10.25 and the vertical dashed line represents the low risk threshold value of 0. ................................ ... 39 3 3 Percent survey responses for the 54 fishes that had a consensus (>51%) of responses in the low risk category. Mean FISK scores are shown by the black bars and are labeled on the secondary Y axis. ................................ ......... 40 3 4 Percent survey responses for the 14 fishes that had a consensus ( 50%) of survey responses in the medium risk category. Mean FISK scores are shown by black bar and lab eled on the secondary Y axis. ................................ ............ 41 3 5 Percent survey responses for the six fishes that had a consensus ( 50%) of survey responses in the high risk category. Mean FISK scores are shown by black bar and labeled on the secondary Y axis. ................................ ................. 42 3 6 Percent survey responses for the 21fishes that did not have a consensus K scores are shown by black bar and labeled on the secondary Y axis. ............... 43 3 7 Top: Receiver operating characteristic (ROC) curves for the three main assessors ( initials ) on 95 fish taxa assessed with FISK v2 for Peninsular Florida. Bottom: Mean ROC curve based on mean scores from all five assessors, with smoothing line and confidence intervals of specificities (see Table 3 3). ................................ ................................ ................................ .......... 44 A 1 New and improved interface of the Species Assessment Menu dialog in FISK v2 (data from Vilizzi and Copp 2013). ................................ ................................ 80 A 2 New and improved interface of the Q&A dialog in FISK v2 (Q 1 for Barcoo g runter Scortum barcoo). ................................ ................................ .................... 81
8 LIST OF ABBREVIATIONS A NSTF Aquatic Nuisance Species Task Force C F Certainty Factor E NSARS European Non native Species in Aquaculture Risk Assessment Scheme F ISK The Fish Invasiveness Scree ning Kit Fwc Florida Fish and Wildlife Conservation Commission N ISC National Invasive Species Council R OC Receiver Operating Characteristic Curve S E Standard Error W RA Australian Weed Risk Assessment
9 Abstract of Thesis Presented to the Graduate S chool of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EVALUATION OF THE FISH INVASIVENESS SCREENING KIT (FISK V 2) FOR IDENTIFYING THE INVASIVENESS RISK OF NON N ATIVE FRESHWATER FI SHES IN PENINSULAR FLORIDA By Larry Leon Lawson Jr. May 2014 Chair: Jeffrey E. Hill Major: Fisheries and Aquatic Sciences The Fish Invasiveness Screening Kit (FISK) is becoming a popular model for rapid risk identification with published applications now spanning the globe. Upgrades (i.e., FISK v2) were completed recently to ensure the incorporation of broader climatic zones for its application to the sub tropical climate of peninsular Florida. The goal of this study was to evaluate the ability of FISK v2 to identify the invasiveness of non native fishes in peninsular Florida. T he 95 fishes selected for screenings using were attributed an a priori invasiveness ranking from information provided by FishBase and ISSG. A n additional estimate of invasiveness was obtained from a survey of experts familiar with peninsular Florida habitats and ichthyofuana. Risk screenings were then completed separately and independently by up to five assessors resulting in one to five screenings hence FISK v2 outcome scores, p er taxon R eceiver operating characteristic (ROC) analysis identified a mean threshold value (i.e., cut off ) of 10.25 which when compared to the a priori invasiveness ranking classified correctly 76% of invasive fishes and 88% of non invasive fishes This threshold value was considerably lower than most other published calibrations of FISK emphasizing the importance of regionally focused risk
10 screening Further supporting these results, responses from the survey identified 9 of the 10 highest scoring speci es to be of elevated risk of invasiveness, the exception being goldfish Carassius auratus Overall, FISK v2 has proven that it would be a valuable tool for natural resource managers trying to make informed decisions related to the risks of non native speci es.
11 CHAPTER 1 INTRODUCTION Background Attempts to prevent the introduction and spread of invasive species have led to the development of science based risk assessment and hazard identification methods. The federal Aquatic Nuisance Species Task Force (ANSTF ) developed t he Generic Nonindigenous Aquatic Organisms Risk Analysis ( ) as a standardized framework to e stimate the risks of non native aquatic organisms (Orr 2003). The Generic Analysis has been the method by which federal risk assessments for non native aquatic species were carried out for the U.S. government (Hill and Zajicek 2007) as well as some states, such as Florida (Zajicek et al. 2009a, 2009b). Risk assessments of barramundi Lates calcarifer (Hill and Thompson 2008; Hardin 2009 ; Hardin and Hill 2012 ), blue tilapia Oreochromis aureus (Hardin et al. unpublished data), triploid grass carp Ctenopharyngodon idella (Zajicek et al. 2009b), and the marine ornamental trade pathway (Zajicek et al. 2009a ) have recently been completed in Florida using the Generic Analysis format T he Generic Analysis has benefits as a risk analysis tool (Orr 2003; Hill and Zajicek 2007); however, it relies heavily on qualitative measures from experts that may be subject to influence (e.g., assessor bias and error) and can cause a misconception that a species has been quantitatively scrutinized ( Kolar and Lodge 2002 ; Simberloff 2005; Simberloff et al. 2005 ) To avoid some of these limitations researchers have developed m ore quantitative methods of risk assessment such as the fish profiling approach of Kolar and Lodge (2002) and the identification of biological and ecological
12 traits of invasive species (Marchetti et al. 2004; Ruesink 2005 ) While t h ese methods show promise the y can be data intensive, time c onsuming, and region specific Risk screening systems are abbreviated form s of risk assessment designed to predict the potential invasiveness of a non native species when introduced in to new regions (Daehler and Carino 2000 ; NRC 2002; Lodge et al. 2006 ). While some risk screenings methods are mor e comprehensive than other s generally, r isk screen predictions are based on a sy nthesis of information covering the biology of the species, biogeographic al and climatic features of the invaded and origin regions, and ecological and evolutionary traits of both the species and regions in question (Pheloung et al. 1999). Components of risk screening methods include specific data requirements (i.e., question based format), peer revie w, repeatability, transparency (e.g. incorpora tes uncertainty and documentation ) and versatility (i.e., applicable to a variet y of differing taxonomic groups ) (Pheloung et al. 1999 ; Daehler and Carino 2000; Kolar and Lodge 2002 ; NRC 2002; Copp et al. 2005 a ; Lodge et al. 2006 ) Risk screens have a broad range of uses and implementation strategies. They can be particularly useful when att empting to distinguish between large suite s of potentially invasive and non invasive species in a timely and cost effective manner (Baker et al. 2008). Risk screens can be additionally beneficial because they are an expedient means of identifying gaps in knowledge and data quality (Copp et al. 2009), thus serving to inform management and research priorities. In some regions A ustralia for example, risk screens have been used to make regulatory decisions for the importat ion status of non native plants ( Pheloung et al. 1999; Weber et al. 2009) More commonly, risk screens are applied as an initial risk identification step to reco mmend whether or not the species in q uestion should be
13 subjected to more comprehensive risk assessment risk analysis, or risk management ( Kolar and Lodge 20 02; Daehler et al. 2004 ; Baker et al. 2008 ; Gordon et al. 2008a; Hill et al. In Press ) Support for the development of new risk screening methods by scientific and regulatory communities is evidenced by the National Research Council (NRC), National Invasive Species Council (NISC) and the ANSTF prioritizing the development of risk screening approaches i n their management plans ( NRC 2002; NISC 2008; ANSTF 2013) The Australian Weed Risk Assessment (WRA) by Pheloung et al. (1999) has become the model for semi quantitative risk screens Originally developed to assess the potenti al invasiveness of terrestri al plants proposed for importation in to Australia the WRA is b ased on the concept that weedy plants in other parts of the world are likely to be weedy when introduced to new areas of similar environmental co nditions (Pheloung et al. 1999). The WRA consist s of a set of 49 questions covering the biogeography, invasive history, biology and ecology, and presence/absence of undesirable traits of plants Responses to the 49 questions are assigned numerical values which are utcom he outcome score is then used to assign the assessed plant species to a risk category The risk categories are based on pre determined score thresholds which re sult in the species either being accepted (score < 1) or rejected (score > 6) for imp ortation int o Australia or recommended for fur ther evaluation (1 < score < 6). The criteria for assigning these score thresholds were (1) a rejection of all historically invasive plants, (2) less than 10% rejection of non invasive plants, and (3) no more t han 30% of assessed plants assigned to the evaluate further category (Pheloung et al. 1999). The success of the WRA in Australia and N ew
14 Zealand has led to its adaptation for use in other regions of the world, including Hawaii, USA ( Daehler and Carino 2000 ), the Pacific Islands (Daehler et al. 2004), the Czech o et al. 2006), and Florida, USA (Gordon et al. 2008a). The accuracy of the WRA across these various geographical locatio ns was subsequently assessed and determined to be sufficient for broad er use to screen non native species to reduce the likelihood of nega tive impacts by invasive plants (Gordon et al. 2008b). Following the successful application of this framework to plan ts, t he WRA w as direct ly adapt ed for fishes and invertebrates (both freshwater and marine) as well as Amphibia (Copp et al. 2005 a ). The first of these to be t est ed w as the Fish Invasiveness Screening Kit (FISK), which was applied to assess the likelihood o f non native freshwater fishes being invasive in the U nited K ingdom Although conceptually unchanged from the WRA model (Copp et al. 2005a, 2005b), some modifications were required during the creation of the FISK due to differences of invasive characterist ics between fishes and plants (Bruton 1986; Ricciardi and Rasmussen 1998; Reichard 2001). Similar to the WRA, responses to each of the 49 species specific FISK questions are assigned a numerical value, gener ally ranging from 1 to 2, which are summed to pr oduce the total outcome score. For the original version of FISK, assessed fish species were assigned to a risk category based on the outcome score and the same WRA score thresholds were used C alibration of FISK score thresholds for regulated non native fishes in the UK was completed by Copp et al. ( 2009) in order to determine the most appropriate sc ore
15 distinguishing between the medium and high risk categories. While adapt ing FISK from the WRA (Copp et al. 2005a), the tool was improved by requiring the assessor to provide a justification for each response (e.g., literature reference, database citation) and to indicate their level of confidence (or uncertainty) with each of their responses (Copp et al. 2009). The FISK program was made publicly available free of charge a s a consequence FISK has been trialed in at least 11 countries spanning five continents (Australia, Belarus, the Balkans, Belgium, Brazil, Finland, Florida, Iberia, Japan, M exico and Turkey ) Furthermore, FISK is used as a screening complement to the Great Britain risk scheme for non native species (Baker et al. 2008) as well as the European Non Native Species in Aquaculture Risk Assessment Scheme (ENSARS) (C opp et al. 2008). Initial comparisons of FISK with other, shorter, pre screening tools (i.e., Parrott et al. 2009; Verbrugge et al. 2010), as well as the applications outside of the U.K. (Mastitsky et al. 2010; Mendoza 2011; Onikura et al. 2011), suggest t hat the FISK provides a viable and consistent means of identifying potentially invasive freshwater fishes across various geographical settings. FISK Evaluation in Peninsular Florida ical to warm temperate), diverse freshwater habitat types, and relatively limited native freshwater ichthyofauna has enabled the successful establishment of many non native fishes that are not found elsewhere in the United States (Hill 2002). To date, ther e are over 100 non native fish species reported as introduced and at least 3 7 are reproducing or established (Fuller et al. 1999; Shafland et al. 2008 ; USGS 2014 ). Additionally, the aquarium trade, aquaculture industry, and live fish markets in peninsular Florida are
16 fueled by the demand for non native freshwater fishes All of these pathways pose a risk of introduction of non native fish that may threaten native aquatic ecosystems (Copp et al. 2005c; Hardin 2007; Zajicek et al. 2009a; Gozlan et al. 2010). Conducting a full risk analysis on every species in regions facing issues similar to Florida would be unrealistic and not a cost effective mean s of assessing risk This assertion is demonstrated by the fact that only a handful of species have undergone a risk analysis specific to Florida (FDACS 2000; Zaji cek et al. 2009a, 2009b; Hardin and Hill 2012) Risk screening s could provide a replicable, transparent, and scientifically supported means to assess a large suite of species (Pheloung et al. 1999; Lodge e t al. 2006) Improved management of non native species would be expedited by more efficient risk screening approaches like the FISK. For many species, a risk screening could be sufficient to determine whether or not they pose an acceptable level of risk (H ill et al. In Press ) F or others, barramundi for example, a more detailed risk analysis may be required to determine the best course of regulatory action (Hardin and Hill 2012) For application in peninsular Florida (i.e., the portion of the state south o f the Suwannee River) the original FISK program developed by Copp et al. ( 2005 a) was recently updated to produce the FISK v2 (Lawson et al. 2013) Although already published, this work was a central aspect of my thesis and is instrumental to this evaluati on of FISK v2, thus it is included here (Appendix A). The FISK format was unchanged in that there remains a total of 49 species specific questions covering two main sections (biogeography/invasive history and biology/ecology) and eight topics: (1) producti on; (2) c limate and distribution; (3) undesirable attributes; (4) invasive traits; (5) feeding guilds; (6) reproduction; (7) dispersal; (8) persistence attributes (Copp et al.
17 2005 a ). However, substantial revisions to the questions and related guidance al ong with upgrades to the software and graphic al user interface produced FISK v2 ( Appendix A ) Several of the original FISK questions were modified specifically for applications in warm temperate and sub tropical areas, i.e., peninsular Florida. For a full list of the updated questions and the new FISK v2 program see Appendix A. Subsequent evaluations of the new FISK v2 (for specific regions) have been completed in multiple countries These evaluations (sum marized in Copp et al. 2013) included Iberia ( Almei da et al. 2013), southern Finland (Puntila et al. 2013), Turkey (Tarkan et al. 2013), the 2013 ), and the Murray Darling Basin, Australia (Vilizzi and Copp 2013 ). The questions in FISK and their guidance have been adapted to accomm odate peninsular Florida and sub tropical climates in general (Lawson et al. 2013). However, the accuracy of FISK to categorize the invasiveness risk of non native fishes in peninsular Florida has not been fully tested. Therefore, t he goal of this study is to undertake the first comprehensive evaluation of FISK v2 in the United States with specific reference to the sub tropical climate of peninsular Florida and to determine its potential use in management. My specific objectives were to : (1) calibrate the F ISK v2 for peninsular Florida by determining t he most appropriate threshold for distinguishing between fishes of medium and high risk of invasiveness ; (2 ) evaluate sources of variation among the scoring and certainty patterns of the assessors and survey re spondents; and (3) i nterpret the FISK assessment outcomes relative to an a priori risk classification according to FishBase and to a risk survey of fisheries professionals.
18 CHAPTER 2 METHODOLOGY Species Selection In total, 95 fish taxa comprising 89 spec ies, four hybrids, and two genera (i.e., those only identified to the genus level), were assessed using the FISK v2 program with peninsular Florida as the risk assessment area of concern (Table 2 1). For brevity scientific names, common names and abbrevia tions are provided (Table 2 1). T he freshwater fish taxa were selected for assessment if (1) they were not considered native to peninsular Florida and (2) they had conclusive documentati on of introduction into public waters (i.e., not on a private facility or fish farm) of the risk assessment area The finalized list was based on a review of literature that detailed specific records of non native fish introductions Primary data sources included Fuller et al. (1999), Shafland et al. (2008) the U.S. Geologi cal Survey Nonindigenous Aquatic Species datab ase (USGS 2014 ) and unpublished data from experts. The available data on some i ntroduced species in Florida was equivocal and revision was necessary. Therefore, species were excluded if they lacked conclusive evidence of legitimate introduction (e.g., unconfirmed taxonomic identification, inconclusive collection data, unusual hybrids, and duplicate records). A Priori Classification and Regional Risk Survey An invasive species is a non indigenous species (a spec ies outside its historical range) the introduction of which is known to cause some form of negative ecological, economic, or social impact (Copp et al. 2005c). The ability of the FISK v2 to identify correctly invasive fishes as high risk and noninvasive fi shes as not high risk provides a measure of the tool H owever, to verify accuracy, there must first be a
19 a prio invasiveness wi th which the FISK v2 outcomes can be tested against. Ideally, for ROC analysis, this a priori invasiveness would be a nvasiveness status of fishes is not currently available. Therefore, a priori estimates of invasiveness were obtained for each of the 95 assessed non native fish es based on information available from the Invasive Species Specialist Group database ( www.issg.org ) and from FishBase ( www.fishbase.org ). While neither database directly lists a species as invasive or non invasive to a specific region, both categorize fishes based on worldwi de documentation of introductions that have establishmen t status in non native regions with information on the ecology, distribution, m anagement and impact The 95 non native fishes assessed here were categorized as either invasive or non invasive qualitatively by combining the information provided in these two databases. In the case of fishes identified only to genus level and hybrids, a priori classification was based on the taxon with highest invasiveness risk. Separate from the a priori invasiveness standard obtained from FishBase and ISSG was a regionally focu sed estimate of invasiveness obtained from a written survey of fisheries scientists, ecologists, and natural resource managers (n = 25) from state and federal agencies and academia who are knowledgeable of non native fishes and habitats of Florida (App endix B) The survey was not a FISK assessment and was not
20 used to calibrate statistically the FISK v2 score thresholds ; instead, it asked explicitly the participants to individually rate (using a combination of their personal expertise and published liter ature) the risk (low, medium, or high) of each non native species to cause moderate to severe negative impacts in peninsular Florida. Impacts were defined as (environmental, economic, asked to rate the fish es based solely on their impacts as they relate to peninsular particular fish species. T he survey allowed participants to write additional comments about the species in question. Finally, the participant s were asked to select the information source used to make their decision. R esponses (at least four per species) to this survey were combined to provide a region specific estimate of invasivenes s (low risk, medium risk or high risk) based on at least a 51% majority of responses in a single risk category. Species that did not have a majority of responses in a single risk category n were then made of the survey outcomes to the FISK assessmen t outcomes and to the a priori estimate of invasiveness risk. Similar projects have used expert opinion to obtain an external es timate of invasiveness to calibrate adaptations of the WRA (Pheloung et al. 1999; Daehler et al. 2004; 2006) and for other non native fish risk assessments (Kolar and Lodge 2002; CISAC 2010). The overall FISK assessments and various risk classification methods were also compared to the Florida Fish and status of non native species in Florida ( http://www.myfwc.com/nonnatives ).
21 FISK Assessments and Scoring U sing the FISK v2 program, assessments of the 95 fish es were completed separately and independently by three main assessors ( Larry Lawson [ LL ] Jeffrey Hill [ JH ] and Scott Hardin [ SH ] ) and two additional assessors ( Gordon Copp [GC] and L orenzo V ilizzi [LV] ) Ultimately, each assessor completed a subset of the total 95 fishes due to k nowledge and time constraints, resulting in one to five assessments per taxon (Table 2 I). Of the 95 fishes analayzed 94 (98.9% of total ) were assessed by LL, 45 (47.4%) by J H, 42 (44.2%) by SH, 11 (11.6%) by G C and 3 (3.2%) by LV (Table 2 1 ). The FISK questions require categorical responses low/medium/high) based on best available data (e.g., peer reviewed journals, textbooks, and internet sources) and in terpretation of the quality, and reliability of the data sources However, a ssessors were not instructed explicitly to use specific sources except when directed by the FISK question guidance. For example, FISK v2 questions and guida nce for climate matching 8) suggest using the Kppen Geiger classifications (Peel et al. 2007) which for peninsular Florida include s three climate types: equatorial summer dry, equatorial monsoonal and warm t emperate fully humid summer hot Ultimately, assessor judgment f or climate matching between the species in question and the risk assessment area (i.e., peninsular Florida) was encouraged when the assessor(s) had extensive experience beyond that of generalized climate models. FISK o utcome scores can range from 15 to 5 7 (Vilizzi, L; personal communication ) and lower scores indicate lower r isk and higher scores indicate greater risk. Although these outcome scores fall within a continuum, past applications of FISK (Copp et al. 2009) FI ISK (Tricarico et al. 2010) and th e WRA (Pheloung et al. 1999)
22 used score thresholds to place species in a specific risk category The calibra ted version of FISK for the UK used a thresh to distinguish high risk species with addition al risk categories of low risk (score < 1) and medium risk < 19) (Copp et al. 2009). Similar risk categories and threshol d scores were determined by recent FISK applications in Bela rus (Mastitsky et al. 2010), Kyushu Isla nd, Japan (Onikura et al. 2011) and more recently the Murray Darling Basin, Australia (Vilizzi and Copp 201 3 ). To facilitate compar ison with these past calibration attempts fishes were placed into risk categories re ferred to as low ris k, medium risk and high risk, and the same calibration procedure (described in Copp et al. 2009) was used to determine the most appropria te threshold score between the medium risk and high risk categories. Calibration of Score Threshol ds Receiver operating characteristic curve (ROC) analysi s (Bewick 2004) was used to assess the predictive ability of FISK to discriminate between invasive and non invasive taxa. Statistically, a ROC curve is a graph of sensitivity (proportion of a priori i nvasive fishes classified by FISK v2 as high risk ) vs 1 specificity (proportion of a priori non invasive fishes classified by FISK v2 as high risk ) or, alternatively, sensitivity vs specificity true negati ves A measure of the correct classifications resulting from the calibration analysis is the area under the ROC curve (AUC). If the AUC is equal to 1.0 that is, of two straight lines, one vertical from 0,0 to 0,1 a nd 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 where non invasive taxa are categorized as invasive nor false negatives where invasive taxa are c ategorized as non invasive. Conversely, if the AUC
23 is equal to 0.5 then the test is of no diagnostic value ( i.e., it is equally likely to indicate a fish is invasive or non invasive ) Typically, the AUC will range between 0.5 and 1.0, where the closer the AUC to 1.0 the better the ability of FISK to differentiate between invasive and non invasive taxa. Separate ROC curves were generated for each of the three main assessors and an overall ROC curve was eve ntually computed on the mean scores from all five assessors. Based on the overall ROC curve, the best FISK threshold which maximi zes the true positive rate ( a priori invasive classified as invasive) and minimizes the false posit ive rate ( a priori non invas ive classified as invasive) was determined using both J statisti c (maximum value of [sensitivity + specificity 1]) (Youden 1950) a nd the point of the ROC curve closest to the point on the axes that indicate perfect sensitivity and specificity. F or both assessor specific and overall ROC curves, bootstrapped confidence intervals were computed for the corresponding AUCs (Delong et al. 1988) F or the overall ROC curve a smoothed curve was also generated including bootstrapped confidence intervals of specificities along the entire range of sensitivity x64 2.13.0 (R Core Team 2008) using 2000 bootstrap replicates Assessor Scoring Variation and Uncertainty For mult iple assessed fishes between maximum and minimum scores, and the Pearson product correlation test was carried out between delta values and number of assessors per species. As each response of FISK for a given taxon was allocated a certainty score
24 (1 = very uncertain; 2 = mostly uncertain; 3 = mostly certain; 4 = very certain), a for the total FISK score was computed as: i )/(4 49) ( i where CQ i is the certaint y for question i 4 is the maximum achievable value for certainty Therefore, t he CF ranges from a minimum of 0.25 where all 49 questions exhibit certainty score equal t o 1 to a maximum of 1 where all 49 questions exhibit a certainty score equal to 4.
25 Table 2 1. List of fishes assessed using FISK v2 with abbreviations for use in tables and figures. # Family Taxon N ame Common Name Abbreviation Authority 1 Cichlidae Ae quidens pulcher Blue acara Apul AFS Gill, 1858 2 Cichlidae Amatitlania nigrofasciata Convict cichlid Anig AFS Gnther, 1867 3 Cichlidae Amphilophus citrinellus Midas cichlid Acit AFS Gnther, 1864 4 Anabantidae Anabas testudineus Climbing perch Ates AFS Bloch, 1792 5 Characidae Aphyocharax anisitsi Bloodfin tetra Aani AFS Eigenmann & Kennedy, 1903 6 Cichlidae Astatotilapia calliptera Eastern happy Acal Vernacular Gnther, 1894 7 Cichlidae Astronotus ocellatus Oscar Aoce AFS Agassiz, 1831 8 Cyprin idae Barbonymus schwanenfeldii Tinfoil barb Bsch AFS Bleeker, 1854 9 Poeciliidae Belonesox belizanus Pike killifish Bbel AFS Kner, 1860 10 Belontiidae Betta splendens Siamese fighting fish Bspl Vernacular Regan, 1910 11 Callichthyidae Callichthys callic hthys Cascarudo Ccal AFS Linnaeus, 1758 12 Cyprinidae Carassius auratus Goldfish Caur AFS Linnaeus, 1758 13 Channidae Channa argus argus Northern snakehead Carg AFS Cantor, 1842 14 Channidae Channa marulius Bullseye snakehead Cmar AFS Hamilton, 1822 15 Notopteridae Chitala ornata Clown knifefish Corn AFS Hamilton, 1822 16 Cichlidae Cichla ocellaris Butterfly peacock bass Coce AFS Bloch & Schneider, 1801 17 Cichlidae Cichla temensis Speckled peacock bass Ctem AFS Humboldt, 1821 18 Cichlidae Cichlasoma bimaculatum Black acara Cbim AFS Linnaeus, 1758 19 Cichlidae Cichlasoma salvini Yellowbelly cichlid Csal AFS Gnther, 1862 20 Cichlidae Cichlasoma trimaculatum Threespot cichlid Ctri AFS Gnther, 1867 21 Cichlidae Cichlasoma urophthalmum Mayan cichli d Curo AFS Gnther, 1862 22 Clariidae Clarias batrachus Walking catfish Cbat AFS Linnaeus, 1758 23 Characidae Colossoma macropomum Black pacu Cmac Cuvier, 1816 24 Cyprinidae Ctenopharyngodon idella Grass carp Cide AFS Valenciennes, 1844 25 Anabantidae Ctenopoma nigropannosum Twospot climbing perch Cnig Vernacular Reichenow, 1875 26 Cyprinidae Cyprinus carpio Common carp Ccar AFS Linnaeus, 1758 27 Cyprinidae Danio rerio Zebra danio Drer AFS Hamilton, 1822 28 Cyprinidae Devario malabaricus Malabar dan io Dmal AFS Jerdon, 1849
26 Table 2 1. Continued # Family Taxon N ame Common Name Abbreviation Authority 29 Clupeidae Dorosoma petenense 1 Threadfin shad Dpet AFS Gnther, 1867 30 Poeciliidae Gambusia affinis Western mosquitofish Gaff AFS Baird and Girar d, 1853 31 Characidae Gymnocorymbus ternetzi Black tetra Gter AFS Boulenger, 1895 32 Helostomatidae Helostoma temminkii Kissing gourami Htem AFS Cuvier, 1829 33 Cichlidae Hemichromis letourneuxi African jewelfish Hlet AFS Sauvage, 1880 34 Cichlidae Her ichthys cyanoguttatus Rio Grande cichlid Hcya AFS Baird & Girard, 1853 35 Cichlidae Heros severus Banded cichlid Hsev AFS Heckel, 1840 36 Erythrinidae Hoplias malabaricus Trahira Hmal AFS Bloch, 1794 37 Callichthyidae Hoplosternum littorale Brown hoplo Hlit AFS Hancock, 1828 38 Cyprinidae Hypophthalmichthys nobilis Bighead carp Hnob AFS Richardson, 1845 39 Loricariidae Hypostomus sp. 2 Suckermouth catfishes Hsp AFS Linnaeus, 1758 40 Cyprinidae Labeo chrysophekadion Black sharkminnow Lchr AFS Bleeker, 1 849 41 Anostomidae Leporinus fasciatus Banded leporinus Lfas AFS Block, 1794 42 Mastacembelidae Macrognathus siamensis Spotfin spiny eel Msia AFS Gnther, 1861 43 Belontiidae Macropodus opercularis Paradisefish Mope AFS Linnaeus, 1758 44 Characidae Met ynnis sp. 3 Silver dollar Msp AFS Kner, 1858 45 Cobitidae Misgurnus anguillicaudatus Oriental weatherfish Mang AFS Cantor, 1842 46 Characidae Moenkhausia sanctaefilomenae Redeye tetra Msan AFS Steindachner, 1907 47 Synbranchidae Monopterus albus Asian sw amp eel Malb AFS Zuiew, 1793 48 Moronidae Morone chrysops x Morone saxatilis Wiper/sunshine bass hybrid Mchr AFS Rafinesque, 1820 49 Cichlidae Oreochromis aureus Blue tilapia Oaur AFS Gnther, 1889 50 Cichlidae Oreochromis mossambicus Mozambique tilapia Omos AFS Peters, 1852 51 Cichlidae Oreochromis niloticus Nile tilapia Onil AFS Linnaeus, 1758 52 Osteoglossidae Osteoglossum bicirrhosum Arawana Obic AFS Cuvier, 1829 53 Doradidae Oxydoras niger Ripsaw catfish Onig AFS Valenciennes 1821 54 Cobitidae P angio kuhlii Coolie loach Pkuh AFS Valenciennes 1846 55 Cichlidae Parachromis managuensis Jaguar guapote Pman AFS Gnther 1867
27 Table 2 1. Continued # Family Taxon name Common Name Abbreviation Authority 56 Cichlidae Paraneetroplus melanurus x P zonatus Redhead x Oaxaca cichlid Pmel AFS Meek, 1905 57 Characidae Piaractus brachypomus Red bellied pacu Pbra AFS Cuvier, 1817 58 Cyprinidae Pimephales promelas Fathead minnow Ppro AFS Rafinesque, 1820 59 Doradidae Platydoras costatus Raphael catfish Pcos AFS Linnaeus, 1758 60 Schilbeidae Platytropius siamensis Siamese schilbeid catfish Psia AFS Sauvage, 1883 61 Poeciliidae Poecilia latipinna x P. velifera Sailfin x Yucatan molly Plve AFS Lesueur, 1821 62 Poeciliidae Poecilia latipunctata lly Plat AFS Meek, 1904 63 Poeciliidae Poecilia petenensis Petn molly Ppet AFS Gnther, 1866 64 Poeciliidae Poecilia reticulata Guppy Pret AFS Peters, 1859 65 Poeciliidae Poecilia sphenops Mexican molly Psph AFS Valenciennes, 1846 66 Polypteridae Poly pterus delhezi Barred bichir Pdel AFS Boulenger, 1899 67 Doradidae Pterodoras granulosus Granulated catfish Pgra AFS Valenciennes, 1821 68 Cichlidae Pterophyllum scalare Freshwater angelfish Psca AFS Schlutze, 1823 69 Loricariidae Pterygoplichthys anisi tsi Southern sailfin catfish Pani AFS Eigenmann & Kennedy, 1903 70 Loricariidae Pterygoplichthys disjunctivus Vermiculated sailfin catfish Pdis AFS Weber, 1991 71 Loricariidae Pterygoplichthys multiradiatus Orinoco sailfin catfish Pmul AFS Hancock, 1828 72 Cyprinidae Puntius conchonius Rosy barb Pcon AFS Hamilton, 1822 73 Cyprinidae Puntius gelius Golden barb Pgel AFS Hamilton, 1822 74 Cyprinidae Systomus tetrazona Sumatra barb Stet Vernacular Bleeker, 1853 75 Characidae Pygocentrus nattereri Red pira nha Pnat AFS Kner, 1858 76 Ictaluridae Pylodictis olivaris Flathead catfish Poli AFS Rafinesque, 1818 77 Pimelodidae Rhamdia quelen South American catfish Rque AFS Quoy & Gaimard, 1824 78 Cichlidae Rocio octofasciata Jack Dempsey Roct AFS Regan, 1903 7 9 Percidae Sander vitreus Walleye Svit AFS Mitchill, 1818 80 Cichlidae Sarotherodon melanotheron Blackchin tilapia Smel AFS Rppell, 1852 81 Characidae Serrasalmus rhombeus Redeye piranha Srho AFS Linnaeus, 1766 82 Cichlidae Thorichthys meeki Firemouth cichlid Tmee AFS Brind, 1918 83 Cichlidae Tilapia buttikoferi Hornet tilapia Tbut Vernacular Hubrecht, 1881 84 Cichlidae Tilapia mariae Spotted tilapia Tmar
28 Table 2 1. Continued # Family Taxon N ame Common Name Abbreviation Authority 85 Cichlidae Ti lapia zillii Redbelly tilapia Tzil AFS Gervais, 1848 86 Belontiidae Trichogaster fasciata Banded gourami Tfas AFS Bloch & Schneider, 1801 87 Belontiidae Trichogaster labiosa Thick lipped gourami Tlab FAO Day, 1877 88 Belontiidae Trichogaster lalius Dwar f gourami Tlal AFS Hamilton, 1822 89 Belontiidae Trichogaster leerii Pearl gourami Tlee AFS Bleeker, 1852 90 Belontiidae Trichopodus trichopterus Three spot gourami Ttri FAO Palla, 1770 91 Belontiidae Trichopsis vittata Croaking gourami Tvit AFS Cuvier, 1831 92 Poeciliidae Xiphophorus hellerii Green swordtail Xhel AFS Heckel, 1848 93 Poeciliidae Xiphophorus hellerii x X. maculatus Red swordtail hybrid Xhma AFS Heckel, 1848 94 Poeciliidae Xiphophorus maculatus Southern platyfish Xmac AFS Gnther, 1866 95 Poeciliidae Xiphophorus variatus Variable platyfish Xvar AFS Meek, 1904 1 Dorosoma petenense is cryptogenic in peninsular Florida 2 A priori invasiveness status for Hypostomus plecostomus 3 A priori invasiveness status for Metynnis argenteus
29 CHAPTER 3 RESULTS Given the newly calibrated high risk threshold (see below) and based on mean scores, FISK v2 assessments resulted in 18 (18.9%) fishes categorized as low risk, 52 (54.7%) as medium risk, and the remaining 25 (26.3%) as high risk (Table 3 1) When compared to the invasive/non invasive a priori determination, FISK v2 assessments classified correctly 76% of the invasive fishes and 88% of the non invasive fishes (Table 3 2) Osteoglossum biccirhosum ) to a high of 42 (common carp Cyprinus carpio ) (Table 3 3 ) Mean o utcome (spotfin spiny eel Macrognathus siamensis ) to 31.8 (common car p ) (Figure 3 1 ) The five highest scoring fishes were common carp, followe d by goldfish Carassius auratu s walking catfish Clarias batrachus sucker catfishes Hypostomus sp. (cf. Hypostomus plecostomus ) and grass carp (Figure 3 1) The five lo west scoring fishes were spotfin spiny eel, banded leporinus Leporinus fasciatus cooli e loach Pangio kuhlii Siamese schilbeid catfish Platytropius siamensis and bloodfin tetra Aphyocharax anisitsi (Figure 3 1) Among the medium risk fishes were members from several families including 13 fishes from Cichlidae, five from Cyprinidae, and six from Poecili i dae When grouped phylogenetically, FISK assessments were carried out on 26 fishes from the Cichlidae family followed by 12 Cyprinid fishes, 11 Poeciliid fishes, and 11 fishes from the Order Siluriformes which resulted in group means and stan dard errors (SE) of 7.5 1.15, 10.8 3.65, 7.3 2.00, and 12.2 3.39, respectively. Another group of fishes that are of interest to fisheries managers and businesses in Florida are the small ornamental fishes that are popular in the aquarium trade. Twe nty nine fishes from
30 five families (i.e., Belontiidae, Characidae, Cyprinidae, Helostomatidae, and Poeciliidae) that could be considered small ornamental fishes were assessed. Mean FISK scores for this group ranged from 4 to 16 with an overall group mean of 2.5 0.83 SE. This resulted in nine (31%) fishes ranked as low risk, 18 (62%) as medium risk, and two (7%) as high risk. All 29 fishes in this group were considered low risk by the consensus rankings from the expert survey and five were considered inva sive by the a priori invasiveness ranking. In total, the a priori classification using database information yielded 74 fishes as non invasive and the remaining 21 fishes as invasive (Table 3 1). For the 74 fishes classified a priori as non invasive mean FISK scores ranged from 6.0 to 25.0 (Table 3 3) with an overall group mean of 4.9 0.79 SE. Mean FISK scores for this group resulted in 18 (24.3%) fishes ranked as low risk, 47 (63.5%) as medium risk, and 9 (12.2%) as high risk ( Figure 3 2 ). For the 21 f ishes classified a priori as invasive, mean FISK scores ranged from 1.7 to 31.8 (Table 3 3) with an overall group mean of 15.3 1.79 SE. Mean FISK scores for the invasive group resulted in no fishes ranked as low risk, 5 (23.8%) ranked as medium risk, and the remaining 16 (76.2%) ranked as high risk (Fig ure 3 2 ). Thus, in general the FISK scores matched the risk categories that were done a priori Eleven of the 25 (44%) invasiveness risk surveys were completed. The returned surveys represented current and retired professionals working for local, state, and federal agencies as well as academia (University of Florida). Each of the 95 fishes received a response from at least four survey respondents and were given a final invasiveness estimate of low risk, medi um risk, high risk, or non consensus. A
31 consensus estimate of the survey responses resulted in 54 (56.8%) fishes categorized as low risk (Figure 3 3 ), 14 (14.7%) fishes as medium risk (Figure 3 4 ), and 6 (6.3%) fishes as high risk (Figure 3 5 ). However, co nsensus (> 50% of survey responses in agreement) was not reached for 21 (22.1%) fishes (Figure 3 6 ). The six high risk fishes were hornet tilapia Tilapia buttikoferi, trahira Hoplias malabaricus red piranha Pygocentrus nattereri redeye piranha Serrasalmu s rhombeus northern snakehead Channa argus and redbelly tilapia Tilapia zillii Of these six fishes only the redbelly tilapia was ranked as invasive by the a priori methodology used for this calibration of FISK v2 Due to the lack of agreement amongst t he survey respondents, the survey invasiveness rankings were not used to calibrate the FISK v2 scoring thresholds; however, it was used for general comparisons to the a priori classifications and to the FISK v2 scoring outcomes. Agreement among the a prior i classification and the survey was highest f or the low risk fishes. Only nine ( 16.7 % ) of the 54 fishes considered low risk by the survey were categorized conversely by the a priori methodology as invasive. M embers of this group that also scored by FISK as high risk sensu lato include goldfish Western mosquitofish Gambusia affinis Guppy Poecilia reticulata fathead minnow Pimephales promelas and convict cichlid A mititalania nigrofasciatia Three (21.4%) of t he 14 fishes that were categorized as medium ri sk by the survey were classified a priori as invasive, of which only the blackchin tilapia Sarotherodon melanotheron had a mean FISK score less than the 10.25 high risk threshold. The remaining 11 medium risk fishes had mean FISK scores ranging from 4.0 t o 24.5 with only a single species (arowana ) attaining a low risk FISK score a nd four fishes scoring above the high risk threshold. Of the 21
32 fishes where survey consensus was not reached, 12 scored high risk by FISK and seven of those were ranked a priori as invasive including jaguar guapote Parachromis managuense blue tilapia, Nile tilapia Oreochromis niloticus vermiculated sailfin catfish Pterygoplichthys disjunctivus bighead carp Hypophthalmichthys nobilis walking catfish, and common carp. The ROCs for the three main assessors were overall similar and this was reflected in the corresponding AUCs (0.803, 0.900 0.705 95% CI for LLL; 0.846, 0.961 0.732 95% CI for JEH; 0.809, 0.993 0.625 95% CI for SH) (Figure 3 7 top). For the overall ROC curve, an AUC of 0.847 (0.943 0.752, 95% CI) was obtained, and this indicated that FISK w as able to discriminate accurately between most fishes classified a priori as invasive and non invasive (Figure 3 7 J and closest point statistics provided a bes t threshold of 10.25, which was then chosen as the calibration threshold to determine FISK risk outcomes for p eninsular Florida (Table 3 3 ). ( fishes with FISK scores bet ween 1 and 10.25 sensu lato fishes ( fishes with FISK scores between 10.25 and 57 ) T fishes ( 15 and 1 ) was retained from previous FISK applications because this threshold D elta scores from multiple a ssessed fishes (n=73) ranged from 0.0 to 24.0 and a statistically significant correlation was found between delta scores and the number of assessors ( p = 0.517, t = 5.156, P < 0.001), with delta values increasing with the number of assessments. For two species, blackchin tilapia and red bellied pacu Colossoma macropomum, the individual assessor scores would have changed the risk
33 ratings from high to low risk ( Figure 3 1) Mea n certainty of all response s using FISK v2 was 3.52 0.11 0.03 SE, ranging from a minimum of 3.15 0.3 SE (CF: 0 .79 0.07 SE) for the silver dollar Metynnis sp. (cf. M. argenteus ) to a maximum of 3.78 (CF: 0.94) for the broadspotted molly Poecilia latipunctata (Table 2 I). Table 3 1. Number of species classified by the three risk ranking methods Low Risk Medium Risk High Risk "Non consensus" FISK v2 18 52 25 A Priori Invasiveness* 74 21 Survey 54 14 6 21 are consider ed high non invasive are combined into low and medium risk categories. Table 3 2. Comparison of risk classification between the mean FISK v2 scores and the a priori invasiveness ranking A Priori Invasiveness Inv asive Non invasive To tal FISK v2 Inv asive 16 (76%) 9 (12%) 25 Non invasive 5 (24%) 65 (88%) 70 Total 21 74 95
34 Table 3 3. FISK v2 assessment results for peninsular Florida Score CF Taxon name A Priori Invasive n Mean Min Max SE Delta Outcome Mean Min Max SE Aequidens pulcher NI 1 3 M 0.85 0.85 0.85 Amatitlania nigrofasciata INV 2 10.5 9 12 1.2 3 M 0.86 0.82 0.89 0.03 Amphilophus citrinellus NI 2 6 4 8 1.6 4 M 0.92 0.91 0.93 0.01 Anabas testudineus NI 2 6.8 4.5 9 1.8 4.5 M 0.83 0.82 0.85 0.01 Aphyocharax anisitsi NI 2 1.6 4 L 0.9 0.86 0.94 0.03 Astatotilapia calliptera NI 1 L 0.89 0.89 0.89 Astronotus ocellatus NI 3 10 8 13 1.5 5 M 0.89 0.84 0.94 0.03 Barbonymus schwanenfeldii NI 1 1 M 0.92 0.92 0.92 Belonesox belizanu s NI 2 9.5 9 10 0.4 1 M 0.89 0.89 0.9 0 Betta splendens NI 2 3 1 5 1.6 4 M 0.85 0.84 0.86 0.01 Callichthys callichthys NI 2 9.5 3 16 5.3 13 M 0.85 0.84 0.86 0.01 Carassius auratus INV 4 28 19 35 4.8 16 H 0.88 0.85 0.9 0.01 Channa argus argus NI 2 19 9 29 8.2 20 M 0.84 0.79 0.9 0.05 Channa marulius NI 2 18.5 9 28 7.8 19 M 0.9 0.88 0.91 0.01 Chitala ornata NI 2 4.8 4 5.5 0.6 1.5 M 0.86 0.81 0.91 0.04 Cichla ocellaris INV 3 11.7 8 15 2 7 M 0.92 0.89 0.95 0.02 Cichla temensis NI 2 6 3 9 2.4 6 M 0.91 0.8 8 0.94 0.02 Cichlasoma bimaculatum NI 2 9.5 7 12 2 5 M 0.88 0.88 0.89 0 Cichlasoma salvini NI 1 9 M 0.89 0.89 0.89 Cichlasoma trimaculatum NI 1 1 M 0.91 0.91 0.91 Cichlasoma urophthalmum NI 2 11 10 12 0.8 2 M 0.89 0.82 0.9 6 0.06 Clarias batrachus INV 4 27.3 19 37 4.8 18 H 0.88 0.83 0.91 0.02 Colossoma macropomum NI 3 6.8 11.5 4.4 13.5 M 0.85 0.84 0.86 0.01 Ctenopharyngodon idella NI 4 24.5 15 30 3.8 15 M 0.91 0.89 0.94 0.01 Ctenopoma nigropannosum NI 1 5 M 0.83 0.83 0.83 Cyprinus carpio INV 5 31.8 26 42 4.4 16 H 0.91 0.85 0.99 0.03 Danio rerio NI 3 1 3 1.2 4 M 0.91 0.87 0.96 0.03
35 Table 3 3. Continued Score CF Taxon name A Priori Invasive n Mean Min Max SE Delta Outcome Mean Min Max SE Devario malabaricus NI 2 2 2.4 6 L 0.86 0.78 0.93 0.06 Gambusia affinis INV 3 19.3 8.5 3 2.5 7 24 M 0.89 0.84 0.97 0.04 Gymnocorymbus ternetzi NI 1 L 0.87 0.87 0.87 Helostoma temminkii NI 1 8 M 0.9 0.9 0.9 Hemichromis letourneuxi NI 2 7.5 7 8 0.4 1 M 0.9 0.9 0.9 0 Herichthys cyanoguttatus NI 2 2 2 2 0 0 M 0.89 0.88 0.89 0 Heros severus NI 1 3 M 0.85 0.85 0.85 Hoplias malabaricus NI 2 7.5 3 12 3.7 9 M 0.86 0.82 0.91 0.04 Hoplosternum littorale NI 3 12.7 8 18 2.9 10 M 0.89 0.86 0.91 0.01 Hypophthalmichthys nobilis INV 2 22.5 11 34 9.4 23 M 0.82 0.81 0.84 0.01 Hypostomus sp. 2 NI 2 25 23 27 1.6 4 H 0.85 0.83 0.87 0.02 Labeo chrysophekadion NI 1 2 M 0.91 0.91 0.91 Leporinus fasciatus NI 1 L 0.8 0.8 0.8 Macrognathus siamensis NI 1 L 0.83 0.83 0.83 Macropodus opercularis NI 1 1 M 0.89 0.89 0.89 Metynnis sp. 3 NI 2 2 2.4 6 L 0.79 0.7 0.88 0.07 Misgurnus anguillicaudatus NI 1 1 3 M 0.86 0.86 0.86 Moenkhausia sanctaefilomenae NI 2 0 0 L 0.88 0.79 0.97 0.07 Monopterus albus NI 2 9 7 11 1.6 4 M 0.88 0.87 0.88 0 Morone chrysops x M. saxatilis NI 2 0.8 2 L 0.94 0.92 0.95 0.01 Oreochromi s aureus INV 2 14.5 12 17 2 5 M 0.91 0.91 0.91 0 Oreochromis mossambicus INV 2 12.5 11 14 1.2 3 M 0.85 0.83 0.86 0.01 Oreochromis niloticus INV 2 15 12 18 2.4 6 M 0.89 0.84 0.94 0.04 Osteoglossum bicirrhosum NI 2 2.4 6 L 0.87 0.85 0.9 0.0 2 Oxydoras niger NI 2 1.5 1 2 0.4 1 M 0.86 0.82 0.89 0.03 Pangio kuhlii NI 1 L 0.92 0.92 0.92 Parachromis managuensis INV 2 13 12 14 0.8 2 M 0.88 0.85 0.9 0.02
36 Table 3 3. Continued Score CF Taxon name A Priori Invasive n Mean Min Max SE Delta Outcome Mean Min Max SE Paraneetroplus melanurus x P. zonatus NI 1 L 0.91 0.91 0.91 Piaractus brachypomus NI 2 2.5 9 5.3 13 M 0.87 0.8 0.94 0.06 Platydoras costatus NI 1 3 M 0.88 0.88 0.88 Platytropius siamensis NI 1 L 0.94 0.94 0.94 Poecilia latipinna x P. velifera INV 1 2 M 0.94 0.94 0.94 Poecilia latipunctata NI 1 L 0.94 0.94 0.94 Poecilia petenensis NI 2 0 0.8 2 L 0.89 0.83 0.94 0.05 Poecilia reticulata INV 3 16 12 23 3.5 11 M 0.86 0.82 0.92 0.03 Poecilia sphenops NI 2 11.5 8 15 2.9 7 M 0.86 0.82 0.91 0.04 Polypterus delhezi NI 2 3.5 3 4 0.4 1 M 0.88 0.82 0.93 0.05 Pterodoras granulosus NI 2 5 5 5 0 0 M 0.89 0.81 0.96 0.06 Pterophyllum scalare NI 3 0.6 2 L 0.92 0.89 0.96 0.02 Pterygoplichthys anisitsi NI 2 21.5 20 23 1.2 3 M 0.9 0.85 0.95 0.04 Pterygoplichthys disjunctivus INV 3 21.7 17 24 2.3 7 M 0.89 0.86 0.92 0.02 Pterygoplichthys multiradiatus NI 2 20.5 17 24 2.9 7 M 0.93 0.92 0 .94 0.01 Puntius conchonius NI 2 3.5 2 5 1.2 3 M 0.88 0.84 0.91 0.03 Puntius gelius NI 2 1 1.6 4 L 0.86 0.79 0.92 0.05 Systomus tetrazona NI 3 1.8 0 4 1.2 4 M 0.85 0.82 0.91 0.03 Pygocentrus nattereri NI 2 9.5 5 14 3.7 9 M 0.8 0.72 0.88 0.06 Pylodictis olivaris NI 3 10 2 16 4.2 14 M 0.89 0.85 0.94 0.03 Rhamdia quelen NI 2 3.5 0 7 2.9 7 M 0.84 0.78 0.91 0.05 Rocio octofasciata NI 3 4.3 9 2.9 10 M 0.87 0.83 0.91 0.02 Sander vitreus NI 2 7 4.5 9.5 2 5 M 0.89 0.84 0.94 0.04 Sarotherodon melanotheron INV 2 6.5 15 6.9 17 M 0.88 0.84 0.91 0.03 Serrasalmus rhombeus NI 2 9.5 7 12 2 5 M 0.85 0.78 0.92 0.06
37 Table 3 3. Continued Score CF Taxon name A Priori Invasive n Mean Min Max SE Delta Outcome Mean Min Max SE Thorichthys meeki NI 2 8.5 6 11 2 5 M 0.86 0.83 0.89 0.03 Tilapia buttikoferi NI 2 4 1 7 2.4 6 M 0.88 0.86 0.9 0.02 Tilapia mariae NI 2 10 6 14 3.3 8 M 0.92 0.86 0.97 0.04 Tilapia zillii INV 3 22.7 11 35 6.9 24 M 0.89 0.86 0.91 0.02 Trichog aster fasciata NI 2 3.5 2 5 1.2 3 M 0.82 0.76 0.89 0.06 Trichogaster labiosa NI 2 1 3 1.6 4 M 0.84 0.81 0.87 0.02 Trichogaster lalius NI 2 2 5 2.4 6 M 0.84 0.81 0.87 0.02 Trichogaster leerii NI 2 2.8 2 3.5 0.6 1.5 M 0.9 0.84 0.95 0.05 Trichopodus trichopterus NI 2 5 5 5 0 0 M 0.87 0.84 0.89 0.02 Trichopsis vittata NI 2 0.4 1 L 0.88 0.85 0.91 0.02 Xiphophorus hellerii INV 2 7.5 7 8 0.4 1 M 0.9 0.88 0.93 0.02 Xiphophorus hellerii x X. maculatus INV 3 1.7 7 3.9 13 M 0.86 0.83 0.89 0.02 Xiphophorus maculatus NI 2 6.5 4 9 2 5 M 0.9 0.85 0.95 0.04 Xiphophoru s variatus INV 2 8.5 8 9 0.4 1 M 0.9 0.88 0.92 0.02 Notes: For each species, selection criterion, a priori invasiveness (INV = invasive and NI = non invasive) (as per http://issg.org and www.fishbase.org ) and, number of assessors (n), along with summary statistics (SE=standard error) for corresponding FISK score, (risk) outcome and certainty factor (CF) are reported. Outcome is based on a calibration threshold of 10.25 between medium r isk and high risk fishes and classified as: low (L) = [ 15, 1[; medium (M) = [1, 10.25[; high (H) = [10.25, 57]. See text for computation of the certainty factor.
38 Figure 3 1. Mean FISK v2 scores for the 95 freshwater fish species assessed using FISK v2 for peninsular Florida with error bars indicating minimum and maxi mum range in scores for multi assessed species Thresholds are: <1 axis labels are abbreviations for species names as appears in Table 2 1
39 Figure 3 2 Frequency distributions of mean FISK v2 scores, plotted separately for a priori non invasive (dashed line) and invasive (solid line) fishes. The vertical solid line represent s the high risk threshold value of 10.25 and the vertical dashed line represents the low risk threshold value of 0.
40 Figure 3 3. Percent survey respons es for the 54 fishes that had a consensus (>51%) of responses in the low risk category. Mean FISK scores are shown by the black bars and are labeled on the secondary Y axis.
41 Figure 3 4 Percent survey responses for the 14 fish es that had a consensus ( 50%) of survey responses in the medium risk category. Mean FISK scores are shown by black bar and labe led on the secondary Y axis.
42 Figure 3 5 Percent survey responses for the six fishes that had a co nsensus ( 50%) of survey responses in the high risk category. Mean FISK scores are shown by black bar and label ed on the secondary Y axis.
43 Figure 3 6 Percent survey responses for the 21 n any of the three risk categories. Mean FISK scores are shown by black bar and label ed on the secondary Y axis.
44 Figure 3 7. Top: Receiver operating characteristic (ROC) curves for the three main assessors ( initials ) on 95 fish taxa assessed with FISK v 2 for Peninsula r Florida Bottom: Mean ROC curve based on mean scores from all five assessors, with smoothing line and confidence interv als of specificities (see Table 3 3 ).
45 CHAPTER 4 CONCLUSION T his study determined that FISK v2 is an effective tool to i nform management of the potential invasiveness risk of non native freshwater fishes for p enins ular Florida with a calibrated high risk threshold value of 10.25. A t least 75% of the fishes that were id entified as high risk by the newly calibrated FISK v2 ar e either establish ed in peninsular Florida or have elevated regulatory status and could potentially cause adverse impacts. T he majority of fishes that scored as either low risk (score < 1) or the low end of medium risk (i 3.5) by FISK v2 we re also identified by the survey (Figure 3 3 ) and the a priori risk identification method to be of minimal risk of becoming invasive in peninsular Florida ( Table 3 3 ) This study adds to the growing number of regions where FISK v2 has been tested but is un ique in evaluating the largest number of species in a FISK application to date and in applying FISK to species of primarily tropical origin. The a priori ranking used to evaluate a diagnostic tool will impact the result of the calibration (i.e., the thre shold values) and the reported accuracy of the tool. In this case, the a priori invasiveness ranking was determined from FishBase and ISSG, global databases which lack the regional focus often needed for risk assessment (Davis et al. 2004; Ruesink 2005). F or example, four of the ten highest scoring species were not reported as invasive by these international databases, even though these fishes are established in peninsular Florida and were identified by the survey to be of elevated ri sk, i .e., medium or hig h risk, (Figure 3 4 and Figure 3 5 ) The characteristics used to assign categories by these databases are not explicitly described and are qualitative in nature. Both databases seem to rely heavily on successful invasion history but do not
46 take into accoun t ecological characteristics of the species and the recipient environment. Invasion history is a good predictor of invasiveness elsewhere (Kolar and Lodge 2002; Deahler et al. 2004; Marchetti et al. 2004) but many established and impactful species lack a h istory of introduction outside of Florida (e.g., pike killifish and African jewelfish). Furthermore, FISK is semi quantitative incorporating a much broader range of specific biological characteristics of the species and ecological characteristics of the ri sk assessment area than does the qualitative nature of the databases that rely the use of global databases to calibrate regional risk screening and assessment methods should be done cautiously and with clear explanation (Hill 2009). The purpose of the expert risk survey was to provide a regionally focused estimate of risk specific to these species in peninsular Florida for comparison with the a priori ratings and the ev entual FISK v2 assessment outcomes. Unfortunately, there was considerable variability in responses by the survey participants. Only 23 (24.2%) of the 95 fishes had a 100% consensus and all of these fishes were considered low risk by the survey respondents. Therefore, incorporating the survey into the statistical analysis for the verification of the risk thresholds would have introduced another element of subjectivity by requiring a qualitative assignment of many of the fishes into a specific risk category. Some of this variability could have been overcome had there been a greater return rate or more experts surveyed; however, this variability may also be indicative of the current status of non native species science. T here is continual debate and often contr asting views within the scientific community as to the impacts that non native fishes have caused (Gozlan et al. 2008; Vitule et al. 2009). Morever, there has
47 been substantial effort detailing the use and proper definition of standard terminology in invasi on ecology (Copp et al. 2005c; Gozlan et al. 2010). In some cases, this survey included, value laden terminology will lead to some amount of qualitative decision making when using expert opinion (Maguire 2004). For example, in the survey we defined high ri likely to cause moderate to severe detectable the question the survey respo ndent is then required to make qualitative judgment s as to what is likely and what constitutes moderate to severe detectable impacts. Similarly Pheloung et al. (1999) found significant differences in the risk sensitivity scores among the expert groups used to evaluate the New Zealand WR A. R esults of the FISK asse ssments suggest that penin sular Florida is a region vulnerable t o the establishment of non nat ive fishes, given that 25 fish es were ranked as high risk and 52 fishes as medium risk. This outcome is supported by the establishment success and elevated regulatory status of the non native fishes in Florida that scored as high risk by FISK Of the se fis hes 14 are reported by FWC to be reproducing or regionally established ; 2 species that are not established northern snakehead Channa argus and redbelly tilapia Tilapia zillii are prohibited ; grass carp is a conditional species with legal possession by t he public only of functionally sterile triploid s ; and convict cichlid was locally established but has since been extirpated P terygoplichthys anisitsi is probably est abli shed in peninsular Florida but is difficult to distinguish from other established cong eners ( USGS 2014; Hill JE, personal communication). The remaining six high risk fishes (i.e., goldfish, broadspotted molly, bighead carp, w estern mosquitofish, guppy, and fathead minnow) do not pose a substantial risk
48 of invasiveness to peninsular Florida as is indicated by their non invasive status from the survey and their failure to successfully establish. H owever, th ey do possess many of the characteristics that cause FISK scores to increase such as extensive invasion histories, documented impacts and overall high climate matches. R ecent applications of FISK v2 (Almeida et al. 2013; Puntila et al. 2013; Sim o Vil izzi and Copp 2013) had some of these same species receiving the highest FISK s cores The goldfish, for example, has been introduced and established worldwide and has been identified as high risk by FISK applications for the United Ki ngdom (Britton et al. 2010), the Murray Darling Basin in Austr alia (Vilizzi and Copp 2013), the Iberian Peninsula (Almeida et al. 2013) and Anatolia and Thrace in Turkey (Tarkan et al. 2014) Clear explanations for these re prov ided for those specific regions but, generally, they have not proven to be invasive in peninsular Florida A potential area of future research that could improve risk assessment is why these species consistently fail to establish in peninsular Florida even after repeated introductions. These cases clearly demonstrate some limitations of predictive models like FISK for evaluating invasiveness potential. L imiting t he number of medium risk outcomes resulting from a risk assessment is desirable from a managemen t perspective (Daehler et al. 2004) Past applications of the WRA focused heavily on the reduction of medium risk species in their evaluation attempts ( Pheloung et al. 1999; Daehler et al. 2004). Daehler et al. (2004) used a decision tree that consisted of In the present case, nearly 55% of the assessed fishes were categorized as medium risk by FISK v2. E ight
49 of these fishes were scored within one point of the high risk threshold Only two of these eight fishes (cascurado Callichthys callichthys and threadfin shad Dorosoma petenense ) were ranked as low risk by a majority of participants from the survey. The spotted tilapia Oscar Astronot us ocellatus pike killifish Belonesox belizanus black acara Cichlasoma bimaculatum yellow belly cichlid Cichlasoma salvini and Asian swamp eel Monopterus albus are all established in peninsular Florida and substantial concern surrounds the potential fu ture establishment of the flathead catfish Pylodictis olivaris that is already established in north Florida While it could be argued for the placement of some of these species into the high risk category like the flathead catfish (Thomas 1993; Pine et al. 2005) and pike killifish ( Greenwood 2012) that have documented negative impacts, others have established with little evidence of negative impacts and should likely remain in the medium risk category Distinguishing which species will turn out to be invasi ve is a notoriously difficult task and brief risk screens are not often capable of distinguishing the difference 100% of the time (Stohlgren and Schnase 2006; Copp et al. 2009). It is, therefore, recommended that FISK and risk screenin g tools of this type be used not as decision making tools but to help inform the development of management strategies and policy decisions related to non native species (Copp et al. 2009) It is inevitable that a risk manager will make the final decision as to what actions are appropriate for the specific species in question and FISK provides a valuable resource for informing this decision. I ncreasing the low risk threshold could limit the number of fishes that score as medium risk without sacrificing the accuracy of identify ing high risk invaders (i.e., sensitivity) The current threshold for low risk fishes is set at a score of < 1 but that
50 threshold has not been thoroughly investigated in recent calibration attempts for the FISK tool. This threshold is a remnant of the tran sformation from the WRA to FISK and was found to be adequate to avoid any invasive species from being scored as low risk (Pheloung et al. 1999; Copp et al. 2009). Eighteen fishes scored between 1.0 and 3.5 and of these fishes only two red swordtail hybrid X iphophorus hellerii x X. maculatus and sailfin molly Poecilia latipinna x P. velifera a priori methodology and none were considered high risk by the survey. Although not statistically tested here, the low risk threshold could hypothetically be increased and actually increase the reported accuracy, in terms of specificity, of the FISK tool without a loss of sensitivity From a practical standpoint, there is little difference between a fish that scores a 10 (medium risk) and one that scores an 11 (high risk), but from a management standpoint a medium risk fish has substantially different implications than does a high risk fish. FISK s cores fall on a continuum and are relative to one another ranging from lowest to highest The outcome score is used to pl ace a fish into a risk category but, more importantly, tells the assessor where along this risk co ntinuum the species is located (Daehler et al. 2004) While it is easy to only focus on the outcome scores and risk ratings, of greater importance to risk managers is how a species got to the score it attained and the level of certainty the responses were attributed The level of certainty is an important aspect to be considered by when making policy and management decisions. Asse ssments with high uncertainty (low confidence) require greater scrutiny and, in many cases, further investigation than do those with low uncertainty (high confidence). The transparency of the specific question and certainty responses and
51 justifications req uired of the assessor provide a more detailed picture of the result and quality of the risk screen than does the outcome score or risk rating (NRC 2002; Lodge et al. 2006) A manager can take all of this information from the FISK assessment in the context of the current risk management goals and risk mitigation strategies to determine appropriate recommendations In the case of the Barcoo grunter (Lawson et al. 2013), the FISK score of five r esulted in a medium risk rating but it was recommended that no fur ther risk assessment was necessary because the current risk management and regulatory actions were suitable. The mean t hreshold value distinguishing between medium risk and high risk species determined from this study is considerably lower than those foun d for many other world regions (thresh old scores ranging from 17 to 23 ) where FISK has been applied (Copp et al. 2009; Onikura et al. 2012; Troca et al. 2012; Almeida et al. 2013; Puntila et al. 2013; Tarkan et al. 2014; Vilizzi and Copp 2013) T he one exc eption is the 9.5 threshold identified for the Balkans region application of FISK (Simonovic et al. 2013). They exp lained this finding to be due to many of the fishes that were assessed were the result of translocations within countries of this region and were not considered invasive on a wide geographic scale (hence lower FISK scores), but were co nsidered invasive in the risk assessment area To some extent this result also occurred in peninsular Florida, where many of the fishes that have successfully established are not found to be widely introduced or invasive on a broader geographical scale Additionally, Vilizzi a nd Copp (2013) found that question modification and the more stringent positive response requirements resulting from the transformation of FISK v1 to FISK v 2 (Lawson et al. 2013) contributed to the overall depression of FISK scores.
52 A ssessor scoring variab ility c ould also potentially dep ress FISK scores as was briefly alluded to by Vilizzi and Copp (2013) Of the 20 species that were assessed both in this study and in the Iberian Peninsula evaluation of FISK v2 (Almeida et al. 2013), mean FISK scores for al l 20 species were lower, as much as 16 points for some species, in the pe ninsular Florida evaluation. With the high number of fishes identified by this study as high or medium risk, it is fortunate that Florida has a pro active philosophy towards non nativ e species management. Current non native species regulations in Florida are designed to protect environmental resources while still considering those businesses that are centered on the trade of non native species (FDACS 2007). All certified commercial aqu aculture facilities in Florida are required to comply with the FDACS Best Management Practices (BMPs) that has provisions designed to minimize the potential for escape of non native fishes from aquaculture farms. Non native fishes seen as particularly high risk are subjected to additional regulations imposed by the FWC (also enforced by FDACS) as part of their conditional and prohibited species lists (www.myfwc.com /wildlifehabitats/nonnatives/conditional prohibited species/ ) The FISK v2 toolkit provides, w ith limited investment, an additional level of risk screening available to non native species managers and can be especially beneficial in providing information to decision makers (Hill et al. In Press ) T h is study represents the first comprehensive evalu ation of FISK v2 in the Unit ed States and the largest evaluation, in terms of number of species, worldwide FISK v2 has now been thoroughly tested on primarily tropical freshwater fish species in the sub tropical/warm temperate climate of peninsular Florid a where it was originally developed
53 (Lawson et al. 2013) and has proven to be successful. A strength of this approach was that all the recorded introduced non native fishes in peninsular Florida were used for this calibration which resulted in a diverse a nd realistic set of species some invasive and some not, being assessed. The calibrated high risk threshold of 10.25 was considerably lower than most other published applications of FISK This result stress es the importance of setting thresholds based on t he region in question and the intended management goals (Puntila et al. 2013). V ariation in risk categorizations of some fishes among the three risk identification approaches (i.e., FISK scores, the survey, and international databases) were likely the resu lt of methodological differences and should be strongly considered and well understood prior to undertaking risk screening or cali bration attempts of this nature. The regionally calibrated high risk threshold, consistency of questions, assessor provided re sponse justifications, and uncertainty ratings incorporated within a FISK v2 assessment provide a detailed and transparent means of estimating the potential invasiveness risk of a fish species These data would be valuable to natural resource managers tryi ng to make informed decisions related to the risks of non native species. Additionally, FISK v2 provides an expedient means of eliminating low risk species from the managers concerns allowing them to focus resources towards species who pose a more substant ial risk. This successful evaluation of FISK v2, specifically designed for sub tropical regions, has substantially expanded the availability of FISK to completely new regions of the world and to a more diverse set of fishes for which it has previously been unavailable
54 APPENDIX A REVISIONS OF THE FISH INVASIVENESS SCORING KIT (FISK) FOR ITS APPLICATION IN WARMER CLIMATIC ZONES, WITH PARTICULAR REFERENCE TO PENINSULAR FLORIDA Risk identification tools are the front line of the risk analysis process, ass isting environmental managers in identifying which non native species are more likely to be invasive and are therefore amenable to detailed risk analysis. Building on the Weed Risk Assessment (WRA) of Pheloung et al (1999) the Fish Invasiveness Scoring Ki t (FISK) was adapte d for use in the United Kingdom (Copp et al. 2005a, 2005b, 2009). As with the WRA, FISK consists of 49 questions within two main sections ( Biogeography / History and Biology / Ecology ), and eight categories ( Domestication / Cultivation ; Climat e and Distribution ; Invasive elsewhere ; Undesirable traits ; Feeding guild ; Reproduction ; Dispersal mechanisms ; Persistence attributes ). The and these have been calibrated (Copp et al. 2009) into three levels of potentia l risk of a species being invasive: low (score <1), medium (1 2010) were developed for the United Kingdom, a temperate zone country, and subsequent applications have been in other temperate z one countries including Belarus (Mastitisky et al. 2010), Belgium (Vandenbergh 2007; Verbrugge et al. 2012), and Japan (Onikura et al. 2012) with additional applications completed or in progress for Reprinted with permission from Laws on, L.L., J.E. Hill, L. Vilizzi, S. Hardin, and G.H. Copp. 2013. Revisions of the Fish Invasiveness Screening Kit (FISK) for its application in warmer climatic zones, with particular reference to peninsular Florida. Risk Analysis 33(8):1414 1431.
55 Australia (Vilizzi and Copp 2013), Brazil (Troca 2012), Iberi a (Almeida et al. 2013), Finland and the Balkan states of Serbia, Macedonia, Montenegro and Bulgaria (P. Simonovi ). The specificity of FISK to the temperate zone was not appreciated until at tempts were made to apply this screening tool in peninsular Florida (LLL, JEH, SH and GHC, unpublished data) and Mexico (R. Mendoza, V.S. Aguirre and L.G. Bermen, unpublished data). To address this shortcoming, the aim of the present study was to improve FISK in terms of its climatic applicability, in particular to regions with tropical/sub tropical environments, and of its user interface The specific objectives of the present study were to: 1) complete a review and revision of FISK questions and related guidance; 2) undertake improvements to the FISK software graphical user interface (GUI) and therefore produce FISK v2; and 3) carry out a trial application of FISK v2 on a non native fish species of relevance to peninsular Florida. T he resulting new versi on of FISK (v2) is applicable over a wider range of environmental conditions, is more user friendly and, ultimately, replaces its predecessor for use as an invasiveness screening tool. Materials and Methods Question and Guidance Refinement Questions and g uidance in FISK v1 were critically reviewed with regard to peninsular Florida south of the Suwanee River, which has a sub tropical climate. Particular attention was directed to characteristics that are considered to facilitate the invasions of tropical/sub tropical environments by freshwater fishes, with the following criteria used in the revision process: 1) improved clarity, where changes reduced ambiguous terminology or uncertainty in interpretation of questions; 2) increased
56 climatic suitability, where changes allowed for increased flexibility of climate match source information or direct incorporation of physiological tolerances; and 3) enhanced ecological applicability, where modifications addressed a wider range of ecological characteristics. FISK question amendments were discussed and executed during a two week meeting in August 2011, with four of the co authors at the University of Florida (LLL, JEH, SH, GHC) and one (LV) via electronic correspondence. During this review process, each of the four co authors explained their interpretation of the questions, the guidance and their likely response under different hypothetical scenarios, with insights from GHC into the intent of each question derived fr om the original adaption of FISK (Copp et al. 2005a ) from the WRA (Pheloung et al. 1999). Once consensus was reached, the question and guidance were revised accordingly, and each of the previously listed criteria was discussed to determine whether or not additional changes were necessary. As a control meas ure, two project collaborators assessed the same two species, discussed the outcomes and made final revisions to the questions and guidance. Software Improvements FISK consists of an electronic toolkit wri tten in Excel for VBA code (Walkenbach 2007). The code was originally developed by Pheloung et al (1999) for their WRA toolkit and later adapted for use with FISK (including its Spanish version S FISK) to accommodate all necessary changes for application to freshwater fish, and eventually to freshwater i nvertebrates (FI ISK), marine invertebrates (MI ISK), amphibians (AM PHISK) and marine fish (M FISK) (Copp et al. 2009) The software
57 architecture of the FISK family of programs is th This represents an advanced level of Excel application development in which virtually all features are controlled by the application itself. In the case of FISK, the user interface consists of tightly controlled data entry user forms that separate the user interface from the underlying data storag e layer, resulting in a fully functional application that involves a high degree of control over user interaction. While FISK v1 already contained all essential features for species listing, questionnaire based user input and report generation, an upgrade to FISK v2 was deemed essential to improve overall computational speed as well as to provide for a higher level of flexibility and user friendliness, with special emph of the two main user forms (or dial ogs), namely: the Species Assessment Menu and the Q&A dialog. Programming was done in Microsoft Visual Basic for Applications 7.0 for Excel 2010 for Windows by LV under on going advice and feedback from the rest of the project team, who also participated in the beta testing phase along with an external expert (E. Tricarico, University of Florence, Italy). Application An example species was screened for peninsular Florida by SH using FISK v2. The Barcoo grunter Scortum barcoo (McCulloch & Waite, 1917), fam ily Terapontidae, is a freshwater species endemic to the Lake Eyre basin in interior Australia and has been increasingly use d in aquaculture as a food fish (QDPI F 2012) This species has been marketed as the jade perch for aquaponics (i.e. integrated fish a nd plant crop culture), and was chosen for assessment in this study because of the recent emergence of small scale commercial and non commercial aquaponics in Florida. Currently, there are
58 no regulations governing the possession of Barcoo grunter in Florid a, so FISK was used to determine whether a full risk assessment is warranted. The FISK assessment was reviewed by JEH to simulate a typical management agency scenario, where an assessment is produced and then reviewed by either an internal or external taxo nomic expert. Thus, screening the Barcoo grunter serves to illustrate the process of assessing a fish species using FISK v2, provides real world answers to FISK questions, and demonstrates its potential use as a practical tool for pro active management in warmer climate zones. Relatively little is published on the Barcoo grunter in its native range; however, there is a growing body of literature on its husbandry. With a maximum size of about 35 cm standard length, the Barcoo grunter in its native range is f ound in low gradient rivers subject to extreme flooding and drying, (Allen et al. 2002) extending over a tem perature range of about 10 30C (Arthington et al. 2010). An omnivorous forager, the Barcoo grunter feeds on a wide variety of plant material, detri tus, invertebrates, and, occasionally, small fish (Balcombe et al. 2005; Reid et al. 2010). Spawning o ccurs during high water periods (Kerezsy et al. 2011). Findings Question Refinement Alterations were made to 36 of the 49 FISK questions and/or guidance n otes in accordance with the revision criteria (Table A 1 ): 27 revisions to improve clarity, 3 to increase climatic suitability, and 14 to enhance ecological applicability. Some questions required multiple changes and were placed into more than one revision criterion.
59 Domestication/ c ultivation (section A1, table A 1) Domestication is considered to have the potential of enhancing the fitness of some freshwater fishes (i.e. increased growth rate, mating success, and/ or fecundity) over wild strains (Muir 1999), and commercial production of ornamental species may also increase the intensity of introduc tions (i.e. propagule pressure) (Copp et al. 2007). However, commercially produced ornamental fish are often small bodied (e.g. variable platyfish Xiphophorus varia tus and swordtail X. hellerii ) and selected for bright coloration (e.g. transgenic zebra danio Danio rerio ) or increased fin length (e.g. long finned zebra danio varieties). This has been shown to increase their vulnerability to predation and decrease thei r likelihood of establish ment in non native environments (Hill et al. 2011; Thompson et al. 2012). Therefore, guidance for question (henceforth produced species that have been s elected for traits that are likely to reduce their fitness. broad term that encompa sses several phases of the invasion process including establishment, colonization, and a meas ure of persistence through time (Copp et al. 2005c), and also implies some level of human acceptance of the non native fish into the native fish community. The ter et al (2008) i.e. non native fishes that have been consistently collected from public waterways, unlikely to be eliminated, and that have persisted for an extended period of time relative to their life span.
60 Climate and d istribution (section A2, table A 1) Climate matching models use computer analysis to match temperature and v1 recommended the use of climate mat ching software to respond to the climate and distribution section with the highest level of certainty (Copp et al. 2009). The certainty response modifies the score given for climate questions and weights other questions related to impact and invasion histo ry (Copp et al. 2009) Climate models use data from terrestrial locations and may not be accurate for predicting water temperatures. Moreover, spatial gaps and differences in elevation between weather station locations and species occurrences may limit usef ulness. The lower lethal temperature is often the most important environmental factor limiting the distribution and spread of non native tropical fish outside their native range, and this has been determined for several non nati ve freshwater fishes in Flor ida (Shafland and Pestrak 1982; Gestring et al. 2010). We modified Qs 2.01 2.04 to allow the FISK assessor to respond to the climate matching section with high certainty without requiring climate matching software (see also Onikura et al (2010 ) and opted to allow for the use of expert knowledge (e.g. laboratory determined lower lethal temperature) or matching climate categories of the Kppen Geiger climate maps (Peel et al. 2007). Invasive e lsewhere (section A3, table A 1) A detailed and accurate account may be unavailable in popular databases as a result of poor data quality, e.g. limited or outdated collection data, species misidentifi cations, or conflicting reports (Fuller et al. 1999; Froese and Pauly 201 2). Therefore, Qs 2.05 3.05 relating to invasive history were changed to allow assessor judgment in determining what constitutes reliable data
61 detailing a non d to describe invasive plants (Richardson et al. 2000) and definition, an invasive aquatic species must be known to cause or l ikely to cause negative impacts (Invasive Species Advisory Committee 2006) Undesirable (or persistence ) traits (section B4, table A 1 ) Minor modifications were made to the terminology used in Qs 4.01, 4.02, 4.06, act as a novel predator in a new environment even where other predators are already present. For example, flathead catfish Pylodictus olivaris is a large bodied piscivorous catfish. The introduction of flathead catfish into Georgia has been implicated in the dec line of native bullhead Ameiurus spp. and redbreast sunfish Lepomis auritus populations, even though those species were not predator nave (Thomas 1993). For Q 4.07, the large body size categorization was increased from 10 to 15 cm total length (TL). Altho ugh fish whose ultimate body length exceeds 10 cm TL have an increased likelihood of being released, the majority of non native fish established in Florida typically reac h an ultimate length > 15 cm TL (Nico and Fuller 1999; Nico 2005). Feeding guild (sect ion B5, table A 1 ) In FISK v1, species accrued an increase in score if categorized as a piscivore, is associated with impacts within the risk assessment area. Therefo re, a modifier was added to each of the feeding guild questions in FISK v1 (Qs 5.01 5.04), so that it is now
62 guild questions) because piscivorous fishes ar e often associated with impacts (Copp et al. 2005a; Moyle and Light 1996). However, there was no similar consi deration given for herbivorous fishes that primarily feed on aquatic macrophytes. In many tropical/sub tropical environments, aquatic macrophytes are essential habitat features and herbivorous fish have the potential to alter this habitat with profound con sequences for native species and ecosystem services. For example, the herbivorous grass carp Ctenopharygodon idella can cause dramatic changes to aquatic ecosystems (Dibble and Kovalenko 2009). Because of the magnitude of these potential impacts, we includ ed herbivorous fishes in Q 5.01. In peninsular Florida, fishes that are primarily algivorous or detritivorous (e.g. blue tilapia Oreochromis aureus and Orinoco sailfin catfish Pterygoplichthys multiradiatus ) have been successful in establishing populations ; therefore, we included these categories in Q 5.03 to allow for the possibility of impacts within the risk assessment area for these types of fishes. Reproduction, dispersal and tolerance attributes (sections B6, 7 and 8, table A 1 ) Q 6.01 was changed to include specific examples of parental care that are exhibited by fishes that have successfully established in Florida. The guidance for Q e. related dispersal of non native fishes. Cultural release is the suspected vector of introduction for multiple non native species(Severinghaus and Chi 1999; Liu et al. 2012), includ ing Asian swamp eels Monopterus albus in Florida (Nico et al. 2011).
63 fishes) is a limiting factor in their ability to establish and spread in new environments (Shafland and Pestrak 1982). Q 8.03 was modified because virtually every species is susceptible to piscicides at high concentrations. However, some species are resistant to piscicides at l egally permitted use levels or capable of limiting exposure due to regulated differently in different regions and countries so no specific chemicals were listed. Softwar e Improvements Species assessment menu The new, improved interface of the Species Assessment Menu dialog was created (Figure A 1), with new features that include: 1. Five additional columns, so that for each species in the list the following information is re ported: i) Answered : number of answered questions from the Q&A dialog out of the total set of 49 questions making up the FISK tool; ii) Score : final, in case of a complete assessment, or partial otherwise; iii) U.K. (1; 19) : risk level outcome (i.e. e to U.K. calibration thresholds (Copp et al. 2009) (only upon completion of the assessment, that is all 49 questions answered); iv) Japan (1; 19.8) : same but relative to Japan calibration thresholds (Onikura et al. 2010) and v) U D ( -; -) : same but relative to user defined (UD) calibration thresholds. 2. Possibility to Edit details of any species in the list, including its scientific and 3. Possibility to Delete any species in the list. 4. Indic ation of the Total number of species assessed and/or under assessment along with the number of species Selected for (multiple) report generation. 5. Possibility to Sort (in ascending order) the species in the list by any of the eight columns therein (hence, e xcluding the unique species ID identifier).
64 6. Possibility to set UD thresholds risk species (e.g. as derived from a new calibration study). 7. Ability to generate a Report for one or more species selected from t he list. If the assessment has not been completed for any of the selected species, then the user is warned whether a report for that species is still required. Q&A The Open Q&A button in the Species Assessment Menu dialog opens up the Q&A dialog (Figure A 2). New features in this dialog include: 1. A title bar indicating: i) the question being answered, ii) the species name, iii) the category for the question. 2. A Go to Question combo box control to move directly to any question in the assessment. 3. A Clear button to erase all question information in the editable fields (i.e. Response, Certainty, and Justification); 4. A Close no Save sment. 5. Context sensitive coloring, with the editable fields (as per above) marked in light green for answered questions, light red for unanswered questions, and light yellow for questions being answered. Application The FISK score for Barcoo grunter placed the species into the lower range of medium risk of bei ng invasive (score = 5; Table A 2 ). Factors increasing the score included use in aquaculture, a climate match between its native range and peninsular Florida, broad climate suitability and tolerance o f temperatures between 10 30C, parental care and a likelihood for introduction and dispersal by humans. Barcoo grunter lacks a history of invasiveness and has few undesirable attribu tes, characteristics that species for aquaponics, the medium risk score of Barcoo grunter suggests the need for further examination of the risk factors in Florida to determine whether mitigation (e.g.
65 restrictions on location, system design, development of best practices, as well as education and outreach) is required. Barcoo grunter was assigned broad climate suitability b ased o n an Australian reference (QDPIF 2012), where its distribution overlapped three Kppen Geiger zones. Barcoo grunter probably is endemic to the Lake Eyre drainage, a large region in central Australia. Although commonl y listed from northern regions draining into the G ulf of Carpentaria (e.g. Unmack 2001 ), these records likely are misidentifications of the gulf grunter S. ogilbyi (Burrows and Perna 2006; Burrows 2008 ). The taxonomy of terapontids is difficult and confuse d. FishBase lists S. ogilbyi as a junior synonym of S. hillii but The Catalog of Fishes lists S. ogilbyi as valid (CAS 2012). Davis et al. (2011) recently used the name S. ogilbyi in ecological studies of terapontid fishes in northern Australia, and among the co authors are noted experts on Australian fishes. Notably, Barcoo grunter is not included in the latter study even though the study area included the disputed range in Australia. Using the more conservative range distribution, Barcoo grunter would be found in only one Kppen Geiger zone and its FISK score would be reduced. Furthermore, reproductive tolerance of Barcoo grunter was considered suited the Kppen Geiger z one occupied by Barcoo grunter in Australia (BWh = arid, desert, hot) does not occur in Florida. Parental care of eggs increased the FISK score, but the type of eggs laid by the Barcoo grunter is disputed in the literature. FishBase references Breder and R (1966) review of reproductive modes in fishes and states that members of the genus
66 Scortum have male parental care of eggs laid in a nest on the substratum. However, the FishBase account for the small headed grunter S. parviceps contradicts itself b y listing pelagic eggs and male parental care of the eggs in the same paragraph on Biology also referencing Breder and Rosen (1966) in the reproduction section (Foese and Pauly 2012). This casts doubt on the validity of the FishBase information on Scortum reproduction. Recent aquaculture references in the Chinese literature describe the eggs of Barcoo grunter as free floating (Chen et al. 2007). Other terapontids such as the silver perch Bidyanus bidyanus clearly lay semi buo yant, pelagic eggs (e.g. Rowlan d 1984 ). Therefore, it is most likely that Barcoo grunter eggs also float or are semi buoyant, and that dispersal of this species occurs via this mechanism. An affirmative response to the FISK question on parental care of eggs was offset by a negative answ er to a subsequent question regarding egg dispersal. Accordingly, subsequent re evaluation of these questions and responses did not have an appreciable impact on the total FISK score. The Barcoo grunter has no clear history of invasiveness, an important tr ait in determining risk categorization in FISK. With the exception of limited translocations within Australia (Koehn and MacKenzie 2004), it is unclear whether or not the Barcoo grunter has been introduced into open waters outside of its native range. This species is listed as introduced into China, but no data are available on the year of in troduction or population status (FAO 2012), casting doubt that the introduction was into open waters rather than simply an aquaculture transfer. Two questions dealing w ith the likelihood of introduction and human mediated
67 body size >15 cm are considered more likely to be abandoned by aquarium hobbyists (Q4.07, Section B4, Table A 1 ). is unlikely that the Barcoo grunter cultured for food would be released. Similarly, it is doubtful that aquaponic s operators would be motivated to introduce this species into public waters as a food source. However, there is a possibility of abandoning home aquaponics systems for a variety of reasons, which might lead to releasing fish into public waters. For this re likelihood of intentional human dispersal.
68 Table A 1. List of the 49 questions making up FISK v2 with highlighted (in italics) changes relative to FISK v1 (cf. Copp et al. 2005a). Section/ (Code) # Criteria Query Guidance A. Biogeography/historical 1. Domestication/Cultivation 1.01 (C) 1 CL, E Is the species highly domesticated or widely cultivated for commercial, angling or ornamental purposes? In order to respond 'Yes', t he taxon must have been grown deliberately for at least 20 generations, or is known to be easily reared in captivity (e.g. fish farms, aquaria or garden ponds) Whereas, if the taxon has been subjected to substantial human selection that has led to reduced fitness and/o r adaptability, then the response should be 'No' despite the species being widely domesticated/cultivated. 1.02 (C) 2 CL, E Has the species established self sustaining populations where introduced? The taxon must be known to have successfully establish ed self sustaining populations in at least one location outside its native range for an extended period of time this 'extended period' is likely to be shorter for short lived species and longer for longer lived species. 1.03 (C) 3 E Does the species ha ve invasive races/varieties/sub species? This question emphasizes the invasiveness of domesticated species 2. Climate and distribution 2.01 (C) 4 CL, CM What is the level of matching between the of the RA area ? The intention of this question is to assess the likelihood of a taxon establishing self sustaining populations in the risk assessment area. If readily available, then a climate matching approach (e.g. Climex, GARP, Climatch) may be used (see summ ary in Venette et al. 2010; BioScience 60: 349 362) If a climate matching model is not available, then make a 'best estimate' through consultation of the Kppen Geiger climate region system (see: www.hydrol earth syst sci discuss.net/4/439/2007/hessd 4 43 9 2007.pdf) and/or local expertise. 2.02 (C) 5 CL, CM What is the quality of the climate match data? Quality' refers to the assessor's evaluation of the information used to determine the climate match. If there are doubts about the quality of the information available, then attribute the minimum score (i.e. low).
69 Table A 1. Continued Section/ (Code) # Criteria Query Guidance 2.03 (C) 6 CL, CM Does the species have self sustaining populations in three or more (Kppen Geiger) climate zones ? O utput from climate matching can help answer this, combined with the known versatility of the taxon as regards climate region distribution. Otherwise the response should be based on natural occurrence in 3 or more distinct climate categories, as defined by Kppen Geiger (see: www.hydrol earth syst sci discuss.net/4/439/2007/hessd 4 439 2007.pdf), or based on knowledge of existing presence in areas of similar climate. 2.04 (C) 7 CL Is the species native to, or established self sustaining populations in r egions with equable climates to the RA area? This issue raised by this question is whether or not the species actually is established in (or originates from) an area where the climate is similar to the risk assessment area. 2.05 (C) 8 CL Does the speci es have a history of being introduced outside its natural range? Should be relatively well documented, with evidence of translocation and introduction. A response of 'Don't know' should be given where uld be given if the taxon is a novel introduction of a single specimen. 3. Invasive elsewhere 3.01 (C) 9 CL Has the species established one or more self sustaining populations beyond its native range? If uncertainty exists regarding the established, s elf sustaining population(s), i.e. whether they constitute a true introduction/translocation or simply a 'range expansion by natural means', then the answer is "Don't know". 3.02 (N) 10 CL In the species' introduced range, are there impacts to wild stoc ks of angling or commercial species? There should be documented evidence of real impacts (i.e. decline of native species, disease introduction or transmission), not just circumstantial or opinion based judgements. 3.03 (A) 11 CL In the species' introdu ced r ange, are there impacts to aquacultural, aquarium or ornamental species? There should be documented evidence of real impacts (e.g. increased control costs, reduced yields), not just circumstantial or opinion based judgements. 3.04 (E) 12 CL In the species' introduced range, are there impacts to rivers, lakes or amenity values? Documented evidence that the species has altered the structure or function of a natural ecosystem. 3.05 (C) 13 CL Does the species have invasive congeners? One or more sp ecies within the genus are known to exert moderate to severe impacts.
70 Table A 1. Continued Section/ (Code) # Criteria Query Guidance B. Biology/Ecology 4. Undesirable (or persistence) traits 4.01 (C) 14 CL Is the species poisonous /venomous, or poses other risks to human health? Applicable if the taxon's presence is known, for any reason, to cause discomfort or pain to animals. 4.02 (C) 15 CL Does the species out compete with native species? There should be documented evidence that the tax on is responsible for suppression of growth or survival, and/or displacement from microhabitat, of native species. 4.03 (C) 16 NC Is the species parasitic of other species? Needs at least some documentation of being a parasite of other species ( e.g. sc ale or fin nipping such as known for Pseudorasbora parva, blood sucking such a s by some lampreys). 4.04 (A) 17 NC Is the species unpalatable to, or lacking, natural predators? This should be considered with respect to the likely level of ambient natural or human predation, if any. 4.05 (C) 18 E Does species prey on a native species previously subjected to low (or no) predation? There should be some evidence that the taxon is likely to establish in a hydrosystem in which predatory fish have never been present, or that is normally devoid of predatory fish (e.g. amphibian ponds), or of a fish species that possesses a predation facilitating biological attribute (e.g. behaviour, large body size, appearance). 4.06 (C) 19 CL Does the species host, and/or is it a vector, for one or more recognized non native infectious agents? The main concerns are non native pathogens and parasites, with the host either being the original introduction vector of the disease or as a host of the disease brought in by another taxon. 4.07 (N) 20 E Does the species achieve a large ultimate body size (i.e. > 15 cm total length) (more likely to be abandoned)? Although small bodied fishes may be abandoned, large bodied fishes are the major concern, as they soon outgrow their aqu arium or garden pond. 4.08 (E) 21 CL Does the species have a wide salinity tolerance or is euryhaline at some stage of its life cycle? Presence in low salinity water bodies (e.g. Baltic Sea Tampa Bay ) does not constitute euryhaline, so minimum salinity level should be about 15 %. 4.09 (E) 22 CL Is the species able to withstand being out of water for extended periods (e.g. minimum of one or more hours)? Examples are lungfishes, walking catfishes and species with desiccation tolerant eggs.
71 Table A 1. Continued Section/ (Code) # Criteria Query Guidance 4.10 (E) 23 NC Is the species tolerant of a range of water velocity conditions (e.g. versatile in habitat use)? Species that are known to persist in both standing and flowing waters over a wide rang e of velocities (0 to 0.7 m per sec). 4.11 (E) 24 CL, E Does feeding or other behaviours of the species reduce habitat quality for native species? There should be evidence of bio engineering behaviour, such as foraging that leads to the destruction of m acrophytes or an increase in suspended solids, reducing water clarity (e.g. as demonstrated for common carp) or burrow construction, which undermines bank character and stability (e.g. armoured sailfin catfishes). 4.12 (C) 25 NC Does the species requi re minimum population size to maintain a viable population? If evidence of population crash or extirpation due to low numbers (e.g. over exploitation or pollution), then response should be: 'yes'. 5. Feeding guild 5.01 (E) 26 E If the species is mainly herbivorous or piscivorous/carnivorous ( e.g. amphibia), then is its foraging likely to have an adverse impact in the RA area ? Obligate herbivores and piscivores (as adults) are most likely to score here, except where there is sufficient documented eviden ce form the RA area (or an area considered very similar) that the species has not exerted adverse impacts and therefore the appropriate response is "No" For a herbivorous species to score here, it must feed primarily on aquatic macrophytes. In the case of some facultative piscivores, they may become more piscivorous when confronted with nave prey. 5.02 (C) 27 E If the species is an omnivore (or a generalist predator), then is its foraging likely to have an adverse impact in the RA area ? There must be evidence of foraging on a wide range of food types, including incidental piscivory For obligate piscivores (as adults) that go through ontogenetic dietary changes (e.g. from zooplankton, to macrobenthos to fish), respond 'Yes' here, but then respond 'No' to the next two questions. 5.03 (C) 28 E If the species is mainly planktivorous or detritivorous or algivorous, then is its foraging likely to have an adverse impact in the RA area? Should be primarily planktivorous detritivorous or algivorous to scor e here For obligate piscivores (as adults) that go through ontogenetic dietary changes that include these food types (e.g. from zooplankton, to macrobenthos to fish), respond 'No' here. Similarly, if there is sufficient documented evidence from the RA are a (or an area considered very similar) that the species has not exerted adverse impacts, then the appropriate response is "No".
72 Table A 1. Continued Section/ (Code) # Criteria Query Guidance 5.04 (C) 29 E If the species is mainly benthivorous, then is its foraging likely to have an adverse impact in the RA area? Should be primarily benthivorous to score here For obligate piscivores (as adults) that go through ontogenetic dietary changes that include these food types (e.g. from zooplankton, to macro benthos to fish), respond 'No' here. 6. Reproduction 6.01 (C) 30 CL Does the species exhibit parental care and/or is it known to reduce age at maturity in response to environment? Needs at least some documentation of expressing parental care includin g nest guarding, mouth brooding, live bearing, etc. 6.02 (C) 31 CL Does the species produce viable gametes? A 'Yes' response requires evidence that the taxon produces viable gametes in the wild (native or introduced range). Functionally sterile hybrids sub 6.03 (A) 32 NC Does the species hybridize naturally with native species (or uses males of native species to activate eggs)? Documented evidence exists of interspecific hybrids occurring, without assis tance, under natural conditions. 6.04 (C) 33 NC Is the species hermaphroditic? Needs at least some documentation of hermaphroditism. 6.05 (C) 34 NC Is the species dependent on the presence of another species (or specific habitat features) to comple te its life cycle? Some species may require specialist incubators ( e.g. unionid mussels used by Rhodeus amarus ) or specific habitat features (e.g. fast flowing water, particular species of plant or types of substrata) in order to reproduce successfully. 6.06 (A) 35 CL, E Is the species highly fecund (>10,000 eggs/kg), iteropatric or has an extended spawning season relative to native species ? Normally observed in medium to longer lived species. 6.07 (C) 36 NC What is the species' known minimum generat ion time (in years)? Time from hatching to full maturity (i.e. active reproduction, not just presence of gonads). Please specify the number of years. 7. Dispersal mechanisms 7.01 (A) 37 CL Are life stages likely to be dispersed unintentionally? Uninten tional dispersal resulting from human activity (e.g. bait buckets, live eggs on anglers' gear). 7.02 (C) 38 CL Are life stages likely to be dispersed intentionally by humans (and suitable habitats abundant nearby)? Taxon has properties that make it att ractive or desirable (e.g. as a food fish or an angling amenity, for ornament or unusual appearance, for cultural reasons). 7.03 (A) 39 NC Are life stages likely to be dispersed as a contaminant of commodities? Taxon is associated with organisms likely to be sold commercially.
73 Table A 1. Continued Section/ (Code) # Criteria Query Guidance 7.04 (C) 40 NC Does natural dispersal occur as a function of egg dispersal? There should be documented evidence that eggs are taken by water currents 7.05 (E ) 41 NC Does natural dispersal occur as a function of dispersal of larvae (along linear and/or 'stepping stone' habitats)? There should be documented evidence that larvae enter, or are taken by, water currents, or can move between water bodies via connecti ons. 7.06 (E) 42 CL Are juveniles or adults of the species known to migrate (spawning, smolting, feeding)? There should be documented evidence of migratory behaviour, even at a small scale ( hundreds or thousands of meters ). 7.07 (C) 43 CL Are eggs o f the species known to be dispersed by other animals (externally)? There should be documented evidence of such movement events, e .g. accidentally by water fowl when they move from water body to water body. 7.08 (C) 44 NC Is dispersal of the species den sity dependent? There should be documented evidence of the taxon spreading out or dispersing when its population density increases. 8. Tolerance attributes 8.01 (C) 45 NC Any life stages likely to survive out of water transport? There should be doc umented evidence of the taxon being able to survive for an extended period (e.g. an hour or more) out of water. 8.02 (C) 46 E Does the species tolerate a wide range of water quality conditions, especially oxygen depletion & temperature extremes ? This is to identify taxa that can persist even in cases of low oxygen and/or elevated toxic levels of normal chemicals (e.g. ammonia) and/or temperature extremes 8.03 (A) 47 CL Is the species readily susceptible to piscicides at the doses legally permitted fo r use in the risk assessment area? To score a 'no' response, there must be documented evidence of the taxon's resistance to chemical control agents at the doses legally permitted for use in the risk assessment area. 8.04 (A) 48 E Does the species tolera te or benefit from environmental disturbance? Growth and spread of taxon may be enhanced by disruptions or unusual events (floods, spates, desiccation), including both short and long term human impacts 8.05 (C) 49 CL, E Are there effective natural en emies of the species present in the risk assessment area ? A known, effective, natural enemy of the taxon may or may not be present in the risk assessment area (this includes infectious agents that would impede establishment). Unless a specific enemy (or en emies) is known, answer 'Don't know'. Note: Sector codes (in parentheses) are: A = Aquaculture; E = Environmental; N = Nuisance; C = Combined. Criteria indicates the reason for making the change; CL = Clarity; CM = Climate, E = Ecological; NC = No Change.
74 Table A 2 FISK v2 protocol with responses given for Barcoo Grunter Scortum barcoo as a real world application of FISK v2 in a management scenario. Species Name: Scortum barcoo Common Name: Barcoo Grunter Assessor: S. Hardin Section/(Code) Qu ery Response Certainty Comments and References A. Biogeography/historical 1. Domestication/Cultivation 1.01 (C) Is the species highly domesticated or widely cultivated for commercial, angling or ornamental purposes? Y 3 Several references in Chinese aq uaculture literature (e.g. Journal of Fisheries Sciences of China; http://en.cnki.com.cn/Article_en/CJFDTOTAL ZSCK200706023.htm); Australian aquaculture references (e.g. http://www.aquaculturequeensland.com/jade_perch.htm) ; popular aquaponics fish (e.g. ht tp://farmingwithfish.com/?p=949). 1.02 (C) Has the species established self sustaining populations where introduced? N 3 Only one introduction referenced in FishBase (China) with outcome unknown and presumably not established 1.03 (C) Does the species have invasive races/varieties/sub species? N 4 No subspecies listed in FishBase or ITIS (http://www.itis.gov/servlet/SingleRpt/SingleRpt?search _topic=TSN&search_value=168075) 2. Climate and distribution 2.01 (C) Is species reproductive tolerance suite d to climates in the risk assessment area (1 low, 2 medium, 3 high)? 3 3 Native range in Australia from Northern Queensland and Northern Territory south to Lake Eyre drainage in central Australia (http://www.dpi.qld.gov.au/28_14677.htm); latitude is from 1 5 S 28 S (equivalent latitude is central Florida and south). FishBase lists temperature range as 10 30 C, but Chinese literature identifies lowest and highest critical temperature of embryonic development as 21 C and 31 C, respectively and optimal temp erature from 24 27 C (Chen et al. 2007). 2.02 (C) What is the quality of the climate match data (1 low; 2 medium ; 3 high)? 2 3 Match based on latitude and aquaculture research rather than water temperature data from native range
75 Table A 2. Continue d Section/(Code) Query Response Certainty Comments and References 2.03 (C) Does the species demonstrate broad climate suitability? Y 3 FishBase lists as temperate. Australian distribution covers 3 Kppen Geiger zones; however none of the three occur i n Florida (map from Peel et al. 2007) 2.04 (C) Is the species native to, or established self sustaining populations in regions with equable climates to the RA area? Y 3 Native range has latitude distribution equivalent to central and southern Florida; temperature range in FishBase similar to portions of peninsular Florida (see earlier questions for references). 2.05 (C) Does the species have a history of being introduced outside its natural range? N 2 FishBase lists one introduction in China witho ut documentation of establishment status. There has been considerable investigation into aquaculture of S. barcoo in China, which could be the source of the putative introduction. 3. Invasive elsewhere 3.01 (C) Has the species established one or more self sustaining populations beyond its native range? N 3 No reference to any established, introduced populations. FishBase reference to Chinese introduction without documentation. 3.02 (N) In the species' introduced range, are there impacts to wild sto cks of angling or commercial species? N 3 The only reference to introduction (FishBase) has no data on establishment or ecological effects. 3.03 (A) In the species' introduced r ange, are there impacts to aquacultural, aquarium or ornamental species? N 3 The only reference to introduction (FishBase) has no data on establishment or ecological effects. 3.04 (E) In the species' introduced range, are there impacts to rivers, lakes or amenity values? N 3 The only reference to introduction (FishBase) has no data on establishment or ecological effects. 3.05 (C) Does the species have invasive congeners? N 4 Three congeners noted in FishBase (S. hillei, S. neili, and S. parviceps) with no reference to introductions. B. Biology/Ecology 4. Undesirable ( or persistence) traits 4.01 (C) Is the species poisonous /venomous, or poses other risks to human health? N 4 FishBase status = Harmless. No references to venom.
76 Table A 2. Continued Section/(Code) Query Response Certainty Comments and References 4.02 (C) Does the species out compete with native species? N 3 No documentation of ecological effects in the only introduction noted in FishBase. No other references located. 4.03 (C) Is the species parasitic of other species? N 4 No reference to thi s in FishBase or Australian summary (http://www.dpi.qld.gov.au/28_14677.htm). 4.04 (A) Is the species unpalatable to, or lacking, natural predators? N 3 Common length of 25 cm would make this species vulnerable to a host of medium to large piscivores, wading birds and crocodilians. Considering its value as a food fish, palatability is likely not an issue for non human predators. 4.05 (C) Does species prey on a native species previously subjected to low (or no) predation? N 3 FishBase lists diet ite ms as fishes, crustaceans, insects and mollusks. In Florida, native predatory fishes occur in virtually every water body where these prey items are found. 4.06 (C) Does the species host, and/or is it a vector, for one or more recognized non native infe ctious agents? N 3 No references found that address vectoring non native parasites and disease. Moreover, no data on the single introduction (China) listed in FishBase. Chinese aquaculture literature does not address this issue. 4.07 (N) Does the speci es achieve a large ultimate body size (i.e. > 15 cm total length) (more likely to be abandoned)? Y 4 Max size 35 cm, common size 25 cm (FishBase). 4.08 (E) Does the species have a wide salinity tolerance or is euryhaline at some stage of its life cycle ? N 3 FishBase lists as freshwater; Appendix table to Australian fish distribution study lists as freshwater only (http://www.docstoc.com/docs/70966492/Appendix I Fish species distribution abundance and trends by catchment). 4.09 (E) Is the species able to withstand being out of water for extended periods (e.g. minimum of one or more hours)? N 3 No references mention unusual desiccation tolerance. 4.10 (E) Is the species tolerant of a range of water velocity conditions (e.g. versatile in habitat use) ? Y 3 Native to Gilbert River, which has the sixth highest discharge of any river in Australia (Wikipedia); FishBase suggests that fish spawn during flooding.
77 Table A 2. Continued Section/(Code) Query Response Certainty Comments and References 4.11 (E) Does feeding or other behaviours of the species reduce habitat quality for native species? N 3 No evidence of bio engineering behavior. 4.12 (C) Does the species require minimum population size to maintain a viable population? ? 1 No references fou nd. 5. Feeding guild 5.01 (E) If the species is mainly herbivorous or piscivorous/carnivorous (e.g. amphibia), then is its foraging likely to have an adverse impact in the RA area ? N 3 Considered an omnivore by FishBase. 5.02 (C) If the species is an omnivore (or a generalist predator), then is its foraging likely to have an adverse impact in the RA area ? N 3 FishBase lists as omnivore; food habits are similar to native Florida fishes and without novel predatory behavior or choice in habitats, impac ts do not seem likely. 5.03 (C) If the species is mainly planktivorous or detritivorous or algivorous, then is its foraging likely to have an adverse impact in the RA area? N 4 Species is omnivorous (FishBase). 5.04 (C) If the species is mainly ben thivorous, then is its foraging likely to have an adverse impact in the RA area? N 4 Species is omnivorous (FishBase). 6. Reproduction 6.01 (C) Does the species exhibit parental care and/or is it known to reduce age at maturity in response to environm ent? Y 2 FishBase considers this a parental guarder, with males fanning eggs per Breder and Rosen 1966 Modes of reproduction in Fishes. However, Chines literature suggests that the egg is buoyant (Chen et al. 2007). 6.02 (C) Does the species produce vi able gametes? Y 4 Reproducing population through much of Australia. 6.03 (A) Does the species hybridize naturally with native species (or uses males of native species to activate eggs)? N 4 No native congeners in Florida; Terapontidae is not native to Florida. 6.04 (C) Is the species hermaphroditic? N 4 No references to this were found.
78 Table A 2. Continued Section/(Code) Query Response Certainty Comments and References 6.05 (C) Is the species dependent on the presence of another species (o r specific habitat features) to complete its life cycle? N 4 No references to this aspect of life cycle were found. 6.06 (A) Is the species highly fecund (>10,000 eggs/kg), iteropatric or has an extended spawning season relative to native species ? ? 2 Could not find references on fecundity; not listed in FishBase; Chinese and Australian aquaculture literature do not mention fecundity. 6.07 (C) What is the species' known minimum generation time (in years)? 2 2 No references, but FishBase doubling ti me listed as 1.4 4.4 years. 7. Dispersal mechanisms 7.01 (A) Are life stages likely to be dispersed unintentionally? N 3 Not a bait species; potential aquaponics species. 7.02 (C) Are life stages likely to be dispersed intentionally by humans (a nd suitable habitats abundant nearby)? Y 2 Potential as aquaponics food fish may lead to stocking; or discards from people who abandon aquaponics systems. 7.03 (A) Are life stages likely to be dispersed as a contaminant of commodities? N 3 Would be ver y unlikely to be sold with other fishes; very low probability as contaminant of non fish commodities. 7.04 (C) Does natural dispersal occur as a function of egg dispersal? N 1 FishBase suggests parental care and pelagic eggs in same paragraph; however, as noted above, Chinese literature suggests floating eggs. 7.05 (E) Does natural dispersal occur as a function of dispersal of larvae (along linear and/or 'stepping stone' habitats)? ? 1 No literature or references found that address this. 7.06 (E ) Are juveniles or adults of the species known to migrate (spawning, smolting, feeding)? ? 1 Only reference is in FishBase which suggests this species may breed during flood events (Allen et al. 2002 Field guide to the freshwater fishes of Australia). Uncl ear if migration is associated with this. 7.07 (C) Are eggs of the species known to be dispersed by other animals (externally)? N 3 No documentation, and unlikely if there is parental care or buoyant eggs. 7.08 (C) Is dispersal of the species dens ity dependent? ? 1 No references found that address stock densities in the wild. 8. Tolerance attributes
79 Table A 2. Continued Section/(Code) Query Response Certainty Comments and References 8.01 (C) Any life stages likely to survive out of wa ter transport? N 4 No references mention desiccation tolerance; not an air breather. 8.02 (C) Does the species tolerate a wide range of water quality conditions, especially oxygen depletion & temperature extremes ? Y 2 FishBase lists temperature of 10 40 C, although aquaculture literature suggests fish are not active at temperatures below 20 C. Species does cover a relatively large portion of Australia. 8.03 (A) Is the species readily susceptible to piscicides at the doses legally permitted for us e in the risk assessment area? Y 4 Rotenone or synthetic should be lethal at maximum label does (5 parts per million), which is allowable in Florida. 8.04 (A) Does the species tolerate or benefit from environmental disturbance? Y 2 FishBase reference to spawning during flooding; species adapted to turbid rivers, likely due to significant run off. 8.05 (C) Are there effective natural enemies of the species present in the risk assessment area ? Y 3 Species does not exhibit unusual morphology or def ensive behaviors. Would be subject to native predators. Statistical summary of scoring Outcome UK: Medium Outcome Japan: Medium Statistical summary of scoring Outcome User defined: Medium Total questions: 49 Total Score: 5 Answered: 49 Section point totals: Biogeography: 4 Unanswered: 0 Undesirable attributes: 2 Biology/ecology: 1 Questions answered: Biogeography: 10 Undesirable attributes: 11 Biology/ecology: 20 Total An swered: 41 Sector affected: Aquacultural: 3 Environmental: 7 Nuisance: 1 Notes: Sector codes (in parentheses) are: A = Aquaculture; E = Environmental; C = Combined. Scoring sub routines for (2) and other responses
80 Figure A 1. New and improved inte rface of the Species Assessment Menu dialog in FISK v2 (data from Vilizzi and Copp 2013).
81 Figure A 2. New and improved interface of the Q&A dialog in FISK v2 (Q 1 for Barcoo grunter Scortum barcoo).
82 Discussion The upgrades and modifications have made FISK v2 more functional than its predecessor (FISK v1) in terms of its use as a practical tool for identifying potentially invasive fish, thus informing the decision making and risk management process undertaken by regulatory agencies. FISK v2 incorporate s a wider range of ecological and environmental characteristics that facilitate invasion on a broader scale of risk environments (i.e. temperate to tropical climates), making FISK v2 applicable to peninsular Florida as well as other warm temperate, sub tro pical and, potentially, tropical climate zones. The changes made to questions and question guidance reduce ambiguity in the terminology while also allowing increased flexibility in assessor judgment where applicable (e.g. determining invasion history, clim ate matching, and evaluating impacts). Noteworthy upgrades were made to the FISK v2 user interface that improved its efficiency and reporting ability and provides the assessor a greater level of control during use. Beyond the application to the Barcoo gr unter in the present study, FISK v2 is being used in two sub tropical regions, namely Florida (U.S.A.) and south eastern Australia. In t he later study, Vilizzi and Copp (2013) evaluated 55 species in the Murray Darling Basin and compared the resulting scor es for 53 of the species with those previously obtained (by LV) for t he United Kingdom using FISK v1 (Copp et al. 2009). They found differences in scores between the two FISK versions, with the main difference due to methodological changes in the Feeding G uild questions (most species) and the question on ultimate body size (four species with total lengths >10 cm but <15 cm), which led to lower scores. Other differences were mainly attributable to
83 climate predictions that suggest Australia, unlike the United Kingdom, will be drier in t he future (see Murphy and Timbal 2008 ). An additional o bjective of the Vilizzi and Copp (2013) study was to compare the risk categories from FISK v2 with those obtained from two Australian based risk assessment protocols (Vilizz i and Copp 2013). The authors proven track record of the FISK tool in several countries and regions world wide compared to screening approaches currently limited to Australi a, its semi quantitative nature as well as its antipodean origin as an a daptation of the Australian WRA (Pheloung et al. 1999). The application of FISK v2 to Barcoo grunter represents the first of a more extensive application of this screening tool to asse ss 98 non native fishes for peninsular Florida (LLL, LV, JEH, SH and GHC, unpublished data). Although the FISK score for Barcoo grunter falls into the that category. Responses to questions were compliant with the FISK guidance and reflected a risk averse approach appropriate for a state conservation agency. Many FISK questions are quantitative, but judgment is required in some cases, e.g. source data quality, degree of climate match, and likeliho od of introduction. Despite increasing use of Barcoo grunter in aquaculture, information in databases and the literature was conflicting or contradictory, even for basic concepts such as native range and reproduction. Moreover, data quality for invasion hi story was poor. These deficiencies increased uncertainty, affected answers to questions, and influenced the final score. references, are valuable for subsequent review and interpretation by environmental
84 managers. A strength of a risk identification (screening) phase is to point out gaps in knowledge that may require additional investigation in the literature, interaction with species experts, or original research, espec ially when evaluating a species that is classified as medium risk. The FISK evaluation of Barcoo grunter suggests that this species may be a suitable candidate for aquaponics in Florida and that restrictions on possession may not be necessary. Variability in assessor judgment is undesirable in risk assessments, in that it may introduce subjectivity; however, in some cases human judgment may improve the quality of the assessment by buffering data deficiencies or inconsistencies (Orr 2003). Nevertheless, wher e data/information are lacking, such as for the Barcoo grunter, increased uncertainty may motivate environmental managers/decision makers to take a precautionary approach. Several responses in our Barcoo grunter screening would have changed if assessor jud gment was not permitted. Alternatively, assessor subjectivity may also be introduced if the screening questions are not specific, clearly written, or require assessor interpretation. The changes made for FISK v2 have addressed these concerns by simultaneou sly improving clarity of the questions and providing the assessor with clear er guidance in selecting the appropriate responses while also allowing the assessor latitude in judging the quality of data and information available and responding accordingly. Ac knowledgements Collaborative authors on this project were Dr. Jeffrey E. Hill, Mr. Scott Hardin, Dr. Lorenzo Vilizzi, and Dr. Gordon Copp. Funding for this research was provided by the U.S. Department of Agriculture T STAR C program, the University of Flor ida College of
85 Agricultural and Life Sciences, the Florida Fish and Wildlife Conservation Commission, and The UK Department of Environment, Food and Rural Affairs through contracts to Cefas.
86 APPENDIX B EXPERT INVASIVENESS SURVEY QUESTIONNAIRE Table B 1 Example of the expert survey of invasiveness for non native fishes in peninsular Florida # Family Scientific Name Common Name Risk Rating Information Source Benefit Rating 1 Anabantidae Anabas testudineus Climbing Perch Don't Know 2 Anabantidae C tenopoma nigropannosum Two spot Climbing Perch Don't Know 3 Anostomidae Leporinus fasciatus Banded Leporinus Don't Know 4 Belontiidae Betta splendens Siamese Fighting Fish 1 LOW RISK Internet 3 MODERATE 5 Belontiidae Colisa fasciata Banded Goura mi 1 LOW RISK Internet 3 MODERATE 6 Belontiidae Colisa labiosa Thicklipped Gourami 1 LOW RISK Internet 3 MODERATE 7 Belontiidae Colisa lalia Dwarf Gourami 1 LOW RISK Internet 3 MODERATE 8 Belontiidae Macropodus opercularis Paradisefish 1 LOW RISK Inter net 3 MODERATE 9 Belontiidae Trichogaster leerii Pearl Gourami 1 LOW RISK Internet 3 MODERATE 10 Belontiidae Trichogaster trichopterus Blue Gourami 1 LOW RISK Internet 3 MODERATE 11 Belontiidae Trichopsis vittata Croaking Gourami 1 LOW RISK Published L it. 2 MINIMAL 12 Callichthyidae Callichthys callichthys Cascarudo Don't Know 13 Callichthyidae Hoplosternum littorale Brown Hoplo 1 LOW RISK Published Lit. 2 MINIMAL 14 Channidae Channa argus Northern Snakehead 3 HIGH RISK Internet 1 NONE 15 Chan nidae Channa marulius Bullseye Snakehead 2 MEDIUM RISK Gray Lit. 3 MODERATE 16 Characidae Aphyocharax anisitsi Bloodfin Tetra 1 LOW RISK Internet 3 MODERATE 17 Characidae Colossoma macropomum Tambaqui/ Pacu 1 LOW RISK Published Lit. 3 MODERATE 18 Charac idae Gymnocorymbus ternetzi Black Tetra 1 LOW RISK Internet 3 MODERATE 19 Characidae Metynnis Sp. Silver Dollar 1 LOW RISK Gray Lit. 3 MODERATE 20 Characidae Moenkhausia sanctaefilomenae Redeye Tetra 1 LOW RISK Internet 3 MODERATE 21 Characidae Piaractu s brachpomus Red bellied Pacu 1 LOW RISK Published Lit. 3 MODERATE 22 Characidae Pygocentrus nattereri Red Piranha 3 HIGH RISK Personal Know 1 NONE 23 Characidae Serrasalmus rhombeus White Piranha 3 HIGH RISK Personal Know 1 NONE 24 Cichlidae Aequidens pulcher Blue Acara 1 LOW RISK Internet 2 MINIMAL 25 Cichlidae Astronotus ocellatus Oscar 1 LOW RISK Published Lit. 3 MODERATE
87 Table B 1. Continued # Family Scientific Name Common Name Risk Rating Information Source Benefit Rating 26 Cichlidae Cichla o cellaris Butterfly Peacock Bass 1 LOW RISK Published Lit. 4 SUBSTANTIAL 27 Cichlidae Cichla temensis Speckled Pavon 1 LOW RISK Published Lit. 1 NONE 28 Cichlidae Cichlasoma bimaculatum Black Acara 1 LOW RISK Published Lit. 2 MINIMAL 29 Cichlidae Cichlas oma citrinellum Midas Cichlid 1 LOW RISK Published Lit. 3 MODERATE 30 Cichlidae Cichlasoma cyanogottatum Rio Grande Cichlid 1 LOW RISK Published Lit. 1 NONE 31 Cichlidae Cichlasoma meeki Firemouth Cichlid 1 LOW RISK Published Lit. 2 MINIMAL 32 Cichlidae Cichlasoma nigrofasciatum Convict Cichlid 1 LOW RISK Published Lit. 2 MINIMAL 33 Cichlidae Cichlasoma octofasciatum Jack Dempsey 1 LOW RISK Published Lit. 3 MODERATE 34 Cichlidae Cichlasoma salvini Yellowbelly Cichlid 1 LOW RISK Published Lit. 1 NONE 3 5 Cichlidae Cichlasoma trimaculatum Threespot Cichlid 1 LOW RISK Internet 1 NONE 36 Cichlidae Cichlasoma urophthalmus Mayan Cichlid 2 MEDIUM RISK Personal Know 3 MODERATE 37 Cichlidae Geophagus sp. Eartheater 1 LOW RISK Published Lit. 2 MINIMAL 38 Cichl idae Haplochromis callipterus Eastern Happy 1 LOW RISK Published Lit. 2 MINIMAL 39 Cichlidae Hemichromis letourneuxi African Jewelfish 2 MEDIUM RISK Published Lit. 3 MODERATE 40 Cichlidae Heros severus Banded Cichlid 1 LOW RISK Published Lit. 2 MINIMAL 41 Cichlidae Oreochromis aureus Blue Tilapia 1 LOW RISK Personal Knowl. 3 MODERATE 42 Cichlidae Oreochromis mossambicus Mozambique Tilapia 1 LOW RISK Gray Lit. 3 MODERATE 43 Cichlidae Oreochromis niloticus Nile Tilapia 1 LOW RISK Colleagues 3 MODERATE 4 4 Cichlidae Cichlasoma managuense Jaguar Guapote 1 LOW RISK Gray Lit. 45 Cichlidae Pterophyllum scalare Freshwater Angelfish 1 LOW RISK Internet 2 MINIMAL 46 Cichlidae Sarotherodon melanotheron Blackchin Tilapia 2 MEDIUM RISK Internet 2 MINIMAL 47 Cic hlidae Theraps melanurus x T. zonatus Theraps Hybrid 1 LOW RISK Published Lit. 2 MINIMAL 48 Cichlidae Tilapia buttikoferi Zebra Tilapia 2 MEDIUM RISK Gray Lit. 2 MINIMAL 49 Cichlidae Tilapia mariae Spotted tilapia 1 LOW RISK Gray Lit. 1 NONE 50 Cichlida e Tilapia zillii Redbelly Tilapia 3 HIGH RISK Personal Knowl. 2 MINIMAL 51 Clariidae Clarias batrachus Walking Catfish 1 LOW RISK Colleagues 1 NONE 52 Clupeidae Alosa sapidissima American Shad 1 LOW RISK Colleagues 3 MODERATE 53 Clupeidae Dorosoma peten ense Threadfin Shad 1 LOW RISK Personal Knowl. 3 MODERATE
88 Table B 1. Continued # Family Scientific Name Common Name Risk Rating Information Source Benefit Rating 54 Cobitidae Misgurnus anguillicaudatus Oriental Weatherfish 1 LOW RISK Published Lit. 2 M INIMAL 55 Cobitidae Pangio kuhlii Coolie Loach Don't Know 56 Cyprinidae Barbonymus schwanenfeldii Tinfoil Barb Don't Know 57 Cyprinidae Danio rerio Zebra Danio 1 LOW RISK Personal Knowl. 3 MODERATE 58 Cyprinidae Carassius auratus Goldfish 1 LOW RISK Personal Knowl. 3 MODERATE 59 Cyprinidae Ctenopharyngodon idella Grass Carp 2 MEDIUM RISK Personal Knowl. 4 SUBSTANTIAL 60 Cyprinidae Cyprinus carpio Common Carp 1 LOW RISK Internet 1 NONE 61 Cyprinidae Danio malabaricus Malabar Danio 1 LOW RISK I nternet 3 MODERATE 62 Cyprinidae Hypophthalmichthys nobilis Bighead Carp 1 LOW RISK Internet 2 MINIMAL 63 Cyprinidae Labeo chrysophekadion Black Sharkminnow 1 LOW RISK Internet 1 NONE 64 Cyprinidae Pimephales promelas Fathead Minnow 1 LOW RISK Personal Knowl. 4 SUBSTANTIAL 65 Cyprinidae Puntius conchonius Rosy Barb 1 LOW RISK Internet 3 MODERATE 66 Cyprinidae Puntius gelius Dwarf Barb 1 LOW RISK Internet 3 MODERATE 67 Cyprinidae Puntius tetrazona Tiger Barb 1 LOW RISK Internet 3 MODERATE 68 Doradidae Platydoras costatus Raphael Catfish Don't Know 69 Doradidae Pseudodoras niger Ripsaw Catfish Don't Know 70 Doradidae Pterodoras granulosus Granulated Catfish Don't Know 71 Erythrinidae Hoplias malabaricus Trahira 3 HIGH RISK Personal Knowl. 72 Helostomatidae Helostoma temmincki Kissing Gourami 1 LOW RISK Personal Knowl. 3 MODERATE 73 Ictaluridae Pylodictis olivaris Flathead Catfish 3 HIGH RISK Published Lit. 3 MODERATE 74 Loricariidae Ancistrus sp. Bristlenosed Catfish 1 LOW RISK Collea gues 2 MINIMAL 75 Loricariidae Hypostomus sp. Suckermouth Catfish 1 LOW RISK Gray Lit. 3 MODERATE 76 Loricariidae Pterygoplichthys anisitsi Southern Sailfin Catfish Don't Know Colleagues 3 MODERATE 77 Loricariidae Pterygoplichthys disjunctivus Vermicula ted Sailfin Catfish 2 MEDIUM RISK Gray Lit. 3 MODERATE 78 Loricariidae Pterygoplichthys multiradiatus Orinoco Sailfin Catfish 2 MEDIUM RISK Gray Lit. 3 MODERATE 79 Mastacembelidae Macrognathus siamensis Spotfin Spiny Eel 1 LOW RISK Colleagues 1 NONE 80 Moronidae Morone chrysops x M. saxatilis Wiper 1 LOW RISK Personal Knowl. 3 MODERATE 81 Notopteridae Chitala ornata Clown Knife 1 LOW RISK Gray Lit. 2 MINIMAL
89 Table B 2. Continued # Family Scientific Name Common Name Risk Rating Information Source Benefit Rating 82 Osteoglossidae Osteoglossum bicirrhosum Silver Arowana 1 LOW RISK Colleagues 2 MINIMAL 83 Percidae Sander vitreus Walleye 1 LOW RISK Gray Lit. 1 NONE 84 Pimelodidae Rhamdia quelen Bagre 1 LOW RISK Internet 85 Poeciliida e Belonesox belizanus Pike Killifish 1 LOW RISK Published Lit. 86 Poeciliidae Gambusia affinis Western Mosquitofish 1 LOW RISK Personal Knowl. 87 Poeciliidae Poecilia latipinna x velifera Black Molly 1 LOW RISK Internet 3 MODERATE 88 Poeciliidae Poe cilia latipunctata Tamesi Molly 1 LOW RISK Internet 2 MINIMAL 89 Poeciliidae Poecilia petenensis Peten Molly 1 LOW RISK Internet 2 MINIMAL 90 Poeciliidae Poecilia reticulata Guppy 1 LOW RISK Personal Knowl. 3 MODERATE 91 Poeciliidae Poecilia sphenops Me xican Molly 1 LOW RISK Internet 3 MODERATE 92 Poeciliidae Xiphophorus hellerii Green Swordtail 1 LOW RISK Colleagues 3 MODERATE 93 Poeciliidae Xiphophorus helerii x maculatus Red Swordtail 1 LOW RISK Colleagues 3 MODERATE 94 Poeciliidae Xiphophorus macu latus Southern Platyfish 1 LOW RISK Colleagues 2 MINIMAL 95 Poeciliidae Xiphophorus variatus Variable Platyfish 1 LOW RISK Colleagues 2 MINIMAL 96 Polypteridae Polypterus delhezi Bichir 1 LOW RISK Gray Lit. 97 Schilbeidae Platytropius siamensis False Siamese Shark 1 LOW RISK Published Lit. 98 Synbranchidae Monopterus albus Asian Swamp Eel 2 MEDIUM RISK Published Lit. 2 MINIMAL
90 L IST OF REFERENCES Allen G R S.H. Midgley, and M. Allen. 2002. Field guide to freshwater fishes of Australia. Perth, We stern Australia: Western Australian Museum. Almeida, D., F. Ribeiro, P.M Leunda, L. Vilizzi, and G.H. Copp. 2013. Effectiveness of FISK, an invasiveness screening tool for non native freshwater fishes, to perform risk identification assessments in the Ib erian Peninsula. Risk Analysis 33(8):1404 1413. ANSTF (Aquatic Nuisance Species Task Force). 1996. Generic nonindigenous aquatic organisms risk analysis review process: For estimating risk associated with the introduction of nonindigenous organisms and ho w to manage for that risk. Aquatic Nuisance Species Task Force, Washington, D.C. ANSTF (Aquatic Nuisance Species Task Force). 2013. Aquatic Nuisance Species Task Force Strategic Plan 2013 2017. Washington, D.C. Available: http://www.anstaskforce.gov/Documents/ANSTF%20Strategic%20Plan%202013 2017.pdf (March 2013). Arthington A H J.D. Olden, S.R. Balcombe and M.S. Thoms. 2010. Multi scale environmental fac tors explain fish losses and refuge quality in drying waterholes of Cooper Creek, an Australian arid zone river. Marine and Freshwater Research 61:842 856. Baker R H R. Black, C opp GH and 23 more authors. 2008. The UK risk assessment scheme for all no n native species. Neobiota 7:46 57 Balcombe A R S.E. Bunn, F.J. McKenzie Smith, and P.M. Davies. 2005. Variability of fish diets between dry and flood periods in an arid zone floodplain river. Journal of Fish Biology 67:1552 1567. Bewick, V., L. Chee k, and J. Ball. 2004. Statistics review 13: receiver operating characteristic curves. Critical Care 8:508 512. Bovey R D. Wallentin, S. Bullen, and J. Green. 2010. Professional Excel development: The definitive guide to developing applications using Mi crosoft Excel, VBA, and .NET 2 nd Edition, Addison Wesley Professional. Breder C M and D.E. Rosen. 1966. Modes of reproduction in fishes. Natural Histor y Press. Garden City, New York. Britton, J.R., J. Cucherousset, G.D. Davies, M.J. Godard, and G.H. Copp. 2010. Non native fishes and climate change: predicting species responses to warming temperatures in a temperate region. Freshwater Biology 55:1130 1141.
91 Bruton M N 1986. Life history styles of invasive fishes in southern Africa. In Macdonald F J A. Kruger, and A. Ferrar ed itors. The ecology and management of i nvasions in South Africa. Oxford University Press, Cape Town Burrows, D. 2008. tropical rivers. Sub project 1 of Australi an integrated data assessment and analysis (DET18) In Lukacs GP and Finlayson CM, editors. A report to Land & Water Australia. Townsville, Queensland: National Centre for Tropical Wetland Research. CAS ( California Academy of Science s ) 2012. Catalog of Fishes, Scortum ogilbyi W hitley 1951. Available: http://research.calac ademy.org/redirect?url=http://researcharchive.calacademy.or g/research/ichthyology/catalog/fishcatget.asp&tbl=species&spid=47977 ( March 2012 ) Chen K X. Zhu, J. Du, G. Xie Y. Liu, G. Zheng, and Y. Chen. 2007. Effects of temperature and salinity on th e embryonic development of jade perch Scortum barcoo Journal of Fishery Sciences of China English abstract available : http://en.cnki.com.cn/Article_en/CJFDTOTAL ZSCK200706023.htm (March 2012 ) CISAC (California Invasi ve Species Advisory Committee). 2010 The California invasive species list. As presented to the Invasive Species Council of California. Available: http://www.iscc.ca.gov/docs/CaliforniaInvasiveSpeciesList.pdf ( March 2014). Cooper A R. Reimann, and D. Cronin. 2007. About face 3: The essentials of interaction design. 3 rd Edition, Wiley. New York, New York. Copp, G.H. 2013. The Fish Invas iveness Screening Kit (FISK) for non native freshwater fishes A summary of current applications. Risk Analysis 33(8):1394 1396. Copp, G.H., J.R. Britton, G. Jeney, and 17 more authors. 2008. Risk assessment protocols and decision making tools for use of a lien species in aquaculture and stock enhancement. Report D3.2 to the European Commission Coordination Priority Action FP6 2005 SSP FA Project IMPASSE Environmental Impacts of Alien Species in Aquaculture. Available: http://www.cefas.defra.gov.uk/media/437410/impasse_44142_d3 2.pdf (March 2014). Copp G H L. Vilizzi J. Mumford, G.M. Fenwick, M.J. Godard, and R.E. Gozlan. 2009. Calibration of FISK, an invasiveness screenin g tool for nonnative freshwater fishes. Risk Analysis 29(3):457 467.
92 Copp G H M. Templeton M, and R.E. Gozlan. 2007. Propagule pressure and the invasion risks of non native freshwater fishes in Europe: a case study of England. Journal of Fish Biology 7 1(Suppl D):148 159. Copp G.H. P.G. Bianco, N.G. Bogutskaya NG and 19 more authors. 2005c. To be, or not to be, a non native freshwater fish? Journal of Applied Ichthyology 21:242 262 Copp G H R. Garthwaite, and R.E. Gozlan. 2005a. Risk identificat ion and assessment of non native freshwater fishes: concepts and perspectives on protocols for the UK. Cefas Science Technical Report. Lowestoft, UK: Cefas Available: http://www.cefas. co.uk/publications/techrep/tech129.pdf (April 2012 ). Copp G.H., R. Garthwaite, and R.E. Gozlan 2005b Risk identification and assessment of non native freshwater fishes: a summary of concepts and perspectives on protocols for the UK. Journal of Applie d Ichthyology 21:371 373 Daehler, C.C., and D.A. Carino. 2000. Predicting invasive plants: prospects for a general screening system based on current regional models. Biological Invasions 2:93 102. Daehler, C.C., J.S. Denslow, S. Ansari, and S. Kuo. 2004 A risk assessment system for screening out invasive pest plants from Hawaii and other Pacific islands. Conservation Biology 18:360 368. Davis, A.M., A.P. Grime, and K. Thompson. 2000. Fluctuating resources in plant communities: a general theory of invas ibility. Journal of Ecology 88:528 534 Davis A M R.G. Pearson B.J. Pusey, C. Perna, D.L. Morgan, and D. Burrows. 2011. feeding classification. Journal of Fish Biology 78:265 286. Delong, E.R., D.M. Delong, and D.L. Clarke Pearson. 1988. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 44:837 845 Dibble E D and K. Kovalenko. 2009. Ecolo gical impact of grass carp: a review of the available data. Journal of Aquatic Plant Management 47:1 15. FAO (Food and Agriculture Organization of the United Nations). 2012. Introduced species fact sheet for Scotum barcoo Av ailable : http://www.fao.org/fishery/introsp/3919/en (April 2012 ) FDACS (Florida Department of Agriculture and Consumer Services). 2007. Aquaculture best management practices rule. Division of Aquaculture, Tallahassee, Florida
93 FDACS (Florida Department of Agriculture and Consumer Services). 2009. Proceedings of the Florida sturgeon culture risk assessment workshop. Metcalf, K., P. Zajicek. eds. Florida Department of Agriculture and Consumer Services, Division of Aquaculture, Florida. Froese R and D. Pauly. 2012. Fishbase. World Wide Web electronic publication, 2011. Available: http://www.fishbase.org (March 2012 ). Fuller P L L.G. Nico, and J.D. Williams. 1999. Nonindigenous fishes introduced into inland waters of the United States. American Fisheries Society Special Publication 27. Bethesda, MD. Gestring K B P.L. Shafland, M.S. Stanford. Status of the exotic Orinoco sailfin catfish ( Pterygoplichthys multiradiatus ) in Florida. The Florida Scientist 73(2):122 137. Gordon D R D.A. Onderdonk, A.M. Fox, R.K. Stocker, and C. Gantz. 2008 a Predicting invasive plants i n Florida using the Australian weed risk a ssessment. Invasive Plant Science and Management 1(2):178 195 Gordon D R D.A. Onderdon k A.M. Fox, and R.K. Stocker 2008b. Consistent accuracy of the Australian weed risk assessment system across varied geographies. Diversity and Distributions 14:234 242 Gozl an R E J.R. Britton, I. Cowx, and G.H. Copp 2 010 Current knowledge on non native freshwater fish introductions. Journal of Fish Biology 76:751 786 Gozlan R E (2008) Introduction of non native freshwater fish: is it all bad? Fish and Fisheries 9:10 6 115 Greenwood M F D 2012 Assessing effects of the nonindigenous pike killifish on indigenous fishes in Tampa Bay, Florida, using a weighted evidence approach. Transactions of the American Fisheries Society 141(1):84 99 Hardin S 2007 Managing no n native wildlife in Florida: state perspective, policy, and practice. USDA National Wildlife Research Symposia: Managing Vertebrate Invasive Species. Available: http://digitalcommons.unl.edu/nwrin vasive/14 (March 2014). Hardin S 2009. Risk analysis of barramundi culture at BRA ROD, LLC. Final Report. Florida Fish and Wildlife Conservation Commission, Tallahassee, Florida Hardin S ., and J.E. Hill 2012. Risk analysis of Barramundi Lates calc arifer aquaculture in Florida. North American Journal of Fisheries Management 32(3):577 585
94 Hill J E 2002. Exotic fishes in Florida. LakeLines (North American Lake Management Society) 22:39 43 Hill J E 2009. Risk analysis for Non Native Species in Aquaculture. Southern Regional Aquacultur e Center. Publication No. 4304. Hill, J.E., and K. Thompson. 2008. Bioprofile of barramundi Lates calcarifer (Latidae). Report submitted to Florida Fish and Wildlife Conservation Commission, Tallahassee. Hill J E and P. Zajicek 2007 National aquatic species risk analysis: a call for improved implementation. Fisheries 32:530 538 Hill J E A.R. Kapuscinsk, and T. Pavlowich. 2011. Flourescent transgenic zebra danio more vulnerable to predators than wild type fish. Transactions of the American Fisheries Society 140:1001 1005. Hil l J E S.E. Hardin, and L.L. Lawson In Press Assessment of risks of transgenic fluorescent ornamental fishes to the United States using the Fish Invasiveness Screening Kit (FISK). North American Journal of Fisheries Management Invasive Species Advisory Committee (US). 2006. Invasive species definition clarification and guidance white paper. Available : http: //www.invasivespeciesinfo.gov/docs/council/isacdef.pdf (April 2012 ) Kato H K. Hata, H. Yamamoto, and T. Yoshioka 2006 Effectiveness of the weed risk assessment system for the Bonin Islands. Pages 65 72 in Koike F M.N. Clout, M. Kawamichi, M. De Poorter, and K. Iwatsuki K. editors. Assessment and control of biological invasion risk. Shoukadoh Book Sellers, Kyoto, Japan and IUCN, Gland, Switzerland, pp 65 72 Kerezsy A S.R. Balcombe, A.H. Arthington, and S.E. Bunn. 2011. Continuous recr uitment underpins fish persistence in the arid rivers of far western Queensland, Australia. Marine and Freshwater Research 62:1178 1190. Koehn J D and R.F. MacKenzie. 2004. Priority management actions for alien freshwater fish species in Australia. New Zealand Journal of Marine and Freshwater Research 38:457 472. Kolar C S and D.M. Lodge 2002 Ecological predictions and risk assessments for alien fish in North America. Science 298:1233 123 Krivnek M and P. 2006 Predicting invasions by woody species in a temperate zone: a test of three risk assessment schemes in the Czech Republic (Central Europe). Diversi ty and Distributions 12:319 327
95 Lawson L L J.E. Hil l L. Vi lizzi, S. Hardin, and G.H. Copp 2013 Revisions of the Fish Invasive ness Screening Kit (FISK) for its application in warmer climatic zones, with particular reference to pen insular Florida. Risk Analysis 33(8):1414 1431 Liu X M.E. McGarrity, and Y. Li. 2012. The influence of traditional Buddhist wildlife release on bio logical invasions. Conservation Letters. DOI:10.1111/j.1755 263X.2011.00215.x. Lodge D M S. Williams, H.J. MacIsaac, K.R. Hayes, B. Leung S. Reichard, R.N. Mack P.B. Moyle, M. Smith, D.A. Andow, J.T. Carlton, and A. McMichael 2006 Biological invas ions: Recommendations for U.S. policy and management. Ecological Applications 16(6):2035 2054 Marchetti M P P.B. Moyle, and R. Levine 2004 Invasive species profiling? Exploring the characteristics of non native fishes across invasion stages in Calif ornia. Freshwater Biology 49:646 661 Aquatic Invasions 5:103 114. Mastitsky S E A.Y. Karatayev, L.E. Burlakova, and B.V. Adamovich 2010 Non native fishes of Belarus: diversity, distribution, and risk classification using the Fish Invasiveness Scree ning Kit (FISK). Aquatic Invasions 5:103 114. Mendoza, R.A., V. A. Segovia, and L.B. Gutierrez. 2011. Anlisis de riesgo de species exticas de peces ornamentals dulceacucolas regularmente importados en Mxico. Anlisis de Riesgo de Especis de Peces Ext icos Potencialmente Invasores Introducidos por la Va del Acuarismo en Mxico Moyle P B and T. Light. 1996. Biological invasions of fresh water: empirical rules and assembly theory. Biological Conservation 78:149 161. Muir W M and R.D. Howard 199 9. Possible ecological risks of transgenic organism release when transgenes affect mating success: sexual selection and the Trojan gene hypothesis. Proceedings of the National Academy of Sciences 96(24):13853 13856. Murphy B F and B. Timbal. 2008. A r eview of recent climate variability and climate change in southeastern Australia. International Journal of Climatology 7:859 879. Nico L G 2005 Changes in the fish fauna of the Kissimmee River Basin, peninsular Fl orida: non native additions. Pages 523 556 in Rinne JN, Hughes RM, Calamusso B, editors. Historical changes in large river fish assemblages of the Americas. American Fisheries So ciety Symposium 45. Bethesda, Maryland Nico L G and P.L. Fuller. 1999. Spatial and temporal patterns of nonindig enous fish introductions in the United States. Fisheries 24:16 27.
96 Nico L G P. Sharp, and T.M. Collins. 2011. Imported Asian swamp eels (Synbranchidae: Monopterus ) in North American live food markets: potential vectors of non native parasites. Aquatic Invasions 6(1):69 76. NISC (Nati onal Invasive Species Council). 2008. 2008 2012 National Invasive Species Management Plan. http://www.invasivespecies.gov/home (March 2014) NRC ( National Research Co uncil ). 2002 Predicting invasions of nonindigenous plants and plant pests. Washington DC. National Academy p. 194 Onikura N J. Makajima I. Ryutei, H. Mizutani, M. Kobayakawa S. Fukuda, and T. Mukai 2011. Evaluating the potential for invasion by al ien freshwater fishes in northern Kyushu Island, Japan, using the Fish Invasiveness Scoring Kit. Ichthyological Research 58:382 387 Orr R 2003. Generic nonindigenous aquatic organisms risk analysis review process. Pages 415 438 in Ruiz, G.M., and J.T. Carlton, editors. Invasive species: vectors and management strategies. Island Press, Washington DC Parrott D S. Roy, R. Baker, J. Mumford, and 13 more authors 2009. Horizon scanning for new invasive non native species in England. Research Report Na tural England, Peterborough. Peel M C B.L. Finlayson, T.A. McMahon. 2007 Updated world map of Kppen Geiger climate classification. Hydrology and Earth System Sciences Discussions 4(2):1633 1644 Pheloung P C P.A Williams, and S.R. Halloy 1999 A weed risk assessment model for use as a biosecurity tool evaluating plant introductions. Journal of Environmental Management 57:239 251 Pine III W E T.J. Kwak, D.S. Water, and J.A. Rice JA 2005 Diet selectivity of introduced flathead catfish in coa stal rivers. Transactions of the American Fisheries Society 134(4):901 909 Puntila R L. Vilizzi, M. Lehtiniemi, and G.H. Copp 2013 First application of FISK, the freshwater Fish Invasiveness Screening Kit, in Northern Europe: example of Southern Fin land. Risk Analysis 33(8):1397 1403 QDPIF ( Queensland Department of Primary Industries and Fisheries ) 2012. Queensland freshwater fish, Barcoo grunter. Available: http://www.dpi.qld.gov.au/28_14677.h tm (April 2012 ) R Development Core Team. 2008 An introduction to R: notes on R, A programming environment for Data Analysis and Graphics. R Core Team
97 Reid A L F. Seebache r and A.J.W. Ward. 2010. Learning to hunt: the role of experienc e in predat or success. Behaviour 147:223 233. Reichard S H 2001. The search for patterns that enable invasion prediction. Pages 10 19 in Groves F D. Panetta, and J.G. Virtue editors. Weed risk a ssessment, CSIRO Publishing, Australia. Ricciardi A and J.B. Ra smussen 1998 Predicting the identity and impact of future biological invaders: a priority for aquatic resource management. Canadian Journal of Fisheries and Aquatic Sciences 55:1759 1765 Richardson D M P. M. Rejmnek, M.G. Barbour, F.D. Panet ta, and C.J. West. 2000. Naturalization and invasion of alien plants: concepts and definitions. Diversity and Distributions 6:93 107. Rowland S J 1984 The hormone induced spawning of silver perch, Bidyanus bidyanus (Mitchell) (Teraponidae). Aquacultur e 42:83 86. Ruesink J L 2005. Global analysis of factors affecting the outcome of freshwater fish introductions. Conservation Biology 19(6):1883 1893 Severinghaus L L and L. Chi. Prayer animal release in Taiwan. Biological Conservation 89(3):301 30 4. Shafland P L and J.M. Pestrak JM. 1982. Lower lethal temperatures for fourteen non native fishes in Florida. Environmental Biology of Fishes 7(2):149 156. Shafland P L K.B. Gestring, and M.S. Stanford. 2008. Categorizing introduced fishes coll ected from public waters. Southeastern Naturalist 7(4):627 636. Shafland P L K.B. Gestring and M.S. Stanford 2008 fishes 2007. Florida Scientist 71:220 245 Simberloff D 2005 The politics of assessing risk for biologi cal invasions: the USA as a case study. Trends in Ecology and Evolution 20(5):216 222 Simberloff D I.M. Parker, and P.N. Windle 2005 Introduced species policy, management, and future research needs. Frontiers in Ecology and the Environment 3(1):12 2 0 P A. M. Vassilev, A. Apostolou, D. Mrdak, M. Ristovska, V. Kostov, V. D. Skraba, L. Vilizzi, and G.H. Copp. 2013 Risk identification of non native freshwater fishes in four countries of the Balkans Region using FISK Mediterra nean Marine Science 14(2):369 376
98 Stohlgren T J and J.L. Schnase 2006 Risk analysis for biological hazards: what we need to know about invasive species. Risk Analysis 26:163 173 Tarkan A S F.G. Ekmekci, L. Vilizzi, and G.H. Copp 2013 Risk scr eening of non native freshwater fishes at the frontier between Asia and Europe: first application in Turkey of the Fish Invasiveness Screening Kit (FISK). Journal of Applied Ichthyology 1:1 7 Thomas M E. 1993. Monitoring the effects of introduced flathe ad catfish on sport fish populations in the Altamaha River, Georgia. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 47:531 538. Thompson K A J.E. Hill, and L.G. Nico. 2012. Eastern mosquitofish resist s invasion by nonindigenous poeciliids through agonistic b ehaviors. Biological Invasions DOI:10.1007/s10530 012 0176 2. Tricarico C L. Vilizzi, F. Gherardi, and G.H. Copp 2010 Calibration of FI ISK, an invasive screening tool for non native freshwate r invertebrates. Risk Analysis 30(2):285 292 Troca D F A and J.P. Vieira. 2012. Potential invasive non native fish farmed in the coastal region of Rio Grande Do Sul, Brazil. Boletim do Insti tuto de Pesca, Sao Paulo 38:109 120. Unmack P J 2001 Biog eography of Australian freshwater fishes. Journal of Biogeography 28(9):1053 1089. USGS (United States Geological Service). 2014. Nonindigenous aquatic species database. U.S. Geological Survey, Gainesville, Florida. Available : http://nas.er.usgs.gov/ ( February 2014 ). Vandenbergh K. 2007. Risicoanalyse voor uitheemse vissorten in Vlaanderen. PhD Thesis, Leuven, Belgium: Departement Biologie, Faculteit Wetenschappen, Kath olieke Universiteit Leuven. Vebrugge L H G. van der Velde, A.J. Hendriks, H. Verreycken, and R.S.E.W. Leuven. 2012. Risk classifications of aquatic non native species: application of contemporary European assessment protocols in different biogeographical se ttings. Aquatic Invasions 7(1):49 58. Verbrugge L H R.S. Leuven, G. Van Der Velde 2010 Evaluation of international risk assessment protocols for exotic species. Final Report, Department of Environmental Science, Faculty of Science, Institu t e for Water and Wetland Research, Radboud Univers ity Nijmegen, Heyendaalseweg
99 Vilizzi L V and G.H. Copp. 2013. Application of FISK, an invasiveness screening tool for non native freshwater fishes, in the Murray Darling Basin (Southeastern Australia). Risk Analysis 33(8):1432 1440 Vitule J R S C.A. Freire and D. Simberloff 2009. Introduction of non native freshwater fish can certainly be bad. Fish and Fisheries 10:98 108 Walkenbach J. 2007. Excel 2007 bible. John Wiley and Sons, Inc. New York, New York. Weber J F.D. Panetta, J. Virtue P. Pheloung 2009 An analysis of assessment assessment system. Journal of Environmental Management 90:798 807 Youden W J 1950. Index for rating diagnostic tests. Cancer 3:32 35 Zajicek P W S. Hardin, and C. Watson 2009a A Florida marine ornamental pathway risk analysis. Reviews in Fisheries Science 17(2):156 169 Zajicek P W T. Weier, S. Hardin, J.R. Cassan, and V. Mudrak 2009b. A triploid grass carp risk analysis spe cific to Florida. Journal of Aquatic Plant Management 47:15 20 Zou K H J. and L. Mauri 2007 Receiver operating characteristic analysis for evaluating diagnostic tests and predictive models. Circulation Journal of the American Heart Associa tion 115:654 657
100 BIOGRAPHICAL SKETCH Larry was born in St. Charles, Missouri where he grew up and attended high school at Francis Howell North His younger years were spent fishing at local ponds and hunting at his family owned farm in Northeastern Miss ouri. His interest in all things fisheries and wildlife grew with his time spent in the outdoors. Upon graduating high school, he turn ed this fondness into a career and began attending the University of Missouri Columbia in August of 2004. While pursuing a Bachelor of Science degree in Fisheries and Wildlife Management from the University of Missouri Larry had the opportunity to experience many different aspects of this career track. Beginning in the summer of 2005, he worked as a field technician samplin g catfish in the tributaries of the Missouri River. This work was followed the next year on the main stem Missouri River, where his interest in fisheries was strengthened. By this time, he chose to emphasize specifically in the fisheries field and graduate d with a degree in 2008 After graduating, he spent several years working for the Missouri Department of Conservation as a fisheries technician. The opportunity to attend the University of Florida for a Master of Science degree in Fisheries and Aquatic Sciences arose in the Summer of 2010. His experiences at UF le d him to focus on aquaculture as a career using his fisheries science background. Currently, he is happily employed by the