Abstract
Invasive species remain a major environmental problem in the world's oceans. Managing the vectors of introduction is the most effective means of mitigating this problem, but current risk assessments and management strategies are largely focused on species, not on vectors and certainly not on multiple simultaneous vectors. To highlight the issue that multiple vectors contribute to invasions, we analyzed the historical and contemporary contributions of eight maritime vectors to the establishment of nonindigenous species in California, where most species were associated with two to six vectors. Vessel biofouling looms larger than ballast water as a major vector and a management opportunity, but aquaculture risk appears reduced from historic levels. Standardized data on species abundances in each vector are lacking for a robust cross-vector assessment, which could be obtained in a proof-of-concept “vector blitz.” Management must shift away from one or two target vectors to coordination across multiple vectors.
Nonindigenous species (NIS) are widespread throughout Earth's oceans and coasts, where they cause environmental impacts and economic damages ( Carlton 1999 ). They have been associated with declines in marine populations ( Kappel 2005 ), alteration of food webs ( Nichols et al. 1990 , Oguz et al. 2008 ), habitat modifications that affect community structure and function ( Neira et al. 2006 , Sousa et al. 2009 ), and the delivery of toxic microorganisms of concern for sea life and human health ( Ruiz et al. 2000 ).
Multiple vectors operating in an increasingly connected world
Increasing global trade, novel trade routes, climate change, habitat modification, fisheries, and invasions themselves can combine to create increasing opportunities for introductions of marine NIS. Global trade increased dramatically in the latter half of the twentieth century, driven by human population growth, changes in policies, and increased efficiency in shipping ( Hulme 2009 ). Previously remote locales, such as Antarctica, where many marine species are endemic, have come under increasing risk from NIS introductions ( Smith et al. 2012 ). At the opposite pole, the Northwest Passage has been sufficiently ice free for navigation since the summer of 2007, which has shortened the path between the Atlantic and Pacific oceans and has increased opportunities for human-mediated species introductions ( Niimi 2004 ). As oceans warm, NIS can gain a foothold over native marine species ( Sorte et al. 2010 ). New conservation strategies, such as assisted colonization or managed relocation, further promote transfers of species beyond their native ranges ( Schwartz et al. 2012 ).
One by one: Single species, single-vector risk assessments
In this increasingly connected world, the most effective means to reduce the future ecological and economic costs of NIS is to prevent their introductions by managing the many vectors that deliver them, rather than by focusing efforts on the management of individual species or even individual vectors ( Ruiz and Carlton 2003 , Reaser et al. 2008 ). Despite this widely accepted recommendation, in most research and management, the problem of biological invasions has been addressed on a species-by-species and vector-by-vector basis.
Species-specific approaches have included black listing on the basis of evidence that a species could be or is an injurious pest. This approach is usually based on previously identified bad actors, and it is laborious, being impractical for the large number of species that circulate in multiple vectors and especially those without an invasion history ( Simberloff 2006 , Reaser et al. 2008 ). A formal risk assessment, involving the evaluation of the probability that a species will be introduced, establish, and cause ecological or economic harm, requires both more data and more effort than is available for most species.
Various types of NIS risk assessments are being considered or applied by management authorities ( Kolar 2004 , Gordon et al. 2012 ). Some assessments favor matching environmental conditions of native and nonnative ranges to predict risk, some take a spatially explicit landscape approach to identify sites that are vulnerable to invasions, some include management actions, and others are focused on propagule pressure or how the recipient community can shape the success or failure of NIS following an introduction. NIS risk assessment makes economic sense ( Keller et al. 2007 , Springborn et al. 2011 ) and has made substantial progress over the past decade, but it is still based primarily on the risk posed by specific species or species characteristics related to invasiveness ( Hayes 2002 , Hayes and Sliwa 2003 , Orr 2003 , Keller et al. 2007 , Campbell 2009 ; see Leung et al. 2012 for a theoretical framework).
In this article, we highlight the challenge that multiple vectors pose in the evaluation and management of invasions, using marine and estuarine (hereafter, marine ) NIS as a model. Maritime vectors are diverse. The major vectors include ballast water, which is already the subject of management to varying degrees ( Miller et al. 2011 ); the biofouling of large and small vessels by a community of sessile and associated mobile organisms that colonize and grow on any wetted surface of a vessel, such as hulls, anchors, storage lockers, and other colonizable locations ( Mineur et al. 2008 , Davidson et al. 2009 , Wanless et al. 2010 ); aquaculture ( Naylor et al. 2001 ); live seafood ( Chapman et al. 2003 ); live bait ( Kilian et al. 2012 ); the ornamental species trade ( Rhyne et al. 2012 ); and marine debris ( Barnes 2002 ). Maritime vectors are also dynamic, with both gradual and punctuated changes—the latter demonstrated by biofouled debris from the 2011 Japanese earthquake and tsunami that washed ashore in North America. The tsunami debris instigated a rapid management response ( RPRW 2012 ), which is crucial to preventing introductions, but also highlighted the ad hoc manner in which limited management and scientific resources are allocated among multiple vectors when no decision has been made about vector prioritization. Our analysis of maritime vectors helps explain why it is difficult to prioritize vectors.
The plethora of existing vectors and temporal changes in operation is not generally recognized, except by the scientists who study them. It is safe to say that the broader public; the media; and the environmental, management, and political communities are largely unaware of how little is known about these vectors. For example, the education campaigns that raised public awareness about ballast water as a possible vector for invasive species contrasts sharply with the lack of awareness about cultured nonnative oysters that have become feral and that can dramatically change marine environments ( Ruesink et al. 2005 ). In a different example, the public is not likely to know that over 100,000 individuals of the quagga mussel ( Dreissena rostriformis bugensis ), over 100 sharks, and transgenic tilapia and salmon—all of which are restricted species in California—were permitted for importation in 2009 for research; exhibition, including at public zoos and aquariums; or aquaculture. Only recently have research facilities and public zoos and aquariums taken steps to prevent the release of NIS. The general ignorance of vector multiplicity is evident in regulatory frameworks based on single species or vectors. For instance, the Lacey Act, administered by the US Fish and Wildlife Service (USFWS), regulates injurious pests by individual species, and the International Maritime Organization is advancing ballast water management. The value of these frameworks cannot be discounted, but they do not address the issue of managing multiple vectors.
How close are we to reliably comparing the relative risk of different maritime vectors or their cumulative risk? Which among the many operating vectors offer management opportunities or merit prioritization, given the limited available resources? The relative importance of maritime vectors has been assessed most often by tallying the number of species attributed to vectors, and these vectors are often defined in different ways ( Bax et al. 2003 , Ruiz et al. 2011 ). This approach is also applied to freshwater and terrestrial NIS vectors (e.g., Keller et al. 2009 ). Although such studies have advanced the management of marine invasive species, they are largely retrospective. A few quantitative risk assessments for single marine species have been accomplished (e.g., Herborg et al. 2007 , 2009 , Therriault and Herborg 2008 ). In a few other studies, a vector-based risk assessment has been undertaken by obtaining the abundances of NIS in invasive pathways ( Hayes 2002 , Hayes and Sliwa 2003 , Hayes and Barry 2008 , Acosta and Forrest 2009 ). Hayes (2002) used shipping records and species distributions to estimate invasion potential; added a Web-based questionnaire to assess the economic, ecological, and health impacts for potential marine invaders in New Zealand; and estimated uncertainty using interval arithmetic. Single-vector management is certainly an improvement over single-species management, but it might not substantially reduce invasions if many other vectors are in place. To our knowledge, there has been no attempt to quantify the relative or combined risk that multiple maritime vectors pose, beyond vector-attribution tallies, or to evaluate their past, present, and future changes.
To illustrate both the need for and the challenge of assessing the risk posed by multiple vectors, we compared available data for eight maritime vectors operating in California: commercial vessel ballast water; commercial vessel biofouling; recreational vessel biofouling; fishing vessel biofouling; aquaculture; and trade in ornamental species, live seafood, and live bait. California provides a good model to illustrate the state of knowledge of multiple vectors for several reasons. California is a major introduction point for marine NIS and serves as a nexus for their establishment and spread. Furthermore, the state supports a relatively high level of activity across diverse vectors ( Ruiz et al. 2011 ). Its large marine economy ( Kildow and Colgan 2005 ) has already been taxed by marine invasive species ( Anderson 2005 , Fernandez 2008 ), which helped spur the state government's interest in a multiple-vector management approach. The vectors considered here operate globally; therefore, their comparison should provide a useful generalized illustration of the issue of multiple vectors.
Multiple vectors characterize California's historical marine invasions
Data often limit quantitative NIS risk assessment ( Gertzen and Leung 2011 , Leung et al. 2012 ). The most readily available data are NIS lists that provide information on the species, introduced locales, and dates of first record (e.g., the National Exotic Marine and Estuarine Species Information System, NEMESIS, http://invasions.si.edu/nemesis/index.html ; the National Introduced Marine Pest Information System, NIMPIS, www.marinepests.gov.au/nimpis ). We analyzed California's marine NIS invasion history on the basis of a subset of data from NEMESIS for all nonfreshwater, nonnative invertebrates (except insects) and algae known to have become established in California between 1853 and 2011. These taxa include the vast majority of marine NIS in California. We updated the work of Ruiz and colleagues (2011) on vector attribution to individual species and expanded it to include multiple introduction events by including the dates and specific bays of introduction within the state, allowing finer parsing of vector attribution. We identified that at least 235 NIS have become established in California's marine and estuarine waters (see supplemental table S1, available online at http://dx.doi.org/10.1525/bio.2013.63.12.8 ), including two species that were subsequently eradicated (the seaweed Caulerpa taxifolia and the polychaete worm Terebrasabella heterouncinata ).
We then addressed vector strength (the number of invasions attributed to a vector; Ruiz and Carlton 2003 ) in several ways. On the basis of life history traits, timing, and the location of NIS detections and vector operations (see Ruiz et al. 2011 for description), 90 of the 235 species were classified as introduced to California almost certainly by only a single vector ( figure 1a , table S1). The single-vector species provided the most conservative estimate of the historical strength of each vector. For these single-vector species, although ballast water was a major vector, non-ballast-water vectors together accounted for more than twice the number of established NIS than ballast water alone. Furthermore, vessel biofouling was identified as the vector for as many established NIS as ballast water and aquaculture combined.
Vector attributions as a function of taxa. (a) Ninety marine nonindigenous species (NIS) established (including two subsequently eradicated) in California were attributed solely to a single vector. The Protozoans include Ciliophora. (b) A total of 145 NIS in California were associated with two or more vectors. The graphed total exceeds 145 species because each species was assigned to all of the vectors considered possible transfer mechanisms at each location (n = 389 events). (c) For multiple-vector species, as in panel (b), vector attribution was apportioned to each species by the number of unique points at which each vector was associated with an introduction. Taxa with fewer than 10 species are included in the Other category. Unascribed indicates that a vector was not assigned to a species in the National Exotic Marine and Estuarine Species Information System database (see supplemental table S1, available online at http://dx.doi.org/10.1525/bio.2013.63.12.8 ).
Vector attributions as a function of taxa. (a) Ninety marine nonindigenous species (NIS) established (including two subsequently eradicated) in California were attributed solely to a single vector. The Protozoans include Ciliophora. (b) A total of 145 NIS in California were associated with two or more vectors. The graphed total exceeds 145 species because each species was assigned to all of the vectors considered possible transfer mechanisms at each location (n = 389 events). (c) For multiple-vector species, as in panel (b), vector attribution was apportioned to each species by the number of unique points at which each vector was associated with an introduction. Taxa with fewer than 10 species are included in the Other category. Unascribed indicates that a vector was not assigned to a species in the National Exotic Marine and Estuarine Species Information System database (see supplemental table S1, available online at http://dx.doi.org/10.1525/bio.2013.63.12.8 ).
The other 145 NIS (61% of the total) were attributed to between two and six possible vectors. For these, we summarized vector strength in two ways. First, for each species, we summed the number of establishment events (the nontransient presence of a species in a bay) attributed to each possible vector. For example, if a multiple-vector species was known from nine different bays and ballast was a sole or possible vector for three of these bays, three invasion events were attributed to ballast for that species. This approach identifies the maximum number of establishment events per vector. Second, we weighted vectors proportionately by the number of events attributed to each for each individual species. When more than one vector was plausible for the presence of a species in a bay, it was treated as multiple events, each by a single vector. For example, if a species was known from 10 bays, including three of those establishment events associated with ballast alone, two with ballast and vessel biofouling together, and five with aquaculture only, ballast was weighted as 0.42 (5/12). Therefore, each species was weighted equally (as 1.0), regardless of the number of invasion events, and each plausible vector for each introduction event has an equal probability of its proportional contribution.
For these multiple-vector species, our unweighted estimate of possible vector importance indicated that ballast and vessel biofouling were the most common vectors involved ( figure 1b ). Of the 77 species introduced by two vectors, 49% were attributed to ballast water and vessel biofouling, 26% to aquaculture and vessel biofouling, and 8% to ballast water and aquaculture. Our weighted estimate highlights patterns that are not apparent from the previous, more traditional, unweighted analysis, alone. Although the unweighted analysis demonstrates the magnitude of possible vector importance (for all invasion events), the weighted analysis is informative for generating hypotheses about probable vector importance on a per species basis. For example, the unweighted estimate indicates that aquaculture was a potent possible vector, but the weighted estimate deemphasized its vector strength relative to ballast water and vessel biofouling ( figure 1c ). To a lesser degree, the same can be said of ballast water, whereby its strength is reduced in the weighted estimate relative to vessel biofouling.
Although the historical data provided some new insights into the relative contributions of vectors introducing marine NIS to California, there are some obvious limitations for a robust comparison of vectors and their relative risk. First, the numbers of species on lists of introduced species such as NEMESIS and NIMPIS are undoubtedly underestimated ( Carlton 2009 ). Second, a more serious limitation is that historical data do not necessarily match the species circulating in vectors today and rarely provide information on species abundances. Third, the nature of individual vectors, themselves, can change over time, both in magnitude and the per capita transfer of organisms. We therefore considered contemporary fluxes of marine NIS to California.
Contemporary fluxes of marine NIS to California
Information on the abundance of organisms circulating in vectors is crucial, because abundance is highly correlated with the probability of an introduction ( Ruiz et al. 2000 , Colautti et al. 2006 , Hayes and Barry 2008 ). For risk assessment, ideally, both the number of individuals released in a single event (the propagule size ) and the number of discrete release events (the propagule number ) would be estimated for each species ( Colautti et al. 2006 ). Although ballast water discharge events are recorded in the United States, data on release rates for biofoulers, marine ornamentals, seafood, and bait are not available ( Weigle et al. 2005 ). Therefore, we focused on propagule flux as the numbers of species and individuals circulating in a vector per unit of time (standardized to 1 year).
To investigate propagule flux, we mined published data and federal and state records and conducted field observations (vessel biofouling, air cargo). Flux estimates for recreational vessel biofouling were obtained from last port of call (LPOC) records of the US Customs and Border Protection agency (CBP) for small-vessel arrivals to California in 2009 and from sampling the hulls of 49 transient vessels between 2010 and 2011 for species identification and abundances. The flux of NIS through commercial fishing biofouling was estimated from the number of arrivals to California in 2008, provided by the Pacific Fisheries Information Network (PacFIN, www.psmfc.org/program/pacific-fisheries-information-network-pacfin ). Flux estimates for commercial vessel biofouling and ballast water discharge were derived from hull surveys of 23 vessels between 2009 and 2011 and from the National Ballast Information Clearinghouse's (NBIC, http://invasions.si.edu/nbic/search.html ) arrival and discharge data and by using the midpoint of zooplankton concentrations in ballast arriving to the United States ( Minton et al. 2005 ). Ornamental species fluxes were estimated from the USFWS's Law Enforcement Management Information System (LEMIS) records for live marine fish and invertebrate (excluding scleractinian corals) importations into San Francisco and Los Angeles in 2009 (the most recent year for which there are complete records) and our observations of air cargo inspections in 2012 and from California Department of Fish and Game (CDFG) permits for restricted species from 1988 through 4 August 2011. We assessed the flux in aquaculture, starting with the CDFG records, which led to other sources (see below for details).
No common currency to compare propagule flux across vectors
At present, no single source of NIS information exists for any of these vectors. Also lacking is a common currency to estimate the flux (i.e., the number of individuals and their identities) circulating annually in each vector, except as order-of-magnitude bounds ( figure 2 ). We first identified a specific unit of NIS delivery for biofouling (vessels arriving), ballast water (vessels discharging), and ornamental species (shipments), but aquaculture import permits did not yield a useful unit of delivery ( figure 2a ). Next, the quantity of organisms associated with each unit of delivery was estimated for recreational and commercial vessels and the ornamental vector. Good records exist for the volume of ballast water discharged, but the number of organisms in ballast water had to be extrapolated from another study ( Minton et al. 2005 ). Vessel biofouling has been measured variously as the number of macroinvertebrate species in the biofouling community on a vessel, the percentage cover of biofouling in areas sampled on vessels, and the biomass of fouling per unit of area ( Davidson et al. 2009 ). LEMIS import permit records provide more indirect but more readily available data for the ornamental species than for biofouling. Although the quantities of ornamental species in the LEMIS records are not always standardized to individual organisms and although the records do not account for subsequent transfers out of the state, they are probably close to actual quantities. Quantities could not be estimated for fishing vessels or aquaculture ( figure 2b ), because sampling access has been restricted to extremely few vessels to date, and the numbers of organisms listed on aquaculture permits lacked sufficient standardization and specificity to derive quantities.
![Annual nonindigenous species (NIS) flux across vectors. These data illustrate the lack of a standardized metric to compare propagule flux across vectors: biofouling (from recreational, fishing, and commercial vessels), the ornamental species trade, aquaculture, and ballast water discharge, represented as (a) the number of arrivals of units that could be managed for each vector (where they were available for biofouling, both the number of unique vessels [the solid bar] and the estimated number of arrivals [the hatched bar] are shown), (b) the propagule flux associated with each arrival unit in panel (a) (the kite diagrams illustrate the distribution of frequencies), and (c) the likelihood of NIS associated with each unit of introduction. Note the log scales in panels (a) and (b). The date range or the number of vessels sampled is indicated at the top of the graphs. The fill indicates certainty; for example, we are confident of the number of ships discharging ballast into California waters (black), but accurate information concerning aquaculture permits in the state was not available (gray). The white fill in panel (b) indicates that the ballast water data are from non-California ports. No data were available for the number of individuals delivered by aquaculture or fishing vessels or the proportion of NIS in ballast water or fishing vessel fouling communities. See tables 1 and 2 for a summary of the vessel data.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/bioscience/63/12/10.1525_bio.2013.63.12.8/1/m_bio.2013.63.12.8-f02.jpeg?Expires=1591684299&Signature=NWUNIgwCRtkJ8rAroe0mq0uv3YzbzNU8f1~s-zpvuwfpv95zG5G9qhiZDciWFErJVIktsQWtdywC76LMLIF6mkNHAmYCLYb0uutJ4XSqRsLthhhGK1wRH130mBCXX0ivKnZUCIbQ1c6IdXdOgLs38tGibqEkhxds~50cYPculr1WX--r0eVJ2d79UCMD9IYO0hn5unssfJ965r44-APcOXH2Bn9JKl0U5CzATVN37gQ2Mu4zChPkHZiLXB7oySpDy339QV5TaPdXwNBadUs3Raa2seA29X8070D6BUnTWgkNgX9-y1icQjdWxlYqtmxgpJ0mUCCa30G9elud6LHAMw__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Annual nonindigenous species (NIS) flux across vectors. These data illustrate the lack of a standardized metric to compare propagule flux across vectors: biofouling (from recreational, fishing, and commercial vessels), the ornamental species trade, aquaculture, and ballast water discharge, represented as (a) the number of arrivals of units that could be managed for each vector (where they were available for biofouling, both the number of unique vessels [the solid bar] and the estimated number of arrivals [the hatched bar] are shown), (b) the propagule flux associated with each arrival unit in panel (a) (the kite diagrams illustrate the distribution of frequencies), and (c) the likelihood of NIS associated with each unit of introduction. Note the log scales in panels (a) and (b). The date range or the number of vessels sampled is indicated at the top of the graphs. The fill indicates certainty; for example, we are confident of the number of ships discharging ballast into California waters (black), but accurate information concerning aquaculture permits in the state was not available (gray). The white fill in panel (b) indicates that the ballast water data are from non-California ports. No data were available for the number of individuals delivered by aquaculture or fishing vessels or the proportion of NIS in ballast water or fishing vessel fouling communities. See tables 1 and 2 for a summary of the vessel data.
Annual nonindigenous species (NIS) flux across vectors. These data illustrate the lack of a standardized metric to compare propagule flux across vectors: biofouling (from recreational, fishing, and commercial vessels), the ornamental species trade, aquaculture, and ballast water discharge, represented as (a) the number of arrivals of units that could be managed for each vector (where they were available for biofouling, both the number of unique vessels [the solid bar] and the estimated number of arrivals [the hatched bar] are shown), (b) the propagule flux associated with each arrival unit in panel (a) (the kite diagrams illustrate the distribution of frequencies), and (c) the likelihood of NIS associated with each unit of introduction. Note the log scales in panels (a) and (b). The date range or the number of vessels sampled is indicated at the top of the graphs. The fill indicates certainty; for example, we are confident of the number of ships discharging ballast into California waters (black), but accurate information concerning aquaculture permits in the state was not available (gray). The white fill in panel (b) indicates that the ballast water data are from non-California ports. No data were available for the number of individuals delivered by aquaculture or fishing vessels or the proportion of NIS in ballast water or fishing vessel fouling communities. See tables 1 and 2 for a summary of the vessel data.
Moreover, comparisons of measures of species richness are also problematic, having been made using different approaches, depending on the vector ( figure 3c ). Taxonomic identification is best for the aquaculture vector in California, because permitted organisms are identified to the species level without exception. Species composition is problematic in live-organism trade ( Rhyne et al. 2012 , Smith et al. 2012 ) but still far better than it is for most ballast water discharge, for which species composition is unknown. USFWS importation records concern only animals, the majority of which fall into generic taxonomic categories (e.g., “marine tropical fishes,” “crustaceans”). Plants fall under the jurisdiction of the US Department of Agriculture, for which the same issues apply regarding labeling and transfers on arrival into the country. Marine plants, in general, are mostly unidentified when they are observed and are underrepresented in vector sampling.
The process for importing animals and marine plants into (or for moving them within) California (CA) for aquaculture purposes. Dashed lines indicate import regulations for plants. Plants and animals imported into the United States must be cleared by the US Fish and Wildlife Service (USFWS) or the US Department of Agriculture (USDA). The California Department of Food and Agriculture (CDFA) maintains a noxious weeds list of species prohibited for importation. There are no restrictions on moving aquaculture species within the state. Marine animals and plants intended for placement in state waters must also be permitted through the California Department of Fish and Game (CDFG) if they are moved from out of state. CDFG private stocking permits are required for organisms used for noncommercial purposes. Commercial facilities must file a management plan with the California Department of Health. The CDFG manages aquaculture leases; lessees must file an annual proof-of-use report. The US Army Corps of Engineers requires permits for structures placed in the water. Despite the numerous permits required in some cases, complete data on the source, species, and number of individuals actually placed in the water are not collected.
The process for importing animals and marine plants into (or for moving them within) California (CA) for aquaculture purposes. Dashed lines indicate import regulations for plants. Plants and animals imported into the United States must be cleared by the US Fish and Wildlife Service (USFWS) or the US Department of Agriculture (USDA). The California Department of Food and Agriculture (CDFA) maintains a noxious weeds list of species prohibited for importation. There are no restrictions on moving aquaculture species within the state. Marine animals and plants intended for placement in state waters must also be permitted through the California Department of Fish and Game (CDFG) if they are moved from out of state. CDFG private stocking permits are required for organisms used for noncommercial purposes. Commercial facilities must file a management plan with the California Department of Health. The CDFG manages aquaculture leases; lessees must file an annual proof-of-use report. The US Army Corps of Engineers requires permits for structures placed in the water. Despite the numerous permits required in some cases, complete data on the source, species, and number of individuals actually placed in the water are not collected.
Small recreational and fishing vessels illustrate the fragmented nature of the data ( tables 1 and 2 ). There were 1182 recreational vessels of foreign origin that entered California and registered with the CBP in 2009. This number is the same order of magnitude as that for commercial vessels (1822 reported in the NBIC), but the actual arrivals of recreational vessels in California and their intrastate movements are unknown but certainly higher in number than those of foreign vessels that reported to the CBP. CBP records include only the initial arrivals in California of vessels with an LPOC outside of the United States. Such boats might travel to additional ports in which further registration is not required, yet NIS might be transferred there. Similarly, vessel arrivals from other US states—either US vessels or foreign vessels that reported to the CBP on their first arrival in a different state—are not captured by the CBP's California records. We also conducted surveys of certain marinas in the state, which revealed that not all foreign arrivals were captured in the records. Fishing vessel data came from a different source; were restricted to where the catch was landed (and excluded other movements during which fish were not landed); and included only vessels reporting to Washington, Oregon, and California. If a fishing vessel travels outside of these states, noncomparable CBP records should capture foreign voyages, but there is no mechanism to capture voyages to other US states.
Vessel statistics used in figure 2 .
| Recreational | Fishing | Commercial | |
|---|---|---|---|
| Number of vessels | 1182 | 134 | 1822 |
| Number of NIS arrivals in California | Unknown | Unknown | 6002 |
| Origin | Foreign | Oregon, Washington | Outside California |
| Date | 2009 | 2008 | 2010 |
| Data source | CBP | PacFIN | NBIC |
| Recreational | Fishing | Commercial | |
|---|---|---|---|
| Number of vessels | 1182 | 134 | 1822 |
| Number of NIS arrivals in California | Unknown | Unknown | 6002 |
| Origin | Foreign | Oregon, Washington | Outside California |
| Date | 2009 | 2008 | 2010 |
| Data source | CBP | PacFIN | NBIC |
Note : No data were available for arrivals of fishing vessels from regions other than Oregon and Washington, although arrivals probably occur. Abbreviations: CBP, US Customs and Border Protection agency; PacFIN, Pacific Fisheries Information Network; NBIC, National Ballast Information Clearinghouse; NIS, nonindigenous species.
Vessel statistics used in figure 2 .
| Recreational | Fishing | Commercial | |
|---|---|---|---|
| Number of vessels | 1182 | 134 | 1822 |
| Number of NIS arrivals in California | Unknown | Unknown | 6002 |
| Origin | Foreign | Oregon, Washington | Outside California |
| Date | 2009 | 2008 | 2010 |
| Data source | CBP | PacFIN | NBIC |
| Recreational | Fishing | Commercial | |
|---|---|---|---|
| Number of vessels | 1182 | 134 | 1822 |
| Number of NIS arrivals in California | Unknown | Unknown | 6002 |
| Origin | Foreign | Oregon, Washington | Outside California |
| Date | 2009 | 2008 | 2010 |
| Data source | CBP | PacFIN | NBIC |
Note : No data were available for arrivals of fishing vessels from regions other than Oregon and Washington, although arrivals probably occur. Abbreviations: CBP, US Customs and Border Protection agency; PacFIN, Pacific Fisheries Information Network; NBIC, National Ballast Information Clearinghouse; NIS, nonindigenous species.
Fouling organism abundance statistics used in figure 2 .
| Number of vessels | |||
|---|---|---|---|
| Abundance per vessel sampled | Recreational | Fishing | Commercial |
| 0 | 7 | Unknown | 2 |
| 1–10 | 11 | Unknown | 2 |
| 11–100 | 6 | Unknown | 0 |
| 101–1000 | 13 | Unknown | 2 |
| 1001–10,000 | 7 | Unknown | 8 |
| 10,001–100,000 | 3 | Unknown | 7 |
| More than 100,000 | 2 | Unknown | 2 |
| Total vessels sampled | 49 | 0 | 23 |
| Percentage with fouling NIS | 86 | Unknown | 68 |
| Arrivals discharging non-California- sourced water | Not applicable | Not applicable | 962 |
| Percentage with ballast NIS | Not applicable | Not applicable | Unknown |
| Number of vessels | |||
|---|---|---|---|
| Abundance per vessel sampled | Recreational | Fishing | Commercial |
| 0 | 7 | Unknown | 2 |
| 1–10 | 11 | Unknown | 2 |
| 11–100 | 6 | Unknown | 0 |
| 101–1000 | 13 | Unknown | 2 |
| 1001–10,000 | 7 | Unknown | 8 |
| 10,001–100,000 | 3 | Unknown | 7 |
| More than 100,000 | 2 | Unknown | 2 |
| Total vessels sampled | 49 | 0 | 23 |
| Percentage with fouling NIS | 86 | Unknown | 68 |
| Arrivals discharging non-California- sourced water | Not applicable | Not applicable | 962 |
| Percentage with ballast NIS | Not applicable | Not applicable | Unknown |
Fouling organism abundance statistics used in figure 2 .
| Number of vessels | |||
|---|---|---|---|
| Abundance per vessel sampled | Recreational | Fishing | Commercial |
| 0 | 7 | Unknown | 2 |
| 1–10 | 11 | Unknown | 2 |
| 11–100 | 6 | Unknown | 0 |
| 101–1000 | 13 | Unknown | 2 |
| 1001–10,000 | 7 | Unknown | 8 |
| 10,001–100,000 | 3 | Unknown | 7 |
| More than 100,000 | 2 | Unknown | 2 |
| Total vessels sampled | 49 | 0 | 23 |
| Percentage with fouling NIS | 86 | Unknown | 68 |
| Arrivals discharging non-California- sourced water | Not applicable | Not applicable | 962 |
| Percentage with ballast NIS | Not applicable | Not applicable | Unknown |
| Number of vessels | |||
|---|---|---|---|
| Abundance per vessel sampled | Recreational | Fishing | Commercial |
| 0 | 7 | Unknown | 2 |
| 1–10 | 11 | Unknown | 2 |
| 11–100 | 6 | Unknown | 0 |
| 101–1000 | 13 | Unknown | 2 |
| 1001–10,000 | 7 | Unknown | 8 |
| 10,001–100,000 | 3 | Unknown | 7 |
| More than 100,000 | 2 | Unknown | 2 |
| Total vessels sampled | 49 | 0 | 23 |
| Percentage with fouling NIS | 86 | Unknown | 68 |
| Arrivals discharging non-California- sourced water | Not applicable | Not applicable | 962 |
| Percentage with ballast NIS | Not applicable | Not applicable | Unknown |
Aquaculture provides an example of both the lack of a common currency for contemporary fluxes and the fragmented nature of the information ( figure 3 ). We initially expected that the NIS flux in aquaculture could be estimated, given that the organisms are intentionally imported or outplanted. We began assessing aquaculture NIS fluxes on the basis of CDFG importation and private stocking permits from 1950 through 2011, only to realize that the records were incomplete and highly dispersed across multiple agencies that regulate various aspects of the industry in California. Different sectors of the CDFG are responsible for import permits, private stocking permits, aquaculture inspection and planting certificates, and bottom lease records (proof-of-use reports for state-registered aquaculture facilities). Commercial aquaculture facilities, which must be registered with the CDFG, must also file a management plan with the California Department of Public Health, a step that we recognized only at the end of the data collection period. Although these plans indicate potentially farmed acreage, they do not provide information on the number of individuals outplanted or even the number of acres actually farmed. The US Army Corps of Engineers (ACE) requires permits for aquaculture businesses placing structures that change the flow of water or affect the substratum in state or federal waters. We examined ACE records held in the Los Angeles office through a Freedom of Information Act request, but the records provided no relevant information on NIS fluxes (the records held in the San Francisco office were not readily retrievable). For aquaculture species imported from other countries, the USFWS's LEMIS recorded no importations of invertebrate species for aquaculture between 2003 and 2011.
Even taken together, these numerous types of records for California marine aquaculture provided little information on NIS fluxes. Abundances were reported as cases, bushels, or thousands of seed individuals. Although California aquaculture import permits must list the intended species and exporter location, there is no postpermit requirement to report the volume or number actually imported. Aquaculture proof-of-use reports specify the number of plantings for a subset of aquaculture facilities ( figure 3 ), but listing the source is voluntary. To further complicate the issue, aquaculture leases were managed variously by the CDFG, conservation districts, cities, and a private energy-generating company. The complexity of the aquaculture regulatory framework is challenging for both the industry and the regulators and is a cogent example of the long-recognized need for an authority to oversee the international and interstate importation of live organisms ( Schmitz and Simberloff 2001 , Lodge et al. 2006 ), which would ideally also include intrastate movements of NIS.
Comparing invasions across multiple vectors: Apples and oranges
Although currently available data on NIS delivery by maritime vectors are too disparate for a rigorous cross-vector risk assessment, this assessment provided some rough comparisons across the vectors. Historical data indicate that ballast water, despite the national and international focus on it as the primary vector for marine NIS, is by no means necessarily the most important vector for established marine NIS in California ( figure 1 ). Indeed, this situation is probably true in other locations, as well ( Bax et al. 2003 ). Managing ballast water, although it is necessary and an example of a vector-based approach, is clearly insufficient to prevent new introductions, given the importance of other vectors. Although we found no common currency to allow highly quantitative comparisons of contemporary NIS presence and abundances circulating in each vector ( figure 2 ), the available data support the conclusion that other vectors must be addressed, in addition to ballast water. Flux data support ornamental species as a potentially risky vector ( figure 2 ), which would not have been evident from historical data ( figure 1 ).
Flux is often positively related to establishment rates ( Colautti et al. 2006 , Hayes and Barry 2008 , Simberloff 2009 ), and reducing or eliminating flux is a prime management target. However, the risk of harm is also shaped in the consecutive stages along the invasion pathways of entrainment, transport, and release into the environment, and the contribution of each stage to overall risk differs among vectors ( figure 4 ). Flux estimates for ornamental, live bait, and live seafood species are not necessarily good predictors of the probability that they will be introduced (released into the environment), because the available data on flux reflect only part of the full journey from source to destination waters (see the points marked as a red 1 or 2 in figure 4 ). Therefore, although these organisms are entering California, there is no information on their release into marine and estuarine habitats until their establishment as NIS is detected. Release rates are also unknown for marine ornamental species; only a few rates exist for freshwater ornamentals (e.g., Strecker et al. 2011 ). Despite the high flux of marine ornamental species and their hardiness, their establishment has been low in California, perhaps because they are rarely released, because most of the species are tropical and do not survive release, or because they are transferred out of the state.
Comparison of maritime vectors and potential management control points. Eight contemporary maritime vectors (the left column) are compared across the relative size of the source propagule pools (the circles on the left), typical transit processes (the polygons in the middle section), and relative inocula sizes during propagule delivery (the circles on the right). The white circles and polygons represent target species pools and deliberate transfer activities during intentional vector processes; the gray circles and polygons represent unintentional transfers of species; the hatched polygons represent intentional transfers of species with associated unintentional transfers. The left-column circles' diameter represents the estimated species richness at the beginning of a typical transfer (the small, medium, and large circles reflect 1–9, 10–99, and 100–1000 species per shipment, respectively). The contemporary shipments of live bait include the target bait species and the associated unintentional entrainment of nontarget species in packing material (marked as a red 1). A large influx of species through the seafood, ornamental, and bait vectors arrive into terrestrial hubs (e.g., airports, wholesalers), followed by an unknown attrition rate, with releases into the sea by end users (marked as a red 2). Biofouling of commercial, recreational, and fishing vessels transfers species into and within the state with invasion opportunities enhanced by constant vector contact with the marine environment. The dashed-line box shows the historical operation of aquaculture (the red 3), to be compared with the present (the red 4), and the historical ballast water operation (the red 5), to be compared with the modern levels (the red 6).
Comparison of maritime vectors and potential management control points. Eight contemporary maritime vectors (the left column) are compared across the relative size of the source propagule pools (the circles on the left), typical transit processes (the polygons in the middle section), and relative inocula sizes during propagule delivery (the circles on the right). The white circles and polygons represent target species pools and deliberate transfer activities during intentional vector processes; the gray circles and polygons represent unintentional transfers of species; the hatched polygons represent intentional transfers of species with associated unintentional transfers. The left-column circles' diameter represents the estimated species richness at the beginning of a typical transfer (the small, medium, and large circles reflect 1–9, 10–99, and 100–1000 species per shipment, respectively). The contemporary shipments of live bait include the target bait species and the associated unintentional entrainment of nontarget species in packing material (marked as a red 1). A large influx of species through the seafood, ornamental, and bait vectors arrive into terrestrial hubs (e.g., airports, wholesalers), followed by an unknown attrition rate, with releases into the sea by end users (marked as a red 2). Biofouling of commercial, recreational, and fishing vessels transfers species into and within the state with invasion opportunities enhanced by constant vector contact with the marine environment. The dashed-line box shows the historical operation of aquaculture (the red 3), to be compared with the present (the red 4), and the historical ballast water operation (the red 5), to be compared with the modern levels (the red 6).
In contrast, ballast, fouling, and aquaculture organisms are released to or directly contact marine environments. Vessel biofouling carries a large number of species and individuals (high flux), but the quality of the vector's habitat is variable for individuals, some of which are undoubtedly lost en route ( Murray et al. 2012 ). Aquaculture differs in having a comparatively low flux, but there is a strong economic incentive to ensure survival. However, data are lacking on the number of permitted aquaculture organisms that are actually placed in the environment.
Marine NIS impacts in California
The ultimate goal of NIS risk assessment is to predict the probability of ecological, economic, and social harm. The perception that harm will occur often drives responses to invasions and motivates NIS management. Just as vectors differ in the number and frequency of nonnative species that they introduce, the impact of those nonnative species might not be distributed evenly among the vectors. To assess whether the ecological and economic impacts of marine NIS in California could be differentiated by vector, we completed BIOSIS ( thomsonreuters.com/biosis-citation-index ) searches for the impacts of mollusc, algal, and crustacean species (including alternate and synonymous species names) from 1926 through 2011 (see supplemental table S2 for the search protocol and references for the results). These taxonomic groups represent the majority of the NIS in California.
Published peer-reviewed information was too limited to assess the impacts of these taxa, let alone by vector (table S2). The majority of the information uncovered was devoted to only three or fewer species in each taxonomic group. Fifty publications concerning impacts were available for 11 of the 41 established molluscan species (table S2), with 34% of these publications devoted to one species, Mytilus galloprovincialis . The impact literature on algal species contained similar results, with three species ( Caulerpa taxifolia , Sargassum muticum , a Codium fragile subspecies) dominating 84% of the 124 impact publications. Impact data were available for just 17 out of 87 crustacean NIS, with 42% of the literature dedicated to the European green crab ( Carcinus maenas ). Because only 17 publications on molluscs, 2 on algae, and 6 on crustaceans were specific to California, the relevance of the impact information may be limited, given that impacts can be highly context dependent ( Thomsen et al. 2011 ). Without better data, impacts cannot be apportioned across vectors, leaving vectors to be singled out for their impacts one at a time. For example, the ornamental trade vector has stood out as being responsible for introducing some of world's worst aquatic invasive species ( Padilla and Williams 2004 , Semmens et al. 2004 ), including the seaweed Caulerpa taxifolia , which cost California over $6 million to eradicate ( Anderson 2005 ), and lionfish ( figure 5 ).
Lionfish (Pterois volitans) in its native Indo-Pacific region. Lionfish are an ornamental species that invaded the Caribbean, the Gulf of Mexico, and US Eastern Seaboard. They can tolerate temperate waters and are being imported into California. Photograph: Bruce Nyden.
Lionfish (Pterois volitans) in its native Indo-Pacific region. Lionfish are an ornamental species that invaded the Caribbean, the Gulf of Mexico, and US Eastern Seaboard. They can tolerate temperate waters and are being imported into California. Photograph: Bruce Nyden.
Vector management led to temporal changes in flux
To the extent that they have been taken, management approaches have also varied by vector. Although aquaculture and ballast water were historically potent vectors for marine NIS introduction, these vectors have been deliberately interrupted to reduce their flux (see the red 3 and 4 in figure 4 ). The drivers of change between the historical and modern transport of bivalves and ballast water differed considerably; aquaculture vector changes were driven by industry practices and profitability, whereas the ballast water mechanism changed because of explicit vector management policy and regulation.
Historically, aquaculture shipments were a stronger vector than they are today because the intentional transfer of adult bivalve species was accompanied by an assortment of unintentional “hitchhiking” species (see the red 3 in figure 4 ), which benefited from the high-quality transport conditions needed for the bivalves. However, consciousness about NIS has heightened in the aquaculture industry, with the result that, under current practices, the accidental introductions of associated species, or “hitchhikers,” which were historically an important source of NIS, are now minimal. The vast majority of contemporary bivalve shipments consist of larvae or juveniles of one species, the commercial Pacific oyster Crassostrea gigas , with few (if any) additional entrained species, which results in a dramatically lower risk of invasions (see the red 4 in figure 4 ). Analogous to current practices for aquaculture shipments, the elimination of biological packing materials (seaweeds; Haska et al. 2012 ) offers an opportunity to reduce the unintentional species that accompany intentional bait shipments, which could slow the transfer of NIS from New England and Asia to California.
Historical ballast water entrained a large pool of planktonic species during a voyage that was reduced in species richness and abundance on discharge by the net effect of interacting biological and environmental factors (the red 5 in figure 4 ). The initiation of midocean ballast water exchange (the red 6 in figure 4 ) has reduced propagule delivery by enhancing the interruption of the transport phase. Today, international, federal, and state ballast water regulations are among the few examples of active vector management in the marine realm; however, establishing such management practices has been the result of an arduous, ongoing, 30-year process.
A few marine and brackish ornamental species intended for sale or display are regulated by California as restricted species (e.g., alligators, sharks, gars), and the state has enacted legislation to ban the importation, sale, and possession of nine species of the seaweed Caulerpa (Assembly Bill 1334, chaptered in 2001, CDFG Code 2300). Neither California's nor the US Department of Agriculture's noxious weed listing of Caulerpa taxifolia (an invasive Mediterranean strain, listed in 1999) has been effective ( Diaz et al. 2012 ). The ornamental trade's involvement in NIS management has not been as strong as industry involvement has been for aquaculture and ballast water, which is both a lesson and an opportunity.
Conclusions
Clearly, the many disparate sources of data available for each vector were not adequate for a rigorous risk assessment of multiple vectors, even for just the initial steps of entrainment and transfer of NIS in the invasion process, let alone their release and impacts.
Way forward: Vector “blitz,” expert judgment, and management opportunities
To contend with the lack of a common currency for a cross-vector comparison of the identity and abundance of organisms arriving in the multiple vectors and in recognition of the fact that managers must prioritize resource allocation in data-poor situations, we recommend a novel type of assessment to obtain comparable flux data rapidly and relatively inexpensively. Rapid assessments, or bioblitzes , in which the number of newly introduced and established NIS is quantified at a single field location over a short time interval (e.g., days), are common ( Delaney et al. 2008 ). In an analogous vector blitz , the abundances and identity of organisms would be quantified in a coordinated manner across vectors over a standard time period and location using a standard sampling unit to be resolved beyond the units that we present in figure 3 . Although marine bioblitzes characterize NIS that have already arrived, the vector blitz would characterize the names and numbers of potential invaders on the way. Vector blitzes would force a resolution of the common currency problem. For example, should the number of vessels, their volume, or their wetted surface area be accounted for—or should all three? The locale selected for the blitz would need to be sufficiently large to encompass the representative vectors. Shipping ports generally do not support aquaculture, because of poor water quality, but aquaculture often occurs in nearby bays. For a Californian example, aquaculture in Agua Hedionda Lagoon and Tomales Bay is within striking distance of Los Angeles and San Francisco, respectively. A rigorous vector blitz would provide a concrete measure of the cost involved in collecting data for a longer-term cross-vector comparison or risk assessment. Coordination of scientists, agency staff, and citizens would be necessary. Access to commercial ports, marinas, and small vessels would need to be secured, which was a problem for our fishing vessel surveys. Taxonomic expertise would be required, as would institutional review board approval (for human subject research) in order to survey vendors of seafood, bait, and ornamental species. A vector blitz is a perfect opportunity for a targeted educational campaign to raise awareness of the multiplicity of vectors in operation. Given that a vector blitz would be the first of its kind, it would constitute a rigorous proof of concept and a step toward defining a common currency for NIS fluxes. Thereafter, however, vector blitzes should be repeated in order to capture the dynamics of NIS introductions and vector operation, with an overall view to evaluating management efficacy and reprioritizing vectors when necessary.
Although understanding the fluxes of organisms in vectors or propagule pressure is critical to managing vectors, other differences among vectors influence the probability of actual introduction, establishment, and impact. As we described earlier, these differences are known only qualitatively. In this data-poor situation, the tool of expert judgment offers a stopgap measure to help inform management decisions about vectors ( Hayes 2002 , Therriault and Herborg 2008 , Acosta and Forrest 2009 ). Although expert judgment is inherently subjective and prone to systematic errors and can lead to false confidence in the result, it can supplement limited data and help quantify uncertainties, as was demonstrated in recent applications of Bayesian models to invasive species ( Kuhnert et al. 2010 ). Unlike more commonly used statistical models, Bayesian models can incorporate prior knowledge about the variables, such as the results of an expert-judgment elicitation on NIS release rates and impacts, into the model's development. A cross-vector Bayesian risk model for California's maritime vectors is under development, on the basis of our results and expert elicitation. The limitations of expert judgment are being addressed ( Burgman et al. 2011 ); for now, expert judgment is a crutch and not a substitute for data-driven management.
Despite major data gaps, there are clear examples of changes in vector operations that have caused sustained vector disruption for both aquaculture and ballast water ( figure 4 ), and further management opportunities exist. Our analysis revealed that vessel biofouling was and is a very strong vector, which supports a compelling need to reduce its NIS flux associated with both commercial vessels and small crafts. Similar to the management progress made for aquaculture and ballast water, managing vessel biofouling to induce more maintenance of the submerged surfaces of commercial vessels and transient boats could reduce propagule delivery ( Johnson and Fernandez 2011 ). A number of countries have recognized that managing the biofouling on ships' hulls can reduce the risk of marine invasions ( Gollasch 2002 , Floerl et al. 2005 , Hewitt et al. 2009 ). In 2011, the International Maritime Organization adopted biofouling management guidelines for ships greater than 24 meters, which are primarily commercial vessels, and is working on guidelines for smaller vessels. California now requires the periodic removal of hull-fouling organisms for vessels over 300 gross registered (imperial) tons and capable of carrying ballast water. An annual hull husbandry report would also be required to enable better understanding of the extent of the hull-fouling flux in California ( www.slc.ca.gov/spec_pub/mfd/ballast_water/Documents/FoulingInfoSheet.pdf ).
Our multivector analysis also revealed that reducing the risk from biofouling requires managing not only large vessels but also small craft. The small craft vector has roughly the same magnitude of vessels traveling from out of state as does the commercial vessel vector (tables 1 and 2, figure 2 ). These vessels can accumulate high biofouling loads when they sit in harbors for long periods ( figure 6 ), but they have not received any substantial management attention or even sustained outreach on NIS transfers to promote cleaner submerged surfaces, with the exception of efforts by the cooperative extension unit in California's Sea Grant Program ( http://ca-sgep.ucsd.edu/focus-areas/healthy-coastal-marine-ecosystems/healthy-ecosystems-boating ). Our analysis highlights the need for better data on the movements of small crafts and the extent of biofouling to determine whether prioritization of this sector of the vector is merited.
Heavy biofouling on the hull of a transient recreational boat in California. Photograph: Ian C. Davidson.
Heavy biofouling on the hull of a transient recreational boat in California. Photograph: Ian C. Davidson.
Given that the majority of established species are associated with multiple vectors, the key to reducing future rates of new NIS introductions is to move away from approaches that target only commercial shipping and toward a more diversified approach that tackles all vectors simultaneously, or, if they must be addressed sequentially, it should do so in a prioritized manner—for example, addressing biofouling of all vessel types next. Efficient, effective management is difficult to envision in the absence of good data and an authority with the resources and accountability for the management of the multiple vectors in operation. Vector blitzes would result in better, more comparable data for estimating the flux of NIS arriving in multiple vectors, as would be required for a quantitative cross-vector risk assessment. Centralized permitting and data collection for NIS-associated sales of live bait, seafood, aquaculture, and ornamental species could also yield more standardized data. The need for a centralized regulatory authority is highlighted by the aquaculture example ( figure 3 ), to echo recommendations that have been repeatedly put forth ( Schmitz and Simberloff 2001 , Lodge et al. 2006 ). Severe impediments to centralization are the expense and effort of reorganizing government bureaucracies and the funding of external contractors, which has contributed to the expiration of several useful NIS Web sites. In California, the Ocean Protection Council is serving at least in a centralized advisory role for maritime NIS. A more flexible and perhaps less costly alternative to centralization is a network of information distributed from nodes representing sites at which data are collected routinely across vectors ( Ruiz and Carlton 2003 ), ideally in standardized vector blitzes. Each node could be funded independently, and the loss of one would be a regrettable data gap but would not cause the collapse of the entire network.
Maritime NIS risk assessment has been moving from single-species approaches to the consideration of entire vectors ( Floerl et al. 2009 , Chan et al. 2013 ). The next opportunity for advancement in theory and practice lies in developing a rigorous assessment of multiple species in multiple vectors that represent the cumulative risk that NIS pose. There is no substitute for good data, and our analysis highlights the need to establish a common currency to compare NIS abundances in different vectors and to collect data on NIS release rates. The invasion process is understood very well, and there are good models and frameworks for how vector management can work, conceptually and practically. We need to join the various disparate components into an integrated system, allowing a strong basis for prioritization and science-based decisions for vector management—leaving behind the current ad hoc approach that leaves the door open to new invasions.
Acknowledgments
This work was supported by Proposition 84 funds made available to the California Ocean Science Trust by the California Ocean Protection Council, with additional funding from the California Ocean Science Trust, the California Sea Grant Program, and the Smithsonian Institution. We thank many staff members of federal and state agencies for their generous assistance, especially Gary Townsend and Kirsten Ramsey. The Ocean Science Trust staff facilitated and helped shape the work—in particular, Skyli McAfee, Rebecca Gentry, Erin Kramer-Wilt, and Ryan Meyer. We thank Paul Fofonoff, Kim Holzer, and Brian Steves for crucial database support.
References cited
