Commercial shipping is the primary pathway of introduction for aquatic nonindigenous species (NIS), mainly through the mechanisms of ballast water and biofouling. In response to this threat, regulatory programs have been established across the globe to regulate and monitor commercial merchant and passenger vessels to assess compliance with local requirements to reduce the likelihood of NIS introductions. Resource limitations often determine the inspection efforts applied by these regulatory agencies to reduce NIS introductions. We present a simple and adaptable model that prioritizes vessel arrivals for inspection using proxies for potential propagule pressure (PPP), namely a ships’ wetted surface area as a proxy for the likelihood of biofouling-mediated PPP and ballast water discharge volume as a proxy for ballast water-mediated PPP. We used a California-specific dataset of vessels that arrived at California ports between 2015 and 2018 to test the proposed model and demonstrate how a finite set of inspection resources can be applied to target vessels with the greatest PPP. The proposed tool is adaptable by jurisdiction, scalable to different segments of the vessel population, adjustable based on the vector of interest, and versatile because it allows combined or separate analyses of the PPP components. The approach can be adopted in any jurisdiction across the globe, especially jurisdictions without access to, or authority to collect, risk profiling data or direct measurements for all incoming vessel arrivals.
Citation: Ceballos-Osuna L, Scianni C, Falkner M, Nedelcheva R, Miller W (2021) Proxy-based model to assess the relative contribution of ballast water and biofouling’s potential propagule pressure and prioritize vessel inspections. PLoS ONE 16(7): e0247538. https://doi.org/10.1371/journal.pone.0247538
Editor: Christopher W. McKindsey, Maurice Lamontagne Institute, CANADA
Received: February 5, 2021; Accepted: June 17, 2021; Published: July 1, 2021
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Commercial shipping is the primary pathway of introduction for aquatic nonindigenous species (NIS) that have established within coastal and estuarine waters globally [1–4]. Commercial vessels transport NIS from one location to another through the uptake and discharge of ballast water and as a result of biofouling associated with the vessel’s wetted surface area (WSA) [5–7].
Ballast water is taken on board a vessel for trim and stability purposes and to offset the mass imbalance of the vessel during cargo loading and unloading operations . However, when ballast water is taken on board in one location, planktonic communities are inadvertently entrained, along with sediments and benthic biota that may have been suspended in the water column by vessel activity. These organisms, which could include NIS, are released when the ballast water is eventually discharged in another port [1, 9].
Biofouling refers to the biota attached to, or associated with, the submerged portions of a vessel (i.e., the WSA) . Biofouling organisms can be sessile or mobile, and range in size from microscopic bacterial biofilms to larger macrofauna. These organisms can either cling to the vessel’s wetted surfaces (e.g., biofilms, barnacles, bivalves, bryozoans, algae) or find refuge in internal cavities where they are protected from shear forces at the hull-water interface (e.g., mobile crustaceans, fish) [11–14]. The biofouling communities travel wherever the vessel travels, transporting organisms to new ports and places. These biofouling organisms can be released as adults, larvae, or other propagules, either through physical forces such as the hull rubbing on a pier piling or displacement from an activated bow thruster, or by natural release and spawning [12, 13].
In response to the biosecurity threat posed by commercial shipping, jurisdictions across the globe have established (or are in the process of establishing) regulatory programs tasked with reducing the likelihood of NIS introductions via ballast water and biofouling (e.g., California’s Marine Invasive Species Program) [15–17]. Each of these local, state, federal, and international programs have the goal of implementing ballast water or biofouling management requirements/regulations to prevent species introductions in their respective jurisdictions. A critical element of these efforts is a robust outreach and inspection program to assess and improve compliance. Active communication between regulators (e.g., Port State Control) and a vessel’s crew is the most straightforward way of getting location-specific requirements into the hands of the people directly responsible for ballast water and biofouling management actions . Ideally, all arriving vessels would be boarded for outreach and inspection, but resource limitations such as staffing levels and funding often require regulatory agencies to make decisions about which vessels to board and which vessels to bypass. For example, vessels make between 10,000 and 11,000 arrivals in California each year at ports that span 700 miles in linear distance. How agencies allocate their limited resources to inspections can vary along a spectrum from arbitrary to deliberate. Using a standardized, data-driven prioritization approach designed for conditions specific to each jurisdiction could greatly improve the decision-making process.
A variety of risk-based vessel prioritization approaches have been described in recent years [18–21], typically with the goal of assessing the likelihood of a vessel introducing NIS through ballast water or biofouling. Such approaches and their reliability differ depending on the quantity and quality of input data available (Fig 1). Agencies with detailed and readily accessible data on ballast water and/or biofouling operational and maintenance practices, or with detailed source and recipient port information, are in the strongest position to use intricate risk-based prioritization approaches for boarding and inspecting vessels to assess compliance. However, other approaches based on available proxies can also be useful when detailed data on vessel operations or the similarities in environmental conditions between ports are limited or not readily accessible.
Reliability of an approach is dependent on the quantity and quality of available resources. Direct measurement approaches may rely on physical sampling of ballast water, surveying of a vessel’s wetted surface area, or similar methods. Risk profiling approaches may rely on environmental matching, vessel operational profile, or other similar methods. Proxy-based approaches may rely on ballast water discharge volumes and wetted surface area as potential propagule pressure parameters, or other appropriate measurements [21, 22].
In this paper, we present a proxy-based model that prioritizes arriving vessels for inspection using fundamental information about ships’ WSA and ballast water discharge (BWD) volumes to enable agencies to make more protective data-driven decisions when resources or detailed data are limiting.
Despite the low numbers of empirical studies able to prove the direct relationship between vector activity and invasion success , the positive relationship between propagule supply (e.g., the release of organisms in BWD or from the WSA of the vessel) and the likelihood of introductions is well recognized [23–26]. Based on the assumptions that ballast water and biofouling are managed consistent with local requirements and that the likelihood of introduction increases with increasing propagule supply (also recognizing the uncertainty of translating proxy-based parameters into invasion success [23, 26, 27]), we present a simplified alternative model to prioritize arriving vessels based on potential propagule supply (i.e., potential propagule pressure, PPP) for jurisdictions with limited resources. In addition, we trial the model using a four-year dataset of vessel arrivals at California ports to illustrate how the model can be applied to any jurisdiction.
The model is based on PPP proxies, as described in Lo et al. , that are readily available to most jurisdictions (e.g., ballast water discharge information via https://nbic.si.edu/database/). The proxies used in this approach are BWD volume as a proxy for ballast water PPP and WSA as a proxy for biofouling PPP. The targeted users for this approach are jurisdictions that do not have access or resources to analyze detailed vessel operational or environmental matching data between source and recipient ports, specific data not available for the jurisdiction, or that do not have the capacity and time to make direct measurements.
Wetted surface area
The use of WSA as a proxy for PPP relies on the assumption that the likelihood of introduction increases as the area of colonizable surface, including niche areas, increases. We used the same vessel dataset as Miller et al. , where the WSA for 373,833 vessel arrivals at United States ports was calculated using the WSA equation and the coefficients reported by Van Maanen and Van Oossanen . A relationship between WSA and gross tonnage (GT) was established via regression analysis (Table 1) for each of the following vessel-types: general (cargo), passenger, RO-RO (auto carriers), bulker, container, and tanker. GT was used here, rather than Net Tonnage as used by Miller et al. , because it is a more readily available metric and to match with an existing California-specific vessel dataset used to trial the model. Unmanned barges (including their respective tug) and articulated tug-barges (ATB) were not included in the regression analysis due to the variability within these groups, but their WSA was calculated directly for each vessel following Van Maanen and Van Oossanen  coefficients.
In addition to creating regression equations to estimate WSA using GT, we also applied the estimated proportion of a vessel’s WSA accounted for by niche areas, as reported by Moser et al.  for each vessel type (Table 1). Niche proportions for ATBs and unmanned barges and their tugs were also estimated using specific niche area values reported by Moser et al.  and adding all niche areas expected for typical ATBs, unmanned barges, and tugs (K. Reynolds, pers comm, 2020). These two metrics, estimated WSA and niche proportion (Np), can be used to calculate Total WSA (TWSA) using Eq 1: (1)
Ballast water discharge volume
The use of BWD volume as a proxy for PPP relies on the assumptions that propagule supply increases as BWD volume increases  and that management is consistently applied in compliance with local requirements. We recognize the limitations that this assumption may have when trying to predict invasion success [23, 25, 27, 29], however, our intent is not to measure probability of species establishment. Instead, our intention is to rely on the positive relationship between BWD volume and propagule supply [23, 24, 29] to prioritize limited resources with the goal of targeting vessels that may have a greater likelihood of introducing NIS, using readily available information.
BWD volume data are available to most jurisdictions in the form of ballast management reports submitted for each vessel arrival. In the U.S., the National Ballast Information Clearinghouse (NBIC) provides vetted BWD and management information for all U.S. arrivals to state and territorial agencies via an online public data portal (https://nbic.si.edu/). Other factors like water origin, environmental matching between source and recipient waters, and management strategy also influence NIS introductions. However, this information is not readily available in most cases, and can be complex and challenging to analyze. As resources increase, additional factors can be included to improve reliability of the analysis (Fig 1) .
Potential propagule pressure: Combined influence of ballast water discharge and biofouling.
For jurisdictions that have minimal resources to accomplish this, individual vessel PPP scoring will allow for a simple prioritization scheme specifically designed for their own population of vessels. The process requires a representative number of historical arrivals (e.g., one month, one year, multiple years, referred to as “Population data” in S2 Dataset) to identify the median individual vessel value of both TWSA (i.e., medTWSAind), calculated using the regression equations in Table 1 for each vessel type and Eq 1, and BWD (i.e., medBWDind) from vessel reported data. A PPP score for each vessel arrival can then be calculated using the estimated TWSA (i.e., TWSAind) and the BWD volume (i.e., BWDind) specific for that arrival as the input values in Eq 2.(2)
(see the R script provided in S1 Script to calculate PPP score)
Each arriving vessel will have an individual PPP score that can be used to sort and prioritize them relative to the other vessels in the population on a daily basis.
PPP scores–component parts and cumulative scores.
The relative contribution of each PPP component (i.e., ballast water and biofouling) can be calculated separately using either of the parenthetical components on the right-hand side of Eq 2. The overall influence of either ballast water or biofouling on PPP over time (i.e., incorporating frequency of arrival) or geographical region (i.e., incorporating all arrivals for specific ports) can then be added appropriately. Similarly, the total PPP score can be added cumulatively across a region or over certain time periods to compare between ports or over time. Likewise, cumulative scores can be assessed by vessel type or myriad other categories.
Because conducting management to meet local requirements is considered the first layer to reduce the likelihood of NIS introductions, the proposed approach assumes that each jurisdiction already has voluntary guidelines or mandatory management requirements for both ballast water (e.g., ballast water exchange or ballast water treatment systems to meet discharge performance standards) and biofouling (e.g., antifouling coatings) in place. When there are local management requirements in place, potential noncompliance should be the first layer of prioritization. The PPP proxy-based prioritization scheme assumes that all vessels are compliant with local management requirements; if they are not, they should be automatically categorized as high priority for inspection independent of the estimated PPP score. The importance given to each parameter influencing the prioritization process (e.g., outreach needs, compliance history) before the use of the PPP score should be determined by each jurisdiction.
Model trial using California data
We used data from four years of arrivals at California ports (2015–2018, S1 Dataset) to trial the proposed model (For clarification, this proposed approach is not the current vessel inspection prioritization scheme used in California). BWD volume data were obtained from Ballast Water Management Reports submitted to the California State Lands Commission and vessel arrival data were obtained from the Marine Exchanges of Southern California and the San Francisco Bay Region.
PPP scoring by individual vessel–California.
Using the TWSA and BWD volume data for each arriving vessel during the first three years of our dataset (2015–2017), we identified the median values for BWD (i.e., medBWDind) and TWSA (i.e., medTWSAind) for future use in Eq 2. We then used the TWSA and BWD volume data for each arriving vessel during the final year of our dataset (2018) in Eq 2 with the median values identified earlier. These calculations produced individual vessel PPP scores for each arrival during 2018 to evaluate how vessels would be prioritized daily according to each arriving vessel’s relative PPP score.
Currently, California has a mandate to inspect at least 25% of all arrivals, corresponding to an average of 6 vessels inspected per day. Using this number as the target for high priority arrivals to inspect (and representing the resources available for the trial), we identified the 6 greatest PPP scores per day and analyzed how each vessel type would be represented under the arrivals prioritized for inspection.
Model results using California data
PPP scores by individual vessel in California.
California’s 2018 dataset produced unitless individual vessel PPP scores ranging from 0.27 to 12.28. The mean PPP score for all California arrivals was 1.56 ± SE 0.012 and the port with the greatest mean PPP score was the Port of Stockton (2.73 ± SE 0.21). For vessel types, bulker and passenger vessels accounted for the greatest mean PPP scores, whereas general cargo vessels and unmanned barges exhibited the lowest mean values (Table 2).
When evaluating how the PPP score could have been used to prioritize daily inspections over the entire set of 2018 arrivals (using the legislative obligation to inspect 25% of arrivals as a resources threshold to categorize high priorities), California would have prioritized container, tanker, bulker and passenger vessels, as these four vessel categories accounted for 96.9% of the high priorities (Table 3). These vessel types presented the highest PPP scores daily, reflecting the large volumes of ballast water frequently discharged by some of these vessels in combination with their large TWSA values.
When comparing the trial results of the proposed model with the actual California vessel inspection prioritization method used during 2018 (based on ballast water discharge activity, outreach opportunities, and compliance history), the proposed prioritization scheme would have resulted in more container, ATBs, and passenger vessels categorized as high priority for inspection and fewer (including zero in some cases) bulkers, tankers, general cargo vessels, RO-ROs, and unmanned barges (Table 3).
In the trial, the proposed prioritization scheme assumes that the greatest 6 individual PPP scores per day will be prioritized as high priority for inspection, regardless of whether the same vessel makes repeated visits. Other jurisdictions that use this approach should decide how often to inspect vessels that frequently arrive and maintain a good compliance record. For example, we could decide to only include the first high priority arrival for each vessel during 2018 and exclude all other arrivals from the same vessel from prioritization to avoid repeatedly inspecting the same vessels. This alternative approach would have resulted in severely reduced numbers of ATBs, container, and passenger vessels categorized as a high priority and more bulkers, tankers, general cargo, and auto carriers than would have been prioritized with the originally proposed PPP model. Reducing the number of repeated inspections would release additional resources that could be used to include other components of the prioritization scheme (e.g., outreach).
Cumulative PPP scores (by location or vessel type).
Cumulative PPP scores calculated during the trial showed that the Los Angeles/Long Beach port complex and the Port of Oakland received the greatest PPP (Fig 2), primarily because of the frequency of container and tanker vessels that arrive at these ports.
Bars represent the contribution of total wetted surface area (TWSA; blue) and ballast water discharge (BWD; green) to the combined cumulative score at all locations (horizontal) and all the different vessel types (vertical). Bubble size represents the cumulative score ranging from 2750.3 (largest bubble) to <1 (smallest bubble). ATB refers to articulated tug-barge. Ports with less than 50 arrivals were removed from the figure (Alameda, Santa Barbara, Monterey, Humboldt and Vallejo).
TWSA outweighed BWD as a factor in the overall PPP scores across most ports and vessel types, with bulkers as the only vessel type and Stockton as the only port where BWD was a larger contributor than TWSA (Fig 2), reflecting the prevalence of ballast water discharges for bulkers and the fact that Stockton receives bulkers almost exclusively .
The proxy-based model described here is a data-driven tool to prioritize ballast water and biofouling inspections of commercial merchant and passenger vessels. The approach is:
- Adaptable because it can be applied within any jurisdiction and relies on the specific characteristics of that jurisdiction’s vessel population
- Scalable because it can be adjusted to capture different segments of the vessel population (e.g., 10%, 15%, 25% of the arriving vessels)
- Adjustable because it can be altered to focus specifically on either ballast water or biofouling instead of taking a combined approach
- Versatile because scores can be added within groups (e.g., vessel types, different ports) or over time periods to differentiate the relative PPP within these groupings
- Simple because it can be set up in minutes and provide data-driven prioritization with only two readily available input values per arriving vessel
In the described form, this model is ideal for programs that regulate the management of both ballast water and biofouling as it considers the combined effect of proxies for both. However, it can be altered into separate or standalone prioritization approaches for programs or locations that are primarily interested in only ballast water (e.g., some U.S. states) or biofouling (e.g., GloFouling Partnerships member countries). The approach also provides jurisdictions with a tool to determine the most efficient use of limited inspection resources on a daily basis or after a retrospective analysis of the patterns (cumulative PPP). Daily arrivals can be prioritized to ensure the inspection of vessels most likely to introduce NIS when additional data are unavailable (assuming compliance with local management requirements is used as first layer of protection).
From a management perspective, the proxy-based PPP model could be particularly useful when assessing cumulative pressure across different ports to guide decisions about where limited resources should be allocated and target those ports that proportionally receive the greatest PPP.
This proxy-based model relies on knowing each arriving vessel’s BWD volume and GT, both readily available to most jurisdictions (e.g., via NBIC, public vessel information websites). Most regulatory programs require submission of ballast water source, management, and discharge activity (e.g., United States, Canada, Australia). In the U.S., vessels are required to submit the Ballast Water Management Report (BWMR) to the NBIC for every arrival at a U.S. port. Under the 2018 Vessel Incidental Discharge Act (VIDA), the NBIC now makes all electronically submitted BWMRs (approx. 99.5% of submissions) available to state programs immediately upon receipt at NBIC. Submission to NBIC is currently required by no later than six hours after arrival, but in practice most forms are submitted prior to arrival. The submission timing requirement under VIDA will be set by a future U.S. Coast Guard rulemaking action.
Vessel GT data are accessible at many free vessel information websites (e.g., vesselfinder.com) and can be used with the regression equations presented here, specific to each vessel type (Table 1), to estimate WSA. The inclusion of niche area WSA in TWSA provides a more realistic proxy for biofouling PPP, as niche areas are known hotspots for biofouling accumulation [14, 31, 32].
Our trial of the proxy-based model with a dataset of California vessel arrivals demonstrates the utility of the method. To meet California’s legislative mandate to inspect at least 25% of arriving vessels, the trial using the proxy-based prioritization scheme focused inspections on container, tanker, bulker, and passenger vessels. All of these vessels either frequently discharge ballast water in California and/or have large TWSA. When compared to the vessels actually categorized as a high priority in California during 2018 (based on outreach opportunities, compliance history, and BWD), the proposed model de-emphasized inspection of bulkers, tankers, RO-ROs, barges and general cargo vessels. Even though compliance history was not included in this trial, it is important to emphasize that California’s prioritization process considers compliance with state requirements a critical part of the risk assessment process, as well as other factors like outreach.
The proxy-based model presented here is a baseline method that, in the absence of more detailed data, allows the detection of those vessels that may have the greatest likelihood of introducing NIS. More detailed data or additional resources, when available, can be incorporated as additional layers to improve this baseline model . The model assumes that management of both vectors has been performed according to the local requirements as the first layer of protection against NIS introductions. In addition, this approach does not consider opportunities for outreach as part of the prioritization process nor does it highlight the value of inspecting lower-scoring arrivals (e.g., for violation follow-up, or distribution of information to new vessels). These are important additional considerations that should be included in any prioritization scheme when possible. The flexibility of this method allows each jurisdiction to define outreach, compliance, or frequency rules (e.g., target new vessels, decrease inspection frequency in response to compliance history, suspicion of potential violations) that can be incorporated into the prioritization approach to reach a more comprehensive distribution of the inspection efforts. The approach also provides the opportunity for each jurisdiction to focus more attention (efforts) on ballast water or biofouling, depending on their priorities and data availability.
The proxy-based prioritization model presented and trialed here is an adaptable, scalable, adjustable, versatile, and simple tool to rapidly identify a subset of vessels for ballast water and/or biofouling inspections (Step-by-step process in S1 Flowchart). The approach can be adopted globally and is especially useful for jurisdictions without resources to collect risk profiling data or direct measurements for all incoming arrivals (Fig 1).
S1 Dataset. California vessels arrivals from 2015 to 2018.
S2 Dataset. Excel data frame template to input “Population Data” and “Arrivals data”.
S1 Flowchart. Step-by-step process to use the proposed approach.
Nicole Dobroski, Jonathan Thompson, Michelle Pelka, and Susie Kropman reviewed draft versions of the manuscript. Jackie Mackay, Dalia Keroles, Craig Fultz, Pearl Tan, Nadia Day, Michelle Pelka, Eric Frisbee, Paulette Boyd, Mary Fakih, Blanca Garcia, Michele Wiebold contributed with the data management of the California dataset. Kevin Reynolds contributed with valuable information and knowledge about vessel engineering for the WSA calculations.
- 1. Ruiz GM, Carlton JT, Grosholz ED, Hines AH. Global invasions of marine and estuarine habitats by non-indigenous species: Mechanisms, extent, and consequences. Am Zool. 1997; 37(6): 621–632.
- 2. Hewitt C, Campbell M. The relative contribution of vectors to the introduction and translocation of invasive marine species. Produced for the Australian Government. 2010.
- 3. Ruiz GM, Fofonoff PW, Steves BP, Carlton JT. Invasion history and vector dynamics in coastal marine ecosystems: A North American perspective. Aquat Ecosyst Health Manag. 2015;18(3): 299–311.
- 4. Ricciardi A. Tracking marine alien species by ship movements. Proc Natl Acad Sci USA. 2016; 113(20): 5470–5471. pmid:27147604
- 5. Bailey S. An overview of thirty years of research on ballast water as a vector for aquatic invasive species to freshwater and marine environments. Aquat Ecosyst Health Manag. 2015; 18(3): 261–268.
- 6. Davidson I, Scianni C, Minton MS, Ruiz GM. A history of ship specialization and consequences for marine invasions, management and policy. J Appl Ecol. 2018; 55: 1799–1811.
- 7. Miller AW, Davidson IC, Minton MS, Steves B, Moser CS, Drake LA, et al. Evaluation of wetted surface area of commercial ships as biofouling habitat flux to the United States. Biol Invasions. 2018; 20: 1977–1990.
- 8. David M. Vessels and ballast water. In: David M, Gollasch S, editors. Global maritime transport and ballast water management. Dordrecht, Invading Nature—Springer Series in Invasion Ecology. 2015, 9. https://doi.org/10.1007/978-94-017-9367-4_2
- 9. Carlton JT. Transoceanic and interoceanic dispersal of coastal marine organisms: the biology of ballast water. Oceanogr Mar Biol Ann Rev. 1985; 23: 313–371.
- 10. Minchin D, Gollasch S. Fouling and Ships’ Hulls: How changing circumstances and spawning events may result in the spread of exotic species. Biofouling. 2003; 19: 111–122. pmid:14618712
- 11. Ruiz GM, Fofonoff PW, Carlton JT, Wonham MJ, Hines AH. Invasion of coastal marine communities in North America: apparent patterns, processes, and biases. Annu Rev Ecol Syst. 2000; 31: 481–531.
- 12. Gollasch S. The importance of ship hull fouling as a vector of species introductions into the North Sea. Biofouling. 2002; 18: 105–121.
- 13. Carlton JT. Patterns of transoceanic marine biological invasions in the Pacific Ocean. Bull Mar Sci 1987; 41: 452–465.
- 14. Moser CS, Wier TP, First MR, Grant JF, Riley SC, Robbins-Wamsley SH, et al. Quantifying the extent of niche areas in the global fleet of commercial ships: the potential for ‘‘super-hot spots” of biofouling. Biol Invasions. 2017; 19: 1745–1759.
- 15. International Maritime Organization: Guidelines for the control and management of ships’ biofouling to minimize the transfer of invasive aquatic species. Marine Environment Protection Committee, MEPCRes.207(62). London 2011. Annex 26. pp. 25.
- 16. Georgiades E, Scianni C, Davidson I, Tamburri MN, First MR, Ruiz G, et al. The role of vessel biofouling in the translocation of marine pathogens: Management considerations and challenges. Front Mar Sci. 2021; 8: 660125.
- 17. Scianni C, Lubarsky K, Ceballos-Osuna L, Bates T. Yes, we CANZ: initial compliance and lessons learned from regulating vessel biofouling management in California and New Zealand. Manag Biol Invasion. Forthcoming 2021.
- 18. Barry SC, Hayes KR, Hewitt CL, Behrens HL, Dragsund E, Bakke SM. Ballast water risk assessment: principles, processes, and methods. ICES J Mar Sci. 2008; 65(2): 121–131.
- 19. Keller RP, Drake JM, Drew MB, Lodge DM. Linking environmental conditions and ship movements to estimate invasive species transport across the global shipping network. Divers Distrib. 2010; 17(1): 93–102.
- 20. Verna DE, Rueb RR, Gantz CA, Gala GT, Green J, Zalusky JA, et al. A decision tree analysis of nonindigenous species risk from ballast water to the lower Columbia River and Oregon coast, USA. Manag Biol Inv. 2018; 9(3): 309–321.
- 21. Bradie JN, Bailey SA. A decision support tool to prioritize ballast water compliance monitoring by ranking risk of non-indigenous species establishment. J Appl Ecol. 2020; 58(3): 587–595.
- 22. Lo VB, Levings CD, Chan KMA. Quantifying potential propagule pressure of aquatic invasive species from the commercial shipping industry in Canada. Mar Pollut Bull. 2012; 64: 295–302. pmid:22206724
- 23. Drake DAR, Monroy-Casas O, Koops MA, Bailey SA. Propagule pressure in the presence of uncertainty: extending the utility of proxy variables with hierarchical models. Methods Ecol Evol. 2015; 6: 1363–1371.
- 24. Drake J, Jerde C. Stochastic models of propagule pressure and establishment. In: Keller R, Lodge D, Lewis MA, Shogren JF, editors. Bioeconomics of Invasive Species. New York: Oxford University Press; 2009. pp. 83–102.
- 25. Wonham JM, Byers JE, Grosholz ED, Leung B. Modeling the relationship between propagule pressure and invasion risk to inform policy management. Ecol App. 2013;27(3):1691–1706.
- 26. Lockwood JL, Cassey P, Blackburn T. The role of propagule pressure in explaining species invasions. Trends Ecol Evol. 2005; 20(5): 223–228. pmid:16701373
- 27. National Research Council. Assessing the relationship between propagule pressure and invasion risk in ballast water. Washington, DC: The National Academies Press. 2011. https://doi.org/10.17226/13184
- 28. Van Maanen JD, Van Oossanen P. Resistance. In: Lewis EV, editor, Principles of naval architecture, volume II. The Society of Naval Architects and Engineers. 1988;1–93.
- 29. Lockwood JL, Cassey P, Blackburn T. The more you introduce the more you get: the role of colonization pressure and propagule pressure in invasion ecology. Divers Distrib. 2009; 15: 904–910.
- 30. Commission. 2021 Biennial Report on the California Marine Invasive Species Program. Produced for the California State Legislature. California State Lands Commission, Marine Environmental Protection Division, 2021.
- 31. Coutts ADM, Moore KM, Hewitt CL. Ships’ sea-chests: an overlooked transfer mechanism for non-indigenous marine species? Mar Pollut Bull. 2003; 46: 1504–1515. pmid:14607549
- 32. Davidson I, Scianni C, Hewitt C, Everett R, Holm E, Tamburri M, et al. Mini-review: Assessing the drivers of ship biofouling management—aligning industry and biosecurity goals. Biofouling. 2016; 32(4): 411–428. pmid:26930397