ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on May 7, 2008
ICES Journal of Marine Science: Journal du Conseil 2008 65(6):841-850; doi:10.1093/icesjms/fsn069
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Management of an invasive marine species: defining and testing the effectiveness of ballast-water management options using management strategy evaluation
CSIRO Marine and Atmospheric Research, Castray Esplanade, Hobart, Tasmania 7001, Australia
Correspondence to P. K. Dunstan: tel: +61 3 6332 5382; fax: +61 3 6232 5485; e-mail: piers.dunstan{at}csiro.au.
Dunstan, P. K., and Bax, N. J. 2008. Management of an invasive marine species: defining and testing the effectiveness of ballast-water management options using management strategy evaluation. – ICES Journal of Marine Science, 65: 841–850.Invasive marine and fresh-water species are being spread around the world in ships' ballast water, damaging industries and natural resources. Management policies are being developed nationally and internationally in response to the threat, but these options are not being rigorously evaluated for their potential to meet management objectives. We used management strategy evaluation (MSE) simulation to compare the performance of different management rules for controlling the spread of an invasive sea star, Asterias amurensis, around the southern coast of Australia. A model incorporating population dynamics, oceanographic patterns, and vessel movement was developed to compare the performance of different ballast-water exchange rules at reducing the likelihood of new populations establishing at locations along the coast over time. Static management rules, where ballast exchange was mandated on all voyages, reduced the median likelihood of new invasions from 0.67 with no ballast control to between 0.36 and 0.42 as distance from the coast was varied. Reducing the volume of high-risk ballast water by 95% did not reduce the likelihood of invasion by 95%, but by an average of 21%. Exchanging ballast farther from the coast did not reduce the likelihood of invasion for any of the static management rules. Feedback management rules using a port monitoring programme to assess the risk of transporting larvae between ports were at least as effective as the static rules, but at a significantly reduced cost for this single-species example. MSE provides a method to compare management options against objectives in this uncertain environment, and can be used to evaluate new and expensive treatment options for their effectiveness and value.
Keywords: ballast-water exchange, invasive species, management strategy evaluation
Received 7 August 2007; accepted 29 March 2008; advance access publication 7 May 2008.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionReferences |
|---|
The movement of species from one location to another via human-mediated transport has had major impacts on many marine ecosystems throughout the world (e.g. Ruiz et al., 1997; Cohen and Carlton, 1998; Carlton, 1999; Bax et al., 2003). Ballast water has been implicated in the introductions of damaging marine and fresh-water invasive species worldwide, causing million of dollars of costs to industry and natural resources (Hallegraeff and Bolch, 1991; Carlton and Geller, 1993; Byrne et al., 1997; Ruiz et al., 1997; Shiganova, 1998). Ballast water is also capable of transporting viral and bacterial pathogens, including the bacteria that cause cholera (Ruiz et al., 2000).
There is a clear pattern linking marine invasions to the growth in maritime trade and changes in shipping activities as new technologies are introduced (Carlton and Geller, 1993; Hewitt et al., 2004). Ballast water is pumped into ballast tanks at a port (the donor port) to stabilize the vessel during unloading and to maintain trim during the subsequent voyage, especially when not fully loaded. This water is released together with any organisms surviving the voyage at or near a recipient port where more cargo is taken on. As ships became faster, the likelihood that marine fauna and flora in ballast water would survive the journey increased. With >35 000 commercial vessels at sea on any given day, it is estimated that 7000 species may be being transported around the world each day in the ballast water of commercial vessels, and up to 10 000 species when other invasion vectors are considered (Carlton, 1999).
Recognition of the threat of commercial shipping introducing invasive marine organisms led Australia to require in July 2001 that all ships arriving from international ports with ballast water deemed high risk by a risk-based decision support system (DSS), to first exchange this water on the high seas (http://www.daff.gov.au/aqis/avm/vessels/ballast/requirements). High seas ballast exchange replaces high-risk aquatic organisms likely to survive in the recipient port with low-risk open-ocean species less likely to survive a port environment and less likely to have a restricted distribution. The International Convention on the Control and Management of Ships' Ballast Water and Sediments introduced by the International Maritime Organization in 2004 awaits ratification and will require all international shipping to undertake high seas exchange, at least until improved treatment technologies for ballast water are available commercially.
Australian legislation and the International Convention aim to manage ballast water coming from international ports, but domestic shipping is also an important pathway for transporting introduced species between ports within Australia and elsewhere. Domestic shipping within Australian waters is the most likely vector for the transfer of several species around Australia, e.g. Asterias amurensis (northern Pacific sea star), Undaria pinnatifida (Wakame), Varicorbula gibba (European clam), and Mytilopsis sallei (black striped mussel). Ballast water has been implicated in the movement of A. amurensis from Hobart (the site of initial establishment) to Port Phillip Bay, a much more active port, where it now threatens all of southern Australia and potentially other trading partners in the southern hemisphere (Bax et al., 2003). Countries typically have many more domestic than international ports, and the potential for domestic transfer of invasive species (and domestic species of limited range) is high. This threat led the Australian government to develop a National System for the Prevention and Management of Marine Pest Incursions to manage all potential domestic vectors of marine organisms (http://www.daffa.gov.au/fisheries/invasive/national-system).
High seas exchange of ballast water is considered a temporary solution until technology for managing ballast water is available, approved by governments, and implemented by industry. Current requirements are that ships exchange at least 95% of their ballast water for open-ocean water between ports, which is typically achieved by pumping the equivalent of three tanks' volume of oceanic water through the tanks—a flow-through exchange (Rigby and Hallegraeff, 1994). An implicit assumption is that the risk of invasion, when all other variables are held constant, will be reduced by a similar amount. Changes in the abundance of planktonic species in the ballast-water tanks of commercial shipping are variable, rarely reaching a reduction of 95% and dependent on a large number of factors (e.g. Rigby and Hallegraeff, 1994; Dickman and Zhang, 1999; McCollin et al., 2007; but see Gray et al., 2007). We do not consider other methods of ballast exchange here because flow-through exchange is the most common method used in Australia (Australian Ship Owners Association, pers. comm.) and will be used until technological solutions are implemented.
Evaluating the impacts and effectiveness of management actions in the marine environment is complicated by uncertainties in biological, physical, and anthropogenic processes. A rigorous approach to dealing with this uncertainty, and one increasingly adopted by fisheries managers evaluating management actions under uncertainty is management strategy evaluation (MSE; Sainsbury et al., 2000). In this approach, a model that defines the system, its components, their relationships, and possible future states over a broad range of input uncertainty is used to test selected management strategies for their effectiveness in achieving defined management objectives against performance indicators. The model is not used for making specific predictions about the outcomes of events, but rather general trends and the relative performance of different management strategies. Management strategies can include feedback strategies containing observation and monitoring processes, scientific assessment, interpretation of results, and the implementation of the management decisions. Uncertainty and time-lags can be incorporated at each stage. Management strategies that meet defined objectives under a broad variety of future system states, and across the range of recognized uncertainty, are identified. Management strategies that do not perform well under these modelled futures are unlikely to perform well in real life.
We have applied the MSE approach to the problem of managing ballast water for domestic commercial shipping in Australia, and in particular its role in spreading the northern Pacific sea star, A. amurensis, a highly fecund predatory sea star from Japan that established in Hobart in the 1980s after adjusting its life cycle timing to the antipodean seasons (Turner, 1992; Ward and Andrew, 1995). Domestic shipping is the likely vector for its spread in the early 1990s to the much busier mainland Port of Melbourne, where it now threatens all of southern Australia and potentially other trading partners in the southern hemisphere (Bax et al., 2003). As the management objectives and monitoring techniques for marine invasives are still in a state of flux (Bax et al., 2006), we cannot fully describe all management objectives or performance indicators. However, one overarching objective of managers is to prevent the further spread of invasive species throughout Australia, and we use this as our basic performance measure.
A daily updated oceanographic current model of southern Australia (Condie et al., 1999) is used to provide the unassisted spread and establishment of the sea star. Superimposed on this unassisted spread is the movement of larvae of the sea star in the ballast water of commercial shipping, with shipping patterns representative of current trading patterns. Given a range of options for managing ballast, including the number of times tanks are exchanged, how far offshore exchange occurs, and whether it is mandatory at all times or only on voyages deemed high risk, we determine the probability that each option will restrict spread of the sea star to that caused by unassisted spread. We compare this benefit with the cost of each option based on direct costs of ballast water exchange, lost opportunity costs of increased transit times, and nominal costs for establishing the monitoring system. We include a feedback management mechanism in the risk-based management system where risk and management requirements are updated based on different levels of in-port monitoring (i.e. adaptive management).
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The operational model for MSE has three components—biological, physical, and anthropogenic. The biological and physical components have been described previously by Dunstan and Bax (2007), so are only described briefly here. Each simulation has a time-scale of 50 years.
Oceanic transport and estuarine exchange
MECO (Model of Estuaries and Coastal Oceans), a three-dimensional, non-linear hydrodynamic model, was used to describe circulation patterns of water across southeastern Australia between 118°E and 153°E and 27°S and 47°S. The oceanographic model covers a grid of 1034 x 2992 km (Figure 1) with a cell grain of 22.2 x 22.2 km, giving a 47 x 136 (6392 total) cell simulation. It has been used previously to describe a wide range of coastal and estuarine systems (Walker, 1996, 1999; Condie et al., 1999; Bruce et al., 2001). Larval densities are advected using van Leer's scheme (1974), with a centred space–time numerical diffusion scheme (Kowalik and Murty, 1993) and a diffusion coefficient set to 400 m2 s–1 (Dunstan and Bax, 2007). The oceanic dispersal includes coupled estuarine retention, so that the larvae can be moved into and out of the 97 estuaries included in the model. Retention times for the 97 estuaries were sourced from the SERM II model (Simple Estuarine Response Model, www.per.marine.csiro.au/serm2/index.htm), which models estuaries as either lagoonal (1 estuarine cell) or salt wedge (5 estuarine cells). Estuarine volumes were obtained from the Ozestuaries database (www.ozestuaries.org/).
|
Calculation of larval mortality and settlement
The biological model was parameterized with data from empirical measurements of adult mortality and density, fecundity, and fertilization dynamics, with probability distributions for adult mortality and fecundity (Dunstan and Bax, 2007). Parameter estimation and the structure of the biological model are described in detail in Dunstan and Bax (2007). The biological model for adult populations is age-structured with a constant rate of mortality. Constant adult and larval mortality generates exponential loss through time. Adult growth and mortality were estimated from data collected in the Derwent Estuary, Tasmania, Australia, where a population of the invasive sea star has been established for 20 years. Reproductive output and the population density were estimated from empirical data from the Derwent Estuary, using simulation models to calculate reproductive output (Morris, 2002). Larvae are produced by the adult populations at rates determined by the density of adults in a particular location, and the age/size structure of the adult population. Larvae are dispersed from the local adult population into estuarine or oceanic cells. The limits of population establishment/growth are determined by the trade-off between adult mortality and new recruitment. Recruitment rates are determined by the retention rates of locally produced larvae and the supply of larvae from external sources, principally through ballast water (Dunstan and Bax, 2007). As there is insufficient information to calculate parameters for each habitat, population parameters are identical across all habitats. There is no evidence of adult dispersal over long distances.
Accurate empirical estimates of the daily rate of larval mortality (Lz) and the probability of successful larval settlement (S) were not available. A simplified model using only the hydrodynamics of the Derwent Estuary and the local population dynamics was constructed, and the rates of Lz and S that allowed a self-sustaining population to continue in the estuary were obtained (Dunstan and Bax, 2007). The effects of these two variables on life history success are multiplicative, so we determined values of Lz that maintained the Derwent population—a self-seeding population in southern Tasmania—while varying S between 1 and 0.1 (Figure 2). The full system model was run with two combinations of larval mortality, and larval settlement was calculated from the simplified model: Lz = 0.1 and S = 0.5 (low mortality/settlement scenario) and Lz = 0.108 and S = 0.9 (high mortality/settlement scenario), giving a total mortality of >99.9% over the entire larval period. Populations with a settlement rate of <0.5 had an extremely stochastic population dynamic, were unstable over any period, and were not considered further, because this pattern did not match observed population trends.
|
Simulation of commercial shipping between ports
The anthropogenic component of the model represented commercial shipping movements around southern Australia, between 104 ports stretching from Esperance in the west to Eden in the east (Figure 1). Shipping movements were calculated from 5 years of Client Place Move (CPM) data from Lloyds Maritime Information Utility (LMIU; 1998–2002). Vessels leaving ports were classified into one of six types; bulk carrier, container, general cargo, roll-on/roll-off (Ro/ro), tanker, and woodchip carrier; each vessel type had different movement and ballast characteristics. The probability of movement between ports, the number of ships leaving each port, and an empirical Gaussian kernel distribution of the dead weight (DWT) in kilogrammes of each class of vessel was estimated from these data and used to provide the number of vessels of different size moving between ports.
Records of the ballast water discharged at ports were used to estimate the (seasonally adjusted) number of larvae that would be entrained in the ballast tanks from the previous port. Expected ballast water discharge (D) from each vessel was determined from a relationship between recorded vessel dead weights and ballast-water discharges from vessels discharging in Australian Ports between 1998 and 2001 (Vessel Management System, Australian Quarantine Information System) for each ship type. Estimated discharge for bulk carriers, general cargo, tankers, and woodchip carriers was estimated with a linear regression (with m and c the linear predictors) as
|
| (1) |
(the error distribution) is distributed as a pert distribution (a modified beta distribution), with parameters min, mode, max, w; where min is the minimum value, mode the most likely, max the maximum value, and w is the weight, determining the spread of the distribution. For container carriers and Ro/ro, the relationship was
|
| (2) |
distributed as above. For ports with a large number of vessel arrivals, it was possible to calculate values of the parameters m and c for those particular ports. However, for ports with low volumes of traffic, generalized parameters calculated from information from all ports were used. All ports within the model region were included in the model, and their positions were assigned in either estuarine or oceanic cells. If the ports were in estuaries, it was assumed that the port would be in the estuarine cell closest to the adjoining oceanic cell. Larvae in the MSE model are distributed as well-mixed particles, spatially homogeneous in estuarine and in oceanic cells (to a depth of 10 m), owing to constraints in information on the hydrodynamic properties of ports and uncertainty in the exact spatial distribution of larvae. The concentration of the larvae in the ballast water entrained is assumed to be the same as at the uptake site. Therefore, if an estuary contains 100 000 m3 of water with 1000 larvae and a vessel takes in 1000 m3 of water, the ship will contain ten larvae after water ballast has been loaded. Upon arrival at the destination port, 80% of the ballast water is discharged along with 80% of the transported larvae, simulating the behaviour of ships that do not completely discharge all water in their ballast tanks (Australian Ship Owners Association, pers. comm.). As the duration of voyages in Southern Australia is typically short (i.e. median time of 14 h; Table 1) and larval mortality for this species is unquantified in ballast tanks, we do not impose additional mortality on larvae during transport in ballast water. Because this assumption is constant across all simulations, it will not influence the comparison of the relative strengths of different management options, the main purpose of this work. The time spent in ballast tanks compared with the entire larval duration is very small and although the instantaneous rate of mortality may be increased in ballast tanks, larval mortality over the whole larval phase will not be significantly higher. This assumption slightly increases the likelihood of invasion, and imposes a more rigorous test in the management options.
|
Exchange of ballast water at sea
There is little information available on the proportion of A. amurensis larvae (or any other life history stages) retained in the ballast tank following exchange at sea (although it has been demonstrated that they are retained; Patil et al., 2004). Simultaneous exchange of ballast water is simulated, because sequential ballast exchange, where the ballast tanks are completely emptied before refilling, is seldom used in Australia because of safety concerns. Three times volumetric exchange of ballast water will lead to the loss of 95% of planktonic organisms, assuming that the larvae are distributed randomly throughout the ballast tanks (Rigby and Hallegraeff, 1994). In the current MSE model, ballast exchange at sea occurs as described in Rigby and Hallegraeff (1994), i.e. a situation where exchange occurs with perfectly mixed conditions in the ballast tanks. The proportion of ballast exchanged per cell as the vessel moves through it is
|
| (3) |
the distance travelled (km), s the speed of the vessel in km h–1, DWT the dead weight of the vessel (kg), and 0.34 the proportion of the ship's DWT carried as ballast (Kerr, 1994). At the beginning of a journey between two ports, a route is selected according to the management rules for ballast water. The route-selection algorithm chooses the two points closest to each port that satisfies the management rule, then plots a route between them that minimizes journey distance but preferentially moves in a direction that satisfies the management rule, where possible. In each cell that satisfies the management options (described below), a proportion of ballast is exchanged according to Equation (3) and the vessel speed. Vessel speed is adjusted to ensure that the required number of tanks of ballast is exchanged over the duration of the voyage.
Static management rules
The static management options considered were shipping with flow-through exchange of ballast water for three equivalent ballast tanks along the coast, shipping with exchange 1 cell from the coast (at least 22 km), shipping with exchange 2 cells offshore (at least 44 km), shipping with exchange in cells at least 200 m deep. The results were compared with shipping without any control on ballast water. In these scenarios, exchange of ballast water was mandated on all trips, irrespective of distance or invasion status of the ports of origin or destination. These simulations were repeated using two different sets of larval mortality and settlement success (low and high mortality/settlement scenarios).
Feedback management rules
One strength of the MSE approach is that it can evaluate feedback management rules. Australia uses a DSS to determine when vessels from international ports are required to undertake ballast-water exchange, and will use a similar system to manage the movement of ballast water around Australia. The DSS is based on a biological risk assessment (Hayes et al., 2007), taking into account infection status of originating and recipient ports, the probability of vessel infection and larvae surviving the journey, and the probability that target species can complete their life cycle in the recipient port. The planktonic period of A. amurensis (90 d) is much longer than the average vessel journey duration for southern Australia, and biogeographic information suggests that its final range in Australia could extend beyond the limits of the area simulated here (http://www.marine.csiro.au/crimp/nimpis/; Hayes et al., 2007). Therefore, we assumed that any larvae entrained in the ballast tanks survived the journey and subsequent discharge (see above), and that A. amurensis could survive and reproduce anywhere within the simulated area, simplifying the risk analysis to a consideration of the infection status of the source and destination ports.
In this simplest case, where journey survival and habitat suitability are assured, the factors determining ballast-water exchange requirements are the invasion status of the ports of origin and destination. Depending on the invasion status of the two ports, there are four alternatives, listed below.
- If no population is detected in the origin or destination ports, then no exchange is necessary: low risk of new invasions;
- If a population is detected in the destination port but not the originating port, then no exchange is necessary: low risk of new invasions;
- If a population is detected in both the destination and originating ports, then no exchange is necessary: low risk of new invasions;
- If a population is detected in the port of origin but not at the destination port, then ballast-water exchange between the two ports is required: high risk of new invasions.
The key factor in changing management requirements from no exchange to exchange is detection of a population in the source port. Similarly, the key factor in changing from exchange to no exchange is detection of a population in the destination port. Successful implementation of this management regime therefore depends on the rapid detection of populations as they establish in ports along the coast. The management response to the detection of a newly detected population is immediate, and the actions the vessels take will change as determined by the management feedback rules. Management is focused on the detection of large populations to limit the spread of the species, rather than early detection to eradicate the population. In these simulations, we assumed that surveys would use the new genetic probe designed for A. amurensis (Deagle et al., 2003), and that detection probability of a single individual adult or larvae is distributed binomially (Hayes et al., 2005). Sampling was simulated as a fixed number of samples taken annually from each port for the entire duration of the simulation. The number of samples, and hence the power of detection, could be varied between simulations. Horizontal tows of a plankton net with 100 µm mesh in the top 10 m of the water column are an effective method of sampling for A. amurensis larvae (Bax et al., 2006). Assuming random distribution of larvae through the top 10 m of the area of interest, i.e. the estuarine or oceanic cell being sampled, the detection probability is then determined by the density of larvae, the volume of water sampled, and the number of samples taken (Hayes et al., 2005). Larval density was calculated from the simulation, the volume of water sampled was held constant at 2772 m3 (equivalent to towing a net with a mouth area of 0.5 m2 for 1 h at 1.54 m s–1), and the detection power was determined for different sampling intensities (Hayes et al., 2005). For example, if larval density is 10–4 larvae m–3, then the corresponding detection probabilities will be 0.43, 0.75, 0.94, and 1 for 2, 5, 10, and 20 samples, respectively. Increases and decreases in larval density or habitat area will result in corresponding increases and decreases in the detection probability for a given sample number. Therefore, a population may be present at low density, but remain undetected for a period. Alternatively, populations in small areas may be detected rapidly. The samples were taken to correspond to the peak of larval production in the model on the 250th day of each year. The feedback rules are implemented as soon as any larvae are detected. Simulations for feedback management rules used the low mortality/settlement rules, and all vessels exchanged their ballast 1 cell from the coast (at least 22 nautical miles). The performance of mandated exchange on all voyages was compared with the feedback rules using 2, 5, 10, or 20 samples in each port for the duration of the simulation, to detect new populations.
Accounting for uncertainty
The power of the MSE approach over a deterministic simulation of the physical, biological, and management systems is that management scenarios are tested over a variety of possible futures, where possible futures arise from probability distributions on adult growth and mortality, settlement success, population density, vessel type, dead weight, and destination. Although we do not claim to have captured all (or even most) uncertainty, much of which by its nature remains unknown, by incorporating many sources of uncertainty, we are able to provide a more robust test of the likely success of alternative management options. Simulations for a particular management option (including unrestricted shipping) were repeated 1000 times to capture system variability and the probability that locations (oceanic cells or estuaries) were invaded over the 50-year duration of each simulation. Reduction in invasion risk for each option and its economic cost was estimated by comparing invasion probability with that resulting from unrestricted shipping. The performance of the various management options was compared using the actual densities of sea stars from all the locations where sea stars were present in the model. We report results for two population densities, the first representing the initial stages of invasion before the population had established (10–4 A. amurensis m–2), and the second after the population had established and reached pest densities (0.1 A. amurensis m–2). These two densities correspond to the interests of scientists interested in early invasions, and managers, interested in preventing pest density populations. Calculations were restricted to sites that were invaded at least once in any scenario. Scenarios were run with only natural dispersal (no shipping) to contrast with the management scenarios.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The probability of invasion depended on destination port characteristics. Median probability of invasion with no ballast-water exchange for sites with any sea stars present in at least one scenario (using low mortality/settlement parameters) was 0.671 for a target sea-star density of 1 x 10–4 m–2 (a detectable density) and 0.256 for 0.1 sea stars m–2 (a density considered a pest population; Table 2). Median probability of invasion to detectable levels was slightly higher for the low mortality/settlement scenario than for the high mortality/settlement scenario, but median probability of invasion to the pest densities was higher for high mortality/settlement, indicating that a greater number of invasions failed to establish over the longer term under low mortality/settlement.
|
Overall reduction in median probability of invasion under ballast-water management and the low mortality/settlement scenario ranged from 10% to 30% depending on target sea-star density (Table 2). However, probability of invasion for some individual ports was reduced by up to 80%, or more, under ballast-water management (Figure 3). Reductions in the probability of invasion were greater at low densities (i.e. 1 x 10–4 m–2; Figure 3a and c) than at high densities (i.e. 0.1 m–2; Figures 3b and 4d), but more variable. There was no discernible difference between the effectiveness of any of the static ballast-water exchange rules in reducing the probability of A. amurensis becoming established at pest densities. A similar pattern was shown for high mortality/settlement scenarios (Figure 3c and d). Ballast-water management had less of an impact on the probability of invasion to detectable levels, but an equivalent impact on the median probability of invasion to pest levels.
|
|
The performance of the risk-based DSS tables (where samples were taken within ports to detect the presence of larvae) was comparable with the performance of static-rule-based ballast-water exchange at 1 cell offshore (Figure 4). Performance at the minimum detectable densities was marginally worse than the static rule, but performance of all management options was essentially identical in reducing the probability of invasion at pest levels. All combinations of ballast-water management and number of samples taken during monitoring significantly reduced the median probability of invasion compared with the median probability without ballast-water control, but the overall performance varied with the number of samples taken. The effect of changing the number of samples taken to inform management is most apparent at sites that have the greatest reduction in invasion probability (i.e. Figure 4a; i
60). Approximately 19% of sites using the static rule of exchange 1 cell from the coast had a reduction in invasion probability of >60%, whereas when 20 samples were taken to inform the DSS, no site had a reduction of >60%.
Changing management requirements changed time taken to travel between ports (as vessels were forced to reduce speed to provide time for flow-through exchange). The median time taken to travel between ports was 14 h for vessels not managing their ballast water (Table 1). Mandating exchange along the coast increased median travel time to 36.8 h, requiring exchange 1 cell (
22 km) from the coast increased this time to 38.5 h, 2 cells (44 km) 40 h, and mandating exchange in 200-m water depth, i.e. off the continental shelf, to 55.6 h. Charter costs for shipping are estimated to be AU$2080 h–1 (Blias and Delforce, 2003). The annual charter cost of shipping in the simulation area increased from AU$95 million with no exchange to AU$139, AU$154, AU$154, and AU$231 million when exchange was required along the coast, 1 cell or 2 cells offshore, or at 200-m depth, respectively. However, using the risk tables with two samples taken annually at each port only increased total charter costs of shipping by AU$4 million, to AU$99 million. The total cost of sampling and processing to obtain presence/absence information for all ports, sampled once a year, is estimated at AU$171 000 for a single species, which does not add significantly to the cost of exchange (Hayes et al., 2005). Actual costs of exchange (i.e. running the pumps) are also negligible compared with charter costs (AU$0.05 h–1; Blias and Delforce, 2003).
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
Ballast-water management is an integral part of efforts to reduce the translocation of marine species into new habitats (Locke et al., 1993; Ruiz et al., 1997). Whether or not a population establishes in a suitable area exposed to translocation via ballast water (or any other vector) depends on two factors: the continued production of larvae from local reproduction to sustain and grow the population, and a continued supply of translocated individuals from established populations. Without at least one of these factors being met, populations cannot overcome environmental and stochastic events that would otherwise lead to local extinction, especially of smaller initial populations (Dunstan and Bax, 2007). Ballast-water management reduces the number of larvae trans-located from existing populations to new populations, particularly in situations where natural transport of larvae is not possible. However, ballast-water management may not be as effective as assumed in our model, particularly in coastal waters (McCollin et al., 2007). Our model represents the best-case scenario for ballast-water control using ballast exchange. Although a reduction in inoculation intensity does decrease invasion probability, the relationship is not linear. Model results indicate that complex interplays between larval supply and larval retention will determine the exact reductions in probability from location to location for this particular species. Differences in habitat suitability would further reduce the probability of invasion in many locations.
Establishing a reproductive population that can produce sufficient larvae to provide positive population growth depends on many factors, but especially in the density of the sea star (which affects fertilization success) and the residence time of water within the port (which affects the retention of newly produced larvae). Fertilization success in sea stars decreases rapidly as distance increases between males and females, and is maximized at high densities (Morris, 2002). This produces an Allee effect (Dunstan and Bax, 2007), where low-density populations do not produce sufficient larvae to compensate for adult mortality. If external supply is not maintained and the population is below a certain density, the population will go extinct. For the same initial inoculation of larvae, the likelihood of a population establishing will be higher at a site that has greater water retention than at a site with less retention (Dunstan and Bax, 2007). Once a population is established and larval production increases, overall population growth rates will be determined by the proportion of larvae retained and recruited to the local population. A small estuary with a long residence time will maximize the probability of a population establishing and growing to high densities—a 95% reduction in supply will often be insufficient to prevent establishment. For larger estuaries with shorter residence times where conditions are not as conducive to establishment or growth, 95% exchange may prevent establishment. These sites are marginal habitat for A. amurensis (i.e. in terms of habitat size and retention time). Overall, the median probability of invasion for all sites following 95% ballast-water exchange is never reduced by 95%.
Larvae in ballast exchanged at sea would be rapidly dispersed away from the release point by prevailing ocean currents over the remaining larval period. Establishment rates along the coast away from the destination ports were insensitive to changing the location of ballast exchange (i.e. next to the coast out to 200-m depth). Larvae rarely remained close to the release point, and the density of larvae at any one location rapidly decreased as larvae were advected and diffused away. Consequently, larval densities along the coast next to exchange points were never high enough to establish the self-sustaining populations necessary to reach pest densities. Pest densities were only achieved in more enclosed ports, which retain larvae for longer (Dunstan and Bax, 2007).
Adaptive management is a frequently espoused goal of environmental management because it uses current information (collected in a passive or active framework) to adapt management intervention to the current situation and knowledge. Australia has used a rigorous risk-assessment process to categorize the risk posed by vessels planning to discharge ballast water sourced internationally (http://www.daffa.gov.au/fisheries/invasive/national-system). Although this will be superseded by the developing IMO Convention (http://www.daffa.gov.au/fisheries/invasive/national-system), a similar risk-assessment process will be used to inform the management of domestically sourced ballast water to the same standard as the IMO standard. The use of a risk-based approach to managing ballast on domestic trips is especially important because it is on these shorter trips that ballast-water management will cause the greatest delays and costs. However, the risk-based approach must be demonstrated to provide a similar level of environmental protection as compulsory exchange. An adaptive approach is required so that the risk categories of trips can be modified following expansion of an invasive species of concern into the source port. In this example, regular plankton surveys of different intensity were used to monitor the arrival of A. amurensis at a new port.
Using a risk-based system to manage the exchange of ballast water was generally as effective as mandating ballast exchange on all trips, with a significant cost reduction. The key difference between the different management options when using adaptive management was the intensity of monitoring. Because the monitoring method used in the model is a plankton survey, it only detects A. amurensis larvae, which are assumed to come from an established adult population. However, the gene probe is very sensitive (Deagle et al., 2003), and could detect larvae released from ballast water when they are at low density, in the absence of an adult population, especially if sampling effort is high. If these results are erroneously attributed to the presence of an established population, and restrictions on discharging ballast water from other infected ports are removed, then the likelihood of an adult population establishing increases with the subsequent increase in larval input. An improved management strategy would be to classify a port as infected for ballast-water uptake, but uninfected for ballast-water discharge, at least until further surveys have confirmed the presence of a self-sustaining population. This is the approach that has been used in situations in Australia where there has been uncertainty whether isolated sightings of an invasive species in a new port have represented an established population.
This MSE application has highlighted the importance of the receiving environment as well as inoculation frequency. By identifying high-risk receiving environments, it should be possible to target monitoring and ballast-water exchange. If ballast-water management was operating only to reduce the risk of A. amurensis spreading, then industry costs could be reduced by relaxing the requirement to complete 95% exchange when proceeding to low-risk environments. Conversely, stricter exchange requirements could be imposed for vessels proceeding to high-risk environments. In this case, basing management on a risk-based approach would cost significantly less than an approach that mandated exchange on all trips, with a comparable long-term risk of invasion.
Management strategies that have a high probability of meeting objectives while minimizing costs to industry will be much harder to determine when we consider the threat of more than one invasive species. Different species will differ in their timing, environmental limits, current distribution, and transferability by different vectors. Managing this complex environment will be far more difficult than managing one vector for one species, as explored here, and it will be much more difficult to understand the trade-offs between risks and cost for vectors as diverse as commercial shipping, recreational boating, and the aquarium industry, especially as new treatment technologies are developed. Although this study focuses on ballast-water exchange for domestic shipping within Australia, the results will have application to shipping in other areas of the world, such as Europe, which has a similar latitudinal and longitudinal spread, and also where management requirements for ballast-water exchange/treatment and designation of acceptable areas for ballast water are still developing. As we have demonstrated for the single species A. amurensis, the MSE approach provides a way to determine the costs and the benefits of the alternative approaches and their interactions. Although the modelling framework will be more complicated, this only reflects the reality of managing multiple invasive species and the difficulty of reaching management objectives and knowing that they have been met for complex environmental problems. Formal evaluation of management strategies on their ability to meet management objectives using this and/or other approaches should be standard practice, when the risks of further invasions are high and the costs to industry substantial.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
-
Bax N., Dunstan P., Gunasekera R., Patil J., Sutton C. Evaluation of National Control Plan management options for the northern Pacific seastar, Asterias amurensis. (2006) Australia: Final Reports for the Department of Environment and Heritage.
Bax N., Williamson A., Aguero M., Gonzalez E., Geeves W. Marine invasive alien species: a threat to global biodiversity. Marine Policy (2003) 27:313–323.[CrossRef][Web of Science]
Blias A., Delforce R. Ballast water management, a comparison of alternative approaches. (2003) Canberra, Australia: ABARE Report Prepared for the Fisheries Resources Research Fund.
Bruce D. B., Condie S. A., Sutton C. A. Larval distribution of blue grenadier (Macruronus novaezelandiae Hector) in south-eastern Australia: further evidence of a second spawning area. Marine and Freshwater Research (2001) 52:603–610.[CrossRef][Web of Science]
Byrne M., Morrice M. G., Wolf B. Introduction of the northern Pacific asteroid Asterias amurensis to Tasmania: reproduction and current distribution. Marine Biology (1997) 127:673–685.[CrossRef]
Carlton J. T. The scale and ecological consequences of biological invasions in the world's oceans. In: Invasive Species and Biodiversity Management—Sandlund O. T., Schei P. J., Viken P., eds. (1999) Dordrecht, The Netherlands: Kluwer Academic Publishers. 195–212.
Carlton J. T., Geller J. B. Ecological roulette: the global transport of nonindigenous marine organisms. Science (1993) 261:78–82.
Cohen A. N., Carlton J. T. Accelerating invasion rate in a highly invaded estuary. Science (1998) 279:555–558.
Condie S. A., Longeragan N. R., Die D. J. Modelling the recruitment of tiger prawns (Penaeus esculentus and P. semisulcatus) to nursery grounds in the Gulf of Carpentaria, northern Australia. Marine Ecology Progress Series (1999) 178:55–68.[CrossRef][Web of Science]
Deagle B. E., Bax N., Patil J. G. Development and evaluation of a PCR-based test for detection of Asterias (Echinodermata: Asteroidea) larvae in Australian plankton samples from ballast water. Marine and Freshwater Research (2003) 54:709–719.[CrossRef][Web of Science]
Dickman M., Zhang F. Mid-ocean exchange of container vessel ballast water. 2: Effects of vessel type in the transport of diatoms and dinoflagellates from Manzanillo, Mexico, to Hong Kong, China. Marine Ecology Progress Series (1999) 176:252–262.
Dunstan P. K., Bax N. J. How far can marine species go: the influence of population biology and larval movement on future range limits. Marine Ecology Progress Series (2007) 344:15–28.[CrossRef][Web of Science]
Gray D. K., Johengen T. H., Reid D. F., MacIsaac H. J. Efficacy of open-ocean ballast water exchange as a means of preventing invertebrate invasions between freshwater ports. Limnology and Oceanography (2007) 52:2386–2397.[Web of Science]
Hallegraeff G. M., Bolch C. J. Transport of toxic dinoflagellate cysts via ship's ballast water. Marine Pollution Bulletin (1991) 22:27–30.[CrossRef][Web of Science]
Hayes K. R., Cannon R., Neil K., Inglis G. Sensitivity and cost considerations for the detection and eradication of marine pests in ports. Marine Pollution Bulletin (2005) 50:823–834.[Web of Science][Medline]
Hayes K. R., McEnnulty F., Green M., Lawrence E., Barry S., Gunasekera R., Sliwa C., et al. Ballast Water Decision Support System (DSS) Service Level Agreement(SLA)—Part 2. (2007) Hobart, Australia: CSIRO Division of Marine and Atmospheric Research.
Hewitt C. L., Campbell M. L., Thresher R. E., Martin R. B., Boyd S., Cohen B. F., Currie D. R., et al. Introduced and cryptogenic species in Port Phillip Bay, Victoria, Australia. Marine Biology (2004) 144:183–202.[CrossRef]
Kowalik Z., Murty T. S. Numerical modeling of ocean dynamics. Advanced Series on Ocean Engineering (1993) 5:61–63.[Medline]
Kerr S. Ballast water ports and shipping study. AQIS Ballast Water Research Series, Report 5, February (1994) 123.
Locke A., Reid D. M., van Leeuwen H. C., Sprules W. G., Carlton J. T. Ballast water exchange as a means of controlling dispersal of freshwater organisms by ships. Canadian Journal of Fisheries and Aquatic Sciences (1993) 50:2086–2093.
McCollin T., Shanks A. M., Dunn J. The efficiency of regional ballast water exchange: changes in phytoplankton abundance and diversity. Harmful Algae (2007) 6:531–546.[CrossRef][Web of Science]
Morris A. Early life history of Asterias amurensis. (2002) PhD thesis, University of Tasmania.
Patil J., Hayes K. R., Gunasekera R., Deagle B. E., McEnnulty F., Bax N., Hewitt C. L. Port of Hastings National Demonstration Project—Verification of the Type II error rate of the Ballast Water Decision Support System (DSS). (2004) Hobart, Australia: Final Report for Victorian Environmental Protection Agency, CSIRO Division of Marine Research. 104.
Rigby G., Hallegraeff G. The transfer and control of harmful marine organisms in shipping ballast water: behaviour of marine plankton and ballast water exchange trials on the MV "Iron Whyalla". Journal of Marine Environmental Engineering (1994) 1:91–110.
Ruiz G. M., Carlton J. T., Grosholz E. D., Hines A. H. Global invasions of marine and estuarine habitats by non-indigenous species: mechanisms, extent, and consequences. American Zoologist (1997) 37:621–632.[Web of Science]
Ruiz G. M., Rawlings T. K., Dobbs F. C., Drake L. A., Mullady T., Huq A., Colwell R. R. Global spread of microorganisms by ships—ballast water discharged from vessels harbours a cocktail of potential pathogens. Nature (2000) 408:49–50.[CrossRef][Medline]
Sainsbury K. J., Punt A. E., Smith A. D. M. Design of operational management strategies for achieving fishery ecosystem recovery. ICES Journal of Marine Science (2000) 57:731–741.
Shiganova T. A. Invasion of the Black Sea by the ctenophore Mnemiopsis leidyi and recent changes in the pelagic community structure. Fisheries Oceanography (1998) 7:305–310.[CrossRef][Web of Science]
Turner E. A northern Pacific seastar, Asterias amurensis, in Tasmania. AMSA Bulletin (1992) 120:18–19.
van Leer B. Towards the ultimate conservative difference scheme. 2. Monotonicity and conservation combined in a second order scheme. Journal of Computational Physics (1974) 14:361–370.[CrossRef][Web of Science]
Walker S. J. Hydrodynamics models of Port Phillip Bay. (1996) Melbourne, Australia: Technical Report 38, CSIRO Port Phillip Bay Environmental Study.
Walker S. J. Coupled hydrodynamics and transport models of Port Phillip Bay, a semi enclosed bay in south-eastern Australia. Marine and Freshwater Research (1999) 50:469–481.[CrossRef][Web of Science]
Ward R. D., Andrew J. Population genetics of the northern Pacific seastar Asterias amurensis (Echinodermata: Asteriidae): allozyme differentiation among Japanese, Russian, and recently introduced Tasmania populations. Marine Biology (1995) 124:99–109.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||