Promoting invasive species control and eradication in the sea: Options ...

Marine Pollution Bulletin 77 (2013) 165–171

Contents lists available at ScienceDirect

Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul

Promoting invasive species control and eradication in the sea: Options for managing the tunicate invader Didemnum vexillum in Sitka, Alaska Linda D. McCann a, Kimberly K. Holzer b,c,⇑, Ian C. Davidson d, Gail V. Ashton a, Marnie D. Chapman e, Gregory M. Ruiz b a

Smithsonian Environmental Research Center, Romberg Tiburon Center, 3152 Paradise Drive, Tiburon, CA 94920, USA Smithsonian Environmental Research Center, 647 Contees Wharf Road, Edgewater, MD 21037, USA Branch of Aquatic Invasive Species, U.S. Fish and Wildlife Service, 4401 North Fairfax Drive, Arlington, VA 22203, USA d Department of Environmental Science & Management, Portland State University & Smithsonian Environmental Research Center, 1025 SW Harrison Street, PO Box 751-ESM, Portland, OR 97207, USA e University of Alaska, Southeast, 1332 Seward Avenue, Sitka, AK 99835, USA b c

a r t i c l e

i n f o

Keywords: Biological invasion Didemnum vexillum Eradication Invasive species management Mortality experiment Biofouling

a b s t r a c t Bioinvasions are a significant force of change – and economic and ecological threat – in marine ecosystems. The threat now encroaches on Alaska, which has had relatively few invasions compared to other global regions, prompting need to develop new incursion response tools. We appraised five ‘eco-friendly’ immersion treatment options (dilute acetic acid, dilute bleach, freshwater, brine and hypoxia) at either minute- or hour-scale exposures to kill the invasive tunicate Didemnum vexillum. Data revealed 100% treatment efficacy after two minutes in acetic acid, ten minutes in bleach, four hours in freshwater and over four hours in brine solution. We also demonstrated the importance of monitoring D. vexillum recovery for at least three weeks, since seemingly destroyed colonies rebounded during this timeframe. Combined, these findings provide insights towards a bay-scale eradication and post-border management plan applicable to the recent D. vexillum incursion in Whiting Harbor, Alaska and other shallow, inshore invasion sites. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Management of marine invasive species has become a policy priority for environmental managers in many countries around the world and includes pre-border actions to prevent invasions and post-border actions to detect, mitigate or eradicate existing invasions (Hewitt and Campbell, 2007; Hewitt et al., 2009). The emergence of vector management policies, notably for vessel ballast water and biofouling, is a proactive approach to lower the numbers of species in transit within anthropogenic vectors to therefore prevent introductions. While such pre-border approaches are being adopted within nations and internationally, the use of preventative tools remains more firmly established than post-border management, especially in marine ecosystems. At present, reactive management activities after introductions have been detected (incursion response) are handled on a case-by-case basis rather than relying on larger scale policy frameworks. ⇑ Corresponding author at: Smithsonian Environmental Research Center, 647 Contees Wharf Road, Edgewater, MD 21037, USA. Tel.: +1 434 249 5824. E-mail address: [email protected] (K.K. Holzer). 0025-326X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.marpolbul.2013.10.011

Incursion response is an important but largely undeveloped component of marine biosecurity. The lack of appropriate eradication protocols and technologies represent obstacles to postborder marine invasive species management (Hewitt et al., 2005). In addition, there is a widespread impression that marine invasive species control and eradication are too difficult, impractical and costly to attempt, despite the growing literature on the topic and the emergence of eradication frameworks (Edwards and Leung, 2009). In contrast, post-border pest management has been prevalent in terrestrial and freshwater systems for many decades. In theory, there is no fundamental reason why eradication or control should be less successful in marine than freshwater or terrestrial systems, as they all involve the same basic parameters in population biology and methodological challenges (Edwards et al., in review). Furthermore, an increasing number of examples of successful eradications in marine systems have been published (polychaete – Culver and Kuris, 2000; seaweed – Wotton et al., 2004; seaweed – Anderson, 2005; bivalve – Hopkins et al., 2011;), demonstrating the potential for broader application. In addition, there are numerous recent cases where control and eradication methods have been considered for marine invasions,

166

L.D. McCann et al. / Marine Pollution Bulletin 77 (2013) 165–171

many focusing on nonnative tunicates (Edwards and Leung, 2009; Arens et al., 2011; Willis and Woods, 2011). This focus on incursion response for tunicates is perhaps best exemplified by Didemnum vexillum. In the late 1980s and early 1990s D. vexillum first emerged as a yet-to-be-identified global introduced species of concern. Since then D. vexillum has been identified and reported from many sites in Europe and New Zealand, as well as the east and west coasts of North America (Bullard et al., 2007; Herborg and O’Hara, 2009). A diversity of methods have been tested to eradicate (remove) or control (suppress) this fouling pest with variable success. A common approach is to dose the tunicate with different solutions (e.g., acetic acid, lime, freshwater, brine, chlorine, bleach) as an immersion dip, injection or spray. Burying (dredge spoil, cement powder), wrapping, burning (petrogen torch), scrubbing, air-drying and pressure washing have been explored for management of D. vexillum (e.g., Coutts and Forrest, 2007; Switzer et al., 2011). Researchers have also tested the biocontrol potential of various native and nonnative benthic predators, including one or more species of urchin, sea star, crab, gastropod and nudibranch with limited success (Carman et al., 2009; Epelbaum et al., 2009). While the consequences of D. vexillum invasions have not been thoroughly investigated, the tunicate fouls shellfish and aquaculture gear (Carman et al., 2010), hampers scallop movement (Dijkstra and Nolan, 2011), and overgrows very extensive areas of benthic habitat (Valentine et al., 2007). After its discovery in Alaska in 2010, we returned to the site to determine boundaries of the infestation, sample benthic communities in infested and uninfested areas of the bay, conduct experiments of potential mortality-inducing treatments, and consider eradication strategies with local and state officials and stakeholders. In this study, we provide (a) an update on the invasion of Whiting Harbor, Sitka, Alaska by the colonial ascidian D. vexillum and (b) results from experimental aquarium-scale trials of eradication treatments conducted at the site that may inform a bayscale management response. D. vexillum was first detected in Whiting Harbor in 2010 as a result of citizen-science-supported monitoring, although it is likely to have become established in the harbor several years prior to the detection date (Cohen et al., 2011). This invasion represents a >1000 km northward leap in distribution from other known occurrences of D. vexillum in the northeast Pacific, and no other populations of the ascidian are known from Alaska. As with many marine introductions on the Pacific coast of North America, the first confirmed detection of D. vexillum occurred in San Francisco Bay, California in 1993. Subsequent incursions over the last 20 years range from Baja California (Mexico), eleven different bays in California, two bays in Oregon, two bays in Puget Sound, Washington and five bays in southern British Columbia (Fofonoff et al., 2013). Probable vectors of introduction to Whiting Harbor, a 0.6 km2 sheltered bay without boat harbor facilities and the site of former oyster farms, include contaminated aquaculture gear or translocated floating docks. A less likely mechanism may have been recreational vessel biofouling since the area was occasionally used for fishing and overnight anchorage before ADFG restricted access in response to the D. vexillum invasion. While vessels cannot be ruled out as vectors, for perspective we note that Sitka’s municipal harbors are within 2–4 km of Whiting Harbor and see extensive vessel traffic, containing 1347 boat slips and hosting one of Alaska’s largest fishing fleets (SEDA, 2013). Yet these nearby areas have shown no evidence of D. vexillum despite intense surveillance efforts including settlement plate studies, dock inspections, public education, and dive inspection of the boats and slips formerly associated with Whiting Harbor aquaculture operations (M. Chapman pers. obs.).

2. Methods 2.1. D. vexillum extent and community ecology in Whiting Harbor In June 2011, we conducted scuba and snorkeling surveys of Whiting Harbor, Sitka, Alaska (57°20 N, 135°210 W) to map the extent of D. vexillum spread on the seafloor. Upon its discovery during the ‘bioblitz’ survey of 2010 (Cohen et al., 2011), it was recorded on floating docks, submerged ropes and lantern nets of the former oyster farm at the site. There were no in-water or intertidal surveys in Whiting Harbor during the bioblitz, so all material was observed and collected from items that could be reached from floating structures. Subsequent to the discovery, the Alaska Department of Fish and Game conducted diver surveys to assess the areal coverage of the ascidian on the seafloor (ADFG unpublished data), and we used these data as guidance for our evaluation of the extent of D. vexillum spread in benthic habitats of the harbor. In September 2011, we conducted scuba diver sampling of epibenthic communities in Whiting Harbor to investigate community differences that may have emerged between areas infested and uninfested with D. vexillum. The sampling protocol was stratified by D. vexillum presence and absence, with 48 random quadrats of primarily boulder substratum divided equally, such that D. vexillum was recorded within half of the samples. All quadrat samples were collected from 3.8 to 7.0 m depth within the same southeastern area of the bay so that samples inside and outside Didemnum-infested areas were adjacent to each other. Divers took images and counted all organisms (>1 cm) within each 0.25 m2 to produce a sample matrix of abundance per species encountered. For the 24 quadrats with D. vexillum present, percent cover was measured by point count using whole-quadrat images. Images and specimens were collected to confirm species identities when necessary. The data were used to produce a non-metric multidimensional scaling plot (nMDS) and analysis of similarities tests (ANOSIM; Primer Ltd, Plymouth, UK) were performed to test for differences in community composition between infested and uninfested areas. The R-statistic from ANOSIM tests fall between zero and one, with comparisons approximating one indicative of communities that are very distinct compositionally while those close to zero have a very non-distinct species composition. We also assessed for correlations between D. vexillum cover and species richness and abundance, and performed t-tests to compare richness and abundance in infested versus uninfested quadrats. 2.2. Effects of treatments on D. vexillum mortality To determine the efficacy of different treatment options to control or remove D. vexillum, aquarium-scale experiments were initiated on site in June and September 2011 (Table 1) and follow-up evaluation continued into November. Nylon mesh (aquaculture netting) covered with the ascidian was cut into pieces, such that replicate units consisted of approximately 5 cm2 encrusting D. vexillum colonies. We conducted trials of six different immersion treatments: freshwater, elevated salinity (brine, 62 ppt

Table 1 Treatment medium, immersion time and start date for Whiting Harbor, Sitka, Alaska field experiments on D. vexillum (n = 4 for each treatment combination). Treatment

Exposure(s)

Month

Acetic acid (10%) Bleach (1%) Freshwater Brine (62 ppt) Hypoxia Brine (62 ppt)

2, 2, 4, 4, 4, 1,

Jun Jun Jun Jun Jun Sep

5, 10 min 5, 10 min 24 h 24 h 24 h 2, 3, 4 h


salinity), hypoxia (0.5 mg L1), acetic acid (10%) and bleach (1%). For the independent immersion treatments, each replicate (n = 4) was placed in a gallon zipped plastic bag with 1–2 L of solution at ambient sea surface temperature (12–13 °C). The sealed plastic bags were submerged in Whiting Harbor to retain ambient seawater temperature through each treatment period. Ascidian colonies were immersed for four and 24 h in freshwater treatments. The brine solution was prepared with pure sodium chloride granules and seawater with immersion times of one, two, three, four and 24 h. The hypoxia condition was created by bubbling nitrogen gas into the zipped bags; colonies were immersed for four and 24 h. The four-hour treatment was bubbled continuously (0.04 ± 0.01 mg L1) and the 24-hour treatment bags were sealed after four hours, maintaining dissolved oxygen levels 0.05). ANOSIM results and the nMDS plot revealed that community composition in infested and uninfested plots was not very distinct, with a global-R value of 0.234 (p < 0.002) indicating a high degree of similarity between groups of samples (Fig. 3). We did not record significant differences between infested and uninfested samples in species richness and abundance of mobile fauna (t-tests; richness, t = 0.47, p > 0.05; abundance, t = 0.36, p > 0.05). We did, however, find significant differences between infested and uninfested plots in abundance and richness of sessile species. Although richness of sessile species was relatively low (ten ascidians and two sponges), these species were approximately twice as prevalent in quadrats without D. vexillum compared to quadrats with the invader (t-tests; richness, t = 3.67, p < 0.001; abundance t = -3.49, p < 0.001). 3.2. Effects of treatments on D. vexillum mortality For treatments tested on minute time-scales, there were significant differences in the outcomes of acetic acid and bleach when compared to controls (Kruskal–Wallis, v2 = 48.7061, p < 0.0001) (Fig. 4A). The ten-minute immersion of D. vexillum in acetic acid and bleach solutions resulted in complete colony mortality three weeks post-treatment. Shorter immersion durations of two and five minutes were only lethal for acetic acid. The shorter immersions in bleach produced an initial decline in colony health, including loss of surface area and pigmentation; however, colonies rebounded and resumed progressive, positive growth for the rest of the monitoring period. The control replicates for these treatments increased to at least a doubling in size over the duration of the monitoring period. There were also significant differences in the outcomes for longer duration treatments of freshwater, brine and hypoxia in June (Kruskal–Wallis, v2 = 26.4412, p < 0.0001) (Fig. 4B). Freshwater immersions for four and 24 h and immersion in brine for 24 h each resulted in complete colony mortality by the completion of the experiment. The shorter immersion in brine of four hours was not 100% fatal to colonies (as discussed below), but treatments did differ significantly from controls (Kruskal–Wallis multiple comparison, one-tailed p < 0.05) with three of the four colonies dying. Hypoxia treatments did not negatively affect D. vexillum colonies. Immersion in brine resulted in significant reductions in tissue of D. vexillum, but was not 100% effective for durations tested 64 h (2-way ANOVA, p < 0.001) in September (Fig. 5). The four-hour brine exposure resulted in substantial dieback of tissue (98% ± 2%) after five weeks, mirroring tunicate response from the June trial (Fig. 4B), and we did not measure changes over longer durations. The two- and three- hour brine immersions caused dieback but not complete mortality, while tunicates immersed for one hour showed growth at the five-week assessment point. Controls for the longer duration treatments increased size by 50–100% over the duration of the monitoring (Fig. 5).

168


Fig. 1. Current D. vexillum distribution on the west coast of North America and in Whiting Harbor, Sitka, Alaska. Known populations of D. vexillum extend from Bahia San Quentin, Mexico, through many sites in California, Oregon, Washington and British Columbia, to its most northerly and isolated population in Whiting Harbor, Sitka, Alaska. In Whiting Harbor, the introduced tunicate primarily lines the southern inner bay. It occurs on floating structures (red dots), shallow boulder scree (solid red) and in patches on the seafloor (diagonal red lines). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. In Whiting Harbor, Sitka, Alaska D. vexillum occurs (A) on floating structures, (B) in continuous encrusting mats overtopping boulder habitat and (C) on isolated boulder or debris ‘islands’ surrounded by shell gravel on the seafloor. Images by K. Holzer (A) and I. Davidson (B and C).

4. Discussion The invasion of Whiting Harbor by D. vexillum provides management challenges relatively novel to Alaska. The state has few established marine invaders compared to regions further south, but the northward trend of invader spread in the northeast Pacific (Ruiz et al., 2011), demonstrated by D. vexillum, suggests additional incursions by marine pests can be expected. In Sitka, Alaska D. vexillum may also present management and scientific opportunities. Despite its proliferation within the bay, dominating benthic habitats in certain areas, the population is isolated to this location and distributed in relatively shallow habitats. Combined with the shutdown of the likeliest vectors of D. vexillum introduction to

the bay, this invasion provides an excellent candidate for scaledup testing of mortality treatments and bay-wide eradication. The current impact of D. vexillum in Whiting Harbor remains somewhat opaque. Despite its noteworthy coverage of rocky benthos (Fig. 2B), we measured relatively minor community differences between infested and uninfested patches, and no differences in overall species richness and abundance of epibenthic macrofauna. These results are somewhat similar to surveys of communities inside and outside D. vexillum cover in Long Island Sound and Georges Bank on the East Coast of North America (Lengyel et al., 2009; Mercer et al., 2009). In Long Island Sound, no differences in benthic diversity and abundance occur between infested and uninfested samples, or contrary to predictions, elevated richness


Fig. 3. A non-metric multi-dimensional scaling plot of epibenthic communities in Whiting Harbor. Epifaunal communities in plots infested and uninfested with D. vexillum (yellow diamonds and blue circles respectively) were not very distinct (ANOSIM, R = 0.234, p < 0.002). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. Mean proportional surface area change (±1 SE) of D. vexillum subject to (A) minute-scale and (B) hour- to week-long treatment immersion times in June (n = 4). Only assessment intervals of three and five weeks are shown for tunicate responses. Surface area change of 1 represents complete mortality, zero is no change from the original extent of colonies, and one represents a doubling of the initial colony area.

and abundance occur inside Didemnum-infested samples (Mercer et al., 2009). Dredge samples from Didemnum-infested areas on Georges Bank also show elevated abundance of two polychaetes compared to uninfested areas, which is the main driver of community differences between impacted and un-impacted habitat patches (Lengyel et al., 2009). On the other hand, we detected significantly fewer sessile species in Didemnum-infested patches of benthos, which matches data from epibenthic video surveys (Lengyel et al., 2009). It was also notable that we found initial evidence for D. vexillum coalescence of pebble-gravel benthos at the fringe of boulder habitats, which may foretell a spread across a much larger areal extent of Whiting Harbor as occurs in Georges Bank (Valentine et al., 2007) and Long Island Sound (Mercer

169

Fig. 5. Mean proportional surface area change (±1 SE) of D. vexillum treated by brine immersion for zero (control), one, two, three and four hours in September, and assessed five weeks post treatment (n = 4 per treatment). Surface area change of 1 represents complete mortality, zero is no change and one represents a doubling of the initial area.

et al., 2009). Thus, while the longer term effects of spatial dominance by D. vexillum at this and other sites remains an important monitoring priority in the absence of control or eradication, the opportunity to enact management interventions should not necessarily wait until these longer term data are collected. The likelihood of successful management action typically diminishes over time, (Couts and Forest, 2009), especially in cases of (currently) isolated populations as in Sitka. Four effective eradication options for D. vexillum were identified that require either a short immersion time (acetic acid, bleach) or long immersion time (freshwater, brine) (Fig. 4). Acetic acid induced complete mortality after a two-minute exposure, while dilute bleach required a ten-minute immersion. Freshwater was lethal after four hours and brine solution substantially reduced D. vexillum coverage after four hours and was completely lethal after 24 h. While others have shown that brine solutions up to 70 ppt are ineffective at removing D. vexillum at minute time-scales (Rolheiser et al., 2012), for the first time we publish that hour time-scales can be 100% successful. Temporal surface area data highlight the importance of extended monitoring, as some injured (bleached), receding and seemingly destroyed D. vexillum colonies, rebounded after three and five weeks (Fig. 4A). Didemnids are known to undergo diebacks due to cold winter water temperatures and low salinity caused by heavy rainfall or freshwater runoff, but have the ability to recover (Kleeman, 2009). Alternatively, greater tolerance for eradication treatments may occur in colder waters (e.g., D. moseleyi – Katayama and Ikeda, 1987), such as the northern geographic extent of its range in Alaska, implying longer monitoring times to verify successful removal. We found that three weeks was required to confirm D. vexillum mortality for all treatments tested. By extension, only two other published studies (five months – Switzer et al., 2011; five weeks – Rolheiser et al., 2012) monitored for a period sufficient to confirm residual treatment effects. This has important implications for any management or eradication plan for this species and possibly other invasive tunicates, and could mean increased monitoring costs for larger scale eradication attempts. These experiments offer insight for potential in situ eradication of D. vexillum in Sitka, Alaska. The ascidian is restricted in Alaska and only known to occur along a 10 m-band of shoreline in one shallow (