==========================================================================================
Department of Marine Resources: Lobster Research Project Final Report
May 22, 2003
Are We Using Herring to Farm Lobsters?
Project Partners:
The Effect of Herring Bait on Lobster Growth and the
Fate of Discarded Bait on Bottom Habitat
Jon Grabowski and Erika Clesceri, Gulf of Maine Aquarium
Phil Yund, University of Maine at Orono
Carl Wilson, Maine Department of Marine Resources
Matt Weber, F/V Griffen
Phil Poland, F/V Charlene Gail
Mike Myrick
Abstract:
Several hypotheses have been generated to explain recent surges in lobster landings in the Gulf of Maine. Research on lobster population dynamics has focused primarily on early postlarval life stages, so that our current understanding of the importance of herring bait on sublegal lobsters remains limited. In order to assess the affects of herring bait on lobster population dynamics, we sampled diet composition (stomach content analyses), tissue production (stable nitrogen isotope ratio analyses), and growth (mark-recapture experiments) of large (66-83 mm CL) and small (45-65 mm CL) lobsters in seasonally closed sites around Monhegan Island and at sites open to fishing in the summer and fall at the Georges Islands in mid-coast Maine during the summer and fall of 2002.
Herring bones were more prevalent in the diet of lobsters at open sites, and were more prevalent in the summer (independent of size). Smaller lobsters from open sites contained more natural prey (i.e., crabs, mussels, clams, polychaetes, urchins, etc.) than those from open sites. Stable nitrogen isotope ratio analyses indicated that larger lobsters derive 33.7 to 54.8% of their tissue from herring bait. Although our estimates suggest that the smaller size class of lobsters generated only 10.9 to 12.6% of its tissue from herring bait, herring bait may remove competitive pressure and could explain why smaller lobsters foraged more effectively on natural prey at open sites.
A higher percentage of recaptured lobsters had molted at open sites (Georges Is.). Of those that did molt, the carapace length of lobsters from open sites increased 16.2% more than for lobsters at closed sites. Quadrat and trap sampling suggested that lobster densities are similar between sites, but lobsters are larger at closed (Monhegan Is.) sites. Video assays that were conducted to determine the fate of discarded bait indicated that the majority of discarded bait is consumed by lobsters.
Collectively, these results support the lobster industry’s notion that herring bait may be very important for lobster population dynamics in the Gulf of Maine, and therefore, the contribution of herring bait should be considered when developing management policies that affect fishing effort.
Project Overview
Research Impetus:
The effects of herring bait on lobster growth and tissue production have not been quantified in mid-coast Maine previous to this study and therefore have not been included in attempts to model lobster population dynamics. Understanding the contribution of herring bait to lobster population dynamics may provide alternative lobster management actions than those based on lobster mortality without consideration of the increased food supply derived from bait.
Objectives:
1. Determine the proportion of lobster diet and lobster tissue derived from herring bait.
2. Assess the impact of a bait-augmented diet on lobster growth.
3. Quantify the initial fate of discarded herring bait in the benthic community.
Introduction
Commercial fishery landings of the American lobster (Homarus arnericanus) in the Gulf of Maine have risen over the past two decades in spite of intense fishing pressure while other regional fisheries have collapsed or failed to recover to historic levels. Because the lobster fishery currently accounts for a disproportionately large percentage of the total fishery value in the Gulf of Maine, research identifying why landings continue to thrive in spite of such intensive fishing effort is critical to the economy of coastal Maine. The thousands of tons of herring annually used as bait to catch lobsters in coastal Maine has been posited as the mechanism driving recent surges in lobster landings. In addition, herring bait is likely to have had consequences for nearshore benthic communities.
This study used a variety of methodological approaches involving multiple scientific disciplines (ecology and biogeochemistry) to address questions of interest to the scientific and fishing communities of Maine in addition to the coastal managers charged with regulating the state’s fishing industries. As awareness has grown that population dynamics of different commercially harvested species are likely to be linked, these possible connections have drawn support for multi-species or ecosystem based approaches to management (Jennings and Kaiser 1998, Botsford et at. 1997). Collaborations among scientists and lobster fishermen in this project have proven successful and increased both the efficiency of research by creating a much more diverse research team with a broader knowledge base and dissemination of results to the potential stakeholders.
White the consequences of the use of herring bait on herring population dynamics (or at least fishing pressures) is fairly clear, the flip side of this interaction has received substantially less attention, although equally important. There are two scientific questions that can be asked to investigate the link among lobster populations, herring populations and the nearshore bottom habitat and organisms. First, has growth of the lobster population been fueled at least in part by the consumption of bait? Herring bait enters the lobster diet in at least three ways: a) consumption by undersize lobsters, b) consumption by adults that subsequently escape traps, and c) consumption of discarded bait. The possible effects of bait consumption on lobster growth and survival have not been quantified, though such effects are widely anticipated by both fishermen and scientists. Second, what are the consequences of relocating that much herring biomass into coastal waters? If discarded bait is not eaten by lobsters, what is its fate, and what are the likely effects on coastal bottom habitats and ecosystems? What are there consequences for coastal groundfish communities? We evaluated by how much herring (Clupea harengus) bait contributes to the production of lobster biornass by quantifying the diet composition and growth of lobsters in areas with vs. without bait. We also quantified the fate of discarded herring bait in shallow nearshore waters in Maine.
Methods
Study site. We conducted investigations in mid-coast Maine near Monhegan Island because the lobster fishery seasonally closes from the end of May until early December. Sites around Monhegan Is.
were compared to sites around the Georges Islands where fishing occurs in the summer and fall. Subtidal (10-15 m depth) sites were selected for lobster dietary, growth, and density analyses. Sites were selected based on bottom type. In particular, we chose sites that contained a mixture of mud/shell bottom and loose cobble rock habitat after conferring with R. Steneck.
1. Determine the proportion of lobster diet and lobster tissue derivedftom herring bait. To determine the prevalence of herring bait in the diet of lobsters and potential changes in lobster diet composition as a consequence of herring bait, stomach content sampling of lobsters was conducted in June 2002 at Monhegan and Georges Is sites. This initial sampling period was scheduled directly after the Monhegan Is. lobster fishery closed on May 30, 2002. Stomach content sampling was conducted in June to determine if lobster diets vary between open and closed areas directly after the closure occurs. Sampling involved collection of ~50 lobsters from each of two size classes (small 45-65 mm carapace length [CL]; large 66-83 mm CL) at each of two sites around Monhegan Is. and the Georges Is. Stomach contents and tissue samples were removed and preserved for analysis in the laboratory. Total stomach contents were weighed, and individual items were identified, weighed, and enumerated. Sampling was also conducted in October to detect seasonal differences in diet composition.
Nitrogen Isotope Ratios
Stable isotope ratios of nitrogen are effective natural tracers of the flow of organic matter in ecosystems (Frye and Sherr 1984, Owens 1987, Peterson and Howarth 1987, Wada et al. 1991, Robinson 2001). Of the nitrogen atoms on earth, approximately 99.6337% are the “normal” ‘4N; the remaining 0.3363% are a heavier 15N form (Robinson 2001). Biogeochemical processes alter the ratio of these two isotopes, and the relative abundance of the heavier isotope is expressed as
on a parts per thousand (%o) scale. We conducted an analysis of the isotope signature of the natural prey of lobsters at our study sites (as reviewed by Lawton and Lavalli 1995) vs. herring. Our results indicate a differential ranging from 2.3 to 5.5%o between herring and three common elements of lobster diet (Table 1). These ~15N differences will propagate up the food chain and should be reflected in lobster tissue.
Table 1. Nitrogen isotope ratios (%o) for natural prey and herring from our study sites in the Gulf of Maine. Values are the mean ± 1 SD for 4-6 replicate individuals.
The likely mechanism for these differences in nuclear isotope ratios for herring vs. other lobster prey items is a process called trophic transfer enrichment (Ehleringer et al. 1986). When processing food, animals excrete 14N disproportionately more than the heavier isotope ‘5N, so that subsequent enrichment of ‘5N in tissues results in heavier nitrogen isotope ratios for these animals relative to their prey. The general expectation is that ~ should be enriched by about 3.4 %o in each transfer between trophic levels (Eggers and Jones 2000); therefore, our preliminary results suggest that herring are about one to one and a half trophic levels above typical lobster prey. Information on the feeding biology of these prey species supports that contention: sea urchins and mussels are herbivores, while herring feed on herbivorous (when juveniles) and carnivorous (when adults) copepods. Because lobsters are a trophic level above their prey (by definition), the absolute value of 6’5N for lobster will be higher than the prey 6’5N. Nevertheless, the relative difference in lobster 6’5N between baited and un-baited areas due to consumption of herring should be easily detectable if herring constitute a major portion of lobster diet. Nitrogen isotope ratios offer a potential advantage over analysis of stomach contents because they measure the integrated assimilated diet (Wainright et al. 1993), which is a function of both average diet throughout the lifespan and actual assimilation (as opposed to simple intake). The percentage of assimilated diet attributable to herring can be calculated from the differences in 6’5N between baited and unbaited sites or over time periods if fishing pressure changes (e.g., Monhegan Is. in the late spring [open] vs. in the fall [closed]).
Lobster tissues were sampled for nitrogen stable isotope ratio analysis in June and October at Monhegan and Georges Is. sites (tissues were sampled from lobsters used for gut content analyses). Lobster sampling was conducted in June directly after the Monhegan Is. fishery closed so that if herring is important to the diet of lobsters, it should be apparent in the tissues after several months of fishing. Conversely, we sampled in October to capture lobster tissue values several months after the fishery was closed in order to determine the isotope values of lobsters that have been feeding on natural prey in the absence of herring for the past several months. Therefore, if herring is important to the diet of lobsters in the Gulf of Maine, maximum contrasts in isotopic signatures between open and closed sites should be apparent by October, and between Monhegan in June (recently open) vs. October (closed). Different lobster tissue types were sampled to characterize the range of tissue turnover rates common within lobsters. For instance, heart muscle (slow turnover), abdominal muscle (intermediate turnover), and midgut gland (quick turnover) tissues were preserved for stable isotope analysis. Samples were freeze dried and analyzed using an elemental analyzer coupled to a mass spectrometer at the University of North Carolina at Chapel Hill.
2. Assess the impact of a bait-augmented diet on lobster growth.
In order to assess the effects of herring bait on lobster growth, we conducted mark-recapture experiments at seasonally closed (Monhegan Island) and open sites (Georges Islands) in 2002. Lobsters were tagged at the beginning of July 2002 prior to initial summer molting, and recapture of lobsters was initiated in October 2002. Lobsters were tagged using streamer tags (FIoy Tag Co. - FTSL-73), which were implanted in the muscle tissue directly between the carapace and abdomen where tag retention through molt cycles is high. We tagged a total of 3117 lobsters in July at Monhegan Island and Georges Islands (-~800 lobsters were measured, sexed, tagged and released at two sub-sites in both closed and open sites). Recapturing lobsters initially involved diver collection and trapping by three industry members using modified traps. In order to bolster recovery of tagged lobsters, we initiated a tag reward through the lobster fishing industry (Monhegan, Cushing, Friendship and Port Clyde) starting in December 2002. We quantified the percentage of recaptured lobsters that molted and percent growth of those that molted for seasonally closed vs. open sites. The effect of herring bait presence on lobster growth was analyzed using an unpaired t-test.
Lobster density sampling was conducted in summer (August) and fall (October) with 1 -in2 quadrat samples, and was conducted again in October. At each site, ten quadrat samples were conducted randomly in each of two habitats (mud bottom vs. rock-cobble), and lobster density, size, and sex ratio were quantified within each quadrat sample. Relative abundance of lobsters at each size stage was quantified at both open and closed sites in the summer (July) and fall (October-November) of 2002 using modified traps. Traps were baited and depth, location, and deployment time were recorded during trap sampling.
3. Quantify the initial fate of discarded herring bait in the benthic community. To quantify the initial fate of discarded bait on benthic communities within the shallow waters of the Gulf of Maine, video assays of discarded herring bait were conducted in late August and early September. In particular, herring bait pieces (ten per deployment) were attached to a grid on the bottom of a metal frame and lowered to the sea floor in 10-15 in depth. Video footage of the grid permitted determination of the fate of each bait piece so that we were able to quantify the proportion of discarded bait consumed by lobsters, crabs and other benthic organisms. Bait trays were filmed for one hour or until all ten bait pieces had been consumed, and we conducted multiple bait tray deployments.
Results
1. Determine the proportion of lobster diet and lobster tissue derived from herring bait. The proportion of lobsters with herring in their stomach was 2 to 3 times higher at open (Georges Islands) than at seasonally closed sites (Monhegan Island) for both large and small lobsters (Figure 1). Herring bait was twice as prevalent during sampling in June than in October regardless of site or size. Lobsters from open sites had up to 7 times more herring (g/stomach) in their diet than did lobsters from closed sites. For smaller lobsters, stomach fullness of lobsters from open sites was greater than that of lobsters at seasonally closed sites (Figure 2). Stomach fullness of larger lobsters did not vary between sites in June, whereas stomach fullness of lobsters from closed sites was greater than at open sites in October. Small lobsters from open sites consumed more natural prey (g/stomach) than small lobsters from closed sites (figure 3). Larger lobsters from open sites consumed more natural prey in June, whereas natural prey was more prevalent in the diet of larger lobsters at seasonally closed sites in October.
Stable nitrogen isotope ratio values were lower for small lobsters from Monhegan Is. than for those at Georges Is. during either sampling period (Figure 4a). Isotope values for small lobsters in June were generally higher than isotope values of lobsters in October regardless of site. For large lobsters, once again isotope values of lobsters from Monhegan Is. were lower than values of lobsters from the Georges Is. during either sampling period (Figure 4b). However, the difference between sites in isotope values of large lobsters was greatest in the fall, with isotope values at open sites (Georges Is.) much greater than seasonally closed sites (Monhegan Is.). Although isotope values of large lobsters in October were generally lower than lobster isotope values in June, the magnitude of this effect was much greater at Monhegan Is.
In order to determine the proportion of lobster production derived from herring bait, we first estimated a fractionation coefficient (measure of trophic transfer enrichment) for nitrogen isotopes because animals tend to disproportionately retain heavier isotopes and have heavier nitrogen isotope values than their dietary members. For each size class, we calculated a fractionation coefficient by subtracting the average value of the natural prey from Monhegan lobsters that were sampled in the fall (Figure 5). Monhegan lobsters sampled in the fall have not had access to herring bait for several months and should have tissue signatures indicative of a diet of natural prey. Because values for natural prey vary (i.e., crab values tended to be slightly higher than urchins, mollusks, etc.) and previous dietary studies have suggested that variability exists in the diet composition of lobsters, we calculated fractionation coefficients for a range of lobster diets. Using dietary information from this study and previous investigations (Elner and Jamieson 1979, Jarnieson et al. 1981, Elner and Campbell 1987, Lawton 1987, Hudon and Lamarche 1989), we determined that crabs account for up to 50% of the diet of lobsters and that crabs, mollusks (i.e., mussels, clams and small scallops) and urchins are the three most important components of the diet. Because crabs tend to account for a larger proportion of the diet of larger lobsters and urchin populations have been widely reduced at our study sites from harvesting efforts, crabs probably account for a greater proportion of the diet of larger lobsters at Monhegan Is.
For larger lobsters, fractionation coefficients varied from 2.1 %~ (50% crab diet) to 0.8%~ (100% crab diet), whereas fractionation coefficients varied from 2.9%~ (crab absent from diet) to 2.2 (50% crab diet). For each size class, the fractionation coefficient was then added to the value of herring to estimate the value of a lobster diet that is comprised solely of herring. We then used this estimate (100% herring diet) and the isotope values for Monhegan lobsters sampled in October (100% natural prey diet) to determine the importance of herring in the diet of lobsters in mid-coast Maine. In particular, we compared these isotope values to Monhegan lobsters from June and calculated the relative proportion of lobster tissue derived by herring bait (Table 6). Utilizing this method for each natural prey diet discussed above, we calculated that herring bait is responsible for deriving 33.7-54.8% of the tissue production of large lobsters and 10.9-12.6% of smaller lobsters.
2. Assess the impact of a bait-augmented diet on lobster growth. We tagged a total of 2257 large and 860 small lobsters (Table 2). We tagged a greater proportion of large lobsters because small lobsters were much more difficult to capture, especially at Monhegan Is. Lobster sex information was determined from tagged lobsters because we sexed a larger volume of lobsters while marking lobsters than during any other sampling technique. Female lobsters always accounted for greater than 50% regardless of size or site. Female lobsters generally accounted for ~60% of the total population, except for small lobsters at Monhegan Is., which accounted for 51.3%.
Table 2. Breakdown of total ~ of lobsters tagged, average size and % female for each size class at each site.
Extensive diving and modified trapping in October 2002 resulted in the recovery of only 10 tagged lobsters. Industry participation from Monhegan, Cushing, Friendship and Port Clyde resulted in the recapture of an additional 47 lobsters between November 2002 and January 2003. Fifty-two of the 57 lobsters recaptured were originally in the small size category (45-65 mm CL); therefore, molt frequency and growth results are only presented for large lobsters (66-83 mi-n CL). In order to tag the largest possible volume of lobsters as quickly as possible at each site in July, we tagged ‘soft’ and ‘hard’ shell lobsters. Because soft shell lobsters have molted more recently, they were less likely to have molted by the fall/winter. Hence, we excluded soft shell lobsters from our analysis of the proportion of lobsters that had molted at each site. Of those recaptured that were hard shell lobsters when tagged, 67.8% of Monhegan Is. lobsters vs. 88.9% from Georges Is. had molted (Figure 7a). Differences in growth increments between soft and hard shell lobsters did not vaiy within each site, so we included both soft and hard lobsters in our analysis of lobster growth. Of the recaptured lobsters that had molted, Georges Is. lobsters grew 16.2% (measured as the percent gain in carapace length) than lobsters from Monhegan Is (Figure 7b).
Lobster densities did not differ between Monhegan Is. and Georges Is. for either habitat type in the summer (Figure 8a). Lobster densities also did not differ between habitats at each site in the summer. In the fall, lobster densities once again did not differ between open and closed sites. However, lobsters were almost completely absent from mud bottom at both sites in the fall. In both sampling periods, lobsters measured during quadrat sampling were substantially larger at Monhegan Is. than at the Georges Is. Trap abundance results also indicated little difference in total lobster abundance between Monhegan Is. and Georges Is. (Figure 9). There was a slight trend (p 0.08) of greater lobster abundances at Georges Is. than at Monhegan Is. in the summer, but abundances were very similar in the fall. Trap sampling also indicated that lobsters were larger on average at Monhegan Is. than at Georges Is (Figure 10). At Monhegan, legal lobsters were more prevalent than any other size category in both the summer and fall. Conversely, lobsters ranging from 76-80 i-nm CL were more abundant than any other size category at Georges Is.
3. Quantify the initial fate of discarded herring bait in the benthic community.
Video assays of discarded bait were used to quantify the initial fate of discarded bait in mid-coast Maine. Discarded bait was rapidly consumed in each assay in under an hour, with lobsters in particular but also crabs (Cancer spp.) consuming the vast majority of bait. Bait trays were also visited by fish (predominately pollock Pollachius virens), but we did not observe any successful foraging events by fish. Attaching bait to a frame might have confounded this experiment by favoring predators with appendages (i.e., lobsters and crabs) that can remove bait from the trays more readily. We also did not observe seal or birds, two common consumers of herring bait, foraging on bait trays during the experiment.
Discussion
Dietary analyses indicated that herring bait has important consequences for the diet of lobsters, for competition between different size classes of lobsters, and potentially for benthic community structure. In particular, our results suggest that up to half of tissue production of larger lobsters in mid-coast Maine may be derived from herring bait. A greater understanding of the lobster diet composition and the fractionation coefficient of lobsters will enhance the resolution of our estimate of the importance of herring bait in deriving lobster tissue production. Although smaller lobsters derived a much smaller proportion of their diet from herring, herring bait could be affecting these lobsters indirectly by removing competitive pressure from larger lobsters for natural prey, explaining why smaller lobsters from fished sites consumed more natural prey than at closed sites. Site-specific differences between Monhegan Is. and Georges Is. could also partially explain why small lobsters from Georges Is. foraged more successfully on natural prey. Stable nitrogen isotope values of both large and small lobsters from June were lower at Monhegan, indicating either that prey communities are different between sites or tissue-turnover rates are more rapid than originally expected. Comparison of the diet composition of lobsters at open and closed sites suggested that herring bait could alter diet composition with consequences for the benthic community. Further analysis of herring bait driven changes in diet composition must be compared to the benthic community at each site to determine how herring bait is influencing benthic communities in coastal Maine.
Because lobsters grew more at open (Georges Is.) than closed (Monhegan Is.) sites, the mark-recapture results suggest that bait augments lobster growth in the Gulf of Maine. The mark-recapture results also indicated that a larger proportion of recaptured lobsters from fished sites had molted. Recaptured lobsters that had not molted probably molted in the winter or early spring when water temperatures are colder. If colder water temperatures adversely affect lobster growth rates, differences in growth rates between open and closed sites might have been even more pronounced (i.e., the effect of herring bait on lobster growth could be even greater). Conducting lobster diet sampling and mark-recapture experiments in open and closed sites a) from additional regions of the Gulf of Maine and b) across multiple years will increase the confidence of these findings and help managers determine the potential implications of management policies (e.g., lobster pot limits, exit-to-entry ratios, etc.) for lobster population dynamics in the Gulf of Maine.
Lobster density (from quadrat sampling)/abundance (from trap sampling) results both suggested that lobster densities/abundances do not differ between open and closed sites. Quadrat sampling indicated that habitat use was dependent on season. Specifically, lobsters were equally distributed between cobble bottom and mud/shell hash bottom in the summer, but were nonexistent in mud/shell hash bottom habitats in the fall. Lobsters may be more vulnerable to predation in the fall because of seasonal patterns in molting frequency (i.e., if a higher proportion of lobsters molted in early fall than in late early summer, a greater proportion of lobsters would have recently molted and hence would be more vulnerable to predation) or abundances of lobster predators. Lobsters may also be responding to seasonal patterns in reproductive and migratory behavior. Both sampling methods indicated that lobsters were larger on average at closed sites at Monhegan Is. than at fished sites in Georges Is. In particular, smaller, sublegal lobsters were caught in greater abundances in fished waters at Georges Is., whereas large sublegals and legal sized lobsters were caught in greater proportions in Monhegan Is. This difference could be a consequence of removal of adult lobsters from fished areas, a result of larger lobsters deterring smaller ones from entering pots, or differences in size-frequencies that are independent of fishing activities. Anecdotal evidence suggests that lobster populations that are further offshore tend to be slightly larger than inshore populations. Because Monhegan Is. is further offshore than Georges Is., this fishery independent factor could partially explain why Monhegan Is. lobsters were larger. In addition to consuming bait in the traps, lobsters were the primary consumer of discarded hening bait. However, this experiment needs to be conducted again at a wide range of sites to increase our understanding of the fate of discarded bait.
Acknowledgements
Successful completion of sampling efforts to achieve the proposed goals has hinged upon collaboration among scientists, government officials, and fishermen. Special thanks to Matt Weber, Phil Poland, and Mike Myrick, whose field knowledge and vessel operation skills have been integral to the successful completion of this research. We are greatly appreciative to the lobster fishing communities from Monhegan, Gushing, Friendship and Port Clyde for reporting recoveiy of tagged lobsters. Thanks to Robert Russel and to the Maine-Department of Marine Resources for the use of video equipment to determine the fate of discarded bait. Thanks to Brian Tarbox and Chris Heinig for loaning us modified traps to collect sublegal lobsters. The Gulf of Maine Aquarium facilitated this project by bringing together this research team and administering the grant (thanks in particular to Laura Taylor Singer, Don Perkins, and Amy Smith). This research project was conducted out of the Darling Marine Center. Funding for this project was provided by the Northeast Consortium, The Davis Foundation, The Lobster Advisory Council, and The Gulf of Maine Aquarium.
Submitted by:
Jonathan Grabowski
Gulf of Maine Aquarium
P.O. Box 7549
Portland, Maine 04112
END
=========================================