Fish species composition and productivity along a freshwater to saltwater ecosystem gradient in the Cart Creek/Parker River system Mairi Poisson Dickinson College, 28 North College Street, Carlisle, PA, 17013 Abstract Stable isotope analyses of 13C, 15N, and 34S were run on various fish and aquatic predators to infer the habits of migration and diet mixing between stream and river systems. Individuals were collected from seven sites along the stream and river and muscle tissue samples were taken to run stable isotopes. By comparing the isotope numbers using the program R, I was able to develop food webs and histograms detailing diet composition of all collected species. I determined that there was no upstream migration from the Parker River into the upper reaches of Cart Creek, probably due to various physical, physiological, and productivity barriers. Introduction The composition of fish species in stream and river systems can vary greatly within a small range. Migrations and movements between systems occur, feeding habits and diets may differ, and because of this the stable isotopes of the animals will fluctuate. To understand how fish move throughout systems, various researchers have applied multiple techniques including PIT tags, tissue assays, and more recently stable isotope analyses of 13C, 15N, and 34S in the muscle tissue of sampled species (Hobson 1999). Stable isotope analysis has been used specifically for determining migration and origins of fish and other species (Hesslein et al. 1991; Cunjak et al. 2004). I focused on 13C and 15N isotopes in my study to understand how fish in a stream and river system in Newburyport, MA move up and downstream. Cart Creek is one of many tributaries of the Parker River, which is influenced by the tides and salt water from the Plum Island Sound (Figure 1). These lotic systems provided the perfect sites for sampling fish and other predatory aquatic animals in both freshwater and saltwater systems. Of the seven sample sites, four were within the freshwater streams of Cart Creek, one was in the downstream freshwater marsh of Cart Creek, and two were in the tidally influenced Parker River (Figure 2). I hypothesized that certain species would move freely between these two systems. Creeks typically act as breeding grounds for anadromous fish, with adults returning to the sea after breeding and the juveniles remaining in the freshwater streams for the summer or the full year (Doucett et al. 1999). Other fish species are capable of migrating within a certain range of salinity to utilize multiple food sources and niches. Due to this movement, I also expected mixed stable isotopic values that would show a diet consisting of multiple food sources from various sites, both fresh and salt influenced. By utilizing information of stable isotopes on various fish species—some typically freshwater and some that were able to mix—I tested my hypothesis and analyzed whether or not the fish I collected from these systems were indeed migrating and eating off a mixed diet. Methods Cart Creek is a tributary of Parker River that is about 3,200 meters long. Interstate 95 cuts off the creek at its head, and there are periodic beaver dams and debris dams that create barriers and ponds throughout the creek. There is a freshwater marsh near the bottom of the creek, where the stream meets a culvert in the road. The Cart Creek eventually flows into the Parker River at the river’s midway point. The Parker River is about 14,000 meters long and connects to the Plum Island Sound, which flows into the Atlantic Ocean. The four upper sites in Cart Creek were chosen to show variation between the beaver pond and stream in terms of composition of fish species. Site 1, a large beaver pond near the top of Cart Creek, is about 400 meters from site 2, a site in the lower part of the stream. The distances between sites 2 and 3, and 3 and 4 are about 170 meters. Site 3 is another beaver pond, smaller than the first, and site 4 was another stream site following the pond. I left three traps at each site for one week, and checked them after a week had passed. The remaining four sites, in the lower Cart Creek and the Parker River, were chosen to examine saltwater influence and to take note of any possible migrations. Site 5, a freshwater marsh, is about 750 meters from site 4, and about 950 meters from where the creek flows into the Parker River. Site 6, the first Parker River site, is about 1,000 meters from the connection of the creek to the river, and about 1,700 meters upstream of the last site. The final collection site in the Parker River, Site 7, is about 7,500 meters from where the river flows into the Plum Island Sound. I left traps in these sites for three days, checking each daily, collecting if there were fish, and returning the traps to the water. All of these collections were done in early November. Upon collection, I brought the fish and other organisms back to the lab and killed them using a clove oil-water mixture, a swift and painless way to die. Each individual was measured in centimeters and massed in grams. I then dissected each fish and collected muscle samples. All individuals from the same site and species were considered of similar dietary habits, so I combined muscle samples into a collective scintillation vial. For example, all 40 threespine sticklebacks were sampled for muscle tissue, and all of these samples were put in the same vial. Stomach samples were taken from representatives of each species. Muscle samples were dried for a 48-hour period, and then I ground up the dried muscle using mortar and pestle. Between each sample, I cleaned the mortar and pestle using kimwipes so as not to cross-contaminate my samples. Afterwards, the samples were desiccated overnight. Stable isotope analyses of carbon and nitrogen were done on all samples (Tieszen 1983). I used the δ13C, δ15N, and 34S results to construct an idea of the food webs in these systems, utilizing the software RStudio (Table 1; Table 2). Along with my collaborators examining benthic species and terrestrial species, we were able to create a food web that considered all parts of these beaver affected areas (Table 3). Separately, I utilized data from the decadal TIDE project to help me understand the food web of the salt water systems (Table 4). I ran individual histograms using a Bayesian mixing model, Stable Isotope Analysis in R (SIAR), to specify the percent of contribution of each source to the tissue of each animal (Inger 2008). Results Of the 81 fish and other species collected, 11 individuals were from the first four upper creek sites, 65 were collected from the freshwater marsh of site five, and 5 were found in the Parker River sites (Table 5). Based off of the creek sources, fish species collected from site 1 (golden shiners, sunfish) relied mostly on the algae-based food web, as most of their muscle isotopes correlated with the range of the algae isotopes (Table 6). Non-fish species, like the crayfish and frog, had mixed diets with different compositions. The predaceous diving beetle, for example, had isotopic values that most correlated with both grasses and algae, nearly equally, based on the percentages of the primary production that correlate with the muscle tissue of the animal (Table 6). The river had different sources from the creek, with species drawing from many different food webs. The one predatory fish found in the Parker River, the white perch, drew mostly from the spartina food web (Table 7). Spartina alterniflora and Spartina patens are two grassy species that are most abundant in saltmarsh systems. The shrimp found in the lower Parker River (site 7) also drew from the spartina food web, whereas the shrimp found in site 6 drew mostly from the particulate organic matter food web. Particulate organic matter is the primary production found in the water column, consisting mostly of phytoplankton. The fish species found in the freshwater marsh in Cart Creek (site 5) were mostly based off of the particulate organic matter food web (Table 8). The mummichog, however, drew mostly from the spartina food web. All of these percentages were determined using RStudio, and are based off of the modes of each organism related to each source. Discussion The food webs developed from this research were chosen based upon the locations the species were found as well as their correlations with isotopes based on sources from the creek and river systems. Because of this, I saw a wide range of diets and was able to better understand the movements and migrations, or lack-thereof, of these species. I expected to see movement by certain species up and down the river and into and out of the creek, whereas others I expected to be more residential in the area they were found. Most of the species found within Cart Creek I expected to be residential, like the sunfish found in big pond of site 1, and the crayfish and predaceous diving beetles found in the streams. Their diets indicated that they were indeed centering their diets on the terrestrial-based creek sources. The histograms generated using the program RStudio can tell us how much of the primary production source contributes to the muscle tissue of each organism (Figure 3; Figure 4). These sources are “terrestrial-based” because there are influences of litter and grasses due to flooding of land areas around the ponds, as the beavers have built dams in these sites. The species found in the Parker River had isotopic values that correlated most with spartina or particulate organic matter. The white perch, for example, had muscle tissue that was 66.9% correlated with spartina isotopes. This indicates that the perch was resident in the river, and was only moving within this saltmarsh system. I found one species in multiple sites and expected to see a mixed-diet based on both the river and creek food webs because of this. The golden shiner was found in both site 1 and site 5, however their diets were not comparable. The site 1 shiner had an algae based diet, whereas the site 5 shiner had a diet centered mostly on particulate organic matter and typha (Table 6; Table 8). These values indicated that the golden shiner from site 5 was resident in the freshwater marsh, even though this species is clearly capable of living in the upper parts of the creek. Other species found in the freshwater marsh of Cart Creek (site 5) had diets that were clearly based off of this system and did not mix with the upper creek. The threespine stickleback and ninespine stickleback were good examples of this. Particulate organic matter composed the highest percentage of each of these species’ muscle tissues (Table 8). Spartina was the second highest contributor, which shows that these species may have migrated into the freshwater marsh from the lower parts of the Parker River, but have not traveled further into Cart Creek as would be indicated by a high percentage of the terrestrial-based sources in the diet. The mummichog, also found in site 5, is a hardy fish species that is capable of moving between many different systems. They are euryhaline, meaning they can survive in fresh and salt water systems (Able 2012). The percent composition of their muscle tissue showed that 61% of their diet was based off of the spartina food web. As stated before, spartina is a saltmarsh primary producer, but the mummichogs were found in the freshwater marsh of Cart Creek. This means that the mummichogs moved upstream from the Parker River into Cart Creek and could potentially move upward of this site. However, based on their isotopes and where the mummichogs were collected, these fish did not move further upstream than site 5, and were blocked by some barrier. There are certain barriers in place preventing or inhibiting movement of fish up and down the river and into the creek. One of these barriers might be productivity, that the river system and the freshwater marsh are more productive than Cart Creek. There also might be a physiological barrier based on the salinities of the saltmarsh compared to the freshwater creek. Certain species might be less tolerant of the freshwater and are unable to travel into the creek due to the lower salinities. The beaver dams and debris dams may act as a physical barrier, blocking the stream with branches and litter. This research was done in late autumn and many fish species may have already migrated out of Cart Creek and habituated to the sites at which they were collected. However, based on the isotopic values, I can infer that most species were resident in the sites they were found and may have only migrated upstream as far as the freshwater marsh of site 5. Stable isotope analyses of the muscle tissue of collected species were a good indicator of the potential movement of these organisms, and demonstrated the trend that there was no migration into Cart Creek. Acknowledgments I would like to thank my mentor, Jimmy Nelson, for helping me along the path to realizing this project. Thank you to my collaborators, Delaney Gibbs, Jessie Moravek, and Julia McMahon for their ceaseless help in the field and in the lab to create a cohesive group. Thank you to John Schade for his help in the field, as well as his photographs utilized in the presentation. And finally thank you to the Semester in Environmental Science faculty and TAs for their help and guidance throughout the semester and beyond. Literature Cited Able, K. W., D. N. Vivian, G. Petruzzelli, and S. M. Hagan. 2012. Connectivity among salt marsh subhabitats: residency and movements of the mummichog (Fundulus heteroclitus). Euastuaries and Coasts 35: 743-753. Cunjak, R. A., J. M. Roussel, M. A. Gray, J. P. Dietrich, D. F. Cartwright, K. R. Munkittrick, and T. D. Jardine. 2005. Using stable isotope analysis with telemetry or mark-recapture data to identify fish movement and foraging. Oceologia 144:636-646. Doucett, R. R., W. Hooper, and G. Power. 1999. Identification of anadromous and nonanadromous adult brook trout and their progeny in the Tabusintac River, New Brunswick, by means of multiple-stable-isotope analysis. Transactions of the American Fisheries Society 128:278-288. Hesslein, R. H., M. J. Capel, D. E. Fox, and K. A. Hallard. 1991. Stable isotopes of sulfur, carbon, and nitrogen as indicators of trophic level and fish migration in the lower Mackenzie River Basin, Canada. Canadian Journal of Fisheries and Aquatic Sciences 48:2258-2265. Hobson, K. A. 1999. Tracing origins and migration of wildlife using stable isotopes: a review. Oceologia 120:314-326. Inger, R., A. L. Jackson, A. Parnell, and S. Bearhop. 2008. SIAR: stable isotope analysis in R. R package version 3. Tieszen, L. L., T. W. Boutton, K. G. Tesdahl, and N. A. Slade. 1983. Fractionation and turnover of stable isotopes in animal tissues: Implications for d13C analysis of diet. Oceologia (Berlin) 57:32-37. Figures and Tables Figure 1. The Parker River and Cart Creek in relation to the Plum Island Sound. Figure 2. Sites 1-7 along Cart Creek and the Parker River in Newburyport, MA. Figure 3. Histogram of the composition of the muscle tissue of sunfish found in site 1, developed from RStudio. Figure 4. Histogram of the composition of the muscle tissue of the crayfish found in site 2, developed from RStudio. Table 1. Stable isotopes of 15N and 13C of the freshwater creek consumers, numbers used in RStudio to develop histograms and plots. Species Site 15N 13C Golden Shiner 1 8.2 -34 Sunfish 1 8.7 -33.4 Tadpole 1 4.6 -30.1 PDB 1 5.9 -31.1 Crayfish 2 7.3 -29.9 Frog 2 4.2 -29.1 PDB 3sp stb 9sp stb Golden Shiner 3 5 5 5 9.8 13.6 13.4 12 -30.1 -24.6 -25.1 -27.6 Table 2. Stable isotopes of 15N and 13C of the tidally influenced creek and river consumers, numbers used in RStudio to develop histograms and plots. Species Site 15N 13C 3sp stb 5 13.6 -24.6 9sp stb 5 13.4 -25.1 Mummichog 5 10.1 -19.4 Shrimp 5 11 -22.3 Golden 5 12 -27.6 Shiner Crab 5 10 -23.3 Shrimp 6 11.3 -22.8 Perch 6 12.1 -17 Shrimp 7 11.3 -18.5 Table 3. Sources and trophic enrichment factors utilized in RStudio to develop histograms and plots for the individuals collected in freshwater creek sites. Sources Concentrations 15 Source Mean SD N Mean SD SD SD 15 13 13 15 %N %C N C C N 13C Grasses 2.95 1 -31.45 1 1.77 0.2 42.8 2 Litter -2.1 1 -29.7 1 0.77 0.2 48.1 2 Algae 4.95 1 -33.45 1 2.27 0.2 24.1 2 Detritus 0.8 1 -28.4 1 0.91 0.2 34.7 2 Typha 5 1 -26 1 1.5 0.2 45 2 Table 4. Sources and trophic enrichment factors utilized in RStudio to develop histograms and plots for the individuals collected in tidally influenced creek and river sites. Source Concentrations Source Mean SD Mean SD %N SD %C SD 15 15 13 13 15 N N C C N 13C Particulate 5.66 0.98 -30.2 0.67 8 1 48 5 Organic Matter Benthic 3.03 0.62 -15.8 1.35 8 1 48 5 Diatoms Spartina 4.7 0.58 -13.1 0.40 1.5 0.5 45 5 Typha 5 1 -26 1 1.5 0.2 45 2 Table 5. All species from all sites, with common and scientific names, and number of individuals collected. Site Common Name Scientific Name Shorthand Number 1 Golden shiner Notemigonus crysoleucas Shiner 1 1 Sunfish Centrarchidae Sunfish 5 1 Green frog tadpole Lithobates clamitans Tadpole 1 1 Predaceous diving beetle Cybister fimbriolatus PDB 1 2 Green frog Lithobates clamitans Frog 1 2 Crayfish Orconectes virilis Crayfish 1 3 Predaceous diving beetle Cybister fimbriolatus PDB 1 5 Threespine stickleback Gasterosteus aculeatus 3sp stb 40 5 Ninespine stickleback Pungitius pungitius 9sp stb 17 5 Mummichog Fundulus heteroclitus Mummichog 3 5 Sand shrimp Crangon septemspinosa Shrimp 2 5 Golden shiner Notemigonus crysoleucas Shiner 1 5 Green crab Carcinas maenas Crab 2 6 White perch Morone americana Perch 1 6 Sand shrimp Crangon septemspinosa Shrimp 2 7 Sand shrimp Crangon septemspinosa Shrimp 1 7 Green crab Carcinas maenas Crab 1 Table 6. Freshwater creek species and percentage of contribution of primary production source to muscle tissue based off of the modes developed from RStudio. Species Site % Grasses % Litter % Algae % Detritus % Typha Golden Shiner 1 1.2 0.9 1 0.5 96.5 Sunfish 1 1.4 0.9 1 0.8 95.9 Tadpole 1 31.9 26.5 3.6 1.2 36.9 PDB 1 46.1 3.2 1.9 1.1 47.8 Crayfish 2 39.4 4 4.4 2.2 50 Frog 2 31 6.8 25.9 1.5 34.9 PDB 3 5.3 2.6 2.7 3.1 86.3 3sp stb 5 0.9 1 0.9 1.1 96.1 9sp stb 5 1 1 1 1.3 95.8 Golden Shiner 5 1.1 1.4 1.8 31.1 64.6 Table 7. Tidally influenced creek and river species and percentages of contribution of primary production source to muscle tissue based off of the modes developed from RStudio. Species Site % POM % BD % Spartina % Typha 3sp stb 5 0.9 18.8 4.2 76 9sp stb 5 1 16.1 4.7 78.2 Mummichog 5 41.7 2.3 5.6 50.4 Shrimp 5 1.5 28.3 25.1 45.1 Golden Shiner Crab Shrimp Perch Shrimp 5 5 6 6 7 94 44.4 46.2 31.2 38.8 0.9 2.2 1.3 0.6 1.1 1.1 18.6 23.6 66.9 57.6 4 34.7 28.8 1.3 2.5 Table 8. Site 5 fish species and percentages of contribution of primary production source to muscle tissue based off of the modes developed from RStudio. Emboldened percentages indicate the most prominent source in the species’ muscle tissue. Species %Grasses %Litter %Algae %Typha %Detritus %POM %BD %Spartina 3sp stb 1.8 2.1 2.0 2.9 2.7 0.8 25.0 62.7 9sp stb 1.8 2.2 2.1 3.2 2.7 1.0 21.6 65.3 Mummichog 1.5 1.6 2.0 2.4 1.9 27.2 2.3 61.2 Shiner 4.7 5.7 6.5 4.0 3.7 1.0 2.0 72.5