Electronic Supplementary Material For s3

Electronic Supplementary Material for:

Role of invasive green crabs on the food web of an intertidal sand flat determined from field observations and a dynamic simulation model

By: Melisa C. Wong1* and Michael Dowd2

1Fisheries and Oceans Canada, Bedford Institute of Oceanography, Dartmouth NS, B2Y 4A2

2Department of Mathematics and Statistics, Dalhousie University, Halifax NS

Methods for data collection for model input

Sampling stations and dates. Macrobenthic infauna (excluding soft shell clams), benthic microalgae, and sediment particulate organic carbon (POC) were sampled at 22 randomly assigned stations spaced at least 20 m apart on the intertidal sand flat. Soft shell clams were sampled at 84 randomly assigned stations spaced at least 10 m apart. Meiofauna samples were widely spaced and taken at 5 of these stations. Suspended POC and phytoplankton were sampled at 3 stations in the main channel on the north side of the intertidal flat (Fig. 1). Zooplankton tows (n = 4) were conducted from the bridge connecting the inner lagoon to the outer lagoon during afternoon ebbing and flooding tides. Fish and epibenthic invertebrates were sampled by seining three 100 m long randomly placed transects on the intertidal flat during one high tide cycle. Birds were enumerated from several fixed locations on the intertidal sand flat. Benthic sampling took place June to July 2008. Sampling of fish, birds, and water column parameters was conducted in the same months in 2009. Data of migrating birds were further refined monthly from June to November in 2010. Logistical constraints did not allow all data to be sampled within the same year. Following is a detailed description of sampling methods used for specific food web components:

Detritus. Sediment POC was sampled using a core measuring 0.03m diameter by 0.05m deep. Samples were frozen until processing. Carbon mass (CM) was determined by drying each sample at 60º C for 48 h, weighing, combusting at 500º C for 7h, and then reweighing. Mass of organic matter (OM) was calculated as total dry mass – mass after combustion, and converted to CM using CM/OM = 0.43 (Luczak et al. 1997).

Suspended POC was determined from water samples taken at 0.5 m depth. A known volume of water was filtered onto pre-weighed glass-fibre filters (0.7 μm). Filters were dried at 60º C for 24 h, weighed, combusted at 450º C for 4 h, and the reweighed. Conversion to CM was determined as for sediment POC.

Sources of DOC (dissolved organic carbon) were calculated as 25% of phytoplankton net primary production (e.g., Anderson and Zeutschel 1970; Choi 1972; as reviewed by Valiela 1995) and 1.5% of seagrass net primary production (Wong, unpub. data; Penhale and Smith 1977). We included DOC from seagrass (Zostera marina) from beds in the inner lagoon because they are highly connected to the intertidal flat by water flow. Furthermore, we also assumed DOC was derived from the epiphytic algae on the seagrass at 50% of that from the seagrass itself, and that DOC from the external ocean was 50% of internal sources (as reviewed by Valiela 1995). Total DOC was determined by summing contributions from the various sources.

Primary producers. Benthic microalgae in the sediments were sampled from cores measuring 0.01 m diameter by 0.01 m deep. Samples were frozen in the dark until processed. To extract chlorophyll-a (Chl-a), samples were vortexed for 1 min, and then left overnight in 90% acetone. Samples were then further vortexed for 1 min to ensure full extraction, and centrifuged for 30 min at 3250 rpm. Chl-a concentration in the supernatant was determined fluorometrically using the acidification technique (Holm-Hansen et al. 1965). Chl-a concentration was converted to CM using CM/Chl-a = 49.9 (de Jonge 1980). Net primary production of benthic microalgae was estimated from the relationship between mg Chl-a and production in Colijin and deJonge (1984).

Water samples to determine biomass of phytoplankton were collected at 0.5 m depth. A known volume of water was filtered onto a glass-fibre filter, and chl-a was extracted overnight using 90% acetone. Chl-a concentration was then determined as for benthic microalgae. Chl-a concentration was converted to CM using CM/Chl-a = 120 (Taylor et al. 1997). Net primary production of phytoplankton was estimated from appropriate photosynthesis-irradiance curves (Kana et al. 1985), using field measurements of light penetration to 0.5 m depth and our measurements of CM.

Heterotrophic bacteria. Bacterioplankton production was estimated as 30% of water column primary production (Cole et al. 1998). Bacterioplankton biomass was estimated as 1/8 of bacterioplankton production (Baird et al. 2004). Production of sediment bacteria was estimated using data of sediment organic matter and water temperature (Sander and Kalff 1993). Biomass of sediment bacteria was estimated using relationships between production and biomass provided by Sander and Kalff (1993). Consumption by bacterioplankton and sediment bacteria was determined assuming a growth efficiency (GE) of 0.4 (Cole et al. 1998) and using the relationships GE = P*Q, and P*Q = (P/B) / (Q/B) (Christensen and Walters 2004) where P = production, Q = consumption, and B = biomass. The proportion of unassimilated food was assumed to be 0.15.

Zooplankton. The zooplankton net was 0.30 m diameter x 1.24 m length with 125 μm mesh, and equipped with a flow meter to determine the volume of water filtered (Omori and Ikeda 1992). Note this mesh size excluded microzooplankton. Wet mass (WM) was determined for each sample and converted to CM using DM/WM = 0.188 and CM/DM = 0.457 (Brey 2001). Zooplankton production and consumption was estimated from the literature and adjusted according to our biomass estimates (Baird et al. 2004). The proportion of unassimilated food was assumed to be 0.40 (Christensen et al. 2008).

Benthic infauna. Meiofauna were sampled using cores measuring 0.03 m diameter by ~0.20 m deep. Cores were fixed in 5% buffered formalin for 2 weeks, and then preserved in 70% ethanol until extraction. Meiofauna were extracted from cores by adding Ludox TM 40 and centrifuging at 1000 rpm for 10 min. Supernatant was poured through a 64 μm sieve, and contents remaining on the sieve were rinsed with distilled water and stored in 70% ethanol until sorting. Meiofauna were sorted under a dissecting microscope aided by rose Bengal stain into broad taxonomic categories (i.e., nematodes, harpacticoids, harpacticoid nauplii, bivalves, ostracods, turbellarians, polychates, oligochates, cladocerans, gastrotrichs, and tanaids). For each sample, the number of individuals per m-2 was determined. Within each sample, the volume of 10 nematodes and harpacticoids (which comprised 95% of the meiofaunal community) was estimated (Danovaro et al. 2002). Volume per individual was converted to WM using an average specific density of 1.13 g cm-3 (Feller and Warwick 1988), WM to DM using DM/WM = 0.25 (Warwick and Gee 1984), and DM to CM using CM/DM = 0.463 (Brey 2001). Average individual CM was multiplied by the appropriate density to determine biomass of nematodes and harpacticoids. Total biomass of the meiofauna community was calculated by summing the biomass of nematodes and harpacticoids, and adding an additional 5% of this biomass to account for the remaining fauna (as determined from previous samples, Wong unpub. data) that were not measured due to logistical constraints. Production of meiofauna was estimated using a production to biomass ratio (P/B) of 10.5 y-1(Danovaro et al. 2002). Consumption was estimated using Q/B = 52 y-1 (Christian and Luczkovich 1999; Baird et al. 2004). Diet composition was obtained from the literature (Moens and Vincx 1997, Leduc et al. 2009; Middelburg 2000; Moens et al. 2002). The proportion of unassimilated food was assumed to be 0.40.

Macroinfauna (with the exception of soft-shell clams, see following paragraph) were sampled using a 0.10 m diameter by 0.12 m deep hand core. Each core sample was gently washed through a 500 μm sieve using sea water. Material remaining on the sieve was fixed in 5% buffered formalin for 2 weeks, and then preserved in 70% ethanol until sorted. Infauna were sorted under a dissecting microscope, identified to species when possible, and enumerated. DM per species for each sample was determined by drying at 60º C for 48 h and weighing. Molluscs were acidified in 10% HCl to remove shells prior to drying. DM was converted to CM using conversion factors from Brey (2001). Production of macroinfauna was determined using ratios estimated using Brey’s (2001) model (v. 4-04), as described in Wong et al. (2011). Consumption of each macroinfauna species was calculated using the equations C = A x 100 / AE and A = P + R, where C = consumption, A = assimilation, AE = assimilation efficiency, P = production, and R = respiration (Baird and Milne 1981). A mean AE of 0.63 was used (Brey 2010), and respiration determined using the empirical relationship as implemented in the spreadsheet provided by Brey (2001) and described in Brey (2010). Diet composition for macroinfauna was obtained from the literature (Fauchald and Jumars 1979; Riisgård 1991; Christian and Luczkovich 1999; Dauer 2000; Baird et al. 2004; McDermott 2005; Riera 2010). The proportion of unassimilated food was assumed to be 0.40.

Soft-shell clams (Mya arenaria) were sampled by Parks Canada during their annual clam survey. Plastic sampling containers measuring 0.38 m diameter by 0.30 m deep were inserted into the sediment (clams bury to a maximum depth of 25 cm, Blundon and Kennedy 1982). All clams were removed by hand from the sampling area and shell length measured. Shell length was converted to DM using the relationship in Moller and Rosenberg (1983). CM, production, consumption, diet composition, and proportion of unassimilated food were determined as described for macroinfauna above.

Epibenthic invertebrates and fish. A beach seine measuring 12 m wide with 0.15 x 0.15 m mesh with a central bag was used to sample epibenthic invertebrates and fish at high tide. At the end of each seine pull, fish and invertebrates were identified to species, counted, and a subset measured. Sand shrimp (Crangon septemspinosa) were measured by carapace length, Green crabs (Carcinus maenas) by carapace width, and fish by total length. Fish data from a previous biological inventory, the trapping program, and personal observations were used to assess and augment the dataset. Using data from Locke et al. (2005), sand shrimp density was increased by 25% to account for smaller shrimp not caught in our seine and by 75% to account for night activity. Green crab density was further informed using data from the crab trapping program conducted by Parks Canada (C. McCarthy unpub. data). Lengths of the shrimp, green crabs, and fish were converted to biomass using relationships from the literature (Corey 1981; Brey 2001; Baeta et al. 2005; FishBase 2010-2011). Total biomass for each species was calculated as mean individual biomass x density. Literature P/B ratios were used to determine production of sand shrimp and green crab (Pihl and Rosenberg 1984; Pihl 1985; Fredette et al. 1990; Heck et al. 1995; Baeta et al. 2005). Consumption rates by sand shrimp and Green crab were determined using empirical relationships as for macroinfauna (Baird and Milne 1981). Production and consumption of fish was determined using empirical relationship from Palomares and Pauly (1998) and using information from fishbase (FishBase 2010-2011). Diet composition of shrimp and fish were obtained from the literature (Pihl and Rosenberg 1984; Taylor and Peck 2004; Feller 2006; FishBase 2010-2011). Diet composition of green crab was from stomach contents analysis of 10 male and 10 female crabs from the field site (see Table 4 below), literature data, and many field observations of crabs consuming Mya arenaria. The proportion of unassimilated food was assumed to be 0.40 (shrimp), 0.30 (Green crabs), and 0.20 (fish) (FishBase 2010-2011).

Birds. Birds observed June through November were considered resident species, while birds observed only in July, August, and September were considered migratory species. Resident birds were included in the model, while migratory birds were an ecosystem perturbation (see main text).

Median individual biomass for each resident bird species was obtained from the literature (Birds of North America online), and multiplied by bird density to determine total CM using DM/WM = 0.33 and CM/DM = 0.4 (King 1974; Grant 1981). Production of each resident bird species was calculated assuming GE = 0.01 (Baird et al. 2004). Consumption of each resident bird species was determined by calculating basal metabolic rate and adjusting by assimilation efficiency and energy content of prey (Scheiffarth and Nehls 1997). The proportion of unassimilated food was assumed to be 0.25 (Castro et al. 1989).

Diet composition of the resident and migrating shorebird diet was determined from the literature (Birds of North America online), by considering prey preferences across habitats and seasons, prey availability at our field site, and from field observations. Total consumption rates of migrating shorebirds were determined using the relationship in Scheiffarth and Nehls (1997), assuming that basal metabolic rate is 8 times that when not migrating (Kvist and Lindström 2003). Total amount of benthic prey consumed by migrating shorebirds was determined from consumption rate data, information on average number of stops made during migration, and the amount of benthic prey available for consumption without driving benthic species to extinction (Birds of North America online; Scheiffarth and Nehls 1997).

REFERENCES

Anderson, G.C. and R.P. Zuetschel. 1970. Release of dissolved organic matter by marine phytoplankton in coastal and offshore areas of the Northeast Pacific Ocean. Limnology and Oceanography 15: 402–407.