Introduction

Tidal flats are gently sloping open areas of soft sediment that undergo regular cycles of inundation and exposure from marine tides (Beninger and Paterson 2018). These productive areas provide a range of ecosystem services (Passarelli et al. 2018), including control of pollutants, shoreline protection, and provision of habitats that support a diversity of species. Tidal flats exhibit spatial heterogeneity on multiple scales (Jesus et al. 2005; Chapman et al. 2010; Gerwing et al. 2015), underpinned by differences in the community composition of intertidal microphytobenthos, driven by a suite of abiotic and biotic factors, including temperature (Kalu et al. 2023), hydrodynamics, and topography (Moccia et al. 2019). The primary productivity of tidal flats from intertidal biofilm passes through the food web to invertebrate and vertebrate consumers, including shorebirds during long-distance migration (Mathot et al. 2018; Hope et al. 2020a, b). As tidal flats remain under threat from multiple stressors, understanding the ecological processes that provide their ecosystem services is a priority for effective conservation (Murray et al. 2019; Mathot et al. 2018).

Migrating shorebirds rely on tidal flats as foraging grounds, and dietary macronutrients (protein, carbohydrate, and lipid) affect the flight performance of migratory birds (Lindström and Piersma 1993; Jenni and Jenni-Eiermann 1998; McWilliams et al. 2004; Guglielmo 2010). Protein can be converted to glucose for fuel, is used in other essential metabolic pathways (Jenni and Jenni-Eiermann 1998), and may play a special role in maintaining water balance during long-distance flights (Landys et al. 2005). Carbohydrate stores are generally only sufficient to fuel short-distance flights, although they can be converted to lipids for storage, whereas lipids are energy dense and the best metabolic fuel for migration (Guglielmo 2010). In general, ∼ 90% of the energy for migratory flight comes from fat (lipid) stores and the remaining 10% from protein (Jenni and Jenni-Eiermann 1998). As diet can determine the flight performance of migrating birds (McWilliams et al. 2004; Guglielmo 2010; Dick et al. 2023), understanding the availability of macronutrients at stopover sites is critical.

Mathot et al. (2018) proposed that the abundance and foraging activity of shorebirds mirror underlying ecological conditions, including temporal and spatial patterns in food supply and community structure within the mudflat ecosystem. Shorebird preferences for particular tidal flats (or areas within tidal flats) may be driven by underlying variation in food availability (Hamilton et al. 2003). Small-bodied shorebirds consume a variety of prey, including marine invertebrates and intertidal biofilm, a thin layer of microalgae, bacteria, and other microorganisms embedded in a mucilaginous matrix composed of extracellular polymeric substances (EPS; Decho 2000; Stal 2010). Intertidal biofilm can account for 45–59% of the shorebird total diet or 50% of their daily energy budget (Kuwae et al. 2012; Hobson et al. 2022). On the Fraser River Estuary, British Columbia, Canada, Jardine et al. (2015) investigated biofilm consumption by Western Sandpiper Calidris mauri and found that sites with the highest proportion of intertidal biofilm in the diet supported larger numbers of birds, suggesting that site use was related to biofilm consumption. Shorebird use may be further associated with spatial variation in the nutritional content of biofilm (Asmus and Bauerfeind 1994; Quinn et al. 2017; Schnurr et al. 2019, 2020; Mogle 2021). In sum, the macronutrient content of biofilm on mudflats may provide a measure of habitat quality tied directly to the nutritional needs of migrating birds.

In this study, we assessed whether the biomass (measured as chlorophyll a), macronutrient content, or both, of intertidal biofilm provide a measure of habitat quality of tidal flats, under the working hypothesis that shorebirds forage at the best quality habitat to meet their requirements for migration. We compared shorebird use and nutritional quality during two migration periods on the Fraser River Estuary at two mudflats known to differ in shorebird use. We predicted that if more nutritious biofilm fosters the best “quality” habitat, then higher macronutrient content should occur at sites with higher use by migratory shorebirds. We also used stable isotopes to assess the relative importance of marine or freshwater inputs to the intertidal biofilm. Carbon (δ13C = 13C/12C) and nitrogen (δ15N = 15N/14N) stable isotope ratios can provide information on the sources of carbon and trophic levels through estuarine food webs, where larger contributions from marine sources are reflected by higher δ13C values (Hobson 1999; Fry 2002; Kristensen et al. 2019; Dias et al. 2023). In addition, we considered how predation risk may differ between the two mudflats by comparing the spatial extent of areas with more relative safety from Peregrine Falcons Falco peregrinus (Pomeroy et al. 2008). These studies are of practical importance in managing tidal flats as dynamic and complex habitats for shorebird conservation, with an emphasis on nutritional ecology of long-distance migration.

Methods

Study Sites

We conducted our study on tidal flats in the Fraser River Estuary on the coast of British Columbia, Canada (Fig. 1). The Fraser River Estuary lies south of the City of Vancouver (49.28° N, 123.12° W), and is an important stopover for migratory birds on the North American Pacific Flyway and a globally significant Important Bird Area (Birdlife International 2022). The study region has a moderate oceanic climate with cool wet springs (mean daily temperatures in April and May range from 9 to 13 °C) and warm dry summers (mean daily temperatures in July to early September range from 15 to 22 °C). The tidal regime on the tidal flats is mixed semi-diurnal with a mean range of 3.1 m. Outflow from the Fraser River fluctuates widely over the year and has a snowmelt-driven discharge regime in the spring, with peak flows typically occurring in June. River discharge has a strong influence on the salinity regime of the intertidal zone, which undergoes a seasonal transition from brackish to nearly fresh water as the outflow increases (Dashtgard 2011; Drever et al. 2023).

Fig. 1
figure 1

Location of Fraser River tidal flats (shown as Intertidal Zone), British Columbia, and our two study sites, Brunswick Point to the south and Iona Foreshore on the north. Panels on the right depict individual sampling stations for biomass and macronutrient content in intertidal biofilm

Full size image

We focused on two sites within the estuary (Brunswick Point and Iona Foreshore) differing in shorebird use (Fig. 1). The first site, Brunswick Point on Roberts Bank, is bracketed between a marsh on the north end and the Deltaport causeway on the south side. The second site, Iona Foreshore, is located on Sturgeon Bank, adjacent to the Vancouver International Airport and a sewage treatment plant, and has two jetties that alter river flow over the mudflats (Levings 1980; Atkins et al. 2016).

The two sites have differing use by shorebirds, with Brunswick Point consistently having higher densities and total numbers of birds relative to Iona Foreshore (Butler 1994, Hemmera Envirochem 2014a, Jardine et al. 2015). Counts of Western Sandpipers at Brunswick Point during northward migration in the spring typically reach peaks of > 150,000 birds and can account for up to 42–64% of their estimated global population (Drever et al. 2014; Canham et al. 2021). Peak counts of Dunlin Calidris alpina typically reach values of > 60,000 birds and can account for up to 30–50% of the flyway population of Pacific dunlin (Drever et al. 2014; Canham et al. 2021). In contrast, peak numbers of both species at Iona Foreshore typically number ~ 10,000 birds during spring migration (Jardine et al. 2015). A similar pattern occurs during the summer southward migration period for Western Sandpipers (the southward migration period of Dunlin occurs later in the fall). Counts of birds and foraging intensity (indexed by fecal dropping counts) during July and August indicate that Brunswick Point receives higher use by Western Sandpipers relative to Iona Foreshore during the southward migration period (Hemmera Envirochem 2014a).

Shorebird Surveys

Shorebird surveys were conducted at Brunswick Point and Iona Foreshore during the two migration periods in 2020 and 2021 to test whether spatial differences in shorebird abundance observed in previous studies held for our study period. We followed protocols outlined in Drever et al. (2014). Experienced observers conducted counts of shorebirds during twelve occasions from July to September 2020, and during nine or ten occasions during April and May 2021 at each site, coinciding with the southward and northward migration periods of western sandpiper, respectively. Observers on shore used binoculars and a spotting telescope to count western sandpiper and dunlin when tide height was approximately 3.5 m, which exposed mudflats to ~ 500 m from shore. The areas surveyed for shorebirds were similarly sized between Brunswick Point (155 ha) and Iona Foreshore (177 ha) study sites. See Supplementary Material for layout of survey areas.

Biofilm Sampling Design

Samples of intertidal biofilm were collected from ten stations in the upper intertidal zone at each of the two study sites during summer 2020 and spring 2021 (Fig. 1). Samples were taken at sites known to be used by shorebirds that lay between elevations of 0 to 0.75-m Geodetic Datum. A boundary was determined between the elevation endpoints, and ArcGIS random site generator (Redlands, CA: Environmental Systems Research Institute, Inc.) was used to select the ten stations within each site (Fig. 1). Distances between individual biofilm sampling locations varied between 50 and 300 m. At the Iona Foreshore site, a further constraint was that five stations were on either side of the Iona Jetty.

The original sampling design focused on capturing site-level differences, with samples repeatedly collected at each individual station over two seasons (although logistical considerations prevented all stations to be sampled each time — see below). Sampling occurred twice in summer 2020 (16 and 17 August, and 6 and 7 September) and twice in spring 2021 (16 and 17 April, and 4 and 5 May). The sample counts were as follows: Brunswick Point northward, 18 samples; Iona Foreshore northward: 18 samples; Brunswick Point southward, 14 samples; Iona Foreshore southward, 15 samples. Based on findings by Mogle (2021) that within a short period, individual day of sampling accounted for little of the variability in biofilm abundance or quality, we included the factor season and pooled across days within seasons in analyses. The quantity and quality of intertidal biofilm can fluctuate among seasons (Jesus et al. 2009; Orvain et al. 2014; Schnurr et al. 2019, 2020), and the spring and summer sampling dates coincided with the southward and northward migration periods of the Western Sandpiper, respectively, through our study region (Butler et al. 1987; Drever et al. 2014). This timing only captures the northward migration of Dunlin, as southward-bound Dunlin typically arrive at the study region later in the fall (October; Lank et al. 2003).

Restrictions to field work occurred in 2020 due to public health measures related to the COVID pandemic. These restrictions resulted in a study design that confounded seasonal (spring vs summer) with annual (2020 vs 2021) differences (we initially planned to conduct all fieldwork in 2020). We acknowledge this disconnect and interpret seasonal differences with due caution.

Intertidal Biofilm Sampling Methods

Intertidal biofilm was collected at each station using a plastic laminate sheet with a 10-cm × 10-cm square opening. The laminate was placed on the surface of the mud, and a paint scraper with 3″ (7.62 cm) steel blade was used to collect the top ∼ 2-mm layer across the opening. Any visible invertebrates (> 1 mm) present during collection were removed from samples with tweezers. Smaller invertebrates (< 1 mm) may have contributed small amounts of biomass and nutrients to each sample, though δ15N signatures suggest that this contribution would have been limited (see “Results”). Each sample was placed into a 120-mL polypropylene container and stored in the dark on dry ice in the field. Samples were maintained at − 20 °C at Environment and Climate Change Canada’s Pacific Environmental Science Centre, North Vancouver, British Columbia, before shipment in dry ice to Mount Allison University, New Brunswick, for laboratory analyses. All biofilm samples were lyophilized (FreeZone 1-l benchtop freeze dryer, Labconco Corporation, Kansas City, MO) and homogenized to allow multiple subsamples for laboratory assays.

Laboratory Analyses

Intertidal Biofilm Biomass (Chlorophyll a)

A subsample of ~ 0.004–0.01 g was taken from each biofilm sample for use in chlorophyll a assays. Because intertidal biofilm is composed of multiple microbial taxa, chlorophyll a content is used as a measure of primary producer biomass across taxa (MacIntyre et al. 1996). Chlorophyll a was extracted from a biofilm subsample in 5 mL of a 3:2 mixture of 90% acetone and dimethyl sulfoxide for 30 min in the dark (Shoaf and Lium 1976), and then chlorophyll a was measured via fluorometry (Turner Trilogy Laboratory Fluorometer, San Jose, California, USA; Welschmeyer 1994) and quantified using a chlorophyll a reference standard (Sigma #96145).

Macronutrient (Lipid, Protein, and Carbohydrate) Content

Protein was extracted from ~ 0.1 to 0.2 g of freeze-dried biofilm samples by bead milling using lysing matrix D bead tubes (MP Biomedicals, Solon, Ohio, USA) and lithium dodecyl sulfate (LDS)–based buffer containing a protease inhibitor (Pefabloc) (Brown et al. 2008). The BioRad DC Protein Assay Kit was used for colorimetric quantitation of protein extract. Although these methods appear to yield accurate measurements of total protein for a diverse range of microalgal samples (Berges et al. 1993; Lourenço et al. 2004; Liefer et al. 2019), protein estimates of biofilm samples were corrected post-quantitation. These methods can have lower colorimetric assay response to protein with sample masses greater than 0.1 g (Mogle 2021), most likely due to excess interfering materials in the sediment (e.g., salts, chelating agents, and reducing agents). Assay response was shown to increase at lower sample masses, giving a higher and presumably more accurate estimate of protein per sample mass, and this relationship between sample mass to assay response was used to correct protein measurement. See Mogle (2021) for details of testing and applied corrections.

Two-stage acid hydrolysis was used to analyze carbohydrates (Pakulski and Benner 1992). About 0.015 g was taken from each biofilm sample for the assay. Carbohydrates were extracted from samples with sulfuric acid and heat, beginning by adding 12 M sulfuric acid to each sample. After resting in darkness at room temperature for 1 h, samples were diluted to 1.2 M sulfuric acid and placed in boiling water for 2 h. After extraction, samples were quantified via a well-established phenol-sulfuric acid method (Dubois et al. 1956) modified for 96-well plates as in Masuko et al. (2005).

Lipids were extracted from biofilm samples using a modified version of the Folch method (Folch et al. 1957). Muffle-furnace combusted or solvent-cleaned glassware was used in all steps. Subsamples of 0.02–0.1 g were taken from each biofilm sample for this assay. In total, 5 mL of Folch (2:1 chloroform/methanol) solvent along with 250 μl of water was used in four rounds of initial extraction and pooled. Samples were disrupted using 30 s of vortex mixing and 2 min of sonication in a sonic bath after each round of extraction. Water was added because our samples were freeze dried, and the Folch method shows greater extraction yield in wet samples (Iverson et al. 2001). Particulate matter was filtered out of the sample using glass microfiber filters (Whatman Grade GF/F) previously combusted in a muffle furnace. Potassium chloride was added to the filtered sample, creating a biphasic system with an upper aqueous phase and lower organic phase (containing extracted lipids). The aqueous phase was discarded. The lower organic phase was evaporated under a gentle stream of nitrogen gas before dissolving the extracted lipids in pure chloroform. Aliquots of the purified lipid extract were quantified via reverse colorimetry using the acid-dichromate method from Pande et al. (1962).

Stable Isotopes

For samples collected in spring 2021 (April and May), we conducted analyses of carbon and nitrogen stable isotopes. These analyses were performed by the EASIL Laboratory at Mount Allison University in Sackville, NB, using an Elementar PyroCube Elemental Analyzer (EA) (Elementar Analysensysteme GmbH, Hanau, Germany) and an Isoprime Precision Isotope Ratio Mass Spectrometer (IRMS) (Elementar UK Ltd, Cheadle, UK). Samples used for carbon analysis were pre-treated with HCl to remove inorganic carbon that may interfere with 13C values. HCl (0.5 M) was added to samples in glass centrifuge tubes, and samples were stirred with glass rods until no more CO2 bubbles were visible. After sitting for approximately an hour, samples were then centrifuged, and liquid was pipetted off. The process was repeated and continued until addition of acid did not lead to CO2 bubble formation.

Values of δ 13C, δ 15N, %C, and %N data were simultaneously generated by combustion analysis of solid samples into CO2 and N2 gas using an elemental analyzer-isotope ratio mass spectrometer (EA-IRMS). A 1.0-mg amount of dry, ground sample was weighed into a tin capsule and carefully encapsulated to remove all air prior to combustion. Certified reference standards for percent composition (mass of element/sample mass) were used for total carbon and total nitrogen. Stable isotopes N and C were also weighed and encapsulated in the same way. Samples were interspersed with at least three replicates of three different certified reference materials purchased from the Reston Stable Isotope Lab of the United States Geological Survey (USGS), Reston, Virginia. These standards were used to normalize data to achieve isotope-ratio-scale-adjustments using the equation below:

$${\delta }^{a}{X}_{\left(Sample\right)}=\left[\left(\frac{{R}_{\left(Sample\right)}}{{R}_{\left(Standard\right)}}\right)-1\right]*1000$$
(1)

where a = the heavier isotope, X = the element of interest (nitrogen or carbon), and R = the ratio of heavy to light isotope. The reported values are expressed relative to international standards where VPDB = Vienna Pee Dee Belemnite, LSVEC = Lithium carbonate, and Air = atmospheric nitrogen. The VPDB-LSVEC scale is normalized such that the δ13C values of NBS 19 calcium carbonate and LSVEC lithium carbonate are + 1.95 ‰ and − 46.6 ‰, respectively (Coplen et al. 2006).

Data Analyses

Site-Level Comparisons

To evaluate spatiotemporal variation in biofilm macronutrient content and biomass relative to site and migration period, we used PRIMER with PERMANOVA (permutational multivariate analysis of variance) add-on (version 6, Clarke and Warwick 2001; Anderson et al. 2008) for univariate and multivariate analyses. Resemblance matrices were constructed using Euclidean distance. Significant PERMANOVA results were followed by PERMDISP tests (permutational analysis of multivariate dispersions) to inform whether differences in multivariate dispersion or univariate variance contributed to the significant result. SIMPER tests (similarity percentages) were conducted to identify the contribution of each response variable to a significant multivariate result. Furthermore, we used estimates of components of variation to assess the relative importance of each source of variation to the multivariate or univariate response.

To visually depict the macronutrient composition across sites and migration periods, a non-metric multi-dimensional scaling (NMDS) plot was created using Euclidian resemblance matrices in PRIMER with PERMANOVA add-on. To analyze biofilm biomass (chlorophyll a content, univariate analysis), we conducted a PERMANOVA with site (two levels: Brunswick Point and Iona Foreshore) and migration period (2 levels: northward and southward) as fixed factors. To analyze biofilm macronutrient content (protein, lipid, and carbohydrate content, multivariate analysis), we conducted a PERMANOVA with site (two levels: Brunswick Point and Iona Foreshore) and migration period (two levels: northward and southward) as fixed factors. Initially, we analyzed the data considering three separate sites (Brunswick Point, Iona North, and Iona South) and found no significant differences between Iona North and South, and so we pooled the data for Iona Foreshore.

Stable Isotopes

We used multivariate analysis of variance (MANOVA, base stats package; R Core Team 2022) to compare stable isotope signatures as a function of site (Brunswick Point/Iona Foreshore) and sample month (April/May 2021).

Biomass (Chl a) and Macronutrient Content

Following our initial appraisal of site-level differences in macronutrients (see “Results”), we examined whether the relationship between biofilm biomass and nutrient content varied between sites and migration periods. We hypothesized that if a difference in “quality” of intertidal biofilm (in contrast to a difference in the “quantity” of intertidal biofilm) accounted for the overall site-level differences in macronutrient content, then the macronutrient content per unit biomass should be higher at Brunswick Point than at Iona Foreshore. Differences in quality of intertidal biofilm may result from differences in taxonomic community composition or the physiological state of component species within the microphytobenthos (Schnurr et al. 2019, 2020). We tested this hypothesis by fitting a series of generalized linear mixed effects models with macronutrient as the response variable, and with explanatory variables including chlorophyll a content, the combination of site/migration period either as two separate 2-level categorical factors, and their interactions with Chl a (see Supplementary Material for model selection details). Unique station ID was included as a random effect. Mixed effects models were fit with the lme4 package (Bates et al. 2015a, b) in R (R Core Team 2022). All statistical tests tested the null hypothesis of no difference between groups and were based on α-level of 0.05.

Results

Shorebird Abundance

Shorebirds were consistently more abundant at Brunswick Point than at Iona Foreshore during both migration periods (Fig. 2). During the southward migration in 2020, an average of 381 Western Sandpipers per survey (SD = 278, n = 12) were counted at Brunswick Point, in contrast to Iona Foreshore with an average of 129 Western Sandpipers per survey (SD = 241, n = 12). No Dunlin were observed at either site during this summer period. Similarly, during the northward migration in 2021, an average of 21,003 Western Sandpipers per survey (SD = 11,426, n = 10) was counted at Brunswick Point, in contrast to an average of 208 Western Sandpipers per survey (SD = 272, n = 9) at Iona Foreshore. For the northward migration period of 2021, an average of 8,136 Dunlin per survey (SD = 6856, n = 10) was counted at Brunswick Point, in comparison to the average of 435 Dunlin per survey (SD = 684, n = 9) observed at Iona Foreshore during the same period.

Fig. 2
figure 2

Abundances of two shorebird species (Western Sandpiper and Dunlin) at the two study sites (Brunswick Point and Iona Foreshore) during southward and northward migration periods on the Fraser River estuary, British Columbia, Canada. Box plots depict distribution of observed values and extend from the lower quartile to the upper quartile, with median values shown as horizontal line. Whiskers indicate 1.5× the interquartile range. Each color dot represents an individual count conducted on separate days over the migration period. Note different y-axes on three panels. Dunlin did not occur at the sites during the summer southward 2020 migration window (they migrate later in the fall)

Full size image

The area surveyed for shorebirds at Iona Foreshore (177 ha) was 14% greater than at Brunswick Point (155 ha), so the higher shorebird abundances at Brunswick Point represented even higher shorebird densities. During the southward migration in 2020, Western Sandpipers had a mean density of 2.5 birds/ha (SD = 1.79, n = 12) at Brunswick Point, in contrast a mean density of 0.7 birds/ha (SD = 1.36, n = 12) at Iona Foreshore. During the 2021 northward migration,Western Sandpipers had a mean density of 135.5 birds/ha (SD = 73.71, n = 10) at Brunswick Point and a mean density of 1.2 birds/ha (SD = 1.54, n = 9) at Iona Foreshore. During the 2021 northward migration, Dunlin had a mean density of 52.5 birds/ha (SD = 44.23, n = 10) at Brunswick Point, and a mean density of 2.5 birds/ha (SD = 3.87, n = 9) at Iona Foreshore.

Biofilm Biomass (Chl a)

The difference in biofilm biomass (Chl a) between sites varied with migration periods (PERMANOVA Migration*Site interaction: Table 1). On average, samples from Brunswick Point had over twice the biomass of biofilm than at Iona Foreshore during southward migration in 2020 (Fig. 2, BP: 0.11 ± 0.08 mg g−1 DW, IF: 0.04 ± 0.02 mg g−1 DW). However, biomass samples were similar at these sites during northward migration 2021 (BP: 0.05 ± 0.01 mg g−1 DW, IF: 0.05 ± 0.03 mg g−1 DW). Over 30% of variation observed in biofilm biomass was dependent on site and migration period (Components of variation and PERMDISP Migration*Site interaction: Table 1), with Brunswick Point during southward migration being the most variable. Of note, nearly 28% of the observed variation in biomass (Chl a) was driven by sample station, indicating the importance of micro- and meso-scale variables in driving biomass patterns (Table 1).

Table 1 PERMANOVA results investigating changes in biofilm biomass (chlorophyll a content, mg g−1 DW) among sites and shorebird migration periods 2020 and 2021. Number of unique permutations = 997–999. P values for significant and interpretable fixed effects are in bold. Components of variation are estimated for all sources of variation. PERMDISP test assesses the variability in dispersion among levels of factors for significant PERMANOVA results
Full size table

Macronutrients

Macronutrient content in intertidal biofilm was generally higher at Brunswick Point than Iona Foreshore during both migration periods, and differences between sites were greater in the southward period than in the northward period (Table 2; Figs. 3 and 4). Protein content was 3.8× and 1.5× higher at Brunswick Point relative to Iona Foreshore during the southward and northward periods, respectively. Carbohydrate content was 1.8× and 1.6× higher at Brunswick Point relative to Iona Foreshore during the southward and northward periods, respectively. Lipid content was 3.2× and 1.4× higher at Brunswick Point relative to Iona Foreshore during the southward and northward periods, respectively.

Table 2 Macronutrient content (mg g−1 DW) and stable isotope values (‰), reported as mean ± standard deviation, for two sites on the Fraser River estuary during two shorebird migration periods (northward = April and May 2021, southward = August and September 2020). Samples sizes in Methods
Full size table
Fig. 3
figure 3

Biomass (measured as chlorophyll a content) and macronutrient content (presented as protein, carbohydrate, and lipid content) of intertidal biofilm at Brunswick Point and Iona Foreshore during northward and southward shorebird migration. Box plots depict distribution of observed values, and extend from the lower quartile to the upper quartile, with median values shown as horizontal line. Whiskers indicate 1.5× the interquartile range. Each color dot represents a unique biofilm sample. Note different y-axes for each plot

Full size image
Fig. 4
figure 4

Non-metric multidimensional scaling (NMDS) graph of biofilm macronutrient content (mg g−1 dry weight) in relation to site and migration period. Symbols closer to each other are more similar in macronutrient content than symbols farther apart. The vector overlay represents the correlations (Pearson correlation coefficients) between each macronutrient and the NMDS axes, and it shows the direction of increase in the NMDS graph. See Fig. 3 for detailed breakdown of macronutrient content

Full size image

Spatiotemporal variation in macronutrient content was complex (significant Migration*Site interaction PERMANOVA, P = 0.026: Table 3), with over 30% and nearly 27% of variation in macronutrient content being driven by migration period and site, respectively (components of variation, Table 3; Fig. 4). These results indicate significant temporal and spatial heterogeneity in macronutrients, which was partially driven by higher carbohydrate content during southward migration at Brunswick Point (SIMPER, Table 4).

Table 3 PERMANOVA results investigating changes in macronutrient content (mg g−1 DW) among sites and shorebird migration periods in 2020 and 2021. Number of unique permutations = 997–999. Components of variation are estimated for all sources of variation. P values for significant and interpretable fixed effects are in bold. PERMDISP test assesses the variation among levels of factors in multivariate dispersion for significant PERMANOVA results
Full size table
Table 4 SIMPER results to identify contributions of lipid, protein, and carbohydrate content to significant differences in macronutrient content observed among sites and migration periods, detected by PERMANOVA (Table 3). The squared distance between the groups represents the amount of dissimilarity between groups, and its average is presented for the assemblage of response variables (i.e., in this case, lipid, protein, and carbohydrate content). Percent contribution of each production variable to the overall dissimilarity is also presented. Bolded contributions (%) indicate that the ratio of the average squared distance to squared distance of the given response variable was > 1, which suggests that response variable was consistently driving observed differences
Full size table

Stable Isotope Signatures

Carbon and nitrogen stable isotope signatures during spring 2021 (δ13C: mean = −20.7, SD = 1.53, n = 35; δ15N: mean = 5.1, SD = 0.65, n = 35 [see Table 2 for site-specific values]) did not differ between the two sites (F[2, 29] = 0.57, P = 0.57), or between the months of April and May (F[2, 29] = 0.04, P = 0.95). The similar δ13C values indicating nutrients at both sites had similar origins from marine and freshwater inputs. The δ15N values of ~ 5‰ most closely correspond with signatures of microphytobenthos and surface sediments (Hobson et al. 2022) and are substantially lower than signatures of marine invertebrates in the study region (which range from ~ 8 to 14‰; Hobson et al. 2022), suggesting surface scrapings are primarily sampling autotrophic components of intertidal biofilm (see Supplementary Material).

Biomass (Chl-a) and Macronutrient Content

Within each site, there was a positive relationship between biomass (Chl-a) and macronutrient content, and the macronutrient content per unit biomass of dry sediment was higher at Brunswick Point than at Iona Foreshore for both migration periods (Fig. 5). Model selection results indicated details of this correlation varied among the three macronutrients (see Supplementary Material for full details). For protein, the most parsimonious model included site, season, and the interaction between Chl a and site (w2 = 0.67), which indicated the correlation between protein content and biofilm biomass was stronger at Brunswick Point, and this difference was similar between the two migration periods (Fig. 5). For carbohydrates, the most parsimonious model included the site/season interaction with Chl a (w4 = 1.00), indicating the positive correlation between carbohydrate content and biofilm biomass differed among all sites and seasons (Fig. 5). For lipid, the most parsimonious model included site, season, and the interaction between season and Chl a, indicating the correlation between lipid content and biofilm biomass was consistent between sites but differed between seasons, being positive in the summer and only weakly positive in the spring (Fig. 5).

Fig. 5
figure 5

The relationship between macronutrient content (presented as lipid, protein, and carbohydrate content) and biomass (measured as chlorophyll a content) of intertidal biofilm at Brunswick Point and Iona Foreshore during southward 2020 and northward 2021 shorebird migration periods. Note different y-axes for each plot, and differing x-axes for top and bottom sets of panels. Each color dot represents a unique biofilm sample. Lines indicate predicted values from general linear mixed effects model (see “Methods”)

Full size image

Discussion

Consistent differences in macronutrient content occurred between the two study sites in the Fraser River estuary during both southward and northward shorebird migration. As in previous studies (Butler 1994; Hemmera Envirochem 2014a; Jardine et al. 2015), the Brunswick Point site had higher abundances of shorebirds relative to Iona Foreshore, and these differences were mirrored by macronutrient content. Following our working hypothesis that shorebirds preferentially forage in higher “quality” habitat, the observed pattern suggests that nutritional content of intertidal biofilm may be an index of habitat quality for migrating shorebirds. The association provides a measure of habitat quality linked to the needs of migrating shorebirds conducting long-distance flights. However, as our comparative study is correlative, we cannot rule out other factors that could be acting in conjunction with the macronutrient content available in intertidal biofilm, including differences in invertebrate prey availability, predation danger, and anthropogenic disturbance. Each of these possible factors is discussed below.

Invertebrates can constitute a large fraction (~ 20–50%) of the diets of western sandpipers and dunlin (Kuwae et al. 2008; Jardine et al. 2015; Hobson et al. 2022), and their abundance, biomass, and diversity might explain the observed differences in shorebird counts between Brunswick Point and Iona Foreshore. Shorebird distribution and abundance has been related to the abundance of benthic invertebrates (Colwell and Landrum 1993) but not always (Horn et al. 2020). At the same Roberts Bank site, Jiménez et al. (2015) determined that the spatial distribution of Western Sandpipers mirrors the distribution of intertidal biofilm more closely than that of invertebrates. Our present results are consistent with Jiménez et al. (2015). Microphytobenthos within biofilm forms the base of the food web on these tidal flats, providing food for benthic invertebrates (Kuwae et al. 2012; Sutherland et al. 2013), which is, again, a reason that shorebirds might focus on sites where the intertidal biofilm has higher macronutrient content. Previous studies on benthic invertebrates using comparable methods at the two study sites yielded ambiguous results. Hemmera Envirochem Inc. (2014b) conducted an area-wide assessment of benthic invertebrates, separately by size classes (meiofauna [63 to 500 μm] and macrofauna [> 500 μm]) during shorebird migration (spring and summer) on the Fraser River Estuary. Biomass, abundance, and diversity of meiofauna were higher at Brunswick Point than Sturgeon Bank (where the Iona Foreshore site is located), with the exception of Polychaeta, which were not statistically significantly different between the two sites. In contrast, biomass, abundance, and diversity of macrofauna (all taxa) were not statistically different between the Brunswick Point and Sturgeon Bank. The lack of difference in macrofauna is consistent with Dashtgard (2011) who mapped the distribution of burrows, tracks, and trails made by intertidal invertebrates of the Fraser River delta. While Dashtgard (2011) found a strong onshore-offshore gradient in diversity and abundance in the traces left by invertebrates, only minor differences occurred between Sturgeon Bank (where the Iona Foreshore study site is located) and Roberts Bank (where the Brunswick Point site is located). Furthermore, Dashtgard (2011) argued that the stronger influence of the river outflow on Roberts Bank relative to Sturgeon Bank can result in relatively lower abundance and diversity of marine invertebrates. Data from an unpublished study (A. Blondin, unpub. data) based on cores collected in 2021 (n = 216) indicate that mean densities of intertidal invertebrates (per each sample volume, 1196.36 mL, collection depth to 10 cm) were similar between the two sites in spring (Brunswick Point: 5.1 vs Iona Foreshore: 5.8), and lower at Brunswick Point (5.9) than Iona Foreshore (11.6) in summer/fall. Thus, the available evidence indicates the abundance, biomass, and diversity of benthic invertebrates vary between the two sites in ways that may or may not correspond to use by shorebirds.

Differences in predation danger or site safety provide another possible explanation for inter-site differences in shorebird counts. Site selection by Western Sandpiper may involve a trade-off between food abundance and predation risk (Pomeroy et al. 2008). The “danger zone” for predation risk at intertidal areas is typically defined as the proportion of total intertidal area that lies within 150 m of the shoreline (Pomeroy 2006). The 150-m threshold is derived from studies estimating the head-start distance needed by sandpipers to achieve flight speeds fast enough to escape attacking Peregrine Falcons (Burns and Ydenberg 2002; Hope et al. 2020a, b). At our two study sites, the danger zone extends to ~ 40% of the total area (Supplementary Material Fig. 2), indicating that both sites present similar predation danger to sandpipers, and yet higher numbers of shorebirds were present at Brunswick Point.

The two study sites differed in the level of anthropogenic changes, with Brunswick Point being generally less disturbed than Iona Foreshore. The Iona Foreshore site has been subject to multiple stressors, including historical effluent discharge from the Iona Island Sewage Treatment plant (Arvai et al. 2002), development of coarser sediment on the north side of the Iona jetty (Lutternaeur and Murray 1973), and the altered hydrological regime over Sturgeon Bank (Levings 1980; Atkins et al. 2016). Also, the Iona Foreshore site is adjacent to the Vancouver International Airport. Shorebirds respond to moving aircraft in complex ways, including their potential for habituation to noises that are regular in occurrence (Van der Kolk et al. 2020). However, based on the working hypothesis that migrating shorebirds stage at estuarine habitats that provide the nutrients they need for long-distance flights, our study was intended to assess whether macronutrient content represented habitat quality, not to assess the cumulative impact of multiple stressors on macronutrient content and site use by shorebirds. Overall, while correlative studies cannot dismiss alternative explanations, our results suggest a positive link between nutritional content of intertidal biofilm and foraging habitat quality for shorebirds during their migratory stopover within the Fraser River estuary.

Factors that Affect the Macronutrient Content of Intertidal Biofilm

The similarity in stable isotope signatures indicates differences in macronutrients between the two sites may not be related to allochthonous inputs, and differences likely result from the endogenous processes by which intertidal biofilm produces macronutrients, similar to the Nanaimo estuary where primary production is mostly from in situ sources (Naiman and Sibert 1979). Macronutrient content in intertidal biofilm may vary due to three non-mutually exclusive mechanisms. First, all else being equal, sites with more biofilm biomass should have higher macronutrient content. We found support for this correlation, particularly in samples collected during the summer of 2020 (Fig. 5). Second, microalgal species differ widely in their relative macronutrient composition (Finkel et al. 2016), and thus, any change in species composition could alter the nutritional quality of biofilm. Intertidal biofilm at the Brunswick Point site comprises a diverse assemblage of diatoms (Beninger et al. 2011; Schnurr et al. 2019). However, given little information exists on taxonomic composition of intertidal biofilm at the Iona Foreshore site, we could only make limited inferences about this mechanism. Third, the physiology and macronutrient composition of microalgal species, such as marine diatoms, can be altered by changes in environmental conditions, such as ambient nutrient levels (e.g., nitrogen and phosphorus), temperature, and salinity (Ma et al. 2020), which can result in an increase in nutrient content per individual cells due to “metabolic overflow” (Stal 2003). All three mechanisms (biofilm biomass, taxonomic composition, physiological changes) are likely operating simultaneously in our study system, and we discuss their likelihood separately by macronutrient type (protein, carbohydrates, and lipid).

Protein

Much of the protein content of microalgae is generally attributed to the cellular architecture, particularly the photosynthetic apparatus (Geider and La Roche 2002; Finkel et al. 2016); thus, the protein and Chl a content of microalgae often show a positive relationship in response to light or nutrient conditions (Geider and La Roche 2002). Therefore, correlations between protein content and biomass (estimated from Chl a) may be stronger than for other macronutrients, a pattern we observed in our study. The correlation was particularly apparent during summer 2020 at the Brunswick Point site where Chl a values at some stations were nearly double the values observed in spring (Figs. 4 and 5). During spring 2021, we found a greater overlap in Chl a values between the two sites, and yet, Brunswick Point had higher protein content (Figs. 3 and 5), suggesting another mechanisms are at play. For example, diatoms are considered more protein-rich than other microalgae (Finkel et al. 2016), and the differences may reflect the underlying community composition of the microphytobenthos. Also, the decoupling of protein and Chl a observed in spring 2021 may reflect differences in primary productivity relative to detrital inputs between sites. Dell’Anno et al. (2002) compared macronutrients in marine sediments across 99 sites in the Mediterranean Sea and argued that protein concentration in marine sediments can be a marker for productivity and disturbance, with higher sediment protein reflecting greater autotrophic production by microphytobenthos and lower sediment protein reflecting greater decomposition or more detritus introduced from deeper sediments. The differences we observed between Brunswick Point and Iona Foreshore support the use of protein content in sediments as an indicator for anthropogenic disturbance of intertidal habitats and their microphytobenthos communities.

Carbohydrate

Carbohydrate content in intertidal biofilm is likely derived from cell walls of the component species and the amount of EPS that surrounds those species (Bellinger et al. 2005; Finkel et al. 2016). Carbohydrate content was consistently higher at Brunswick Point than at the Iona Foreshore site during both migration periods (Fig. 3), which may reflect taxonomic differences between the two sites, e.g., cells walls of cyanobacteria and dinoflagellates can be richer in carbohydrates than diatoms with silica-based cell walls (Finkel et al. 2016). Furthermore, carbohydrate content can vary widely depending on how microalgae respond to changes in environmental conditions (Bellinger et al. 2005). The secretion of EPS is considered a byproduct of unbalanced growth and overflow metabolism in which growing microbial cells divert excess absorbed energy to avoid oxidative damage (Stal 2003). EPS protects microbial cells from adverse conditions and enables cells to move within sediments (Smith and Underwood 1998). In addition, carbohydrate content in EPS can vary depending on whether microalgae are in exponential or stationary growth phases. When cells enter the stationary phase under unfavorable conditions, a shift in the metabolic pathways occurs that allows diatoms to continue photosynthesis by secreting photosynthetic products (Orvain et al. 2003). We found the correlation between carbohydrate content and biofilm biomass varied by site and season (Fig. 5), reflecting the complex relationship between EPS production and biofilm, suggesting multiple mechanisms are responsible for the spatial and seasonal differences in carbohydrate content.

Lipid

Lipid is the preferred fuel for migrating birds (Guglielmo 2010) and being the least plentiful of the three macronutrients in intertidal biofilm (Table 2), deserves special attention. Lipid content was consistently higher at Brunswick Point than Iona Foreshore (Fig. 3), reflecting the difference in shorebird abundance (Fig. 2). Furthermore, the correlation between lipid content and biofilm biomass differed between the two migration periods (Fig. 5). In summer 2020, the highest lipid content was found at stations with higher Chl a, indicating the higher lipid likely resulted from differences in biomass. In contrast, lipid content was only weakly associated with biofilm biomass during the spring 2021 period, suggesting differences in lipid content are likely due to physiological changes or taxonomic composition, e.g., diatoms are considered more lipid-rich than cyanobacteria (Finkel et al. 2016). Microalgae would experience changes in physiology depending on their exponential or stationary growth phases. Lipid accumulation in response to variable conditions is widespread in eukaryotic microalgae, which form cytoplasmic lipid bodies composed primarily of triacylglycerols (Maeda et al. 2017). Spring is a time of rapid change over the intertidal mudflats of the Fraser River Estuary, with increasing solar radiation, rising mean temperature, decreasing salinity, and longer tidal exposure times (Canham et al. 2021). The rapid discharge of freshwater from the river increases turbidity in the water column and pushes marine diatoms in estuarine biofilm towards their lower salinity tolerances (Admiraal 1984; Underwood and Provot 2000; Scholz and Liebezeit 2012). These changes, acting separately or in combination, can be predicted to incite a lipid accumulation response in the microalgae found in intertidal biofilm.

The difference in bird use between sites (6–100×) was a higher magnitude than the difference in biofilm quantity and quality (1.5–3.8 × in macronutrient content), indicating either other factors also play a role, or that birds could be focusing on specific nutritional attributes in a non-linear manner, e.g., focusing on specific compounds. The lipid content of diatoms is a major source of long-chained polyunsaturated fatty acids (Passarelli et al. 2015), essential nutrients for long-distance migration of shorebirds (Maillet and Weber 2006; Baird 2022; Dick et al. 2023). Schnurr et al. (2020) found abundances of fatty acids in intertidal biofilm peaked during the northward migration of shorebirds in spring. Similarly, lipid content in intertidal biofilm at Roberts Bank in 2017 differed between spring and summer (Schnurr et al. 2020), with the decreased abundance of lipid in the summer being associated with a lower abundance of diatoms and higher abundance of bacteria, cyanobacteria, and dinoflagellates, as inferred from fatty acid profiles (Schnurr et al. 2020). In contrast, we found lipid content at Brunswick Point was higher during the summer than in the spring (Fig. 3), although our study confounded seasonal with annual differences. Furthermore, the study by Schnurr et al. (2020) was primarily located in the Westham Island area that lies adjacent to the north of our sampling locations, and thus, comparisons with Schnurr et al. (2020) may confound differences with both area and season. Further replication between sites and seasons is needed to resolve these questions. Abiotic conditions can cause intertidal biofilm communities to be highly variable year-to-year (Mogle 2021). We expect summer macronutrient composition to depend partially on what occurred during the previous spring, which we could not observe given we sampled intertidal biofilm in non-consecutive migration seasons. The spatial and annual differences in seasonal patterns underscore the complexity and natural variability of these intertidal systems.

Our study demonstrates that more nutritious biofilm with higher macronutrient content occurs at tidal flats where Western Sandpiper and Dunlin forage in higher numbers. The spatial difference suggests that biofilm quality is an important aspect of tidal flats that staging shorebirds target to meet their long-distance migration needs. These findings show that tidal flats in the Fraser River Estuary are not all “equal” in terms of their ecological processes or services. Further testing of the importance of biofilm macronutrient content could involve assessing the total nutrient intake by shorebirds given typical foraging rates and comparing to availability, e.g., if total nutrient demand exceeds the nutrient concentration in intertidal biofilm at one site but not the other.

Macronutrient content of intertidal biofilm provides a measure that can be linked directly to the dietary needs of migrating shorebirds. Spatial differences in macronutrient content, particularly lipids, demonstrate the need to consider the nutritional “quality” of biofilm and not just biomass (i.e., chlorophyll a) when determining restoration goals for tidal flats or assessing coastal areas for development projects. The technical attributes for artificial intertidal flats that promote use by small-bodied shorebirds include a depositional environment, complex shoreline, gentle slope, gradient of grain sizes from muddy to sandy, maximum water depth at the lowest tide 5 cm or less, freshwater inflow, and unobstructed sight-lines (Kuwae et al. 2021). However, several critical questions remain for effective intervention in intertidal flat ecosystems, including the provision of essential nutrients during key periods in the life cycle of shorebirds. The drivers of biofilm macronutrient content, such as biofilm community composition and physiological responses induced by environmental changes (e.g., lipid accumulation response), have to be more fully understood to better inform future management decisions regarding tidal flats and to facilitate more effective conservation for shorebirds (Mathot et al. 2018). Moreover, the urgency is exacerbated by climate change affecting the biofilm community, and consequently migratory shorebirds (Ollinik et al. 2021; Mogle 2021), through effects of increased temperatures and earlier spring phenology. As such, enhanced measures to complement shorebird surveys are needed to assess coastal development proposals and enact restoration plans to conserve critical tidal flats for staging shorebirds.