Knowing what birds eat is fundamental to understand the ecology and distribution of individuals and populations. Often, diet is assessed based on field observations and excrement analyses, which has previously been the case for Sanderling Calidris alba. This may have biased their known diets towards large prey with indigestible body parts that can still be recognized in faeces or regurgitations. A literature review of Sanderling diet worldwide showed that Sanderlings exploit a large diversity of prey. We carried out DNA metabarcoding on Sanderling faeces to get a complete view of their diet in the Wadden Sea during staging and moult from late July to early October. Given the diversity of available prey in the Wadden Sea, it was remarkable that 94% of the samples contained Brown Shrimp Crangon crangon which, next to the Shore Crab Carcinus maenas, were also the most abundant species in the samples. This study shows that whereas Sanderling can feed on a large variety of invertebrates, in the Wadden Sea during southward staging they primarily rely on Brown Shrimp
Knowing the diet of birds is an essential component of understanding their ecology, behaviour, distribution and population trajectories. Clearly, this also holds for shorebirds (Charadriiformes; Goss-Custard et al. 1977, van de Kam et al. 2004, Piersma 2012) but knowledge on the diet of coastal sandpipers (Scolopacidae) is undoubtedly incomplete. In sandpipers, the most common methods to assess their diet are focal observations (Sutherland et al. 2000, Choi et al. 2017) and excrement analyses (Durell & Kelly 1990, Piersma et al. 1994, Duijns et al. 2013). However, focal observations come with the risk that small or quickly handled prey will remain unnoticed or unidentified (Vanermen et al. 2009) and dropping analyses only include species that contain indigestible parts (Dekinga & Piersma 1993). For example, a combination of microscopy of their tongue, isotope analysis and high-resolution filming of foraging behaviour analytical techniques was necessary to discover that Western Sandpipers Calidris mauri can graze biofilm (Elner et al. 2005, Kuwae et al. 2008), which can not be identified as prey using the conventional techniques. Insights from a combination of new techniques may lead to a different view on the ecology of the species in question (Kuwae et al. 2012).
Sanderlings Calidris alba are also small sandpipers whose diets have been studied mainly by visual observation (Evans et al. 1980, Maron & Myers 1985, Masero 2003, Vanermen et al. 2009, Grond et al. 2015) and microscopic analyses of their faeces (Nuka et al. 2005, Castro 2009, Lourenço et al. 2015, 2016). Sanderlings inhabit a diverse range of habitats: from freshwater estuaries to coastal beaches in the non-breeding season and High Arctic tundra on the breeding grounds (Reneerkens et al. 2009, Conklin et al. 2015).
The East Atlantic flyway population of Sanderlings has increased three-fold since the 1980s (van Roomen et al. 2018). During that period, peak numbers at the most important staging location in the Dutch Wadden Sea increased even faster, highlighting the increased importance of this area (Loonstra et al. 2016). Field observations on the Wadden Sea mudflats suggested that Sanderlings foraged on crustaceans, such as Brown Shrimp Crangon crangon (hereafter ‘shrimp’; Loonstra et al. 2016). However, as in other sandpipers, field observations unlikely give a complete overview of Sanderling diet (Jouta et al. 2017). Furthermore, soft-bodied prey will leave little remains in the faeces, making dropping analyses also unsuitable to determine Sanderling diet. Finally, Sanderlings have been reported to regurgitate pellets (Petracci 2002, Kelly 2008), but such pellets are produced considerably less frequently than faeces, and diet reconstructions based on examined pellets will be biased towards prey with indigestible parts (Worrall 1984, Dekinga & Piersma 1993).
We reviewed the scientific literature to get an overview of the Sanderlings' diet worldwide and listed the available (potential) benthic prey of Sanderlings in the Wadden Sea, based on an existing local benthic fauna sampling programme. In addition, we specifically described the diet of Sanderling during southward migration in the Wadden Sea using DNA metabarcoding of the faeces (De Barba et al. 2014). This noninvasive method uses the interspecific variation in DNA sequences of a small part of the genome to taxonomically identify prey items, using a reference database (Hebert et al. 2003). As such, we can place the diet of Sanderlings in the Wadden Sea into worldwide perspective.
METHODS
Study area
The study was conducted near the uninhabited island Griend (53°14′N, 5°15′E; Figure 1), which is the most important site for Sanderlings in the Dutch Wadden Sea (van den Hout & Piersma 2013). An estimated 11–14% of the flyway population makes use of the Griend area as a staging site (Loonstra et al. 2016). Up to 20,000 Sanderlings at one time use the island's beaches to roost (EP pers. obs.). The island is surrounded by extensive mudflats where Sanderlings forage during low tide (van den Hout & Piersma 2013). The area is characterized by a semi-diurnal tide with an amplitude of 1.5–2.5 meters.
Literature review
We searched the literature for original papers on Sanderling diet. For each publication, prey species groups and methodology were identified. Additional information was noted on the location, habitat and the season in which the study was carried out. When multiple methods or habitats were used in a study, each was included in our review.
Database of prey items
A list of possible prey items of Sanderlings was constructed based on all species that were found by a large-scale monitoring campaign of intertidal macrofauna that covers the Dutch Wadden Sea (Compton et al. 2013). Sediment samples were taken using a sampling corer (25 cm depth and 0.018 m2 surface) on a 500-m grid with additional random points (Bijleveld et al. 2012). For this study, only the 238 sampling points around Griend, which were sampled in July 2018, were included (Figure 1A). Samples were collected during low tide on foot or during high tide from a small boat. Each sample was sieved on a 1-mm mesh sieve and taken back to the NIOZ laboratory to identify all organisms up to species level or the finest taxonomic level possible (Compton et al. 2013). The sampling programme detects sedentary species best and likely underestimates mobile species like the Shore Crab Carcinus maenas (hereafter ‘crab’) and shrimp. We considered all benthic species that were encountered on at least five sampling stations, independent of their burying depth, to compile the most generous list of possible prey.
Faecal sample collection
Between late July and early October 2016–2018, we collected faecal samples at 35 locations (Figure 1B). During low tide, flocks of Sanderlings were followed on the mudflat and samples were collected immediately after a group had left the area. To avoid the collection of droppings from other small shorebird species, we collected droppings instantly after a flock had left the area and from flocks consisting of more than 90% Sanderlings only. The samples were picked up with a metal spoon such that we collected as little as possible uric acid (the white part of a dropping) and sediment to avoid deterioration of the DNA and/or contamination of the sample. On each location between 2 and 20 faeces were collected. Each sample was put in a separate plastic bag and stored in the freezer at –10°C within 5 h after collection and remained frozen until DNA extraction. From all the collected samples (n = 310) a random selection of 110 samples was taken for analysis. For one set of samples that was collected at the same time (n = 12), the year and location of collection is known but the day of sampling was lost.
DNA isolation and amplification
DNA was extracted with the Invitrogen PureLink Microbiome DNA Purification Kit, which is especially effective at reducing levels of uric acids, which are present in bird faeces and cause PCR inhibition (Jedlicka et al. 2013). Faecal samples were homogenized and subsampled to arrive at a sample weight of <1 g. The sample, together with c. 0.1 g extra 0.1 mm Zircona/Silica beads, was added to a manufacturer-provided bead tube and subjected to bead beating in 5×2-min bouts pausing 30 s between bouts. Subsequently the manufacture's protocol was followed. The elution step with 50 ml Ambion purified DNA-free water had an extended incubation time of 4–5 min, and eluted DNA was re-applied to the filter and incubated 2 min extra before final elution.
For PCR amplification AccuStart II PCR ToughMix was used to guarantee amplification success, as the targeted prey DNA may have been degraded. The reaction volume was 10 ml including 5 ml AccuStart, 1ml of each primer (10 mM), 1ml ddH2O and 2ml DNA template. The PCR profile was 3 min at 94°C followed by 35 cycles of 1 min at 94°C, 30 s at 48°C and 1 min at 72°C, and a final extension at 72°C for 10 min. Annealing temperature was set as low as 48°C to minimize taxonomic bias (following Ishii & Fukui 2001). Each sample was amplified in triplicate to avoid initial-cycle template bias. Samples were randomized before DNA extraction and extractions were randomized before PCR. Negative DNA extractions and negative PCR controls were included to track possible contamination of reagents. Before pooling the PCR replicates, 5 ml PCR product was assessed in a standard gel electrophoresis. The 18S rRNA gene was targeted using generic primers SSU-F04 and SSU-R22mod as primer pair with an amplicon length of 450 bp (Sinniger et al. 2016, Fonseca et al. 2010). Choosing generic primers, we aimed to target the full spectrum of species in Sanderling excrements and limit observation bias. This meant that non-Animalia taxa were also targeted (e.g. Fungi).
DNA sequencing and bioinformatics
PCR products of samples (n = 110), negative extraction controls (n = 3) and PCR controls (n = 4) were sequenced on the MiSeq Sequencer (Illumina) at the Department of Human Genetics, Leiden University Medical Centre, aiming for a read depth of 50,000 per sample. Libraries were prepared with the MiSeq V3 kit, generating 300-bp paired-end reads. The V3-kit does not normalise, which means that it leaves the relative presence of initial PCR product intact, and therefore this library preparation method allows assessing the relative contribution of prey taxa (Verkuil et al. 2022).
Low-quality reads with a quality score ≤ 30 over 75% of the nucleotide positions were discarded using the fastq_quality_filter script in the FASTX-Toolkit ( http://hannonlab.cshl.edu.fastx_toolkit). The quality filtered reads were front and end clipped to remove the primers. Subsequently, reads were dereplicated and unique reads were discarded. The remaining sequences were clustered in operational taxonomic units (OTUs) using a 98% similarity cut off in VSEARCH (Rognes et al. 2016) and singleton OTUs were omitted. The final OTU table was adjusted for the negative extraction controls. Samples were corrected using the nearest extraction control or the mean of two extraction controls for samples that were located between two controls. The sum of all reads in negative extraction controls made up 1.3% of total number of reads (Table S4 and S6). Human DNA, found in negative controls and field samples, made up 0.16% of all reads.
Taxonomic assignments
All OTUs were taxonomically assigned based on a reference database build from the SILVA 18s rRNA database (release 128; Pruesse et al. 2007) and our own local Wadden Sea reference database for marine benthic species (GenBank accession numbers MZ709983-MZ710042). OTUs from the taxonomic kingdom of Animalia were identified up to species level or the finest taxonomic level possible. OTUs from other taxa were clustered at the level of phylum or kingdom because we a priori rendered it unlikely that these organisms (e.g. Fungi) are part of true Sanderling diet. Taxonomic assignment was performed using the RDP Classifier (Wang et al. 2007) with a minimum confidence of 0.8.
Filtering of results
After taxonomic assignment, all non-macrofaunal species groups (i.e. those that are not retained by a 1-mm sieve) were excluded from further analyses. For these species groups it could not be excluded that they ended up in the sample as a result of contamination during sample collection, for example sand containing environmental DNA, or secondary consumption, i.e. prey within the prey on which Sanderlings fed.
Data analyses
The species composition found in the Sanderling faeces was described based on two approaches: frequency of occurrence (presence/absence of taxa) and sequence counts (relative read abundance). The frequency of occurrence is simply the frequency at which a taxon occurred in the samples expressed as a percentage of all samples. The relative read abundance across all samples was calculated by dividing the number of reads of a taxon by the total number of reads. We compared the species composition between faecal samples using the relative read abundance of taxa per sample. The relative read abundances were first Hellinger transformed using function ‘decostand’ from the vegan package v. 2.5.7 in R (Oksanen et al. 2018) followed by the calculation of dissimilarity distances using the Bray-Curtis equation. The resulting dissimilarity matrix was used for analysis of variance species assemblage between (1) sampling locations and (2) month of sampling, using the adonis2 function (permanova) and nonmetric multidimensional scaling (nMDS) plots. Non-metrical dimensional scaling was used to see if the samples from the same location or month of collection were more similar than other samples. The dissimilarity matrix was also used to perform a similarity percentage (Simper) analysis to discriminate between the effect of each species and to answer the question of whether certain species were more important in explaining the variation in a specific time of the year. The relative read abundance of the ten most abundant taxa were calculated per taxa and per sample to show species-specific changes over time. To improve the readability, we fitted a local regression (LOESS function in package ggplot2 v. 3.3.5 (Wickham 2016) to smooth the variation in relative read abundance over time. All statistical data analyses were carried out in R v. 4.1.2 (R Core Team 2021).
RESULTS
Literature overview
We found 28 publications that described Sanderling diet (Table S1) based on behavioural observations, dropping analyses, stable isotope analyses, pellet analyses, stomach content and stomach flushing. Most of the studies used behavioural observations (n = 15) or dropping analysis (n = 9) to determine Sanderling diet. Other non-invasive methods that were used were pellet analysis (n = 2) and DNA metabarcoding of faeces (n = 1) from Arctic breeding grounds. Invasive methods such as stomach flushing (n = 2) and the dissection of Sanderlings to obtain the stomach content (n = 4) were not often applied. Stable isotope analyses of toenails and blood was used in one study. Most studies were carried out in winter and spring (n = 17 and 10, respectively) and along the East Atlantic Flyway (n = 17) and American flyways (n = 11; Figure 2A). Crustaceans and polychaetes were common prey, as well as bivalves, insects, gastropods and fish (Figure 2B). Although Sanderlings have tongue spines that facilitate the consumption of biofilm (Kuwae et al. 2012), biofilm turned out to be not of importance in the Sanderling diet (Lourenço et al. 2017).
Figure 2.
The worldwide diet of Sanderling. (A) World map with the study locations of the reviewed studies. Dot size indicates the number of studies that were conducted. (B) Frequency distribution of prey groups present in examined studies. Only prey groups that occurred in more than two studies are shown. A complete list of Sanderling prey is available in Table S1.
Occurrence of available prey in the Wadden Sea
In 2018, a total of 38 species were found more than five times as part of the sampling programme of benthic fauna in the Wadden Sea (Table S2). Identification up to species level (n = 24) and genus level (n = 9) was most common followed by family (n = 4) and class level (n = 1). Most encountered species around Griend were polychaetes (n = 23), followed by bivalves (n = 8), crustaceans (n = 5) and the Mudsnail Peringia ulvae (gastropod).
Metabarcoding
DNA was extracted and amplified from 110 samples collected between 29 July and 6 October in three years (2016–2018; Table S3), which approximately spans the entire autumn period during which Sanderlings use the Griend area (Loonstra et al. 2016). The metabarcoding analysis resulted in 3,258,278 reads (excluding the controls). Raw Illumina sequences were deposited in the European Nucleotide Archive (ENA accession number: PRJEB55071). After clustering and removing 818,566 singletons, 7585 OTUs remained. Taxonomic assignment of these OTUs resulted in 133 different taxa from seven taxonomic kingdoms (all taxa listed in Table S4). The sum of reads in the pooled negative PCR controls (n = 4) was 26 reads with a maximum of 10 per OTU. The sum of reads in the negative extraction controls was 41,922 reads with a maximum of 11,917 reads per OTU. 86% of the reads in the negative extraction controls were of non-macrobenthic taxa. In field samples with DNA template there is a lot of competition for the primers, so very tiny amounts of contaminants would also lead to tiny number of reads. In controls however, the tiniest bit can create many reads. One of the negative extraction control samples contained a very high number of springtail (Collembola) reads (n = 5859, total number of Collembola reads in controls: 5866; Table S6) and another had an exceptionally high number of shrimp reads (n = 79, total number of shrimp reads in controls: 93; Table S6). Therefore, we corrected the number of reads of Collembola and shrimp based on the negative extraction controls (see Table S6 for results of negative controls). In total, 14.6% of all Animalia reads could not be assigned up to species level (Animalia only: 0.09%, Phylum: 0.64% (n = 4), Class: 0.35% (n = 11), Order: 9.39% (n = 28), Family: 0.79% (n = 11), Genus: 3.33% (n = 18)).
Figure 3.
Diet of Sanderlings in the western Dutch Wadden Sea. The frequency of occurrence of taxa that occurred in more than 10% of all samples (n = 20). Taxa were grouped by higher taxonomic levels if possible. Colours refer to the taxonomic class or sub-phylum (Crustacea). Images of species are chosen to represent higher taxonomic groups.
Frequency of occurrence
Focusing on macrobenthic species only, we found 47 taxa from five different phyla (see Table S4) in Sanderling excrements. Shrimp were found in 94% of all samples (Figure 3). Another crustacean, Shore Crab, occurred in 91% of all samples, followed by four groups of polychaetes (Lanice sp. 84%, Orbiniidae 77%, Arenicola sp. 68%, Hediste sp. 60%; Figure 3). Twenty-seven of the taxa occurred in less than 10% of the samples. Scoloplos armiger was the only species of Orbiniidae found in the sediment cores in our study area. Therefore, Orbiniidae will hereafter be referred to as Scoloplos armiger. Similarly, there were only single species of the following genera: Lanice conchilega, Arenicola marina, Hediste diversicolor, Heteromastus filiformis, Eteone longa, Mytilus edulis, Mya arenaria and Ensis leei. Taxa with unknown Family or lower taxonomic level, will be referred to by their Class: Arachnida (Trombidiformes), Insecta (Diptera).
Relative read abundance
The bulk (97.5%) of all reads originated from ten species (Figure 4) that belonged to the following three phyla: Arthropoda, Annelida and Mollusca (respectively 82.3%, 10.2%, 4.7%). Shrimp were most abundant and represented 61.7% of all the reads in the samples followed by crabs, Bristle Worms Scoloplos armiger and Razor Clams Ensis leei (respectively, 15.8%, 7.6%, 3.8%).
Figure 4.
The relative read abundance of the ten most abundant macrobenthic taxa found in the Sanderling diet in the western Dutch Wadden Sea. Each colour refers to a different taxon. Images of species are chosen to represent higher taxonomic groups.
Figure 5.
Nonmetric dimensional scaling plots comparing communities in Sanderling faeces. The plots are based on Bray-Curtis dissimilarities of relative read abundance of species within samples for (A) sampling locations and (B) month. Colours represent different sampling location (A) and months (B). Black triangles indicate the position of the five most dominant prey species according to the simper analysis. Small differences in point position are due to differences in sample size (A: n = 110, B: n = 98; see Methods).
Inspection of nMDS plots did not show clear clusters based on either location but suggested differences between months (Figure 5). Permanova analysis confirmed that there was no difference between locations (F1,108 = 1.706, P = 0.1), but indicated a significant difference in the diet composition between months (F1,96 = 3.278, P = 0.004). The average dissimilarity between months was 54.6% ± 0.02 (SD), with little variation in the dissimilarity between compared months. The lowest dissimilarity was between August and September and the highest dissimilarity was between July and September (Table S5). The greatest difference between months was due to the higher contribution of crab in the diet in July compared to October (Table S5).
We did not detect large seasonal changes in the relative read abundance for different species, except for crabs, whose relative read abundance slightly decreased with the progressing season (Figure 6). Shrimp showed a high relative read abundance throughout the season (Figure 6).
DISCUSSION
Worldwide, Sanderlings exploit a variety of benthic prey: arthropods, gastropods, shellfish, polychaetes and crustaceans. Along the East Atlantic Flyway in which the Wadden Sea is a central wetland, the staple food for Sanderlings wintering in Ghana was the small bivalve Donax pulchellus (Grond et al. 2015, Quartey 2018). In Mauritania, Sanderlings foraging on intertidal mudflats mainly ate Mudsnails (Lourenço et al. 2016). The diet was complemented with bivalves, polychaetes, crustaceans, insects and to a lesser extent plant material. Sanderlings wintering in Southern Europe had a diet that consisted of bivalves, gastropods, polychaetes and insect larvae (Perez-Hurtado et al. 1997, Masero 2003, Lourenço et al. 2015). On Belgian shores, Sanderlings ate a variety of prey, of which the polychaete Scolelepis squamata and the flesh of dead bivalves or those that were washed up ashore were important components (Vanermen et al. 2009). On Dutch beaches Sanderlings mainly fed on Scolelepis squamata in winter (Grond et al. 2015). Like most other Arctic-breeding shorebirds, Sanderlings ate arthropods on the breeding grounds (Wirta et al. 2015, Reneerkens et al. 2016).
Figure 6.
The relative read abundance per species per sample over time. Dots indicate the means, bars the SE and the black line is the LOESS smoothed mean (span = 0.75) with 95% CI (grey). The colours of the banners refer to the taxonomic class in correspondence with Figures 2–4.
Given that species from all taxa commonly consumed by Sanderlings were present in the Wadden Sea, it is striking that DNA metabarcoding revealed a predominance of shrimp in Sanderling faeces (94% of the samples). The shrimp showed a steady presence in the diet throughout the entire season of southward migration. In terms of the number of sequence reads, shrimp were also the most abundant species (61.8% of all reads). Crabs occurred in 91% of the samples and were the second most common and abundant species with 15.8% of all reads. Perhaps surprisingly, shrimp and, to a lesser extent, crabs, did not frequently occur in the benthic sampling program (Table S2), but we believe that this is because mobile species such as shrimp and crabs, can escape before the sample is taken. We did not detect an effect of sampling location on Sanderling diet composition, so the diet was similar across the sampling locations. Even though shrimp and crabs were by far the most important prey for Sanderlings in the Wadden Sea, these species have only been described as Sanderling prey species in the Wash in England (Kelly 2008), an intertidal area close to and quite similar to the Wadden Sea (Bocher et al. 2007).
Why do Sanderlings in the Wadden Sea eat shrimp and crabs? Shrimp and crabs are relatively soft-bodied prey, which are easy to digest compared with hard-bodied prey such as bivalves and gastropods (van Gils et al. 2003). Hard-bodied prey forces birds to take digestive breaks, leaving less time for foraging (van Gils et al. 2005a). Sanderlings ingesting whole bivalves in Ghana indeed spent a lot of time seemingly inactive, presumably to process the shell material (Grond et al. 2015). Therefore, in terms of digestibility, shrimp and crabs could be a more profitable prey because they contain little indigestible ballast mass compared to bivalves (van Gils et al. 2005b). In addition to that, shorebirds that forage on hard-bodied prey need to maintain a larger gizzard to crush the shell material (Dekinga et al. 2001, Piersma et al. 1993). Considering that a significant proportion of the Sanderling population in our study area was fuelling up for migration (Loonstra et al. 2016), foraging on hard-bodied prey, which would involve the carrying of a heavy organ, would not be favourable (Piersma 1998).
Furthermore, Brown Shrimp are rich in polyunsaturated fatty acids, specifically docosahexaenoic acid (C22:6n-3) and eicosapentaenoic acid (C20:5n-3; Mika et al. 2012 & Turan et al. 2012), which were shown to be quickly incorporated into body lipids and to increase the activity of flight muscle enzymes important for prolonged flight (Maillet & Weber 2006). For that reason, it may be that shrimp are a very suitable prey for Sanderlings prior to migration. Still, levels of polyunsaturated fatty acids are lower in autumn (when Sanderlings stage in the Wadden Sea) and in juveniles (the age group eaten by Sanderlings) and specific values for the Wadden Sea are unknown so further research is required to determine the nutritional value of shrimp for Sanderlings (Moore 1976, Mika et al. 2014).
Shrimp that occur on mudflats are mostly juveniles that leave the intertidal to become adults in the subtidal (Tiews 1970). The young shrimp rapidly disappear from mudflats from August onwards (Beukema 1992). Small shrimp (<20 mm) were swallowed whole by Sanderlings while the scarcer large shrimp (>20 mm) were stabbed and only their flesh was picked out (EP & JR pers. obs.). This implies that all sizes of shrimp may be edible for Sanderlings, which could explain the high relative read abundance of shrimp throughout the study period (Figure 6). Still, there may be differences in profitability for shrimp of different sizes. Large shrimp may be more difficult to catch for Sanderlings compared to small shrimp, while the gain in terms of flesh mass is higher. Foraging experiments with captive Sanderlings that are fed with different-sized shrimp could reveal whether such a trade-off exists.
The importance of crabs in the Sanderling diet decreased over the season (Figure 6), contributing to the slight change in diet composition across months, according to the simper analysis. Like shrimp, crabs decrease in abundance after July and they reach larger sizes (Beukema 1991, van Gils et al. 2005b). Large crabs (>20mm) are less common on mudflats than small crabs (<20 mm; Beukema 1991) and like large shrimp, probably require more handling time. Moreover, the strength of the crab carapace may not allow to be opened by a Sanderling bill at all. Therefore, Sanderlings foraged mainly on small crabs (EP pers. obs). As crabs were mostly swallowed whole by Sanderlings, at some point they may have outgrown the edible size for Sanderlings (Zwarts & Blomert 1992).
Most encountered species by the benthic sampling program were polychaetes. Still, not many polychaetes were common or abundant in the Sanderling diet, possibly because they were buried deeper than 2 cm and therefore unavailable for Sanderlings (Gerritsen & Meiboom 1986). Scoloplos armiger was the only taxon of polychaetes that was both common and abundant in their diet. An average Scoloplos armiger contains 0.004 g ash free dry mass (Duijns et al. 2013). This is 3–4 times less than the biomass of an average sized shrimp or crab from the mudflat (respectively 0.011 g and 0.014 g; Duijns et al. 2013, Zwarts & Wanink 1993). In order to obtain the same energy uptake, Sanderlings would therefore need to eat 3 to 4 times more Scoloplos armiger compared to shrimp or crabs. Prey density had the largest effect on the intake rate of Sanderlings at a beach along the Pacific coast of California, USA (Myers et al. 1980). To learn more about the profitability of a Wadden Sea prey, we would need information on prey density as well as handling times of specific prey items to compare the functional responses of Sanderling foraging on different prey items.
Razor clams Ensis spp. have been described as prey for Sanderlings in winter (Vanermen et al. 2009); Sanderlings picked out the meat from clams that stranded on beaches (Kelly 2008). In our results, the importance of razor clams in the diet increased in October due to a single sample (Figure 6). Perhaps, an increased storm frequency in this time of year lead to strandings of razor clams on mudflats, which could serve as food for Sanderlings. Interestingly, the contribution of prey that were not shrimp or crab, remained relatively stable over the season (Figure 6). It may be that Sanderlings miss out on specific nutrients by foraging exclusively on shrimp and crab. Therefore Sanderlings, like other vertebrate carnivores, may selectively forage on other prey to balance their diet (Kohl et al. 2005).
Biofilm is part of the diet of North American Western Sandpipers, but Lourenço et al. (2017) found that biofilm was not of importance in the Sanderling diet. Even though this study was conducted on wintering grounds and on more southern latitudes, there are good reasons to extend their conclusions to the Wadden Sea. First, the biofilm mats that other smaller shorebirds use to forage on elsewhere (Elner et al. 2005), do not occur in the places where we observed Sanderlings foraging and where faecal samples were collected in the Wadden Sea. The water and sand dynamics in our study area are probably too high for substantial biofilm to develop. Secondly, we did not observe foraging behaviour that is associated with eating biofilm while we did observe Sanderlings eating shrimp and crabs. The searching behaviour of Sanderlings was characterized by relatively few probes in the sediment, moving fast through the habitat, switching often between movements with the bill pointing to the ground and standing up straight, then turning the head and body to follow the moving prey (EP pers. obs). For Sanderlings to obtain enough meiofauna or biofilm, we expected to see foraging behaviour that matches those prey types. Western Sandpipers foraging on meiofauna showed a high frequency pecking mode and slow movement through the habitat (Kuwae et al. 2008), which we did not observe in Sanderlings on mudflats in our study area.
Over the past decades, shrimp have become more abundant during the spring staging period of Sanderlings (May) in the Wadden Sea (Penning et al. 2021). It is unknown whether the availability of shrimp has also increased during the autumn staging period. Sanderlings use the Wadden Sea to fuel up during northward and southward migration and as such, the availability of benthic prey on the intertidal mudflats are important for a significant part of the East Atlantic Flyway population of Sanderlings (Loonstra et al. 2016). Sanderlings wintering in Africa use multiple staging sites during northward migration, of which the longest time is used for re-fuelling in the Wadden Sea compared with more southerly locations (Reneerkens et al. 2020). This suggests that a large proportion of the flyway population fuels up predominantly preying on shrimp. Perhaps, the (current) presence of large densities of shrimp and crabs make the Wadden Sea a more suitable staging site than more southerly wetlands along the flyway. It is currently unknown to what extent changes in shrimp abundance have affected the local and flyway-wide population growth of Sanderlings.
ACKNOWLEDGEMENTS
This study was part of the project ‘Behoud van een dynamisch Griend’ managed by Natuurmonumenten and funded by Waddenfonds, the provinces of Fryslân, Groningen and Noord-Holland, and Rijkswaterstaat. We thank Jop van Beek who helped with the lab work and Sölvi Rúnar Vignisson for sending additional data on Icelandic diet of Sanderlings. We are indebted to Allert Bijleveld and Anita Koolhaas for maintaining the SIBES database and kindly providing the data on prey availability. Theunis Piersma and Laura Govers provided (encouraging) comments on an earlier draft.
REFERENCES
Appendices
SAMENVATTING
Voor een beter begrip van de ecologie en verspreiding van vogels is kennis van hun voedsel fundamenteel. In de meeste studies wordt het voedsel bepaald aan de hand van veldobservaties en analyses van uitwerpselen. Dat is ook bij Drieteenstrandlopers Calidris alba het geval. Het is daardoor mogelijk dat we een vertekend beeld hebben van de prooikeuze, doordat voornamelijk grote prooien met onverteerbare delen herkend kunnen worden in de uitwerpselen of braakballen. Een literatuuroverzicht van wereldwijde onderzoeken naar het voedsel van Drieteenstrandlopers liet zien dat deze strandlopers een grote diversiteit aan prooisoorten benutten. Met behulp van DNA metabarcoding van uitwerpselen van Drieteenstrandlopers hebben we het volledige menu van in de nazomer in de Waddenzee opvettende en ruiende Drieteenstrandlopers in kaart gebracht. Gezien de grote diversiteit aan aanwezige prooisoorten in de Waddenzee was het opmerkelijk dat 94% van de monsters de Gewone Garnaal Crangon crangon bevatte. Gewone Garnalen waren naast de Strandkrab Carcinus maenas bovendien het meest talrijk in de verzamelde uitwerpselen. Hoewel Drieteenstrandlopers zich dus kunnen voeden met een grote diversiteit aan evertebraten, benutten ze in de Waddenzee tijdens de zuidwaartse trek voornamelijk de Gewone Garnaal.
Supplementary Material is available online www.ardeajournal.nl/supplement/s110-187-#199.pdf
SUPPLEMENTARY MATERIAL
Table S1.
Summary of the 28 studies on Sanderling that were reviewed. Per study, species were sorted on a higher taxonomic level if possible. Horizontal lines separate studies and within studies lines separate method, location, habitat or season.
Continued.
Continued.
Continued.
Continued.
Continued.
References
Alexander S.A., Hobson K.A., Gratto-Trevor C.L. & Diamond A.W. 1996. Conventional and isotopic determinations of shorebird diets at an inland stopover: The importance of inverterbrates and Potamogeton pectinatus tubers. Can. J. Zool. 74:1057–1068. doi.org/10.1139/z96-117
Castro M., Suazo C.G., Quiroga E., et al. 2009. Diet selection of Sanderlings (Calidris alba) in Isla Guamblin National Park in the Chilean fjords. Ornitol. Neotrop. 20:.247–253.
Evans P.R., Brearey D.M. & Goodyer LR 1980. Studies on sanderling at Teesmouth, NE England. Wader Study Gr. Bull. 30: 18–20.
Grond K., Ntiamoa-Baidu Y., Piersma T. & Reneerkens J. 2015. Prey type and foraging ecology of Sanderlings Calidris alba in different climate zones: are tropical areas more favourable than temperate sites? PeerJ 3: e1125. doi.org/10.7717/peerj.1125
Harrington B.A., Antas P. de T.Z. & Silva F. 1986. Northward shorebird migration on the Atlantic coast of Southern Brazil. Vida Silv. Neotrop. 1: 45–54.
Hernández M.A. & Bala L.O. 2005. Diet of the Sanderlings (Calidris alba) at Punta Norte, Peninsula Valdés, Argentina. Wader Study Gr. Bull. 30: 60–62.
Kelly C. 2008. Sanderling studies on the Wash. Wash Wader Ringing Group 2006–2007 Report.
Kober K. & Bairlein F. 2006. Shorebirds of the Bragantinian Peninsula I. Prey availability and shorebird consumption at a tropical site in northern Brazil. Ornitol. Neotrop. 17: 531–548.
Loonstra A.H.J., Piersma T. & Reneerkens J. 2016. Staging duration and passage population size of sanderlings in the Western Dutch Wadden Sea. Ardea 104: 49–61. doi.org/10.5253/arde.v104i1.a4
Lourenço P.M., Alves J.A., Catry T. & Granadeiro J.P. 2015. Foraging ecology of sanderlings Calidris alba wintering in estuarine and non-estuarine intertidal areas. J. Sea Res. 104: 33–40. doi.org/10.1016/j.seares.2015.06.013
Lourenço P.M., Catry T., Lopes R.J., et al. 2017. Invisible trophic links? Quantifying the importance of non-standard food sources for key intertidal avian predators in the Eastern Atlantic. Mar. Ecol. Prog. Ser. 563: 219–232. doi.org/10.3354/meps11979
Lourenço P.M., Catry T., Piersma T. & Granadeiro J.P. 2016. Comparative feeding ecology of shorebirds wintering at Banc d'Arguin, Mauritania. Estuaries and Coasts 39: 855–865. doi.org/10.1007/s12237-015-0029-1
Maron J.L. & Myers J.P. 1985. Seasonal changes in feeding success, activity patterns, and weights of nonbreeding sanderlings (Calidris alba). Auk 102: 580–586. doi.org/10.1093/auk/102.3.580
Masero J.A. 2003. Assessing alternative anthropogenic habitats for conserving waterbirds: Salinas as buffer areas against the impact of natural habitat loss for shorebirds. Biodivers. Conserv. 12: 1157–1173. doi.org/10.1023/A:1023021320448
Myers J.P., Williams S.L. & Pitelka F.A. 1980. An experimental analysis of prey availability for sanderlings (Aves: Scolopacidae) feeding on sandy beach crustaceans. Can. J. Zool. 58: 1564–1574. doi.org/10.1139/z80-216
Ntiamoa-Baidu Y., Piersma T. & Wiersma P. et al. 1998 Water depth selection, daily feeding routines and diets of waterbirds in coastal lagoons in Ghana. Ibis (Lond 1859) 140: 89–103. doi.org/10.1111/j.1474-919x.1998.tb04545.x
Nuka T., Miyazaki T., Norman C.P. & Kuwabara K. 2005. Feeding behavior and effect of prey availability on Sanderling Calidris alba distribution on Kujukuri Beach. Ornithol. Sci. 4: 139–146. doi.org/10.2326/osj.4.139
Perez-Hurtado A., Goss-Custard J.D. & Garcia F. 1997. The diet of wintering waders in Cádiz Bay, southwest Spain. Bird Study 44: 45–52. doi.org/10.1080/00063659709461037
Petracci P.F. 2002. Diet of Sanderling in Buenos Aires Province, Argentina. Waterbirds 25: 366–370. doi.org/10.1675/1524-4695(2002)025[0366:dosiba]2.0.co;2
Piersma T. & Ntiamoa-Baidu Y. 1995. Waterbird ecology and the management of coastal wetlands in Ghana. NlOZ Report 1995. 6: 1–105.
Quartey J.K. 2018. Foraging ecology of sanderling Calidris alba on the western coast of Ghana. PhD Thesis. University of Ghana, Ghana.
Reeder W.G. 1951. Stomach analysis of a group of shorebirds. Condor 53: 43–45. doi.org/ https://doi.org/10.2307/1364586
Reneerkens J., Benhoussa A., Boland H. et al. 2009. Sanderlings using African-Eurasian flyways: A review of current knowledge. Wader Study Gr. Bull. 116: 2–20.
Sallaberry M., Tabilo E., Klesse C. & Abarca J. 1996. The Chilean shorebird network (RECAP). Int. Wader Stud. 8: 71–78.
Tsipoura N. & Burger J. 1999. Shorebird diet during spring migration stopover on Delaware Bay. Condor 101: 635–644. doi.org/10.2307/1370193
Vanermen N., Stienen E.W.M., de Meulenaer B. et al. .2009. Low dietary importance of polychaetes in opportunistic feeding Sanderlings Calidris alba on Belgian beaches. Ardea 97: 81–87. doi.org/10.5253/078.097.0110
Vignisson S.R. 2019. Fæðuval vaðfugla í fjörum á fari um Reykjanesskaga Fæðuval vaðfugla í fjörum á fari um Reykjanesskaga. MSc Thesis. University of Iceland, Iceland.
Wirta H.K., Vesterinen E.J., Hambäck P.A. et al. 2015. Exposing the structure of an Arctic food web. Ecol Evol 5: 3842–3856. doi.org/10.1002/ece3.1647
Table S2.
List of species that were encountered more than five times by the benthic sampling programme near Griend. Counts refer to the number of times the species was encountered. Species are sorted by count within Class or in the case of Crustacea, by sub-phylum (separated by horizontal lines).
Table S3.
Date and location of the collected faecal samples.
Table S4.
Taxa detected by DNA barcoding of Sanderling excrements. Bold cells mark the macrobenthic species that were included in the study. ‘Total reads’ indicates the sum of the number of reads of all samples. FOO is the frequency of occurrence.
Continued.
Continued.
Continued.
Table S5.
Results of the simper analysis by the months that were compared. ‘average’ gives the average contribution of a species to the overall dissimilarity; SD is the standard deviation of the contribution. ‘ratio’ gives the ratio between average and SD. ‘ava’ and ‘avb’ give the average abundance of each month in the comparison. ‘cumsum’ indicates the cumulative contribution of each species (%) to the dissimilarity between two months.
Continued.
Continued.
Continued.
Continued.
Continued.
