Study area and sample collection

The Bijagós Archipelago, located off the coast of West-Africa in Guinea-Bissau is composed of 88 islands and is well known for its rich biodiversity and cultural heritage (Campredon & Catry 2017). The climate is tropical and bi-seasonal, with a dry season from November to May followed by an intense wet season characterized by the heavy rains common in the region (Pennober 1999, Campredon & Catry 2017). Tidal amplitude is very high in this continental archipelago, reaching up to 4.5 m in spring tides and exposing up to 450 km² of intertidal flats twice a day during low tide (Pennober 1999, Granadeiro et al. 2021, Hill et al. 2021, Henriques et al. 2022b). In this study, data was collected across seven study sites at three islands (Figure 1) in order to capture the variability of this ecosystem. Given wader specific associations with intertidal sub-habitats, no species was sampled in every site in each season (Table S1), and therefore site effect is not investigated here as the sampling regimen was not designed to capture any potential spatial variation.

Flock of Ringed Plovers and Curlew Sandpipers gathering with the incoming tide at Bubaque island, Bijagós Archipelago, Guinea-Bissau ???(photo APC, January 2019).

Sanderlings feeding in the surf of Bruce beach, Bijagós Archipelago, Guinea-Bissau ???(photo APC, March 2019).

Eight shorebird species were targeted: Common Ringed Plover Charadrius hiaticula, Curlew Sandpiper Calidris ferruginea, Sanderling Calidris alba, Red Knot Calidris canutus, Common Redshank Tringa totanus, Grey Plover Pluvialis squatarola, Bar-tailed Godwit Limosa lapponica and Eurasian Whimbrel Numenius phaeopus. Five of the sampled sites were visited monthly from October to May during three consecutive winter seasons (2017–2018, 2018–2019 and 2019–2020), and faeces were collected opportunistically from the mudflats when a mono-specific flock was observed foraging (searching the area for fresh faeces once individuals were observed defecating and/or immediately after the flock left). Additionally, fresh faeces were also collected from single-species keeping cages when birds were captured with mist nets during ringing operations (Alves et al. 2021; Table 1). In both cases only intact faeces were collected, hence, those trampled were not retrieved. Collected samples were stored separately in Eppendorf tubes or paper envelopes, labelled according to species and date, and then taken to the laboratory for analysis.

Figure 1.

Map of the Bijagós Archipelago, off the west coast of Africa, highlighting the locations of study sites where droppings were collected.

Table 1.

Total number of analysed droppings in each month (all winter seasons aggregated), for each shorebird species.

Diet reconstruction

Diet was reconstructed based on identifiable prey remains found in birds' faeces. Air dried samples were sorted using a 30× magnification stereo microscope (Leica Zoom 2000) and all identifiable prey remains (jaws, pincers, hinges, claws, body whorls or whole intact individuals) were separated from the remaining content. Prey remains from the droppings were identified by comparison with a hard-structure reference collection of individuals collected at the same sampling sites in a previous study (Coelho et al. 2022). To build this reference collection, the diagnostic structures of identified, intact prey individuals were manually extracted, labelled and stored separately.

The number of individuals in each sample was attained by matching left and right structures for most prey (pincers for crustaceans, hinges for bivalves and jaws for most polychaetes), while for gastropods the number of individuals was attained by using the number of body whorls found and for Glycera sp. polychaetes by dividing the number of similar sized jaws by four, as species of this genera have four jaws (Day 1967, Duijns et al. 2013). Identified diagnostic prey remains were measured to the nearest 0.1 mm and the biomass of each ingested individual estimated using previously published regression equations (Lourenço et al. 2016b, 2017), that first relate the size of the prey remain with the individual total body length and then to its respective biomass (mg of Ash Free Dry Mass; AFDM). Fish remains were not converted into biomass, as no equation converting remain size to body size was available; however, these never comprised more than 0.03% of the number of prey items of any of the three species where these were found (see below).

Data analysis

We distinguished three periods to decompose the shorebird season in the Bijagós (Table 1): (1) the arrival period (Oct–Nov), corresponding to the progressive arrival of shorebirds in the archipelago (Zwarts 1988, Salvig et al. 1997) during the end of the local wet season, after shorebirds have depleted their body stores during the post-breeding migration while simultaneously having to finding the best quality patches in a new environment and complete a full body moult (Conklin et al. 2013, Aharon-Rotman et al. 2016), (2) the mid-winter period (Dec–Feb), in the beginning of the dry season and when shorebirds are in a relative stasis, and (3) the period of fuelling (Mar–May), which corresponds to the end of the local dry season, when the climate is the driest and freshwater availability is reduced (Campredon & Catry 2017), and when shorebirds increase their food intake in order to store energy for their migratory flight towards the breeding areas (Zwarts & Dirksen 1990, Zwarts et al. 1990, Lindström & Piersma 1993).

Intraspecific dietary variation between periods was assessed for each shorebird species by comparing the proportions of individuals and biomass of each main prey group (polychaetes, bivalves, gastropods and crustaceans) found in the droppings. This was done using Generalized Linear Models (GLMs) with a Binomial error structure (or Quasibinomial when encountering overdispersion; Zuur et al. 2009). The response variable was a matched pair (i.e. matrix) of successes (number of items of the focal prey group or biomass of the focal prey group) and failures (number of items of the remaining prey groups or biomass of the remaining prey groups), and the explanatory variable was period. GLMs indicating significant differences between periods were followed by pairwise post-hoc Tukey tests. Red Knot was excluded from this analysis as this species was not commonly recorded in the field sites, which limited sample collection to a single period (arrival; Table 1).

In order to quantify the level of intraspecific dietary change between periods, we calculated the Schoener index of overlap using the average proportion of biomass (i.e. mean percentage from all droppings from that species) of each prey group (polychaetes, bivalves, crustaceans and gastropods) in each season (Lourenço et al. 2017). A mean Schoener overlap index value (± SD) was obtained through bootstrapping 999 times, using the ‘niche.overlap.boot’ function from the R package ‘spaa’ (Gotelli 2000, Zhang 2016). While this index varies between 0 and 1, overlaps were considered low (0–0.39), intermediate (0.4–0.6) or high (0.61–1), with low overlap levels indicating that diet has changed substantially (Grossman 1986, Lourenço et al. 2017, Maitra et al. 2020). All statistical analyses were conducted using the R environment for statistical programming v. 3.6.1 (R Development Core Team 2011).

Wintering shorebird diet

The overall diet of the eight shorebird species studied was mostly composed of four prey groups (polychaetes, bivalves, crustaceans and gastropods), while three also included a small percentage of fish (Redshank, Bar-tailed Godwit and Whimbrel; Table 2, Figure 2A). The diet of the smaller species was composed mainly of polychaetes, contributing 83% to the diet of Curlew Sandpiper, and close to 70% in Sanderling and Ringed Plover. The latter two species also consumed crustaceans (17% and 13%, respectively). Redshank and Bar-tailed Godwit had the most diverse diet, with a larger proportion of crustaceans (around 40%), but also considerable percentages of bivalves and polychaetes (both between 20 and 30%). Grey Plover also had a large proportion of crustaceans in its diet (33%) but the most abundant prey were polychaetes (44%). In the remaining two species, the diets consisted mostly of one prey group alone: 78% bivalves in Red Knots and 76% crustaceans in Whimbrel.

Regarding biomass, the percentages were similar to the number of individuals consumed by the three smaller species (Curlew Sandpiper, Sanderling and Ringed Plover), with polychaete biomass corresponding to between 60 and 70% of the total (Figure 2B). Despite also consuming more individual polychaetes than any other prey, most biomass in the Grey Plovers' diet consisted of crustaceans (c. 60%). This was also the case for Redshank, but not for Bar-tailed Godwit, as most of the biomass in their diet consisted of polychaetes (around 50%). The proportion of biomass for Red Knot and Whimbrel, reflected the proportion of individual prey in their diet, with 80% and 90% of biomass originating from bivalves and crustaceans respectively.

Table 2.

Composition of shorebird diet based on number of individuals found on droppings, expressed as the percentage of each taxon in each shorebird's diet.

Figure 2.

(A) Diet composition of eight shorebird species based on prey remains found in droppings, and (B) the reconstructed biomass (mg AFDM), expressed as the proportion of each prey group (%) on each shorebird's diet.

Seasonal variation of diet

In almost all shorebird species the proportion of prey groups in their diet varied significantly between periods, for at least one of the prey groups (Figures 3 and 4, Table S2). Only Whimbrels did not show significant differences between periods, maintaining a high proportion of crustaceans in their diet throughout the season (Table S3). Two of the smaller species (Ringed Plover and Curlew Sandpiper) had significantly more polychaetes in their diet during fuelling than during arrival, and while this pattern was also noticeable for the biomass values in both species, it was only significant for Curlew Sandpipers (Figure 3, Table S3). Furthermore, both species had lower crustacean biomass in the fuelling period, but the proportion of crustaceans in their diet was only significantly lower for Ringed Plovers, which also had lower proportion of gastropods. Sanderlings did not show any major changes in their diet, except for the proportion of polychaetes that was significantly lower in fuelling compared to mid-winter (Figure 3). None of the smaller species showed variation in the proportion of bivalves, which was very low in their overall diets (less than 5%).

Of the larger shorebirds, both Redshank and Bar-tailed Godwit had a higher proportion of bivalves in their diet during fuelling than mid-winter (mid-winter was also lower than during arrival for Bar-tailed Godwits), and the same occurred regarding biomass, despite not being significant for Redshanks (Figure 4, Table S3). Both these species had lower polychaete biomass in their diet in fuelling, in comparison to the remaining periods, but only Redshanks showed significantly lower proportion of polychaetes in fuelling compared to mid-winter. Grey Plovers showed significant variation only in the proportions of bivalve and gastropod biomass, which was lower during fuelling than during arrival (Figure 4). The proportion of consumed crustaceans (both in individuals and biomass) did not vary between periods for these three larger shorebirds, despite representing a considerable part of their overall diet (more than 30% of both individuals and biomass).

Diet change between periods was considerable for most species, with only two (Curlew Sandpiper and Bar-tailed Godwit) showing intermediate values of diet overlap, both between fuelling and another period (Table 3). Interestingly, half of the analysed shorebird species (Ringed Plover, Curlew Sandpiper and Redshank) showed the lowest overlap, and therefore the largest seasonal change, between the two periods furthest in time: arrival and fuelling. For the remaining larger species, Grey Plover had the lowest overlap, i.e. largest seasonal change, between arrival and midwinter, whilst for both Bar-tailed Godwit and Whimbrel this was recorded between mid-winter and fuelling.

Table 3.

Mean diet overlap (± SD) between seasons for each shorebird species quantified using Schoener index. Periods on the top row of comparisons were abbreviated as: A (arrival) and W (mid-winter). The pair of periods with the lowest overlap for each species is outlined in bold.

Figure 3.

Seasonal variation in diet composition across three periods of the non-breeding season, for the smaller shorebirds (Ringed Plover, Curlew Sandpiper and Sanderling, indicated in the top row), expressed as the proportion of individuals (‘Freq', light grey boxplots) and proportion of biomass (‘Biom’, estimated as mg AFDM, dark grey boxplots) of each given prey group (one per row), as indicated to the left of the plots (‘POLY’ = polychaetes, ‘BIV’ = bivalves, ‘GAS’ = gastropods, ‘CRU’ = crustaceans). Each species therefore has two columns of plots: ‘Freq’ and ‘Biom’. Letters above each plot indicate significant differences between periods. See Table S2 and S3 for model and pairwise test results and associated P-values.

Shorebird diet reconstruction in the Bijagós

The diet of most species was in line with the few previous descriptions for the Bijagós and also with the much wider contributions from other wintering sites (Pienkowski 1982, Moreira 1996, Lourenço et al. 2016b, 2017). For example, Ringed Plovers and Curlew Sandpipers were found to consume a high proportion of polychaetes (Puttick 1978, Pienkowski 1982, Perez-Hurtado et al. 1997, Lourenço et al. 2017), while Red Knots almost exclusively consumed bivalves, just as they do in most parts of the world (Piersma 1991, Moreira 1994, van Gils et al. 2013, Yang et al. 2013, Sturbois et al. 2015, Lourenço et al. 2017). Redshanks, on the other hand, are more generalist predators, targeting different prey groups according to local availability at different sites throughout the flyway (Goss-Custard 1977, Moreira 1996, Sánchez et al. 2005). This is in line with the present findings, as Redshank diet consisted of wide range of prey of the most abundant groups, from polychaetes to crustaceans and bivalves. The diet of the crab specialist Whimbrel was also similar to what had previously been reported for this area, feeding almost exclusively on West African Fiddler Crabs Afruca tangeri (Zwarts 1985, 1988, Lourenço et al. 2017).

However, likely due to the larger spatio-temporal sampling achieved in this study, with seven mudflats sampled throughout the non-breeding season, some results deviated considerably from what was expected from literature. Bar-tailed Godwits mostly target polychaetes in wintering areas in Europe (Scheiffarth 2001, Duijns et al. 2013), however, in the Bijagós they consumed as many polychaetes as Ghost Shrimps Balsscallichirus balssi (23% of total consumed individuals for each prey). In terms of biomass, this translated to about 50% polychaetes but only about 30% crustaceans, including other species besides the low energy but easy digestible ghost shrimps, such as the more energetic but harder to digest crabs. Another high-Arctic breeding species, the Grey Plover is known to feed mostly on polychaetes in European estuaries (Kersten & Piersma 1984, Pienkowski 1982, Moreira 1996), whilst in the Bijagós it consumed almost as many crustaceans as polychaetes (33% and 44%, respectively).

Some authors have demonstrated the importance of Fiddler Crabs, a widely available species in the Bijagós (Paulino et al. 2021), as a key prey in the diet of the most of shorebirds in the archipelago (Zwarts 1985, Lourenço et al. 2017, Carneiro et al. 2021). Our findings corroborate this, but their relative importance appears to be much lower than previously suggested. In particular, Curlew Sandpiper and Sanderling diets contained a proportion of crustacean biomass estimated to be less than half of what was previously reported (only 8% and 17%, respectively). For Redshank and Grey Plover, crustaceans consistently represented about 60% of consumed biomass, including not only Fiddler Crabs but also other crustaceans such as Ghost Shrimp and Anthuridae, revealing a much lower proportion of this prey group than the c. 85% previously reported by Lourenço et al. (2017). These differences in diet composition are likely due to the much wider sampling effort of the present study, which encompasses seven different sites on different islands, with distinct macrozoobenthic community compositions (Coelho et al. 2022), and consequently different prey availability and abundance, thus likely representing more completely shorebird diet in the Bijagós Archipelago. Our results also indicate that the Ringed Plover's main prey are polychaetes, as recently reported (Lourenço et al. 2017) and not (or no longer) the Fiddler Crab as reported several decades ago (Zwarts 1985). Interestingly, Whimbrel, Redshank and Bar-tailed Godwit also occasionally include fish in their diet, which suggests that despite the difficulty in catching and handling such prey, these shorebirds are able to take advantage of its very high quality when conditions allow, for example when fish get trapped in small pools after the receding tide.

This was a surprising finding particularly for Whimbrel, which have previously been reported to feed exclusively on crabs in the Bijagós (Lourenço et al. 2017), although a more recent study indicates that Ghost Shrimp is also consumed (Carneiro et al. 2021). Our results show that despite crabs being indeed the main source of biomass (90%), Whimbrel also occasionally included a much more varied prey set in their diet, specifically mud snails (Hydrobiidae), Ghost Shrimps and fish (Table2).

Seasonal variation of shorebird diet

Shorebird diet changed seasonally with some species showing opposite patterns of change. Two of the smaller sized species, the Ringed Plover and Curlew Sandpiper, included more polychaetes in their diet towards the end of the season, while simultaneously decreasing the amount of gastropods and crustaceans. Soft-bodied polychaetes are theoretically very profitable and are in fact preferred by many shorebirds even when other prey with a high shell:flesh ratio are available in higher densities (Kalejta 1993). While both these shorebird species consume mostly polychaetes, this result appears to be intuitive if seen as an increase of the most preferred and profitable prey for these species, the polychaetes. Interestingly, the opposite pattern was found for larger bodied species, in particular for Redshank and Bar-tailed Godwit that displayed a change in diet from polychaetes to bivalves by the end of the season. This is an interesting result, as it was expected that polychaetes would also be the most profitable prey item for these two larger species, specifically for the polychaete specialist Bar-tailed Godwit (Scheiffarth 2001, Duijns et al. 2013). Even though polychaete abundance does not decrease in the Bijagós towards the end of winter, the average size of polychaetes found in the archipelago is very small (mean size of 5 mm; Coelho et al. 2022). Given such small size, in relation to polychaetes found in other sites, this prey may still be relatively profitable for smaller sized shorebirds, but less so for larger ones. Instead, the increase in the proportion of bivalves in the diet of these larger shorebirds may be due to the increasing abundance of these prey in the Bijagós towards the end of the season, possibly due to recruitment during this period (Coelho et al. 2022).

The macrozoobenthic community in the Bijagós is diverse and spatially variable, with different sites having distinct macrozoobenthic compositions (Lourenço et al. 2018, Coelho et al. 2022). Bivalves represent the majority of the biomass available and increase in density during the shorebirds' presence in the archipelago, but other prey groups show spatial differences (Coelho et al. 2022). Polychaetes, the most abundant group, as well as gastropods and crustaceans, increase in density in only a few sites, while remaining constant in others (Coelho et al. 2022). The observed changes in shorebird diet could be influenced by this spatial variation in prey availability, and future studies should take this into consideration. While this study was not designed to address spatial variation in shorebird diet (Table S1), the fact that several species showed opposite patterns in diet variation (e.g. some increasing consumption of bivalves and other decreasing), and that prey density does not decrease (Coelho et al. 2022), suggests that diet variation is caused by the responses of shorebirds to increased energetic requirements.

Some species however showed much smaller dietary variation. For example, Grey Plover maintained the proportion of its two main prey groups constant throughout the season (polychaetes and crustaceans), but there was a significant reduction of the other two less important prey groups (bivalves and gastropods) in the fuelling period compared to the arrival period, despite both increasing in abundance in the archipelago (Coelho et al. 2022). These results suggest that despite having the capacity to change their diet in accordance with the available prey species in a certain site (Pienkowski 1982), Grey Plover wintering in the Bijagós actively select their preferred prey groups during the fuelling period. However, as the proportion of those least preferred prey was so small to begin with, these results should be interpreted with caution. Similarly, Sanderling diet varied only in the proportion of polychaetes, that was lower during fuelling compared to mid-winter, although it is unclear which prey group replaced polychaetes in the fuelling period, possibly due to the consumption of soft-bodied prey that was therefore missed.

The shorebird species that demonstrated least dietary variation throughout the season was the Whimbrel, which may be explained by the extremely high local availability of its most consumed prey, the Fiddler Crab (Paulino et al. 2021). This may also be an important factor contributing to the Whimbrel's increasing population in the Bijagós, whereas on the rest of the flyway the species remains stable (Henriques et al. 2022a).

Dietary overlap between the three periods

Intraspecific seasonal overlap between periods was often high, corroborating the previous results and indicating the existence of only small adjustments in the proportion of consumed prey groups. Given that shorebirds tend to display high levels of site-fidelity (Alves et al. 2013, Lourenço et al. 2016a), dietary changes within the same season are expected to be minimal. However, the reported temporal change in the macrozoobenthic community in the Bijagós towards the end of the shorebird season (Coelho et al. 2022), contemporaneous with the fuelling phase of these species, could lead to dietary changes. In fact, the most pronounced seasonal changes in diet were indeed recorded between fuelling and another period for all shorebird species (except Grey Plovers), indicating that almost all species adjusted their diet closer to their departure, thus likely responding to the higher energetic requirements of this period. While some species made this dietary adjustment gradually over the season, with the largest differences found between the arrival and fuelling periods, others changed more rapidly, between mid-winter and fuelling. The lowest overlap, between Grey Plovers' diet in arrival and mid-winter, suggests that it may have been less selective upon arrival, consuming some of its least preferred prey (gastropods and bivalves), but changing shortly after, to the preferred polychaetes and crustaceans, which it kept consuming in similar proportions during mid-winter and fuelling.

Understanding dietary flexibility in shorebirds may be important to better assess their resilience to environmental change. However, despite their general ability to change their diet in the Bijagós, particularly during fuelling, it remains unknown if such a dietary shift provides the required energetic input. The fact that the biomass densities of the most consumed prey groups do not decrease in the course of the season in the Bijagós (Coelho et al. 2022) may point towards a frequent renewal of resources, but this merits further investigation. Daily energy budget estimates are necessary in order to understand how the shorebirds make ends meet.