Exploring the variation in spoonbill chick diet and trophic niches between traditional and recently colonised sites in Portugal

The Eurasian Spoonbill Platalea leucorodia is currently experiencing an increasing population trend in Western Europe and colonising areas in its breeding range. In this study we assessed the diet and trophic ecology of spoonbill chicks in Portuguese colonies (Ria Formosa and Tagus Estuary) occupied at different years during the population expansion in the country. We accomplished this by combining diet analysis with blood and feather stable isotopic analysis. Spoonbills from the Portuguese colonies relied mainly (but not exclusively) on crustaceans (Ria Formosa = 75%, Tagus Estuary = 72.5%) and fish (Ria Formosa = 17.11%, Tagus Estuary = 37.50%) to feed their chicks. Crustaceans seems to be the most important prey group. Nevertheless, chicks from Ria Formosa were fed with prey from a higher trophic level than chicks from Tagus Estuary. In both colonies chicks were fed with available in habitats near the colony, although in Tagus Estuary most prey originated from freshwater habitats. Diet of chicks remained constant through the development period and the proportions of crustaceans and fish consumed differed from other colonies elsewhere in Western Europe (Spain and the Netherlands). We conclude that in Portugal, spoonbills feed their chicks with prey commonly found in the vicinity of colonies. Prey originating from artificial habitats, particularly the invasive Louisiana crayfish Procambarus clarkii in the more recently occupied colony, were the most important items in the diet of chicks.


Introduction
Several waterbird species are increasing in numbers in Europe. 1 There are multiple potential drivers for the expansion of waterbirds, with the main factors suggested being the increment in legal protection of species and habitats. 2,3 Another likely contributor to this trend is the creation of artificial wetland habitats such as rice fields, saltpans and fishponds that can act as alternatives to natural habitats. 4 These habitats are particularly relevant as foraging habitat, once they can provide high abundance of food resources during the breeding season. [5][6][7][8] Finally, the occurrence of exotic species that have the potential to become profitable food resources can also be regarded as a potential contributor to waterbird expansion. A remarkable example being the Louisiana Crayfish Procambarus clarkii that has become part of the diet of several European Ardeidae species. 9,10 The Eurasian Spoonbill Platalea leucorodia is a migratory waterbird, and the subspecies Platalea l. leucorodia breeds in Europe and winters across a wide latitudinal range, from south Europe to Senegal (Africa), using mostly coastal wetlands in both continents. 11 Spoonbills are one of these waterbird species that are increasing in numbers 3 and (re)colonising areas across the breeding range. 12 In Portugal, breeding spoonbills were absent until 1988, despite historical records of breeding pairs in the 17 th century. 13 Portugal was always a key area for spoonbills during stopover and winter, 14 and in 1988 the first returning breeding pairs were recorded in the area of Paúl do Boquilobo. 15 Ria Formosa was the next site to be recolonised only five years later, 14 and currently ten breeding colonies are known to exist in Portugal. 16 The breeding ecology and dietary choices of Eurasian spoonbills is well documented, especially for populations from the traditional European colonies, in the Netherlands Wadden Sea and in Spain (e.g. [17][18][19][20]. It is well known that the species forages in shallow waters, varying from saline to freshwater, mostly feeding on small fishes, crustaceans, amphibians, and other small invertebrates. 17,18,20 In contrast, information from other colonies remains extremely scarce, with only one more study available from Germany, but also referring to the Wadden Sea. 21 In order to investigate potential drivers of the increasing numbers in Portugal, as well as the contribution of dietary sources in making recently colonised potentially sites attractive for the species, we aimed to unravel chick diet and trophic ecology in traditional and recently established colonies in the country. Specifically, we assessed chick diet in Ria Formosa, one of the first Portuguese colonies (occupied since 1993), and in Tagus Estuary, a more recently occupied location (established in 2013), characterized for being close to an extensive area of rice fields. The major aim was to assess the diet and trophic ecology of spoonbill chicks. Using traditional dietary methods, we (1) determine diet composition for chicks of both colonies and predict that in Tagus Estuary spoonbills would rely on the invasive Louisiana Crayfish that is abundant in the nearby rice fields. We then use stable isotope analysis on blood and feathers to (2) provide a more comprehensive assessment of the diet during the entire chick growth period and to determine the origin of prey (freshwater or saltwater) and the dietary niche width. We hypothesise that chicks from the Tagus Estuary would be fed with prey items from both saline and freshwater habitats (as adults are known to use the rice fields) and, therefore, should present a larger niche width than birds from Ria Formosa. Finally, we compare the diet of chicks in these two colonies to that of European colonies in Spain and the Netherlands (compiled from published sources) and explored possible changes in the diet of chicks along their development (using the stable isotope date) as this seems to be important in the Netherlands (at least in late born chicks).

Study sites and sample collection
We selected two Portuguese breeding colonies ( Figure 1) considering accessibility to the nests while minimizing the disturbance to breeding birds. The colony in Ria Formosa Natural Park (RF) (36°59'N, 7°55'W) represents the oldest colonies in Portugal (occupied since 1993). It is located inside a coastal protected area, separated from the Atlantic Ocean by a system of small barrier islands and sand banks, encompassing several habitats such as marshes, intertidal flats, fresh and brackish lagoons, saltpans, dune banks and agricultural fields. During the 2017 breeding season we visited the Ria Formosa colony twice (in May and late June) and the Tagus Estuary colony once (in early July), and again both colonies once in 2020 (in June). In 2017 we collected 150µl of blood from the brachial vein (TE= 11, RF=13) and the 5th primary feathers (cut at the base) of pre-fledging spoonbill chicks (TE=4, RF=28) from different nests in the colony. Blood and feathers were stored in eppendorfs (containing alcohol 70%) and plastic bags, respectively, until stable isotope analysis (see below). During these visits we also collected regurgitations and fresh faecal samples opportunistically, as spoonbill chicks usually regurgitate spontaneously when handled. 19 Whilst the number of fresh faecal samples collected was larger in 2017 (TE=5, RF=5) than in 2020 (TE = 1, RF = 0), this was the opposite regarding regurgitations (2017: TE = 1, RF = 6; 2020: TE = 3, RF=11). In Ria Formosa we could also collect dry faecal samples from the nests (RF=11) in 2017. These dry faeces are usually deposited in nest cups and are clearly different than fresh faeces. All regurgitation and faeces samples were stored in plastic bags and frozen (at -20°C) until laboratorial analysis.

Diet
To reconstruct the diet of chicks we searched for prey items with a stereomicroscope (Wild Heerbrugg M3B) in regurgitations and faeces. When fish and crustaceans were found almost intact, they were identified to the lowest taxonomic level (using a stereomicroscope at 40x magnification). Prawns were identified based on the identification key proposed by González-Ortegón and Cuesta, 22 and fish otoliths were identified following published sources. 23,24 Other smaller remains (often very small fragments) were grouped by colour: dark red parts were considered as belonging to Louisiana Crayfish (that has a red exoskeleton) and orange parts considered as remains of other crustaceans (shrimps, prawns or crabs). These small remains were only considered for frequency of occurrence. Small gastropods, bivalves and vegetal matter were not identified to lower taxa since they were possibly ingested accidentally. We calculated numerical frequency (number of estimated individual prey based on prey fragments) using only the regurgitation samples due to the low detectability of some prey on faeces samples 25,26 which could bias the results. Faecal samples were thus only used to estimate frequency of occurrence. Results present prey at family level, unless otherwise indicated.

Stable isotope analysis
We used stable isotope analysis of carbon (C) and nitrogen (N) to infer the dietary space of chicks during the entire chick development. Stable isotopic ratio of carbon (δ 13 C) and nitrogen (δ 15 N) present in consumer's tissues are correlated to those of their diet items in a predictable way: δ 15 N increases approximately 3 to 5 parts per thousand (‰) between successive trophic levels [27][28][29] ; while δ 13 C discriminates between freshwater and marine items, with lower values originating from freshwater habitats. [30][31][32] We sampled ca. 25 day-age chicks, and analysed the stable isotopes form: blood, to attain information concerning their entire growth period until that moment 33,34 ; and of different parts of the primary feathers (base and tip) reflecting different periods within the initial 25 days of live, [35][36][37] that allow us to compare between earlier (tip) and later (base) phases of chick development. We did not analyse the middle part of the feather to ensure the necessary transition from early to late phase of chick growth.
Blood samples were collected into an empty eppendorf tube and dried at 50°C during 48h leaving whole blood solid particles. From each 5 th primary feather we extracted two sections: one from the base of the feather (close to its calamus), reflecting the later chick development phase, and another from the tip, reflecting the earlier chick development phase. Feather sections were cleaned in 2:1 chloroform and methanol solution to remove surface impurities and lipids and dried at 60°C for 24 hours. Finally, nitrogen and carbon isotopic ratios were measured by continuous flow-isotope ratio mass spectrometry (CF-IMRS). Results were expressed in δ notation as result of the equation: where R= 13 C/ 12 C or 15 N/ 14 N and standards were the PeeDee Belemnite (PDB) in the case of δ 13 C, and atmospheric nitrogen (N 2 ) in the case of δ 15 N. Carbon and nitrogen measurements had an analytical precision of 0.2‰.

Data analysis
Feather isotopic results were corrected to account for different isotopic assimilations of the tissues using the procedure described in Cherel 2014, 34 thus enabling to compare those with the results of blood. We first compared δ 13 C and δ 15 N values of isotope samples from Ria Formosa with the ones from Tagus Estuary using one way analysis of variance (ANOVA) for Carbon and the Kruskal-Wallis test to compare values of Nitrogen (as this data was not parametric). Next, we compared (within colony) the values of δ 13 C and δ 15 N in the different tissues, using ANOVA in all cases except for Carbon in Ria Formosa and Nitrogen in Tagus Estuary were Kruskal-Wallis test was used. Significant differences between groups were investigated with post-hoc tests (Tukey test for ANOVA and Pairwise Wilcoxon tests for Kruskal-Wallis).
To assess the isotopic niches of spoonbill chicks from the different colonies and periods (i.e. sampled tissues), "SI-BER", Stable Isotope Bayesian Ellipses in R, 38 was used to calculate the area of standard ellipses (SEAc) corrected for unbalanced and small sample sizes. 39 In addition, we calculated the Bayesian estimates of the standard ellipses (SEA B ) to test for differences in size of niche widths. 39 All statistical analyses were performed in R environment. 40

Diet
Samples from both regurgitations and faeces originating from Ria Formosa and Tagus Estuary were mostly composed by crustaceans (of order Decapoda) and fish (Table 1). Other prey items were gastropods, isopods, and insects. Crustaceans occurred in all samples from both colonies and years. Overall, all unidentified crustacean from regurgitations belonged to the order Decapoda and were mostly shrimps (91.67%) and were always most frequent in the regurgitations samples than fish. In Ria Formosa, the most frequent identified crustaceans (at species level) in 2017 was Carcinus maenas (50%) whereas in 2020 it was Palaemon elegans (62.96%) (Figure 2). Procambarus clarkii (the only identified species from the Cambaridae family), which was absent Table 1. Percentage of occurrence of each prey item (at family level) in spoonbill chick regurgitations and faeces (N RF = 33, N TE = 10) and numerical frequency of prey items in regurgitations (N RF = 152, N TE = 48) from the Ria Formosa (RF) and the Tagus Estuary (TE). Category "Others" includes gastropods, isopods and insects.  from Ria Formosa, was present in all dietary samples from Tagus Estuary. The only other crustacean identified in the Tagus was P. varians ( Figure 2). The occurrence of fish in Ria Formosa samples was constant in both years (54.55%), yet the numerical frequency differed between 2017 (22.64%) and 2020 (14.14%). All regurgitations from the Tagus Estuary of 2020 included fish, whereas the single sample from 2017 was only composed by P. clarkii. The numerical frequency of fish was higher in regurgitations from Tagus Estuary (2020) than from Ria Formosa. In the Tagus, identified fish belonged to the Gobiidae and Poeciliidae families whereas in the Ria Formosa, the most frequent identified prey belonged to the Gobiidae family in both years (2017: 75%, 2020: 66.67%) (Figure 2). The families Atherinidae and Mugilidae were only identified in faeces samples, however, Atheneridae was the most frequent identified family of fish in faeces from Ria Formosa (69.23%). In these samples we were only able of identifying fish otoliths, as other prey items were very digested.
Niche width estimated as SEA C was wider in the Tagus (Figure 3). Indeed, SEA B size of blood was significantly larger in the Tagus than in Ria Formosa (p=0). Regarding niche width of feathers there was no difference between the tip and the base in Tagus (p = 0.09) or in Ria Formosa (p = 0.25).

Diet
Spoonbills from Ria Formosa feed their chicks mostly with crustaceans of the genus Palaemon, which are present in the lagoons and saltpans in the vicinity of the colony, likely Table 2. Stable isotope ratio values for δ 13 C and δ 15 N (mean ± SD) from blood and feathers from spoonbill chicks of Ria Formosa and Tagus Estuary. N is the number of each tissue samples analysed.  making this an easy and predictable prey. In fact, many other birds, such as Calidris alpina and Sterna albifrons, also feed on these prawns in Ria Formosa. 41,42 Nevertheless, and probably due to the small biomass of these crustaceans, 43 spoonbills also provided their chicks with fish, that are of higher energetic value. 20,44 The most consumed fish families, Atherinidae and Gobbidae are common in Ria Formosa 45,46 suggesting that these may also be a relatively easy to catch prey due to their availability. On the Tagus Estuary the general chick dietary pattern was similar, with high occurrence of crustaceans. The main difference was the presence of the Louisiana Crayfish in all samples, which is known to be very common in the rice fields nearby the colony. 47 This result is in accordance with the reported diet from other waterbird species in southern Europe that also consume this invasive species. 9,10,48 To the best of our knowledge, only two other studies confirmed the ingestion of this crayfish by spoonbills, both in South-West Spain: in Odiel Estuary and in Guadalquivir marshes. 10,18 The widespread consumption of this previously absent prey may be advantageous, as some waterbird species like the Eurasian Bittern Botaurus stellaris and the Grey Heron Ardea cinerea, show positive population trends as the consumption of this prey increased. 10,49 However, for other species, such as the Glossy Ibis Plegadis falcinellus in Doñana (South Spain), the consumption of this crayfish alone cannot underline the reported population increase, as this has not been described as an important prey in is diet, 50 although this may not be the case in more recent years and at other wetlands where the species has been frequently observed consuming this crayfish (pers. obs). Other seemingly important prey in Tagus Estuary are fish of the families Gobidae and Poeciliidae, and the shrimp Palaemon varians. Both prey types are very common in Tagus Estuary, especially the family Gobidae 51,52 and P. varians, is present both in intertidal mudflats and in saltpans. [52][53][54] Stable isotope analysis Trophic niche of earlier and late chick development phase (assessed by sampling different parts of feathers) revealed no differences on both colonies, implying a similar diet of the chicks since feathers start to be developed. However, blood δ 15 N in Ria Formosa differed from feather base. We did not expect this result but it may be possible that spoonbills chicks in Ria Formosa changed their diet only few days before sampling collection and the quicker turn over of blood 55 reflected that. The small difference in δ 15 N values between colonies may corroborate the differences found in dietary analysis. Ria Formosa highest δ 15 N indicates that spoonbills chicks from there are fed with prey of a higher trophic level than chicks from Tagus Estuary. [27][28][29] Indeed, the higher occurrence of the Louisiana crayfish may result in lower δ 15 N values in Tagus Estuary, as this crustacean occupies a lower trophic level than fish. Regarding the differences in δ13C values between colonies suggest that spoonbills feed their chick from prey items obtained in the vicinity of the colonies. This is because higher values of δ13C indicate marine prey, which will occur around the saltmarsh island in Ria Formosa. i.e. around their colony, whereas the signature of freshwater prey is indicated by lower δ13C values, 30,32,56 suggesting that spoonbills in the Tagus Estuary depend less on the intertidal areas that also surround their colony, but more so on the vast rice-fields present in the vicinity of the colony. However, this pattern is less clear in the dietary analysis, where P. varians were common, suggesting that prey from more saline habitats is also taken. In accordance, the difference in sizes of niche width between colonies suggest indeed a more diverse diet in the Tagus Estuary colony 57,58 that was not detected based on the diet samples alone. This is likely a limitation of such analysis particularly when sample sizes are relatively small. In fact, the use of regurgitations to infer diet is always limited, representing a snapshot of the diet as it only includes items consumed on that day. It is also further constrained by different digestive rates of prey, which leads to the impossibility of identifying some prey items and the total absence of others. 25,59 Regarding faeces samples, this disadvantage is even larger because only hard parts are preserved. 25,59 The analysis of dry faeces constitutes another uncertainty of this study because they can remain in the nests for several years. This means that the dry faeces that we collected in Ria Formosa can originate from previous seasons, which may justify the presence of two families of fish (Atherinidae and Mugilidae) in only this type of samples. In order to further understand diet variation a larger number of samples should be collected and analysed, combined with isotopic SIAR mixing models 20,60,61 to estimate the percentage of each prey type.

Comparison with European traditional colonies
We found that in Portuguese colonies spoonbills appear to consume more crustaceans than fish, contrary to what has been described for traditional Dutch and Spanish colonies, where spoonbills heavily rely on small fish. 18,20 Also, in Germany, Enners 2020, 21 found a diet similar to the traditional colonies with an higher consumption of fish over crustaceans. The results of stable isotopic analysis were in accordance with this because the values of δ15N from chicks of Portuguese colonies were overall lower than values found in studies from Dutch and German spoonbill chicks. 20,21,37 However, Veen 2012, 62 found that in Mauritania, the diet of the subspecies Platalea leucorodia balsaci varied between years, with a varying dominance of fish or crustaceans in the diet. Still, we only have two years of data for Ria Formosa (the first sampling year for Tagus Estuary was very incomplete) and the proportion of crustacean was in both years higher than fish.
In addition, the fact that in Portugal the diet of chicks is almost constant during the entire chick development period contrasts with what was reported for the Dutch colonies, where a shift from freshwater to marine diet in the middle of chick rearing season occurs. 37 This suggests that rather than being a species-specific pattern, diet switching may be related to prey availability in the vicinity of colonies.
Our results are, however, in accordance with the results from studies in other colonies in the sense that, even considered as a specialist waterbird, the diet of spoonbills seems to be driven by food availability 21,37,62 and by feeding in the surrounding of the colonies. 20 In the Tagus Estuary, the rice field seems to constitute an important freshwater foraging habitat where spoonbills can feed on Louisiana crayfish and small fish. In Ria Formosa, P. varians, seems to be a very important prey, that can be easily accessed in the intertidal lagoons or in saltpans in the vicinity of the breeding colony.

Conclusions
Our study reveals differences in dietary and tropic ecology of spoonbills chicks on recently established and traditional colonies. Furthermore, the reliance on prey available in artificial habitats, such as saltpans, where prawns are abundant, or rice fields where new prey as the Louisiana Crayfish is present, may be important in the establishment of new spoonbill breeding colonies across its range. However longterm diet studies are necessary in order to further investigate such possibility.