Background: Black-faced spoonbill (BFS) is a global endangered species, distributed only in the coastal zones of East Asia. Xinghua Bay is one of the main wintering sites and migration stopovers of BFS in mainland China. However, with the reduction and degradation of natural wetlands, it is uncertain whether the constructed wetland can provide habitat for the endangered BFS. Research on diet of BFS will help to understand their preference between natural and artificial wetlands, and also provide reference for their conservation and habitat restoration.

Results: In the early winter, the proportion of Palaemonidae in BFS’s food was as high as 74.4%, while that of other food was only 3.0% to 6.0%. In the late winter, the food contribution of BFS was as follow: Portunidae 39.3% > Palaemonidae 26.1% > Cyprinidae 8.8% > Mugilidae 8.5% > Gobiidae 7.3% > Crucian 5.1% > Whiteshrimp 4.8%. The proportion of Portunidae exceeded that of Palaemonidae, and together with Palaemonidae, it has become the main food of BFS in late winter.

Conclusion: The diet composition of BFS between the early and late winter was significantly different, which may be due to seasonal changes in food resources. Natural wetlands are the main feeding grounds of BFS, but artificial wetlands also provide them with supplementary feeding grounds and resting places. Aquaculture ponds play an important ecological function in maintaining the overwintering population of BFS in Xinghua Bay.

From December 2017 to February 2020, 45 potential food samples and 199 fecal samples of BFS were collected during six sampling period, of which Cyprinidae, Mugilidae, Portunidae, Gobiidae and Palaemonidae were collected from natural wetlands and Crucian (Carassius auratus) and Whiteshrimp (Litopenaeus Vannamei) were collected from artificial wetland. Their stable isotope values (δ13C, δ15N) were measured to obtain the food composition information of BFS.

The mean and standard deviation of stable isotope values of samples were calculated, and the outliers that deviated from the mean by more than twice the standard deviation were screened out. The 14 potential food sources were divided into 7 groups, of which Cyprinidae, Mugilidae, Portunidae, Gobiidae and Palaemonidae were collected from natural wetlands and Crucian and Whiteshrimp were collected from artificial wetland.

The differences of δ13C and δ15N values between fecal samples and potential food samples were compared by One-Way ANOVA analysis. All isotope data were expressed as Mean± standard deviation (Mean ± SE), and p < 0.05 was considered as significant difference.

The most commonly used model to study food composition by stable isotopic analysis is Bayesian mixing model. It can take into account the variable nondeterministic factors with relatively high accuracy and a wide range of application. We used the R package simmr, which can be used to process the mixing equations of stable isotope data within the framework of the bayesian mixing model (Parnell, Inger, Bearhop, & Jackson, 2010; Parnell et al., 2013). Fecal values were calibrated by the discrimination factors (δ13C=1.15‰, δ15N=2.91‰), which were estimated for fecal samples of birds in general (Caut, Angulo, & Courchamp, 2009).

Fecal values were calibrated by the discrimination factors (δ13C=1.15‰, δ15N=2.91‰), which were estimated for fecal samples of birds in general (Caut, Angulo, & Courchamp, 2009. doi:10.1111/j.1365-2664.2009.01620.x).