Dabbling ducks choose a multitude of habitats throughout their life cycle. This choice depends on the abundance, diversity, and accessibility of food resources. Wetlands such as the Marais breton and Marais poitevin (Atlantic coast, France) are common habitats for several Anatidae, especially during their prenuptial migration. However, studies on the food ecology of Anatidae at stopover sites are limited. Therefore, this study focused on the Northern shoveler Spatula clypeata, a species that regularly inhabits the two marshes during the breeding and migration periods and is highly dependent on freshwater invertebrates as the food resource because of its bill morphology. Fifteen Northern shoveler were equipped with a GPS/GSM tag and monitored during their prenuptial migration. The study objectives were to understand the strategies used by the Northern shoveler to select the foraging sites and describe the characteristics of these sites (such as freshwater invertebrates’ abundance and diversity and the habitat type). In addition, home range (HR), space use, and habitat selection were studied.

Northern shoveler capture and tagging

This study was conducted on two large wetlands on the French Atlantic coast in Vendée: the MB (N2000 FR5212009 and Ramsar 2283) and the MP (N2000 FR5200659). During the prenuptial migration period, shovelers were captured using cage traps and live male or female shovelers as decoys. A camera (NATURACAM – STDX2) was positioned near each trap to monitor the presence of birds in the traps, which were kept every day from March 01 to March 17, 2020, and from March 01 to April 10, 2021. Overall, 15 shovelers were caught and equipped with a GPS-GSM tag (Ornitela, OrniTrack-E10, solar-powered GPS-GSM); these included 4 females (F) and 11 males (M) and 8 juveniles (less than two years old) and 7 adults (more than two years old). Eight individuals were caught in the MB and 7 in the MP. The equipment (GPS-GSM tag, harness, and metal tag) weighed less than 3% of the body mass. The shovelers were ethically captured and handled (Authorization from Ministry of Ecological Transition by Research Center on the Biology of Bird Populations PP: 1821).The movements of the individuals were monitored during the prenuptial period (from March 01, 2020, to April 30, 2020, and from March 01, 2021, to April 30, 2021) with a recording frequency of once every 5 minutes. Thus, the data analysed were obtained from the 15 birds that used either the MB or MP over a period of at least 15 days after being tagged (Table 1). This period corresponds to the minimum amount of time a majority of the individuals spend at the stopover site before leaving for the breeding site.

Trophic Resources and Environmental Parameter Measurement at the Feeding Sites

To describe the abundance and energy value of the freshwater invertebrates at the shovelers’ main feeding sites, invertebrates more than 200µm in size were sampled from March 1 to April 30, 2021, in some sites exploited by the tagged individuals. The sites, 31 in total, were selected for sampling only if the tagged birds exploited their freshwater sites for a minimum of 72 consecutive hours. The sites were classified into three categories according to the habitat type: wet meadow, pond, and channel (Fig. 2). In each site, freshwater invertebrates were sampled thrice using a plankton net (mesh of 200 µm, frame size of 35.5 x 15.0 cm) on a transect of 2 m at 35 cm of the water column, which corresponded to the shovelers’ feeding depth. The contents obtained using the net were preserved in 70° ethanol (Balcombe et al. 2005) and quickly analysed. The individuals from each sample were counted and identified using a binocular magnifier (Euromex, Series Z, 7-45 x). The invertebrates’ family was identified (except copepods, subclass; Cladocera, superorder; Hydrachnidia, suborder; and Ostracods, class) (Moisan 2010, Thorp and Rogers 2011).

The feeding sites were characterized by the environmental variables. The sediment height, water level, salinity (using a multiparameter probe VWR MU 6100 H Multimeter), and percentage cover of riparian helophytes and of emerged and submerged aquatic vegetation were noted in each site. Other factors were coded into three classes: sediment type (class 1: loamy sediment; 2: loamy / muddy; 3: muddy), slope (class 1: < 5% soft slope; 2: 5%–10% moderate slope; 3: > 10% steep slope), and habitat (class 1: meadows; 2: ponds; 3: channels). The surface of the sites was measured using the geographic information system QGIS (QGIS Development Team 2009). The site boundary corresponded to the hydraulic entity of the birds.

This two dataset describes the 15 Northern shoveler captured in stopover during the prenuptial migration (File :"data_shovelers.txt") and their feeding sites characteristics (File :"data_sites.txt") .