Prey availability and foraging activity by tundra-nesting sea ducks: Strong preference for specific wetland types

Abstract

Wetlands in Arctic tundra support abundant breeding waterbirds. Wetland types differing in area, depth, vegetation, and invertebrate biomass density may vary in importance to birds, and in vulnerability to climate change. We studied availability and use of different wetland types by prelaying females of four species of sea ducks (Mergini) breeding on the Arctic Coastal Plain of Alaska, USA: long-tailed ducks (Clangula hyemalis) and Steller's (Polysticta stelleri), spectacled (Somateria fischeri), and king eiders (Somateria spectabilis). All four species preferred shallow vegetated wetlands versus deeper lakes. The ducks spent almost all their active time feeding, but their occurrence in different wetland types was not affected by the relative biomass density of known prey or of all invertebrates that we sampled combined. Sea ducks strongly preferred wetlands dominated by emergent and submersed Arctophila fulva over those dominated by the sedge Carex aquatilis, despite the much greater number, total area, and invertebrate biomass density of Carex wetlands. The hens depend heavily on local invertebrate prey for protein to produce eggs; thus, their preference for Arctophila wetlands likely reflects greater accessibility of prey in the near-surface canopy and detritus of Arctophila. Such shallow wetlands decreased substantially in number (-17%) and area (-30%) over 62 years before 2013 and appear highly susceptible to further declines with climate warming. Impacts on sea ducks of climate-driven changes in availability of important wetland types will depend on their adaptability in exploiting alternative wetlands.

Keywords: Arctic birds; climate change; eider; long‐tailed duck; thawing tundra; tundra wetlands; wetland invertebrates.

FIGURE 1

(a) Study location (black star) near Utqiaġvik, Alaska. (b) Study area showing locations of ground‐surveyed 400 × 400 m randomly‐located wetland plots (black rectangles) relative to human developments and major water bodies (gray). (c) Examples of adjacent wetlands of variable types in a landscape of polygonal tundra and thaw lake basins. Note extensive emergent Arctophila in some wetlands.

FIGURE 2

Percent occurrence of female sea ducks (black bars, interannual means in mid–late June 1999–2019 ± SD) versus (a) aggregate percent availability of wetland types in terms of total area (gray bars for different wetland types), or relative invertebrate biomass density in (b) cores (mg C/m² of benthic surface area) and (c) sweeps (mg C/m³ of water column within emergent vegetation). Numerical values for wetland use relative to surface area are in Tables 2 and A5. Numbers in parentheses indicate the number of years for which habitat use was sampled for each species. Selection trends are indicated (+, use > availability; =, use equal to availability; and –, use < availability [χ ², α = 0.05]).

FIGURE 3

Percentages of time spent in different activities by female sea ducks during the prebreeding period, 2017–2019. Values are annotated for each broad behavioral category, rescaled after removing time spent out of view (see Section 2.6); values annotated with ^ indicate the category immediately above the annotated value (e.g., reproductive behaviors of long‐tailed ducks). Values may not sum to 100% due to rounding. Means and SD are shown in Table A7, and specific behaviors within major categories are in Figure A4.

FIGURE A1

2017 ground‐based survey subunits. The 2017 survey was completed over 12 days in mid‐late June, surveying from west to east over this period.

FIGURE A2

Remote‐sensing workflow.

FIGURE A3

Invertebrate sampling locations, 2017–2018. Black triangles indicate wetlands sampled for invertebrates and estimate biomass density. Pair survey subareas are delineated in dashed lines.

FIGURE A4

Category‐specific breakdown of specific behaviors by focal sea duck females. Annotated values are percentages of a specific behavior comprising the overall category (generalized activity budgets are in Figure 3). Asterisks indicate very rare behavior categories (≤1%) not otherwise annotated. Means and standard deviations are in Table A7. Values may not sum to 100% due to rounding.