To identify factors potentially limiting the abundance of ducks at the northern latitude of Lake Ontario, we investigated whether the lipid levels and foraging behavior of sea ducks wintering there were best explained by endogenous or exogenous mechanisms. We used an information-theoretic approach to evaluate if date (14 December–15 March; endogenous mechanism), percentage ice cover of Lake Ontario (long-term exogenous mechanism), short-term ambient temperature (exogenous mechanism), or some combination of these factors best explained variation in lipid reserves and foraging effort of the Bufflehead (Bucephala albeola), Common Goldeneye (B. clangula), and Long-tailed Duck (Clangula hyemalis). In the Long-tailed Duck, date explained lipid declines best, whereas in the Bufflehead and Common Goldeneye lipid reserves varied negatively with percentage ice cover. Conditions at our study site appeared to necessitate foraging because all species foraged, on average, ≥69% of daylight hours, and they increased foraging through the winter and when temperatures decreased. We calculated the number of days until lipid reserves should reach zero (i.e., fasting potential). Fasting potential differed in December, when Buffleheads (11 fasting days) had proportionately more lipid than did Common Goldeneyes (8 days) and Long-tailed Ducks (5 days). By March, the fasting potentials of the three species were similar (4 days), suggesting their survival strategies during similar winter conditions differed. Spatial and temporal modeling of temperature and ice cover with sea duck abundance at Lake Ontario and other Great Lakes may clarify factors that limit sea duck use of the region during winter.