Winter hypoxia in shallow, ice‐covered lakes can be a significant limiting factor for overwintering fish populations. In this study we tested the hypothesis that low overwinter survival due to winter hypoxia is a limiting factor for a rare, adfluvial population of native Arctic Grayling Thymallus arcticus inhabiting Upper Red Rock Lake, Montana. We used a combined laboratory and telemetry study to document the extent of hypoxia over two winters and to assess the physiological tolerance, behavioral response, and winter survival in relation to hypoxia. In the laboratory, we observed a significant behavioral and physiological response to dissolved oxygen (DO) levels ≤ 4.0 mg/L and determined acute 24‐h LC50 values (concentration lethal to 50% of test fish) of 0.75 mg/L DO for adults and 1.50–1.96 mg/L for juveniles at temperatures of 1–3°C. In the field study, we observed dynamic DO concentrations (DO < 1.0 to 10.0 mg/L) during winter ice cover, ranging from persistent near‐anoxic conditions near the bottom to DO concentrations > 4.0 mg/L in the epilimnion. Radiotelemetry indicated adult winter survival rate was high (0.97 in 2014, 0.95 in 2015) and that telemetered fish selected deeper (>1 m), more oxygenated habitat during ice cover. Our study demonstrated that Arctic Grayling have a high tolerance to acute hypoxia exposure and exhibit a physiological and behavioral stress response to DO concentrations ≤ 4.0 mg/L. Although hypoxia was present in parts of the lake, sufficient suitable habitat with DO > 4.0 mg/L was available in the lake epilimnion in both study winters. However, winter conditions more severe than those observed during our 2‐year study occur periodically in the lake, and thus winterkill could still be a limiting factor for the population.

Low dissolved oxygen, or hypoxia, is a common phenomenon in ice-covered lakes in winter. We measured dissolved oxygen (DO) before, during, and after ice-over to characterize the timing, severity, and spatial variability of winter hypoxia in Upper Red Rock Lake, Montana, home to one of the last remaining lacustrine populations of endemic Montana Arctic Grayling (Thymallus arcticus). Unlike most previous investigations of winterkill-prone lakes, we observed considerable horizontal spatial variability in DO, a non-linear winter oxygen depletion rate, and lake-wide re-oxygenation 2–4 weeks prior to spring ice loss. Parts of the upper 1 m of the lake and near stream mouths remained well-oxygenated even during late winter. DO levels were strongly associated with maximum daily air temperature. Our analysis of a 28-year weather record revealed large interannual variability in risk of winter hypoxia, with a slight declining trend in winter severity (number of days with maximum air temperatures ≤ 0°C) in Upper Red Rock Lake. The approach we used in our study provides a useful framework for quantifying and mapping the seasonal dynamics of the extent and severity of winter hypoxia, and for identifying critical winter habitats.

No abstract