An empirical model was developed that describes the influence of lake productivity, climate, and morphometry on coldwater fish oxythermal habitat. An oxythermal habitat variable called temperature at 3 mgÁL -1 of dissolved oxygen (T DO3 ) was developed by interpolating the water temperature at a benchmark oxygen concentration (3 mgÁL -1 ) from a temperature-oxygen profile. Coldwater habitat was most available in the least productive lakes (total P < 25 mgÁL -1 ) with the greatest relative depths (geometry ratios < 2 m -0.5 ) and where mean July air temperatures were less than 17 8C. Species response curves were developed from values of TDO3 measured during the greatest period of oxythermal stress in late summer (maxTDO3). Lake trout was present in lakes with the lowest values of maxTDO3, while cisco was present in lakes with the highest and broadest range of maxTDO3. Projections for a scenario where climate warming (+4 8C in mean July air temperature) was accompanied by eutrophication (doubling of total P) indicated that coldwater fish oxythermal habitat would be devastated in a subset of lakes typical for Minnesota. Protecting deep, unproductive lakes from eutrophication will be a necessary management strategy to ensure that coldwater fish persist in at least some Minnesota lakes after climate warming.Résumé : Nous mettons au point un modèle empirique qui décrit l'effet de la productivité lacustre, du climat et de la morphométrie sur l'habitat oxythermique des poissons d'eau froide. En interpolant la température de l'eau à une concentration repère d'oxygène (3 mgÁL -1 ) dans un profil de température-oxygène, nous obtenons une variable de l'habitat oxythermique, soit la température à 3 mgÁL -1 d'oxygène dissous (TDO3). L'habitat d'eau froide se retrouve le plus couramment dans les lacs les moins productifs (P total < 25 mgÁL -1 ) avec les profondeurs relatives les plus grandes (rapports géométri-ques < 2 m -0,5 ) et une température moyenne de l'air en juillet de moins de 17 8C. Nous avons mis au point des courbes de réaction des espèces à partir des valeurs de TDO3 mesurées durant la période de stress thermique maximal en fin d'été (TDO3 maximum). Les touladis se retrouvent dans les lacs qui ont les valeurs les plus basses de TDO3 maximum, alors que les ciscos vivent dans les lacs avec les étendues des valeurs de TDO3 maximum les plus élevées et les plus larges. Des projections selon un scénario dans lequel le réchauffement du climat (+4 8C de température moyenne de l'air en juillet) est accompagné d'une eutrophisation (dédoublement du P total) indiquent que l'habitat oxythermique des poissons d'eau froide serait dévasté dans un sous-ensemble de lacs typiques du Minnesota. Il sera nécessaire d'adopter une stratégie de gestion pour protéger les lacs profonds et peu productifs de l'eutrophisation afin d'assurer que les poissons d'eau froide survivent au moins dans certains lacs du Minnesota après le réchauffement climatique.[Traduit par la Rédaction]

Ciscoes Coregonus artedi are coldwater stenotherms that are sensitive indicators of ecological stressors, such as eutrophication and climate warming, that reduce coldwater habitat. Temperature and oxygen profile data were collected during cisco mortality events at 17 lakes during an unusually warm summer in Minnesota in 2006. Combinations of temperature and oxygen from the profiles were mapped onto a twodimensional niche space to directly quantify an oxythermal lethal niche boundary for ciscoes. Quantile regression was used to estimate the niche boundary that described lethal combinations of oxygen and temperature for cisco. The fitted lethal oxythermal niche boundary exhibited a curvilinear interaction between lethal temperatures and lethal oxygen concentrations. The lethal temperature under normoxia (about 8 mg/L) was estimated to be 248C. Lethal temperatures were progressively less at lower lethal oxygen concentrations (e.g., the lethal temperature was 23.08C at 5.0 mg/L, 22.08C at 3.0 mg/L, and 19.58C at 1.0 mg/L). The lethal oxythermal niche boundary provides a basis for quantifying coldwater cisco habitat under current and future climate warming and water quality conditions.

We developed a fish‐based index of biotic integrity (IBI) for a set of Minnesota lakes having similar geophysical and chemical features. Fish data were collected by means of trap nets, gill nets, shoreline seines, and backpack electrofishing. Of 30 evaluated metrics, we identified 16 metrics of three types: species richness, community assemblage, and trophic composition. In contrast to lotic IBIs, where a single sampling gear is usually used, data from all four sampling gears were necessary for IBI development. We selected metrics based on responses to measures of human‐induced stress based on watershed land use patterns and human population density. Species richness and community composition metrics describing intolerant or habitat specialist species were most sensitive to differences in human‐induced stress. Because these species were found in the nearshore zone of lakes, effective sampling of the nearshore fish community was essential to the development and performance of the IBI. Trap‐net‐ and gill‐net‐based metrics, however, were essential to the development of trophic composition metrics, the trap nets providing the best insectivore and omnivore metric and the gill nets providing the best top carnivore metrics. Lake IBI scores reflected differences in land use patterns, trophic state, and aquatic vegetation. Sensitivity analysis indicated that selected metrics contributed to IBI performance, but the relative contribution of individual metrics varied among human‐induced stress categories. Additional work is needed to test the performance of this IBI on an independent group of lakes, to set the range of applicability, and to determine the effectiveness of the IBI as an assessment tool for lakes.