Oxythermal habitat parameters and identifying cisco refuge lakes in Minnesota under future climate scenarios using variable benchmark periods

Fish habitat in lakes is strongly constrained by water temperature (T) and available dissolved oxygen (DO). With climate warming both of these water quality parameters will change. A fish habitat simulation model was developed to calculate an oxythermal habitat variable, TDO3, i.e. water temperature at 3mg/L of DO, from simulated daily T and DO profiles and used to determine potential refuge lakes for lake herring or cisco (Coregonus artedi) in Minnesota. Daily T and DO profiles were simulated using a deterministic, unsteady, one-dimensional (vertical) year-round lake water quality model (MINLAKE2010) that was developed specifically for transparent deep lakes. Profiles for thirty Minnesota cisco lakes were simulated with the MINLAKE2010 model using recorded (1961–2008) weather data or projected future climate scenarios (CGCM 3.1 and MIROC 3.2) as model input. The average of daily TDO3 values (called ATDO3VB) over a 31-day sliding or variable benchmark period (VB), was first calculated from daily TDO3 values. The annual maximum of ATDO3VB was retained for each simulated lake and year, and the average of 47 ATDO3VB values (called AvgATD3VB) obtained for the 1962–2008 simulation period was compared to oxythermal stress criteria for cisco determined from the analysis of field data. Isopleths of AvgATD3VB were interpolated for the 30 simulated lakes on a plot of Secchi depth vs. lake geometry ratio used as indicators of trophic state and summer mixing conditions, respectively. Marking the 620 Minnesota lakes with identified cisco populations on the plot of AvgATD3VB allowed to partition the 620 lakes into the three tiers depending on where they fell between the isopleths: lakes with AvgATD3VB≤11°C (Tier 1 lakes) were selected to be most suitable for cisco; lakes with 11°C17°C (Tier 3 lakes) would support cisco only at a reduced probability of occurrence or not at all. About 160 existing cisco lakes were thus projected to be suitable refuge lakes (Tier 1 plus Tier 2) under the two future climate scenarios. The remaining lakes were projected to loose their cisco supporting ability. The methodology to identify the refuge lakes was applied to a cisco kill that occurred in 18 cisco lakes in Minnesota in the hot summer of 2006, and was able to identify the majority of these lakes as non-refuge lakes.