Identifying Potential Cisco Refuge Lakes in Minnesota under Future Climate Warming Scenarios

Abstract

Fish habitat models were developed and applied to investigate the impacts of future climate change on cisco oxythermal habitat in Minnesota lakes. Cisco Coregonus artedi is the most common cold-water stenothermal fish species in lakes over the several northern states in USA. Daily water temperature (T) and dissolved oxygen (DO) profiles were simulated for different types of virtual lakes (surface area from 0.05 to 50 km2) representing cisco lakes in Minnesota and under the past climate conditions (1961–2008) and projected future climate scenarios. A process-oriented, dynamic and one-dimensional lake water quality model (MINLAKE2010/2012) was applied for the temperature and DO simulations using daily time steps. The parameters used to describe a lake in the model are surface area (As), maximum depth (Hmax), and Secchi depth (as a measure of radiation attenuation and trophic state). Two projected future climate scenarios, predicted by the Canadian Centre of Climate Modeling and Analysis (CCCma) coupled general circulation model (CCCma CGCM 3.0) and the Model for Interdisciplinary Research on Climate (MIROC 3.2) were used. Both climate scenarios lead to a longer period of hypoxic hypolimnetic conditions in stratified lakes. The study was able to identify potential refuge lakes important for sustaining cisco habitat under climate warming scenarios. To project cisco’s chances of survival under future warmer climate conditions, using simulated daily T and DO profiles in 74 virtual lakes, FishHabitat2013 model was developed with three oxythermal habitat modeling options: (1) constant lethal T and DO limits, (2) lethal-niche-boundary curve (variable lethal T and DO limits), and (3) an oxythermal habitat variable, TDO3, i.e., water temperature at DO = 3 mg/L. The fish habitat model implementing the first two options was validated against cisco mortality/habitable observations in 23 Minnesota cisco lakes of which 18 had cisco mortality while 5 had no cisco mortality in the unusually warm summer of 2006. Cisco lethal and habitable conditions in the 23 lakes simulated by the model had overall good agreement with observations in 2006. After model validation, T and DO in the 44 virtual lakes were first simulated under past and future climate scenarios to determine cisco lethal days. Polymictic shallow lakes with lake geometry ratio As0.25/Hmax > 5.2 m-0.5 (As in m2 and Hmax in m) were simulated to typically not support cisco habitat under past and future climate scenarios. Medium-depth lakes are projected to be most vulnerable to climate warming with most increase in the number of years with cisco kill. The mean daily TDO3 values over a 31-day fixed and variable benchmark periods were calculated for each of simulated years and then averaged over the simulation period for each of 30 virtual deep lakes. Projected increases of the multi-year average TDO3 (called AvgATD3) under the two future climate scenarios and relative to the past 47-year simulation period (1962–2008) had averages from 2.6 to 3.4 oC. Isopleths of AvgATD3 were interpolated for the 30 simulated virtual lakes on a plot of Secchi depth vs. lake geometry ratio used as indicators of trophic state and summer mixing conditions, respectively. Using their lake geometry ratios and Secchi depths as coordinates, existing 620 Minnesota cisco lakes were marked on the plot of AvgATD3, and this allowed to partition these 620 cisco lakes into the three tiers depending on where they fell between the isopleths: lakes with AvgATD3 ≤ 11oC (Tier 1 lakes) were selected to be most suitable for cisco; lakes with 11 oC < AvgATD3 ≤ 17 oC (Tier 2 lakes) had suitable habitat for cisco; and non-refuge lakes with AvgATD3 > 17 oC (Tier 3 lakes) would support cisco only at a reduced probability of occurrence or not at all. About 208 (one third) and 160 (one fourth) of the 620 Minnesota cisco lakes that were known to have cisco populations are projected to maintain viable cisco habitat under the two projected future climate scenarios using the fixed and variable benchmark periods, respectively. These selective lakes have a Secchi depth greater than 2.3 m (mesotrophic and oligotrophic lakes) and are seasonally stratified (geometry ratio less than 2.7 m-0.5).