Cisco population characteristics in Wisconsin lakes in relation to lake- and landscape-level factors
Objective Declines in Cisco Coregonus artedi populations in some inland lakes have prompted assessments of Cisco occurrence and extirpation risk in relation to various stressors to identify refuge lakes and factors that promote Cisco persistence. However, most previous assessments have focused on presence–absence of Cisco rather than examining how population characteristics, such as relative abundance or growth, might change in relation to lake- and landscape-level environmental factors. Consequently, our specific objectives were to identify important environmental factors explaining variation in Cisco relative abundance and growth and to determine whether population metrics describing size and age distributions were related to relative abundance in Wisconsin inland lakes. Methods Cisco were collected from 48 inland Wisconsin lakes during 2011–2015 using vertical monofilament gill nets and population-specific relative abundance estimates (catch per unit effort [CPUE]) were quantified as the number of individuals per gill-net night. Sagittal otoliths were removed from a subsample of Cisco for age estimation and growth was indexed as mean total length (TL; mm) at age 2. Length and age data were used to develop a suite of metrics describing size and age distributions of each population. Random forest models were used to evaluate relationships between 10 biologically relevant predictor variables representing variation in physical, climatic, catchment, and limnological characteristics and Cisco CPUE and growth. Pearson correlations were used to determine whether population characteristics were related to CPUE. Result Cisco populations exhibited large variation in relative abundance, growth, and size and age distributions. Best-fit random forest models explained approximately 25% of the variation in Cisco CPUE and 46% of the variation in growth. Growing degree-days and variables associated with availability, quality, and quantity of suitable oxythermal conditions were identified as important predictors of both Cisco CPUE and growth; CPUE was also identified as an important predictor of growth. Mean TL and mean TL at age 2 were negatively related to Cisco CPUE, whereas mean age, number of age-classes present, and maximum observed age were positively related to CPUE. Conclusion Our results suggest that maintenance of suitable oxythermal habitat conditions may be critical to conserving abundant Cisco populations. Our assessment also provides insights on how Cisco populations may respond to environmental and anthropogenic stressors, which could aid ongoing and future conservation and management efforts in Wisconsin and elsewhere.