Communicating uncertain futures for management: A case study with eastern brook trout

Background/Question/Methods

Changes in temperature and precipitation are expected to have strong influences on fish and wildlife populations, especially on ectothermic organisms. Brook trout, Salvelinus fontinalis, are a species of ecological and cultural importance that are expected to decline as temperatures increase in the northeastern United States because of their thermal sensitivity. Brook trout populations have already been locally extirpated in much of the southern portion of their native range. Despite knowledge relating climate and trout populations, there is uncertainty in future persistence. This uncertainty stems from numerous sources including different GCMs, downscaling global climate predictions, carbon emission scenarios, and imperfect estimates of relationships between trout and climate. It can be a challenge for managers to make decisions in the face of these compounded sources of uncertainty.  We used presence-absence data for stream reaches from across the northeastern US to model weather, landscape, and land-use effects on brook trout occupancy probability. We calculated the tolerance of brook trout to changes in climate and land-use conditions (i.e. amount of environmental change while maintaining threshold occupancy probability). We then compared this approach to the more typical forecasting of occupancy probability given a climate-emissions ensemble.

Results/Conclusions

The best model of brook trout occupancy contained flow, stream temperature, forest cover, wetland area, and number of dams. Flow, temperature, and wetlands had the largest effects on occupancy. There was also a moderate amount of variance among watersheds (10-digit HUC) that was not explained by the other model parameters. Additionally, the effects of forests and wetlands varied slightly depending on which watershed the survey stream was in.

In all GCM-Emission combinations brook trout are expected to occur in fewer streams. Probabilities of declines can be extracted for any future year. This is helpful for some managers, but others find the uncertainty to be daunting and a hindrance to action. We also applied an alternative approach by examining the tolerance of populations to environmental change. For all stream reaches, we calculated the maximum level of warming that could occur and still maintain a set threshold of occupancy (e.g. 30%, 50%, 70%). This approach was useful for managers whose objectives and decisions were based on maximizing stream kilometers with a goal probability of occupancy. We also evaluated the amount of forest that could be removed or restored to achieve set occupancy thresholds and apply these in combination with temperature tolerances.