What Controls the Spatial Variation in Thermal Sensitivity in a Cascades River Basin During an Extreme Heat Year?

Increasing stream temperatures resulting from climate change and consequent detrimental impacts on temperature-dependent aquatic species have signalled the need to better understand how different streams respond to changes in air temperature. This research generates a new assessment of thermal sensitivity in the Clackamas River Basin, Oregon, USA by analysing how underlying geology, climate, hydrology and land cover explain spatial variation in stream thermal sensitivity. Multiple linear regression analysis using spatially explicit landscape predictors across 72 sites identified upstream High Cascades geology, elevation, baseflow index, total upstream area and upstream shrubland area as the most significant and stable predictors determining thermal sensitivity. Spatial stream network and random forest models further validated and clarified spatial influences on stream temperature and each improved performance relative to multiple linear regression. Predictors derived at the total upstream area scale were most consistently associated with thermal sensitivity when compared to local reach contributing area or 100-m wide buffer area scales. Streams underlain by High Cascades geology, with higher baseflow and greater upstream area were substantially less sensitive to air-temperature changes (β = −0.27 to −0.61). Meanwhile, higher elevation (β = −0.51 to −0.59) and lower shrubland cover (β = 0.19 to 0.30) were associated with lower thermal sensitivity. Variation in predictors retained among models appears to reflect different indicators of shared underlying processes. Furthermore, this study confirms that these predictors are statistically robust drivers of stream thermal sensitivity via regularized regression techniques. Analysing why streams are more or less climate-resilient can inform management and restoration efforts and protect cold-water habitats.