A novel approach for examining downstream thermal responses of streams to contemporary forestry

Temperature is a fundamental driver of aquatic environments. Changes in thermal regimes due to timber harvest may be detrimental for cold-water instream biota. Although it is understood that stream temperature may increase immediately below timber harvest operations, the understanding of how thermal responses propagate downstream is less clear. Here, we examine the effects of timber harvest on stream temperature pre- (2-3 years) and post-harvest (5 years) at 16 sites (average annual streamflow rates <0.283 m³ s^-1) located in the Coast Range, Oregon, USA. At each site, an array of temperature sensors were deployed on the extremes of three consecutive reaches: an upstream unharvested reference reach, a treatment reach, and a downstream unharvested reach. We used several metrics to describe and evaluate changes over time and space focusing on the responses of downstream reaches. Primarily, we evaluated the differences over time in daily maximum temperature between the two sensors located at the downstream unharvested reach. Furthermore, using a statistical ordination technique, we examined the spatial and temporal variability of an array of sensors for daily maximum temperature. Moreover, we assessed distributional shifts (statistical moments) of hourly temperature differences between the two sensors at the downstream unharvested reach over time. Lastly, we used a combination of statistical moments and the ordination technique to provide an index that describes the behavior of site-specific thermal disturbance over time. We found that stream reaches responded differently to upstream timber harvest operations between pre and post-harvest summer seasons. In addition, we showed distinct patterns of longitudinal variability of temperature across sites and summer seasons with increases, decreases or mixed responses including no change downstream. Overall, the net change of daily maximum temperature at the downstream reach revealed that the highest difference occurred during the first and second year post-harvest and, in some cases, a distinctive shift in stream warming and cooling occurred between the day and the night. Observed temperature patterns in downstream reaches were most similar to the pre-harvest conditions at the fifth year post-harvest. Collectively, we offer a novel approach for assessing stream temperature regime change using multiple metrics that can improve our understanding of thermal effects downstream of timber harvest, taking in consideration idiosyncratic responses across sites and time.