Statistical Modeling of Historical Daily Water Temperatures in the Lower Columbia River

Water temperature affects numerous aspects of aquatic life, and its stability is critical to cold water fish such as salmonids. With rising anthropogenic influence in natural environments, the future existence of these organisms is becoming less certain. In this study, I examined the evolution of historical water temperatures in the lower Columbia River by creating a statistical model to estimate daily historical water temperatures between the 1850s and 2010s. Daily air temperature and river flow measurements were used as inputs to the model, which estimated daily water temperatures at Bonneville Dam. The model used time lags between air temperature and water temperature to determine how the river\u27s memory, or capacitance, has changed over time. Since the 1850s, there has been an increase of 2.2 ± 0.2°C in mean annual water temperature in the lower Columbia River. The largest seasonal rate of increase of about 2.2 ± 0.2°C/century occurs in the fall and early winter while the smallest seasonal increase rate of about 0.04 ± 0.2°C/century occurs in the spring. Furthermore, I sought to identify and quantify the causes of these long-term changes by distinguishing what fraction of the water temperature changes were due to climate effects and how much of the changes were due to the reservoir system and irrigation withdrawals. I found the hydropower and reservoir system has contributed approximately twice as much as climate change to warming river temperatures. While these are simplified approximations of attribution, they have important implications for conservation management