Summarizing and comparing annual stream temperature signals from mountain streams across the U.S.

Heat has long been used as a tracer of various watershed processes, but only a few studies have explored the patterns of long-period temperature signals (e.g. annual) and their relation to these processes. In previous research, metrics describing annual air and water temperature patterns combined with a deterministic model were used to describe groundwater influence on stream temperatures for streams in Shenandoah National Park, Virginia. Here, we examine the transferability of these results and patterns to other mountain systems and climates (e.g., the Pacific Northwest), and whether these signals are stable over time. Additionally, we evaluate the effects of "missing data", a common issue facing collectors of multi-year stream temperature data, on thermal process inferences. Our results indicate that, in general, groundwater influence on stream thermal regimes was slightly lower in the Shenandoah National Park sites but that groundwater was sourced from similar depths for the study areas despite regional differences in temperature and precipitation patterns. In terms of data demands, we found that differences in the minimum required number of data points within a year were dependent on the arrangement of the missing data (e.g., random versus consecutive). This research expands our understanding of how simple techniques using multi-year temperature data can be used to explore the spatial and temporal variations in watershed physical processes, including groundwater discharge, across regions.