SPATIAL AND TEMPORAL VARIATION OF WATER TEMPERATURE REGIMES ON THE SNOQUALMIE RIVER NETWORK

Although mean temperatures change annually and are highly correlated with elevation, the entire thermal regime on the Snoqualmie River, Washington, USA does not simply shift with elevation or season. Particular facets of the thermal regime have unique spatial patterns on the river network and at particular times of the year. We used a spatially and temporally dense temperature dataset to generate 13 temperature metrics representing popular summary measures (e.g., minimum, mean, or maximum temperature) and wavelet variances over each of seven time windows. Spatial stream-network models which account for within-network dependence were fit using three commonly used predictors of riverine thermal regime (elevation, mean annual discharge, and percent commercial area) to each temperature metric in each time window. Predictors were strongly related (r(2) > 0.6) to common summaries of the thermal regime but were less effective at describing other facets of the thermal regime. Relationships shifted with season and across facets. Summer mean temperatures decreased strongly with increasing elevation but this relationship was weaker for winter mean temperatures and winter minimum temperatures; it was reversed for mean daily range and there was no relationship between elevation and wavelet variances. We provide examples of how enriched information about the spatial and temporal complexities of natural thermal regimes can improve management and monitoring of aquatic resources.