Strategies to characterize water quality vary widely, but regulatory approaches mainly consider single-value thresholds for individual parameters (e.g. minimum dissolved oxygen concentration). Distributions of parameter values across multiple temporal and spatial scales, commonly referred to as regimes, add greater context and interpretability to point measurements. Although uncommon among monitoring programmes, continuous water quality data collected at high frequency (e.g. hourly) can characterize waterbody health more accurately than infrequent point measurements. We used multivariate analysis to describe water quality regimes based on hourly measurements of dissolved oxygen, pH, specific conductance and water temperature from three divergent stream types in Southeast Alaska national parks. We also assessed whether less frequent measurements drawn from the original hourly data set resulted in similar water quality regime descriptions. The monthly means and standard deviations of the four water quality parameters created ordinations with interpretable, stream-specific environmental gradients. Procrustean analysis revealed that ordination results were strikingly similar across all temporal monitoring frequencies. Univariate medians and distributions of weekly, twice monthly and monthly measurements were similar across all parameters, but hourly monitoring was necessary to accurately characterize extreme values. These analyses demonstrated the ability of commonly collected water quality parameters to define unique physical–chemical properties across regional stream types and present scientists with common analytical tools to determine appropriate monitoring scales for accurately characterizing water quality regimes.
