资源环境科技发展态势分析平台

Catchments act as a sequence of hierarchical filters, damping input signals and generating less variable streamflow and stream solute concentration output signals. Spectral analysis across a wide array of solutes and catchments has revealed similar to 1/f scaling behavior (i.e., spectral power inversely proportional to frequency) spanning periods of hours to decades, attributed to hillslope filtering. We hypothesize that additional filtering by stream channel processes should occur. However, most catchments in which solute scaling behavior has been resolved have channel travel times too short for this to be evident from the relatively low sampling rates utilized. We use high-frequency (1-4 hr(-1)) sampling in larger catchments (up to 3 x 10(6) km(2)) to show that solute signals are indeed multifractal, steepening from similar to 1/f spectral at low frequencies to similar to 1/f(2) at higher frequencies. Across conservative, reactive, and gaseous solutes we demonstrate that departure from 1/f scaling occurs at frequencies that correspond with metrics of catchment size.

Plain Language Summary Concentrations of elements in stream water are less variable than concentrations in rainfall or other sources. This damping occurs due to mixing during transport through the subsurface and in the stream channel. Subsurface travel times are typically much longer than channel travel times. So while subsurface filtering has been well recognized, subsequent channel filtering only becomes apparent when frequent measurements are taken, or in large catchments. Here we demonstrate that this channel filtering also occurs and produces concentrations in large rivers, which are proportionally less variable than those in small streams over similar time scales. The results have important implications for how we measure and interpret water quality measurements.