Combining Empirical Relationships with Data Based Mechanistic Modeling to Inform Solute Tracer Investigations across Stream Orders

Abstract:

Solute transport studies in streams and rivers often begin with the introduction of conservative and reactive tracers into the water column. Information on the transport of these substances is then captured within tracer breakthrough curves (BTCs) and used to estimate, for instance, travel times and dissolved nutrient and carbon dynamics. Traditionally, these investigations have been limited to systems with small discharges (< 200 L/s) and with small reach lengths (< 500 m), partly due to the need for a priori information of the reach’s hydraulic characteristics (e.g., channel geometry, resistance and dispersion coefficients) to predict arrival times, times to peak concentrations of the solute and mean travel times. Current techniques to acquire these channel characteristics through preliminary tracer injections become cost prohibitive at higher stream orders and the use of semi-continuous water quality sensors for collecting real-time information may be affected from erroneous readings that are masked by high turbidity (e.g., nitrate signals with SUNA instruments or fluorescence measures) and/or high total dissolved solids (e.g., making prohibitively expensive the use of salt tracers such as NaCl) in larger systems. Additionally, a successful time-of-travel study is valuable for only a single discharge and river stage. We have developed a method to predict tracer BTCs to inform sampling frequencies at small and large stream orders using empirical relationships developed from multiple tracer injections spanning several orders of magnitude in discharge and reach length. This method was successfully tested in 1^st to 8^th order systems along the Middle Rio Grande River Basin in New Mexico, USA.