Solute transport in multiple-reach experiments: Evaluation of parameters and reliability of prediction

We used two stream tracer studies to address issues related to using the transient storage model (TSM) in simulating solute transport in streams containing riffle-pools and step-pools. The injection durations were short such that one may consider the storage zones to consist mainly of surface water storage zones. We found that the TSM was not able to reproduce the high portion (near the peak) of the breakthrough curve in the first subreach (85 m) regardless of the functional form of the objective function. We attributed this to the presence of a large pool 35 m downstream of the injection location that caused segregation of solute flow that was completely mixed across the stream upstream of the pool. The model was able to closely fit the breakthrough curves at x=185 and 323 m, most likely because the signature of the pool diminished downstream. The fitting method did not always ensure that the parameter values represented physical reality, and we believe that physical measurements should be made and incorporated into the parameter estimation procedure utilizing Bayesian estimation. Our findings indicate that the representative values of TSM parameters in a reach do not generally fall between the representative values of its subreaches. Hence, the expression 'reach-averaged' should be used cautiously. The results indicated that it is more reliable to extrapolate (going downstream) than to interpolate. We believe this is due to local physical controls that could not be accounted for with parameters estimated far downstream from the controls.