An evaluation of catchment transit time model parameters: A comparative study between two stable isotopes of water

Abstract

Using δ18 O and δ2 H in mean transit time (MTT) modeling can ensure the verifiability of results across catchments. The main objectives of this study were to (i) evaluate the δ18 O-and δ2 H-based behavioral transit time distributions and (ii) assess if δ18 O and δ2 H-based MTTs can lead to similar conclusions about catchment hydrologic functioning. A volume weighted δ18 O (or δ2 H) time series of sampled precipitation was used as an input variable in a 50,000 Monte Carlo (MC) time-based convolution modeling process. An observed streamflow δ18 O (or δ2 H) time series was used to calibrate the model to obtain the simulated time series of δ18 O (or δ2 H) of the streamflow within a nested system of eight Prairie catchments in Canada. The model efficiency was assessed via a generalized likelihood uncertainty estimation by setting a minimum Nash–Sutcliffe Efficiency threshold of 0.3 for behavioral parameter sets. Results show that the percentage of behavioral parameter sets across both tracers were lower than 50 at the majority of the studied outlets; a phenomenon hypothesized to have resulted from the number of MC runs. Tracer-based verifiability of results could be achieved within five of the eight studied outlets during the model process. The flow process in those five outlets were mainly of a shallow subsurface flow as opposed to the other three outlets, which experienced other additional flow dynamics. The potential impacts of this study on the integrated use of δ18 O and δ2 H in catchment water storage and release dynamics must be further investigated in multiple catchments within various hydro-physiographic settings across the world.