What Can Catchment Transit Time Distributions Tell Us About Runoff Mechanisms? Exploring "Age Equifinality" with an Integrated Surface-Groundwater Model.

The backward transit-time distribution (bTTD) is the time-varying, probabilistic distribution of water travel times or, equivalently, water ages in catchment outflow. The bTTD is increasingly seen as a master variable of catchment hydrology that links flow and transport processes, in part because it is believed to embed information about runoff generation mechanisms (RGMs) that are difficult to directly observe. The ability to use water age to make inferences about RGMs depends on the degree of "age equifinality" in a watershed, defined here as the phenomenon where significant volumes of similarly-aged water are delivered to the outlet by different RGMs at the same time. When age equifinality is low (e.g., all discharge is old groundwater), the mapping of water age to the RGM may be simple; when age equifinality is high (e.g., discharge is a mix of old groundwater and old interflow), this mapping may be impossible. In this study we conduct experiments in a virtual watershed to (1) understand the hydrologic conditions that lead to age equifinality, (2) identify relationships between water age and RGMs that are particularly obscured/unobscured by age equifinality, and (3) test the generalizability of these relationships in other watersheds. Our experiments used the fully-distributed surface-groundwater model ParFlow, which simulates a suite of RGMs, plus SLIM-FAST particle tracking. To improve realism, the watershed model was parameterized and forced using extensive field data from the USDA's Mahantango Creek experimental catchment in PA, USA. The model output is being interrogated to understand the time-varying relationships between the composition of RGMs and the bTTD at the outlet. We are also testing the robustness of these relationships by re-running our model with controlled differences in climate, topography, and scale. Initial results suggest high age equifinality at peak flows due to overlapping young water contributions from infiltration- and saturation-excess overland flow, as well as high age equifinality at moderate flows due to overlapping middle-aged water contributions from perched saturation flow and local groundwater. The final results are intended to provide guidance on how to account for age equifinality when using bTTDs to make inferences about dominant runoff generation mechanisms.