The spatial and temporal evolution of contributing areas

Abstract

Predicting runoff source areas and how they change through time is a challenge in hydrology.Topographically induced lateral water redistribution and water removal through evapotranspiration lead to spatially and temporally variable patterns of watershed water storage. These dynamic storage patterns combined with threshold mediation of saturated subsurface throughflow lead to runoff source areas that are dynamic through time. To investigate these processes and their manifestation in watershed runoff, we developed and applied a parsimonious but spatially distributed model (WECOH-Watershed ECOHydrology). Evapotranspiration was measured via an eddy-covariance tower located within the catchment and disaggregated as a function of vegetation structure. This modeling approach reproduced the stream hydrograph well and was internally consistent with observed watershed runoff patterns and behavior. We further examined the spatial patterns of water storage and their evolution through time by building on past research focused on landscape hydrologic connectivity. The percentage of landscape area connected to the stream network ranged from less than 1% during the fall and winter base flow period to 71% during snowmelt. Over the course of the 2 year study period, 90% of the watershed areas were connected to the stream network for at least 1 day, leaving 10% of area that never became connected. Runoff source areas during the event shifted from riparian dominated runoff to areas at greater distances from the stream network when hillslopes became connected. Our modeling approach elucidates and enables quantification and prediction of watershed active areas and those active areas connected to the stream network through time.

Characterizing and reducing equifinality by constraining a distributed catchment model with regional signatures, local observations, and process understanding

“…The three models to which we compare our results demonstrate a range of model frameworks that can be used to evaluate model behavior: conceptual , lumped (Ahl et al, 2008), and distributed without physically based parameters (Nippgen et al, 2015). As is shown in this study, all of these models are able to accurately simulate the hydrograph for this catchment.…”

“…Nippgen et al (2015) Simulation of Stringer Creek with CCM achieved similar levels of fit to our study (NSE of 0.81 for Box-Cox transformed streamflow; Smith et al, 2013), as did model simulations of LTC (NSE of 0.903) and MSC (NSE of 0.856; Smith et al, 2016). Simulations of streamflow by Nippgen et al (2015) and Smith et al (2016) also underestimated peak streamflow for the 2008 water year, suggesting that uncertainty in the meteorological forcing data may be responsible for poor performance during the calibration period for snowmelt. Ahl et al (2008) modeled Tenderfoot Creek streamflow using the soil and water assessment tool (SWAT) for the period 1995-2002 and achieved an average NSE of 0.86 during calibration and 0.76 during validation.…”

“…While any of these approaches may be used to simulate streamflow, each will enable researchers to answer different questions related to hypotheses about catchment functioning, the use of field information to inform model parameter constraints, and predictions of spatiotemporal hydrologic processes. Finally, these contrasting models also illustrate the differences between a model like DHSVM that may be applied to many different catchments versus the models introduced by and Nippgen et al (2015), in which the model framework and structural equations were developed only for this catchment. In this study, we specifically evaluate the application of physically based, distributed models to simulate experimental catchments, though we encourage researchers to select the right tool, and therefore the appropriate model, for a given study objective.…”

“…Both field investigations and modeling at Tenderfoot Creek have focused on observation and prediction of water table response with the goal of improving our interpretation of streamflow generation and the connection between rainfall and runoff (Jencso et al, 2009;Jensco and McGlynn, 2011;Nippgen et al, 2015). Thus, we chose to evaluate simulations of water table depth for our case study.…”

Modelling storage‐driven connectivity between landscapes and riverscapes: towards a simple framework for long‐term ecohydrological assessment

“…This, in turn provides a robust basis for projecting the possible impacts of climate or land use change (Capell et al, 2013). Crucially, as well as being useful tools for predicting and understanding catchment function, such modelling approaches also have largely unrealized potential in providing a simple landscape hydrology context, based on storage-driven connectivity dynamics, for in-stream ecohydrological studies (Nippgen et al, 2015). The advantage of this is that it allows us to track the state of catchment storage and contextualize high flows and wet periods in terms of the longevity of positive storages.…”