We introduce EcH2O-iso, a new development of the physically based, fully distributed ecohydrological model EcH2O where the tracking of water isotopic tracers (2H and 18O) and age has been incorporated. EcH2O-iso is evaluated at a montane, low-energy experimental catchment in northern Scotland using 16 independent isotope time series from various landscape positions and compartments, encompassing soil water, groundwater, stream water, and plant xylem. The simulation results show consistent isotopic ranges and temporal variability (seasonal and higher frequency) across the soil profile at most sites (especially on hillslopes), broad model–data agreement in heather xylem, and consistent deuterium dynamics in stream water and in groundwater. Since EcH2O-iso was calibrated only using hydrometric and energy flux datasets, tracking water composition provides a truly independent validation of the physical basis of the model for successfully capturing catchment hydrological functioning, both in terms of the celerity in energy propagation shaping the hydrological response (e.g. runoff generation under prevailing hydraulic gradients) and flow velocities of water molecules (e.g. in consistent tracer concentrations at given locations and times). Additionally, we show that the spatially distributed formulation of EcH2O-iso has the potential to quantitatively link water stores and fluxes with spatiotemporal patterns of isotope ratios and water ages. However, our case study also highlights model–data discrepancies in some compartments, such as an over-dampened variability in groundwater and stream water lc-excess, and over-fractionated riparian topsoils. The adopted minimalistic framework, without site-specific parameterisation of isotopes and age tracking, allows us to learn from these mismatches in further model development and benchmarking needs, while taking into account the idiosyncracies of our study catchment. Notably, we suggest that more advanced conceptualisation of soil water mixing and of plant water use would be needed to reproduce some of the observed patterns. Balancing the need for basic hypothesis testing with that of improved simulations of catchment dynamics for a range of applications (e.g. plant water use under changing environmental conditions, water quality issues, and calibration-derived estimates of landscape characteristics), further work could also benefit from including isotope-based calibration.
Bibliographical noteThis work was funded by the European Research Council (project GA 335910 VeWa). Marco P. Maneta acknowledges support from the US National Science Foundation (project GSS 1461576) and US National Science Foundation EPSCoR cooperative agreement no. EPS-1101342. We thank the topical editor (Min-Hui Lo), Jr-Chuan Huang, two anonymous reviewers for their suggestions which significantly improved the paper, and Aaron A. Smith for fruitful discussions regarding the model development. We also acknowledge the support of the Maxwell computer cluster funded by the University of Aberdeen. Finally, we are grateful to the many people who have been involved in establishing and continuing data collection at the Bruntland Burn, particularly Christian Birkel, Maria Blumstock, Jon Dick, Josie Geris, Konrad Piegat, Bernhard Scheliga,
Matthias Sprenger, Claire Tunaley, and Hailong Wang.