H131-08
Constraining water age estimations for runoff and evapotranspiration fluxes via multi-objective parameterization of StorAge Selection functions based on stable isotopes sampled at the catchment outlet and in trees

Monday, 14 December 2020: 04:28
Virtual
Matthias Sprenger1, Pilar Llorens2, Francesc Gallart2, Scott T Allen3, Paolo Benettin4 and Jérôme Latron2, (1)North Carolina State University Raleigh, Raleigh, NC, United States, (2)Institute of Environmental Assessment and Water Research (IDAEA-CSIC), Barcelona, Spain, (3)University of Nevada Reno, Reno, NV, United States, (4)École Polytechinque Fédérale de Lausanne (EPFL), Institute of Environmental Engineering, Laboratory of Ecohydrology, Lausanne, Switzerland
Recorded Presentation
Abstract:
Understanding how rainfall is stored in a catchment and released back to the atmosphere via evapotranspiration (ET) or flows to the stream is essential to assess the impact of climate and land use changes. Here we investigate the dynamics of rainfall partitioning based on stable isotope data (2H & 18O) at the Vallcebre catchment in the Spanish Pyrenees.

We use an extensive isotope data set covering 4 years with >550 rainfall and >980 stream samples. Sampling was conducted at high temporal resolution, with highest frequencies (30 min.) during rainfall-runoff events. We calibrated time variant StorAge Selection (SAS) functions on hourly time steps optimizing the simulation of the stream isotope data. In addition to this conventional approach, we also included tree xylem isotope data, sampled fortnightly over 8 months to better constrain the representation of the selection function for the ET flux.

Our results showed that the conventional approach, limiting the model calibration to stream isotopes, lead to problems of identifiability and the SAS parameter for the ET component did not lead to an acceptable simulation of the observed xylem isotope ratios. When adding the isotope information about the ET flux to the calibration via a multi-objective approach, both the simulation of the stream and the xylem water isotopes were well reflected, while the parameter space was greatly constrained, which reduced equifinality. Successfully transferring the calibrated SAS parameters to a validation period underlined that the calibration was robust.

Water age dynamics resulting from the conventional and the multi-objective calibration approaches differed considerably, because the latter resulted in older water being used by trees as compared when the SAS functions were calibrated solely to the stream isotopes. Thus, our modeling results support recent findings in ecohydrological field studies that highlighted both subsurface heterogeneity of water storage and fluxes and the use of relatively old water by trees. Comparisons of the water ages derived from SAS simulations with other approaches such as young water fractions or endmember-splitting show a consistent pattern of high variations of young water stream contributions as a function of runoff and ET being partly sourced from winter precipitation stored in the subsurface.

<< Previous Abstract | Next Abstract