Tracer-aided modelling with in-situ isotope data to advance understanding of ecohydrological partitioning in urban areas

Quantifying ecohydrological partitioning is important to understand water balance changes in the context of land cover change, and requires better integration of novel field data and improved models. Stable water isotopes are invaluable in understanding how green water fluxes are partitioned in evapotranspiration (ET) as they allow to constrain estimates of the depths of root water uptake that sustain interception I, transpiration T and soil evaporation E. Recent developments allowing for in-situ monitoring now provide prolonged periods of near-continuous isotope time series for multiple landscape compartments of the soil-plant-atmosphere continuum. Such high-resolution isotope data provide an invaluable resource for improving isotope-aided ecohydrological models by allowing to inform model structure, enhance process understanding and constrain flux estimates.

Here, we use concurrent in-situ isotope time series of entire growing periods of soil water, xylem water and atmospheric water vapour in an intensively monitored urban green space in Berlin, Germany. These data were integrated into tracer-aided ecohydrological models with, e.g., isotopes in xylem and atmospheric moisture as simulation targets. Both variables were found to be informative, although xylem isotopes were less stable with ambiguities in terms of the influence of internal ecophysiological processes or methodological problems. Atmospheric vapor sampled at 1, 5 and 10m heights was logistically much simpler and captured well the isotopic signals of I, E and T from the xylem, as well as more regional influences.

We could resolve ET fluxes revealing seasonal changes in dominant sources of root water uptake, as well as time-variant changes in the relative important of E, I and T to ET losses. Inter-species differences between willow (Salix) and maple (Acer) trees were also captured. The study demonstrated the complementarity of different isotope approaches and highlighted the under-utilised potential of atmospheric water vapour in ecohydrological models. We also demonstrated the importance of using non-isotope ecohydrological data (sap-flow and dendrometers) in conjunction as calibration constraints.