Stable water isotope simulation by current land-surface schemes: results of iPILPS Phase 1

Phase 1 of isotopes in the Project for Intercomparison of Land-surface Parameterization Schemes (iPILPS) compares the simulation of two stable water isotopologues (¹H₂¹⁸O and ¹H²H¹⁶O) at the land-atmosphere interface. The simulations are offline, with forcing from an isotopically enabled regional model for three locations selected to offer contrasting climates and ecotypes: an evergreen tropical forest, a sclerophyll eucalypt forest and a mixed deciduous wood. Here, we report on the experimental framework, the quality control undertaken on the simulation results and the method of intercomparisons employed. The small number of available isotopically enabled land-surface schemes (ILSSs) limits the drawing of strong conclusions, but, despite this, there is shown to be benefit in undertaking this type of isotopic intercomparison. Although validation of isotopic simulations at the land surface must await more and much more complete, observational campaigns, we find that the empirically based Craig-Gordon parameterization (of isotopic fractionation during evaporation) gives adequately realistic isotopic simulations when incorporated in a wide range of land-surface codes. By introducing two new tools for understanding isotopic variability from the land surface, the isotope transfer function and the iPILPS plot, we show that different hydrological parameterizations cause very different isotopic responses. We show that ILSS-simulated isotopic equilibrium is independent of the total water and energy budget (with respect to both equilibration time and state), but interestingly the partitioning of available energy and water is a function of the models' complexity.