Simultaneous estimation of soil hydraulic properties and surface evaporation using inverse modeling for a large field-lysimeter

A sound prediction of water and energy fluxes at the soil-atmosphere interface is important for many practical questions regarding e.g. irrigation and salinity management. Precise knowledge of soil hydraulic properties (SHP) is mandatory for such predictions. The SHP can be measured either in the laboratory within a wide moisture range or at the field scale, e.g. by inverse simulation techniques based on in situ matric potential and water content measurements. Depending on the installation depth of the sensors, soil texture, and boundary conditions, field-determined SHP are often limited to a quite narrow range of moisture conditions. Prediction of actual surface fluxes on basis of this limited information is highly uncertain. With well-instrumented large weighable lysimeters, systems are now available that allow to measure very precisely surface (and bottom) water fluxes under natural atmospheric conditions. In particular, they can be used to quantify the difference between potential evaporation, Ep, and observed actual evaporation, Ea. The difference (Ep-Ea) increases during the drying process when the soil hydraulic conductivity becomes limiting for the evaporation process. Thus, our hypothesis was that this information can be used to improve the identification of SHP of soils.

Accordingly, the aim of this study was to see whether the information on (Ep-Ea), measured during a calibration period and supplemented by water content and matric potential data measured inside of a lysimeter, is sufficient to inversely estimate the SHP. Furthermore, we were interested to see if the prediction of Ea was possible and reliable also for time periods beyond the calibration period.  For a proof-of-concept study, we conducted forward simulations with Hydrus-1D where we generated synthetic data of actual surface fluxes and soil hydraulic internal state variables. The atmospheric boundary was given by natural precipitation and potential evaporation rates in a semi-arid climate. The study showed that it was possible to identify SHP by inverse modeling, and prediction of the cumulative actual evaporation after the calibration period was successful. In a second step, the methodology was applied to data of a real large bare-soil field-lysimeter. Our simulation results showed also here a good match between observed and predicted cumulative evaporation.