Influence of Data Availability and Pedotransfer Functions on the Modeling of Soil Moisture Dynamics.

Hydrological models based on physical properties of soils require reliable information about soil hydraulic properties. Therefore, a reliable set of parameters for describing the soil water retention curve (SWRC) is required to get an estimate of the soil moisture dynamic (SMD) and water balance (WB). Direct measurement of soil properties provides the best estimate of the actual parameters needed for modeling, but this approach is typically time and cost demanding. Thus, the pedotransfer functions (PTF) are used to predict these parameters when predicting SMD and WB without labor-intensive fieldwork. HYDRUS-1D is one of the most known software used to predict these processes and it is also coupled with Rosetta, a model that estimates the parameters of the van Genuchten equation through PTF. These PTFs, that are included in Hydrus-1D, and require an increasing accumulative amount of known soil data as follows: particle size distribution (PSD), bulk density (BD), field capacity (FC) and wilting point (WP). Therefore, this work evaluates how the HYDRUS-1D model coupled with Rosetta performs when available data is limited and when the missing data is computed with PTFs (FC and WP). Annual simulations of the first two soil horizons collected from three non-fine textured soils from Central Chile were conducted using a HYDRUS 1D + Rosetta using available models with actual data and different PTFs. SMD and WB estimates based on actual SWRC were stated as control. Additionally, the saturated hydraulic conductivity (Ksat) was evaluated according to the Rosetta model and experimental runs. This approach allows to quantify the differences in the results from Hydrus-1D when data is predicted in comparison with the real SWRC. The results are discussed in detail in this paper.