Manifestation of spatial and temporal variability of soil hydraulic properties in the uncultivated Fluvisol and performance of hydrological model

The aim of this study was to document the spatial and temporal variability of soil hydraulic properties (SHP) present in an uncultivated soil formed by fluvial sediments (Fluvisol in our case). Specifically, a small scale experimental plot (10 x 10 m) was used during four field works evenly distributed in one vegetation season (8 months) in order to demonstrate the importance (representativeness) of sufficient sampling in time and space. During each field work, six soil water retention curves (SWRC) for two depths (25 and 60 cm; reflecting the position of installed tensiometers and the presence of two layers of the soil profile) and ten saturated hydraulic conductivities (K-sat) were determined. The measured values of SHP were used for the HYDRUS-1D model in order to observe the influence of the spatial and temporal variability on the modelled soil water regime. The vegetation season was for modelling purposes split into four two-month sub-periods (reflecting air temperature and precipitation conditions of the season), each represented by specific soil hydraulic properties. The results indicated that the spatial variability of SWRC completely covered for any possible temporal variability as the coefficients of variation (C-v) of the SWRC parameters were up to 127% (in the case of residual soil water content -theta(r)) and from 62% to 88% concerning the alpha shape parameter (alpha). The remaining two SWRC parameters (theta(s) and n) were less variable. The spatial variability of K-sat expressed by C-v equalled 128%. The modelling results showed a significant improvement of the soil water regime prediction when the proper SWRC was utilized (compared to its areal average). The improvement was reflected by the overall increase of the Nash-Sutcliffe efficiency (from 0.20 to 0.68) and a decrease in root mean square error (RMSE) (by 38%) for the measurement at 25 cm. The improvement of the model prediction was less significant at the depth of 60 cm. In summary, the high spatial variability of soil hydraulic properties in the uncultivated Fluvisol (representing permeable sandy fluvial sediments) proved to be superior to their temporal changes and it is therefore worth quantifying its extent for modelling purposes. The described variability of SHP will serve as a benchmark for the further investigation of the organic matter influence on the soil water regime in field conditions.