Analysis of Soil Hydraulic Parameters effects on soil water modelling based on Danish soil water monitoring systems. 

Abstract

Accurate soil hydraulic parameters are crucial in effectively modeling agricultural field processes where soil water plays a crucial role (e.g. nitrate leaching processes, crop growth, etc.). Three horizons soil maps, called AC maps, have been developed for Denmark including related standard soil profile descriptions including three soil horizons (A, B and C) with data on soil texture, organic matter and soil hydraulic parameters (SHPs) which have been determined using a Danish developed pedotransfer function (PTF) predicting the vanGenuchten Mualem SHP called P10 . However, the standard soils have not been validated on simulating temporal change in soil water content (SWC). This study aims to evaluate the accuracy of the SHP set up obtained from the P10 and HYPRES (European developed PTF model predicting the predicting vanGenuchten Mualem SHP) as inputs for the Daisy model simulations (Danish soil-water-plant-atmosphere system model based on solving Richards equation in 1-D), by comparing simulations with measured SWC. Soil water content was measured at four different soil depth (25, 60, 90, and 110 cm, respectively) using TDR equipment for three different soil texture classes at three different experimental fields in Denmark for a period of 20 years. Statistical parameters including root mean square error (RMSE) and normalized root mean square error and Nash-Sutcliffe efficiency coefficient (NSE) were used to analysis the precision of the soil water content simulations. The results of the study indicate that the differences between simulated soil water content using the P10 PTF and measured values were not statistically significant. However, it was significant using HYPRES PTFs in Jyndevad (sandy soil). The NRMSE, RMSE (%), and NSE values varied between 0.14-0.38, 4.56-10.8 (%), and 0.78-0.98, respectively. In comparison, simulations using the HYPRES model had NRMSE, RMSE, and NSE values of 0.16-0.69, 4.16-7.39 (%), and 0.09-0.98, respectively. Our results suggest that P10 PTFs provided more accurate simulations of soil water content in Denmark compared to HYPRES. High values of NRMSE are related to the simulations in 60 cm depth for the site where the soil had the highest percentage of clay (ranging from 30-43%). However, high values of NSE indicate that the model successfully simulates the pattern of soil water variation during different days using AC map data. It is important to mention that soils with clay content exceeding 20% are rarely found in Denmark. In conclusion, the SHPs obtained from AC maps (P10 PTFs model) appear to be reliable and suitable for soil water simulations.  Key words: Soil maps, clay content, soil water content, Daisy model, Hydraulic parameters, Pedotransfer functions.