Predicting soil moisture condition in arbitrary agricultural lands using the digital soil map and soil physical properties database

To predict soil moisture condition in arbitral agricultural lands by taking spatial distribution of soil type in a watershed into account, we investigated the possible proper methods of determining the soil hydrological parameters using available soil physical properties database of Japanese soils, “SolphyJ”, and the digital soil map. In this study, simulation of soil moisture was conducted at three neighboring locations, including monitoring sites each in Toyama and Fukui cities. The simulated results of soil moisture appeared to be improved when hydraulic parameters were obtained by fitting water retention data of SolphyJ compared to the parameters estimated by neural network with soil texture and dry bulk density data. It is probably because peculiar soil structure in each field could affect the hydraulic properties more than the soil texture. Simulation results also indicated that soil moistures are much different even if they are located in neighbors (<5 km). We concluded that combination of these two databases is useful for estimating soil hydraulic parameters and to predict soil moisture condition in arbitrary agricultural lands.