Applicability of soil pore size distribution derived from digital microscopy images to determination of water retention curve

Information on the water retention curve (WRC) is required for the numerical simulation of water flow; however, its acquisition is time- and cost-consuming. In this study, the applicability of two-dimensional (2D) image of soil taken by a digital microscope to WRC measurement was assessed. For this purpose, four undisturbed soil samples were collected at surface (10–15 cm) and undersurface (22.5–27.5 cm) depths to acquire 2D images and measure WRCs from the conventional method, pressure plate method. To derive the WRCs, soil pore-size distribution in soil images was manually extracted using the following three methods: In Methods 1 and 2, the fuzzy region and the dark region, respectively, in the image were assumed to be a soil pore, and in Method 3, the pore boundary, being assumed to be a bright part, connected the lines and divided the region into pores. Method 3 had much large-size pores (>75 µm in radius) at deeper depths than at surface depths. This trend was the same as the pressure plate method. The WRCs in three methods were well fitted to the lognormal model. In all methods, the values of parameter of θ_e were larger at surface layers than at deeper depths, of which similar result was also seen in the pressure plate method. We discussed applicability of the parameters. It is effective to determine θ_e from the pressure plate method and the other parameters (ψ_m and σ) from the image-based method. From those analyses, it could be said to prove to obtain WRC from an image-based method.