ИСТИНА

Different sphere packs were used by soil scientistsand hydrologists to model structure and assess hydraulic properties. It is now quite clear that spheres alone can not account for soil complexity. However, modified sequential methods can be extremely useful for sandy/ unstructured soil structure reconstructions and different models verification (pore-network, correlation functions, modified Kozeny-Carman, etc.). Here we consider a couple of possible practical implications where 3D soil structure or hydraulic properties can not be assessed via conventional approaches due to sample size or shape, or imaging resolution limitations. Mainly, two types of sequential models are considered: 1) process-based (slow, but somewhat more realistic), and 2) discrete numerical model (fast with millions of particles). We show that both approaches can be successfully applied to yield structure reconstructions and hydraulic properties predictions based on grain size distribution data and/or high resolution 2D cut images