Geometrical and statistical analysis of dynamic crack morphology in shrink-swell soils with addition of maize roots or salinity (NaCl)

Abstract

Cracks generated in shrink-swell soils detrimentally affect hydraulic and mechanical properties of soils, and alter soil-water processes, such as gas diffusion, water flow, solute transport, and root penetration. Root additives and salinity beget physicochemical changes in soils, resulting in variations of crack morphology. Understanding these variations provides effective means of crack utilization or prevention. In this study, we investigated the effects of maize roots and sodium chloride (NaCl) on crack dynamics and morphology in shrink-swell light clays. Drying experiments were conducted in soils containing roots (root densities: 0, 0.5, 1.0 and 1.5 cm/cm3) or NaCl (mass contents: 0.03, 1.0, 1.5 and 2.0 %). Crack evolution patterns were photographed at regular intervals and crack parameters were calculated from the resulting images using morphological algorithms. We investigated the dynamic geometry, topology, and statistical behaviors of cracks. Results indicate that incorporating maize roots effectively restrained crack area and width, but increased crack length, tortuosity, and number of cracks. Crack networks in root-treated soils contained a vast number of dead-ends and displayed pronounced fragmentation and low connectivity. Distribution of crack intersection angle well approximate normality with a mean around 120°. The bonding and friction between roots and particles, the bridging effects, stiffness, and ductility of roots improved the soil tensile strength and altered the intrinsic release criteria of tensile stress. The NaCl treatments increased crack area and width but decreased total crack length and tortuosity. With a slight amount of NaCl, the crack network tended to be highly connected with no dead-ends. Distribution of crack intersection angle in sodic soils exhibited asymmetric bimodality and was well described by a mixture of two Gaussian distributions, with primary and secondary peaks around 105° and 165°, respectively. We proposed a statistical function that describes crack area probability distribution considering water content as a variable, which proved to be a good fit in modeling dynamic statistics of crack patterns. Our results provide compelling evidence of how roots and salinity alter crack evolution, dynamics, and morphology in shrink-swell soils, which will be beneficial in crack inhibition and utilization, and in hydrological use such as water management in saline-sodic soils and prevention of preferential flow. However, cracking behaviors are highly susceptible to sample size effects and basal boundary. We hypothesize cracking behaviors (in roots- or salinity-treated soils) in small scale and in-situ scale are dimensionally similar, and the validity of these conclusions being generalized to in-situ cracked soils should be further investigated.