A pore network model of soil water repellency: Model implementation and experimental validation

点击翻译按钮获取中文摘要

Soil water repellency has a big impact on soil quality: it reduces the water holding capacity and it enhances overland flow, soil erosion and preferential leaching of agrochemicals. Better understanding of the mechanisms controlling soil water repellency is therefore needed for the sustainable use of water and soil resources. Existing studies demonstrated that soil hydrophobicity is strongly related to two factors: soil organic matter and soil water content. The experiments indicated a threshold behaviour: above a critical water content soils with a given percentage of soil organic matter are wettable (Contact Angle<90); below this critical water content, the contact angle rapidly increases and the soil turns hydrophobic (CA>90). The critical water content decreases with increasing soil organic content. Although there are empirical models that are able to effectively mimic this hydraulic behaviour, a mechanistic model that is able to predict occurrence of soil water repellency for varying soil properties is missing. Objective of this project is to implement and experimentally validate a pore-scale model that is capable to predict the occurrence of water repellency in soils of various texture, water content and soil organic content. Our central hypothesis is that the microscopic distribution of water repellent surfaces underpins the wettability at the macroscopic scale. We hypothesize that a soil turns water repellent when the fraction of pores with CA>90 is above the percolation threshold, which corresponds to the fraction of pores that need to be disconnected to block the macroscopic flow through the soil. We plan to develop a 3D pore-network model to simulate soil rewetting. Soil organic matter (SOM) will be distributed either uniformly or preferentially in the small pores. The wettability of each pore will depend on the amount of SOM per soil surface, its spatial distribution and the matric potential. We will test and validate our model using a soil mixed with mucilage from maize roots, which was shown to make the soil water repellent. We will parameterize the model measuring the CA of glass plates covered with mucilage. The undisturbed distribution of mucilage in the pore space will be captured using the environmental scanning electron microscope condensation technique. Then we will apply our model to natural soils showing a varying degree of water repellency. We will use X-ray photoelectron spectroscopy to correlate wettability with chemical interfacial data. Simulation of water drop infiltration will be compared to water drop penetration time tests captured with a camera and neutron radiography to visualize the infiltration and distribution over time. Capillary rise experiments will be performed to estimate the dynamic contact angle under varying matric potentials.