Mapping of cracked soils and lateral water flow characteristics through a network of cracks

• A model to predict the lateral flow rate through a network of cracks is proposed.
• The model incorporates head losses due to changes in the crack aperture.
• A mapping procedure to obtain an idealized network of cracks is proposed.
• In a flow through a cracked soil, steady and transient state flows occur.
• Seepage into the soil matrix is smaller than flow through the network of cracks.

When a cohesive soil is dried, its volume tends to shrink in three directions. Shrinkage of the soil causes tensile stresses to develop and desiccation cracks will start to develop. The occurrence of cracks can significantly influence the lateral flow of water through the soil. It is important to understand the characteristics of lateral flow through a cracked soil. A model to predict the lateral flow rate through a network of cracks in the soils is proposed in this paper. In the proposed model, the actual network of cracks was idealized into a set of linear cracks. The flow through a single crack was modeled as a flow through parallel plates and the flow rate through the idealized network of cracks was calculated by incorporating the conservation of mass principle and the additional head losses due to the change in crack aperture. Laboratory experiments were performed to investigate the predictive performance of the model. Experiments were performed consisting of two main parts; namely, performing a desiccation test and performing a lateral flow test to measure the lateral flow rate through a cracked soil specimen followed by measuring water contents along the cracked soil specimen following the completion of the test. The laboratory test results indicated that during the lateral flow through the unsaturated soil specimens, two types of flow occurred which can be described as the steady state water flow through the network of cracks and the transient state seepage into the soil matrix. A comparison of the predicted and measured lateral water flow rates showed that the proposed model was able to predict the lateral flow rate through the network of cracks quite well.