Climate-Induced Multimodality of Soil Moisture Distribution of Water Balance Cover

The characterization of in situ spatial variability of soil moisture content (θ) is highly relevant for understanding the landfill’s water balance (WB) cover hydrological processes, better modeling the processes, and applying them to effective cover design and planning. Considerable in situ soil heterogeneities and uneven climatic events result in unanticipated yield in the cover performance. Hence, it becomes necessary to probabilistically characterize the θ in the field for enhanced reliability. This study used Gaussian distribution analysis to produce probability density functions (PDF) of soil moisture at variable depths (0.30 and 0.76 m) of a pilot-scale WB cover under a semi-humid climatic region. In addition, the kernel density estimate (KDE) was used to produce a smooth estimate of the PDFs to define the multimodality of soil moisture distribution. The dimensions of the WB cover were 12 m × 12 m and 1.2 m in depth. The cover was constructed with local fine-grained soil (CH). Moisture sensors were installed at 0.3 and 0.76 m for continuous measurement of θ. The data collected from the instrumentation were analyzed using the R program. Results indicated that the soil moisture distribution at 0.30 m is more spread out than at 0.76 m depth. Also, the soil at 0.30 m depth showed a relatively more degree of normality than the soil at 0.76 m. The statistical analyses indicated that the soil at 0.76 m depth was significantly negatively skewed (skewness = −1.87). The θ distribution at both depths showed multimodality. The KDE-generated PDFs showed a distinguishable probability difference compared to the Gaussian distribution, especially at 0.76 m depth. The performance of Gaussian distribution was greatly improved using the KDE, especially when the soil moisture PDF showed highly skewed features. The results also indicated the initial soil moisture content affects the overall distribution of θ under the field conditions.