Correlating Site-Scale Erodibility Parameters from Jet Erosion Tests to Soil Physical Properties

Abstract

Abstract. One of the most commonly used methods of measuring erodibility parameters, i.e., critical shear stress (τc) and erodibility coefficient (kd), of cohesive soils is the Jet Erosion Test (JET). While numerous factors influence the erodibility parameters, the JET provides an in situ measurement technique. However, in many cases where erodibility parameters are required for simulating channel erosion processes, the erodibility parameters are not characterized in situ but estimated empirically based on soil physical properties with relationships that may not be good predictors for all streambanks. The objectives of this study were to investigate the correlation between the erodibility parameters measured with JETs and soil physical properties at a site-specific scale and across three unique streambanks. A total of 74 JETs were conducted within visually homogeneous streambank layers at three sites in Oklahoma along with measurements of soil physical parameters such as texture, bulk density, moisture content, and water and soil temperatures. At the site scale, τc and kd varied by up to three orders of magnitude. Neither multiple linear regressions nor principal components regressions suggested any consistent strongly correlated variables. Therefore, erodibility parameters measured in this study could not be predicted based solely on soil physical properties. It was concluded that τc and kd must be measured in situ and cannot be estimated from empirical relationships due to the heterogeneous nature of soil and the variability in subaerial processes, even within visually homogeneous streambank layers. More research is needed in order to correlate erodibility parameters to other soil parameters and quantify the role of subaerial processes, such as seepage, soil desiccation, and freeze-thaw cycles, on erodibility in order to incorporate spatial variability of erodibility parameters into stability and channel evolution models.