A Comprehensive Model for Single Ring Infiltration II: Estimating Field‐Saturated Hydraulic Conductivity

Abstract

Core Ideas Four approaches were tested to estimate Kfs from single ring infiltration data. Highest accuracy occurred when capillary length was appropriately constrained. Kfs estimates improved if capillary length was overestimated vs. underestimated. Assuming a universal capillary length can also be useful for determining Kfs. Kfs can be accurately estimated using both early‐time and steady‐state data. In this study, we explored four approaches to infer field‐saturated hydraulic conductivity (Kfs) from both early‐time and steady‐state infiltration measurements using an explicit expression for three‐dimensional flow. All approaches required an estimate of the soil capillary length, λ. Approach 1 estimated Kfs via optimization, in which all other infiltration parameters (9 in total) were known. The remaining approaches constrained λ through different interpretations of coefficients generated by linear regression between infiltration and time. Approach 2 utilized these coefficients plus estimated soil water content to simultaneously quantify both λ and Kfs. Approach 3 used an analytical expression in which λ was estimated based on water retention/unsaturated hydraulic conductivity parameters, while Approach 4 adopted a universal λ value of 15 cm. The accuracy of these four approaches were tested using numerical and laboratory infiltration data. Approach 1 had the highest accuracy but also required the most auxiliary data, making it most suitable for laboratory and numerical experiments. Approach 2 was the least consistent, providing negative estimates for λ and Kfs under certain conditions. Approach 3 also gave accurate predictions of Kfs, but may be inaccurate in instances where the water retention model parameters are uncertain or do not describe soil hydraulic behaviors well. Approach 4 provided reasonable estimates of Kfs (within a factor of three from the actual value in most cases), while not requiring additional observational data. The optimal approach for interpreting Kfs will thus vary depending on the type and quality of available auxiliary data.