LAUSR

Abstract Soil infiltration, an important component of hillslope hydrology, is widely measured with ring infiltrometers. In the numerical algorithms used for soil infiltration measurement, the time step length (STL) for reading water level and positioning location (PL) of average infiltration rates within a time step considerably affect the measured infiltration rate curves. In this study, four TSLs (1, 2, 5 and 10 min) were used to record falling water depth and three PLs (initial, mid, and end points of the time step) were applied to position the measured average infiltration rate, to evaluate the effects of TSL and PL on the measurement accuracy of infiltration rate. For a specific TSL, three infiltration rate curves were obtained by positioning the average infiltration rate at initial point ( f i ), midpoint ( f m ), and end point ( f e ), respectively. Results show that the infiltration rates of f e increases with TSL increasing. A short TSL reduces measurement errors caused by the TSL. However, a short TSL produces high measurement errors caused by reading the Mariotte bottle scale and increases practical difficulties. The f m of different TSL were the closest to the true soil infiltration process, regardless of the TSL (i.e. 2, 5 or 10 min), with a maximum error in cumulative infiltration of approximately 11.14%. The TSL could be reasonably long, such as 5 or 10 min, as long as the measured average infiltration rates are positioned at the midpoint of a TSL. This strategy can avoid short-time measurement errors and reduce operational difficulties.