Field experimental studies on the erosion characteristics and dynamic mechanisms of rills that have already eroded (eroded rills) when rainfall occurs are limited, especially compared to the dynamics of those rills that have experienced erosion for the first time (eroding rills). The field runoff scouring experiments included two intermittent sub-experiments: flow discharge of 9, 13, 17, 21, 25, and 29 L min-1 (E1) with a scour duration of 60 min and flow discharge of 6, 9, 12, 15, 18, and 21 L min-1 (E2) with a scour duration of 10 min. Two antecedent soil moistures (15 % and 25 %) were used for the runoff plot (18 x 1 m) before starting E1. The results from E1 indicated that flow discharge and antecedent soil moisture increased the average runoff rate (19.46 % with discharge, 4.63 % with soil moisture), soil loss rate (63.94 % with discharge, 58.80 % with soil moisture), rill width (16.49 % with discharge, 4.66 % with soil moisture) and rill depth (13.40 % with discharge, 6.60 % with soil moisture). The average flow velocity varied from 0.231 to 0.344 m s-1, and the flow regimes were basically subcritical turbulent flows. The impact of rill depth on cumulative sediment yield was greater than that of rill width. Moreover, the critical values of flow velocity, shear stress, stream power, unit stream power, and unit energy under eroded rills were 0.183 m s-1, 7.532 Pa, 1.278 W m- 2, 0.032 m s-1 and 0.007 m from E2, respectively, and the rill erodibility parameters for E1 and E2 were 0.252 x 10-2 and 0.574 x 10-3 s m- 1, respectively. The flow unit stream power and stream power could be effectively used to simulate the dynamic mechanism under conditions E1 and E2, respectively. The study aids in understanding the dynamic mechanism of rill erosion under intermittent rainfall conditions in the field.
