Spatial-temporal changes in flow hydraulic characteristics and soil loss during gully headcut erosion

Abstract. The temporal-spatial changes in flow hydraulics and energy consumption and their associated soil erosion remain unclear during gully headcut retreat. A simulated scouring experiment was conducted on five headcut plots consisting of upstream area (UA), gully headwall (GH) and gully bed (GB) to elucidate the temporal-spatial changes in flow hydraulic, energy consumption, and soil loss during headcut erosion. The flow velocity at the brink of headcut increased as a power function of time, whereas the jet velocity entry to plunge pool and jet shear stress logarithmically or linearly decreased over time. The jet properties significantly affected by upstream flow discharge. The Reynold number, runoff shear stress, and stream power of UA and GB increased as logarithmic or power functions of time, but the Froude number decreased logarithmically over time. The flow of UA and GB was supercritical and subcritical, respectively, and transformed to turbulent with inflow discharge increased. The Reynold number, shear stress and stream power decreased by 56.0 %, 63.8 % and 55.9 %, respectively, but the Froude number increased by 7.9 % when flow dropped from UA to GB. The accumulated runoff energy consumption of UA, GH and GB positions linearly increased with time, and their proportions of energy consumption are 18.3 %, 77.7 % and 4.0 %, respectively. The soil loss rate of the UA-GH-GB system initially rose and then gradually declined and levelled off. The soil loss of UA and GH decreased logarithmically over time, whereas the GB was mainly characterized by sediment deposition. The proportion of soil loss at UA and GH are 11.5 % and 88.5 %, respectively, of which the proportion of deposited sediment on GB reached 3.8 %. The change in soil loss of UA, GH and GB was significantly affected by flow hydraulic and jet properties. The critical energy consumption initiating soil erosion of UA, GH, and GB are 1.62 J s⁻¹, 5.79 J s⁻¹ and 1.64 J s⁻¹, respectively. These results are helpful to reveal the mechanism of gully headcut erosion and built headcut migration model.