Submitted to: American Society of Agri Engineers Special Meetings and Conferences Papers
Publication Type: Proceedings
Publication Acceptance Date: 12/10/2000
Publication Date: N/A
Citation: Interpretive Summary: Gully erosion destroys large amounts of agricultural land in the USA, and leads to problems of river pollution and the filling of reservoirs. Much of the sediment eroded from gullies comes from slumping of the channel head. The gully head is the upslope end of the gully, marked by a vertical step in the slope between the non-eroded area upslope and the eroded gully downslope. The process of gully head collapse is not fully understood, making prediction and control of gullies difficult. We generally assume that slumps occur when gully heads are saturated with water. However, computer simulations of gully heads show that during a rainstorm, many gully heads remain dry, and are strong enough not to collapse. Collapse is only possible when the gully heads are cracked, allowing water to seep deeper into the head wall, weakening the soil. Our work has shown that cracking can occur when a gully head is undermined by the erosion of the slope below it. The main effect of the crack is as a zone for water movement, rather than as a physical weakness. This work is relevant to ARS and to the US public interest because it allows us to make more accurate predictions of how much soil erosion will occur under different circumstances.
Technical Abstract: Head-cut migration is a common phenomenon in many gullies and channels in loess areas, and occurs by a mixture of particle erosion and mass failure. The mechanics of head-cut mass failure are complicated by the fact that, below a relatively shallow wetting front, many gully heads maintain unsaturated pore conditions throughout rainfall and flow events, considerably increasing soil strength. This additional strength is enough to maintain head-cut stability in many cases. Finite element hydrology modeling has shown that pore suction can be destroyed by macropores and tension cracks in the head, and that in many cases strength can only be overcome through the presence of such preferential flow paths. Computer simulations show that negative pore water pressures of - 4 KPa can be maintained in a 1 meter high uncracked gully head despite several hours of overland flow, but that positive pressures of 4KPa develop in head-cuts with cracks behind the head-cut. Combining the hydrologic modeling with slope stability analysis shows that the difference in pore pressure is sufficient to account for head-cut instability in many instances, and that the effect of the crack is primarily hydrological rather than mechanical. Additional analysis with a finite element stress-strain model has shown the presence of tension cracks and basal hollows to be related to unloading of the slope downstream of a head-cut. The results suggest that incorporation of macro porous, as well as matric, hydrologic processes is an essential prerequisite for successful simulation of head-cut retreat by mass failure.