Densely-packed traffic pan at hillslope fields can restrict downward soil water percolation. The hydraulic conductivity contrast between the top and traffic pan layers may lead to the perched water table (PWT) and downslope subsurface flow ensues above the pan layer. The baffling observation of upward advance of gullies or erosion channel at Tsumagoi (Japan) hillslope catchments has explained that erosion is often caused due to the return of infiltrated rainwater and rainfalls onto the saturated areas. Potential for such hillslope saturation and erosion is therefore an attribute located by the breakout intercept of the PWT on hillslope surface as Return Flow Generating Point (RFGP). The objectives of the study were: (1) to identify the mechanism of hillslope saturation formation by observing experimentally and numerically the soil water movement in a uniform hillslope layered with traffic pan, and (2) to understand the possible starting point of erosion by observing experimentally and numerically RFGP on the variably saturated hillslope surface. To observe the two-dimensional soil water movement, twenty experimental runs of the 8°, 12°, 16° and 20° volcanic ash soil model hillslopes of each 100 cm long, 20 cm deep (15 cm top and 5 cm pan layers were packed at 60~70% mass-basis water content), and 5 cm width were conducted under the simulated rainfalls of 50, 80, 100, 125, and 150 mm h^-1. In addition, two experimental runs with shallow soils (top and pan were 10 cm each) and two runs with relatively dry initial conditions (10~15% water content) prior to applied rainfall were performed for the 8° and 20° under 50 mm h^-1 rainfall, respectively. Modified HYDRUS-2D, a two-dimensional numerical model solving Richards' equation (for Darcian flow under isothermal and variably saturated conditions) with Galerkin finite element method and incorporating dynamic surface boundary conditions switched between seepage and atmospheric to consider water returned and/or saturation-excess rainfall, was used to simulate hillslope saturation and RFGP. Soil water pressure profiles, the PWT positions, saturation extents, and Darcian flux distributions were the fundamental aspects to assess experimentally and numerically the model hillslope soil water movement and saturation. Results of the each experimental runs generalized the pattern of soil water movement such that traffic pan caused hillslope saturation because of the rising saturated wedge-shaped PWT fed by infiltration into the upslope unsaturated zone, and by laterally inflowing unsaturated-saturated subsurface flow above traffic pan, leading to RFGP occurrence at steady state. Outflow hydrographs showed that saturation occurred from below, indicating that return flow coupled with saturation-excess rainfalls caused saturation overland flow. Visual inspection in all the experimental runs confirmed that RFGP extended potentiality to identify the location at which erosion was most likely to initiate on the downslope saturated surface. Comparison of numerical simulation results with the observed soil water movement patterns, saturation extents, and outflow hydrographs for hillslope saturation and the prediction of RFGPs indicated that modified HYDRUS-2D reproduced the phenomena and RFGP reasonably well. Experimental and numerical simulation runs notably clarified that hillslope with relatively smaller slope inclination, shallower soils, and drier initial conditions under the higher rainfall rates decreased RFGPs from the upslope, indicating more hillslope saturation. Confining attention to the PWT intersection as RFGP along a long hillslope surface under steady state, RFGP from the upslope was analytically estimated by equating the net subsurface flow passing through the evolved saturated RFGP section to the influx of rainfall above that RFGP. Comparisons of analytically-estimated RFGPs with the observed and simulated RFGPs were in agreement and thus generalized the occurrence of RFGP on a long uniform hillslope layered with traffic pan.  

Key words: Hillslope saturation, traffic pan, Return Flow Generating Point, laboratory experiment, numerical simulation