SIMULATION OF NEAR-SURFACE HYDROLOGIC RESPONSE AND SLOPE FAILURE ASSESSMENT AT THE COOS BAY, OREGON EXPERIMENTAL CATCHMENT

Shallow landslides present a natural hazard to human life, the built environment, and ecosystems. While prevailing conceptual models of failure initiation rely primarily on the reduction of effective stress by positive pore-water pressure development, recent research has demonstrated that capillary stress reduction from declines in matric suction can also trigger shallow slope failure. The growing field of hydrogeomorphology emphasizes the role of hydrologic drivers of slope instability, regardless of whether a given failure is triggered by unsaturated or saturated zone mechanisms. The most promising protocol for improving the methodology of hydrologically-driven slope-stability assessment is the “measure and model” approach, which employs field observations and measurements to parameterize and evaluate models.

Here we present an application of the measure and model paradigm at the Coos Bay Experimental Catchment in the Oregon Coast Range, USA (see Figure 1).  The hydrologic response to three controlled sprinkling experiments and six years of natural storms were simulated using the Integrated Hydrology Model (InHM). InHM employs the finite-element method to solve the fully-coupled governing equations of 3D variably-saturated subsurface and 2D surface water flow and solute transport. The hydrologic model was parameterized using estimates of saturated hydraulic conductivity, porosity, and in-situ hysteretic soil-water retention curves. Detailed comparisons of simulated results against observations of discharge, matric potential, soil-water content, total head, and tracer concentrations facilitate a diagnostic assessment of model performance. Slope stability is evaluated using the relatively simple infinite slope model for variably-saturated soils driven by the simulated pore-water pressures. The slope-failure assessments focus on an 18 November, 1996 storm with a total rainfall of 225 mm and a peak intensity of 40 mm / hr, which initiated a debris flow.

Hydrologic-response simulations were conducted for the failure-initiating storm using a hysteretic soil-water retention curve in addition to the primary wetting curve, the primary drying curve, and an intermediate (or “mean” curve) between the primary wetting and drying relationships. Comparison between simulated slope-failure potential at the Coos Bay site for the different soil-water retention curve scenarios suggests that (i) employing the drying soil-water retention curve or a mean of the primary wetting and drying soil-water retention curves underestimates failure likelihood, (ii) the wetting soil-water retention curve, which is seldom measured, is more appropriate when hysteresis cannot be considered, (iii) unsaturated zone storage provides an important control on failure initiation and sets the stage for fracture-flow driven triggering of instability at the Coos Bay site.

Figure 1. Map showing the location of the Coos Bay Experimental Catchment in coastal Oregon, USA and a diagram of the topography, instrumentation-monitoring platforms, and the approximate observed slope failure extent in 1996. Note that the topography was surveyed before the failure and does not reflect the post-failure topography.