In order to help assess the threat posed by an earthflow complex to a critical municipal water source (Liddell Spring) located just within the toe of the complex, we investigated landslide geometry and stability, the hydrologic behavior of the earthflow complex, and the possible hydrologic connection between the landslide and the spring. Our combined, state-of-the-art geotechnical and hydrologic monitoring approach led to a far more nuanced, accurate model than either approach alone.

We drilled, instrumented, and monitored the earthflow landslide complex and downslope spring system through water year (WY) 2001. Subsurface investigations included small-diameter borings with drive sampling of landslide debris, coring of bedrock, and large-diameter borings penetrating into bedrock. Monitoring instrumentation included slope inclinometers (SIs); time domain reflectometry; vibrating wire piezometers and stream gaging instrumentation down drainage from Liddell Spring. All information was recorded by dataloggers at 15- and 20-minute intervals (except SIs). Additional analysis included general mineral analyses of water samples and x-ray diffraction (XRD) of spring turbidity samples.

The large-diameter borings provide evidence that a granodiorite body (located in the toe region of the slide) controls ground-water flow at the site and possibly the location of Liddell Spring. Water within the landslide complex appears to be derived from shallow storm-driven ground-water pulses and deeper continually-fed ground water that is forced up through the fractured marble and schist bedrock. Continuous water temperature and specific conductance data suggest ground-water addition to the karst system during storm events with leading-edge transit times of 4 to 6 hours following the onset of rain (Schmidt, C., 2001, consulting report). Preliminary XRD results indicate the spring is fed by ground-water sources upslope from the spring and the marble body. Slope stability analysis indicates that water levels control slide movement, and water levels varied markedly, however rainfall during WY01 was 66 % of normal and no landslide movement was detected. Monitoring continues in order to capture the conditions accompanying detectable movement of the landslide and refine our geohydrologic model of the system.