Paper No. 17-6
Presentation Time: 3:05 PM
FINITE ELEMENT MODELING TO CONSTRAIN UNKNOWN GROUNDWATER CONDITIONS FOR LARGE LANDSLIDES IN THE SIGUAS RIVER VALLEY, PERU
Since the 1980s, the Majes-Siguas irrigation project has irrigated thousands of hectares of arid landscape in a plateau area of the pampas of southern Peru. Since the beginning of irrigation, several large, bedrock landslides in the adjacent Siguas River Valley have destroyed cropland and buildings and threatened road infrastructure and an industrial facility. While many researchers agree that these landslides are triggered by irrigation water infiltrating into the subsurface, the effect of the irrigation water influx on groundwater conditions and the relationship between the irrigation water and landslide triggering are poorly understood. Improving knowledge of these relationships will help stakeholders to better manage irrigation water and reduce the likelihood of triggering more landslides. Using finite element modeling, this work 1) quantifies the increase in water table height to trigger landslide movement, 2) tests the model response of different slope geometries to changes in groundwater conditions, and 3) evaluates the effect of water infiltration into the unsaturated zone on model instability. Using calibrated strength parameters, estimated stratigraphy, and DEM topography, models are computed with varied water table geometries and moist unit weights and compared using the critical Strength Reduction Factor (mathematically equivalent to the Factor of Safety). Models show greater sensitivity when groundwater is represented as a single water table, sub-parallel to the topography, rather than as a distinct perched zone above a deeper, separate water table. Results of modeling indicate that large landslides in this valley can initiate (i.e. the critical SRF drops below 1) when the water table is as deep as 200m below the top of the slope for simple profiles with one large slope, and as deep as 60m below the top of the slope for stepped slope profiles. Implications for landslide behavior and irrigation management are discussed.