Effect of Climate Change On Slope Stability Problems - Geological ...

UNIVERSITY COLLEGE DUBLINSchool of Architecture, Landscape and Civil EngineeringEffect of ClimateChange On SlopeStability ProblemsDr. Ken Gavin

Outline of the Presentation• Glacial Till slopes in Ireland• Description of Probabilistic Analysis• Benchmarking & Conclusions• Case Studies of rainfall induced slopefailures

Survey of 150 year old railway embankments9080Average slope angle = 387060No. of Slopes504030201000 8 16 24 32 40 48 56 64 72 80Embankment Slope Angle (degrees)Jennings & Muldoon (2003)

Slopes are Partially Saturated-• Air voids act like minicapilliary tubes, drawingwater up from the saturatedzone below the water tableby suction• This suction providesinherent stability+

Stability is affected by rainfallFinalSuctionInitialSuctionwetted soilwetting frontdrier soilairwatersoil particleAs rain falls, suctionsreduce. A wetting frontdevelops as the groundbecomes almost saturatedfrom the top down

Effect of suction on slope stabilityFOSCapacityDemand• Stability is providedby near surface suction(-pwp)• Rainfall causesformation of wettingfront (increased pwp)reduces suction andmay cause failure –typically shallow slip

Reliability Index• Consider alternative form ofperformance function g(x):FOS= C / Dg(x) = C – DIf C – D ≤ 0, the slope hasfailed• is the mean of theperformance function, dividedby the number of standarddeviations between the meanand origin – usually = 2 isminimum acceptable

UCD Reliability Model - GASSAGeotech reliability analysis ofslopes usually assumes fixed soilproperties to determine criticalslip surface and the varies soilproperties along this fixed surfaceto determine reliabilityBy using polar transformationUCD consider soils properties andslip coordinates as variables andsolves simultaneously, invariantand computationally efficient

Performance ofUCD model• Method applied tobenchmark 1:2 slope from theliterature• A range of soil propertiesCases A-N were considered• The relative performance ofUCD model was best in all butone (Case H) of the fourteensoil profiles analysed, and inCase H all methods werebroadly similar

Partially Saturated Slope AnalysesDepth of failure surface (m)00.20.40.60.811.21 1.5 2 2.5 3FOSMMonte CarloGASSA• Model used to asses theperformance of an instrumentedslope in the Swiss Alps• Measurements of suction,rainfall intensity and soil strengthparameters were available. 0.5 mdeep failure occurred after 45hours into infiltration test.• GASSA and Monte-Carlo givevery similar results. Both suggestcritical conditions (2) when thewetting front depth reaches 0.5m.

For Additional Uncertainty – Increase FOSFOS2.62.52.42.32.22.12.01.91.81.71.61.51.4Slope angle25 o30 o35 o40 oCOV(h)=0.15 o COV()=0.10.10 0.15 0.20 0.25 0.30 0.35 0.40COV(C)Possible reasons for additionaluncertainty:(i) Untypical groundconditions(ii) Poor site investigation(iii) Complicated stratigraphy(iv) Lack of rainfall or otherdata(v)Natural variability

Conclusions - Analyses of slopes• A number of highly variable parameterscontrol the stability of slopes• Stability analysis is more suited toprobabilistic rather than deterministicmethods• Probabilistic models are available• Hybrid design approaches are also beingdeveloped

Stability of Infrastructure – Smart Slopesrainfall• In-situ monitoring ofsuction and in-situstressesPressureSuction• Real time assessment ofinfrastructure reliability• We cannot monitorevery slope, we need toidentify critical sections–Geophysics?

Recent Slope Failures

Rainfall RecordsOctober• no exceptional storms• overall rainfall 50%above normalNovember• Rainfall 84% abovenormal• Included storm on25 th with 58mm ofrainfall over 6 hourperiod.

Cross-Section

Geophysical InvestigationResistivityinvestigation acrossfailure plane andalong track to confirmsoil stratigraphy andinvestigate suspectedlateral flow of wateralong embankment

ERT profile across track through failure plane

ERT Profile Along Track

Finite Element AnalysisModel of post-construction behaviour

Seepage Analyses• Rainfall data was used tomodel infiltration intoembankment• FOS reduced to 1.2• Regrading of embankmentcaused increase in FOS to1.5

Stress distribution beneath embankment• Geophysical investigationhighlighted extent of softorganic clay present beneaththe slope• Tube samples were obtainedas part of a forensicinvestigation• Samples exhibited significantcreep at stress levels 80 – 100kPa

Conclusions - Case HistoryComplex groundconditions. The failurewas triggered byheavy antecedentrainfall and the effectswere amplified bycreep in the soft claydeposit. Geophysicsoffered key insightsfor low-cost

Acknowledgements• Irish Rail who sponsor our work on stabilityof steep slopes• Dr. Jianfeng Xue• Geophysical Research Unit at UCD andApex Geoservices• Arup Consulting Engineers