Monitoring and modeling of slope movement on rock cliffs prior to failure

摘要

In this paper we examine the use of terrestrial laser scan data for monitoring large rock cliffs inorder to assess the dynamics of the slope condition prior to failure. The assessment presented uses a combinationof this terrestrial data and air-borne mapping and imagery to explore the degree to which small-scale precursorsto slope failure can be identified,and ultimately used to predict failure. The degree to which data is projectedfrom any given perspective which represents the development of the slope is suggested to be critical. Here wepresent data that has been collected for four years at monthly intervals from over 35,000 m3 of near-verticalcoastal rock face. This high-resolution data is used to identify characteristic spatial and temporal patterns inrockfall activity notably in the periods leading to large-slope failures. To date in excess of 60,000 discreteevents have been recorded ranging from 0.00001 m3 to 2,500 m3. The patterns in the data show similaritywith time-dependant models of failure mechanisms,which may enable the prediction of failure occurrence bothin time and in space. This interpretation is only reached however with appropriate and,in terms of survey,unconventional treatment. Analysis of the data derived from the laser scanning suggests that given sufficientmeasurement precision,precursory behavior can be identified and monitored but that the relative viewing angleof the observation compared to the vector of deformation is critical. In this instance pre-failure deformations aremanifest as the rate of rock fall activity prior to failure,in addition to the direct measurement of the accumulationof the strain field across the failing rock face. The monitoring data implies a time-dependent sequence in theoccurrence of rock falls in the period leading to the largest failures recorded,often mirroring the hyperbolicincreases in movement witnessed in standard displacement monitoring or laboratory simulation prior to failure.The implication is that combining this data with models of failure mechanisms may allow failure time to beforecast from wide-area monitoring of precursory behavior. This ultimately means that a small number ofinstruments can be employed to monitor large expanses of rock face,providing both spatial and temporal data.These approaches,combined with the time-dependant modeling of precursory activity,have implications for themanagement of potentially unstable slopes,the understanding of slope failure mechanisms and potentially feedinto a new generation of slope failure warning systems applicable to both natural and man-made situations.