The application of GIS-based binary logistic regression for slope failure susceptibility mapping in the Western Grampian Mountains, Scotland

Abstract: Slope failure has resulted in significant disruption to the Scottish road network in recent years and failure processes are widely considered to pose a very real risk to both infrastructure and road users. The manifestation of proposed regional climate variations could increase the hazard posed by landslide and debris flow activity within upland environments. It is therefore in the interests of decision makers and land managers to delineate the susceptibility of these areas to failure activity. The availability of accurate and high resolution geophysical data presents an opportunity to conduct a susceptibility analysis of proposed risk areas based on existing sites of failure. It is considered that failure sites are identifiable prior to activity and that events are triggered by external forcing in the form of excessive antecedent precipitation conditions. Binary logistic regression analysis is utilized to identify independent geophysical parameters that have been most associated with instances of past failure events. This technique facilitates the delineation of locations characterized by key parameter conditions most inductive to failure given the occurrence of an external trigger. It is proposed that when exposed to external forcing these locations are most susceptible to failure. To identify these locations is paramount to the successful application of any monitoring and/or preventative strategy. A Geographical Information System (GIS) is the ideal platform from which to undertake such a susceptibility analysis as it facilitates the precise identification of key independent parameter data associated with recorded instances of existing failure locations. The preparation, storage, extraction and analysis of intrinsic geophysical parameters promotes the development of a consistent modelling approach which can be applied to additional regions in the future.