An analysis of landslides in Great Britain using soil texture, rainfall, and topography reveals contrasting failure conditions between organic and mineral soils

Jane Elliot, Siul Ruiz, Daniel McKay Fletcher

Rainfall-induced landslides cause millions of pounds in damage to infrastructure in Great Britain (GB) annually and occasionally result in human fatalities. However, there are limited guidelines or policies aimed at reducing landslide risk in GB and few studies have broadly characterized landslide incidence across the region. Furthermore, peat landslides, which are a phenomenon that occur almost exclusively in the British Isles, have recently made headlines for their destructive impacts and degradation of a nonrenewable terrestrial carbon store. Given the environmental significance of peat, we explore the knowledge gaps surrounding the mechanical conditions that induce peat failures. We start by empirically characterizing landslide incidence in GB considering landslide events recorded in the British Geological Survey (BGS) database. Soil texture, topographic, and antecedent rainfall data were acquired for the considered landslides. Key distinctions in spatial, temporal, and triggering conditions between organic and mineral soil landslides were identified using a heuristic analysis, statistical testing, and dimension reduction techniques. Organic landslides had significantly steeper slopes and higher antecedent rainfall sums than mineral landslides and occurred most frequently in late summer and early autumn months. Using a K-means clustering analysis, landslide groups exhibiting similar slope, soil, and rainfall characteristics were identified revealing unique intra-cluster spatial and temporal patterns. The spatial distribution and temporal patterns of organic landslides are interpreted to suggest that antecedent rainfall, and by proxy soil moisture content, play key roles in rainwater infiltration and organic soil landslide susceptibility in GB. Seasonal drops in peat moisture content may facilitate rainwater infiltration via desiccation cracks and increase peat landslide susceptibility in late summer months. Our results highlight contrasting mechanisms for peat landslides, which can be used to guide more accurate landslide risk management considering region and preconditioning factors which is pertinent for recent peatland restoration activities in GB.