Failure of peat-covered hillslopes at Dooncarton Mountain, Co. Mayo, Ireland: Analysis of topographic and geotechnical factors

Shallow landslides involving loss of blanket peat are relatively uncommon but can nevertheless be environmentally and geomorphologically significant. This paper describes a cluster of 40 shallow landslides, including significant peat failures, which occurred within 2.5 km2 on Dooncarton Mountain (Republic of Ireland) in the evening of 19 September 2003 during an exceptional rainfall event. It examines two hypotheses: (i) that the combination of topography and local soil catena characteristics were the primary site factors that led to the occurrence of most of the failures, and (ii) that anthropogenic disturbance to the slopes in the form of boundary ditches or peat-cutting activities may have contributed to three of the largest peat failures. All of the landslides were mapped and sampled within 4 months of the event. Physical, hydrological and geotechnical properties of the slope materials were determined from samples obtained from nine representative failures. Possible controls on the stability of the slopes were investigated by modelling seven of the landslides and a series of hypothetical slopes using standard limit equilibrium methods. The humified blanket peat was underlain by a dominantly sandy mineral material, including a buried soil horizon, derived from weathering of the mainly schist bedrock. Direct shear tests established the respective shear strengths to be ϕ′ = 50°, c′ = 0 (buried soil) and ϕ′ = 21–25°, c′ = 8–11 kPa (basal peat), although back-analyses suggested that the average peat cohesion was lower than this on the failed slopes. Analyses of separate slope segments of the landslides showed that the main slope convexity, where the peat cover gives way to thin peaty soil, defined the zone of minimum stability. Failure of the slope segment immediately above the convexity was controlled by the depth of peat and the hydrostatic or possibly artesian water pressures within the slopes. Thus catena characteristics coupled with local topography (gradient and peat depth) were important determinants of slope instability. Two of the failed slopes were crossed by boundary ditches but, contrary to accounts from elsewhere, stability analyses suggested that these ditches did not contribute significantly to the landslides. Stability analysis of a slope affected by peat extraction also suggested no direct causal association, but the hydrological conditions developed in the vertical tine cuts did probably contribute to a large peat slide. At other locations, therefore, similar anthropogenic factors should be incorporated in landslide hazard assessments.