A novel triggerless approach for mass wasting susceptibility modeling applied to the Boston Mountains of Arkansas, USA

This research deploys a novel mass wasting susceptibility modeling approach for cases where temporal information is unavailable and circumstances are prejudiced to merit applying traditional susceptibility modeling strategies. Conventional models typically employ approaches deemed problematic for this study, e.g., biased weighted input; a “more is better” approach pertaining to voluminous inputs; neglecting geologic structural influence; and establishing temporal linkages between cause (trigger) and effect (failure) with a trigger being defined as a catalyst for failure, such as timed events like earthquakes or precipitation as well as physical changes like vegetation removal or slope disturbance. Road bias may also influence modeling dramatically when event data are derived from observations of road-related failures, which become unreliable at predicting susceptibility in regions with no roads. However, a triggerless approach can extrapolate naturally occurring susceptibility via priori knowledge of local topography and structural geology factors. Two models are then created for comparison: One model has integrated empirical Bayesian kriging and fuzzy logic considering basically local topography and structural geology, while the second model has employed a standard implementation of a weighted overlay using all available (8) input data layers. Statistical comparisons show that the first model has identified ~ 83%, compared to only ~ 28% for the latter model, of the 47 documented mass wasting events in the selected study site. These results demonstrate that the introduced triggerless approach is efficiently capable of modeling mass wasting susceptibility in areas lacking temporal datasets, which in turn can help in mitigating future geohazards.