LINEAR REGRESSION ANALYSIS OF LANDSLIDE MORPHOMETRIC PARAMETERS FROM LARGE-SCALE LANDSLIDES IN NEVADA, USA

Nevada’s climate and prevalent tectonism through time have provided suitable conditions for large-scale landslides. However, the mechanism(s) driving the spatial distribution of landslides is unclear. Previous statistical analysis of factors influencing large-scale landslide abundance produced inconclusive results. To identify variables linked to the formation and frequency of large-scale landslides in Nevada, we evaluated correlations between morphometric parameters from landslides in Nevada.

Linear regression analyses were performed on data gathered from 694 large-scale (>1km²) landslide deposits exposed in Nevada. Data on the landslides were gathered from published USGS and Nevada Bureau of Mines and Geology maps. These data include morphometric parameters for the landslide deposits (landslide runout distance, drop, azimuth, and deposit length, width, and area), lithology, age of rock material and failure, and tectonic regime. Initially, analyses were conducted on the totality of data which showed no clear correlation between the factors. To account for differences in landslide type we separated the data into five subsets based on tectonic or geomorphic province, including basalt plateaus, Basin and Range (extensional), Oligo-Miocene calderas, Walker Lane (translational), and ash-flow tuff provinces. In each province, we used linear regression to evaluate paired comparisons for all geomorphic characteristics. After separating by geomorphic province, we found stronger correlations (R²>0.5), particularly for runout length vs. drop. Paired characteristics with moderate correlation (R²<0.5) were drop vs. deposit area, slide deposit length vs. slide deposit width, runout length vs. flow direction, and drop vs. flow direction. However, some variability exists in correlation strength between geomorphic provinces, and outliers (especially for the largest landslides) decrease some of the correlation values. Future work will include applying more advanced statistical techniques to assess correlation of other data types to evaluate the drivers of landslides in Nevada.