Gravity erosion on the steep loess slope: Behavior, trigger and sensitivity

• Many sentences and words in vague meaning are revised.
• Numerous words have been re-ordered to improve flow and understandability.
• Numerals are normally spelled out, especially when starting a sentence.
• Abbreviations, e.g., h and min, are not used in full narrative.
• The term topography meter needs not to be capitalized for it is not really a proper title.

Gravity erosion is a dominant geomorphic process on the widespread steep loess slopes, yet it is not well understood due to the complexity of failure occurrence and behavior. This study conducted a series of experiments in the laboratory to test the stability of different slope geometries and rainfalls and then performed a sensitivity analysis to quantitatively explore the triggering mechanisms of mass failure on the steep loess slope. A topography meter designed by the authors was used to quantitatively measure the process of gravity erosion, and the increase-rate-analysis method presented by the authors was also used to analyze the sensitivity of gravity erosion. The following three types of gravity erosion were observed: landslide, avalanche, and mudslide. In an event of rainfall, various types of gravity erosion might emerge in the same period, and mass failures with the same mode and similar size often adjacently appeared. Sometimes, a group of mass failures might happen on a large, slowly slipping block. Then the increase-rate-analysis method was used to evaluate variations in the gravity erosion with respect to changes in other causal parameters of rainfall duration–intensity and slope height-gradient. Climate-driven factors and topography triggers had prominent influences on gravity erosion. Whether for the total amount or the peak amount in an experiment, the largest sensitivity parameter on both landslides and mudslides was that of rainfall duration. In comparison, topography was relatively less influential. For the total amount in an experiment, the sensitivity parameters of rainfall duration on the landslide and mudslide were 24.9 and 19.5, respectively, while the sensitivity parameter of rainfall intensity on the avalanche was 2.2. For the peak amount in an experiment, the sensitivity parameter of rainfall duration on the landslide and mudslide were 5.5 and 15.6, respectively. Meanwhile the sensitivity parameter of slope gradient on the avalanche was 4.6. The experimental results obtained here provide an insight into the pre-failure mechanisms and processes of steep loess slopes.