Rainfall at Collapse Tells Only Half the Story of Failing Rock Slopes

A new study argues that this final-window figure can be deeply misleading. In an analysis of twelve natural rock-slope failures with resolvable event timing and meteorological evidence, researchers found that five of the six events with quantitative event-window rainfall recorded essentially no rain in the 24 hours before collapse — five had a 24-hour precipitation of 0 millimeters, and one, at Pengshui, recorded just 0.9 millimeters. Yet behind those quiet final windows lay strikingly different histories of antecedent forcing.

The study, published in the journal Frontiers in Earth Science, was led by Meng-Lun Li and colleagues at the Beijing Institute of Geological Hazard Prevention and partner institutions in China. Its central point is methodological: the meteorological state at the moment of release may not represent the pre-failure history of a rock slope. Rainfall is central to slope-hazard assessment and early warning, but a single standardized number such as 24-hour precipitation, often abbreviated P24, compresses a complicated story of wetting and drying into one immediate descriptor.

To unpack that story, the team reconstructed failure times, rolling precipitation windows, antecedent rainfall forcing, and documented precursor evidence for a purposively selected set of twelve natural rock-slope failures. The events were chosen because their timing could be resolved and meteorological evidence existed — a deliberate selection that the authors stress makes the set non-representative. Because the search and retention criteria favored events with resolvable timing and meteorology, the sample cannot be used to estimate how common any particular pattern is.

Within the six quantitative cases, the antecedent histories ranged widely. Some slopes experienced multi-week dryness before failing; others had been subjected to earlier extreme rainfall, to cycling reservoir water levels, or to rainfall arriving on a short lag before collapse. A sensitivity check strengthened the case against reading too much into P24.

The researchers examined rolling windows of 12, 24, 48, and 72 hours across the event set and found that earlier rainfall re-entered the longer windows for some events. This confirms, the authors argue, that P24 is a standardized immediate descriptor rather than a trigger threshold — a convention for reporting, not a physical boundary between safe and dangerous conditions. The six remaining events in the twelve-case set served as boundary cases, used to separate physical-process boundaries from evidence-definition boundaries.

The event-level comparison leads to the study's key conclusion: a weak or absent final rainfall signal can coexist with contrasting pre-failure histories and monitoring evidence. What it does suggest is that a quiet final window should not be read as evidence that rainfall played no role. The authors are careful about the limits of their evidence.

The twelve events were purposively selected, not sampled at random, so no prevalence claims follow from the analysis. The findings are event-level comparisons rather than statistical generalizations, and the article's final formatted version is forthcoming.

Subject of Research: Antecedent rainfall histories and precursor evidence in natural rock-slope failures

Article Title: Contrasting slope histories under similar final-window rainfall in rock-slope failures: implications for monitoring

Article References: Li, M.-L., Yin, H., Liu, Y., Zheng, Y., Yu, Z., Han, J., Jiao, R., & Hu, F. (2026). Contrasting slope histories under similar final-window rainfall in rock-slope failures: implications for monitoring. Frontiers in Earth Science, 14, Article 1980922. https://doi.org/10.3389/feart.2026.1980922

Image Credits: AI Generated

DOI: 10.3389/feart.2026.1980922

Keywords: rock-slope failure, rainfall, antecedent forcing, P24 precipitation, geohazard monitoring, early warning, precursor deformation, slope instability, rockfall, reservoir water level, precipitation windows, landslide triggers