In alpine canyon areas, rock mass landslides play an important role in processes such as surface erosion, chemical weathering, hydrological processes, carbon cycling and geomorphological evolution. At the same time, rock mass landslides, especially large rock avalanches, are a serious natural disaster, threatening the safety of human life and property, and causing serious damage to infrastructure. Therefore, in order to better quantify the flow of material caused by landslides and mitigate the hazards of landslides, the control factors of the scale of rock mass landslides have always been research difficulties and hot issues in the field of earth science and engineering disaster reduction.
Previous studies have shown that the landslide area, depth and width related to the scale of the landslide show the characteristics of power law distribution, showing that the larger the landslide area, the deeper the failure depth and the wider the failure width, and the scale of the landslide is related to the rock mass strength of the bedrock (Larsen et. al, 2010）。 Other studies have revealed that the strength of the rock mass of the bedrock is mainly affected by the structural surface and weathering, so the above two factors may also be the main factors controlling the scale of the rock landslide (Clarke and Burbank, 2010).
Theoretical and observational studies have shown that there are significant differences in the distribution of structural surfaces of underground bedrock due to tectonic and gravitational field disturbances on the terrain. These studies show that the topographic stress field changes with the size and direction of tectonic stresses, gravitational stresses and topographic shapes, thus affecting the horizontal and vertical range of bedrock subsurface structural surface development. These open structural surfaces influence various physical and chemical weathering processes of geological bodies and control the bottom boundaries of the weathered bedrock. Therefore, the extent of these open structural surfaces controlled by topographic stress provides constraints on the occurrence of bedrock landslides and controls the strength of the rock mass of the bedrock, thereby limiting the scale of bedrock landslides. However, due to the difficulty of characterizing these structural surfaces on a regional spatial scale, the above related studies remain blank.
Research by Gen K. Li of UCLA and his collaborator Selgi Moon (Li and Moon, 2021) shows that topographic stress is the result of the interaction of tectonic stress and topography, which coordinates with fracture and weathering to adjust the strength of surface rocks and soils to control the occurrence of bedrock landslides of different scales.
The authors selected the Longmen Mountains on the eastern edge of the Qinghai-Tibet Plateau as the study area, which has steep terrain, frequent monsoons and seismic activity leading to dense distribution of landslides, and selected a study area with a total area of 525 km2 in the middle of Longmen (Figure 1). Based on the 0.5m resolution satellite image, the earthquake landslide catalogue during the main earthquake and aftershock of the 2008 Mw7.9 Wenchuan earthquake and the landslide list triggered by rainfall before the Wenchuan earthquake were established. The authors selected 982 rock landslides with a source area of more than 20,000 m2 in the catalogue, of which 121 were rainfall-type landslides and 861 were seismic landslides.
Figure 1 Basic information map of the study area (Li and Moon, 2021). (a) Geomorphological location map of the study area; (b) Geological formation map of the study area; (c) Elevation map of the study area; (d) Landslide distribution map of the study area; (e) Fracture tendency hierarchy map of the study area
Considering the effects of tectonic and gravitational fields on topographic stress fields, the authors simulated the underground stress field in the study area using the 3D boundary element code Poly3D (Clair et.al, 2015), and the data for the tectonic stress field was derived from the collected hydraulic fracturing data. The magnitude and direction of the principal stresses (first compressive stress (σmcs), second compressive stress (σics) and third compressive stress (σlcs)) in the study area were obtained, and two indicators were calculated to express the tendency of the fracture surface of the rock mass, namely the rupture tendency (FP) and σlcs, which represented the tendency to generate or reactivate shear and open structural surfaces, respectively. FP is calculated from (σmcs- σlcs)/(σmcs + σlcs).
The authors plotted the spatial distribution of 500m depth FP (FP500m), 500m depth minimum compressive stress (LCS500m) and σlcs=10MPa (D10MPa) (Figures 2b-2d), thereby inferring the horizontal and vertical ranges of structural surface development. The indicators FP and σlcs show significant horizontal changes in the geomorphology of the study area near the surface (Figure 2). Based on high FP500m values, low LCS500m values, and deep D10MPa values, it is inferred that the deep open structure surface develops in the ridge area and the shallow open structure surface develops in the valley area (Figure 2b-d). As the depth increases, the FP decreases and the σlcs increase (Figure 2d-f). FP500m was positively correlated with both FP1000m and FP1500m, but less correlated with FP1500m (Figure 2g). This confirms that the effect of near-surface terrain disturbances on the corresponding force field is large and decreases with increasing depth (Figures 2h-2i). According to the area of the study area and the simulation time, the authors selected the average and maximum values represented by the stress at a depth of 500m, and compared them with the measured size of the rock landslide.
Fig. 2 Stress model result plot (Li and Moon, 2021)
The authors compared the area of 982 bedrock landslides with the distribution of open structural surfaces induced by topographic stress, and obtained the following key results:
(1) In earthquake- and precipitation-induced landslides, there was a strong positive correlation between the upper limit of bedrock landslide area and the maximum value of FP500m (FPmax); this correlation was better than the weak correlation, but it was statistically significant, and the average and maximum values of FP500m were positively correlated with the source area of all bedrock landslides (r = 0.17-0.35, P
(2) The relationship between the observed earthquake and precipitation induced landslide area and FPmax largely overlaps (Figure 3). This result shows that the control of terrain stress over the upper limit of landslide size may be achieved by adjusting the strength of fracturing and weathered materials, and is independent of the landslide trigger mechanism. When FPmax exceeds ~0.4, precipitation-induced landslides tend to be larger than earthquake-induced landslides. During periods that were wetter than the Wenchuan earthquake (May 12, 2008), precipitation-induced landslides were more likely to occur. In this case, a deeper open structural surface may cause groundwater seepage deeper, increased pore pressure, and higher FPmax, resulting in larger landslides.
(3) The correlation between landslide area and FPmax is stronger than the mean of FP500m (Figure 3). This may indicate that the weakest part of the underground, whether from the maximum depth of the open structural surface or the densest part of the open structural surface, has a greater impact on the scale of the landslide and less on the average landslide scale. This is consistent with previous experimental studies that the non-uniformity of matter (such as fractures or weak layers of stratigraphy) affects the size distribution of large landslides.
In further studies, the authors investigated other potential controls for the upper limit of landslide size, including topographic indicators (i.e., gradient, local topography, mean negative curvature, elevation, and distance from the backwall of the landslide to the ditch), seismic seismic indicators (i.e., peak ground acceleration (PGA)," precipitation-related indicators (i.e., average annual precipitation and precipitation intensity of 90 percentiles), and slope instability indicators, taking into account topographic slope and groundwater hydrology. The authors found that the correlation between these factors and the upper limit of rock landslide size was not as strong as that of stress indicators (Figures 4 and 5). These factors (such as slope and PGA of earthquake-induced landslides, elevation of precipitation-induced landslides) correlated better with median landslide size or landslide polygon density compared to stress indicators. These different correlations may indicate that the size and density of the landslide may be controlled by different factors.
Figure 3 Comparison of source area and FP area of different types of rock landslides (Li and Moon, 2021)
Figure 4 Relationship between landslide area and topographic indicators (Li and Moon, 2021)
The results show that the extent of the underground open structural surface induced by topographic stress affects the scale of large bedrock landslides in the steep mountainous areas of the eastern Qinghai-Tibet Plateau. The findings are similar to landslides caused by earthquakes and precipitation. As the bedrock erodes, the structural surface begins to open due to the influence of surface conditions such as terrain disturbances and other corresponding force fields. The opening fracture of the bedrock further facilitates the penetration and circulation of fluids into the ground, enhancing physical and chemical weathering processes such as freeze cracking and transport. In addition, subcritical crack propagation caused by changes in near-surface environmental conditions may enhance the opening of the fracture. While there may be multiple mechanisms at play, the authors' study suggests for the first time that the extent of the deep open structural surface caused by topographic stress determines the maximum size and depth that a bedrock landslide can reach.
While the authors revealed strong correlations between landslide area and stress indicators, the authors also found correlations between landslide area and other topographic indicators such as slope, elevation, and distance to the river channel (Figure 4). Due to the terrain's involvement in underground stress disturbances, slope stability, and seismic amplification, there is moderate spatial covariance (Figures 4, 6), and these correlations are to some extent to be expected. For example, terrain with a lower curvature or higher elevation (ridge area) may have deeper structural surfaces opening due to greater terrain disturbances and amplification of seismic vibrations. The authors found some degree of synergy between FPmax and other controls (Figure 6). Most large bedrock landslides (backwall area > 40,000 > 100,000 m2) occur in FPmax in excess of 0.4, higher than the linear fit between FPmax and other controls based on all landslides (Figure 5). These results show that even under wide-ranging covariance control, the effect of FPmax on landslide magnitude is obvious.
Area density and slope of earthquake-induced landslides (r = 0.56, P
Figure 5 Comparison of source areas of rocky landslides with seismic and precipitation control (Li and Moon, 2021)
Fig. 6 Comparison of conditions for FPmax with other potential controls (Li and Moon, 2021)
Overall, the authors' findings reveal connections between surface and subsurface tectonic stresses, topography, structural surfaces, and erosion. Topographically induced bedrock fissures are not only key factors in controlling the scale of landslides, but also induce feedback mechanisms between surface and subsurface processes in tectonic activity areas. The authors' method can be applied to other landslide-prone environments with different environmental conditions to assess the control of terrain stresses over landslide scale. Future studies will compare ground measurement data and bedrock material properties in different tectonic areas, including fracture patterns and degrees, to further address the factors that contribute to landslide occurrence and magnitude.
Key References (Swipe up and down to view)
Clair J S, Moon S, Holbrook W S, et al. Geophysical imaging reveals topographic stress control of bedrock weathering[J]. Science, 2015, 350（6260）: 534-538.
Clarke B A, Burbank D W. Bedrock fracturing, threshold hillslopes, and limits to the magnitude of bedrock landslides[J]. Earth and Planetary Science Letters, 2010, 297（3-4）: 577-586.
Li G K, Moon S. Topographic stress control on bedrock landslide size[J]. Nature Geoscience, 2021, 14: 307-313. （原文链接）
Larsen I J, Montgomery D R, Korup O. Landslide erosion controlled by hillslope material[J]. Nature Geoscience, 2010, 3（4）: 247-251.
Written by: Qi Shengwen/Shale Gas and Engineering Room
Editor: Chen Feifei
Proofreader: Zhou Xingxing Liu Qi county