Impact-Detection Algorithm That Uses Point Clouds as Topographic Inputs for 3D Rockfall Simulations

Abstract

Numerous 3D rockfall simulation models use coarse gridded digital terrain model (DTM raster) as their topography input. Artificial surface roughness is often added to overcome the loss of details that occurs during the gridding process. Together with the use of sensitive energy damping parameters, they provide great freedom to the user at the expense of the objectivity of the method. To quantify and limit the range of such artificial values, we developed an impact-detection algorithm that can be used to extract the perceived surface roughness from detailed terrain samples in relation to the size of the impacting rocks. The algorithm can also be combined with a rebound model to perform rockfall simulations directly on detailed 3D point clouds. The abilities of the algorithm are demonstrated by objectively extracting different perceived surface roughnesses from detailed terrain samples and by simulating rockfalls on detailed terrain models as proof of concept. The results produced are also compared to that of rockfall simulation software CRSP 4, RocFall 8 and Rockyfor3D 5.2.15 as validation. Although differences were observed, the validation shows that the algorithm can produce similar results. With the presented approach not being limited to coarse terrain models, the need for adding artificial terrain roughness or for adjusting sensitive damping parameters on a per-site basis is reduced, thereby limiting the related biases and subjectivity.

Highly energetic rockfalls: back analysis of the 2015 event from the Mel de la Niva, Switzerland

“…In this way, smaller terrain geometrical features are also processed by SfM to obtain the detailed surface roughness of the main terrain types affected by rockfalls: (1) the active scree slopes of colluvium accumulated mostly from rockfall events and partly remobilized by snow avalanches and debris flows; (2) the blocky meadows where the active processes are less energetic or less frequent, and fines can slightly accumulate and thin vegetation settle in; and (3) the grassy alpine montane meadow where the morainic gray fine sediments are present, partly covered here and there by rockfall block fragments, and reshaped in small terrasses by livestock grassing. The perceived surface roughness affecting the initial geometry of rockfall impacts prior to scarring for the three samples is described in Noël et al (2021).…”

“…Rockfall runout distances depend on many factors, such as the geometry of the terrain and the rock's perceived surface roughness of the encountered materials (Noël et al 2021;Bourrier et al 2012;Jones et al 2000;Pfeiffer and Bowen 1989;Pfeiffer and Higgins 1990). The resulting reach angles from the horizontal and vertical traveled distances of the rock can thus vary greatly from site to site (Labiouse 2004;Volkwein et al 2011).…”

“…After performing rockfall experiments and sensitivity analyses, they noted that the slope profile geometry and perceived surface roughness at impact are the main elements controlling the runouts of their model (Jones et al 2000;Pfeiffer and Higgins 1990). These geometric controlling factors can be objectively considered from detailed terrain models, as shown in Noël et al (2021). The Pfeiffer and Bowen (1989) model, used with slight adjustments in CRSP 4 (Jones et al 2000) and in RocFall 8 (Rocscience Inc 2022) and from which Rockyfor3D 5.2 (Dorren 2015) is derived, should be valid for simulating rockfalls for sites and rocks sharing features in the same range as the tested sites.…”

“…The Pfeiffer and Bowen (1989) model, used with slight adjustments in CRSP 4 (Jones et al 2000) and in RocFall 8 (Rocscience Inc 2022) and from which Rockyfor3D 5.2 (Dorren 2015) is derived, should be valid for simulating rockfalls for sites and rocks sharing features in the same range as the tested sites. However, the observed differences between CRSP 4 and RocFall 8 (Noël et al 2021) remind us of the importance of independently calibrating the parameters for each software program using back analysis from events on sites sharing the same features (Volkwein et al 2011;Labiouse 2004).…”

Contribution of High-Resolution Digital Twins for the Definition of Rockfall Activity and Associated Hazard Modelling

“…From the different methods available to assess susceptibility and hazard for rockfall events [48][49][50], and characterise the 3D rockfall events [51], we selected a well-established strategy to quantify the probability for a rock fall to reach vulnerable elements and generate negative consequences. Probabilistic rockfall trajectory analysis tries to minimise both ontic and epistemic uncertainty by introducing a certain degree of variability in the inputs [52].…”