[Back]

A. Preh, E. Fleris:
"Process based rockfall modelling in 3D using data sets from RPAS.";
Poster: EGU General Assembly 2019 (EGU 2019), Vienna, Austria; 2019-04-07 - 2019-04-12; in: "Geophysical Research Abstracts", European Geosciences Union, Vol. 21, Vienna (2019), Paper ID 16942, 1 pages.

A number of methodologies for the quantitative rockfall hazard assessment and the definition of hazard or dangerzones have been proposed in the scientific literature (e.g. Mölk & Rieder, 2017). In general, the classification andranking of the factors contributing to hazard is an intrinsically uncertain task and different researchers proposedifferent techniques to tackle the problem. Despite some major differences, what is common in these approaches,is the use of 3D physically based rockfall modelling for the generation of the necessary spatially distributed data,in the form of 3D rockfall trajectories. Since this data serves as the basis for any subsequent hazard assessment, itis evident that data quality and its resolution will dictate the results of the task.The use of RPAS (Remotely Piloted Aircraft Systems), employing different data acquisition techniques (e.g.Photogrammetry - SfM (Structure from Motion), LiDAR) is constantly increasing in the field of geosciences. Inrelation to rockfall modelling, RPAS provide with advanced methods for obtaining remotely sensed data. DigitalElevation Models of fine resolution can be generated, Rockmass characteristics can be remotely investigated,overcoming difficulties imposed by physically inaccessible and/or dangerous terrains. A rockfall dedicatedfield-mapping campaign can be enriched with valuable information. A question that emerges is to what extent ourexisting numerical tools can utilize this information. What could be useful and what is not.For the past few years we have been working on developing numerical algorithms for rockfall modeling andfurther exploring the idea of a simple but yet effective hybrid lumped mass model, using a deterministic methodto mathematically treat rockfall impacts (Goldsmith, 1960), modified by the introduction of stochastic surfaceroughness and the calculation of hyperbolic restitutions factors (Bourrier & Hungr, 2011). WURF (Fleris & Preh,2016) is a PYTHON numerical code based on the aforementioned principles, creating through its functions avirtual environment for the study of rockfall in 3D.As resolution and the preciseness of remotely sensed data increase, finer geomorphological detail is being capturedin generated DEM´s. This can be problematic to high resolution 3D rockfall numerical modelling since it is toaffect the range of artificial roughness that should be stochastically introduced to rockfall simulations. Therecurrently exists limited information on the direct use of high resolution data sets (i.e. LiDAR) in rockfall numericalmodels.We are to present results from numerical modelling in 3D using WURF and utilizing data arriving from RPAS,both at regional and smaller topographic scales. Data sets have been acquired in the context of the NoeTALUSproject (Melzner & Preh, 2019).We are also to address how the remotely acquired data may assist in solving prob-lems while preparing for and conducting rockfall numerical modelling in 3D such as i) the correct representationof topography ii) the uncertainty of identifying rockfall release positions and measuring rockfall release volumes,iii) the spatial distribution of several key model input parameters (e.g. definition of homogenous regions of surfaceroughness and restitution) iv) model calibration and v) model efficiency.

rockfall, modelling, RPAS

https://publik.tuwien.ac.at/files/publik_286724.pdf