Rockfall Processes in Alpine Catchments
Research Team: PD Dr. Florian Haas, Thomas Klein, Prof. Dr. Michael Becht
In high mountain regions, rockfall is an important geomorphic process in terms of sediment budgets, and as a natural hazard. High-resolution Terrestrial Laserscanning (TLS, Riegl LMS Z420i) was applied to study:
- the location and size of rock fall events within steep rockfaces
- the characteristics of rockfall deposits such as surface roughness, size dis- tribution and fragment morphology
- and their influence on rockfall runout length
Beside the quantification of rockfall activity and the spatial analyses of the the rock fans, the measurements should help to develop and to improve existing rockfall models. With the high reolution spatiotemporal data it should be possible to model the disposition for rockfall activity, and so to differentiate active areas on a rock face from inactive areas (e.g. dependency between lithological conditions and weathering). By knowing the starting areas of rock fall and knowing the runout distances of rocks on the rock fans, it should be possible to improve existing process models which model runout distances (e.g. dependency between surface roughness on the rock fans and runout distance of rocks).
To consider different climatic an lithologic conditions, rockfaces and the corresponding talus cones were scanned twice a year with two scanning resolutions in three alpine study areas (Val di Funes, Northern Dolomites/Italy; Horlachtal, Central Alps/Austria; Höllental, Northern Calcareous Alps/Germany).
Rock faces are scanned by TLS in two resolutions. To record larger events, large areas of the rockfaces and the corresponding talus cones were scanned from a great distance (~200-600 m) in a resolution of 0,1° (point density). In contrast, detailed scans from shorter distances (50-200 m) in high reolution (0,05°) were used to investigate the capability of the approach to detect smaller events.
The rock faces are scanned twice a year in spring and in the late summer. The registration of both scans (spring and summer) is done not by marked tie points but by a Multistation Adjustment (MSA) with the Scanner software RiscanPro (finding corresponding normal vectors of planes in both data sets). The accuracy of this MSA is very high (Standard Deviation 3 mm for short distance scans and 2-3cm for great distance scans), so it is possible to carry out a very accurate cut and fill analysis for both scans to evaluate the rock fall activity.
The registered point clouds are exported from RiscanPro as ASCii-files (x,y,z) and then imported into ArcView by changing the axes. After this the point clouds are triangulated and transformed into grid-format (cell size 2 x 2 cm short distance scans, 20x20cm great distance scans). The grid files are imported into the open source GIS SAGA to carry out cut and fill analyses. With this approach, three large and several smaller events were recorded in the three catchments. The largest event occurred 2008 in the Northern Dolomite Alps (Val di Funes/Italy) with a volume of nearly 3300 cubic meters (8700 tons). The figure shows as an example the result of the Cut and Fill - Analysis for the Val di Funes Rockfall event. For better visulisation the result is draped on an Analytic Hillshading (derived in SAGA) of this area. The two photographs show the rockface (detachment zone) before and after the event.
Beside roughness analyses the interrelationship between rock fragment morphology (characterised by shape parameters) and runout distance was analysed at the site of a large rockfall event (>10 000 cubic meters) from the year 2003 in the Northern Dolomite Alps. For these analyses the talus cone was scanned from 6 positions in order to minimize shadowing effects in the TLS point cloud.
The axial ratio (axial ratio = a:b:c; b = 1) of 618 rocks (>50 cm long axis) in the depositional zone of this rockfall and their corresponding runout distances were measured using TLS data and the software RiscanPro.
The results of these analyses show that there is no significant correlation between runout distance of rocks an their size. But there is a significant correlation between the axial ratio of the rocks and their runout distances. This result shows that rocks with a "round" shape (axial ratio around 1) have a longer runout distance than elongated or irregularly shaped particles (axial ratio greater than 1).
The figure shows a 3-D view of a cirque in the Zwieselbachtal (Central Alps), derived from Terrestrial Laserscanning data (Resolution of the point cloud is 0.05 degree). The following analyses are carried out for the red marked area. According to the roughness-length-method (Fardin et al. 2001, Fenq et al. 2002, Rahman et al. 2006), the dataset was fragmented in a regular raster. Each cell of the used raster has an extent of 5 x 5 meters.The value of the Standard Deviation is used as information for the roughness of the surface. A high Standard Deviation means a rough and a lower Standard Deviation means a smooth surface. The figure (right side) shows the derives Standard Deviation for each cell of the selected talus cone. It is clearly visible, that there is a higher roughness at the bottom (large rocks have a higher runout distance) and at the upeer part of the talus cone (influenced by debris flows). This map of surface roughness will be used as input data for future rockfall modeling.
The cell size of 5 x 5 meters (25 sqm) has been chosen, because the existing rockfall models of the working group are calibrated by using DEMs with this resolution. In ArcInfo a best-fit-plane through the point cloud has been generated for each single cell. This plane is used as a kind of regression plane to calculate the Standard Deviation of all measured points (from TLS) belonging to each single cell.
To validate the results of the roughness analysis, the point cloud of the talus cone was triangulated an the colorized with the calculated roughness data (figure on the right).