Search

Federal Institute for Geosciences and Natural Resources, Hanover, GermanyCorrespondence to: P. Reichenbach (paola.reichenbach@irpi.cnr.it)The special issue of Natural Hazards and Earth SystemSciences entitled “Progress in landslide hazard and risk eval-uation” contains 9 out of more than 30 oral and poster con-tributions originally presented in the “NH3.11 Landslide haz-ard and risk assessment, and landslide management” sessionheld at the General Assembly of the European GeosciencesUnion, in Vienna (Austria), on 22–27 April 2012.The session was aimed at comparing qualitative or quanti-tative landslide hazard and risk estimates in different physio-graphic and geographical settings for different kinds of pro-cesses affecting the environment at different spatial scales.Presentations focusing on the adoption and integration ofdifferent modeling approaches for quantitative assessmentswere welcomed, and papers providing information on thequality, reliability and limitations of process-oriented or sta-tistical models were encouraged.The meeting proved to be a valuable opportunity to discussand compare methods, techniques and tools for the recog-nition, evaluation and mitigation of landslide hazards andthe related risks. In particular, the quality, the reliability andthe limitation of models, input variables and output mapshave been discussed for presented case studies at differentgeographical scales and in different physiographic environ-ments. This is important information for filling the gap be-tween academic research and application of the results inenvironmental planning and management. The special issuecontains the following selected papers presented and dis-cussed at the session.Voumard et al. (2013) have calculated risk along roadsusing a dynamic traffic approach that simulates the durationof the presence of vehicles inside hazardous areas during agiven time interval. The risk is analyzed along three roadsections threatened by different natural hazards (active land-slide, debris flows and dolines) along a mountain road: Aigle– Col du Pillon (western Switzerland). Two different scenar-ios were simulated: (1) a road without obstacles and (2) aroad regulated by traffic lights. Results of the dynamic riskassessment were compared with the static methodology thatconsiders an average number of vehicles per time unit and aconstant vehicle speed. The main advantage of the dynamicapproach is a better representation of the real traffic regardingthe interaction between different vehicles of different types(cars, trucks, coaches). The dynamic approach with a micro-scopic traffic simulator was well designed to analyze the riskon relatively short road sections (up to a few kilometers) indetail. At the regional scale, the risk estimations would beaveraged over the entire network (with large parts at no risk),and differences between static and dynamic risks may not beso pronounced. Thus, the interest of this method is to analyzehotspots, i.e., strongly hazardous short road sections, and tosee for example how the location of traffic lights can increaseor reduce the risk.Catani et al. (2013) have performed different tests to un-derstand how model tuning and model parameters can af-fect landslide susceptibility mapping. In the paper, the au-thors have adopted the random forest (RF) technique to pro-duce an ensemble of landslide susceptibility maps for a setof different models, input data types and observation scales.The RF model was initially applied using the complete set ofinput variables, then an iterative process was implemented,and progressively smaller subsets of the parameter spacewere considered to estimate the relative importance of sin-gle input parameters and to select the optimal configurationof the classification model. The main results are that (i) theoptimal number of parameters varies with scale and resolu-tion, (ii) the importance of each conditioning variable is in-fluenced by the model settings and the available data, andPublished by Copernicus Publications on behalf of the European Geosciences Union.

Rock slides and rock falls are the most frequent types of slope movements in Yosemite National Park, California. In historical time (1857–2002) 392 rock falls and rock slides have been documented in the valley, and some of them have been mapped in detail. We present the results of an attempt to assess rock fall hazards in the Yosemite Valley. Spatial and temporal aspects of rock falls hazard are considered. A detailed inventory of slope movements covering the 145-year period from 1857 to 2002 is used to determine the frequency-volume statistics of rock falls and to estimate the annual frequency of rock falls, providing the temporal component of rock fall hazard. The extent of the areas potentially subject to rock fall hazards in the Yosemite Valley were obtained using STONE, a physically-based rock fall simulation computer program. The software computes 3-dimensional rock fall trajectories starting from a digital elevation model (DEM), the location of rock fall release points, and maps of the dynamic rolling friction coefficient and of the coefficients of normal and tangential energy restitution. For each DEM cell the software calculates the number of rock falls passing through the cell, the maximum rock fall velocity and the maximum flying height. For the Yosemite Valley, a DEM with a ground resolution of 10 × 10 m was prepared using topographic contour lines from the U.S. Geological Survey 1:24 000-scale maps. Rock fall release points were identified as DEM cells having a slope steeper than 60°, an assumption based on the location of historical rock falls. Maps of the normal and tangential energy restitution coefficients and of the rolling friction coefficient were produced from a surficial geologic map. The availability of historical rock falls mapped in detail allowed us to check the computer program performance and to calibrate the model parameters. Visual and statistical comparison of the model results with the mapped rock falls confirmed the accuracy of the model. The model results are compared with a previous map of rockfall talus and with a geomorphic assessment of rock fall hazard based on potential energy referred to as a shadow angle approach, recently completed for the Yosemite Valley. The model results are then used to identify the roads and trails more subject to rock fall hazard. Of the 166.5 km of roads and trails in the Yosemite Valley 31.2% were found to be potentially subject to rock fall hazard, of which 14% are subject to very high hazard.

Identification and mapping of landslide deposits are an intrinsically difficult and subjective operation that requires a great effort to minimise the inherent uncertainty. For the Staffora Basin, which extends for almost 300 km2 in the northern Apennines, three landslide inventory maps were independently produced by three groups of geomorphologists. In comparing each map with the others, large positional discrepancies arise (in the range of 55–65%). When all three maps are overlain, the locational mismatch of landslide deposit polygons increases to over 80%. To assess the impact of these errors on predictive models of landslide hazard, for the study area discriminant models were built up from the same set of geological-geomorphological factors as predictors, and the occurrence of landslide deposits within each terrain-unit, derived from each inventory map, as dependent variable. The comparison of these models demonstrates that statistical modelling greatly minimises the impact of input data errors which remain, however, a major limitation on the reliability of landslide hazard maps.