Search

Empirical models are frequently applied to produce landslide susceptibility maps for large areas. Subsequent quantitative validation results are routinely used as the primary criteria to infer the validity and applicability of the final maps or to select one of several models. This study hypothesizes that such direct deductions can be misleading. The main objective was to explore discrepancies between the predictive performance of a landslide susceptibility model and the geomorphic plausibility of subsequent landslide susceptibility maps while a particular emphasis was placed on the influence of incomplete landslide inventories on modelling and validation results. The study was conducted within the Flysch Zone of Lower Austria (1,354 km 2 ) which is known to be highly susceptible to landslides of the slide-type movement. Sixteen susceptibility models were generated by applying two statistical classifiers (logistic regression and generalized additive model) and two machine learning techniques (random forest and support vector machine) separately for two landslide inventories of differing completeness and two predictor sets. The results were validated quantitatively by estimating the area under the receiver operating characteristic curve ( AUROC ) with single holdout and spatial cross-validation technique. The heuristic evaluation of the geomorphic plausibility of the final results was supported by findings of an exploratory data analysis, an estimation of odds ratios and an evaluation of the spatial structure of the final maps. The results showed that maps generated by different inventories, classifiers and predictors appeared differently while holdout validation revealed similar high predictive performances. Spatial cross-validation proved useful to expose spatially varying inconsistencies of the modelling results while additionally providing evidence for slightly overfitted machine learning-based models. However, the highest predictive performances were obtained for maps that explicitly expressed geomorphically implausible relationships indicating that the predictive performance of a model might be misleading in the case a predictor systematically relates to a spatially consistent bias of the inventory. Furthermore, we observed that random forest-based maps displayed spatial artifacts. The most plausible susceptibility map of the study area showed smooth prediction surfaces while the underlying model revealed a high predictive capability and was generated with an accurate landslide inventory and predictors that did not directly describe a bias. However, none of the presented models was found to be completely unbiased. This study showed that high predictive performances cannot be equated with a high plausibility and applicability of subsequent landslide susceptibility maps. We suggest that greater emphasis should be placed on identifying confounding factors and biases in landslide inventories. A joint discussion between modelers and decision makers of the spatial pattern of the final susceptibility maps in the field might increase their acceptance and applicability.

Many regions worldwide are threatened by multiple natural hazards with the potential to cause high damages and losses. However, the modeling of multiple hazards in a joint analysis scheme is still in the early stages of development as a range of serious challenges emerges in the multihazard context such as differing modeling approaches in use for contrasting hazards; the time- and data-demanding conduct of each single preparative, intermediate and analysis step; and the clear visualization of the modeling outcome. Under consideration of these difficulties, a regional multihazard exposure analysis concept is developed for five natural hazards: debris flows, rock falls, shallow landslides, avalanches, and river floods, complemented by a visualization scheme to present the modeling outcome. An automation of the two schemes resulted in a beta version of the MultiRISK modeling and the MultiRISK visualization software tool forming together the MultiRISK platform. To test the analysis scheme and the software implementation of MultiRISK a case study is performed in the Barcelonnette basin in France with a worst-case parameterization of the models on the basis of extensive literature reviews. Experiences from this case study offered many insights into the multihazard topic and even more questions, e.g. with respect to coherent multihazard model parameterization, validation or the comparability and interpretation of single-hazard modeling results, respectively. Although analysis schemes can be proposed and software tools can be provided to facilitate many steps, a well-conceived and reflective approach to multihazard settings is essential. The worst-case analysis based on literature values apparently leads to an overestimation of the susceptible areas and the number of exposed elements. Nevertheless, depending on the data situation of an area, especially in areas without any information on past events, this approach may offer the determination of general hazard distributions, overlaps, and areas of potential risk without data-demanding calibration.