Geoenvironmental conditioning of landsliding in river valleys of lowland regions and its significance in landslide susceptibility assessment: A case study in the Lower Vistula Valley, Northern Poland
Dariusz Grabowski, Izabela Laskowicz, Anna Małka, Jacek Rubinkiewicz
Abstract
This paper presents an attempt to analyse landslide susceptibility of the Vistula River Valley in lowland Poland within young morainic areas. Although morphological and geological conditions of Northern European Lowland areas are different from those of mountainous and hilly areas, both of which are highly prone to mass movements, the frequency of landslides may also be very high on the slopes of large, entrenched river valleys. The frequency ratio (FR) and predictor rate (PR) methods were chosen to calculate the landslide susceptibility index (LSI) for entire landslides and detachment zones. Quantitative evaluation of the susceptibility model was performed by calculating area under the curve (AUC) for success rate curves (SRCs) and prediction rate curves (PRCs). To assign LSI values to individual classes the equal areas and the natural breaks methods were used. Identification and prioritization of causal factors show that the topographic attributes such as terrain ruggedness index (TRI), topographic wetness index (TWI), topographic position index (TPI) play a particularly important role. Most landslides develop on the slopes with a complex geological layering (i.e. two-layer or multi-layer setting) and those where the permeable layer (sand) separates less permeable layers (glacial till, clay/silt). In addition, landslides activate more frequently on slopes lacking erosional gullies. 81 % of unstable areas are on slopes with a small amount of erosional gullies, and the number of landslides decreases on slopes with a large number of erosional gullies. The selection of the study area turned out to be an important methodological issue; its geomorphological-geological homogeneity and low variability of the landslide types may have contributed to the very high values (AUC > 95) of SRC and PRC curves.