Landslide of contaminated soil into rivers: Environmental impacts and risks

Sammanfattning: Landslide risk assessments and environmental risk assessments are routinely carried out in many countries exposed to landslide hazards and/or contaminated soil hazards, but separately from each other. By combining results from different types of risk assessments and superimposing the results, possible multi-hazards may be revealed, which is not the case when the results are only displayed independently. Such a combined assessment was done for one part of the river Göta Älv in Sweden and it was found that several contaminated sites adjacent to the river were also exposed to a landslide risk. A qualitative description and a conceptualisation of the problem domain were then carried out to identify the system under study and the governing factors. The unique observation of a landslide-generated increase in turbidity from a minor landslide in Göta Älv allowed for the testing of analytical solutions to the advection-dispersion equation (ADE) for describing the transport of suspended sediment and associated contaminants from a landslide. Although analytical solutions simplifies a complex system, it facilitates quick estimation of suspended sediment and contaminant concentrations as well as the arrival time of the peak and the decrease in maximum concentration from the landslide location. Such solutions to ADE are simple enough so that a large number of scenarios can be evaluated and it provides quantitative outputs that may be easy to include in socio-economic models. Data from Göta Älv was further analysed with respect to turbidity to be able to distinguish landslide-generated suspended sediment peaks from other factors. The river flow is governed by the regulated outflow from a large lake and further controlled at several hydropower stations, making the response of the turbidity to river flow and precipitation slow and complex. A step-wise statistical analysis proved to be a good approach to analyse variation in turbidity with flow and precipitation in such a system. In addition, ship traffic causes short turbidity peaks due to wave-induced bed and bank erosion. A study was therefore conducted to analyse the impact from ship waves on turbidity. Ship-induced bed and bank erosion is also one of the triggering mechanisms behind landslides in the area. Finally, a methodology for risk estimation was developed and tested in the Göta Älv river valley. The approach is probabilistic and allows for the analysis of datasets with large uncertainties and the use of expert judgements, providing quantitative estimates of probabilities for defined failures. The approach was illustrated by a case study along Göta Älv, where failures were defined and probabilities for those failures were estimated. The approach provides a quantitative analysis of the risks associated with landslides in contaminated areas; thus, the methodology makes the problem of landslides in contaminated areas visible. It can also act as a basis for communication and discussions between stakeholders, thereby contributing to intersectoral management solutions. The approach is meant to be a complement to existing landslide risk assessment methodologies as well as to environmental risk assessment methodologies.