Rockslides in a Changing Climate: Establishing Relationships Between Meteorological Conditions and Rockslides in Southwestern Norway for the Purposes of Developing a Hazard Forecast System

Abstract

The steep, mountainous terrain of southwestern Norway is prone to a high frequency of rockslides. It is known that many of these rockslides are triggered by meteorological conditions, yet there have been few studies dedicated to quantifying the link between rockslides and the runoff conditions and freeze/thaw processes that trigger failure. With recent climate research indicating that southwestern Norway will experience warmer temperatures and increased precipitation, it has become apparent that a better understanding of this link is required to help prepare for future events.

Rockslides in Norway lead to road closures, property damage and fatalities every year, and one of the biggest challenges for Norwegian authorities is to react to rockslides as they happen and to reopen roads as soon as possible. This is especially true when several rockslides occur on the same day in multiple locations. As a result, authorities wish to implement a hazard mapping system that uses a weather forecast to predict when and where geohazards are likely to occur. To this end, this thesis is aimed at providing a rockslide forecast map that changes every day based on the weather forecast.

By comparing a rockslide database to historic weather records, the work carried out for this thesis has indicated that extreme runoff during winter storms is responsible for triggering the majority of rockslides in the region. Using this knowledge as a basis, two potential hazard mapping systems are proposed, one based on trigger threshold exceedance and the other based on weights-of-evidence susceptibility mapping. Both of these methods operate by mapping areas experiencing extreme runoff conditions. Several runoff parameters were tested for possible inclusion, and it was found that 48-hr antecedent runoff, normalized by mean monthly precipitation had the best correlation with rockslide occurrence. Verification of these methods indicates that both approaches are successful in predicting days with extreme conditions, thereby alerting authorities that a high frequency of rockslides is likely.

Due to the complex nature of rockslide triggering, it is not fully understood how climate change will affect future rockslide activity; however, this thesis attempts to answer these questions and to provide a basis for future studies.