300 years of adaptation to unintended and cascading consequences of an early river engineering work

Abstract:

In 1714 AD, the Kander River (Switzerland) was deviated to the Lake Thun with the intention of making use of the retention effect of the lake and reducing the flood peaks for the city of Bern, located downstream. The river deviation got out of control within a few days by unintended river incision and made a spontaneously decided restoration impossible. This pioneer geoengineering measure triggered cascading effects on water resources and flood risk management. The downstream settlements of the old river course lost all their water supplies from one day to the other. The catchment area contributing to the Lake Thun doubled and the lake level raised. At the lake outflow in the city of Thun, all hydropower facilities and bridges collapsed and the castle moat had to be opened in the following years as a second river course to enhance the lake outflow capacity. The downstream reaches of the Aare River lost their sediment supply and changed their morphology. After 1815, the rivers in the Aare River basin were straightened by lateral levees. This led to river incision that is still posing problems for bridges and groundwater extraction facilities today. The problem of the increased flood risk of the city of Thun was finally resolved in 2009 by a relief tunnel that allows preemptive lowering of the lake level before an expected flood event. In a model experiment, I show and quantify the combined effects of all the anthropogenic interventions to the river system. Due to the availability of high-resolution digital elevation models and historic topographic surveys, the digital elevation models could have been reconstructed before any anthropogenic interventions took place. A coupled hydrological-hydraulic model shows how the water was routed through the system in historical times. The comparison between the historic state and the actual state of the system allows disentangling the effects river engineering works to flood hydrology and analyzing the spatiotemporal shifts of flood risk within the river basin including downstream-upstream flood risk transfer. The study shows how coupled component modelling can be used to evaluate changes in floodplain processes and adaptation in coupled human and natural systems.