The Effect of River Training Since the Year 1800 Along an Alpine River by the Help of Two-dimensional Hydrodynamic Modeling

More and more efforts for floodplain restoration both at large and small rivers have been made during recent years for ecological or flood mitigation purposes. Whereas the effects of single measures on floodwave propagation and retention have been well evaluated at a small scale, studies for long river sections are still missing. Generally, one would expect significant damping and deceleration effects on flood waves in completely natural riverscapes compared to nowadays anthropogenically modified river courses.

With this case study of an alpine river, we want to contribute to the assessment of retention potential of large natural rivers. The Inn River in southeastern Germany flowing from the Alps northwards has been modified and channelized during the last two centuries. In the selected 50 kilometers for the case study, the river, which had originally braided gravel beds, islands and wide floodplains, was transformed into a channel almost completely impounded by run-of-the-river hydropower plants. The width of the inundated area for retention was reduced from several kilometers to hardly 200 to 800 meters. To conduct the study, two-dimensional hydrodynamic models of these two different states were set up: one representing today’s river and one a historic state from around the year 1800. For today’s model, airborne laser-scanning data and cross-section measurements were used; it was calibrated and validated against observed water levels and discharge hydrographs of various flood events. Based on the terrain of the first model and historical maps drawn in the first half of the 19th century, the position of the river channel without any man-made alterations was reconstructed for the second model. Available data found in archives was used to estimate the bathymetry.

Flood events typical for this alpine region and with different magnitudes ranging from 10 to 100 years return periods were simulated with both models to show the effect of river training on flood discharges. The results confirm the findings of other studies: the potential of natural flood retention decreases with the increasing magnitude of flood events. Flood peak discharge was reduced in the historic state by not more than 10 % compared to the actual state at minor flood events. Flood wave deceleration was significant in all scenarios but has to be furthermore evaluated with regard to the superposition of flood waves from tributaries.