Identification of synoptic precursors to extreme precipitation events in the Swiss Alps by the analysis of backward trajectories

Abstract:

One of the most expensive natural disasters in Switzerland consists in floods related to heavy precipitation. The occurrence of heavy rainfall may induce landslides and debris flows, as observed during three major precipitation events that occurred recently in the Swiss Alps (August 1987, September 1993 and October 2000). Even though all these events took place under a southerly circulation, especially in autumn, not all southerly circulations lead to heavy precipitation. Although many studies tried to understand them, they are still difficult to forecast. Therefore, this work aims to identify synoptic precursors to such events throughout backward trajectories analysis.

Part one tests as many combinations of tools, datasets and methods as possible in order to compare the trajectories in the case of heavy precipitations in the Alps and to reduce the number of models to be assessed for the second part by selecting the most relevant. As a result, we removed models yielding to similar results by using an absolute horizontal transport deviation measure (ATEH). The trajectories were processed with tools (HYSPLIT, a Matlab script developed at the University of Lausanne, and METEX) based on different Reanalysis (NCEP/NCAR, ECMWF, Japanese and NASA). Moreover, different types of trajectories (3D, isobaric, isentropic, isosigma, and constant density) have been used. As a result, 21 trajectory models were compared, and 9 were selected. Results show that most of the differences between trajectories are mainly related to the dataset rather than to the model.

In part two, the 9 selected models were used to search precursors leading to heavy precipitations. 10-days backward trajectories were processed for the Binn station at different pressure levels, for all the days between 1961 and 2014 characterized by a southerly circulation in autumn. Based on these trajectories, two analysis for the identification of precursors were conducted. First, the ATEH was used to assess similarities between extreme events and non-extreme events. Results show that trajectories leading to heavy precipitation tend to be more similar than trajectories leading to non-rainy days. Finally, further analysis showed that trajectories run at 700 hPa for a short duration (up to 3 days) are the most relevant for precipitation analysis.