Guest blog by Benjamin Poschlod

Southern Norway is regularly exposed to floods triggered by heavy precipitation and snow-melt. A combination of high temperature leading to snow-melt and heavy rainfall is an example of a compound event, which leads to severe flooding when run-off due to snow-melt adds to river discharge from the rainfall. These compound events typically occur in southern Norway during spring, when the snow masses accumulated during the winter have not melted yet, but temperature starts to rise. This combination of rainfall- and snowmelt-induced flood has resulted in the largest floods in southern Norway, for instance in south-eastern Norway, in 1995 and 2013. The co-occurrence of heavy rainfall and snow-melt can also lead to several types of mass movements such as landslides, debris flow or slush flow.

In southern Norway, flood hazards and mass movements often severely impact infrastructure, economy, personal property, and may even cause fatalities. Therefore, it is of great relevance to investigate the effects of a changing climate on the frequency of these compound events.

Short-term scientific mission (STSM)

In preparation of the cooperation between the University of Munich (LMU) (Germany) and CICERO (Norway) focusing on climate extremes and compound events, an STSM conducted in March 2019 within the COST action DAMOCLES. Benjamin Poschlod (LMU) collaborated with Jana Sillmann (CICERO) to explore the compound event of heavy rainfall and snow-melt.

The effect of climate change on the frequency of this compound event was investigated by comparing the frequency of occurrences during the period 1980 – 2009 to that in the far future in 2070 – 2099, using climate model data from the ClimEx project. Normally, the statistical analysis of compound events is limited by a small sample size. In our case we search for heavy rainfall and big amounts of snow-melt occurring simultaneously, which may happen quite rarely within limited time periods. The special feature of the ClimEx dataset is that the Canadian Earth System Model CanESM2 was run 50 times with slightly different initial conditions for the period 1950 – 2099 under the emission scenario RCP 8.5. These 50 climate simulations were dynamically downscaled with the Canadian Regional Climate Model version 5 (CRCM5) to a spatial grid of 0.11 degrees (~ 12km) and hourly resolution. Hence, this database consists of 50 x 30 years of data for both periods (1980 – 2009 and 2070 – 2099), which is expected to provide enough compound events and therefore contributes to the statistical robustness of the results. 

For 1980 – 2009, the average return period of a precipitation-snow melt compound event for southern Norway is 7.4 years, but the spatial distribution is very inhomogeneous.

The highest number of events can be found near the west coast. The windward sides of the mountains as well as the area around the Oslo fjord also show a higher occurrence rate. Due to climate change, the amount and frequency of heavy precipitation will increase. Rising temperatures on the other hand will result in less snowfall and an earlier onset of snow-melt. In the ClimEx dataset, this results in a 47% decrease in frequency of the investigated compound event, leading to a return period of 14 years in the future. The spatial pattern is changing as well by 2070 – 2099. Due to reduced snow cover, hardly any events will occur in the southern parts of the west coast. However, the northern parts of the west coasts, the windward sides of the mountains and the area around the Oslo fjord will show an increased number of events.

We have seen, that a changing climate significantly affects the occurrence of compound events. While the investigated event decreases in frequency, other events, such as heavy precipitation on saturated soils are very likely to occur more frequently in the future. This research topic is scientifically challenging and of great relevance for civil security. We are looking forward to contributing to this research field through further cooperation between the LMU and CICERO.