Mountainous regions are highly sensitive to climatic changes. The warming trend of the last decades and the related retreat of glaciers and permafrost consequently imply strong changes regarding sediment transport patterns due to intensive melting and the (re-)activation of potential sediment sources. The present study aims to present a holistic approach to analyze sediment cascades in this climatic and geomorphological context.

The study area, the partly glaciated Solda / Sulden catchment in the area of Val Venosta / Vinschgau (Italian Alps), was mapped geomorphologically to define the spatial distribution of source, transport and deposition areas. Subsequently, potential sediment pathways between topographically connected geomorphological landforms form the base to analyse the network of transport processes within the catchment in detail. The identification of spatial and temporal hot-spots as well as triggering factors of these sediment transport processes will be obtained by combining the results of the network analysis with results of DEMs of Difference (DoDs) representing different time intervals (months/years). Furthermore, measurements of fluvial sediment transport play an important role to qualitatively validate the findings. The monitoring combines regularly conducted direct measurements of both bedload and suspended sediments and continuous indirect measurements (e.g. geophones, turbidimeters). Sampling points are located at glacier outlets, in the proglacial river and at the outlet of the catchment.

Finally, based on the overall result of the study, the significance of extreme meteorological events, like heavy rain storms or heat waves, for the intensity and spatial distribution of sediment transport will be estimated to outline potential risks for infrastructure or population.

Processes like fluvial flows (floods with or without intensive bedload transport) or debris flows represent extreme events in torrential catchments that can transfer large amounts of sediment to the receiving river or alluvial fan. In contrast to fluvial flows, the formation of debris flows is the result of a critical combination of water, sufficient sediment that can be mobilized, and topographic slope. In alpine regions, excessive water input typically results from long lasting rainfall (LLR), short duration storms (SDS) or intense snow melt (SM). In a recent study, we derived multiple hydro-meteorological variables from long-term hydrological simulations to investigate the trigger conditions of fluvial flows and debris flows in six regions in Austria. We found distinct regional and seasonal differences, but overall roughly 2/3 of the debris flows were initiated by SDS events and 1/4 by LLR events, the remaining by SM. Fluvial flows tend to results from a higher total precipitation input than debris flows. When ex-post predicting the susceptibility for debris flow initiation with a Naïve Bayes Classifier model, we found that multi-variable trigger models outperformed simple intensity-duration thresholds. The remaining uncertainty we largely attribute to the unknown sediment dynamics in the watersheds that were not included in the susceptibility assessment. We conclude that the consideration of different hydro-meteorological trigger types improves engineering risk assessment.

Prediction of a watersheds temporal susceptibility to debris flows using multiple hydro-meteorological variables

Prenner D., Kaitna R., Mostbauer K., Hrachowitz M.

Debris flows represent a threat for societies in alpine regions and are typically triggered by excessive water input from long lasting rainfall (LLR), short duration storms (SDS) or intense snow melt (SM) into torrential watersheds. The prediction of debris flow events mostly relies on rainfall intensity and duration (I-D) alone, which is often less reliable for practical applications because of the high spatial variability of precipitation. To overcome this limitation, we utilize multiple hydro-meteorological variables like snow melt, evapotranspiration, soil moisture from a hydrological simulation besides station data of precipitation and temperature to predict the temporal susceptibility of the Montafon watershed to debris flows between 1953 and 2013. Therefore, we setup four Naive Bayes Classifier models of different complexity, ranging from simple rainfall only to multi-variable, multi-trigger type as well as a classical I-D curve and evaluate the performance using Receiver Operating Statistics. Results show that the watershed is in very different states in dependence of the trigger of either LLR, SDS or SM on the 38 documented debris flow event days in the region. The multi trigger-type models outperform the simpler models as well as I-D curve by showing both, higher true positive rates and lower false alarm rates. We conclude that, the consideration of hydro-meteorological variables can help to improve debris flow prediction in future.

Alpine regions are exposed to different mass wasting processes, including debris flows, landslides, and rock fall. Debris flows are highly mobile gravity driven mixtures of sediment and water, that can exhibit different flow states, and can alter the flow behavior during a single event. The combination of high velocities and the capacity to carry large boulders endangers human lives and infrastructures. Small-scale models are often limited to reproduce the complex dynamics. For the design of mitigation measures and for back-calculations, realistic values of the range of impact pressures are required, but rarely available. In this contribution, we present first results of in-situ measurements of the natural debris-flow impact pressures onto a debris flow breaker at the Gadria torrent, Italy. The measured forces support the notion to differentiate between dynamic bulk impact and impacts by single boulders. The data also highlight the problem at low Froude numbers to define the empirical coefficient in the hydrodynamic models currently used. Our study contributes to data densification of real-scale debris flow impact forces and aims for a better understanding of the complexity of the interactions between the flow process and engineering structures.

Debris flows are dangerous natural hazards that every year cause fatalities and damages to infrastructures. One of the challenges of the last decades is to predict the dynamics of such events using specific numerical models. The aim of this study is to satisfactorily reproduce the progressive erosion dynamic of a debris flow starting from a channelized water flux. This kind of process was reproduced exploiting the functionality of the r.avaflow model that is capable to simulate a two-phase mixture of gravity mass flows.

The investigated study case occurred near the village of Cortina D’Ampezzo (Veneto Region, Italy) during the 2017 summer. The debris flow was triggered by an extreme rainfall event that caused an intense runoff, entraining sediment materials from the channel bed and then generating a mature granular debris flow. To reproduce the erosion pattern, the modelling approach developed in r.avaflow was improved. Specifically, the input erosion coefficient was varied according to a calibrated exponential function of the local slope. Another simulation was performed using the best constant erosion coefficient to assess the performance of the improved model.

The results of these two simulations were then compared with the observed (LiDAR surveys) erosion pattern. In particular, using a slope dependent erosion coefficient, the entrained volumes and the erosion trend resulted in a more reliable topography than those produced by a constant erosion coefficient. The same good performance was obtained in terms of bulked peak discharge flowing 1.4 km downstream of the input water hydrograph.

In conclusion, the research has highlighted the satisfactorily simulation of debris flow triggering for r.avaflow model and provided a supplementary methodology of model implementation for a more accurate prediction of debris flow erosion.

Understanding morphology changes and sediment spreading along a debris-flow channel is a key step in hazard mitigation planning.
This research analyses a 10 years evolution of erosion/deposition patterns in an active debris-flow upper channel located on the Dolomites (rio Soial, Val di Fassa, Trento, Italy). The morphologic evolution of the channel has been analysed performing a Difference of DEM (DoD) and comparing the 2008 LiDAR-derived DTM of the Autonomous Province of Trento with a DTM created from a UAV-based point cloud of the July 2018. This data set was also used to determine the changes of the sediment Connectivity Index (CI), which explains the existing degree of linkage between sediment sources and channel network. During the period 2008-2018 five debris flow events have occurred. Each associated rainstorm was analysed in order to assess the evolution of the threshold rain intensities for the triggering in relation to the evolution of the channel-valley morphology.
The results on the CI analysis show a general decrease in CI values, meaning an increased disconnection between the head basin areas and the outlet at the end of the transport reach. Also, the rain thresholds show a slight increase after the lasts event, indicating a gradual stabilization of the basin and a possible reduction of the expected frequency of debris flow events.