Marius Funk

Marius Funk

Auto-generated profile, based on publicly available information

Marius Funk

University of Augsburg, Augsburg, Germany

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 November 2019 in Water Resources Research
Verified
Journal Article
Open access
PDF file
CAC. Ancey
C. Ancey
EBE. Bardou
E. Bardou
MFM. Funk
M. Funk
MHMatthias Huss
Matthias Huss
MWM. A. Werder
M. A. Werder
TTT. Trewhela
T. Trewhela
In the spring of 1818, ice avalanches from the Giétro Glacier created an ice dam, which in turn formed a glacial lake in the Drance Valley (Canton of Valais, Switzerland). Today, its maximum volume is estimated to have been 25×106 m3. Cantonal authorities commissioned an engineer named Ignaz Venetz to mitigate the risk of the ice dam's failure. He supervised the construction of a tunnel through which a large volume of water was drained as the lake rose (9×106 m3 according to his estimates and 11×106 m3 according to our model). After 2.5 days of slow drainage, the ice dam failed on 16 June 1818 and caused major flooding in the Drance Valley up to 40 km downstream, resulting in about 40 deaths. Venetz's lake monitoring notes, numerous testimonies gathered in the disaster's aftermath, and our field survey have made it possible to collect a wealth of information on this event, which is one of the world's major documented glacial lake outburst floods. Reconstructing major outburst floods remains challenging because not only do they involve enormous volumes of water spreading over long distances but they are also associated with additional physical processes such as massive erosion; intense transport of ice, sediment, and debris; and damage to vegetation and buildings. This paper attempts to reconstruct the 1818 Giétro flood by focusing on its water component. We develop a simple model to estimate the initial hydrograph during the slow drainage and failure phases. The flood's features are deduced by solving the shallow‐water equations numerically. The computational framework involves six free parameters, of which five are constrained by physical considerations. Using iterative manual parameter adjustments, we matched the numerical simulations to the historical data. We found that the peak discharge was close to 14,500 m3/s, the flood's front velocity was about 6 m/s, and flow depth varied considerably along the River Drance's bed (from 30 m just downstream of the ice dam to 2 m on the alluvial fan, 24 km west of the dam). To achieve a good agreement between computations and historical data, we had to select a high value for the Manning friction coefficient n (with n as large as 0.08 s/m1/3). As the Drance Valley is narrow, high flow resistance caused the flood's leading edge to behave like a plug, moving at a fairly constant velocity, with little dependence on what happened behind it. This result may explain why a simple flood routing model is able to reproduce the flood's features, because in an Alpine valley, a lateral spreading of the water volume is limited.

4 November 2024
Verified
Conference Paper
Open access
MFMarius Funk
Marius Funk
Shogo OkadaShogo Okada
Shogo Okada
EAElisabeth André
Elisabeth André
Non-verbal behavior is a central challenge in understanding the dynamics of a conversation and the affective states between interlocutors arising from the interaction. Although psychological research has demonstrated that non-verbal behaviors vary across cultures, limited computational analysis has been conducted to clarify these differences and assess their impact on engagement recognition. To gain a greater understanding of engagement and non-verbal behaviors among a wide range of cultures and language spheres, in this study we conduct a multilingual computational analysis of non-verbal features and investigate their role in engagement and engagement prediction. To achieve this goal, we first expanded the NoXi dataset, which contains interaction data from participants living in France, Germany, and the United Kingdom, by collecting session data of dyadic conversations in Japanese and Chinese, resulting in the enhanced dataset NoXi+J. Next, we extracted multimodal non-verbal features, including speech acoustics, facial expressions, backchanneling and gestures, via various pattern recognition techniques and algorithms. Then, we conducted a statistical analysis of listening behaviors and backchannel patterns to identify culturally dependent and independent features in each language and common features among multiple languages. These features were also correlated with the engagement shown by the interlocutors. Finally, we analyzed the influence of cultural differences in the input features of LSTM models trained to predict engagement for five language datasets. A SHAP analysis combined with transfer learning confirmed a considerable correlation between the importance of input features for a language set and the significant cultural characteristics analyzed.

27 October 2025
Verified
Conference Paper
Open access
DDDaksitha Senel Withanage Don
Daksitha Senel Withanage Don
MFMarius Funk
Marius Funk
MBMichal Balazia
Michal Balazia
HQHuajian Qiu
Huajian Qiu
Shogo OkadaShogo Okada
Shogo Okada
PMPhilipp Müller
Philipp Müller


12 October 2025
Verified
Conference Paper
Open access
ISIslam J A M Samiul
Islam J A M Samiul
KZKhalid Zaman
Khalid Zaman
MFMarius Funk
Marius Funk
MUMasashi Unoki
Masashi Unoki
YNYukiko Nakano
Yukiko Nakano
Shogo OkadaShogo Okada
Shogo Okada

5 October 2026
Verified
Conference Paper
Open access
MFMarius Funk
Marius Funk
Shogo OkadaShogo Okada
Shogo Okada
EAElisabeth André
Elisabeth André