Predictive modeling of hourly probabilities for weather-related road accidents

Becker, Nico; Rust, Henning W.; Ulbrich, Uwe

doi:https://doi.org/10.5194/nhess-20-2857-2020

Articles | Volume 20, issue 10

https://doi.org/10.5194/nhess-20-2857-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/nhess-20-2857-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 20, issue 10

Research article

|

29 Oct 2020

Research article |

| 29 Oct 2020

Predictive modeling of hourly probabilities for weather-related road accidents

Nico Becker, Henning W. Rust, and Uwe Ulbrich

Supplement

https://doi.org/10.5194/nhess-20-2857-2020-supplement

Data sets

Projekt RADOLAN – Routineverfahren zur Online-Aneichung der Radarniederschlagsdaten mit Hilfe von automatischen Bodenniederschlagsstationen (Ombrometer) H. Bartels, E. Weigl, T. Reich, P. Lang, A. Wagner, O. Kohler, and N. Gerlach https://www.dwd.de/DE/leistungen/radolan/radolan.html

A novel convective-scale regional reanalyses COSMO-REA2: Improving the representation of precipitation (https://reanalysis.meteo.uni-bonn.de/?Download_Data___COSMO-REA2) S. Wahl, C. Bollmeyer, S. Crewell, C. Figura, P. Friederichs, A. Hense, J. D. Keller, and C. Ohlwein https://doi.org/10.1127/metz/2017/0824

Download

Article (3570 KB)
Full-text XML

Short summary

A set of models is developed to forecast hourly probabilities of weather-related road accidents in Germany at the spatial scale of administrative districts. Model verification shows that using precipitation and temperature data leads to the best accident forecasts. Based on weather forecast data we show that skilful predictions of accident probabilities of up to 21 h ahead are possible. The models can be used to issue impact-based warnings, which are relevant for road users and authorities.