Atmospheric fronts play a major role in day-to-day weather and are well known for sharp changes in local weather conditions. Especially in regions of high mountains, such as the European Alps, fronts are typically deformed and retarded when they interact with the orography. As a result of retardation, fronts can become nearly stationary which then often leads to heavy precipitation. Usually, fronts are still drawn subjectively by meteorologists. Automated detection of fronts based on atmospheric conditions, however, is an objective and reproducible method to identify frontal lines. In the framework of the project Non-Hydrostatic Climate Modelling, Part II (NHCM-2) (www.nhcm-2.eu), funded by the Austrian Science Fund (FWF), an algorithm to identify frontal line location, motion and type is developed. For testing purposes, an ERA-Interim driven hindcast with 0.11 (? 12.5 km) grid spacing from the COSMO model in CLimate Mode (CCLM) covering the European continent and conducted by Brandenburg University of Technology Cottbus (BTU), is used. Additionally, a CCLM hindcast on convection permitting scale 0.0275 (? 3 km) is investigated. It covers the Alpine region and is driven by the 0.11 simulation. The aim of this work is to provide the necessary steps for detection of fronts based on atmospheric variables. Different vertical detection levels (pressure levels, terrain- following coordinates, and geometric height) are discussed. Model and resolution inde- pendent implementation is also addressed. The application of the algorithm is demon- strated by analyzing different effects on fronts in the presence of mountains (dissolution, deformation, and lee cyclogenesis) provide a successful application of the algorithm.