Ecological studies commonly demand elaborate knowledge of the distribution pattern of humous horizons, especially in forest ecosystems with high humus variability. Conventional investigative methods are timeconsuming and surface-destructive. Ground penetrating radar (GPR), in contrast, offers a fast and non-destructive method of portraying the shallow subsurface. Against this background, the reliability of GPR-derived metric data on terrestrial humous horizons was tested at a study site in the altimontane zone of the Bavarian Alps (1,260 m a.s.l.). The GPR survey, using a shielded bistatic antenna device at 800 MHz, is based on eight parallel lines within a 30 9 30 m plot. The metric interpretation of the GPR results is validated by a trench, which records the thickness of humous horizons at a 10-cm interval, and is compared to conventional sampling methods, represented by 18 soil monolith profiles of both random and grid-based distribution. The results show that it is not possible to resolve the borders of single humous horizons. The delineation of humous against non-humous horizons, however, shows correlation (R = 0.553) with the reference data and delivers reliable data on the average thickness and variation of humous horizons. The observed thickness deviations between GPR and the trench line range between -9.5 and 7.5 cm, with remarkably narrow deviations of -2 to 2 cm or less in 50% of the cases. In summary, supported by the high number of data records (n = 1,727), GPR creates a much better and none destructive image of the actual humus distribution pattern compared to conventional methods.