Abstract. On 60 m² hillslope plots, at 18 mainly grassland locations in Switzerland rain was applied at rates of 50–100 mm/h for between 3 and 6 h. The generated flows were measured, including overland flow, near surface and subsurface flow 0.5–1.3 m below the surface. At some locations less than 2% of the rain flowed down the slope either on or below the surface, whereas at some others more than 90% of the rain ran off. At the majority of sites most runoff was overland flow, though at a few sites subsurface flow, usually via macropores was dominant. Data collected during each of 48 high intensity sprinkling experiments were used to distinguish, which processes were dominant in each experiment. Which dominant and subsidiary processes occurred depended on interactions between infiltration rate, change in soil water storage and drainage of the soil water. These attributes were often not directly linked to parameters usually considered important like vegetation, slope, soil clay content and antecedent soil moisture. Considering the structure of the soil in combination with these attributes, process determination was in many cases fairly straightforward, indicating the possibility of reliably predicting runoff processes at a site. However, at some sites, effects occurred that were not easily recognizable and led to surprising results.