Testing models of step formation against observations of channel steps in a steep mountain stream

Steep streams often feature a step-pool morphology where the steps determine channel stability and dissipate the stream's energy, and thus are important for local flow hydraulics and bedload transport. Steps also play a key-role for the coupling of channels and adjacent hillslopes by controlling hillslope stability. Although step-pool systems have been investigated in various modelling and experimental efforts, the processes of step formation and destruction remain under debate. Theories of step formation consider a wide range of dominant drivers and fall into three groups favouring hydraulic controls (HC), granular interactions during flow (GI) or random drivers (RD) as relevant factors for step location. A direct evaluation of these models with field observations is challenging, as step formation cannot be directly observed. Based on the physical mechanisms of the various formation models we derive diagnostic morphometric parameters and test them with a field data set from a steep stream in Switzerland. Our results suggest that one class of alluvial steps form due to jamming in narrow and narrowing sections of the channel, while steps in wide and widening sections form around rarely mobile keystones. These two models of step formation apply in our study reach at the same time in different locations of the channel. A third class of steps is forced by logs. Such steps are typically located close to the original growth position of the tree and therefore reflect strong channel-hillslope coupling. Wood-forced steps make up a minor fraction of the step population, but contribute significantly to the cumulative step height and are therefore relevant to reach-scale flow resistance of the channel. (c) 2019 John Wiley & Sons, Ltd.