AUTOMATIC SELECTION OF KNICKPOINTS IN LONGITUDINAL RIVER PROFILES USING OPEN-SOURCE R SCRIPT AND A COMBINATION OF NORMALIZED STEEPNESS, SMOOTHED SLOPE, AND SUM OF THE DIFFERENCES

Many studies have focused on fluvial geomorphology as a tool for understanding landscape evolution and the connections between river profiles, tectonics, climate, rock strength, and sediment size. River-profile analysis often requires the manual identification of anomalously steep sections of the stream, commonly referred to as knickpoints or knickzones. There are multiple studies that use the existence of migrating knickpoints as indicators of base-level change due to uplift, but this type of analysis is only valid if the knickpoints are migrating and not a result of lithologic variation or other structural controls. To automate the determination of knickpoint locations, I present an algorithm that uses normalized steepness (commonly abbreviated as ‘ksn’), smoothed slope, and a new function called ‘sum of the differences.’ This function works by finding the vertical change between contiguous elevation values along the profile and calculating the values of summed differences 50 m downstream and summed differences 50 m upstream for each point along the stream. The ratio of downstream differences to upstream differences will be approximately equal to one where the stream is at equilibrium, and will spike at the location of a knickpoint, where the downstream differences are far greater than those upstream. While this test alone will not always identify knickpoints, a combination of normalized steepness, smoothed slope, and sum of the differences identifies knickpoints with an accuracy comparable to manual selection, with the advantage of being faster, repeatable, and objective. Once the knickpoints are selected, they can be plotted on a geologic map to see if they are located at lithologic boundaries. The elevations of multiple knickpoints within a single drainage network can also be plotted to see if they are similar, which is expected if they are transient and migrating through the drainage network.