Steven E. Yochum

Spatial characterization of roughness elements in high-gradient channels of the Fraser Experimental Forest, Colorado, USA

We collected high-resolution LiDAR-based spatial and reach-average flow resistance data at a range of flows in headwater stream channels of the Fraser Experimental Forest, Colorado, USA. Using these data, we implemented a random field approach for assessing the variability of detrended bed elevations and flow depths for both the entire channel width and the thalweg-centered 50% of the channel width (to exclude bank effects). The spatial characteristics of these channels, due to bedforms, large clasts and instream wood, were compared with Darcy-Weisbach f and stream type through the use of the first four probability density function moments (mean, variance, skewness, kurtosis). The standard deviation of the bed elevations (σz) combined with depth (h), as relative bedform submergence (h/σz), was well correlated with f (R2 = 0.81) for the 50% of channel width. The explained variance decreased substantially (R2 = 0.69) when accounting for the entire width, indicating lesser contribution of channel edges to flow resistance. The flow depth skew also explained a substantial amount of the variance in f (R2 = 0.78). A spectrum of channel types is evident in depth plots of skew versus kurtosis, with channel types ranging from plane bed, transitional, step pool/cascade, to cascade. These results varied when bank effects were included or excluded, although definitive patterns were observed for both analyses. Random field analyses may be valuable for developing tools for predicting flow resistance, as well as for quantifying the spectrum of morphologic change in high-gradient channel types, from plane bed through cascade.

Longitudinal variability of geomorphic response to floods: Geomorphic response to floods

Morphodynamic response of channels and floodplains to flooding reflects interactions of erosive and resistive forces with sediment transport capacity and supply at multiple scales. Monotonic relationships between reach-scale response to floods with independent variables such as flood stream power and channel confinement can be confounded by longitudinal variations in these variables at longer scales. In these cases, channel response depends on both local and upstream drivers. Using high resolution preand post-flood digital elevation models, we calculate reach-scale (0.5 to 1 km) and segment scale (10 km) longitudinal variations in channel widening and sediment balance. We relate these responses to longitudinal variations of unit stream power and channel confinement for selected streams impacted by the 2013 Colorado Front Range regional flood event. These streams transition from steep and relatively confined in the canyons of the foothills to less steep and unconfined on the high plains. The channel widening response is more closely linked with reach scale gradients in unit stream power: abrupt widening typically occurred within reaches where a large drop in unit stream power occurred relative to upstream. Sediment balance followed segment scale trends in unit stream power, exhibiting a net erosional trend within the foothills that switches to a net depositional trend within the transition to the plains. These findings indicate that unit stream power gradients mediate channel response at reach to segment scales. Predictive modeling of stream response to floods and fluvial hazards assessments that only consider absolute values of reach-scale stream power may under-estimate fluvial hazards in some settings by ignoring unit stream power gradients.