Comparison of cross-sectional to continuous measures of channel dimensions in Mid-Atlantic Piedmont streams

Abstract

Measurement of stream channel dimensions provides valuable information toward understanding geomorphic processes. A recently developed geomorphon-based stream mapping algorithm enables fast, continuous generation of channel dimension estimates anywhere a high-resolution digital elevation surface is available. This study compared automated, continuous estimates of channel width and bank height to cross-sectional measurements placed manually on the same digital surface and field transects. Both types of digital measurements performed best on channels that were of sufficient size to be visible on a 1 x 1 m digital elevation model (DEM), had banks that appeared as distinct features, and could not be mistaken for valley walls. Geomorphon-based measurement required refinement for very small and very large channel reaches. For channels of moderate size (approximately 2-30 meters wide), especially those wider than about four meters, 25ᵗʰ percentile geomorphon-based bank height estimates corresponded reasonably well with cross-sectional bank height estimates, and median geomorphon-based channel width correlated with, but underestimated, cross-sectional width measurements. Additionally, the range of width or height values detected by geomorphon-based estimates tended to encapsulate mean cross-sectional estimates of each reach. The interquartile range (IQR) of geomorphon-based estimates of channel width included mean cross-sectional field top widths for 29% of reaches between two and eight meters wide. The IQR of geomorphon-based estimates of bank height included mean cross-sectional field bank heights for 22% of channels wider than two meters. Despite necessary areas for improvement, automated, digital techniques show promise for describing channel dimensions in streams of moderate size.