Methodological sensitivity of morphometric estimates of coarse fluvial sediment transport

The estimation of fluvial sediment transport rate from measurements of morphological change has received growing recent interest. The revival of the 'morphological method' reflects continuing concern over traditional methods of rate determination but also the availability of new survey methods capable of high-precision, high-resolution topographic monitoring. Remote sensing of river channels through aerial digital photogrammetry is a potentially attractive alternative to labour intensive ground surveys. However, while photogrammetry presents the opportunity to acquire survey data over large areas, data precision and accuracy, particularly in the vertical dimension are lower than in traditional ground survey methods. This paper presents results of recent research in which digital elevation models (DEMs) have been developed for a reach of a large braided gravel-bed river in Scotland using both digital photogrammetry and high-resolution RTK GPS ground surveys. A statistical level of change detection is assessed by comparing surfaces with independent check points. The methodological sensitivity of the annual channel sediment budget (1999-2000) to the threshold is presented. Results suggest that while the remote survey methods employed here can be used to develop qualitatively convincing, moderate precision DEMs of channel topography (RMSE = +/-0.21 m), the remaining errors imply significant limits on reliable change detection which lead to important information losses. Tests at a 95% confidence interval for change detection show that over 60% of channel deposition and 40% of erosion may be obscured by the lower level of precision associated with photogrammetric monitoring when compared to ground survey measurements. This bias reflects the difficulty of detecting the topographic signature of widespread, but shallow deposition on bar tops. (C) 2003 Elsevier Science B.V. All rights reserved.