Detection of surface change is a fundamental task in geomorphology. Terrestrial laser scanners are increasingly used for monitoring surface change resulting from a variety of geomorphic processes, as they allow the rapid generation of high-resolution digital elevation models. Irrespective of instrument specifics, survey design or data processing, such data are subject to a finite level of ambiguity in position measurement, a consideration of which must be taken into account when deriving change. The propagation of errors is crucial in change detection because even very small uncertainties in elevation can produce large uncertainties in volume when extrapolated over an area of interest. In this study we propose a methodology to detect surface change and to quantify the resultant volumetric errors in areas of complex topography such as channels, where data from multiple scan stations must be combined. We find that a commonly proposed source of error – laser point elongation at low incidence angles – has a negligible effect on the quality of the final registered point cloud. Instead, ambiguities in elevation inherent to registered datasets have a strong effect on our ability to detect and measure surface change. Similarly, we find that changes in surface roughness between surveys also reduce our ability to detect change. Explicit consideration of these ambiguities, when propagated through to volume calculations, allows us to detect volume change of 87 ±5m3, over an area of ∼ 4900m2, due to passage of a debris flow down a 300m reach of the Illgraben channel in Switzerland.