Assessing UAV survey performance for geomorphological monitoring of mountain rivers

Characterising morphological changes in mountain areas is of fundamental importance for science
and engineering. Intense floods usually involve massive sediment transport, which may significantly
alter basin and river characteristics. Sediment erosion and deposition control the dynamics
of morphological structures such as alternate bars and meanders. By using unmanned aerial vehicles
(UAV), it has been possible to obtain high-precision bed elevation data at the sediment scale.
Our project aims to develop a consistent and optimised methodology for monitoring morphological
changes in an Alpine watershed using an UAV. Since 2017, we have been monitoring the Plat de la
Lé area drained by the River Navisence (Zinal, canton Valais, Switzerland). In mountainous regions,
poor accessibility and light conditions make it difficult to set control points on the ground. We first
analysed the relevance and influence of certain ground control points (GCP) on the the accuracy of
the digital elevation model (DEM) obtained from the UAV’s images. Errors in the GCP localisation
were much larger than the DEM resolution. Not only did the GCP number and flight height affect
these errors, as expected, but their positions and orientations also played a part. We then carried
out an additional monitoring campaign to understand the influence of these parameters on the DEM
accuracy. This campaign was ran on two areas: a steep-slope area with irregular topography and
a low-slope area that comprises the river channel and its floodplain. We built DEMs for each area
considering different GCP numbers (in the 3–18 range with 14 additional checkpoints) and flight
heights (in the 40–140-m range). The present study provides guidelines, including an optimal combination
of parameters that significantly reduce errors in the DEM, and a methodology that can be
used for monitoring Alpine watersheds on a regular basis.