Assessment of reliability of multi-beam echo-sounder bathymetric uncertainty prediction models

Abstract

Nowadays Multi-Beam Echo-Sounder (MBES) systems are used for obtaining information of the sea/river bed bathymetry and sediment composition. For the latter, use is usually made of the backscatter strength and depth derivatives, such as depth residuals. However, the depth derivatives are affected by the uncertainties inherent to the MBES varying with the sensors used, survey configuration and operational environment. Although models are available for the vertical uncertainty prediction, the question is how well these models can capture the estimated uncertainties of real observations. The present contribution addresses this issue by comparing the measured with modelled depth uncertainty accounting for the most recent insights of the error contributors. Data was acquired in water depths of around 2m, 10m and 30m with pulse lengths of 27 μs, 54 μs and 134 μs in the Oosterschelde estuary, the Netherlands, enabling the assessment of depth and pulse length dependence of the uncertainties. In general, the predicted and measured uncertainties are in the same order of magnitude. With increasing depth the discrepancy between the modelled and measured uncertainties increases. The effect of changing pulse length is found to be captured by the model, except for the angles close to nadir. The most dominant contributors to the vertical uncertainty are those induced by the angle of impact and range measurements. These contributors thus require further investigation to obtain a more realistic estimate of the vertical uncertainties.