Acoustic Doppler current profilers (ADCP) were deployed to investigate the backscattering (BS) signal in three navigable rivers, in different bedload transport conditions. This study aims to demonstrate that the BS strength, as an additional variable to the apparent bedload velocity, improves the characterization of the bedload transport using ADCPs. The M9 –3 MHz and the vertical beam M9 – 0.5 MHz showed decline of the BS strength as the bedload intensity increased, whereas the RDI –1.2 MHz was relatively insensitive. The correlation between the median grain size and the BS strength for the 0.5 MHz was linear, for the 3 MHz the BS strength was attenuated in the active layer, and for 1.2 MHz, it revealed a parabolic distribution. Moreover, the analyses of the ADCP measured variables, using wavelet transformations and unsupervised machine learning, highlighted the importance of the spatial and temporal variance and transient nature of the bedload transport.