n situ three-dimensional (3D) fish school descriptors were compared using two multibeam sonars, a Simrad SM20 200-kHz and a Reson SeaBat 6012 455-kHz, deployed together in a lake to observe same fish schools simultaneously. School-specific 3D parameters observed by SeaBat and SM20 were extracted using the SBIViewer software for the Reson, and a customized IRD Matlab algorithm (IMA) for the SM20. This study shows that for most of the school parameters considered (school length, width, height, surface, volume, mean energy, energy standard deviation), the estimates from the SeaBat and SM20 sonars converged. The divergences observed for three other school parameters (school roughness, number of holes, and hole volume) may have been due to differences between the instrumental and/or the extraction methods. The source of each divergence was identified, and its impact discussed.