River restoration has become a priority goal in many countries. The interest in rivers as valuable habitats for floral and faunal diversity and as an ecosystem with important functions has increased. For evaluating the structural changes from a river restoration project, quantitative methods are needed to support engineers and resource managers in decision making. The Hydromorphological Index of Diversity (HMID) is a metric using the statistical values of flow depths and flow velocities measured at several points at multiple cross-sections of a river reach. For restoration project planning, however, numerical models often are applied offering a rapid calculation of flow depth and flow velocity for multiple variants of a study reach.This is a case study of a 2-km meandering residual flow stretch in the Sarine River in Switzerland downstream of a dam with a constant discharge of 2.5 m3/s. In this river, flow depths and flow velocities are measured at 27 cross-sections. Further, a numerical model is created to generate flow depths and flow velocities using BASEMENT.The roughness of the numerical model is first estimated based on grain characteristics. The model is then calibrated. Analyses show that HMID changes substantially between the values generated with the physically feasible roughness and the calibrated roughness value. It turns out to be less profound for higher discharges. Analysis of the influence of extreme values then shows a strong dependence of the HMID on them. Therefore, extreme values from numerical models may have significantly lesser weight due to the large sample size compared to field measured data, where only the values of 27 cross-sections are taken into account.