Exploiting GPS and RFID field data to validate numerical modelling of large wood transport in the Tagliamento River (Italy)

Abstract This study evaluates the reliability of the ORSA2D_WT model, a Eulerian–Lagrangian model, in simulating large wood (LW) transport in the Tagliamento River. The model implements a literature strategy to account for sliding and rolling entrainment modes, besides floating. Overall, the model demonstrated an acceptable level of accuracy in replicating LW entrainment with a successful prediction of the behaviour of 8 out of 11 entrained logs, and 29 out of 37 stable logs were observed during field surveys. The findings are based on a limited number of comparisons, including 36 logs in the Cornino reach and 12 in the Flagogna reach, with only 2 GPS trajectories available, emphasising the exploratory nature of the study and the preliminary validation of the model. While the model effectively predicted LW dynamics under simplified conditions, discrepancies in trajectories near islands and areas of complex flow dynamics highlighted challenges in capturing intricate LW transport. Sensitivity analysis revealed the significant influence of wood density on LW transport, with wet density (WD) conditions showing notable deviations from the observed data. These findings emphasise the complex interplay among density, buoyancy, and hydrodynamic forces, underscoring the need for precise density estimates in LW transport modelling. Additionally, the initial orientation of logs was found to significantly affect transport dynamics, with logs aligned parallel to the flow experiencing longer displacements, while perpendicular or oblique orientations increased hydrodynamic forces, anticipating entrainment, and also fostered early deposition because of the higher interaction with the riverbanks. The model displayed an overestimation of LW mobility, compared to field surveys observations. This limitation highlights the need for a more realistic representation of log interactions, partial burial and structural features such as root wads and branches. To enhance the model's accuracy and reliability, future improvements should focus on better representing wood accumulations and partial burial, as well as optimising computational efficiency. These advancements will enable more comprehensive analyses and improve the model's applicability and robustness for real‐world scenarios.