Simulating Flow Dynamics in Shallow Lowland Rivers: A Computational Approach to Indicate the Effect of Natural Obstacles on Flow Patterns

Computational analysis of river flow remains a cornerstone of mathematical hydrological research, particularly in ecologically sensitive and morphologically complex environments. This study introduces an integrated workflow that combines aerial image‐based obstacle detection in the riverbed with two‐dimensional (2D) depth‐averaged simulations to assess meso‐scale flow structures in morphologically complex lowland rivers. It presents a novel framework for modelling shallow rivers that is able to capture meso‐ to micro‐scale flow dynamics influenced by instream boulders, spatially heterogenous vegetation types, and bottom roughness. In situ measurements from four lowland river reaches in Lithuania were used to parameterize and calibrate flow resistance based on two boulder types and three distinct vegetation cover classes. A finite element (FE) approach was applied to solve the shallow water equations (SWE) with site‐specific roughness coefficients. The model simulations were validated against observed velocity data, with relative errors of less than 20% at most observation points. The proposed approach demonstrated strong potential for accurate reproduction of fine‐scale hydrodynamic behaviour in shallow lowland rivers, offering a flexible environment and scalable platform for future integration with automated AI‐based remote sensing data. The precision and adaptability of the model make it a valuable tool for ecological assessments, flood risk studies, and restoration planning.