An improved pressure–velocity correction-based 2D flood simulation model on non-staggered grid arrangement

ABSTRACT Real-time flood forecasting requires computationally efficient models, and implicit flood models are often preferred due to their ability to use larger time steps, enhancing computational efficiency. However, these models may suffer from spurious checkerboard oscillations, causing unrealistic predictions of water depth and velocity. While some models employ the Rhie and Chow interpolation technique to mitigate these oscillations, this approach depends on the under-relaxation factor, affecting convergence and computation time. This study introduces IITMFLO-2D, an implicit 2D flood model developed using the SIMPLE-Consistent (SIMPLEC) algorithm on a non-staggered grid system, with two novel modifications to reduce checkerboard oscillations. The first modification uses an explicit equation, similar to Roe-type averaging, for the momentum weighting factor based on local flow depths. The second introduces a Courant–Friedrichs–Lewy (CFL)-based weighting factor to determine interfacial water levels, adapting to local flow dynamics. These adaptive weighting factors vary spatially and temporally. Performance of IITMFLO-2D is verified through analytical test cases, hypothetical flood simulations, and experimental events, showing significant reductions in computation time and improved accuracy compared to other models such as HECRAS-2D and TELEMAC-2D. These innovations make IITMFLO-2D suitable for real-time flood forecasting and applicable to other numerical models with complex meshes.