Compound flooding (CF) poses significant risks to human and natural systems. The urgent need for an improved predictive understanding of CF, and its impact on sediment dynamics and riverine saltwater intrusion, at global scales under a future climate is well recognized. The current version of the Department of Energy’s (DOE’s) Energy Exascale Earth System Model (E3SM) addresses these needs, however, several shortcomings including those related to model structure, parameterization, and coarse spatial resolution, pose limits. Efforts towards improved predictions using E3SM are also restricted by the lack of portability of the current E3SM’s river model on DOE’s exascale-class supercomputers, which have heterogeneous computing architectures.

To overcome existing shortcomings of E3SM for the study of CF and its impacts on sediment dynamics and riverine saltwater intrusion in changing climate, a rigorously verified and validated river dynamical core (RDycore), which includes efficient and scalable solvers for DOE’s exascale-class supercomputers, is being developed. The initial development uses two-dimensional shallow water equations coupled with advection-diffusion transport equations to provide a mechanistic model for the transport of water and sediments across the land—river—ocean interface. RDycore uses two numerical libraries, PETSc and libCEED, to provide scalable solvers for heterogeneous computing architectures. The JIGSAW meshing library is used for generating variable resolution meshes for RDycore. In this work, we present the current development status of RDycore and its application to study flooding during Hurricane Harvey. We also report the computational performance of the model on DOE’s supercomputers at the National Energy Research Scientific Computing Center and the Oak Ridge Leadership Computing Facility.