Numerical modeling of the Columbia River plume dynamics

River plumes are an important feature of the shelf seas: They bring freshwater and nutrients to the self, and can significantly impact coastal circulation, stratification, and upwelling. Modeling of estuary-river plume systems, however, is challenging due to the wide range of spatial and temporal scales and their complex interactions. We simulated the Columbia River plume (USA) and the adjacent US west coast with Thetis. Thetis is a next-generation 3D unstructured-grid circulation model based on finite element discretization (Kärnä et al., Geosci. Model Dev., 2018). Due to the flexible mesh resolution, we can represent the small-scale dynamics in the estuary and the large-scale coastal sea in the same model. The second-order discontinuous Galerkin discretization implies low numerical mixing and therefore better representation of the plume front and eddies. We present validation of the plume model, and comparison against a previous unstructured grid model, SELFE. The comparison shows that frontal features and baroclinic dynamics of the plume are indeed greatly improved; Thetis can accurately reproduce plume front traversals seen in mooring observations. The plume responds rapidly to wind conditions, alternating between a narrow coastal current (southerly winds) and south-west oriented plume (northerly, upwelling-favorable winds). The plume spreading rate and bulge circulation are found to be in good agreement with theoretical estimates. We also assess the vertical structure of the plume and mixing rates from the k-epsilon turbulence closure model. This work lays a foundation for more accurate coastal modeling and a better description of the Columbia River plume dynamics.