Investigation of Two Methods for Including Precipitation in a 2D Hydrodynamic Model in Coastal and Lake Environments: Direct Rainfall vs. Lateral Discharge

Heavy precipitation combined with surges and tides can contribute to significant flooding in the coastal and lakeside communities of the United States. We investigated and compared two methods for incorporating precipitation into the D-Flow Flexible Mesh (D-Flow FM) model for the Total Water Level (TWL) prediction. The two methods considered in this study included: 1) Direct Rainfall Method (DRM), also known as rain-on-grid, where precipitation rates are applied at mesh nodes. In this case, D-Flow FM performs internal calculations to convert the precipitation rates into discharges at every grid cell and apply them individually as a source term; and 2) Lateral Discharge Method (LDM), where lateral inflows from precipitation and snowmelt are routed through the National Water Model (NWM) channel network. Within the 2D hydrodynamic model, the lateral inflows may be applied as a point source or may be aggregated in a user-defined polygon.

We tested both DRM and LDM using the D-Flow FM model and compared the effectiveness of each method. Model results are presented, discussed, and compared with measurements in terms of total water level. With a focus on how we incorporate precipitation forcing into our coastal models for the TWL prediction, the relative strengths and weaknesses of the DRM and LDM are discussed, and recommendations for additional research are provided.