Comparasion of finite difference and finite element hydrodynamic models applied to the Laguna Madre Estuary, Texas

Abstract

The U.S. Geological Survey Surface Water Flow and Transport Model in Two-Dimensions (SV*9FT2D) model was applied to the northern half of the Laguna Madre Estuary. SW=D is a two dimensional hydrodynamic and transport model for well-mixed estuaries, coastal embayments, harbors, lakes, rivers, and inland waterways. The model numerically solves finite difference forms of the vertically integrated equations of mass and momentum conservation in conjunction with transport equations for heat, salt, and constituent fluxes. The finite difference scheme in SW=D is based on a spatial discretization of the water body as a grid of equal sized, square cers. The model includes the effects of wetting and drying, wind, inflows and return flows, flow barriers, and hydraulic structures. The results of the SW=D model were compared to results from an application of the TXBLEND model by Texas Water Development Board to the same part of the estuary. TXBLEND is a two-dimensional hydrodynamic model based on the finite element method. 'Me model employs triangular elements with linear basis functions and solves the generalized wave continuity formulation of the shallow water equations. TXBLEND is an expanded version of the BLEND model to additional features that include the coupling of the density and momentum equations, the inclusion of evaporation and direct precipitation, and the addition tributary inflows. The TXBLEND model simulations discussed in this study were performed by personnel at the TWDB. The two models were calibrated to a June 1991 data set from a TWDB intensive inflow survey of the Laguna Madre. Velocity and water quality data were available for the three days of the survey. Tide data for a much longer period were available from TCOON network stations. Results of the two models were compared at seven tide stations, eight velocity stations, and eleven flow cross sections. Simulated water surface elevations, velocities, and circulation patterns were comparable between models. The models were also compared on the basis of the ease of application and the computational efficiencies of the two models. The results indicate that, in the case of the Laguna Madre Estuary, TXBLEND is the more efficient of the two models.