Three‐dimensional water quality modelling with TELEMAC‐3D

Abstract

In multidimensional water quality modelling the spatial and temporal distributions of dissolved substances and suspended particles are simulated in order to account for inhomogeneities in the transport and reactive processes. The exchange processes at the water surface, e.g. photosynthesis and oxygen transfer, and at the water–sediment interface, e.g. erosion, deposition and oxygen transfer, are basically multidimensional. A homogeneous distribution can be expected only if the mixing processes generated by the hydrodynamics occur in a short time period compared with the reaction rates. This article presents sample results of a three-dimensional water quality model that includes the lake hydrodynamics, sedimentation and the proper representation of physical and reactive processes at the bottom and surface boundaries. The goal is to use this insight in order to obtain a better description of water quality with lower-dimensional models using spatial averaging techniques. For this purpose, the program TELEMAC-3D, which is able to calculate three-dimensional free surface flows and transport, was extended to include reactive processes. Ordinary differential equations remain to be solved following the solution of the non-reactive transport equations using operator-splitting techniques. These are in turn solved using a fourth-order Runge–Kutta method. As the reaction rates are in part considerably long compared with the hydrodynamics, a larger time-step is used for solving the transport equations. The model was applied to a lake in Berlin, Germany, the Großer Müggelsee. Although comparatively shallow, it experiences extensive multidimensional hydrodynamics as a result of wind-induced currents and the Spree River, which flows through it. Numerous water quality measurements undertaken by the Institüt für Gewässerökologie und Binnenfischerei (IGB) demonstrate the influence of the hydrodynamics on the concentration distributions. A comparison between calculations with the spatial average of the 3-D model and directly calculated 0-D model results using the same reactive–kinetic formulations clearly demonstrates that the multidimensional effects, including the current distribution, have a significant effect on the water quality. Therefore, care must be taken to correctly include multidimensional effects for detailed assessment of possible remedial measures to improve the water quality. An averaging technique for including these effects is presented. Copyright © 2000 John Wiley & Sons, Ltd.