Optimizing hydraulic routing at the continental scale: identifying where and when dynamic wave routing is required

Abstract:

One-dimensional open channel hydraulic flow routing is achieved with the Saint-Venant equations (SVE). The most complete form of the SVE equations, the dynamic wave equation, maintains fidelity to the energy terms included in the SVE: inertia, pressure, gravitational and friction forces. However, while capable of adequately resolving transient flow conditions, the dynamic wave equation is computationally expensive and prone to numerical instabilities. For these reasons, various terms of the SVE are frequently assumed to be negligible for many flow conditions. Many hydraulic modeling software programs include dynamic wave algorithms, however, due the computational expense, only simplified versions of the SVE are currently operational in the continental-scale National Water Model. This presentation examines where and when dynamic wave hydraulic routing is required for accurate representation of river flood waves; and aims to determine, a priori, where fidelity to the dynamic wave equation should be maintained for accurate flood wave forecasting at the continental scale.

Previous work has identified a number of dimensionless parameters that adequately encapsulate the relative significance of the inertia, pressure, gravitational, and friction forces on hydraulic waves. These parameters are used here to quantify the conditions in terms of channel properties (slope, width, and channel roughness) and flow regimes (baseflow, bankfull flow, backwater, drawdown, normal) where there is no loss in predictive skill by using simplified forms of the SVE. Several case studies will show the behavior of the various dimensionless parameters under a variety of physical and flow regime conditions. In addition these parameters will then be summarized for all major river basins in the conterminous United States. The results of this analysis can be used by operational continental-scale models, such as the National Water Model, to prioritize computational resources to river reaches that are identified, under certain flow conditions, to require fidelity to the full dynamic wave representation of the SVE.