Significant river improvement projects in the early 1900s provided the impetus for development of an array of simplified flow routing methods. These have been termed storage routing models. They are based on the conservation of mass equation (1) written in the following form:
in which AS is the change in storage within the routing reach during a At time increment, I = 0.5[/(r) + I(t + Ai)], and 0 = 0.5[0(i) + 0(t + Ai)]; the storage (S) is assumed to be related to inflow (I) and/or outflow (O), i.e.,
in which K is a storage constant with dimensions of time, and X is a weighting coefficient, 0 < X < 1. Storage routing models are limited to typical flood routing applications where the outflow and water-surface elevation relation is essentially single valued, and the waterways are not mild sloping (S0 > 0.002). Thus, backwater effects from tides, significant tributary inflow, and dams or bridges are not considered by these models, nor are they well-suited for rapidly changing unsteady flows such as dam-break flood waves, reservoir power releases, or hurricane storm surges. Generally, storage routing models have two parameters that can be calibrated to effectively reproduce the flood wave speed and its attenuated peak. The calibration requires that most storage routing model applications be limited to where observed inflow-outflow hydrographs exist. When using the observed hydographs to calibrate the routing coefficients, variations in flood wave shapes within the observed data set are not considered, and only the average wave shape is reflected in the fitted routing coefficients.
Was this article helpful?