One-dimensional channel network modelling and simulation of flow conditions during the 2008 ice breakup in the Mackenzie Delta, Canada

摘要

This paper presents recent developments to River1D's (the University of Alberta's public-domain hydrodynamic and river ice process model) channel network modelling capabilities. While previous versions of the model assumed equal water surface elevations through the junction, the approach presented in this paper takes into account the significant physical effects at channel junctions (such as gravity and flow separation forces, and channel resistance). The adapted approach is also equipped with the ability to dynamically change junction configurations (i.e. diverging to converging or vice versa) as the result of flow reversals. The intent here was to develop an approach that would permit the simulation of flow in a channel network that includes the more important physical effects at junctions but without the need to adjust model parameters or redefine junctions should a flow reversal occur. This momentum based approach to simulate junctions is assessed using a series of steady and unsteady tests using a 2D model, the University of Alberta's River2D, for comparison. For the diverging junction tests, the proposed 1D network model performed very well with respect to the discharge split. The model accurately simulated the water surface elevation in the two receiving channels but tended to overestimate the water surface elevation immediately upstream of the junction, perhaps attributable to model discretization and/or neglecting the centrifugal forces acting on the main channel as the lateral channel branches off. For the two parallel channels with a perpendicular connecting channel tests, the 1D model agreed well with the 2D model for all steady and unsteady tests. The unsteady test results demonstrated how capable the 1D model is at handling transient flow reversals. The model is then applied to a network of channels in the Mackenzie Delta. The model was calibrated and validated for three open water events, and subsequently used to simulate flow conditions during the 2008 breakup. Model results agreed well with observed water level data collected using data loggers. A comparison of the model flows for the pre-jam and ice-jam conditions suggests that ice jamming can significantly impact the distribution of flow in the upper Mackenzie Delta. For the ice-jam conditions, the simulated flow reversal in the Peel Mackenzie Connector is consistent with observations in this channel during the 2008 breakup.