Evaluations of Lagrangian egg drift models: From a laboratory flume to large channelized rivers

To help better interpret computational models in predicting drift of carp eggs in rivers, we present a series of model assessments for the longitudinal egg dispersion. Two three-dimensional Lagrangian particle tracking models, SDrift and FluEgg, are evaluated in a series of channels with increasing complexity. The model evaluation demonstrates that both models are able to accommodate channel complexity and provide a wide range of dispersion coefficients: K_l=0(1 − 100)Hu_∗ with H being water depth and u_∗ being shear velocity. In a straight channel with K_l=0(1)Hu_∗ SDrift predicts weaker longitudinal dispersion than FluEgg in the early stage as a result of weak vertical mixing associated with smooth wall turbulence. With sufficient time, SDrift and FluEgg predict similar egg dispersion, accounting for the differential advection due to the vertical velocity profile. In an idealized curved channel with K_l=0(10)Hu_∗, dispersion is driven by both vertical and transverse velocity profiles. SDrift yields slightly larger dispersion coefficients than FluEgg. In a real river with channel-training structures and having K_l=0(100)Hu_∗ SDrift predicts a stronger longitudinal dispersion than FluEgg due to substantial local turbulent eddies and velocity gradients. To summarize, FluEgg shows good performance in capturing dispersion due to vertical velocity profiles and cross-channel velocity gradients. SDrift shows excellent model capabilities of revealing various dispersion mechanisms in addition to the vertical and cross-channel velocity variations. They include the initial turbulent diffusion stage with growing dispersion coefficients and strong dispersion due to in-stream hydraulic structures and localized turbulence.