Physical and Numerical Model Testing of Boulder Cluster Configurations In Urban Channels

The Bureau of Reclamation’s Hydraulics Laboratory in Denver, Colorado constructed a Froude-scale physical hydraulic model of the low-flow channel portion of the Los Angeles River. The purpose of the model was to investigate the placement of boulders in urban channels to provide the best habitat suitability for steelhead. The physical model was designed with a vertical scale of 1:4 and horizontal scale of 1:8 (distortion factor of 2) to provide sufficient water depth for data collection around the boulders. Selected boulder configurations were also analyzed at prototype scale with a two-dimensional (2D) numerical model.

Large and small boulders were tested in a physical hydraulic model using Acoustic Doppler Velocimetry and Large-Scale Particle Image Velocimetry. For the large boulder tests, a baseline and four unique rock configurations were tested at two different flow rates. For the small boulder testing, a baseline and three unique rock configurations were tested at three different flow rates. The configurations were: Single Rock, Upstream V, Downstream V, and Diamond. The Diamond configuration was only tested for large rocks. These configurations were tested at four to five different densities: high, medium-high (only for Single Rocks), medium, low, and very low. Water surface elevation data was collected for small boulder configurations in the streamwise direction through the test section. Dimensionless analysis was performed for boulder and flow properties to assess: 1) percent plan view area blocked; 2) percent cross-section area blocked; 3) percent volume blocked. The velocity ratios of each configuration were derived from cumulative distribution functions (CDF) for minimum, quartiles, and maximum values.

Upstream V and Downstream V boulder configurations create the greatest velocity reduction with the lowest density of boulder clusters. For all rock configurations, the more cross-sectional area obstructed by rocks, the more effectively the velocity is reduced in the channel. However, the trend does not significantly improve after 35% blocked. Therefore, the ideal amount of cross-sectional channel obstructed is between 30-40%. The higher the percent volume blocked by boulders in the modeled restoration section, the more velocity reduction was achieved for all configurations between 0 and 4% blocked. It is unclear, however, what the optimal percent blockage by volume should be, since larger percent blockages were not tested.

2D numerical model results indicate that boulders installed within the low-flow channel of the Los Angeles River provide suitable fish passage up to about the 1% exceedance mean daily flow. Boulders steer the flow by creating an upstream backwater effect, a downstream wake zone, and local flow acceleration around the boulders. Boulders create morphologic and hydraulic diversity in an otherwise uniform channel. The largest hydraulic effect occurs when the water surface is near the elevation of the boulder crests. Further increases to the discharge and water surface elevation cause flow to overtop the boulders, which disrupts the downstream wake zone. Deeply submerged boulders may provide local velocity refuge near the bed, but do not have a significant effect on the depth-averaged velocity.