Near-bank processes, bioengineering, and the dormant willow post method

Abstract

Bioengineered systems were constructed along Harland Creek, Mississippi, using the dormant willow post method. The systems were constructed using live posts harvested from sections of the trunks of native black willows (Salix nigra). The posts were planted vertically along the banks in a 1 m by 1 m rectangular grid. Rows in the grid were aligned parallel to the edge of water, and three to five rows of willow posts were planted. In time, willow posts develop roots, stems, branches, and leaves; features that enhance the ability of willow posts to stabilize streambanks and reduce erosion. The objectives of this research were to study near-bank processes and the influence of planted black willow posts on those processes. It is hypothesized that willow posts planted along an eroding bank will thrive only if the posts are able to induce net sediment deposition. Soil samples derived from the banks of Harland Creek were analyzed but no indication that the willow posts are inducing deposition was found. Three study reaches were established. Planted willow posts were present in two of the reaches; the third reach was a control reach in which no manmade bank stabilization was present. Vegetation density data were collected at each study reach in August 1997, January 1998, and March 1998 using a horizontal point frame. Vegetation density was found to be as much as 44% greater for the reaches with willow posts than for the control reach. This observation suggests that willow posts help create a suitable environment for the establishment of volunteer vegetation. Physical model test results revealed that willow posts significantly change flow patterns atop the bank. Depth-averaged velocities amid the simulated willow posts were reduced by more than 50% in the model. It was determined that computed Manning's roughness coefficients (n-values) did not correlate with the configuration of the simulated willow posts. An analytical model for predicting flow through emergent vegetation in trapezoidal channels is presented. The model predicted depth-averaged velocities behind the dowels to within 2%, on average, of velocities computed from the model study data.