Development of a monitoring protocol for bioengineered sediment mitigation related to stream adjacent bank failures on Fourmile and Sholes Creeks in the Mattole River Watershed

The North Coast Watershed Assessment Program's March 2003 assessment of the eastern subbasin of the Mattole River Watershed found that high sediment and other parameters current limit salmonid populations (California Resources Agency 2003). A second study, the Total Maximum Daily Load assessment for the Mattole River Watershed, authored in 2003 by the United States Environmental Protection Agency confirmed this concern, noting the 79 square miles of the eastern subbasin contribute 1870 tons per square mile per year due to natural mass wasting and stream bank erosion (EPA 2003). In an attempt to reduce erosion in the Mattole River watershed, the Mattole Restoration Council implemented a restoration project during the summer of 2004. The project consisted largely of installing bioengineered willow fences at the base of stream adjacent bank failures present along the banks of Fourmile and Sholes Creek, both located in the eastern subbasin of the Mattole River Watershed. The purpose of this thesis was to develop and test a replicable monitoring protocol for the Mattole Restoration Council to be used to assess the success of bioengineered sediment mitigation projects. The protocol was evaluated at identified treatment sites of channel adjacent bank failures along Fourmile Creek and contributing tributaries and Sholes Creek in the eastern subbasin of the Mattole River watershed and hypothesized that no difference would be detected between erosion contributions for control and treatment sites. The primary field method for the protocol was surveying twenty (ten control and ten treatment) channel adjacent bank failures using a total station to obtain topographic x, y, and z coordinates that were used to calculate volume, slope, and area for the twenty landslides before and after treatment of two field seasons. Data from both seasons' pairs of control and treated sites were overlaid using ESRI ArcMap 9.1. The overlaid data was used to statistically compare sediment contributions from treated and control sites. The results of the protocol from this project will be used by the Mattole Restoration Council to help assess the effectiveness of their sediment reduction treatments, as well as to continue monitoring of these sites in the future. First-year monitoring results concluded that the applied treatment did not mitigate erosion. Resulting data was statistically insignificant and demonstrated that there was no real difference between the net gain or loss of soil volume from either control or treated landslides. A Two-Sample Hotelling's T Test of the net gain or loss of soil volume from control and treated landslides resulted in a P value of 0.0238. Similarly, a Hotelling's Paired-Sample T Test also comparing the net gain or loss of soil volume from control from pairs of control and treated landslides yielded a P value of 0.3246. Finally, a Paired T Test, again testing the same parameters as in the Hotelling's Paired-Sample T Test and Two-Sample Hotelling's T Test, demonstrated a P value of -1.042. In all three statistical tests appropriate for this data set, the resulting P values were significantly out of range of the 0.05 target value to ensure statistical significance. Furthermore, the resulting volumes that can be retained by the restoration technique was minimal compared to the overall volumes of sediment displaced in these landslides.