Assessment of Stream Restoration Efforts and Development of Effectiveness Monitoring Protocol, Ohop Creek, Pierce County, Washington

Abstract

Between 2008 and 2015, the Nisqually Indian Tribe, Nisqually Land Trust, and South Puget Sound Salmon Enhancement Group conducted a multiphase restoration project on a two-km section of Ohop Creek, located west of the town of Eatonville, WA. The creek had been channelized in the early part of the 19th century to develop Ohop Valley for agriculture. The restoration focused on constructing a new channel to provide improved aquatic and riparian habitat. Restoration on the upstream section, Phase I, occurred between 2008 and 2010 and the downstream section, Phase II, between 2014 and 2015. A key element of the restoration was the use of engineered wood structures (LWM) to create flow complexity and provide cover for anadromous fish. To evaluate the success of the restoration project in meeting objectives, I assessed how the channel morphology and aquatic habitat distribution have evolved since completion. The assessment focused on four key components: the stability of the channel dimensions, the amount and distribution of in-channel habitat, the physical effects on the channel from the engineered wood structures, and the frequency of floodplain inundation. From observations and analysis, I developed a protocol to monitor habitat and to help land owners and future researchers continue to assess the effectiveness of the restoration treatment. The purpose of the monitoring protocol is to inform stakeholders in future restoration on other sections of Ohop Creek. The protocol implements benchmarked cross-sections, longitudinal profile and pool survey, discharge measurements, and an extensive photo survey that documents LWM and channel bank interaction. I measured channel widening from aerial photos and orthographic images between 2011 and 2017 using transects oriented perpendicular to the channel centerline. Greatest average annual rate of widening occurred in areas with engineered wood structures in Phase I, with a consistent average rate of 0.3 m/yr from 2011 to 2017. Widening in the Phase II reach was also greatest for transects located in areas containing engineered wood structures, with an annual average rate of 0.2 m/yr from 2015-2017. Channel incision was determined from comparing 25 surveyed cross-sections with design cross-sections derived from as-built drawings and construction plans. Deepening from LWM-induced bed scour is relatively active throughout most of the project reach, averaging 0.2 m/yr in the Phase II reach from 2015 to 2017. The magnitude of deepening in the Phase I channel could not be confidently quantified because it was not possible to verify if the Phase I reach had been constructed according to design plans. Channel habitat was investigated by conducting a pool survey using a survey-grade GPS with real time kinematic (RTK) corrections, recording the depth, location, and type of each pool. Pools formed from scour around LWM and free-form pools augmented by nearby LWM structures constituted 51% of the pools in the Project reach. LWM scour pools were on average deeper than other pool types, with an average residual depth of 0.4 m. LWM-induced bank scour was surveyed and measured from orthographic images. Forty percent of the structures were observed to be inducing lateral bank scour. LWM structures composed of several vertical log piles interwoven with horizontal logs and rootwads (referred to as ìcomplexî or ìtype 6î structures) positioned at the inflection point between meanders were associated with the highest frequency and magnitude of bank scour, averaging 25 m2 of lateral scour per structure between 2015 and 2017. Discharge measurements made in the restoration area were on average 6.5% higher than discharge measurements recorded six kilometers upstream at USGS gage #12088000 ìOhop Creek near Eatonville, WAî, which is 9% lower than the increase in flow derived by engineers by using a regional scaling relationship between drainage area and discharge. During a site visit in February of 2018 I observed the Phase II reach to be at or just above bankfull flow at 9.1 m3/s. Using reach average cross-sectional bankfull area and water-surface slope, I calculated a bankfull discharge of 12.4 m3/s for the Phase I reach from application of Manningís equation. Flood duration analysis of USGS mean daily discharge records from 1994 to 2010 indicate that the Phase I and II bankfull flows are expected to inundate the floodplain an average of 2.2 and 5.4 days per year, respectively, which is 50-60% less than predicted by Project engineers. Based on the findings presented in the current study, the LWM structures are providing bed scour for fish habitat in the Project reach. In addition, channel widening was shown to be strongly correlated with lateral bank scour induced by LWM-structures, indicating that the engineered wood structures are providing both in-channel habitat and flow complexity necessary for achieving the Project objective of enhancing aquatic habitat for rearing fish. However, initial channel widening from LWM-induced bank scour has decreased the frequency of floodplain inundation predicted by Project engineers, which negatively impacts the Project goal of riparian zone rehabilitation at least in the short-term. Monitoring the LWM structures is a key component of the protocol developed from this study and will help to resolve this issue in the future.