Dam Removal: Adaptive Management & Bed Sediment Monitoring Before and After

Dam removal is crucial for reconnecting river habitats, restoring passage of fish and other aquatic organisms, and restoring the free flow of water and sediment. However, removal of obsolete dams is often resisted due to concerns of releasing sediment and initiating channel instability. Two dams on the Musconetcong River in northern New Jersey have been removed by Princeton Hydro LLC in 2008 and 2009 as part of a watershed-wide effort to remove or breach all major obstructions to restore the river to its original free-flowing state. The two dams were consecutively situated 0.6 miles apart and their removals provided lessons in design and construction as well as offered an opportunity to study the geomorphic response in the form of bed elevation changes and sediment size through pre- and post-removal monitoring. The Musconetcong River, within the Delaware River watershed, is 89 km long and has a watershed of approximately 400 km². The watershed has mostly rural (forest and agriculture) land use with a few small towns, the largest being Hackettstown (population ~10,000). The Musconetcong flows southwest and is predominantly an alluvial stream with mostly shale and carbonate bedrock lithologies (Drake et al., 1996). The importance of measuring, predicting, and mitigating the effects of these dam removals is augmented by 24 miles of the Musconetcong River being federally designated as "National Wild and Scenic" in 2006 (Musconetcong Advisory Committee, 2003). For the removal of the downstream dam, the permitted engineering design called for the installation of boulder rock vanes that would maintain a low, in-stream, grade-control to stabilize some of the sediment, allow fish passage, direct flow away from downstream bridge abutments, and blend in with the character of the waterway. Unforeseen conditions that arose during construction necessitated in-field adjustments, or adaptive management. The dam structure itself was notched weeks prior to the start of construction to allow for partial dewatering of the impoundment. Due to downcutting and channel formation, this process drastically altered upstream bed elevations, such that the proposed design was no longer appropriate. Through collaboration with NJDEP Dam Safety and other project partners, the in-field redesign of not only the elevations, but the configuration and layout of the weir, was crucial to moving the project forward. Ultimately, two boulder rock vanes were installed to maintain an in-channel grade control to limit channel incision and retain some of the impounded sediments, while allowing for fish passage and directing flow away from abutments of a downstream bridge. In the course of construction, sediment was managed in several ways. First, some impounded sediments were permitted to migrate downstream after notching the dam to dewater the impoundment. Second, loose fine sediments (dominated by silt and sand) were excavated from the channel center and placed on the channel margins during the installation of the boulder rock vanes. Sediments were excavated to the approximate depth of the original stream bottom where highly resistant, consolidated layers were exposed to serve as a solid foundation for the footer stones of the boulder rock vane. These design details are important in understanding initial local patterns of erosion and deposition and may be important in interpreting the long-term geomorphic response of the reach. The upstream dam was crudely constructed of concrete poured over a base of river cobbles most likely gathered from the impounded reach. The impoundment was created to provide a swimming area and beach, which was periodically enhanced with imported sand. The structure had two 36-inch (approx) concrete culverts on the right and left sides of the channel that were set nearly flush to the riverbed. As a result, the impoundment was limited in extent, contained entirely within the river channel and sediments had only accumulated in the center of the channel. Over-topping flows had caused a bank failure on river right of the structure. Concerns of project partners and regulators varied widely regarding the fate of impounded sediments, the risk of channel instability and the need for in-stream grade control structures. After much discussion, the final design avoided unnecessary and costly in-stream structures and allowed for gradual geomorphic adjustment. Permits required the rebuilding of the eroded bank and placement of cobbles and boulders in the channel. To minimize the use of imported materials, the bank was built-out with cobbles from the dam and impounded sediments immediately upstream of the dam. Initial geomorphic surveys of the riverbed in the vicinity of and between the two dams have shown areas of erosion and deposition. These surveys have established a set of control points along the river channel between the two dams, and confirm the downstream movement of a sediment plume and localized areas of erosion. At the upstream dam, comparisons pre- and post-dam removal surveys show greater than 100 cubic meters of sediment being both eroded and deposited within the site. Most but not all of the erosion occurred around the newly exposed sediment bar upstream of the former dam, where the thalweg has reestablished itself following the dam's removal. Areas that were excavated during removal have experienced deposition. Most of the deposition occurred downstream and on the left-hand bank. Due to the two low flow culverts in the former dam, a mid-channel sediment bar remains. At the downstream dam site, erosion has removed up to 1.1 m of sediment from the bed in places while depositing up to 0.5 m sediment in others. As sediment from the former impoundment migrated through the project site, areas excavated during the removal became areas of deposition following the removal, and; alternately, areas in the channel margins where sediments were placed experienced gradual erosion. Grain size analysis shows a coarsening of the riverbed over the first nine months since removal. Grain size analyses were done upstream and downstream of the dam sites as well as at two locations between the sites. Wolman pebble counts (Wolman, 1954) were completed using the random walk method at each of the six sites. The largest change in grain sizes at the four sites occurred upstream of the downstream dam site, where there was a significant coarsening of the sediment from October 2008 to June 2009. This has most likely occurred from the increase in energy upstream of the dam post-removal, which has transported many of the fine-grained sediments downstream. Downstream of this dam site sediment size has not significantly changed, suggesting that the fine sediments have been transported downstream far enough to leave the site. Surveys of the channel thalweg above and below both dams also show a pulse of sediment migrating slowing from the uppermost impoundment areas. Long-term monitoring of the channel thalweg may reveal reach-level changes in channel slope.