Geomorphic Effects and Habitat Impacts of Large Wood at Restoration Sites in New England

Abstract

Large wood (used interchangeably with the term “instream wood”), which refers to trees, logs and other wood within a channel, is beneficial to river ecosystems and is being used more frequently as a component of river restoration projects. The process of large wood becoming stable within a river channel, inducing floodplain formation, and eventually providing large wood back to the system is known as the ‘floodplain large-wood cycle’ hypothesis (Collins et al., 2012). In a stream restoration context, this process can be viewed as an indicator of a self-sustaining cycle. The ‘floodplain large-wood cycle’ hypothesis was formulated in the Pacific Northwest. To investigate this process in other regions, I used the Merrimack Village Dam (MVD) study site in southern New Hampshire. The study site provided a location where instream wood was recruited to the river from an adjacent terrace as a consequence of erosion associated with a dam removal. Assessment of wood in this scenario was used to evaluate the ‘floodplain large-wood cycle’ (Collins et al., 2012), and to compare MVD to “passive” large wood restoration and deliberate, and potentially engineered, large wood restoration sites throughout New England. To assess multiple sites, I identified metrics to evaluate the effectiveness of large wood to promote ecological and geomorphic complexity within channels. The metrics were quantified at the MVD site and several other sites in New England with natural or placed large wood. I also collected additional data at the MVD site using methods implemented during previous studies, including cross section surveys and repeat photographs (Collins et al., 2017; Pearson et al., 2011). The study assessed habitat and geomorphic effects of large wood within river systems in the northeastern U.S. and provided information to evaluate the use of large wood during river restoration. Overall, only 33%, 33%, and 20% of surveyed sites are consistent with hypotheses formulated regarding significant differences in depth variability, velocity variability, and median velocity between test and reference reaches, respectively. With evidence for and against each hypothesis at both passive and active sites, large wood structures did not cause the geomorphic and hydraulic changes I expected to see. The availability of sand in a channel and the stream slope influencing sediment transport seem to be important factors in determining whether or not large wood has the ability to impact the geomorphic and hydraulic characteristics of a channel. At the MVD site, where sand is available, up to 0.90 m of sediment deposition is seen on top of the surface eroded by a March 2010 flood, surrounding recruited trees. Evaluation of historical aerial imagery further indicates that evidence of the ‘floodplain large-wood cycle’ hypothesis is present at the MVD06 cross section on the Souhegan River in New Hampshire.