A HINDCAST COMPARING FIELD OBSERVATION OF THE RESPONSE OF THE SOUHEGAN RIVER TO DAM REMOVAL WITH SIMULATIONS OF THE DAM REMOVAL EXPRESS ASSESSMENT MODEL-1

Dam removal is a widely used river restoration technique. Historically, dams produced hydropower, controlled flooding, and provided water storage, but currently many dams in the United States, specifically low head dams in New England, are obsolete. Early research produced conceptual models of upstream and downstream channel response after dam removal, and recent studies quantified channel response with cross section surveys and grain size analysis. This study aims to assess the accuracy of a simple morphodynamic sediment transport model, Dam Removal Express Assessment Model (DREAM-1), developed by Cui et al. (2008). We compare simulations to a dam removal monitoring project that quantified the physical response of the Souhegan River to the removal of the Merrimack Village Dam, Merrimack, NH. Pearson et al. (2011) reported results of field monitoring from August 2007-May 2010 and found that the Souhegan River responded to dam removal in two phases: initial rapid incision of impoundment sediment induced by immediate base level drop of 3.9 m (~50% of impounded sediment eroded in ~2 months), followed by an event-driven phase in which impoundment sediment eroded primarily during floods. The reach downstream of the dam showed a similar two-phase response, with rapid deposition in the first three weeks after dam removal, followed by bed degradation to the pre-removal elevation profile within a year. We extend the work of Pearson et al. (2011) by continuing field monitoring and quantifying channel response in June 2011 and July 2012. Results of the most recent two surveys show that the channel reached a quasi-equilibrium state in which the former impoundment lost only 1-2% of remaining sediment per year. We hypothesize that in the future high discharge events will be necessary to access the remaining impoundment sediment. Using five years of comprehensive field data, we conduct a hindcast to compare the sediment erosion and deposition patterns predicted by DREAM-1 to the observed downstream response of the Souhegan River. We model the change in bed elevation for the downstream channel reach at intervals that correspond with the dates of four longitudinal profile surveys. Results of initial runs show DREAM-1 can capture the downstream two-phase response of rapid deposition followed by bed degradation to pre-removal elevation.