EP31E-07:
Dam Breach Release of Non-Cohesive Sediments: Channel Response and Recovery Rates
EP31E-07:
Dam Breach Release of Non-Cohesive Sediments: Channel Response and Recovery Rates
Wednesday, 17 December 2014: 9:30 AM
Abstract:
Dam removals featuring unchecked releases of non-cohesive sediments are excellent opportunities to learn more about stream channel response to abrupt increases in bed material supply that can occur deliberately or by natural processes like landslides and volcanic eruptions. Understanding channel response to sediment pulses, including response rates, is essential because human uses of river channels and floodplains are impacted by these events as are aquatic habitats. We had the opportunity to study a dam removal site at the Simkins Dam in Maryland, USA, that shares many important geophysical attributes of another well-studied dam removal in the humid northeast United States [Merrimack Village Dam, New Hampshire; Pearson et al., 2011]. The watershed sizes are the same order of magnitude (102 km2), and at both sites relatively low head dams were removed (~ 3-4 m) and ~60,000 m3 of dominantly sand-sized sediments discharged to low-gradient reaches immediately downstream. Analyzing four years of repeat morphometry and bed sediment grain size surveys at the Simkins site on the Patapsco River, as well as continuous discharge and suspended sediment gaging data, we clearly document a two-phase response in the upstream reach as described by Pearson et al. [2011] for their New Hampshire site and noted at other dam removals [e.g., Major et al., 2012]. In the early phase, approximately 50% of the impounded sediment mass was eroded rapidly over a period of about three months when flows were very modest (Figure 1). After incision to base level and channel widening in the former impoundment, a second phase began when further erosion depended on floods large enough to access impounded sediments more distant from the newly-formed channel. We also found important differences in the upstream responses at the Maryland and New Hampshire sites that appear to be related to valley type (non-glaciated versus glaciated, respectively). Response variances immediately downstream between the respective sites are potentially related to local gradient and hydraulics.