RIVER RESTORATION ON THE CHICKLEY RIVER, PREDICTIONS OF CHANNEL CHANGE, AND THE ROLE OF A MAJOR FLOOD AS A NATURAL AGENT OF RIVER RESTORATION

Two key questions emerge from field studies of river restoration projects: 1) can we predict changes that will occur in altered reaches, and 2) was the restoration successful? In this study, we examine the Chickley River in western Massachusetts, which was affected by a series of condition-altering events, including channelization (~1950), debris removal (~1950 to 2010), extreme flooding (Tropical Storm Irene, 2011), channelization and debris removal (2011), and subsequent restoration (2012). Using a combination of field observations, repeat cross-section measurements, hydraulic modeling to compute shields parameter, and analysis of aerial imagery, we examine changes in woody debris, physical complexity, and channel stability (the ability of a stream to maintain channel geometry without severe erosion or deposition during water and sediment transport). In a novel application of shields parameter, we predicted that channels would be unstable if the shields parameter was outside the range of 0.04 to 0.09 and stable if the shield parameter was within the range of 0.04 to 0.09, which is the characteristic range of shields parameter in natural, gravel-bedded rivers. We find that the flood created stable channels, the channelization created unstable channels, and the restoration improved channel stability. These changes in stability were successfully predicted through hydraulic modeling of shields parameter, with shields parameter values of 0.02, 0.12 and 0.15 for unstable channels. For woody debris, the flood improved conditions from ~ 30 pieces per km pre-flood to ~330 pieces per km post-flood. Channelization worsened conditions to 0 pieces per km. Restoration improved conditions to ~ 30 pieces per km but did not recreate post-flood conditions. Physical complexity followed a similar trajectory as woody debris, with the flood creating beneficial heterogeneous habitat, channelization worsening conditions by homogenizing the bed and banks, and restoration improving physical complexity but not to the post-flood condition. Overall the restoration was clearly successful, but these results question the need to channelize and remove wood from rivers immediately after floods. Moreover, the results suggest that the extreme flood was an agent of river restoration on a regional scale.