The geomorphology and management of a dynamic, unstable gravel-bed river: the Feshie/Spey confluence, Scotland

The alluvial fan that has developed at the confluence of the Rivers Feshie and Spey over the past 13,000 years provides an exceptional example of an unstable, gravel-bed river in the Scottish Highlands protected under UK and EU environmental law. River engineering extending back to the early 19th century has only registered a modest impact on this dynamic system up to the 1980s. Since then a flood rich period (1989-1994) generated a series of avulsions caused by local aggradation of the main channel. Initially the channel switched to the western side of the fan (in 1989-90) and was restored to its formed position following there-instatement of flood backs and channel regrading in 1991. By 1997 the main channel had shifted to the eastern side of the fan following a further avulsion and the re-occupation of palaeo-channels, triggered in part by further repairs of the flood banks on the west side of the fan but mainly in response to natural processes. In doing so the channel now occupies a position akin to that proposed in a river engineering scheme proposed in 1991 but not implemented. Re-sectioning of the Spey to provide a flood-relief channel immediately downstream of the confluence was completed in 1992, but this has not adversely impacted on water levels in the Loch Insh marshes (a internationally protected wetland upstream of the Feshie-Spey confluence). In seeking to reconcile the conflicting demands of nature conservation and generating an economic return from the land, a map of the geomorphological sensitivity of the site (based on Brunsden's "landscape change safety factor" concept) has been developed. This reports three zones with contrasting sensitivities and recommended management strategies: (1) highly sensitive and dynamic areas (HS-5) where no engineering works should be attempted; (2) areas of medium sensitivity (M-20) where permitted engineering works need careful management and monitoring; and (3) areas of low sensitivity (LS-100) where appropriate river engineering should be permitted. The assessment of environmental risk is based on the probability of each of the zones being de-stabilised by floods with return periods of 5, 20 and 100 years, respectively. Paradoxically, this imprecise guidance to river engineers provides the best framework for combining conservation sensitive management with cost-effective engineering.