The feasibility of developing a river channel classification for use in river management is examined.
Substantial progress has been made towards a geomorphological classification of semi-natural rivers from the ‘dimension’ variables. The classification is useful in its own. right as a statistically-justified delineation of-rivers on the basis of substrate type and it has the potential further to classify channels as sub-groups of substrate types on the basis of geomorphological features. A classification based on substrate type makes it relatively easy to conceptualise for the casual user. It also demonstrates that the -TWINSPAN technique has applications for deriving ,classes from this type of mixed .data - subject always to constraints imposed by the quality and appropriateness of the datasets themselves.