Classifying urban rivers

Classification systems have been developed over the last century as a tool to aid managers in the preservation, conservation, enhancement and management of rivers. The classification systems developed to date have been designed to differentiate between relatively unimpacted, mainly rural rivers. Urban rivers typically show poor water quality and biological diversity, and so most current classification systems tend to group urban rivers into a single "poor" category. In this paper we describe a hierarchical framework for recording information about urban rivers that allows a more sensitive description of these rivers enabling subdivision into several classes according to the purpose of the classification. The different levels in the hierarchy, the types of attributes that are to be recorded at each level, and the relational database structure for storing the data are described. The 100-500 m river stretch level in the hierarchy relates to the engineered modification of urban rivers and is the key to their classification. An example classification at this scale illustrates a link between engineering modification, bank and bed materials and the number and diversity of physical habitats present. This classification underlines the importance of adopting a hierarchy of nested spatial scales for data collection, classification and interpretation since it illustrates a clear link between characteristics at the stretch scale and at the finer habitat scale. The classification also illustrates the varied nature of urban rivers and the fact that even quite heavily engineered stretches can contain a diversity of habitat types.