Habitat survey and classification of urban rivers

Abstract

AbstractRiver surveys are undertaken for a variety of purposes including (i) to establish inventories of particular features and their changes, (ii) to collect data to underpin the classification of river types or to assess resources according to particular criteria, and (iii) to identify sites that have particular qualities or may require particular types of management. In this paper we describe a new reach‐scale survey technique, a range of synthetic indices, and a series of classifications specifically developed for application to urban rivers.The Urban River Survey (URS) is developed from the River Habitat Survey (RHS) which is applied routinely to UK rivers. A number of important differences between the URS and RHS allow the former to provide improved discrimination between urban river channels to support management decision‐making. Urban river stretches are identified for survey according to their engineering type (a combination of planform, cross‐sectional form and level of reinforcement). The URS is then applied to stretches of a single engineering type and incorporates recording of (i) additional variables to the RHS that are particularly relevant to urban channels (e.g. indicators of pollution); (ii) improved resolution in the recording of some variables in comparison with the RHS (e.g. habitat features); and (iii) separation of layers of information that relate to the engineered (e.g. artificially introduced materials) and more natural (e.g. bank materials and morphological features) channel properties so that the interaction between these properties can be identified.The URS is applied during two surveys of approximately 50 stretches of the River Tame, West Midlands, UK. The data are used to estimate a range of synthetic indices describing ‘Materials’, ‘Physical Habitat’ and ‘Vegetation’ attributes of urban river stretches. Cluster analysis is then applied to these indices to derive three classifications of urban river stretches. The similarity in classifications based on measurements from two different surveys indicates their robustness. Because the type of engineering applied to a stretch appears to have a significant influence on the class to which the stretch is allocated in each of the three classifications (with the strongest associations being apparent in the Materials classes and the weakest in the Vegetation classes), they can be used to explore the consequences of changed engineering, and the influence of scenarios of vegetation and water quality management can be additionally explored in relation to the Vegetation classification. Copyright © 2004 John Wiley & Sons, Ltd.