Traditional catchment-based approaches to runoff prediction suffer from the problem that it is difficult to interpret the collective physical significance of a large number of intercorrelated drainage basin variables. This has highlighted the need for a sensitive and meaningful index to relate the basin character to the discharge produced. Network routing models also require an appropriate descriptor of drainage basin form to relate to hydrologic response characteristics. An index of the drainage network is potentially the most valuable because it responds to precipitation, reflects the characteristics of the basin and affects runoff. Although a large number of drainage network indices have been developed they have proved inadequate in their failure to describe the dynamic nature of drainage networks. Future research into the use of drainage networks for rainfall- runoff modelling and runoff prediction needs to have as its central aim the development of a dynamic network index which has physical meaning for drainage basins of all sizes and which is quick and easy to calculate from data that are rapidly obtainable. Despite the improved resolution of satellite imagery its high cost still prevents the widespread application of satellite remote sensing techniques to monitoring storm-specific drainage network changes. However, the increased availability of topographic data in digital format and the recent developments in digital elevation models (DEMs) have demonstrated the potential for the rapid derivation of both perennial and extended drainage networks from which network expansion potential can be calculated.