Abstract:
Current hydrologic predictions are notoriously unreliable if we lack sufficient observations of the hydrologic variable of interest (most often strearnflow). In this paper we show how catchment classification can contribute to increase the reliability of ungauged predictions. The basis for a classification framework relying on catchment similarity with respect to physical characteristics, climatic regime and response behavior is discussed. Response behavior can be quantified through signatures (hydrologic indices) that can be regionalized resulting in constraints on the expected behavior of ungauged catchments. Examples of studies utilizing such regional information for the prediction of continuous streamfiow are provided. Finally, the discussion is extended to include the concept of catchment services to explicitly address the need for improved support of water resource management. Ultimately such an approach will advance our ability to achieve water security for humans and ecosystems through improved predictions.