科研之友
微信
新浪微博
Facebook
分享链接
摘要
The data-based mechanistic (DBM) approach to modelling has been proposed as an alternative to the conceptual rainfall-runoff (CRR) models commonly used in operational flood forecasting systems. The approach offers a number of potential advantages, but questions remain over its applicability to lowland UK catchments, its use with 15 min data, the most appropriate model identification and real-time updating approaches, as well as how the flood forecasting performance of DBM models compares with that of CRR models. This paper applies the DBM approach to two catchments in south-east England, finding that it is as appropriate for lowland catchments and 15 min data as for the upland catchments and hourly data of previous studies. It also finds that continuous-time model identification is superior to discrete-time model identification, and that such models can be applied within a Kalman filter updating framework. Overall, however, the relatively simple application of DBM modelling could not improve upon the performance of a more complex CRR model, indicating that further work is required to exploit the potential of the DBM modelling framework to operational flood forecasting.
- 出版日期2012-2
