Back to Search
Start Over
Uncertainty analysis of hydrological multi-model ensembles based on CBP-BMA method
- Source :
- Hydrology Research. 49:1636-1651
- Publication Year :
- 2018
- Publisher :
- IWA Publishing, 2018.
-
Abstract
- Quantification of the inherent uncertainty in hydrologic forecasting is essential for flood control and water resources management. The existing approaches, such as Bayesian model averaging (BMA), hydrologic uncertainty processor (HUP), copula-BMA (CBMA), aim at developing reliable probabilistic forecasts to characterize the uncertainty induced by model structures. In the probability forecast framework, these approaches either assume the probability density function (PDF) to follow a certain distribution, or are unable to reduce bias effectively for complex hydrological forecasts. To overcome these limitations, a copula Bayesian processor associated with BMA (CBP-BMA) method is proposed with ensemble lumped hydrological models. Comparing with the BMA and CBMA methods, the CBP-BMA method relaxes any assumption on the distribution of conditional PDFs. Several evaluation criteria, such as containing ratio, average bandwidth and average deviation amplitude of probabilistic application, are utilized to evaluate the model performance. The case study results demonstrate that the CBP-BMA method can improve hydrological forecasting precision with higher cover ratios more than 90%, which are increased by 4.4% and 3.2%, 2.2% and 1.7% over those of BMA and CBMA during the calibration and validation periods, respectively. The proposed CBP-BMA method provides an alternative approach for uncertainty estimation of hydrological multi-model forecasts.
- Subjects :
- Mathematical optimization
Calibration and validation
Computer science
0208 environmental biotechnology
Bayesian probability
Bandwidth (signal processing)
Probabilistic logic
Probability density function
02 engineering and technology
Bayesian inference
020801 environmental engineering
Copula (probability theory)
Uncertainty analysis
Water Science and Technology
Subjects
Details
- ISSN :
- 22247955 and 00291277
- Volume :
- 49
- Database :
- OpenAIRE
- Journal :
- Hydrology Research
- Accession number :
- edsair.doi...........1176ccc02faac6bf43f259b0f6952699