Comparison of uncertainty analysis methods for a distributed rainfall-runoff model

Pao Shan Yu, Tao Chang Yang, Shen Jan Chen

Research output: Contribution to journalArticlepeer-review

80 Citations (Scopus)

Abstract

A rainfall-runoff model is normally applied to storm events outside of the range of conditions in which it has been successfully calibrated and verified. This investigation examined the uncertainty of model output caused by model calibration parameters. Four methods, the Monte Carlo simulation (MCS), Latin hypercube simulation (LHS), Rosenblueth's point estimation method (RPEM), and Harr's point estimation method (HPEM), were utilized to build uncertainty bounds on an estimated hydrograph. Comparing these four methods indicates that LHS produces analytical results similar to those of MCS. According to our results, the LHS only needs 10% of the number of MCS parameters to achieve similar performance. However, the analysis results from RPEM and HPEM differ markedly from those from MCS due to the very small number of model parameters.

Original languageEnglish
Pages (from-to)43-59
Number of pages17
JournalJournal of Hydrology
Volume244
Issue number1-2
DOIs
Publication statusPublished - 2001 Apr 2

All Science Journal Classification (ASJC) codes

  • Water Science and Technology

Fingerprint Dive into the research topics of 'Comparison of uncertainty analysis methods for a distributed rainfall-runoff model'. Together they form a unique fingerprint.

Cite this