Sampling methods and sensitivity analysis for large parameter sets

Andres, Terry

doi:10.1080/00949659708811804

Cited by 73 publications

(50 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…IFFD works well when first and second order parameter effects dominate (Andres, 1997). Using IFFD in ANOVA allows users to neglect higher order interactions not included in the model (Liang and Guo, 2003;Andres, 1997) while generating highly repeatable results (Saltelli et al, 1995).…”

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

“…Using IFFD in ANOVA allows users to neglect higher order interactions not included in the model (Liang and Guo, 2003;Andres, 1997) while generating highly repeatable results (Saltelli et al, 1995). Consequently, the number of model runs required can be reduced substantially.…”

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

“…The parameter levels are defined as equally spaced intervals within the predefined parameter ranges. Using a small number of factor levels enables the sampling scheme to attain statistically significant results efficiently and accurately (Mokhtari and Frey, 2005;Andres, 1997). IFFD extends the basic orthogonal fractional factorial design (FFD) by conducting multiple iterations.…”

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

“…If one parameter or parameter interaction is analyzed at a time in succession, the total number of model runs will be excessively large and most hydrologic applications would be computationally intractable. In this study, the iterated fractional factorial design (IFFD) sampling scheme (Andres and Wayne, 1993;Andres, 1997;Saltelli et al, 1995) was used to limit the computational burden posed by ANOVA while seeking high quality results.…”

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

See 3 more Smart Citations

Comparing sensitivity analysis methods to advance lumped watershed model identification and evaluation

Tang

Reed

Wagener

et al. 2007

Hydrol. Earth Syst. Sci.

312

220

View full text Add to dashboard Cite

Abstract. This study seeks to identify sensitivity tools that will advance our understanding of lumped hydrologic models for the purposes of model improvement, calibration efficiency and improved measurement schemes. Four sensitivity analysis methods were tested: (1) local analysis using parameter estimation software (PEST), (2) regional sensitivity analysis (RSA), (3) analysis of variance (ANOVA), and (4) Sobol's method. The methods' relative efficiencies and effectiveness have been analyzed and compared. These four sensitivity methods were applied to the lumped Sacramento soil moisture accounting model (SAC-SMA) coupled with SNOW-17. Results from this study characterize model sensitivities for two medium sized watersheds within the Juniata River Basin in Pennsylvania, USA. Comparative results for the 4 sensitivity methods are presented for a 3-year time series with 1 h, 6 h, and 24 h time intervals. The results of this study show that model parameter sensitivities are heavily impacted by the choice of analysis method as well as the model time interval. Differences between the two adjacent watersheds also suggest strong influences of local physical characteristics on the sensitivity methods' results. This study also contributes a comprehensive assessment of the repeatability, robustness, efficiency, and ease-of-implementation of the four sensitivity methods. Overall ANOVA and Sobol's method were shown to be superior to RSA and PEST. Relative to one another, ANOVA has reduced computational requirements and Sobol's method yielded more robust sensitivity rankings.

show abstract

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

Section: Analysis Of Variance Using Iterated Fractional Factorial Desmentioning

confidence: 99%

See 2 more Smart Citations

Comparing sensitivity analysis methods to advance lumped watershed model identification and evaluation

Tang

Reed

Wagener

et al. 2007

Hydrol. Earth Syst. Sci.

312

220

View full text Add to dashboard Cite

show abstract

“…Also, specific techniques can be applied to validate SA conclusions, e.g. the visual test proposed by Andres (1997) to validate factor screening or the quantitative test based on the KolmogoroveSmirnov statistic presented in Pianosi and Wagener (2015). Here conditional (on either sensitive or insensitive parameters) and unconditional output distributions are compared to check whether all insensitive factors have been identified.…”

Section: Assess Credibilitymentioning

confidence: 99%

Sensitivity Analysis of Environmental Models: A Systematic Review with Practical Workflow

Pianosi

Wagener

Rougier

et al. 2014

Vulnerability, Uncertainty, and Risk

257

416

View full text Add to dashboard Cite

a b s t r a c tSensitivity Analysis (SA) investigates how the variation in the output of a numerical model can be attributed to variations of its input factors. SA is increasingly being used in environmental modelling for a variety of purposes, including uncertainty assessment, model calibration and diagnostic evaluation, dominant control analysis and robust decision-making. In this paper we review the SA literature with the goal of providing: (i) a comprehensive view of SA approaches also in relation to other methodologies for model identification and application; (ii) a systematic classification of the most commonly used SA methods; (iii) practical guidelines for the application of SA. The paper aims at delivering an introduction to SA for non-specialist readers, as well as practical advice with best practice examples from the literature; and at stimulating the discussion within the community of SA developers and users regarding the setting of good practices and on defining priorities for future research.

show abstract

Using Unreplicated 2k−p Designs for Characterizing Moderately Dimensioned Deterministic Computer Models

Houston

Ferreira

Montgomery

2005

Qual. Reliab. Engng. Int.

View full text Add to dashboard Cite

Deterministic computer simulation models (DCSMs) have been used extensively by two very different schools of researchers. One has produced highly dimensioned models (hundreds of parameters) of very complex physical phenomena and has explored various means of sensitivity analysis. The other school typically produces low-dimensioned models (dozens of parameters) and has very limited experience with sensitivity analysis. However, the sophisticated techniques used in the large models are not readily accessible to users of the smaller models. The approaches of both schools are reviewed before describing how classical DOE plans can be applied to sensitivity analysis of moderately dimensioned (a dozen to a couple hundred parameters) DCSMs. Because the usual analysis of classical DOE requires randomness, alternative means of experimental analysis are discussed. Four techniques for DCSM experimental analysis, including per cent contribution to variation, are described and the results are statistically compared.

show abstract

Sampling methods and sensitivity analysis for large parameter sets

Cited by 73 publications

References 10 publications

Comparing sensitivity analysis methods to advance lumped watershed model identification and evaluation

Comparing sensitivity analysis methods to advance lumped watershed model identification and evaluation

Sensitivity Analysis of Environmental Models: A Systematic Review with Practical Workflow

Using Unreplicated 2k−p Designs for Characterizing Moderately Dimensioned Deterministic Computer Models

Contact Info

Product

Resources

About