Predictive uncertainty in water-quality modeling

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

A general and integrated approach to parameter identification, model calibration, and estimation of predictive uncertainty in water-quality models is proposed and validated. The proposed approach determines the maximal conditional likelihood functions of each of the model parameters using a transformation that forces the model errors to be normally distributed, with predictive uncertainty characterized by random normally distributed and homoscedastic model errors in the transform space. The proposed approach is demonstrated using a watershed-scale model to predict the fecal coliform levels in a third-order stream within the Little River Experimental Watershed in Georgia. Maximal conditional likelihood functions were identified for all parameters in the log, square root, and no-transformation cases. The key results are: (1) the number of sensitive parameters and the optimal parameter values can depend on the transformation; (2) only in the case of the log-transformation are the errors normally distributed and consistent with the assumed Gaussian likelihood function; (3) the standard error in the model is least for the no-transform case and highest for the log-transform case; and (4) the observed model errors are most predictable using the log-transform and least predictable using the no-transform approach.

Original languageEnglish
Pages (from-to)1315-1325
Number of pages11
JournalJournal of Environmental Engineering
Volume135
Issue number12
DOIs
StatePublished - Nov 27 2009

Fingerprint

Water quality
water quality
transform
modeling
Mathematical transformations
Watersheds
watershed
Uncertainty
fecal coliform
integrated approach
Identification (control systems)
Rivers
parameter
Calibration
calibration
river

Keywords

  • Computer models
  • Rivers
  • Streams
  • Uncertainty principles
  • Water quality
  • Watersheds

ASJC Scopus subject areas

  • Environmental Engineering
  • Environmental Science(all)
  • Environmental Chemistry
  • Civil and Structural Engineering

Cite this

Predictive uncertainty in water-quality modeling. / Chin, David A.

In: Journal of Environmental Engineering, Vol. 135, No. 12, 27.11.2009, p. 1315-1325.

Research output: Contribution to journalArticle

@article{4308f6ddcbac4133b9abb76b494f02bc,
title = "Predictive uncertainty in water-quality modeling",
abstract = "A general and integrated approach to parameter identification, model calibration, and estimation of predictive uncertainty in water-quality models is proposed and validated. The proposed approach determines the maximal conditional likelihood functions of each of the model parameters using a transformation that forces the model errors to be normally distributed, with predictive uncertainty characterized by random normally distributed and homoscedastic model errors in the transform space. The proposed approach is demonstrated using a watershed-scale model to predict the fecal coliform levels in a third-order stream within the Little River Experimental Watershed in Georgia. Maximal conditional likelihood functions were identified for all parameters in the log, square root, and no-transformation cases. The key results are: (1) the number of sensitive parameters and the optimal parameter values can depend on the transformation; (2) only in the case of the log-transformation are the errors normally distributed and consistent with the assumed Gaussian likelihood function; (3) the standard error in the model is least for the no-transform case and highest for the log-transform case; and (4) the observed model errors are most predictable using the log-transform and least predictable using the no-transform approach.",
keywords = "Computer models, Rivers, Streams, Uncertainty principles, Water quality, Watersheds",
author = "Chin, {David A.}",
year = "2009",
month = "11",
day = "27",
doi = "10.1061/(ASCE)EE.1943-7870.0000101",
language = "English",
volume = "135",
pages = "1315--1325",
journal = "Journal of Environmental Engineering, ASCE",
issn = "0733-9372",
publisher = "American Society of Civil Engineers (ASCE)",
number = "12",

}

TY - JOUR

T1 - Predictive uncertainty in water-quality modeling

AU - Chin, David A.

PY - 2009/11/27

Y1 - 2009/11/27

N2 - A general and integrated approach to parameter identification, model calibration, and estimation of predictive uncertainty in water-quality models is proposed and validated. The proposed approach determines the maximal conditional likelihood functions of each of the model parameters using a transformation that forces the model errors to be normally distributed, with predictive uncertainty characterized by random normally distributed and homoscedastic model errors in the transform space. The proposed approach is demonstrated using a watershed-scale model to predict the fecal coliform levels in a third-order stream within the Little River Experimental Watershed in Georgia. Maximal conditional likelihood functions were identified for all parameters in the log, square root, and no-transformation cases. The key results are: (1) the number of sensitive parameters and the optimal parameter values can depend on the transformation; (2) only in the case of the log-transformation are the errors normally distributed and consistent with the assumed Gaussian likelihood function; (3) the standard error in the model is least for the no-transform case and highest for the log-transform case; and (4) the observed model errors are most predictable using the log-transform and least predictable using the no-transform approach.

AB - A general and integrated approach to parameter identification, model calibration, and estimation of predictive uncertainty in water-quality models is proposed and validated. The proposed approach determines the maximal conditional likelihood functions of each of the model parameters using a transformation that forces the model errors to be normally distributed, with predictive uncertainty characterized by random normally distributed and homoscedastic model errors in the transform space. The proposed approach is demonstrated using a watershed-scale model to predict the fecal coliform levels in a third-order stream within the Little River Experimental Watershed in Georgia. Maximal conditional likelihood functions were identified for all parameters in the log, square root, and no-transformation cases. The key results are: (1) the number of sensitive parameters and the optimal parameter values can depend on the transformation; (2) only in the case of the log-transformation are the errors normally distributed and consistent with the assumed Gaussian likelihood function; (3) the standard error in the model is least for the no-transform case and highest for the log-transform case; and (4) the observed model errors are most predictable using the log-transform and least predictable using the no-transform approach.

KW - Computer models

KW - Rivers

KW - Streams

KW - Uncertainty principles

KW - Water quality

KW - Watersheds

UR - http://www.scopus.com/inward/record.url?scp=60749100599&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=60749100599&partnerID=8YFLogxK

U2 - 10.1061/(ASCE)EE.1943-7870.0000101

DO - 10.1061/(ASCE)EE.1943-7870.0000101

M3 - Article

AN - SCOPUS:60749100599

VL - 135

SP - 1315

EP - 1325

JO - Journal of Environmental Engineering, ASCE

JF - Journal of Environmental Engineering, ASCE

SN - 0733-9372

IS - 12

ER -