Uncertainty analysis of the relationship between discharge and nitrate concentration in the Lower Illinois River using a copula model

Daeryong Park, Momcilo Markus, Kichul Jung, Myoung Jin Um

Research output: Contribution to journalArticle

Abstract

The complex relationship between water quality and river discharge has long been an active area of research. This research applies bivariate statistics to river nitrate concentration and discharge, and develops a procedure for selecting the best bivariate distribution for the chosen models. Nitrate concentration (NO 3 ) data from five stations on the Lower Illinois River, USA, collected from 1972 to 2012 were used for demonstration of this procedure. To explore the statistical relationship between discharge and nitrate concentration, three copula statistical models were applied to determine optimal bivariate distributions. The generalized Pareto (GPA), generalized extreme value (GEV), and lognormal two- (LN2) and three-parameter (LN3) distributions were used as marginal distributions, and were combined with the Clayton, Frank, and Gumbel copula models. The probability plot correlation coefficient (PPCC) test and the S n statistic were used to select the most suitable copula models. As a result, the GPA distributions for streamflow and nitrate concentration in the Frank copula model were more accurate for representing the data from the two mainstem stations, Havana (D-31) and Valley City (D-32), and those from the Oakford station along the Sangamon River (E-25). However, the Clayton copula model using NL3 for representing streamflow and GPA for representing nitrate concentration was more accurate for stations along the Spoon (DJ-08) and La Moine Rivers (DG-01). This study enabled the development of a procedure for determining an optimal copula for application in the environmental sciences and also provided a case study for applying copula models using discharge and nutrient concentration as hydrologic and water quality variables, respectively.

Original languageEnglish (US)
Pages (from-to)310-323
Number of pages14
JournalJournal of Cleaner Production
Volume222
DOIs
StatePublished - Jun 10 2019

Fingerprint

Uncertainty analysis
uncertainty analysis
Discharge (fluid mechanics)
Nitrates
Rivers
nitrate
river
Water quality
streamflow
Statistics
water quality
Nutrients
Copula
Nitrate
river discharge
Demonstrations
distribution
valley
station
nutrient

Keywords

  • Copula model
  • Discharge
  • Lower Illinois river
  • Nitrate concentration

ASJC Scopus subject areas

  • Renewable Energy, Sustainability and the Environment
  • Environmental Science(all)
  • Strategy and Management
  • Industrial and Manufacturing Engineering

Cite this

Uncertainty analysis of the relationship between discharge and nitrate concentration in the Lower Illinois River using a copula model. / Park, Daeryong; Markus, Momcilo; Jung, Kichul; Um, Myoung Jin.

In: Journal of Cleaner Production, Vol. 222, 10.06.2019, p. 310-323.

Research output: Contribution to journalArticle

@article{2eb0946ebc7a4587a5600249a17ee97d,
title = "Uncertainty analysis of the relationship between discharge and nitrate concentration in the Lower Illinois River using a copula model",
abstract = "The complex relationship between water quality and river discharge has long been an active area of research. This research applies bivariate statistics to river nitrate concentration and discharge, and develops a procedure for selecting the best bivariate distribution for the chosen models. Nitrate concentration (NO 3 ) data from five stations on the Lower Illinois River, USA, collected from 1972 to 2012 were used for demonstration of this procedure. To explore the statistical relationship between discharge and nitrate concentration, three copula statistical models were applied to determine optimal bivariate distributions. The generalized Pareto (GPA), generalized extreme value (GEV), and lognormal two- (LN2) and three-parameter (LN3) distributions were used as marginal distributions, and were combined with the Clayton, Frank, and Gumbel copula models. The probability plot correlation coefficient (PPCC) test and the S n statistic were used to select the most suitable copula models. As a result, the GPA distributions for streamflow and nitrate concentration in the Frank copula model were more accurate for representing the data from the two mainstem stations, Havana (D-31) and Valley City (D-32), and those from the Oakford station along the Sangamon River (E-25). However, the Clayton copula model using NL3 for representing streamflow and GPA for representing nitrate concentration was more accurate for stations along the Spoon (DJ-08) and La Moine Rivers (DG-01). This study enabled the development of a procedure for determining an optimal copula for application in the environmental sciences and also provided a case study for applying copula models using discharge and nutrient concentration as hydrologic and water quality variables, respectively.",
keywords = "Copula model, Discharge, Lower Illinois river, Nitrate concentration",
author = "Daeryong Park and Momcilo Markus and Kichul Jung and Um, {Myoung Jin}",
year = "2019",
month = "6",
day = "10",
doi = "10.1016/j.jclepro.2019.03.034",
language = "English (US)",
volume = "222",
pages = "310--323",
journal = "Journal of Cleaner Production",
issn = "0959-6526",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Uncertainty analysis of the relationship between discharge and nitrate concentration in the Lower Illinois River using a copula model

AU - Park, Daeryong

AU - Markus, Momcilo

AU - Jung, Kichul

AU - Um, Myoung Jin

PY - 2019/6/10

Y1 - 2019/6/10

N2 - The complex relationship between water quality and river discharge has long been an active area of research. This research applies bivariate statistics to river nitrate concentration and discharge, and develops a procedure for selecting the best bivariate distribution for the chosen models. Nitrate concentration (NO 3 ) data from five stations on the Lower Illinois River, USA, collected from 1972 to 2012 were used for demonstration of this procedure. To explore the statistical relationship between discharge and nitrate concentration, three copula statistical models were applied to determine optimal bivariate distributions. The generalized Pareto (GPA), generalized extreme value (GEV), and lognormal two- (LN2) and three-parameter (LN3) distributions were used as marginal distributions, and were combined with the Clayton, Frank, and Gumbel copula models. The probability plot correlation coefficient (PPCC) test and the S n statistic were used to select the most suitable copula models. As a result, the GPA distributions for streamflow and nitrate concentration in the Frank copula model were more accurate for representing the data from the two mainstem stations, Havana (D-31) and Valley City (D-32), and those from the Oakford station along the Sangamon River (E-25). However, the Clayton copula model using NL3 for representing streamflow and GPA for representing nitrate concentration was more accurate for stations along the Spoon (DJ-08) and La Moine Rivers (DG-01). This study enabled the development of a procedure for determining an optimal copula for application in the environmental sciences and also provided a case study for applying copula models using discharge and nutrient concentration as hydrologic and water quality variables, respectively.

AB - The complex relationship between water quality and river discharge has long been an active area of research. This research applies bivariate statistics to river nitrate concentration and discharge, and develops a procedure for selecting the best bivariate distribution for the chosen models. Nitrate concentration (NO 3 ) data from five stations on the Lower Illinois River, USA, collected from 1972 to 2012 were used for demonstration of this procedure. To explore the statistical relationship between discharge and nitrate concentration, three copula statistical models were applied to determine optimal bivariate distributions. The generalized Pareto (GPA), generalized extreme value (GEV), and lognormal two- (LN2) and three-parameter (LN3) distributions were used as marginal distributions, and were combined with the Clayton, Frank, and Gumbel copula models. The probability plot correlation coefficient (PPCC) test and the S n statistic were used to select the most suitable copula models. As a result, the GPA distributions for streamflow and nitrate concentration in the Frank copula model were more accurate for representing the data from the two mainstem stations, Havana (D-31) and Valley City (D-32), and those from the Oakford station along the Sangamon River (E-25). However, the Clayton copula model using NL3 for representing streamflow and GPA for representing nitrate concentration was more accurate for stations along the Spoon (DJ-08) and La Moine Rivers (DG-01). This study enabled the development of a procedure for determining an optimal copula for application in the environmental sciences and also provided a case study for applying copula models using discharge and nutrient concentration as hydrologic and water quality variables, respectively.

KW - Copula model

KW - Discharge

KW - Lower Illinois river

KW - Nitrate concentration

UR - http://www.scopus.com/inward/record.url?scp=85062848349&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062848349&partnerID=8YFLogxK

U2 - 10.1016/j.jclepro.2019.03.034

DO - 10.1016/j.jclepro.2019.03.034

M3 - Article

AN - SCOPUS:85062848349

VL - 222

SP - 310

EP - 323

JO - Journal of Cleaner Production

JF - Journal of Cleaner Production

SN - 0959-6526

ER -