Flexible regression models over river networks

David O'Donnell, Alastair Rushworth, Adrian W. Bowman, E. Marian Scott, Mark Hallard

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

Many statistical models are available for spatial data but the vast majority of these assume that spatial separation can be measured by Euclidean distance. Data which are collected over river networks constitute a notable and commonly occurring exception, where distance must be measured along complex paths and, in addition, account must be taken of the relative flows of water into and out of confluences. Suitable models for this type of data have been constructed based on covariance functions. The aim of the paper is to place the focus on underlying spatial trends by adopting a regression formulation and using methods which allow smooth but flexible patterns. Specifically, kernel methods and penalized splines are investigated, with the latter proving more suitable from both computational and modelling perspectives. In addition to their use in a purely spatial setting, penalized splines also offer a convenient route to the construction of spatiotemporal models, where data are available over time as well as over space. Models which include main effects and spatiotemporal interactions, as well as seasonal terms and interactions, are constructed for data on nitrate pollution in the River Tweed. The results give valuable insight into the changes in water quality in both space and time.

LanguageEnglish
Pages47-63
Number of pages17
JournalJournal of the Royal Statistical Society: Series C
Volume63
Issue number1
Early online date11 Jul 2013
DOIs
Publication statusPublished - Jan 2014

Fingerprint

Regression Model
Penalized Splines
Spatio-temporal Model
Confluence
Covariance Function
Main Effect
Kernel Methods
Water Quality
Nitrate
Spatial Data
Euclidean Distance
Pollution
Interaction
Statistical Model
Exception
Regression
Water
Path
Rivers
Regression model

Keywords

  • flexible regression
  • kernels
  • network
  • penalized splines
  • smoothing
  • spatial separation
  • spatiotemporal models
  • water quality

Cite this

O'Donnell, D., Rushworth, A., Bowman, A. W., Marian Scott, E., & Hallard, M. (2014). Flexible regression models over river networks. Journal of the Royal Statistical Society: Series C, 63(1), 47-63. https://doi.org/10.1111/rssc.12024
O'Donnell, David ; Rushworth, Alastair ; Bowman, Adrian W. ; Marian Scott, E. ; Hallard, Mark. / Flexible regression models over river networks. In: Journal of the Royal Statistical Society: Series C. 2014 ; Vol. 63, No. 1. pp. 47-63.
@article{48749eab3744447bb75dd5921202cf69,
title = "Flexible regression models over river networks",
abstract = "Many statistical models are available for spatial data but the vast majority of these assume that spatial separation can be measured by Euclidean distance. Data which are collected over river networks constitute a notable and commonly occurring exception, where distance must be measured along complex paths and, in addition, account must be taken of the relative flows of water into and out of confluences. Suitable models for this type of data have been constructed based on covariance functions. The aim of the paper is to place the focus on underlying spatial trends by adopting a regression formulation and using methods which allow smooth but flexible patterns. Specifically, kernel methods and penalized splines are investigated, with the latter proving more suitable from both computational and modelling perspectives. In addition to their use in a purely spatial setting, penalized splines also offer a convenient route to the construction of spatiotemporal models, where data are available over time as well as over space. Models which include main effects and spatiotemporal interactions, as well as seasonal terms and interactions, are constructed for data on nitrate pollution in the River Tweed. The results give valuable insight into the changes in water quality in both space and time.",
keywords = "flexible regression, kernels, network, penalized splines, smoothing, spatial separation, spatiotemporal models, water quality",
author = "David O'Donnell and Alastair Rushworth and Bowman, {Adrian W.} and {Marian Scott}, E. and Mark Hallard",
year = "2014",
month = "1",
doi = "10.1111/rssc.12024",
language = "English",
volume = "63",
pages = "47--63",
journal = "Journal of the Royal Statistical Society: Series C",
issn = "0035-9254",
number = "1",

}

O'Donnell, D, Rushworth, A, Bowman, AW, Marian Scott, E & Hallard, M 2014, 'Flexible regression models over river networks' Journal of the Royal Statistical Society: Series C, vol. 63, no. 1, pp. 47-63. https://doi.org/10.1111/rssc.12024

Flexible regression models over river networks. / O'Donnell, David; Rushworth, Alastair; Bowman, Adrian W.; Marian Scott, E.; Hallard, Mark.

In: Journal of the Royal Statistical Society: Series C, Vol. 63, No. 1, 01.2014, p. 47-63.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Flexible regression models over river networks

AU - O'Donnell, David

AU - Rushworth, Alastair

AU - Bowman, Adrian W.

AU - Marian Scott, E.

AU - Hallard, Mark

PY - 2014/1

Y1 - 2014/1

N2 - Many statistical models are available for spatial data but the vast majority of these assume that spatial separation can be measured by Euclidean distance. Data which are collected over river networks constitute a notable and commonly occurring exception, where distance must be measured along complex paths and, in addition, account must be taken of the relative flows of water into and out of confluences. Suitable models for this type of data have been constructed based on covariance functions. The aim of the paper is to place the focus on underlying spatial trends by adopting a regression formulation and using methods which allow smooth but flexible patterns. Specifically, kernel methods and penalized splines are investigated, with the latter proving more suitable from both computational and modelling perspectives. In addition to their use in a purely spatial setting, penalized splines also offer a convenient route to the construction of spatiotemporal models, where data are available over time as well as over space. Models which include main effects and spatiotemporal interactions, as well as seasonal terms and interactions, are constructed for data on nitrate pollution in the River Tweed. The results give valuable insight into the changes in water quality in both space and time.

AB - Many statistical models are available for spatial data but the vast majority of these assume that spatial separation can be measured by Euclidean distance. Data which are collected over river networks constitute a notable and commonly occurring exception, where distance must be measured along complex paths and, in addition, account must be taken of the relative flows of water into and out of confluences. Suitable models for this type of data have been constructed based on covariance functions. The aim of the paper is to place the focus on underlying spatial trends by adopting a regression formulation and using methods which allow smooth but flexible patterns. Specifically, kernel methods and penalized splines are investigated, with the latter proving more suitable from both computational and modelling perspectives. In addition to their use in a purely spatial setting, penalized splines also offer a convenient route to the construction of spatiotemporal models, where data are available over time as well as over space. Models which include main effects and spatiotemporal interactions, as well as seasonal terms and interactions, are constructed for data on nitrate pollution in the River Tweed. The results give valuable insight into the changes in water quality in both space and time.

KW - flexible regression

KW - kernels

KW - network

KW - penalized splines

KW - smoothing

KW - spatial separation

KW - spatiotemporal models

KW - water quality

UR - http://www.scopus.com/inward/record.url?scp=84884306823&partnerID=8YFLogxK

UR - http://onlinelibrary.wiley.com/journal/10.1111/(ISSN)1467-9876

U2 - 10.1111/rssc.12024

DO - 10.1111/rssc.12024

M3 - Article

VL - 63

SP - 47

EP - 63

JO - Journal of the Royal Statistical Society: Series C

T2 - Journal of the Royal Statistical Society: Series C

JF - Journal of the Royal Statistical Society: Series C

SN - 0035-9254

IS - 1

ER -

O'Donnell D, Rushworth A, Bowman AW, Marian Scott E, Hallard M. Flexible regression models over river networks. Journal of the Royal Statistical Society: Series C. 2014 Jan;63(1):47-63. https://doi.org/10.1111/rssc.12024