Modeling moisture fluxes using artificial neural networks: Can information extraction overcome data loss?

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Eddy covariance sites can experience data losses as high as 30 to 45% on an annual basis. Artificial neural networks (ANNs) have been identified as powerful tools for gap filling, but their performance depends on the representativeness of data used to train the model. In this paper, we develop a normalization method, which has similar performance compared to conventional training approaches, but exhibits differences in the timing of fluxes, indicating different and previously unused information in the data record. Specifically, the differences between half-hourly model fluxes, especially during summer months, indicate that the structure of the information content in the data changes seasonally, diurnally and with the rate of data loss. Extracting more information from data may not improve model performance and indicates the need for improved data and models to address flux behavior at critical times. We advise several approaches to address these concerns, including use of separate models for day and nighttime processes and the use of alternate data streams at dawn, when eddy covariance may be particularly ineffective due to the timing of the onset of turbulent mixing.

Original languageEnglish (US)
Pages (from-to)359-368
Number of pages10
JournalHydrology and Earth System Sciences
Volume15
Issue number1
DOIs
StatePublished - 2011

Fingerprint

moisture flux
artificial neural network
modeling
eddy covariance
loss
turbulent mixing
train
summer

ASJC Scopus subject areas

  • Earth and Planetary Sciences (miscellaneous)
  • Water Science and Technology

Cite this

@article{bea7d01a253a4f26b27fd0b256919d35,
title = "Modeling moisture fluxes using artificial neural networks: Can information extraction overcome data loss?",
abstract = "Eddy covariance sites can experience data losses as high as 30 to 45{\%} on an annual basis. Artificial neural networks (ANNs) have been identified as powerful tools for gap filling, but their performance depends on the representativeness of data used to train the model. In this paper, we develop a normalization method, which has similar performance compared to conventional training approaches, but exhibits differences in the timing of fluxes, indicating different and previously unused information in the data record. Specifically, the differences between half-hourly model fluxes, especially during summer months, indicate that the structure of the information content in the data changes seasonally, diurnally and with the rate of data loss. Extracting more information from data may not improve model performance and indicates the need for improved data and models to address flux behavior at critical times. We advise several approaches to address these concerns, including use of separate models for day and nighttime processes and the use of alternate data streams at dawn, when eddy covariance may be particularly ineffective due to the timing of the onset of turbulent mixing.",
author = "Neal, {A. L.} and Gupta, {Hoshin Vijai} and Shirley Papuga and Paul Brooks",
year = "2011",
doi = "10.5194/hess-15-359-2011",
language = "English (US)",
volume = "15",
pages = "359--368",
journal = "Hydrology and Earth System Sciences",
issn = "1027-5606",
publisher = "European Geosciences Union",
number = "1",

}

TY - JOUR

T1 - Modeling moisture fluxes using artificial neural networks

T2 - Can information extraction overcome data loss?

AU - Neal, A. L.

AU - Gupta, Hoshin Vijai

AU - Papuga, Shirley

AU - Brooks, Paul

PY - 2011

Y1 - 2011

N2 - Eddy covariance sites can experience data losses as high as 30 to 45% on an annual basis. Artificial neural networks (ANNs) have been identified as powerful tools for gap filling, but their performance depends on the representativeness of data used to train the model. In this paper, we develop a normalization method, which has similar performance compared to conventional training approaches, but exhibits differences in the timing of fluxes, indicating different and previously unused information in the data record. Specifically, the differences between half-hourly model fluxes, especially during summer months, indicate that the structure of the information content in the data changes seasonally, diurnally and with the rate of data loss. Extracting more information from data may not improve model performance and indicates the need for improved data and models to address flux behavior at critical times. We advise several approaches to address these concerns, including use of separate models for day and nighttime processes and the use of alternate data streams at dawn, when eddy covariance may be particularly ineffective due to the timing of the onset of turbulent mixing.

AB - Eddy covariance sites can experience data losses as high as 30 to 45% on an annual basis. Artificial neural networks (ANNs) have been identified as powerful tools for gap filling, but their performance depends on the representativeness of data used to train the model. In this paper, we develop a normalization method, which has similar performance compared to conventional training approaches, but exhibits differences in the timing of fluxes, indicating different and previously unused information in the data record. Specifically, the differences between half-hourly model fluxes, especially during summer months, indicate that the structure of the information content in the data changes seasonally, diurnally and with the rate of data loss. Extracting more information from data may not improve model performance and indicates the need for improved data and models to address flux behavior at critical times. We advise several approaches to address these concerns, including use of separate models for day and nighttime processes and the use of alternate data streams at dawn, when eddy covariance may be particularly ineffective due to the timing of the onset of turbulent mixing.

UR - http://www.scopus.com/inward/record.url?scp=79551503552&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79551503552&partnerID=8YFLogxK

U2 - 10.5194/hess-15-359-2011

DO - 10.5194/hess-15-359-2011

M3 - Article

AN - SCOPUS:79551503552

VL - 15

SP - 359

EP - 368

JO - Hydrology and Earth System Sciences

JF - Hydrology and Earth System Sciences

SN - 1027-5606

IS - 1

ER -