Inferential ecosystem models, from network data to prediction

James S. Clark, Pankaj Agarwal, David M. Bell, Paul G Flikkema, Alan Gelfand, Xuanlong Nguyen, Eric Ward, Jun Yang

Research output: Contribution to journalArticle

22 Citations (Scopus)

Abstract

Recent developments suggest that predictive modeling could begin to play a larger role not only for data analysis, but also for data collection. We address the example of efficient wireless sensor networks, where inferential ecosystem models can be used to weigh the value of an observation against the cost of data collection. Transmission costs make observations "expensive"; networks will typically be deployed in remote locations without access to infrastructure (e.g., power). The capacity to sample intensively makes sensor networks valuable, but high-frequency data are informative only at specific times and locations. Sampling intervals will range from meters and seconds to landscapes and years, depending on the process, the current states of the system, the uncertainty about those states, and the perceived potential for rapid change. Given that intensive sampling is sometimes critical, but more often wasteful, how do we develop tools to control the measurement and transmission processes? We address the potential of data collection controlled and/or supplemented by inferential ecosystem models. In a given model, the value of an observation can be evaluated in terms of its contribution to estimates of state variables and important parameters. There will be more than one model applied to network data that will include as state variables water, carbon, energy balance, biogeochemistry, tree ecophysiology, and forest demographic processes. The value of an observation will depend on the application. Inference is needed to weigh the contributions against transmission cost. Network control must be dynamic and driven by models capable of learning about both the environment and the network. We discuss application of Bayesian inference to model data from a developing sensor network as a basis for controlling the measurement and transmission processes. Our examples involve soil moisture and sap flux, but we discuss broader application of the approach, including its implications for network design.

Original languageEnglish (US)
Pages (from-to)1523-1536
Number of pages14
JournalEcological Applications
Volume21
Issue number5
DOIs
StatePublished - Jul 2011

Fingerprint

ecosystem
prediction
sensor
cost
ecophysiology
network design
sampling
biogeochemistry
energy balance
soil moisture
learning
infrastructure
carbon
modeling
water

Keywords

  • Bayesian prediction
  • Carbon-energy-water balance
  • Ecosystem data
  • Research networks
  • Sensor networks

ASJC Scopus subject areas

  • Ecology

Cite this

Clark, J. S., Agarwal, P., Bell, D. M., Flikkema, P. G., Gelfand, A., Nguyen, X., ... Yang, J. (2011). Inferential ecosystem models, from network data to prediction. Ecological Applications, 21(5), 1523-1536. https://doi.org/10.1890/09-1212.1

Inferential ecosystem models, from network data to prediction. / Clark, James S.; Agarwal, Pankaj; Bell, David M.; Flikkema, Paul G; Gelfand, Alan; Nguyen, Xuanlong; Ward, Eric; Yang, Jun.

In: Ecological Applications, Vol. 21, No. 5, 07.2011, p. 1523-1536.

Research output: Contribution to journalArticle

Clark, JS, Agarwal, P, Bell, DM, Flikkema, PG, Gelfand, A, Nguyen, X, Ward, E & Yang, J 2011, 'Inferential ecosystem models, from network data to prediction', Ecological Applications, vol. 21, no. 5, pp. 1523-1536. https://doi.org/10.1890/09-1212.1
Clark JS, Agarwal P, Bell DM, Flikkema PG, Gelfand A, Nguyen X et al. Inferential ecosystem models, from network data to prediction. Ecological Applications. 2011 Jul;21(5):1523-1536. https://doi.org/10.1890/09-1212.1
Clark, James S. ; Agarwal, Pankaj ; Bell, David M. ; Flikkema, Paul G ; Gelfand, Alan ; Nguyen, Xuanlong ; Ward, Eric ; Yang, Jun. / Inferential ecosystem models, from network data to prediction. In: Ecological Applications. 2011 ; Vol. 21, No. 5. pp. 1523-1536.
@article{19a0a6f5b0b949ef8434a6fb9b759d9b,
title = "Inferential ecosystem models, from network data to prediction",
abstract = "Recent developments suggest that predictive modeling could begin to play a larger role not only for data analysis, but also for data collection. We address the example of efficient wireless sensor networks, where inferential ecosystem models can be used to weigh the value of an observation against the cost of data collection. Transmission costs make observations {"}expensive{"}; networks will typically be deployed in remote locations without access to infrastructure (e.g., power). The capacity to sample intensively makes sensor networks valuable, but high-frequency data are informative only at specific times and locations. Sampling intervals will range from meters and seconds to landscapes and years, depending on the process, the current states of the system, the uncertainty about those states, and the perceived potential for rapid change. Given that intensive sampling is sometimes critical, but more often wasteful, how do we develop tools to control the measurement and transmission processes? We address the potential of data collection controlled and/or supplemented by inferential ecosystem models. In a given model, the value of an observation can be evaluated in terms of its contribution to estimates of state variables and important parameters. There will be more than one model applied to network data that will include as state variables water, carbon, energy balance, biogeochemistry, tree ecophysiology, and forest demographic processes. The value of an observation will depend on the application. Inference is needed to weigh the contributions against transmission cost. Network control must be dynamic and driven by models capable of learning about both the environment and the network. We discuss application of Bayesian inference to model data from a developing sensor network as a basis for controlling the measurement and transmission processes. Our examples involve soil moisture and sap flux, but we discuss broader application of the approach, including its implications for network design.",
keywords = "Bayesian prediction, Carbon-energy-water balance, Ecosystem data, Research networks, Sensor networks",
author = "Clark, {James S.} and Pankaj Agarwal and Bell, {David M.} and Flikkema, {Paul G} and Alan Gelfand and Xuanlong Nguyen and Eric Ward and Jun Yang",
year = "2011",
month = "7",
doi = "10.1890/09-1212.1",
language = "English (US)",
volume = "21",
pages = "1523--1536",
journal = "Ecological Appplications",
issn = "1051-0761",
publisher = "Ecological Society of America",
number = "5",

}

TY - JOUR

T1 - Inferential ecosystem models, from network data to prediction

AU - Clark, James S.

AU - Agarwal, Pankaj

AU - Bell, David M.

AU - Flikkema, Paul G

AU - Gelfand, Alan

AU - Nguyen, Xuanlong

AU - Ward, Eric

AU - Yang, Jun

PY - 2011/7

Y1 - 2011/7

N2 - Recent developments suggest that predictive modeling could begin to play a larger role not only for data analysis, but also for data collection. We address the example of efficient wireless sensor networks, where inferential ecosystem models can be used to weigh the value of an observation against the cost of data collection. Transmission costs make observations "expensive"; networks will typically be deployed in remote locations without access to infrastructure (e.g., power). The capacity to sample intensively makes sensor networks valuable, but high-frequency data are informative only at specific times and locations. Sampling intervals will range from meters and seconds to landscapes and years, depending on the process, the current states of the system, the uncertainty about those states, and the perceived potential for rapid change. Given that intensive sampling is sometimes critical, but more often wasteful, how do we develop tools to control the measurement and transmission processes? We address the potential of data collection controlled and/or supplemented by inferential ecosystem models. In a given model, the value of an observation can be evaluated in terms of its contribution to estimates of state variables and important parameters. There will be more than one model applied to network data that will include as state variables water, carbon, energy balance, biogeochemistry, tree ecophysiology, and forest demographic processes. The value of an observation will depend on the application. Inference is needed to weigh the contributions against transmission cost. Network control must be dynamic and driven by models capable of learning about both the environment and the network. We discuss application of Bayesian inference to model data from a developing sensor network as a basis for controlling the measurement and transmission processes. Our examples involve soil moisture and sap flux, but we discuss broader application of the approach, including its implications for network design.

AB - Recent developments suggest that predictive modeling could begin to play a larger role not only for data analysis, but also for data collection. We address the example of efficient wireless sensor networks, where inferential ecosystem models can be used to weigh the value of an observation against the cost of data collection. Transmission costs make observations "expensive"; networks will typically be deployed in remote locations without access to infrastructure (e.g., power). The capacity to sample intensively makes sensor networks valuable, but high-frequency data are informative only at specific times and locations. Sampling intervals will range from meters and seconds to landscapes and years, depending on the process, the current states of the system, the uncertainty about those states, and the perceived potential for rapid change. Given that intensive sampling is sometimes critical, but more often wasteful, how do we develop tools to control the measurement and transmission processes? We address the potential of data collection controlled and/or supplemented by inferential ecosystem models. In a given model, the value of an observation can be evaluated in terms of its contribution to estimates of state variables and important parameters. There will be more than one model applied to network data that will include as state variables water, carbon, energy balance, biogeochemistry, tree ecophysiology, and forest demographic processes. The value of an observation will depend on the application. Inference is needed to weigh the contributions against transmission cost. Network control must be dynamic and driven by models capable of learning about both the environment and the network. We discuss application of Bayesian inference to model data from a developing sensor network as a basis for controlling the measurement and transmission processes. Our examples involve soil moisture and sap flux, but we discuss broader application of the approach, including its implications for network design.

KW - Bayesian prediction

KW - Carbon-energy-water balance

KW - Ecosystem data

KW - Research networks

KW - Sensor networks

UR - http://www.scopus.com/inward/record.url?scp=79960367631&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79960367631&partnerID=8YFLogxK

U2 - 10.1890/09-1212.1

DO - 10.1890/09-1212.1

M3 - Article

VL - 21

SP - 1523

EP - 1536

JO - Ecological Appplications

JF - Ecological Appplications

SN - 1051-0761

IS - 5

ER -