Ensemble modeling of transport and dispersion simulations guided by machine learning hypotheses generation

Andreas D. Lattner, Guido Cervone

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

In this article an approach is presented where machine learning classifiers are used to drive an ensemble modeling method of multiple atmospheric transport and dispersion simulations. The goal is to achieve a higher spread of the results with a lower number of ensemble simulations. Symbolic machine learning algorithms are used to define choices for the variation of meteorological input data, model parameters, model physics, based on their combined effects on the final dispersion calculations (i.e., construction of ensembles). The methodology uses an iterative approach with the aim to identify ensemble members leading to a more balanced distribution of results.The methodology is tested using real meteorological data from Istanbul, Turkey, simulating atmospheric releases along the Bosphorus channel. In an extensive evaluation, different settings of the approach are compared in a series of experiments. The results indicate that the desired effect of more balanced results of the ensemble members can be achieved by the approach.

Original languageEnglish (US)
Pages (from-to)267-279
Number of pages13
JournalComputers and Geosciences
Volume48
DOIs
StatePublished - Nov 1 2012

Fingerprint

Learning systems
Learning algorithms
modeling
simulation
Data structures
Classifiers
Physics
methodology
atmospheric transport
physics
Experiments
machine learning
experiment
effect

All Science Journal Classification (ASJC) codes

  • Information Systems
  • Computers in Earth Sciences

Cite this

@article{fe5525e3b4a04c3d83ea63af8fb4fb46,
title = "Ensemble modeling of transport and dispersion simulations guided by machine learning hypotheses generation",
abstract = "In this article an approach is presented where machine learning classifiers are used to drive an ensemble modeling method of multiple atmospheric transport and dispersion simulations. The goal is to achieve a higher spread of the results with a lower number of ensemble simulations. Symbolic machine learning algorithms are used to define choices for the variation of meteorological input data, model parameters, model physics, based on their combined effects on the final dispersion calculations (i.e., construction of ensembles). The methodology uses an iterative approach with the aim to identify ensemble members leading to a more balanced distribution of results.The methodology is tested using real meteorological data from Istanbul, Turkey, simulating atmospheric releases along the Bosphorus channel. In an extensive evaluation, different settings of the approach are compared in a series of experiments. The results indicate that the desired effect of more balanced results of the ensemble members can be achieved by the approach.",
author = "Lattner, {Andreas D.} and Guido Cervone",
year = "2012",
month = "11",
day = "1",
doi = "10.1016/j.cageo.2012.01.017",
language = "English (US)",
volume = "48",
pages = "267--279",
journal = "Computers and Geosciences",
issn = "0098-3004",
publisher = "Elsevier Limited",

}

Ensemble modeling of transport and dispersion simulations guided by machine learning hypotheses generation. / Lattner, Andreas D.; Cervone, Guido.

In: Computers and Geosciences, Vol. 48, 01.11.2012, p. 267-279.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Ensemble modeling of transport and dispersion simulations guided by machine learning hypotheses generation

AU - Lattner, Andreas D.

AU - Cervone, Guido

PY - 2012/11/1

Y1 - 2012/11/1

N2 - In this article an approach is presented where machine learning classifiers are used to drive an ensemble modeling method of multiple atmospheric transport and dispersion simulations. The goal is to achieve a higher spread of the results with a lower number of ensemble simulations. Symbolic machine learning algorithms are used to define choices for the variation of meteorological input data, model parameters, model physics, based on their combined effects on the final dispersion calculations (i.e., construction of ensembles). The methodology uses an iterative approach with the aim to identify ensemble members leading to a more balanced distribution of results.The methodology is tested using real meteorological data from Istanbul, Turkey, simulating atmospheric releases along the Bosphorus channel. In an extensive evaluation, different settings of the approach are compared in a series of experiments. The results indicate that the desired effect of more balanced results of the ensemble members can be achieved by the approach.

AB - In this article an approach is presented where machine learning classifiers are used to drive an ensemble modeling method of multiple atmospheric transport and dispersion simulations. The goal is to achieve a higher spread of the results with a lower number of ensemble simulations. Symbolic machine learning algorithms are used to define choices for the variation of meteorological input data, model parameters, model physics, based on their combined effects on the final dispersion calculations (i.e., construction of ensembles). The methodology uses an iterative approach with the aim to identify ensemble members leading to a more balanced distribution of results.The methodology is tested using real meteorological data from Istanbul, Turkey, simulating atmospheric releases along the Bosphorus channel. In an extensive evaluation, different settings of the approach are compared in a series of experiments. The results indicate that the desired effect of more balanced results of the ensemble members can be achieved by the approach.

UR - http://www.scopus.com/inward/record.url?scp=84866151975&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84866151975&partnerID=8YFLogxK

U2 - 10.1016/j.cageo.2012.01.017

DO - 10.1016/j.cageo.2012.01.017

M3 - Article

AN - SCOPUS:84866151975

VL - 48

SP - 267

EP - 279

JO - Computers and Geosciences

JF - Computers and Geosciences

SN - 0098-3004

ER -