Reduced rank covariances for the analysis of environmental data

Orietta Nicolis, Doug Nychka

Resultado de la investigación: Chapter

Resumen

In this work we propose a Monte Carlo estimator for non stationary covariances of large incomplete lattice or irregularly distributed data. In particular, we propose a method called “reduced rank covariance” (RRC), based on the multiresolution approach for reducing the dimensionality of the spatial covariances. The basic idea is to estimate the covariance on a lower resolution grid starting from a stationary model (such as the Mathérn covariance) and use the multiresolution property of wavelet basis for evaluating the covariance on the full grid. Since this method doesn’t need to compute the wavelet coefficients, it is very fast in estimating covariances in large data sets. The spatial forecasting performances of the method has been described through a simulation study. Finally, the method has been applied to two environmental data sets: the aerosol optical thickness (AOT) satellite data observed in Northern Italy and the ozone concentrations in the eastern United States.

Idioma originalEnglish
Título de la publicación alojadaStudies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies
EditorialSpringer International Publishing AG
Páginas253-263
Número de páginas11
DOI
EstadoPublished - 1 ene 2012

Serie de la publicación

NombreStudies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies
ISSN (versión impresa)2194-7767
ISSN (versión digital)2194-7775

Huella dactilar

Reduced Rank
Multiresolution
Grid
Ozone
Wavelet Bases
Wavelet Coefficients
Aerosol
Large Data Sets
Dimensionality
Forecasting
Simulation Study
Estimator

ASJC Scopus subject areas

  • Statistics and Probability

Citar esto

Nicolis, O., & Nychka, D. (2012). Reduced rank covariances for the analysis of environmental data. En Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies (pp. 253-263). (Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies). Springer International Publishing AG. https://doi.org/10.1007/978-3-642-21037-2_23
Nicolis, Orietta ; Nychka, Doug. / Reduced rank covariances for the analysis of environmental data. Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies. Springer International Publishing AG, 2012. pp. 253-263 (Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies).
@inbook{1949d3bec32748739166063a27d716b7,
title = "Reduced rank covariances for the analysis of environmental data",
abstract = "In this work we propose a Monte Carlo estimator for non stationary covariances of large incomplete lattice or irregularly distributed data. In particular, we propose a method called “reduced rank covariance” (RRC), based on the multiresolution approach for reducing the dimensionality of the spatial covariances. The basic idea is to estimate the covariance on a lower resolution grid starting from a stationary model (such as the Math{\'e}rn covariance) and use the multiresolution property of wavelet basis for evaluating the covariance on the full grid. Since this method doesn’t need to compute the wavelet coefficients, it is very fast in estimating covariances in large data sets. The spatial forecasting performances of the method has been described through a simulation study. Finally, the method has been applied to two environmental data sets: the aerosol optical thickness (AOT) satellite data observed in Northern Italy and the ozone concentrations in the eastern United States.",
keywords = "Aerosol optical thickness, Conditional covariance, Gaussian random field, Local solar time, Ozone concentration",
author = "Orietta Nicolis and Doug Nychka",
year = "2012",
month = "1",
day = "1",
doi = "10.1007/978-3-642-21037-2_23",
language = "English",
series = "Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies",
publisher = "Springer International Publishing AG",
pages = "253--263",
booktitle = "Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies",
address = "Switzerland",

}

Nicolis, O & Nychka, D 2012, Reduced rank covariances for the analysis of environmental data. En Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies. Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies, Springer International Publishing AG, pp. 253-263. https://doi.org/10.1007/978-3-642-21037-2_23

Reduced rank covariances for the analysis of environmental data. / Nicolis, Orietta; Nychka, Doug.

Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies. Springer International Publishing AG, 2012. p. 253-263 (Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies).

Resultado de la investigación: Chapter

TY - CHAP

T1 - Reduced rank covariances for the analysis of environmental data

AU - Nicolis, Orietta

AU - Nychka, Doug

PY - 2012/1/1

Y1 - 2012/1/1

N2 - In this work we propose a Monte Carlo estimator for non stationary covariances of large incomplete lattice or irregularly distributed data. In particular, we propose a method called “reduced rank covariance” (RRC), based on the multiresolution approach for reducing the dimensionality of the spatial covariances. The basic idea is to estimate the covariance on a lower resolution grid starting from a stationary model (such as the Mathérn covariance) and use the multiresolution property of wavelet basis for evaluating the covariance on the full grid. Since this method doesn’t need to compute the wavelet coefficients, it is very fast in estimating covariances in large data sets. The spatial forecasting performances of the method has been described through a simulation study. Finally, the method has been applied to two environmental data sets: the aerosol optical thickness (AOT) satellite data observed in Northern Italy and the ozone concentrations in the eastern United States.

AB - In this work we propose a Monte Carlo estimator for non stationary covariances of large incomplete lattice or irregularly distributed data. In particular, we propose a method called “reduced rank covariance” (RRC), based on the multiresolution approach for reducing the dimensionality of the spatial covariances. The basic idea is to estimate the covariance on a lower resolution grid starting from a stationary model (such as the Mathérn covariance) and use the multiresolution property of wavelet basis for evaluating the covariance on the full grid. Since this method doesn’t need to compute the wavelet coefficients, it is very fast in estimating covariances in large data sets. The spatial forecasting performances of the method has been described through a simulation study. Finally, the method has been applied to two environmental data sets: the aerosol optical thickness (AOT) satellite data observed in Northern Italy and the ozone concentrations in the eastern United States.

KW - Aerosol optical thickness

KW - Conditional covariance

KW - Gaussian random field

KW - Local solar time

KW - Ozone concentration

UR - http://www.scopus.com/inward/record.url?scp=85060273924&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-21037-2_23

DO - 10.1007/978-3-642-21037-2_23

M3 - Chapter

AN - SCOPUS:85060273924

T3 - Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies

SP - 253

EP - 263

BT - Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies

PB - Springer International Publishing AG

ER -

Nicolis O, Nychka D. Reduced rank covariances for the analysis of environmental data. En Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies. Springer International Publishing AG. 2012. p. 253-263. (Studies in Theoretical and Applied Statistics, Selected Papers of the Statistical Societies). https://doi.org/10.1007/978-3-642-21037-2_23