Maximum likelihood estimation of error covariances in ensemble-based filters and its application to a coupled atmosphere-ocean model

Genta Ueno, Tomoyuki Higuchi, Takashi Kagimoto, Naoki Hirose

    Research output: Contribution to journalArticle

    14 Citations (Scopus)

    Abstract

    We propose a method for estimating optimal error covariances in the context of sequential assimilation, including the case where both the system equation and the observation equation are nonlinear. When the system equation is nonlinear, ensemble-based filtering methods such as the ensemble Kalman filter (EnKF) are widely used to deal directly with the nonlinearity. The present approach for covariance optimization is a maximum likelihood estimation carried out by approximating the likelihood with the ensemblemean. Specifically, the likelihood is approximated as the samplemean of the likelihood of each member of the ensemble. To evaluate the sampling error of the proposed ensemble-approximated likelihood, we construct a method for examining the statistical significance using the bootstrap method without extra ensemble computation. We apply the proposed methods to an EnKF experiment where TOPEX/POSEIDON altimetry observations are assimilated into an intermediate coupled model, which is nonlinear, and estimate the optimal parameters that specify the covariances of the system noise and observation noise. Using these optimal covariance parameters, we examine the estimates by the EnKF and the ensemble Kalman smoother (EnKS). The effect of smoothing decreases by 1/e approximately one year after the filtering step. One of the properties of the smoothed estimate is that westerly wind anomalies over the western Pacific are not reproduced around the period of an El Niño event, while those over the central Pacific are strengthened. From additional experiments, we find that (1) the westerly winds in the western Pacific are phenomena outside of the coupled model and are not necessary tomodel ElNiño, (2) themodel El Niño is maintained by thewesterlies over the central Pacific, and (3) the modelled evolution process of the sea-surface temperature (SST) requires improvement to reproduce the westerly winds over the western Pacific.

    Original languageEnglish
    Pages (from-to)1316-1343
    Number of pages28
    JournalQuarterly Journal of the Royal Meteorological Society
    Volume136
    Issue number650
    DOIs
    Publication statusPublished - Jul 1 2010

    Fingerprint

    Kalman filter
    westerly
    filter
    atmosphere
    ocean
    bootstrapping
    altimetry
    smoothing
    nonlinearity
    sea surface temperature
    experiment
    anomaly
    method
    sampling
    parameter

    All Science Journal Classification (ASJC) codes

    • Atmospheric Science

    Cite this

    Maximum likelihood estimation of error covariances in ensemble-based filters and its application to a coupled atmosphere-ocean model. / Ueno, Genta; Higuchi, Tomoyuki; Kagimoto, Takashi; Hirose, Naoki.

    In: Quarterly Journal of the Royal Meteorological Society, Vol. 136, No. 650, 01.07.2010, p. 1316-1343.

    Research output: Contribution to journalArticle

    @article{34d0d354a2c54442b3431c5b7ba35c0e,
    title = "Maximum likelihood estimation of error covariances in ensemble-based filters and its application to a coupled atmosphere-ocean model",
    abstract = "We propose a method for estimating optimal error covariances in the context of sequential assimilation, including the case where both the system equation and the observation equation are nonlinear. When the system equation is nonlinear, ensemble-based filtering methods such as the ensemble Kalman filter (EnKF) are widely used to deal directly with the nonlinearity. The present approach for covariance optimization is a maximum likelihood estimation carried out by approximating the likelihood with the ensemblemean. Specifically, the likelihood is approximated as the samplemean of the likelihood of each member of the ensemble. To evaluate the sampling error of the proposed ensemble-approximated likelihood, we construct a method for examining the statistical significance using the bootstrap method without extra ensemble computation. We apply the proposed methods to an EnKF experiment where TOPEX/POSEIDON altimetry observations are assimilated into an intermediate coupled model, which is nonlinear, and estimate the optimal parameters that specify the covariances of the system noise and observation noise. Using these optimal covariance parameters, we examine the estimates by the EnKF and the ensemble Kalman smoother (EnKS). The effect of smoothing decreases by 1/e approximately one year after the filtering step. One of the properties of the smoothed estimate is that westerly wind anomalies over the western Pacific are not reproduced around the period of an El Ni{\~n}o event, while those over the central Pacific are strengthened. From additional experiments, we find that (1) the westerly winds in the western Pacific are phenomena outside of the coupled model and are not necessary tomodel ElNi{\~n}o, (2) themodel El Ni{\~n}o is maintained by thewesterlies over the central Pacific, and (3) the modelled evolution process of the sea-surface temperature (SST) requires improvement to reproduce the westerly winds over the western Pacific.",
    author = "Genta Ueno and Tomoyuki Higuchi and Takashi Kagimoto and Naoki Hirose",
    year = "2010",
    month = "7",
    day = "1",
    doi = "10.1002/qj.654",
    language = "English",
    volume = "136",
    pages = "1316--1343",
    journal = "Quarterly Journal of the Royal Meteorological Society",
    issn = "0035-9009",
    publisher = "John Wiley and Sons Ltd",
    number = "650",

    }

    TY - JOUR

    T1 - Maximum likelihood estimation of error covariances in ensemble-based filters and its application to a coupled atmosphere-ocean model

    AU - Ueno, Genta

    AU - Higuchi, Tomoyuki

    AU - Kagimoto, Takashi

    AU - Hirose, Naoki

    PY - 2010/7/1

    Y1 - 2010/7/1

    N2 - We propose a method for estimating optimal error covariances in the context of sequential assimilation, including the case where both the system equation and the observation equation are nonlinear. When the system equation is nonlinear, ensemble-based filtering methods such as the ensemble Kalman filter (EnKF) are widely used to deal directly with the nonlinearity. The present approach for covariance optimization is a maximum likelihood estimation carried out by approximating the likelihood with the ensemblemean. Specifically, the likelihood is approximated as the samplemean of the likelihood of each member of the ensemble. To evaluate the sampling error of the proposed ensemble-approximated likelihood, we construct a method for examining the statistical significance using the bootstrap method without extra ensemble computation. We apply the proposed methods to an EnKF experiment where TOPEX/POSEIDON altimetry observations are assimilated into an intermediate coupled model, which is nonlinear, and estimate the optimal parameters that specify the covariances of the system noise and observation noise. Using these optimal covariance parameters, we examine the estimates by the EnKF and the ensemble Kalman smoother (EnKS). The effect of smoothing decreases by 1/e approximately one year after the filtering step. One of the properties of the smoothed estimate is that westerly wind anomalies over the western Pacific are not reproduced around the period of an El Niño event, while those over the central Pacific are strengthened. From additional experiments, we find that (1) the westerly winds in the western Pacific are phenomena outside of the coupled model and are not necessary tomodel ElNiño, (2) themodel El Niño is maintained by thewesterlies over the central Pacific, and (3) the modelled evolution process of the sea-surface temperature (SST) requires improvement to reproduce the westerly winds over the western Pacific.

    AB - We propose a method for estimating optimal error covariances in the context of sequential assimilation, including the case where both the system equation and the observation equation are nonlinear. When the system equation is nonlinear, ensemble-based filtering methods such as the ensemble Kalman filter (EnKF) are widely used to deal directly with the nonlinearity. The present approach for covariance optimization is a maximum likelihood estimation carried out by approximating the likelihood with the ensemblemean. Specifically, the likelihood is approximated as the samplemean of the likelihood of each member of the ensemble. To evaluate the sampling error of the proposed ensemble-approximated likelihood, we construct a method for examining the statistical significance using the bootstrap method without extra ensemble computation. We apply the proposed methods to an EnKF experiment where TOPEX/POSEIDON altimetry observations are assimilated into an intermediate coupled model, which is nonlinear, and estimate the optimal parameters that specify the covariances of the system noise and observation noise. Using these optimal covariance parameters, we examine the estimates by the EnKF and the ensemble Kalman smoother (EnKS). The effect of smoothing decreases by 1/e approximately one year after the filtering step. One of the properties of the smoothed estimate is that westerly wind anomalies over the western Pacific are not reproduced around the period of an El Niño event, while those over the central Pacific are strengthened. From additional experiments, we find that (1) the westerly winds in the western Pacific are phenomena outside of the coupled model and are not necessary tomodel ElNiño, (2) themodel El Niño is maintained by thewesterlies over the central Pacific, and (3) the modelled evolution process of the sea-surface temperature (SST) requires improvement to reproduce the westerly winds over the western Pacific.

    UR - http://www.scopus.com/inward/record.url?scp=77956199369&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=77956199369&partnerID=8YFLogxK

    U2 - 10.1002/qj.654

    DO - 10.1002/qj.654

    M3 - Article

    AN - SCOPUS:77956199369

    VL - 136

    SP - 1316

    EP - 1343

    JO - Quarterly Journal of the Royal Meteorological Society

    JF - Quarterly Journal of the Royal Meteorological Society

    SN - 0035-9009

    IS - 650

    ER -