Initial perturbations based on the Ensemble Transform (ET) technique in the NCEP Global Ensemble Forecast System
Advanced Search
Select up to three search categories and corresponding keywords using the fields to the right. Refer to the Help section for more detailed instructions.

Search our Collections & Repository

All these words:

For very narrow results

This exact word or phrase:

When looking for a specific result

Any of these words:

Best used for discovery & interchangable words

None of these words:

Recommended to be used in conjunction with other fields



Publication Date Range:


Document Data


Document Type:






Clear All

Query Builder

Query box

Clear All

For additional assistance using the Custom Query please check out our Help Page


Initial perturbations based on the Ensemble Transform (ET) technique in the NCEP Global Ensemble Forecast System

Filetype[PDF-522.58 KB]


  • Alternative Title:
    NCEP Global Ensemble Forecast System
  • Description:
    Limitations on the initial perturbations used in the global operational ensemble forecast system at the National Centers for Environmental Prediction (NCEP) include the use of a climatologically fixed estimate of the analysis error variance and nonorthogonal paired bred vectors. In order to address these shortcomings, we introduced initial perturbations generated by the Ensemble Transform (ET) and ET with rescaling methods and compared them with the breeding ensemble in the NCEP operational environment. Both ET and ET with rescaling are second generation methods and generate initial perturbations that are consistent with the operational data assimilation (DA) system. The best possible initial analysis error variance from DA is used to restrain the initial perturbations that are orthogonal with respect to an inverse analysis error variance norm. In addition, a simplex transformation (ST) is imposed to ensure that the initial perturbations are centered and span a subspace with the maximum number of degrees of freedom. The variance is maintained in as many directions as possible within the ensemble subspace. It is shown that the perturbations are uniformly centered and distributed in different directions. The more ensemble members we have, the more orthogonal the perturbations will become. In the limit of infinite number of ensemble members, the perturbations will be orthogonal. Results show that the overall difference is not large although ET with rescaling performs best in almost all probabilistic scores and in terms of the forecast error explained by the perturbations. The forecast error variance can be explained best by pure ET with ST, which also has the highest time consistency between the analysis and forecast perturbations. The anomaly correlation of the ensemble mean from the breeding ensemble is slightly higher than that of the others for longer forecast lead times. A new one-sided breeding, which is centered by removing the mean of all perturbations, is tested for the same experimental period. It shows higher probabilistic scores than the paired ensemble. It is also found that the onesided bred perturbations span a high number of degrees of freedom and show a strong high time consistency.
  • Content Notes:
    Mozheng Wei, Zoltan Toth, Richard Wobus, and Yuejian Zhu.

    "September 19, 2006."

    Title from online resource (viewed on May 23, 2016).

    System requirements: Adobe Acrobat Reader.

    Includes bibliographical references (pages 17-19).

  • Document Type:
  • Place as Subject:
  • Rights Information:
    Public Domain
  • Compliance:
  • Main Document Checksum:
  • File Type:

Supporting Files

  • No Additional Files

More +

You May Also Like

Checkout today's featured content at

Version 3.26