Adaptive calibration of (u,v)‐wind ensemble forecasts

Publication: Research - peer-reviewJournal article – Annual report year: 2012

Documents

DOI

View graph of relations

Ensemble forecasts of (u,v)‐wind are of crucial importance for a number of decision‐making problems related to e.g. air traffic control, ship routeing and energy management. The skill of these ensemble forecasts as generated by NWP‐based models can be maximised by correcting for their lack of sufficient reliability. The original framework introduced here allows for an adaptive bivariate calibration of these ensemble forecasts. The originality of this methodology lies in the fact that calibrated ensembles still consist of a set of (space–time) trajectories, after translation and dilation. In parallel, the parameters of the models employed for improving the stochastic properties of the generating processes involved are adaptively and recursively estimated to accommodate smooth changes in the process characteristics and to lower computational costs. The approach is applied and evaluated based on the adaptive calibration of ECMWF ensemble forecasts of (u,v)‐wind at 10 m above ground level over Europe over a three‐year period between December 2006 and December 2009. Substantial improvements in (bivariate) reliability and in various deterministic/probabilistic scores are observed. Finally, the maps of translation and dilation factors are discussed. Copyright © 2012 Royal Meteorological Society
Original languageEnglish
JournalRoyal Meteorological Society. Quarterly Journal
Publication date2012
Volume138
Issue666
Pages1273-1284
ISSN0035-9009
DOIs
StatePublished
CitationsWeb of Science® Times Cited: 9

Keywords

  • Ensemble prediction, Probabilistic calibration, Bivariate processes, Recursive estimation, Near-surface wind
Download as:
Download as PDF
Select render style:
APAAuthorCBEHarvardMLAStandardVancouverShortLong
PDF
Download as HTML
Select render style:
APAAuthorCBEHarvardMLAStandardVancouverShortLong
HTML
Download as Word
Select render style:
APAAuthorCBEHarvardMLAStandardVancouverShortLong
Word

Download statistics

No data available

ID: 6380594