Spatial Ensemble Post-Processing with Standardized Anomalies
Markus Dabernig (),
Georg J. Mayr (),
Jakob W. Messner () and
Achim Zeileis ()
Working Papers from Faculty of Economics and Statistics, University of Innsbruck
To post-process ensemble predictions to a particular location, often statistical methods are used, especially in complex terrain such as the Alps. When expanded to several stations, the post-processing has to be repeated at every station individually thus losing information about spatial coherence and increasing computational cost. Therefore, we transform observations and predictions to standardized anomalies. Site- and season-specific characteristics are eliminated by subtracting a climatological mean and dividing by the climatological standard deviation from both observations and numerical forecasts. Then ensemble post-processing can be applied simultaneously at multiple locations. Furthermore, this method allows to forecast even at locations where no observations are available. The skill of these forecasts is comparable to forecasts post-processed individually at every station, and even better on average.
Keywords: statistical post-processing; ensemble post-processing; spatial; temperature; standardized anomalies; climatology; generalized additive model (search for similar items in EconPapers)
JEL-codes: C53 C61 Q50 (search for similar items in EconPapers)
Pages: 18 pages
New Economics Papers: this item is included in nep-for
References: View references in EconPapers View complete reference list from CitEc
Citations: Track citations by RSS feed
Downloads: (external link)
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
Persistent link: https://EconPapers.repec.org/RePEc:inn:wpaper:2016-08
Access Statistics for this paper
More papers in Working Papers from Faculty of Economics and Statistics, University of Innsbruck Contact information at EDIRC.
Bibliographic data for series maintained by Janette Walde ().