Postprocessing: Difference between revisions
From Glossary of Meteorology
imported>Rbrandt (Created page with "== Postprocessing == <div class="definition"><div class="short_definition">The use of statistical techniques to transform [[numerical forecasting|numerical weather predic...") |
imported>Rbrandt |
||
| Line 3: | Line 3: | ||
<div class="definition"><div class="short_definition">The use of [[statistical]] techniques to transform [[numerical forecasting|numerical weather prediction]] ([[Nwp|NWP]]) [[output]] (possibly ensemble output) into a prediction of a meteorological [[variable]] or variables, often with the aim of improving meteorological guidance and decision-making. The transformation may incorporate data from sources outside the NWP (e.g., [[weather observation|observations]] or [[climatology]]) and can take many forms. For example, it may include a first-[[moment]] correction (i.e., bias correction) or a correction of higher moments (e.g., ensemble resampling or [[dispersion]] correction), or it may produce the [[weather forecast|forecast]] of a derived variable not explicitly part of the NWP output (e.g., [[fog]] or maximum [[wind]] over some period of time). Postprocessing can be a multistep process and involve a variety of statistical techniques (e.g., [[regression analysis]] or [[discriminant analysis]]). Most postprocessing techniques fall into the categories of [[model output statistics]] ([[MOS]]) and [[perfect prognostic|perfect prog]] (PP).</div><br/></div> | <div class="definition"><div class="short_definition">The use of [[statistical]] techniques to transform [[numerical forecasting|numerical weather prediction]] ([[Nwp|NWP]]) [[output]] (possibly ensemble output) into a prediction of a meteorological [[variable]] or variables, often with the aim of improving meteorological guidance and decision-making. The transformation may incorporate data from sources outside the NWP (e.g., [[weather observation|observations]] or [[climatology]]) and can take many forms. For example, it may include a first-[[moment]] correction (i.e., bias correction) or a correction of higher moments (e.g., ensemble resampling or [[dispersion]] correction), or it may produce the [[weather forecast|forecast]] of a derived variable not explicitly part of the NWP output (e.g., [[fog]] or maximum [[wind]] over some period of time). Postprocessing can be a multistep process and involve a variety of statistical techniques (e.g., [[regression analysis]] or [[discriminant analysis]]). Most postprocessing techniques fall into the categories of [[model output statistics]] ([[MOS]]) and [[perfect prognostic|perfect prog]] (PP).</div><br/></div> | ||
<p>''Term edited | <p>''Term edited 2 March 2020.''</p> | ||
{{TermIndex}} | {{TermIndex}} | ||
Latest revision as of 13:30, 2 March 2020
Postprocessing
The use of statistical techniques to transform numerical weather prediction (NWP) output (possibly ensemble output) into a prediction of a meteorological variable or variables, often with the aim of improving meteorological guidance and decision-making. The transformation may incorporate data from sources outside the NWP (e.g., observations or climatology) and can take many forms. For example, it may include a first-moment correction (i.e., bias correction) or a correction of higher moments (e.g., ensemble resampling or dispersion correction), or it may produce the forecast of a derived variable not explicitly part of the NWP output (e.g., fog or maximum wind over some period of time). Postprocessing can be a multistep process and involve a variety of statistical techniques (e.g., regression analysis or discriminant analysis). Most postprocessing techniques fall into the categories of model output statistics (MOS) and perfect prog (PP).
Term edited 2 March 2020.
