Forecast verification for multicategory discrete predictands related to the Common Air Quality Index

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00020699%3A_____%2F11%3A%230000436" target="_blank" >RIV/00020699:_____/11:#0000436 - isvavai.cz</a>

Result on the web

—

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Forecast verification for multicategory discrete predictands related to the Common Air Quality Index

Original language description

When forecast concentrations of modelled substances are used to calculate an air quality index, such as the Common Air Quality Index (CAQI), the forecast index classes must be validated. Here, we focus on the Peirce Skill Score, the Heidke Skill Score and the Gandian-Murphy Skill Score which allow assessment of how well CAQI classes were forecast by a statistically adapted forecast model. The techniques are applied to 70 air quality forecasts at European cities produced within the CITEAIR II project. The value of the statistical measures, to the verification process, is discussed. A key effect of using the scores is that a loss of information occurs by distilling the data into one value and this can disguise some aspects of forecast skill. The Gandian-Murphy Skill Score was found have the potential to lead to non-intuitive skill conclusions which may be unsuitable for air quality forecast verification.

Czech name

—

Czech description

—

Type

A - Audiovisual production

CEP classification

DG - Atmospheric sciences, meteorology

OECD FORD branch

—

Project

—

Continuities

V - Vyzkumna aktivita podporovana z jinych verejnych zdroju

Publication year

2011

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

ISBN

—

Place of publication

Praha, Česká republika

Publisher/client name

—

Version

—

Carrier ID

—