Review article: Observations for high-impact weather and their use in verification

Chiara Marsigli, Elizabeth Ebert, Raghavendra Ashrit, Barbara Casati, Jing Chen, Caio Coelho, Manfred Dorninger, Eric Gilleland, Thomas Haiden, Stephanie Landman, Marion P. Mittermaier

Publications: Contribution to journalReviewPeer Reviewed

Abstract

Verification of forecasts and warnings of high-impact weather is needed by the meteorological centres, but how to perform it still presents many open questions, starting from which data are suitable as reference. This paper reviews new observations which can be considered for the verification of high-impact weather and provides advice for their usage in objective verification. Two high-impact weather phenomena are considered: thunderstorm and fog. First, a framework for the verification of high-impact weather is proposed, including the definition of forecast and observations in this context and creation of a verification set. Then, new observations showing a potential for the detection and quantification of high-impact weather are reviewed, including remote sensing datasets, products developed for nowcasting, datasets derived from telecommunication systems, data collected from citizens, reports of impacts and claim/damage reports from insurance companies. The observation characteristics which are relevant for their usage in forecast verification are also discussed. Examples of forecast evaluation and verification are then presented, highlighting the methods which can be adopted to address the issues posed by the usage of these non-conventional observations and objectively quantify the skill of a high-impact weather forecast.
Original languageEnglish
Pages (from-to)1297–1312
Number of pages16
JournalNatural Hazards and Earth System Sciences
Volume21
Issue number4
DOIs
Publication statusPublished - 26 Apr 2021

Austrian Fields of Science 2012

  • 105206 Meteorology

Keywords

  • FOG
  • FORECAST VERIFICATION
  • MODEL
  • OBJECT-BASED VERIFICATION
  • SATELLITE
  • THUNDERSTORMS

Cite this