All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 709 297 14
PDF Downloads 457 172 9

Operational Cloud-Motion Winds from Meteosat Infrared Images

Johannes SchmetzEuropean Space Agency, European Space Operations Centre, Darmstadt, Germany

Search for other papers by Johannes Schmetz in
Current site
Google Scholar
PubMed
Close
,
Kenneth HolmlundEuropean Space Agency, European Space Operations Centre, Darmstadt, Germany

Search for other papers by Kenneth Holmlund in
Current site
Google Scholar
PubMed
Close
,
Joel HoffmanMETEO France, Ecole Nationale de la Météorogie, Toulouse, France

Search for other papers by Joel Hoffman in
Current site
Google Scholar
PubMed
Close
,
Bernard StraussEuropean Centre for Medium-Range Weather Forecasts, Reading, Berkshire, United Kingdom

Search for other papers by Bernard Strauss in
Current site
Google Scholar
PubMed
Close
,
Brian MasonEuropean Space Agency, European Space Operations Centre, Darmstadt, Germany

Search for other papers by Brian Mason in
Current site
Google Scholar
PubMed
Close
,
Volker GaertnerEuropean Space Agency, European Space Operations Centre, Darmstadt, Germany

Search for other papers by Volker Gaertner in
Current site
Google Scholar
PubMed
Close
,
Arno KochEuropean Space Agency, European Space Operations Centre, Darmstadt, Germany

Search for other papers by Arno Koch in
Current site
Google Scholar
PubMed
Close
, and
Leo Van De BergEuropean Space Agency, European Space Operations Centre, Darmstadt, Germany

Search for other papers by Leo Van De Berg in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The displacement of clouds in successive satellite images reflects the atmospheric circulation at various scales. The main application of the satellite-derived cloud-motion vectors is their use as winds in the data analysis for numerical weather prediction. At low latitudes in particular they constitute an indispensible data source for numerical weather prediction.

This paper describes the operational method of deriving cloud-motion winds (CMW) from the IR image (10.5–12.5 µm) of the European geostationary Meteostat satellites. The method is automatic, that is, the cloud tracking uses cross correlation and the height assignment is based on satellite observed brightness temperature and a forecast temperature profile. Semitransparent clouds undergo a height correction based on radiative forward calculations and simultaneous radiance observations in both the IR and water vapor (5.7–7.1 µm) channel. Cloud-motion winds are subject to various quality checks that include manual quality control as the last step. Typically about 3000 wind vectors are produced per day over four production cycles.

This paper documents algorithm changes and improvements made to the operational CMWs over the last five years. The improvements are shown by long-term comparisons with both collocated radiosondes and the first guess of the forecast model of the European Centre for Medium-Range Weather Forecasts. In particular, the height assignment of a wind vector and radiance filtering techniques preceding the cloud tracking have ameliorated the errors in Meteostat winds. The slow speed bias of high-level CMWs (<400 hPa) in comparison to radiosonde winds have been reduced from about 4 to 1.3 m s−1 for a mean wind speed of 24 m s−1. Correspondingly, the rms vectors error of Meteosat high-level CMWs decreased from about 7.8 to 5 m s−1. Medium- and low-level CMWs were also significantly improved.

Abstract

The displacement of clouds in successive satellite images reflects the atmospheric circulation at various scales. The main application of the satellite-derived cloud-motion vectors is their use as winds in the data analysis for numerical weather prediction. At low latitudes in particular they constitute an indispensible data source for numerical weather prediction.

This paper describes the operational method of deriving cloud-motion winds (CMW) from the IR image (10.5–12.5 µm) of the European geostationary Meteostat satellites. The method is automatic, that is, the cloud tracking uses cross correlation and the height assignment is based on satellite observed brightness temperature and a forecast temperature profile. Semitransparent clouds undergo a height correction based on radiative forward calculations and simultaneous radiance observations in both the IR and water vapor (5.7–7.1 µm) channel. Cloud-motion winds are subject to various quality checks that include manual quality control as the last step. Typically about 3000 wind vectors are produced per day over four production cycles.

This paper documents algorithm changes and improvements made to the operational CMWs over the last five years. The improvements are shown by long-term comparisons with both collocated radiosondes and the first guess of the forecast model of the European Centre for Medium-Range Weather Forecasts. In particular, the height assignment of a wind vector and radiance filtering techniques preceding the cloud tracking have ameliorated the errors in Meteostat winds. The slow speed bias of high-level CMWs (<400 hPa) in comparison to radiosonde winds have been reduced from about 4 to 1.3 m s−1 for a mean wind speed of 24 m s−1. Correspondingly, the rms vectors error of Meteosat high-level CMWs decreased from about 7.8 to 5 m s−1. Medium- and low-level CMWs were also significantly improved.

Save