Physics
Scientific paper
May 1997
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=1997adspr..19..443p&link_type=abstract
Advances in Space Research, Volume 19, Issue 3, p. 443-446.
Physics
Scientific paper
A scheme is currently being devised at the UK Met. Office to identify cloud fields and synoptic scale cloud systems using pattern recognition techniques, in an effort to find new observational products which might be assimilated into numerical weather prediction (NWP) models. The scheme uses artificial neural networks at three different scale lengths to calculate appropriate spectral, textural, spatial and contextual features. This paper presents results from one of the networks, which uses simple spectral features to obtain a classification by airmass into one of clear, dynamic, shallow or deep convection. The network was trained using 2190 Meteosat visible and infrared samples taken during the course of 1994, validated from UK Met. Office analyses. A product from the airmass classifier is described, already in use in a trial operational version of a thunderstorm flood forecasting system. The assimilation of similar products into NWP models is discussed, including discriminators of cloud fields, synoptic scale cloud systems and objectively labelled parameters such as precipitation rate.
No associations
LandOfFree
New METEOSAT pattern recognition products for use in weather forecasting does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with New METEOSAT pattern recognition products for use in weather forecasting, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and New METEOSAT pattern recognition products for use in weather forecasting will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-832424