AMV Cross Correlation Matrices

Graeme Kelly, MetOffice@Reading, University of Reading

  • Javier García Pereda, NWCSAF/AEMET, Spain

This study explores the structure of the cross correlation matrices that are used to compute the initial trajectories at the beginning of the AMV process. A simple method based on identification and separation of the first two maxima. This gives a measure of the AMV quality.

This study used the AMV processing code from the NWCSAF package together with a large sample of AMV innervations from the MetOffice meso model for validation.