Atmospheric motion vectors (AMV)

Atmospheric motion vectors are used to estimate horizontal winds by tracking cloud or water-vapor tracers across consecutive calibrated and navigated satellite imager scenes.^[1][2][3]

The method is an analysis layer on top of visible, infrared, or water-vapor radiance imagery. Operational processing selects trackable targets, matches them between image times, assigns a representative vector height, and applies quality indicators before the wind field is used in atmospheric analysis or numerical weather prediction.^[1][4]

The requirements envelope is temporal rather than hardware-bound: at least two time-ordered, well-navigated imager scenes are needed, and the scene pair must contain cloud or water-vapor structure that can be followed. If the tracer cannot be tracked or assigned a usable height, the methodology does not produce a dependable vector for that location.^[1]

Operational AMV production is not limited to one orbit class. Meteosat Second Generation provides a geostationary AMV implementation, while Metop AVHRR provides a polar-orbiting global implementation.^[5][6]

Because AMV depends on image sequences rather than a dedicated wind sensor, it is best treated as an analysis technique that consumes imager radiance products and returns speed, direction, representative height, and quality metadata.^[1]