AMVs

by Mary Forsythe — last modified Feb 08, 2012 09:50 AM

• Derivation
• Advice for research users
• Real-time plots
• Monitoring
• NWP usage and impact
• Other users
  

• AMV example

AMVs are produced by tracking clouds or areas of water vapour in a sequence of satellite images. Traditionally geostationary imagery was used due to the frequent viewing of the same area of the Earth's atmosphere, but more recently polar orbiter imagery (MODIS and AVHRR) has also been used to provide AMVs in the polar regions, where the successive overpasses overlap.

AMVs are produced by tracking in several channels including the infrared window at 11 μm (IR), the WV absorption (WV), the visible (VIS) and the infrared 3.9 μm. The main derivation steps are:

1. correct and rectify the raw data
2. locate a suitable tracer within the image
3. perform a cross-correlation to locate the same feature in an earlier or later image
4. calculate the vector from the displacement in tracer location
5. assign a height to the vector
   

1. perform quality control.

The final AMV is an average of two or three component vectors calculated from a sequence of three of four images. For further information on AMV derivation, see Schmetz et al. (1993, JAM 32 p1206-1225) and Nieman et al. (1997, BAMS 78 p1121-1133).

AMVs are routinely produced by several centres including: EUMETSAT, NESDIS, CIMSS, JMA, IMD, CMA, KMA, CPTEC and BoM. Some links to further details are provided below:

Following discussion at IWW10, we plan to include a new wiki page to highlight where we may benefit from further work / inter-comparisons.  This is likely to involve:

• summarising the main AMV derivation steps e.g. target selection, tracking etc.
• for each one, identifying the different methods that have been tested e.g. for tracking: optical flow, Euclidean distance, cross-correlation - providing details as necessary
• summarising whether the different approaches have been well-compared in the past (reference as appropriate) or it they would benefit from further inter-comparison to determine best practise.  Also identify whether we would benefit from further research to develop new approaches (e.g. if thought to be an error-prone step).
  

This section will be expanded, as agreed at IWW10, to provide more guidance for research users who are interested in accessing the AMV data or developing the AMV derivation.

Derivation software

The EUMETSAT-funded NWCSAF have developed a portable software package which includes all the source code and libraries needed to preprocess input data (MSG SEVIRI images and NWP input), calculate AMVs and write output to BUFR or HDF5 files. The source code is written in C, with some Fortran functions.  Currently this can be used to produce MSG IR and HRVIS AMVs, but the intention is to extend to the WV channels (~2012) and (funding-dependent) other satellites (~2015).

To receive the NWCSAF software, you need to register as a user by applying through mafernandeza@aemet.es or gsanchezf@aemet.es. You will receive a user/password and have access to the Helpdesk restricted area, where the software can be downloaded and all the NWCSAF technical and  scientific documentation can be reached.  The installation of NWCSAF software is very easy. A test dataset is provided so that the user can verify they are doing everything correctly.
 
Further information about the NWCSAF project and HRW  product is available via the NWCSAF webpage including:
   - A detailed description of the HRW algorithm
   - A real time display of HRW output, so that the user can have a clear idea of what the product calculates and can compare his own local version with the official Helpdesk output.
 

Visualisation software
A useful tool for the visualisation of satellite imagery and products is the University of Wisconsin developed McIDAS software package.   For further details and free access to the software visit the McIDAS website.

Plots of AMVs, often overlain on satellite imagery, are provided from a number of centres. These can be useful for forecasters and as a tool for better understanding the AMV data.

• UW/CIMSS Tropical Cyclones and Real Time Winds
• UW/CIMSS Polar Winds from MODIS and AVHRR
• NOAA/NESDIS Operational Satellite Derived Winds
• NOAA/NESDIS/STAR/OPDB Experimental Satellite Derived Winds

Monitoring is very important to identify real-time problems e.g. data outages, changes in data quality.  More tailored statistical analysis can be used to help characterise the AMV errors.  This is an important activity for AMVs, which tend to have quite complicated errors. 

• NWP SAF AMV monitoring - provides 3 years of comparable monthly O-B monitoring from the Met Office and ECMWF, together with biennial analysis reports, results of one-off investigations and information on how AMV data is used in NWP.  Useful for better understanding sources of error in the AMV data
• ECMWF real-time monitoring - useful for identifying real time problems with the AMV data.

For information on NWP usage see the NWP SAF AMV web pageshttp://research.metoffice.gov.uk/research/interproj/nwpsaf/satwind_report/amvinfo

For information on NWP impact see the winds impact study under the IWWG activities section. 

• AMVs can be used directly by forecasters, normally by overlaying on satellite imagery or model fields.   They may be particularly helpful in meteorologically interesting areas (e.g. tropical cyclones….) or in cases of increased NWP model uncertainty.  
• AMVs can be used as a research tool to better understand cloud processes and ….