wind vectors change when satellite imagery displayed

[kbedka1]

See attached graphics, the plot with only yellow wind vectors are the "true wind vectors (derived from u and v)" at 250 mb from a GRIB file. The plot with the satellite image is the same wind vector field. The vectors looked the same as the the previous plot until I included the MSG 10.8 satellite image. The vectors became very small in size ( I tripled the size for this graphic) and the data changed such that the wind field became dominated by northerly winds instead of the "true" flow represented in the plot with the yellow vectors. Any idea as to what is going on here?

[joleenf]

Kris,

This sounded familiar, I pulled this explanation from Don Murray which was posted to the IDV user's email list

It's unwanted, but as HP notes, expected behavior. ... The problem is that the zoom factor is not taken into account when the barbs are first drawn. A similar problem was encountered with the point obs display, and the solution is to account for the zoom factor when calculating the scale that the user has chosen.

http://www.unidata.ucar.edu/mailing_lis ... 00022.html

Joleen

[kbedka1]

thanks, that explains the smaller vectors but not the incorrect data which is the more troubling issue here

[joleenf]

Kris,

This is speculation. I have noticed in scripts that the centerAndZoom command does not always work as expected. In fact, I am not sure if there is a concrete explanation for needing to call it three times for it to work in a script. I am wondering if there is a combination of the centerAndZoom partially working and the scale factor issue with the wind barbs. Could it be possible for the image to zoom, and not the gridded field, or for the gridded field to centerAndZoom to the wrong location?

Joleen

[kbedka1]

I noticed that the vectors are also wrong if I load the satellite image first and the vectors second but I have autoset projection turned off so the vectors are forced to project onto the satellite view. If I allow the projection to shift to the NWP model (i.e. autoset projection on) then the vectors look OK. So it is something with the satellite projection...

[Rick]

Hi Kris,

I agree that the bad data is definitely a bug. I created an inquiry for this to be resolved:

http://mcidas.ssec.wisc.edu/inquiry-v/?inquiry=1150