satellite channel misalignment

[kbedka1]

I'm working with GOES-16 data acquired as a MUG member via ADDE and I've noticed that when I display the 0.64 micron and 10.4 micron channel data, the images are misaligned, i.e. the IR data is 2 visible pixels up and 4 pixels to the left of the corresponding visible channel feature. You can take any GOES-16 image, zoom in, and you will see this. I'd like to think that this is just a pre-operational GOES-16 issue, but this misalignment is occurring even for current GOES-13 and SEVIRI data...basically any time one displays imagery of different resolution. I thought I recall hearing that navigation is defined by pixel corners, not centers, but I could be wrong about this.

Bob Carp had provided a formula to enable shifting of the imagery by a user-specified lat/lon increment which is useful to correct this for cases here and there, especially for addressing parallax and getting severe weather reports or radar to align with satellite. But for example, when I create an RGB and seek to preserve the 0.5 km visible data, I would have to shift 2 of the 3 channels to get them to align with the visible. If one doesn't do this shift for an RGB, they will see weird artifacts caused by misalignment. This process will start to get frustrating when trying to look at many different cases and having to do this manual shift each and every time.

Are there plans to automatically correct for this issue internally in McIDAS-V, i.e. the software can automatically know to shift the data based on the sensor type and resolution? This would be incredibly helpful and is overdue.

[jayh]

Hi Kris-

I did some comparisons loading some realtime GVAR data (GOES-15) using -V and -X and see the similar shifts in both. I did various displays with lat/lon and line/ele data loads, center vs. upper left corner, etc. I don't see any inquiry for either software package. I'll send out my findings to the programmers and gather some information for you. Stay tuned...

Thanks, Jay