LoneSky

GOES-16 and GOES-17 have no green band so one has to be created from the blue, red, and near-infrared channels. But no amount of simple blending of those channels will produce a truly accurate synthetic green band so additional processing—e.g. a look up table (LUT)—is usually necessary to further compensate. For demonstration purposes I'll be using mostly Himawari-8 imagery in this post. Himawari and GOES share the same imager with reasonably similar blue, red, and near-infrared bands. The big difference, of course, is Himawari includes a visible wavelength green filter (though at a wavelength that is not ideal for true color). GOES prioritizes a unique near-infrared filter in place of a green one. By combining blue, red, and near-infrared channels Himawari will produce approximately what GOES is capable of. This can then be compared to Himawari imagery using its RGB filters (with a little near-infrared added to boost the vegetation signal). Here is a commonly cited GOES synt...

This is a two-day color-enhanced composite from GOES-16 taken June 2-3, 2017 (17:45:00 UTC) with simple cloud reduction processing. The individual image bands were provided by SSEC RealEarth . Credit: NASA / NOAA / SSEC - Modified by LoneSky.

This is an approximately true color composite image of the contiguous United States as captured by GOES-16 on May 17, 2017 (17:15:00 UTC) and remapped to Mercator projection for SSEC RealEarth . Credit: NASA / NOAA / SSEC - Modified by LoneSky.