A snow event in southwest Kansas on 28/29 January 2020 led to widespread snow amounts of 4 to 8 inches, with locally higher amounts (above 12″). On the 29th, a thin layer of fog developed over the snow covered land. Inspect the GOES-16 imagery:
Upper left: Visible (0.64 micron) band
Upper right: CIRA Snow Cloud Layers RGB
Lower left: Day Snow Fog RGB
Lower right: Day Cloud Phase Distinction RGB
In the visible imagery, the fog is undetectable due to the lack of contrast (the snow cover and fog are the same color).
In the CIRA Snow Cloud Layers RGB, snow cover is white while low-cloud/fog is yellowish-green. Note how we can see through the fog since it is thin, the snow cover on the ground can be seen.
In the Day Snow Fog RGB, snow cover is red, low clouds / fog are light purple. The transparency in this product is less than the previous RGB so that the snow cover under the fog is somewhat more subtle.
In the Day Cloud Phase Distinction RGB, snow cover is green, low-cloud/fog is cyan and there is sufficient transparency to view the snow cover under the thin fog.
The key takeaway point is to make use of RGB products to discern fog from snow cover, the visible imagery alone makes it much more challenging to make this discrimination.
As an example of data fusion, note this tweet from the NWS WFO in Dodge City which confirms the fog via web-cams:
— NWS Dodge City (@NWSDodgeCity) January 29, 2020