Earthdata Forum

When analysing the MOD06 cloud_top_height_1km data in the area of 89°W-92°W, 2°N-2°S, I encountered a large number of artefacts. The artefacts appear in the form of "patches" of approx. 1°x1° in size, where the cloud_top_height deviates significantly from each other.
These artefacts can be clearly seen, for example, at:

MOD06_L2.A2021260.1615.061.2021261020203.hdf
MOD06_L2.A2021278.1600.061.2021279015745.hdf

These artefacts appear in almost every scene from about September 2011 onwards, whereby the sea pixels seem to be more affected than the land pixels. I have also attached a plot of the average of the MOD06 cloud_top_height_1km data from 2021 as an example.

Where can these artefacts come from?
And is there a quality flag for this scene that I can use to find out which pixels are still usable?

Many thanks in advance!

Below is the response from the MOD06 product developer/SME:
"That is due to MOD_PRLCAT being unable to spatially interpolate ancillary data. The 1 degree boxes are because all relevant ancillary there is at 1 degree resolution. Not much we can do about that without doing some serious work on rewriting some of the core routines of MOD_PRLCAT. I could be mistaken, but I don’t think there is anyone in Wisconsin who would be willing/interested in tackling something like that at this time."

More from explanation product science team :
"This artifact in the 1-km CT height products relates to the input GDAS (1-degree) ancillary data. It may relate to an outlier pixel value or interpolation failure.
Unfortunately, no quality flag in the current code filters out these pixels. However, it does not affect the 5-km product, which is suggested to use instead over that area.
To understand the source of the blockiness, we need to investigate this artifact further, and we plan to fix it for the next collection (Collection 7). "