MODIS MOD06 cloud top height = 0?

[captainjasper]

I am working on a project with MODIS MOD06_L2 data sets and I have noticed that in the cloud_top_height and cloud_top_height_1km data sets there are a number of pixels with cloud top heights of 0m. I assume that is some sort of error (especially since most of these pixels are located above (or should I say "below") the Galápagos Islands - my research area - and the elevation there already is above 0m a.s.l.) and I was wondering how this happened and how one is to treat these pixels. Just omit them or is there a way to still get some usable data out of these pixels?

Thanks a lot in advance for any help!
Jasper

[LAADSx_UserServices_M]

Dear Jasper:
In order to get actual values you need to apply the scale_factor and the add_offset for the parameter/SDS you are intersted in (e.g., cloud_top_height). The scale_factor and offset values are provided in the header metadata for each SDS/layer or in the product user guide. The following equation should be used to get to the real value:

value = scale_factor * ( stored_value - add_offset )

If you have already applied the equation and still think that the values are not accurate, send us the file name of the file that you think is showing in accurate values for us to investigate.

[captainjasper]

Hi,

Thanks for the quick response! For the cloud_top_height_1km data set the scale factor is 1 and there is no offset - I had checked that already, so I doubt that this is the source of the problem. The file I am working with is MOD06_L2.A2021156.1705.061.2021268214743.hdf, but I have noticed this issue with all the files I looked at so far (MOD06 daylight files from 2021-06-01 to 2021-06-30, showing the Galápagos Islands). Even the pixels that show values > 0m often have cloud top heights below the height of the DEM, which seems weird to me.

Thanks a lot in advance!
Jasper

[LAADSx_UserServices_M]

Just to let you know we sent your question to our MODIS Cloud product expert and waiting for the response. We will let you know as soon as we here back. We appreciate your patience.

[LAADSx_UserServices_M]

Dear Jasper:
Our Cloud SME just sent this:
"We have been carefully looking at this but haven’t yet reached a point where we have a clear answer for the user. Within the next few days we should be able to answer."

[LAADSx_UserServices_M]

Dear Jasper:
The Cloud_Top_Heights in the MOD06/MYD06 products are derived from mapping Cloud_Top_Pressure retrievals to heights using the GFS reanalysis Geopotential Heights. Even though there are Cloud_Top_Pressure data values in the data files it does not mean that all of the data are of good quality. For the 5km product one must look at the Qaulity_Assurance_5km product and decode bits 1-3 to this table:
0 Fill
1 Marginal Confidence
2 Good Confidence (not currently used)
3 Very Good Confidence
[Source https://atmosphere-imager.gsfc.nasa.gov/sites/default/files/ModAtmo/QA_Plan_C61_Master_2017_03_15.pdf]

Essentially I have found that all the high Cloud_Top_Pressure, that correspond to 0 value in the Cloud_Top_Heights are from the ‘Marginal Confidence’ values.

The 1km retrievals use a different algorithm on the back end, so I will look to see if there is a similar quality flag for the 1km cloud_top_heights. An easy method would be to simply filter out heights that are identically 0, or cloud_top_pressures that are greater than or a few millibar from the Surface_Pressure product also given in the MOD06 data files.

[captainjasper]

Hello,

Thanks for the response! I guess that I will then simply omit the 0m height pixels, but please let me know, if you find anything on the quality flags for the cloud_top_height_1km data set.

Have a great day!
Jasper