Decoding n_pctaudist in HGG-Basic

[Cole,Jason (ECCC)]

Hello,

 

  I seem to be misunderstanding how to decode the variable n_pctaudist in the HGG-Basic product. 

 

For the attached figure, I am using the file ISCCP-Basic.HGG.v01r00.GLOBAL.2005.12.31.0000.GPC.10KM.CS00.EA1.00.nc for which the metadata is:

 

        short n_pctaudist(time, levtau, levpc, lat, lon) ;

                n_pctaudist:long_name = "Cloudy pixel fraction in each cloud-top pressure (PC)/cloud optical depth (TAU) level" ;

                n_pctaudist:units = "1" ;

                n_pctaudist:valid_min = -500s ;

                n_pctaudist:valid_max = 500s ;

                n_pctaudist:description = "Cloud detected by either IR or VIS threshold, cloud top location adjusted for optically thinner clouds" ;

                n_pctaudist:standard_name = "isccp_cloud_area_fraction" ;

                n_pctaudist:add_offset = 50.f ;

                n_pctaudist:scale_factor = 0.1f ;

 

I would expect that the data should be decoded as 0.1*n_pctaudist+50.0.  However, when I do this I get the upper plot in the attached figure which seems incorrect, the night-time areas are ~40 and sunlit areas seem to have a background value of ~50.  I also tried decoding without the offset, 0.1*pctaudist, (lower plot) which seems more plausible, background of ~0 cloud fraction for sunlit areas and ~-10 for night-time areas (missing value?) but it is not consistent with the metadata.

 

It is possible to confirm how to decode this variable? 

 

On a related note, is there the equivalent of n_pctaudist in the monthly ISCCP data?  I could not find it. I only found the variable “cloud_type” which I assume is a reduced version of n_pctaudist and insufficient for my application.

 

Thanks,

 

Jason

[Ken Knapp - NOAA Federal]

Hi Jason

Sorry for the late reply.

It appears that the tool that you are using to read the data is already performing the conversion from stored integer data to floats. 

If you use a netCDF utility (like ncdump) to look at the data, the missing data has the form of -1500 (which results in missing value of -100.0 after being scale). Please look at the documentation of the utility that you are using as it seems the scale_factor and add_offset are already applied.

I have verified that the file mentioned:  ISCCP-Basic.HGG.v01r00.GLOBAL.2005.12.31.0000.GPC.10KM.CS00.EA1.00.nc is indeed correct. The resulting data (when read in) can be scaled using the variable attributes with valid values of cloud fraction (ranging from 0-100 with -100 being the missing data).

Hope this helps you and other users-

-Ken