Earthdata Forum

With MOD11 data from OPeNDAP, and with MOD21A1D HDF files from EarthData Search, I've been able to interpret values of the `QC` variable with bitwise operations and the appropriate user guide. In these cases, `QC` is provided as integers. But in the case of MOD21A1D data from OPeNDAP, `QC` is given as floating-point numbers (`Float32` in OPeNDAP parlance), and I can't see how it's meant to represent the bit patterns described in the user guide (Table 10 on pages 24-26 of [link] ://lpdaac.usgs.gov/documents/620/MOD21_User_Guide_V61.pdf ).

For example, in the file `MOD21A1D.A2003001.h12v04.061.2020084062741.hdf` I got from EarthData Search, the first five `QC` values are all 54 (hexadecimal 36). The corresponding values from OPeNDAP, available at [link] ://opendap.cr.usgs.gov/opendap/hyrax/MOD21A1D.061/h12v04.ncml.ascii?QC%5B766:1:766%5D%5B0:1:0%5D%5B0:1:4%5D , are all -0.509091 (hexadecimal 86b56fbf in 32-bit IEEE floating point). So how can I interpret the OPeNDAP version?

@kodiologist ,

Thank you for pointing out this issue. The OPeNDAP URL you are accessing is one that is configured to automatically apply scale factors and offsets to the science datasets. This issue was unknown to us. We will investigate this behavior for that OPeNDAP URL. We do have another OPeNDAP URL that is configured to not apply the scale factors and offsets. That endpoint is, in fact, returning correct data for the QC layer. The following link will provide correct, unscaled data for M*D21A1D and M*D21A1N: https://opendap.cr.usgs.gov/opendap/uns/MOD21A1D.061/h12v04.ncml.html