Implementation of unpackDoubleIEEE

Open Martin-Imrisek opened this issue 10 months ago • 1 comments

Dear Andrew,

this are my changes regarding the reading and unpacking Double IEEE data . As you can see the dirty fix is converting the double values to float as they are unpacked in method Grib2RecordDS.unpackDoubleIEEE(). (for geopotential there is no loss of precision)

As was mentioned, we are using 64-bit interpretation for geopotential, witch is unnecessary and may be changed in future. However, I believe the unpacking may be useful for ECMWF gribs https://apps.ecmwf.int/ifs-experiments/rd/hej6/ and the double precision should be preserved.

What changes do you suggest to the data flow?

Regards, Martin

Feb 24 '25 09:02 Martin-Imrisek