The timezone strings produced by BESUtil::get_time() are not compatible with the UNIDATA client code (java-netcdf?). The client code rejects the time string because of the time zone.
The error from Panoply is:
We can see that the correctly constructed ISO-8601 time from Hyrax is being rejected because the regex only accepts Z or z for the timezone value.
Note: A to-lowercase() operation is obviously being performed on the time string prior to applying the regex.
We can reach out to UNIDATA about this and they may chose to fix it. Even if they address the problem this leaves existing clients hanging. It seems that we might need to parse these values in the OLFS and then convert them to an encoding that makes it past that regular expression.