IBM Manuals give this...
If nn is 16, match the identifier in message ICE606I or ICE627I for this operation to the same identifier in DFSORT message ICE200I (in the DFSMSG data set) to find the set of DFSORT messages for this operation. Correct the error indicated in the DFSORT error message.
As the message description says, you would look "in the DFSMSG data set". That's where the DFSORT messages appear.
When I tried the equivalent of your job, I got this message in DFSMSG:
ICE027A 9 END OF OUT FIELD BEYOND MAXIMUM RECORD LENGTH
The reason is that you have LRECL=80 for the T1 data set. Since you're adding the 2-byte id at 81-82, the T1 data set must have LRECL=82. If you hadn't specified LRECL=80, DFSORT would have automatically set LRECL=82 and the job would have worked. But because you set LRECL=80 for T1, you forced LRECL=80. So your INCLUDE=(81,2,CH,EQ,C'11') tries to go beyond the end of the 80-byte records in T1 and you get the ICE027A message.
It's best NOT to specify DCB parameters for DFSORT output data sets unless you really need to override what DFSORT sets automatically.