The motivation
The PMP annual cycle calculations required the time range of CMIP data from 1981-01 to 2005-12. Data from all the piControl and other idealized experiments general start from year 1 and end in several hundred years. So there is no time overlapping between them, thus, all the executions with the piControl experiments failed with the index error.
The proposed solution
Filter the ingested data, so the piControl experiments won't be used for the annual-cycle calculation. Or change the code to account for the simulations with the time starting from 1.
Alternatives
Additional context