limit taxmode is too strict?

[anton-seaice]

I've been struggling to understand how best to configure data streams files, and it basically comes down to first and last time steps.

Easiest to show with an example:

• We have forcing data which ends in 2023, at 9pm
• therefore, setting taxmode to "limit" means the model stops at 9pm.
• We would like to run to the end of 2023, but not run in 2024.

The limit calculations appear to ignore dtlimit when there is no data to interpolate between - e.g. setting dtlimit to 2 or 3 has no impact. Using extend means the model will run forever with the same data from the last timestep of 2023 (rather than stop at the end of 2023).

[uturuncoglu]

@anton-seaice there might be other solutions but you could copy the last time step until 12pm in your netcdf (using nco or cdo) and keep using limit in CDEPS side. BTW, you could finish 2023 with constant data at the end.

[anton-seaice]

you could copy the last time step until 12pm in

Thanks!

Also - trying to avoid this so we can use unmodified data from the source files