Checking Land Surface Diagnostics in AMIP Run


I’m doing some UKESM AMIP runs on Monsoon using ancillary files with data from a range of sources and created with xancil following Jeff’s advice in ticket #311(Using xancil and xconv to create timeslice ancillary files).

In particular, as I am changing the vegetation PFT distribution (stash 216), I would like to it output it to check the land surface is doing what I want it to be doing. As a test I am running u-ck407, a PI AMIP run with just the surface fraction ancillary (stash 216) altered. The model runs fine but I am unable to output stash 216 or 19013 to check the new ancillary file is working properly. I think it is working properly looking at the isoprene emissions but I would like to be certain. As stash items 216 and 19013 are in sections I have not used before, are there other changes to the model code I need to make to output these? I see that stash 216 has a non-zero option code value whereas 19013’s option code is all zeros.

Many thanks for your help,


Hi James

I think to output either of those STASH items you need to be using the TRIFFID dynamic vegetation model, which it doesn’t look like you are.

The option code for 0216 has a 2 in position 11 (from the right) which according to UM document C04 means

Item included only for direct vegetation parametrization with or without competition

Which I believe means TRIFFID.

The version mask for 19013 has a 1 in position 2 (from the right) which means the diagnostic is only available for version 2 of section 19 and if you look in the jules source code the are 2 versions 1b and 2b, 1b is with triffid off and 2b with triffid on.

I guess this is because this field is not changing, other than being read from the ancil file and updated, unless TRIFFID is used, hence there is no point in writing it out.

You could look at the field in the dump file and see if it looks correct.


Thanks, Jeff, I had wondered if TRIFFID was required for these. I will check the dump files as you suggest.