search

coecms / esm-pmip-last-mil

0 stars 1 forks source link

Ozone #4

Open HoWol76 opened 4 years ago

HoWol76 commented 4 years ago

O3: “for models without interactive ozone chemistry, we suggest that O3 modulation is derived in a similar way from the modulation of the UV part of the solar spectrum as in the historical simulations” – so should be able to follow what was done in the historical runs

we don’t have interactive O3 chemistry (no UKCA), but we have prescribed O3 concentrations. Need to figure out how those were created.

ScottWales commented 4 years ago

https://gmd.copernicus.org/articles/10/2247/2017/gmd-10-2247-2017.pdf states that some modulation should be added to ozone based on the solar radiation, as is done in the provided historical forcings, but precalculated files or the details of how to do so are not available

nickywright commented 4 years ago

Section A4.3 from https://gmd.copernicus.org/articles/10/4005/2017/gmd-10-4005-2017.pdf might help

HoWol76 commented 4 years ago

See section A4.3 of the document Nicky mentioned.

O3(t) = O3(tpicontrol)·(1 + 0.01 · A · 1JUV(t)).

J_UV(t) might be available from Solar Radiation?

A contains the linear regression coefficients of the ozone mole fraction with respect to changes in solar UV irradiance given in % ozone change per unit (i.e. 1 W m−2 ) irradiance change in the spectral range from 200 to 320 nm (file available at https://pmip4.lsce.ipsl.fr/doku.php/data:solar).

ScottWales commented 4 years ago

@nickywright Actually, would you like to do the calculation of the ozone field as a netcdf file, and then I can convert it into the format that the model needs? It looks like the coefficients need to be calculated, they are not provided.

I've put the historical run ozone data and the solar irradiance source data for the PMIP run at /g/data/w35/saw562/access-esm/ozone.

nickywright commented 4 years ago

Yep @ScottWales, I'll take a look and see how I go.

nickywright commented 4 years ago

Hi @ScottWales, are you able to do the final netcdf calculation for the ozone?

I've calculated and put the different parts of the equation in /g/data/w35/nw3057/access-esm_past1000/ozone.

O3(tpicontrol) and A are in monthly res, while the solar file (and so the calculated ΔJUV) is in annual, and I'm not quite sure what the model does when converting this to monthly and don't want to add any artefacts here! Otherwise can always do a linear interp to resample time using cdo or something.

filenames are: O3(tpicontrol) is vmro3_input4MIPs.... A is cmip6_solar-o3_coeff... ΔJUV is pmip4_past1000_dJUV.nc

ScottWales commented 4 years ago

Ozone files are now under /g/data/access/payu/access-esm/input/pmip-lm/atmosphere

Due to limitations of input file lengths the ozone data has been split into 200 year sections, we'll need to set up the model to update which file it reads from as it goes along.

The data in netcdf format can be found at /g/data/w35/saw562/access-esm/ozone/intermediate/mmro3_zlev_zonal_esm.nc if you'd like to take a look and make sure values are sensible @nickywright