[ERROR] NorESM 1pct co2mass data incorrect

[YanchunHe]

Describe the error
An inquiry from a PhD student from the University of Leeds:

I am a PhD student at the university of Leeds (UK). Currently, I’m working on developing a simple carbon cycle model, for which I am using data from 1pct runs from multiple CMIP6 models. Among those is NorESM-LM, and I have a few queries about it that I’d very grateful if someone with better knowledge of the model could help me understand:

First (and main) query: The CO2 concentration (co2mass variable) data seems wrong to me. If I aggregate the data to global-average annual-mean quantities I get around 190 ppt for the first few years, which is far lower that the pre-industrial values (around 280 ppt) I get for all other models. I thought the concentrations were prescribed for this specific run, so I fail to understand how the values can be so low. To make sure there is nothing wrong with the treatment of the data on my side I’ve also plotted the co2mass variable directly from the datafile, without any conversions, which looks like this:

As you can see, the values in the y axis are different from what I see in other models, see here:

NorESM values seem to be around 40% smaller. I thought maybe this was a mistake and you were publishing values of Carbon instead of Carbon dioxide, but I think the value conversion would be around 4 (44.97/12.011) to go from one to another, so it wouldn’t’ help. Any idea what might be going wrong here? Or was the experiment just run with much lower values of co2 concentration than other ESMs?

Second query: I added all the soil, litter and coarse wood debris carbon pools into a general “soil” carbon pool and I differentiated its amount to obtain the net flux between soil and atmosphere. If I compare that implicit flux to the addition of litter flux (fVegLitter) and heterotrophic respiration there’s still some flux unaccounted for:

          Any idea what that missing flux could be? I haven’t been able to find any variable that may explain that. Initially I thought fLitterFire (so the flux from litter to the atmosphere due to fire) would be that, but it seems to be 0 at all gridpoints and timesteps (for the 1pct run, anyway).

Third query: I’ve tried to do a similar thing as above but with the total flux between atmosphere and land (so the nbp variable). I’ve added GPP, Ra, Rh and fVegFire, but there is still some missing flux, see below. Any idea what this could be? I tried using fharvesttoproduct and fdeforesttoproduct, but they were 0. No other published variable seemed relevant, so I’m wondering if this is something that hasn’t been published? Its quite a small discrepancy anyway.

Thanks for your help. As I said, the question I most want answered is the first one, the other two are secondary, but it would still be good to understand.

Best,

A.R.P

[JorgSchwinger]

@YanchunHe I have checked the noresm2cmor var_CMIP6_NorESM2_default.nml, which says:

'co2mass ','co2vmr*1.5172413793 ','timesmass ',

this seems correct to me (with the 1.5172413793 being 44/29, the ratio of CO2 molar weight to dry air molar weight), and assuming that 'timesmass' means that this is multiplied by the total atmospheric mass?

[ffroeb]

The output variable 'co2mass' is correct; it is in fact a mass and not a measure of concentration. Dividing 'co2mass' by the mass of the atmosphere, 'airmass' (I used a mean over the preindustrial control, assuming that total airmass changes due to increasing CO2 are small), results in the mixing ratio of CO2 in ppm. This is identical to 'atmco2' from the ocean output, see Figure below. Converting to PgC using 2.124 conversion factor.

Depending on what the CMIP output variable should be, the cmorization script should maybe be updated.