Discontinuity between PI control and historical ozone

[MartinDix]

Dataset source ID(s)

FZJ-CMIP-ozone-1-2

Describe the issue

There is a discontinuity between the PI control climatology and the historical ozone. The 1829-1849 mean from the v20251025 dataset matches the 1850 climatology but it's discontinuous at 1850.

Expected data

I expected a smooth transition from climatology to historical. Data description for the 1829-1849 data says "This transient piControl forcing includes the QBO signal but uses average solar forcing and repeating 1850 emissions.". Why does switching to evolving solar forcing and emissions give a systematic shift in ozone?

Screenshots

[znichollscr]

@hegglinm any advice here for @MartinDix ?

[hegglinm]

@MartinDix @znicholls Thanks for bringing this issue to my attention. I will need a few days to get to the bottom of it. Will update you as soon as I know more.

[MartinDix]

We've spun up our PI control with the PI ozone climatology and are about to start our simulation for CMIP7.

Can we at least have confirmatation that the PI control ozone is correct.

[tlurton]

Since the solar pi climatology was calculated using the 1850-1873 period, would it make sense to yield the O3 climatology over the same period?

[DominikStiller]

The stratospheric volcanic aerosols, used in the simulations from which the ozone forcing is derived, are also discontinuous at 1850 (piControl uses 1850-2021 climatology). See Figure 10 and Section 4.4 in https://doi.org/10.5194/egusphere-2025-4990. These describe the prescribed optical properties though, I don't know if and how emissions were implemented for piControl.

[vnaik60]

@DominikStiller - the use of 1850-2021 climatology for prescribing volcanic aerosols in piControl follows the protocol used in CMIP6.

On your point on "if and how emissions were implemented for piControl." - is the question for models used to generate the ozone forcings?

[DominikStiller]

@vnaik60 Yes, that point was for models used to generate the ozone forcings. My thought was that a potential discontinuity in volcanic aerosols at 1850, depending on how they're implemented those models, could have resulted in the discontinuity in ozone.

[vnaik60]

Possibly, provided models used volcanic aerosol emissions rather than prescribing optical properties. @hegglinm and @ccmi1-test could provide more information.

[ccmi1-test]

My concern is that the offset between the pre-industrial control and the historical is too large to be explained by a change in forcing. From the plot of PI total column shown above, the pre-industrial column ozone falls outside the variability found over 1850 - 1900, when there will be several solar cycles and the Krakato eruption.

We are in the process of setting up the forcings for our group's (CCCma) pre-industrial control and my judgement has been that the PI ozone is correct but that there seems to be problems with the historical. For me the striking feature of the PI control climatology that suggests it is okay is that the equatorial maximum ozone mixing ratio sits at 10 hPa and is about 10 ppm. This is something of a canonical feature of the ozone distribution. From the very beginning of the historical timeseries the ozone maximum is still about 10 ppm but is located one level above - at 7 hPa. This strikes me as unrealistic under any kind of forcing.

This is just my judgement, but the definitive answer will require tracing back the origins of the differences.