The quantification of the climate impacts exerted by stratospheric ozone changes in abrupt 4 × CO2 forcing experiments is an important step in assessing the role of the ozone layer in the climate system. Here, we build on our previous work on the change of the ozone layer under 4 × CO2 and examine the effects of ozone changes on the climate response to 4 × CO2, using the Whole Atmosphere Community Climate Model. We show that the global-mean radiative perturbation induced by the ozone changes under 4 × CO2 is small, due to nearly total cancellation between high and low latitudes, and between longwave and shortwave fluxes. Consistent with the small global-mean radiative perturbation, the effect of ozone changes on the global-mean surface temperature response to 4 × CO2 is negligible. However, changes in the ozone layer due to 4 × CO2 have a considerable impact on the tropospheric circulation. During boreal winter, we find significant ozone-induced tropospheric circulation responses in both hemispheres. In particular, ozone changes cause an equatorward shift of the North Atlantic jet, cooling over Eurasia, and drying over northern Europe. The ozone signals generally oppose the direct effects of increased CO2 levels and are robust across the range of ozone changes imposed in this study. Our results demonstrate that stratospheric ozone changes play a considerable role in shaping the atmospheric circulation response to CO2 forcing in both hemispheres and should be accounted for in climate sensitivity studies.

This dataset consists of a set of meteorological and chemical fields produced by the NCAR CESM climate model, which have been used to produce the analysis described in the paper. The data is in NETCDF format and has been post-processed and formatted using the NCO command language (see http://nco.sourceforge.net/ for more details).

Each field is separately saved as NETCDF file, containing a header and metadata (header). Each field is show on geographic coordinates, which can be longitude,latitude (for surface fields, such as surface temperature, wind at 850 hPa and precipitation rate), or latitude,vertical hybrid-pressure level (for zonal mean fields, such as temperature and wind). To open the data, NETCDF libraries should be installed in the user's own operating system. See https://www.unidata.ucar.edu/software/netcdf/ for more information.

Each NETCDF file is named using a naming convention which identifies the model used (1), the experiment (2), the time resolution (3), the field name (4), and the time covered.

Each file is named as follows:

$MODEL.$EXP.$RESOLUTION.$FIELD.$START_YEAR-$END_YEAR.nc

The data is provided in 3 sections, following the sequence described in the paper. Th first data-set section is the ozone field, the second section is the forcing data, the third section is the data from (free-running) simulations (PiControl and abrupt4xCO2) from SC-WACCM, forced with the ozone forcing.

section 1) The monthly-mean zonal mean ozone data (e.g. to produce Fig.1) is arranged as follows:

$MODEL identifies the climate model from which the ozone is taken from (CESM-WACCM; SOCOL-MPIOM; GFDL-CM3). See paper for details.

section 2) The adjusted forcing data (to produce Fig. 2) is arranged as follows:

$MODEL is "port" (i.e., CESM-PORT), $EXP identifies the model from which the ozone forcing has been derived from (corresponding to the field from section 1). Each NC file contains the annual mean SW and LW flux differences (between PERT and CTRL) at the tropopause, called "FSNR" and "FLNR", respectively. For example, port.waccm.4xCO2.annual.FORCING.0001.zm.nc identifies the forcing arising from the ozone response in CESM-WACCM. See paper for details.

section 3) The output from the climate model simulations (Fig.3-8) is arranged as follows:

$MODEL is "scwaccm" (i.e., SC-WACCM4, as described in the paper). $EXP identifies the experiment (PiControl and 4xCO2), as listed in Table 1 of the paper. 

$RESOLUTION is monthly

$FIELD can be as follows:

T --> zonal mean atmospheric temperature (3 dimensions: lat,lev,time)

U --> zonal mean zonal wind (3 dimensions: lat,lev,time)

PRECT --> precipitation rate in m/s (3 dimensions: lat,lon, time)

TREFHT --> surface temperature (3 dimensions: lon,lat,time)

U850 --> zonal wind component at 850 hPa (3 dimensions: lon,lat,time)

$START_YEAR-$END_YEAR is the time period covered by the simulation, in years.

The dimensions in each field are:

lon --> longitude

lat --> latitude

lev --> hydrid pressure level (NOTE: conversion into pressure levels is possible by using the a/b coefficients provided, and the equation provided in each NC file)

time --> days of the simulation