This paper describes the recommended solar forcing dataset for CMIP6 and highlights changes with respect to CMIP5. The solar forcing is provided for radiative properties, namely total solar irradiance (TSI), solar spectral irradiance (SSI), and the F10.7 index as well as particle forcing, including geomagnetic indices Ap and Kp, and ionization rates to account for effects of solar protons, electrons, and galactic cosmic rays. This is the first time that a recommendation for solar-driven particle forcing has been provided for a CMIP exercise. The solar forcing datasets are provided at daily and monthly resolution separately for the CMIP6 preindustrial control, historical (1850–2014), and future (2015–2300) simulations. For the preindustrial control simulation, both constant and time-varying solar forcing components are provided, with the latter including variability on 11-year and shorter timescales but no long-term changes. For the future, we provide a realistic scenario of what solar behavior could be, as well as an additional extreme Maunder-minimum-like sensitivity scenario. This paper describes the forcing datasets and also provides detailed recommendations as to their implementation in current climate models. For the historical simulations, the TSI and SSI time series are defined as the average of two solar irradiance models that are adapted to CMIP6 needs: an empirical one (NRLTSI2–NRLSSI2) and a semi-empirical one (SATIRE). A new and lower TSI value is recommended: the contemporary solar-cycle average is now 1361.0 W m⁻². The slight negative trend in TSI over the three most recent solar cycles in the CMIP6 dataset leads to only a small global radiative forcing of −0.04 W m⁻². In the 200–400 nm wavelength range, which is important for ozone photochemistry, the CMIP6 solar forcing dataset shows a larger solar-cycle variability contribution to TSI than in CMIP5 (50 % compared to 35 %). We compare the climatic effects of the CMIP6 solar forcing dataset to its CMIP5 predecessor by using time-slice experiments of two chemistry–climate models and a reference radiative transfer model. The differences in the long-term mean SSI in the CMIP6 dataset, compared to CMIP5, impact on climatological stratospheric conditions (lower shortwave heating rates of −0.35 K day⁻¹ at the stratopause), cooler stratospheric temperatures (−1.5 K in the upper stratosphere), lower ozone abundances in the lower stratosphere (−3 %), and higher ozone abundances (+1.5 % in the upper stratosphere and lower mesosphere). Between the maximum and minimum phases of the 11-year solar cycle, there is an increase in shortwave heating rates (+0.2 K day⁻¹ at the stratopause), temperatures ( ∼ 1 K at the stratopause), and ozone (+2.5 % in the upper stratosphere) in the tropical upper stratosphere using the CMIP6 forcing dataset. This solar-cycle response is slightly larger, but not statistically significantly different from that for the CMIP5 forcing dataset. CMIP6 models with a well-resolved shortwave radiation scheme are encouraged to prescribe SSI changes and include solar-induced stratospheric ozone variations, in order to better represent solar climate variability compared to models that only prescribe TSI and/or exclude the solar-ozone response. We show that monthly-mean solar-induced ozone variations are implicitly included in the SPARC/CCMI CMIP6 Ozone Database for historical simulations, which is derived from transient chemistry–climate model simulations and has been developed for climate models that do not calculate ozone interactively. CMIP6 models without chemistry that perform a preindustrial control simulation with time-varying solar forcing will need to use a modified version of the SPARC/CCMI Ozone Database that includes solar variability. CMIP6 models with interactive chemistry are also encouraged to use the particle forcing datasets, which will allow the potential long-term effects of particles to be addressed for the first time. The consideration of particle forcing has been shown to significantly improve the representation of reactive nitrogen and ozone variability in the polar middle atmosphere, eventually resulting in further improvements in the representation of solar climate variability in global models.

Temperature and ozone changes in the upper troposphere and lower stratosphere (UTLS) are important components of climate change. In this paper, variability and trends of temperature and ozone in the UTLS are investigated for the period 2002–2017 using high-quality, high vertical resolution Global Navigation Satellite System radio occultation (GNSS RO) data and improved merged satellite data sets. As part of the Stratosphere-troposphere Processes And their Role in Climate (SPARC) Reanalysis Intercomparison Project (S-RIP), three reanalysis data sets, including the ERA-I, MERRA2 and the recently released ERA5, are evaluated for their representation of temperature and ozone in the UTLS. The recent temperature and ozone trends are updated with a multiple linear regression (MLR) method and related to sea surface temperature (SST) changes based on model simulations made with NCAR's Whole Atmosphere Community Climate Model (WACCM). All reanalysis temperatures show good agreement with the GNSS RO measurements in both absolute value and annual cycle. Interannual variations in temperature related to Quasi-Biennial Oscillation (QBO) and the El Niño–Southern Oscillation (ENSO) processes are well represented by all reanalyses. However, evident biases can be seen in reanalyses for the linear trends of temperature since they are affected by discontinuities in assimilated observations and methods. Such biases can be corrected and the estimated trends can be significantly improved. ERA5 is significantly improved compared to ERA-I and shows the best agreement with the GNSS RO temperature. The MLR results indicate a significant warming of 0.2–0.3 K per decade in most areas of the troposphere, with a stronger increase of 0.4–0.5 K per decade at midlatitudes of both hemispheres. In contrast, the stratospheric temperature decreases at a rate of 0.1–0.3 K per decade, which is most significant in the Southern Hemisphere (SH). Positive temperature trends of 0.1–0.3 K per decade are seen in the tropical lower stratosphere (100–50 hPa). Negative trends of ozone are found in the Northern Hemisphere (NH) at 150–50 hPa, while positive trends are evident in the tropical lower stratosphere. Asymmetric trends of ozone can be found in the midlatitudes of two hemispheres in the middle stratosphere, with significant ozone decrease in the NH and increase in ozone in the SH. Large biases exist in reanalyses, and it is still challenging to do trend analysis based on reanalysis ozone data. According to single-factor-controlled model simulations with WACCM, the temperature increase in the troposphere and the ozone decrease in the NH stratosphere are mainly connected to the increase in SST and subsequent changes of atmospheric circulations. Both the increase in SSTs and the decrease in ozone in the NH contribute to the temperature decrease in the NH stratosphere. The increase in temperature in the lower stratospheric tropics may be related to an increase in ozone in that region, while warming SSTs contribute to a cooling in that area.

The quasi-biennial oscillation (QBO), as the dominant mode in the equatorial stratosphere, modulates the dynamical circulation and the distribution of trace gases in the stratosphere. While the zonal mean QBO signals in stratospheric ozone have been relatively well documented, the zonal (longitudinal) differences in the QBO ozone signals have been less studied. Using satellite-based total column ozone (TCO) data from 1979 to 2020, zonal mean ozone data from 1984 to 2020, three-dimensional (3-D) ozone data from 2002 to 2020, and ERA5 reanalysis and model simulations from 1979 to 2020, we demonstrate that the influences of the QBO (using a QBO index at 20 hPa) on stratospheric ozone are zonally asymmetric. The global distribution of stratospheric ozone varies significantly during different QBO phases. During QBO westerly (QBOW) phases, the TCO and stratospheric ozone are anomalously high in the tropics, while in the subtropics they are anomalously low over most of the areas, especially during the winter–spring of the respective hemisphere. This confirms the results from previous studies. In the polar region, the TCO and stratospheric ozone (50–10 hPa) anomalies are seasonally dependent and zonally asymmetric. During boreal winter (December–February, DJF), positive anomalies of the TCO and stratospheric ozone are evident during QBOW over the regions from North America to the North Atlantic (120^∘ W–30^∘ E), while significant negative anomalies exist over other longitudes in the Arctic. In boreal autumn (September–November, SON), the TCO and stratospheric ozone are anomalously high from Greenland to Eurasia (60^∘ W–120^∘ E) but anomalously low in other regions over the Arctic. Weak positive TCO and stratospheric ozone anomalies exist over the South America sector (90^∘ W–30^∘ E) of the Antarctic, while negative anomalies of the TCO and stratospheric ozone are seen in other longitudes. The consistent features of TCO and stratospheric ozone anomalies indicate that the QBO signals in TCO are mainly determined by the stratospheric ozone variations. Analysis of meteorological conditions indicates that the QBO ozone perturbations are mainly caused by dynamical transport and also influenced by chemical reactions associated with the corresponding temperature changes. QBO affects the geopotential height and the polar vortex and subsequently the transport of ozone-rich air from lower latitudes to the polar region, which therefore influences the ozone concentrations over the polar region. The geopotential height anomalies associated with QBO (QBOW–QBOE) are zonally asymmetric with clear wave number 1 features, which indicates that QBO influences the polar vortex and stratospheric ozone mainly by modifying the wave number 1 activities.

Solar signals in the atmosphere and the ocean, especially in tropopause temperatures and lower stratospheric water vapour are investigated using recent observational and reanalyses data sets for the period from 1958 through 2013. Previous observational and modeling studies demonstrated solar influences in the lower stratosphere resembling a positive Northern Annular Mode due to the top-down mechanism involving enhanced solar UV radiation in the stratosphere during solar maxima and dynamical amplification mechanisms in the atmosphere. We found that these stratospheric changes might propagate down to the troposphere and become zonally asymmetric with characteristic pressure and wind pattern over the North Atlantic and North Pacific. Such changes in tropospheric circulation are related to anomalous positive SST anomalies in the central Pacific which resemble an El Niño Modoki event. We show for the first time with ocean reanalysis data that these SST anomalies are amplified by a positive feedback through oceanic subsurface currents and heat transport in the equatorial Pacific. Anomalous warm SSTs in the equatorial central Pacific change the zonal SST gradient and lead to anomalous westerly winds and currents in the western Pacific and easterly winds and currents in the eastern Pacific. This indicates a convergence and less upwelling and therefore enhances the positive SST anomalies in the equatorial central Pacific. Such a positive feedback results in a peak of El Niño Modoki events about 2 years after the solar maximum. These solar-induced signals in the ocean in turn modify the circulation and convection in the troposphere, resulting in lagged solar signals of anomalous high tropopause heights and negative anomalies in tropopause temperatures as well as in lower stratospheric water vapour over the equatorial Pacific which are in agreement with a time evolving solar-induced El Niño Modoki-like SST pattern. We demonstrate a solar modulation of intrinsic decadal climate variability over the Pacific which is amplified by positive feedbacks between the ocean and the atmosphere.