Natalia Calvo

The springtime variability of the stratospheric polar vortex over Antarctica is an important source of predictability for the Southern Annular Mode (SAM) and associated Southern Hemisphere (SH) regional climate anomalies in austral spring–summer seasons. Variations in the Antarctic spring vortex often result from anomalous meridional shifts of the SH winter stratospheric jet maximum near the stratopause and associated wave–mean flow feedback. In this study, we evaluate 1) the ability of the models participating in the Coupled Model Intercomparison Project phase 6 (CMIP6) under historical forcings to simulate the observed dynamical chain of processes linking wintertime upper-stratospheric vortex anomalies to springtime surface climate anomalies; and 2) the influence of this ability on future projections of the SH stratospheric vortex and SAM under the highest greenhouse gas emission scenario [shared socioeconomic pathway (SSP) 5-8.5]. Our results show that many CMIP6 models are deficient in reproducing the observed dynamical evolution of SH vortex anomalies, mainly due to a systematic poleward bias in the mean position of the SH winter stratospheric jet maximum. These biases appear to significantly influence projected changes in the stratospheric zonal circulation and the associated SAM by the end of the twenty-first century. Therefore, model deficiencies in simulating the correct position of the SH winter stratospheric jet near the stratopause and the dynamical evolution of its anomalous meridional shift to a spring polar vortex strength anomaly constitute a key source of uncertainty in projections of the SAM and associated SH climate changes in austral spring. Significance Statement: The Southern Annular Mode (SAM) is the leading mode of extratropical circulation variability in the Southern Hemisphere (SH) and significantly influences SH regional climate, ocean circulations, and Antarctic sea ice variations. Austral spring SAM variability is substantially driven by the SH springtime stratospheric polar vortex variability, which often follows anomalous meridional shifts of the preceding winter stratospheric jet and associated wave–mean flow interactions. We show that many climate models do not accurately simulate these stratospheric processes largely due to a systematic poleward bias in the mean position of the SH winter stratospheric jet maximum. These model deficiencies appear to be a key source of uncertainty in future projections of the SAM and associated SH climate changes in austral spring.

Accurate representation of large‐scale stratospheric transport in chemistry–climate models is essential for understanding and interpreting observed atmospheric variability and for making reliable projections. The subtropical transport barriers (STBs) separate the tropical stratosphere, influenced by slow upwelling, from the midlatitudes, where rapid mixing occurs. Long‐lived tracer contours reflect the combined effects of advection and mixing, making them valuable for identifying these barriers. This study compares a set of tracer‐based and dynamical diagnostics of the STBs in the CESM1‐WACCM chemistry‐climate model with various observational data sets. Tracer‐based barriers in the model show excellent agreement with observations between approximately 50‐5 hPa in seasonal climatology, variability associated with the Quasi‐Biennial Oscillation (QBO), and trends. The STBs exhibit a QBO‐driven displacement due to secondary meridional circulation in the winter hemisphere and due to the enhanced isentropic mixing associated with waves crossing the equator in the summer hemisphere. Tracer‐based barriers reveal a statistically significant southward shift of the tropical pipe from 2005 to 2012 consistent with prior findings. We show that this southward shift is a non‐linear effect of the QBO and disappears over longer periods. Dynamical estimates of the STBs show overall good agreement between model and reanalysis in terms of the climatology but are unreliable in the summer hemisphere. Moreover, they fail to capture the QBO‐related variability and trends observed in the tracer‐based boundaries. This discrepancy highlights that dynamical metrics isolate specific instantaneous transport processes, whereas tracer‐based metrics capture their cumulative impact, displaying a lagged and damped seasonality compared to the dynamical estimates.

Arctic ozone is projected to recover over the 21st century in compliance with the Montreal Protocol. Chemistry‐climate models show a large spread in recovery rates and future changes in key dynamical features such as the polar vortex and the residual circulation. Here, we quantify the spread in Arctic ozone recovery explained by uncertainties in these dynamical changes, using simulations from the Chemistry‐Climate Model Initiative, phases 1 and 2022. About 60% of the spread in projected ozone recovery is attributed to differences in polar vortex trends and about one‐third to the intermodel spread in residual circulation trends. Ozone‐depleting substances and greenhouse gas‐driven global warming contribute similarly to the spread in ozone recovery, although uncertainties in dynamical trends primarily affect the warming component. While dynamical changes clearly separate the simulated ozone recovery among different models, the large variability in the Arctic stratosphere precludes establishing an observational constraint on future ozone recovery.

The wintertime stratospheric polar vortex in the Northern Hemisphere is an important source of predictability of surface weather and climate variability in subseasonal-to-seasonal time scales. Yet, climate models suffer from long-standing biases in the representation of the mean state of the vortex, as well as large uncertainties in the future evolution and variability under scenarios of increasing greenhouse gas concentrations. This study analyzes historical simulations from the Coupled Model Intercomparison Project phase 6 and investigates the dynamics associated with the winter-mean biases of the polar vortex strength. Our findings reveal a linear relationship between the biases in the winter-mean strength of the vortex and the meridional wave activity flux in the stratosphere, which is also partly connected to tropospheric circulation biases. A large cold bias in subtropical Pacific sea surface temperatures is associated with an overly strong subtropical jet over the region, and poleward meridional fluxes in both troposphere and stratosphere, contributing to a weaker climatological vortex. In addition, we find a systematic bias toward a too weak stationary wave forcing in the polar stratosphere, which appears in almost all models independently of the vortex strength bias. The main contributor to this bias in wave forcing is a too weak upward wave activity flux, which is partly linked to a systematically weakened Aleutian low in the upper troposphere. Significance Statement: Understanding how the stratospheric polar vortex behaves during winter is important because it can influence weather patterns across the Northern Hemisphere weeks in advance and modulate year-to-year climate variations. The purpose of this study is to better understand the dynamical mechanisms responsible for the large spread in the simulation of the mean strength of the polar vortex by climate models. Our results highlight the connection of model errors in the polar vortex to temperature patterns on the surface of the oceans, a connection that is mediated by large-scale atmospheric waves.

Sudden Stratospheric Warmings (SSWs) are the most important variability phenomena of the wintertime polar stratosphere. SSWs impact tropospheric circulation and surface temperature, even influencing the occurrence of surface extreme events. Still, there is a lack of consensus on how SSWs will change in the future. In this study, we investigate how the SSW impacts on temperature surface extreme events are affected by an abrupt quadrupling of CO ₂ , compared to pre‐industrial conditions, in CMIP6 models. We analyze two subsets of models regarding their changes in SSWs frequency in response to 4xCO ₂ (models with decreasing SSWs frequency (NEG), and models with an increase frequency (POS)). Under 4xCO ₂ conditions, NEG models project a strengthened polar vortex, have more persistent SSW signals in the lower stratosphere and have larger surface impacts than POS models. NEG models also show a higher probability of days affected by SSW‐related cold spells under a 4xCO ₂ increase with longer cold spells over Scandinavia (and to a lesser extent, Siberia). This would be related to the larger persistence of the SSW signal in the lower stratosphere and the intensified stratosphere‐troposphere coupling in response to 4xCO ₂ in this group of models. SSWs seem to have less influence on warm extreme events over Greenland and so, they are not affected by changes in the downward coupling after SSWs under a 4xCO ₂ increase. These results improve the current understanding of how the stratosphere‐troposphere coupling is modified under 4xCO ₂ conditions.

Lower stratospheric ozone between 60°S and 60°N has continued to decline since 1998, despite the reduction of ozone‐depleting substances following the Montreal Protocol. Previous studies have shown that, while chemistry‐climate models reproduce the negative ozone trend in the tropical lower stratosphere as a response to increased upwelling, they fail to capture the ozone decline in northern midlatitudes. This study revisits recent lower stratospheric ozone trends over the period 1998–2018 using two types of simulations from the new Chemistry Climate Model Initiative 2022 (CCMI‐2022): REF‐D1, with observed sea surface temperatures, and REF‐D2, with simulated ocean. The observed negative trend in midlatitudes falls within the range of model trends, especially when considering simulations with observed boundary conditions. There is a large spread in the simulated midlatitudes ozone trends, with some simulations showing positive and others negative trends. A multiple linear regression analysis shows that the spread in the trends is not explained by the different linear response to external forcings (solar cycle, global warming, and ozone‐depleting substances) or to the main variability modes (El Niño‐Southern Oscillation and the quasi‐biennial oscillation) but is instead attributed to internal atmospheric variability. Moreover, the fact that some models show very different trends across members, while other models show similar trends in all members, suggests fundamental differences in the representation of the internal variability of ozone transport across models. Indeed, we report substantial intermodel differences in the ozone‐transport connection on interannual timescales and we find that ozone trends are closely coupled to transport trends.

Changes in the Brewer‐Dobson circulation (BDC) in response to increasing CO ₂ concentrations can arise from the direct effect of radiative cooling in the stratosphere or the indirect effects induced by warmer sea surface temperatures (SSTs). This study aims to disentangle these two contributions in the Whole Atmosphere Community Climate Model (WACCM) by analyzing the timescales of the tropical upwelling response to an abrupt quadrupling of CO ₂ . Transient atmosphere‐ocean climate model simulations of 100 years under 4xCO ₂ conditions are compared to preindustrial control simulations and to simulations with an atmosphere‐only version of WACCM that uses preindustrial SSTs. We find that most of the response in both shallow and deep branches of the BDC occurs on fast timescales (first 2‐3 decades). In the shallow branch of the BDC, the response is mainly driven by changes in SSTs in the well‐mixed shallow ocean, which cause tropospheric warming and an intensification and upward displacement of the subtropical jets, and alter wave forcing in their vicinity. The contribution from stratospheric radiative cooling is almost negligible. In the upper stratosphere, the response of tropical upwelling begins earlier and develops faster than in the shallow branch, owing to the larger contribution from the rapid adjustments. At 1 hPa, 70% of the fast response relates to stratospheric radiative cooling and 30% to warmer SSTs. The modulation of the filtering of non‐orographic gravity waves (mainly of frontal origin) in the subtropics explain most of the response in tropical upwelling in the deep branch, on both fast and slow timescales.

Nonlinearities and asymmetries of El Niño Southern Oscillation (ENSO) stratospheric pathway to the North Atlantic and Europe are examined in large ensembles conducted with fully coupled climate models during wintertime. The analysis is centered on historical experiments of the Max Planck Institute Grand Ensemble (MPI‐GE, 95 members) and expanded to six other ensembles of more limited size. In MPI‐GE, significant responses are obtained for each ENSO phase and three different intensities (weak, moderate and strong). Overall, linear relationships are found for either El Niño or La Niña key diagnostics that characterize the pathway. These relationships are generally weaker for the cold La Niña than for the warm El Niño so that asymmetries between them develop as the events intensify. Specifically for strong events, the extra‐tropical North Pacific and stratospheric responses are asymmetric, with larger responses for El Niño. In addition, the stratospheric asymmetry in strong events seems to contribute to the asymmetry in strong events in the North Atlantic—Europe response in the troposphere in late winter. The extra‐tropical North Pacific response shows general agreement between MPI‐GE and the other large ensembles. However, this agreement is not as large when other parts of the pathway are compared. Relatively high inter‐model response spread confirms the typical model uncertainty found when examining atmospheric circulation responses which include the stratosphere in state‐of‐the‐art climate models.

Major sudden stratospheric warmings (SSWs) are the most important phenomena of the wintertime boreal stratospheric variability. During SSWs, the polar temperature increases abruptly, and easterlies prevail in the stratosphere. Their effects extend farther from the polar stratosphere, affecting near‐surface circulation. According to observations, SSWs are not equally distributed in time, with decades experiencing very few events, while others experiencing SSWs almost every winter. Some sources of this SSW multidecadal variability can be traced back to sea surface temperature changes. Here, we investigate the effects of Pacific decadal variability (PDV) and Atlantic multidecadal variability (AMV) on SSWs. We use for the first time a large ensemble of historical experiments to examine the modulation of the frequency, tropospheric precursors, and impact of SSWs by the PDV and AMV. We find a strong impact of the PDV on the occurrence of SSWs, with a higher SSW frequency for the positive phase of the PDV. This PDV influence is mediated by constructive interference of PDV anomalies with tropospheric stationary waves. The main effect of AMV is, instead, a modulation of the tropospheric response to SSWs, a finding that can be useful for predicting the tropospheric fingerprint of SSWs.

The Brewer-Dobson circulation (BDC) is a key feature of the stratosphere that models need to accurately represent in order to improve the representation of surface climate variability. For the first time, the Climate Model Intercomparison Project includes in its phase 6 (CMIP6) a set of diagnostics that allow for careful evaluation of the BDC. Here, the BDC is evaluated against observations and reanalyses using historical simulations. CMIP6 results confirm the well-known inconsistency in BDC trends between observations and models in the middle and upper stratosphere. The increasing CO2 simulations feature a robust acceleration of the BDC but also reveal large uncertainties in the deep branch trends. The very close connection between the shallow branch and surface temperature is highlighted, which is absent in the deep branch. The trends in mean age of air are shown to be more robust throughout the stratosphere than those in the residual circulation.