The hydroxyl radical (OH) is chemically coupled to other atmospheric constituents including water vapor, NOx, ozone, CO, and methane that provide the sources and sinks of OH. These species have longer lifetimes than OH itself and consequently undergo atmospheric transport, allowing dynamics to indirectly affect OH. We investigated whether a combination of meteorological variables and idealized tracers can predict the OH distributions for 40 degrees S-40 degrees N simulated by multiple models. We find that they can explain 70% or more of the variance in July spatial anomalies in OH with the zonal mean removed at 400 hPa, and 59% or more for tropospheric column OH (tcolOH). We find two constituents observed from space, water vapor and NO2, can together serve as proxies for much of the 40 degrees S-40 degrees N spatial variability in OH at 400 hPa, especially over the ocean. Multiple linear regression (MLR) on water vapor and NO2 columns versus tcolOH results in r 2 > 0.5 for the interannual variability in January tcolOH over more than half of the 40 degrees S-40 degrees N domain in most models. These results highlight the value of satellite observations of water vapor and NO2 for constraining simulated OH variability. However, the relative sensitivity of OH to each of these two variables differs between models. Consequently, understanding individual models' relative sensitivities can help maximize the value of these observational constraints. The results of our proof-of-concept study are encouraging and justify additional research to fully explore the potential of other satellite-observable variables for the development of process-based diagnostics and constraining the spatiotemporal variations of tropospheric OH.

We evaluate the climate response of the NASA Goddard Earth Observing System (GEOS) coupled Atmosphere-Ocean General Circulation Model (AOGCM) to idealized quadrupling of CO2 concentrations. A pre-industrial simulation and an abrupt 4 & times; CO2 simulation are conducted with the GEOS AOGCM following the Coupled Model Intercomparison Project (CMIP) approach. The 4 & times; CO2 run shows expected tropospheric and stratospheric climate response features, including Arctic amplification, intensification of the hydrologic cycle, Hadley Cell expansion, a poleward shift of the Southern Hemisphere jet stream, acceleration of the Brewer-Dobson circulation, and a weaker Quasi-Biennial Oscillation. At the ocean surface, we see a complete disappearance of Arctic sea ice in July-December, and in the Southern Hemisphere, there is a poleward shift of the surface wind stress that drives the Southern Ocean meridional overturning circulation poleward. In the ocean, there is enhanced heat content but reduced heat export to the high northern latitudes coincident with a slowing of the Atlantic Meridional Overturning Circulation. The Equilibrium Climate Sensitivity (ECS) of the GEOS AOGCM is estimated to be 2.55 K, in the lowest quartile of the CMIP5 and CMIP6 ranges. The low ECS is due to a strong negative feedback and a weak effective radiative forcing.

The role of the stratosphere for decadal surface cooling in subpolar and midlatitude Eurasia over the last few decades despite Arctic amplification is isolated using two sets of simulations in four different coupled ocean-atmosphere climate models. In the first set, the stratosphere is nudged to observations (NUDGED) while allowing for the rest of the atmosphere to evolve freely, while in the second set the ocean-atmosphere system is free-running (FREE) and the stratospheric polar vortex does not exhibit long term trends. By comparing NUDGED to FREE, we attribute to the stratosphere the anomalously cold conditions in the 2000s in high latitude Eurasia, and also the contemporaneous warm conditions in Eastern Canada. Furthermore, anomalously rainy conditions in much of Southern Europe in the 2000s can also be largely attributed to the stratosphere. This cold Eurasia fingerprint from the stratosphere overwhelmed or strongly suppressed the forced signal from anthropogenic emissions in all four models, demonstrating the crucial role the stratosphere has for decadal surface-climate variability even in a warming climate.

This study uses an ensemble of climate model experiments coordinated by the Quasi-Biennial Oscillation initiative (QBOi) to analyze the Madden-Julian Oscillation (MJO) in the presence of either perpetual El Ni & ntilde;o or La Ni & ntilde;a sea surface temperatures during boreal winter. In addition to the prescribed El Ni & ntilde;o Southern Oscillation (ENSO) conditions, the nine models internally generate QBOs, meaning each may influence the MJO. Objectives of our analyses are to assess the response of the MJO to strong idealized ENSO forcing and look for evidence of a QBO influence on the MJO in a multi-model context. The diagnostics used include wavenumber-frequency spectra of tropical convective and dynamical fields, measures of MJO lifetime, an evaluation of MJO diversity and visualization of MJO vertical structure, as well as an assessment of QBO morphology and the QBO's impact on tropical convection. Kelvin wave spectral power increases in the El Ni & ntilde;o simulations whereas equatorial Rossby waves power is stronger in the La Ni & ntilde;a simulations. All models simulate faster MJO propagation under El Ni & ntilde;o conditions. This change in speed is corroborated by the MJO diversity analysis, which reveals that models better reproduce the observed "fast propagating" and "standing" MJO archetypes given perpetual El Ni & ntilde;o and La Ni & ntilde;a, respectively. Regardless of ENSO, QBO descent into the lower stratosphere is underestimated and we detect little QBO influence on tropical tropopause stability and MJO activity. With little influence from the QBO on the MJO activity in these runs, we can be confident that the aforementioned changes in the MJO indeed arise from the different ENSO boundary conditions.

The quasi-biennial oscillation (QBO) is the main mode of variability in the tropical stratosphere, influencing the predictability of other regions in the atmosphere through its teleconnections to the stratospheric polar vortices and coupling to surface tropical and extratropical variability. However, climate and forecasting models consistently underestimate QBO amplitudes in the lower stratosphere, likely contributing to their failure to simulate these teleconnections. One underexplored contributor to model biases is missing representation of ozone-radiative feedbacks, which enhance temperature variability in the lower stratosphere, particularly at periods at and greater than the QBO (>28 months). While previous studies suggest that ozone-radiative feedbacks can impact QBO periods, amplitudes and the associated secondary circulation in the lower stratosphere, the reported impacts differ widely among models and are hard to interpret due to differences in methodology. To this end, here we propose a coordinated experimental protocol - held joint between the Atmospheric Processes and their Role in Climate (APARC) Quasi-Biennial Oscillation Initiative (QBOi) and Chemistry Climate Modeling Initiative (CCMI) activities - which is aimed at assessing the coupling between stratospheric ozone, temperature and the circulation. We use the proposed experiments to define the ozone feedback on the QBO in both present-day and idealized (abrupt quadrupling of carbon dioxide) climates. While primary focus is on the QBO, the proposed protocol also enables analysis of other aspects of ozone-radiative-dynamical coupling in the atmosphere, including impacts on the Brewer-Dobson Circulation and tropospheric eddy-driven jet responses to future climate change. Here we document the scientific rationale and design of the QUOCA Phase 1 experiments, summarize the data request, and give a brief overview of participating models. Preliminary results using the NASA Goddard Institute for Space Studies E2-2 climate model are used to illustrate sensitivities to certain methodological choices.

The acceleration of the Brewer-Dobson circulation is one of the most robust impacts on the atmospheric circulation of increasing levels of carbon dioxide (CO2). However, a complete understanding of the mechanisms leading to that acceleration is as yet incomplete. Here, using a single-model framework, we separate and quantify three largely independent pathways that lead to Brewer-Dobson circulation (BDC) acceleration under an abrupt 4 3 CO2 forcing: the warming of sea surface temperatures (SSTs), the cooling of the stratosphere from direct radiative forcing, and the composition feedbacks associated with changes of the ozone layer, each of which is caused by increased CO2. We accomplish this by contrasting NASA Goddard Institute for Space Studies (GISS) model E2.2 simulations in fully coupled and atmosphere-only configurations. First, we validate our methodology and demonstrate that the response in the fully coupled model can be simulated as the linear sum of contributions from warmer SSTs, direct radiative effects, and ozone changes. Second, we show that while surface warming induces ;85% of the BDC acceleration, its impact is limited to the lower stratosphere. By comparison, in the upper-and midstratosphere, the BDC response is dominated by changes due to direct radiative forcing from CO2 (80% of the acceleration at 10 hPa). Third, we find that changes in ozone lead to a deceleration of the BDC, nearly canceling the acceleration by the CO2 direct radiative forcing in the mid-lower stratosphere (30-70 hPa).

This study investigates Quasi-Biennial Oscillation (QBO) teleconnections and their modulation by the El Niño–Southern Oscillation (ENSO) using a multi-model ensemble from the Atmospheric Processes And their Role in Climate (APARC) QBO initiative (QBOi). Analyzing observed QBO–ENSO teleconnections is challenging because it is difficult to separate the respective influences of QBO and ENSO outside the QBO region due to aliasing in the historical record. To isolate these signals, simulations were conducted with annually repeating prescribed sea-surface temperatures (SSTs) representing idealized El Niño and La Niña conditions (the QBOi EN and LN experiments, respectively), and results are compared with the QBOi control experiment (CTL) under ENSO-neutral conditions. The strength of the Holton-Tan relationship between the phase of the QBO and the strength of the polar vortex seen in observations is reproduced in fewer than three models in CTL and by one model in EN. In LN, three out of nine models reproduce the observed Holton–Tan relationship, but with less than half of the observed amplitude. In the Arctic winter climate, sudden stratospheric warmings (SSWs) occur more frequently in EN than in LN; however, unlike in observations, there is no discernible difference in SSW frequency between QBO westerly (QBO-W) and QBO easterly (QBO-E) phases. The Asia-Pacific subtropical jet (APJ) shifts significantly equatorward during QBO-W compared to QBO-E in observations, but this shift is not robust across models, regardless of ENSO phases. In the tropics, the sign and spatial pattern of the QBO precipitation response vary widely across models and experiments, indicating that any potential QBO signal is strongly modulated by the prevailing ENSO phases. Overall, the QBOi models exhibit unrealistically weak QBO wind amplitudes in the lower stratosphere, which may explain the weak polar vortex and APJ responses, as well as the weak precipitation signals in the tropics. In contrast, the QBO teleconnection with the Walker circulation during boreal summer and autumn shows consistent signals in both observations and most models. Specifically, the QBO-W phase is characterized by upper-level westerly and lower-level easterly anomalies over the Indian Ocean–Maritime Continent relative to QBO-E, although the amplitude and timing of these anomalies remain model-dependent. Notably, the influence of QBO phase on the Walker circulation appears insensitive to the ENSO phase.

The Atmospheric Processes And their Role in Climate (APARC) Quasi-Biennial Oscillation initiative (QBOi) has conducted new experiments to explore the modulation of the QBO by El Ni & ntilde;o-Southern Oscillation (ENSO). This paper provides an overview of the experimental design and investigates the modulation of the QBO by ENSO using nine climate models used in QBOi. A key finding is a consistent lengthening of the QBO period during La Ni & ntilde;a compared to El Ni & ntilde;o across all models, aligning with observational evidence. Although several models simulate QBO periods that deviate from the observed mean of approximately 28 months, the relative difference between La Ni & ntilde;a and El Ni & ntilde;o remains interpretable within each model. The simulated QBO periods during La Ni & ntilde;a tend to be longer than those during El Ni & ntilde;o, although, in most models, the differences are small compared to that observed. However, the magnitude of this lengthening shows large inter-model differences. By contrast, even the sign of the ENSO effect on QBO amplitude varies among models. Models employing variable parameterized gravity wave sources generally exhibit greater sensitivity of the QBO amplitude to the presence of ENSO than those models using fixed sources. The models capture key observed ENSO-related characteristics, including a weaker Walker circulation and increased equatorial precipitation during El Ni & ntilde;o compared to La Ni & ntilde;a, as well as a characteristic response in zonal mean zonal wind and temperature. All models also simulate stronger equatorial tropical upwelling in El Ni & ntilde;o compared to La Ni & ntilde;a up to similar to 10 hPa, consistent with ERA5 reanalysis. These modulations influence the propagation and filtering of gravity waves. Notably, models with variable parameterized gravity wave sources show stronger wave forcing during El Ni & ntilde;o, potentially explaining the shorter QBO period modulation in these models. Further investigation into the complex interplay between ENSO, gravity waves, and the QBO can contribute to improved model formulations.

The North Atlantic Oscillation (NAO) is the principal mode of atmospheric variability over the North Atlantic, modulating the weather and climate of neighboring regions in both winter and summer. While Earth System Models generally project a more positive NAO under 21st century high-emission scenarios, uncertainties persist as to the precise response of the NAO to increased CO2 levels, owing to large internal variability. In this study we investigate the response of the NAO to a wide range of CO2 forcings, from two to eight times the preindustrial values. Analyzing a large sample of present-generation climate models, we find that the NAO likely becomes more positive with increasing CO2 concentrations. Moreover, we find a reduction in NAO variability. This leads to a smaller increase in the likelihood of extremely positive NAO events than would be expected based solely on the shift in the mean. On the other hand, we also find a reduction in extremely negative NAO events, which is attributable to both the shift toward more positive values and the decrease in variance. Finally, our analysis reveals that the distribution of the NAO response at high CO2 forcing is negatively skewed. This fact partially offsets the decrease in extremely positive NAO events associated with reduced variability. Ultimately, our results suggest a greater increase in positive NAO events compared to the decrease in extremely negative NAO events at higher CO2 forcing.

The phase and amplitude relationships between dynamical quantities and ozone within the quasi-biennial oscillation (QBO) are explored. An initial assessment of this is done by applying a principal oscillation pattern analysis to observations (SWOOSH for Ozone) and reanalysis data (ERA5). This analysis highlights features of the ozone and temperature variability including two peaks in amplitude at different heights in the QBO region as well as more subtle phase differences that cannot be explained by a simple QBO theory. We also quantify the sizes of the ozone and temperature advection terms and show that the contribution of background upwelling on variations in the ozone gradient is not negligible (approximate to 25%) $(\approx 25\%)$. A radiative-convective equilibrium and photochemical equilibrium model, with the imposed ERA5 QBO variation in upwelling and OSIRIS NOx $\mathrm{N}{\mathrm{O}}_{\mathrm{x}}$ variations, is used to further understand ozone and temperature changes. The results show that photochemistry and transport are important at all levels, and it is misleading to divide the QBO into separate regimes. Prominent aspects of the variability can be reproduced if ERA5 upwelling is reduced by similar to 60% ${\sim} 60\%$ between 15 and 50hPa $50\,\mathrm{h}\mathrm{P}\mathrm{a}$ where ERA5 is likely overpredicting the strength of the secondary meridional circulation. Finally, we demonstrate that nonlocality in the vertical plays a major role in QBO dynamics. This arises from ozone transport, the dependence on column ozone of photochemical production, and radiative transfer between layers.

Accurately modeling the large‐scale transport of trace gases and aerosols is critical for interpreting past (and projecting future) changes in atmospheric composition. Simulations of the stratospheric mean age‐of‐air continue to show persistent biases in chemistry climate models, although the drivers of these biases are not well understood. Here we identify one driver of simulated stratospheric transport differences among various NASA Global Earth Observing System (GEOS) candidate model versions under consideration for the upcoming GEOS Retrospective analysis for the Century (GEOS‐R21C). In particular, we show that the simulated age‐of‐air values are sensitive to the so‐called “remapping” algorithm used within the finite‐volume dynamical core, which controls how individual material surfaces are vertically interpolated back to standard pressure levels after each horizontal advection time step. Differences in the age‐of‐air resulting from changes within the remapping algorithm approach 1 year over the high latitude middle stratosphere—or about 30% climatological mean values—and imprint on several trace gases, including methane () and nitrous oxide (O). These transport sensitivities reflect, to first order, changes in the strength of tropical upwelling in the lower stratosphere (70–100 hPa) which are driven by changes in resolved wave convergence over northern midlatitudes as (critical lines of) wave propagation shift in latitude. Our results strongly support continued examination of the role of numerics in contributing to transport biases in composition modeling.

Under climate change driven by increased carbon dioxide (CO2) concentrations, stratospheric ozone will respond to temperature and circulation changes, leading to chemistry–climate feedback by modulating large-scale atmospheric circulation and Earth's energy budget. However, there is significant model uncertainty since many processes are involved and few models have a detailed chemistry scheme. This work employs the latest data from Coupled Model Intercomparison Project Phase 6 (CMIP6) to investigate the ozone response to increased CO2. We find that in most models, ozone increases in the upper stratosphere (US) and extratropical lower stratosphere (LS) and decreases in the tropical LS; thus, the total column ozone (TCO) response is small in the tropics. The ozone response is mainly driven by slower chemical destruction cycles in the US and enhanced upwelling in the LS, with a highly model-dependent Arctic ozone response to polar vortex strength changes. We then explore the ozone–climate feedback by combining offline calculations and comparisons between models with (“chem”) and without (“no-chem”) interactive chemistry. We find that the stratospheric temperature response is substantial, with a global negative radiative forcing ranging from −0.03 to −0.19 W m−2. We find that chem models consistently simulate less tropospheric warming and a stronger weakening of the polar stratospheric vortex, which result in a larger increase in sudden stratospheric warming (SSW) frequency than in most no-chem models. Our findings show that ozone–climate feedback is essential for the climate system and should be considered in the development of Earth system models.

Most research on the state dependence of climate sensitivity has focused on radiative feedbacks, with less attention given to radiative forcing. However, recent studies show that the carbon dioxide (CO2) radiative forcing depends not only on the CO2 concentration but also on the base state, particularly the stratospheric temperature profile. Hence, we here carry out atmosphere-only experiments with prescribed sea surface temperatures using Community Earth System Model, version 1, Large Ensemble (CESM1-LE), broadband radiative transfer calculations, and a one-dimensional radiative- convective equilibrium model to thoroughly investigate the dependence of effective radiative forcing (ERF) on varying levels of CO2 forcing and base-state stratospheric temperatures from 1/16X to 16XCO2. Using both the CESM1-LE and a radiative- convective equilibrium model, we demonstrate that ERF strongly depends on the CO2 value of the underlying base state, deviating significantly from a simple logarithmic relationship with CO2 concentration. Specifically, doubling CO2 from a base state of 8XCO2 results in an ERF value that is 50% higher than doubling CO2 from a 1/16XCO2 state. By decomposing ERF into instantaneous radiative forcing (IRF) and radiative adjustments, we show that the IRF is largely responsible for the state dependence of ERF. We attribute the increase in IRF with CO2 concentrations to the stratospheric cooling at 10 hPa. Furthermore, we find that the radiative adjustments are not constant with each CO2 doubling and halving, and their magnitude depends on the method used to compute them, be it via radiative kernels or via offline radiative transfer calculations. A significant implication of our findings is that the state dependence of ERF needs to be taken into account when studying climate sensitivity under large CO2 perturbations within the feedback-forcing framework.

Rising greenhouse gases (GHG) and decreasing anthropogenic ozone-depleting substances (ODS) are the main drivers of stratospheric climate evolution in the 21st century. However, our understanding of the coupling between stratospheric composition, radiation and dynamics is still a subject to many uncertainties, partly because of the simplified representation of ozone in many current climate models. In our work, we study stratospheric ozone-climate interactions using idealized CMIP6 DECK experiments (pre-industrial control, abrupt quadrupling of CO2, and 1 % yr−1 CO2 increase). This set up provides longer time-series and stronger GHG forcing than in the historical period. The 6th phase of CMIP has a larger number of participating models with interactive chemistry (“CHEM”) to be contrasted against the models where it is prescribed (“NOCHEM”) than in previous generations of CMIP models. Our findings show that CMIP6 models broadly exhibit a similar ozone response to CO2 with increased ozone in the upper stratosphere (US), driven mostly by rapid adjustments (chemistry), and slow transport-driven decrease in the tropical lower stratosphere (LS), and increase in the extratropical LS. The total column ozone response is small in the tropics and positive at high latitudes, with large inter-model discrepancy, possibly arising from model biases in polar vortex dynamics. We also quantify, for the first time, the radiative and dynamical impacts of ozone and quantify their inter-model uncertainty, by means of radiative transfer calculations and careful comparison of chem vs nochem models. First, we find that CHEM models are colder than NOCHEM models in the UTLS region, consistent with the ozone changes in these regions. Second, we find that the large-scale circulation response is systematically different in CHEM and NOCHEM. Lastly, climate sensitivity tends to be lower in CHEM than NOCHEM models, although the uncertainty across models is large and processes that are not tied to ozone cannot be ruled out. Taken together, our work demonstrates that ozone changes can potentially modulate the modeled response to elevated CO2 levels, stressing the importance of interactive chemistry in the future generation of models, in order to correctly simulate the coupling between chemistry, radiative and dynamical processes under climate change.

In the Southern Hemisphere, Earth system models project an intensification fi cation of winter storm tracks by the end of the twenty-first fi rst century. Previous studies using idealized models showed that storm track intensity saturates with increasing temperatures, suggesting that the intensification fi cation of the winter storm tracks might not continue further with increasing greenhouse gases. Here, we examine the response of midlatitude winter storm tracks in the Southern Hemisphere to increasing CO2 2 from two to eight times preindustrial concentrations in more realistic Earth system models. We fi nd that at high CO2 2 levels (beyond 43CO2), 2 ), winter storm tracks no longer exhibit an intensification fi cation across the extra tropics. Instead, they shift poleward, weakening the storm tracks at lower midlatitudes and strengthening at higher midlatitudes. By analyzing the eddy kinetic energy (EKE) budget, the nonlinear storm-track response to an increase in CO2 2 levels in the lower midlatitudes is found to stem from a scale-dependent conversion of eddy available potential energy to EKE. Specifically, fi cally, in the lower midlatitudes, this energy conversion acts to oppositely change the EKE of long and short scales at low CO2 2 levels, but at high CO2 2 levels, it mostly reduces the EKE of shorter scales, resulting in a poleward shift of the storms. Furthermore, we identify a "tug of war" between the upper and lower temperature changes as the primary driver of the nonlinear-scale-dependent EKE response in the lower midlatitudes. Our results suggest that in the highest emission scenarios beyond the twenty-first fi rst century, the storm tracks' response may differ in magnitude and latitudinal distribution from projected changes by 2100.

Observational studies have shown that the El Niño–Southern Oscillation (ENSO) exerts an influence on the Quasi-Biennial Oscillation (QBO). The downward propagation of the QBO tends to speed up and slow down during El Niño and La Niña, respectively. Recent results from general circulation models have indicated that the ENSO modulation of the QBO requires a relatively high horizontal resolution, and that it does not show up in the climate models with parameterized but temporally constant gravity wave sources. Here, we demonstrate that the NASA Goddard Institute for Space Studies (GISS) E2.2 models can capture the observed ENSO modulation of the QBO period with a horizontal resolution of 2∘ latitude by 2.5∘ longitude but with its gravity wave sources being parameterized interactively. This is because El Niño events lead to more vigorous gravity wave sources generating more absolute momentum fluxes over the equatorial belt, as well as less filtering of these waves into the tropical lower stratosphere through a weakening of the Walker circulation. Various components of the ENSO system, such as the sea surface temperatures, the convective activities, and the Walker circulation, are intimately involved in the generation and propagation of parameterized gravity waves, through which ENSO modulates the QBO period in GISS E2.2 models.

Trace gases and aerosols play a crucial role in shaping Arctic climate through their impacts on radiation and chemistry. The concentration of these substances over the Arctic is largely determined by long-range transport originating from midlatitude and tropical source regions. In this study, we explore how atmospheric circulation modulates the interannual variability of long-range transport into the Arctic by utilizing a chemistry–climate model. Idealized tracers, which have fixed lifetimes and spatially varying but temporally fixed surface emissions corresponding to the climatology of anthropogenic emissions of the year 2000, are employed to isolate the role of atmospheric transport from emission and chemistry in modulating interannual variability. Tracers emitted from different source regions are tagged to quantify their relative contributions. Model simulations reveal that tracers from Europe, East Asia, and North America contribute the most to Arctic tracer mass, followed by those from the Tibetan Plateau and South Asia, as well as the Middle East. These regional tracers are predominantly transported into the Arctic middle to upper troposphere, with the exception of tracers from Europe during winter, which are transported into the Arctic lower troposphere. Our analysis shows that the interannual variability of transport into the Arctic for each regional tracer is determined by the atmospheric circulation over the corresponding emission region; i.e., anomalous poleward and eastward winds over the source region promote transport into the Arctic. Considering tracers with global emissions, a southward shift of the midlatitude jet during winter favors increased transport into the Arctic, particularly for tracers emitted over Asia, aligning with previous studies. Comparisons of tracers with different lifetimes indicate that the interannual variability of shorter lifetime tracers is predominantly influenced by regional tracers with shorter transport pathways into the Arctic (e.g., Europe), while the interannual variability of longer lifetime tracers is more contributed by regional tracers with higher emissions (e.g., East Asia).

Previous studies found many climate properties such as northern hemisphere (NH) surface temperature and precipitation respond non-monotonically when CO2 is increased from 1 x to 8 x CO(2 )relative to pre-industrial levels. Here, we explore the robustness of the non-monotonicity in the NH precipitation response in 11 coupled climate models. Eight models show a decrease in NH precipitation under repeated CO2 doubling, indicating that the non-monotonic response is a common but not universal result. Although common, the critical CO(2 )level where the NH precipitation decrease first occurs differs widely across models, ranging from 2xCO(2) to 8xCO(2). These models also show a prominent weakening in the Atlantic meridional overturning circulation(AMOC) at the same critical CO(2 )level, with the AMOC weakening leading the precipitation decrease. The sensitivities of NH precipitation and the AMOC to CO(2 )doublings are positively correlated, especially when the AMOC weakens beyond 10 Sv. This suggests that the differences in models' AMOC response can explain their contrasting NH precipitation responses, where models with a large AMOC weakening have decreased NH precipitation. Regionally, this decrease in NH precipitation is the most prominent over the North Atlantic, Europe and the tropical Pacific. Our results suggest that special care must be taken with the use of pattern scaling to inform regional climate decision-making.