We present the first results of the Water Isotope Model Intercomparison Project (WisoMIP), with Phase 1 focused on modern simulations (1979-2023) from a suite of isotope-enabled atmospheric general circulation models nudged to ERA5 reanalyzes. Water sources, mixing, and rainout history influence the isotopic composition of vapor and precipitation, making these simulations powerful tools for tracing the global water cycle. By prescribing identical winds, sea surface temperatures, and sea ice conditions, we isolate differences in water isotope behavior across models, controlling for variability in atmospheric dynamics and mean climate. Our analyses show that the ensemble mean best matches observations, as individual model errors cancel out to yield a more accurate representation of Earth's isotope distributions. We also evaluate trends and responses to major climate modes during the recent warming period, highlighting regional and temporal sensitivities in the isotope signals. These diagnostics extend beyond traditional model evaluation metrics (e.g., temperature, precipitation) to reveal uncertainties in physical processes and guide improvements in model parameterizations. The resulting modern nudged ensemble data set serves as a benchmark for isotope-enabled model development, satellite product comparison, and understanding of water cycle changes in a warming climate. Given its standardized design and broad participation, WisoMIP provides a valuable "isotope reanalysis" product for applications ranging from paleoclimate reconstruction to model tuning. Our work demonstrates the importance of coordinated isotope model evaluation in advancing the use of water isotopes as a diagnostic tool in climate science.

Water isotopes are widely used as proxies in climate research. However, our understanding of the processes that control their composition is still incomplete. In comparison to other major ocean basins, the Mediterranean Sea is well suited to improve our understanding of the factors influencing oxygen isotope variability and to refine current modelling approaches. This study successfully implemented and simulated stable water isotopes (δ18O and δD) across the Mediterranean basin using the high-resolution NEMO-MED12 dynamical model. The model accurately simulates the well-documented east–west δ18O gradient in Mediterranean water masses. Moreover, the results show a good agreement between simulated and observed δD. δD exhibits a strong linear relationship with δ18O (r2 = 0.98) and salinity (r2 = 0.94) across the Mediterranean basin. Furthermore, the modeled δ18O/salinity relationships align with observations, showing a weaker gradient in the eastern basins compared to the western basins. This development offers significant potential for paleoclimate-related applications.

Deep Convective Systems (DCSs) reaching scales of 100–1000 km play a pivotal role as the primary precipitation source in the tropics. Those systems can have large cloud shields, and thus not only affect severe precipitation patterns but also play a crucial part in modulating the tropical radiation budget. Understanding the complex factors that control how these systems grow and how they will behave in a warming climate remain fundamental challenges. Research efforts have been directed, on one hand, towards understanding the environmental control on these systems, and on the other hand, towards exploring the internal potential of systems to develop and self-aggregate in idealized simulations. However, we still lack understanding on the relative role of the environment and internal feedbacks on DCS mature size and why. The novel high-resolution global SAM simulation from the DYAMOND project, combined with the TOOCAN Lagrangian tracking of DCSs and machine learning tools, offers an unprecedented opportunity to explore this question. We find that a system’s growth rate during the first 2 h of development predicts its final size with a Pearson correlation coefficient of 0.65. Beyond this period, growth rate emerges as the strongest predictor. However, in the early stages, additional factors–such as ice water path heterogeneity, migration distance, interactions with neighboring systems, and deep shear–play a more significant role. Our study quantitatively assesses the relative influence of internal versus external factors on the mature cloud shield size. Our results show that system-intrinsic properties exert a stronger influence than environmental conditions, suggesting that the initial environment does not strictly constrain final system size, particularly for larger systems where internal dynamics dominate.

The past four decades have witnessed a strengthening of the winter anticyclonic circulation over the Barents-Kara Sea (BKS), a change that has contributed substantially to amplified local warming and sea ice loss, as well as to Eurasian cooling. However, the cause of this trend in the BKS atmospheric circulation remains unknown. Here we show that anthropogenic greenhouse gases are the primary driver of the strengthening of the BKS anticyclonic circulation, with anthropogenic aerosols playing a secondary role, both together accounting for about 86% of the observed circulation trend. Both forcings induce an amplified BKS low-tropospheric warming through coupling with strong sea ice loss. This amplified warming raises geopotential height aloft through thermal expansion, causing an anomalous anticyclonic anomaly, which in turn enhances warming and sea ice loss, forming a positive feedback loop. Our work provides a theoretical framework for understanding Arctic atmospheric circulation responses to anthropogenic warming and may have implications for climate and environment in the Arctic and beyond.

Mass spectrometry and laser spectroscopy have been widely employed for precise water vapor isotope measurements. Nevertheless, these techniques are limited by logistical challenges in fieldwork, consequently constraining the temporal and spatial resolution of measurements. Specifically, water vapor isotope measurements are primarily limited to near-surface levels, while measurements in the free troposphere are notably scarce. Portable sampling devices, such as air bags and glass bottles, have therefore become necessary alternatives for collecting, storing, and transporting gaseous samples in diverse environments prior to analysis with less portable instruments. In drone-based high-altitude vapor sampling, air bags are preferred for their lighter weight and greater flexibility compared to glass bottles. Nevertheless, they present specific challenges, such as potential sample contamination and isotopic fractionation during storage, primarily due to the inherent permeability of air bags. Here, we developed a theoretical model for water vapor diffusion through the sampling bag surface, with parameters calibrated through laboratory experiments. This model enables the reconstruction of the initial isotopic composition of sampled vapor based on measurements obtained within the bag and from the surrounding environment. This diffusion model underwent rigorous validation through experiments conducted under varying humidity and isotopic composition differences between the inside and outside of the air bag, confirming its reliability. We applied this correction method to air samples collected at various pressures up to the upper troposphere using an air bag-mounted drone that we developed, thereby estimating the initial isotopic composition and uncertainty based on our observations. Our correction method enhances the reliability and applicability of water vapor isotope observations conducted using drones equipped with air bags, and provides a detailed assessment of all potential sources of error and quantifies the uncertainty range of the observations. This approach leverages the strengths of drone-based air bag sampling while mitigating its limitations, thus facilitating the convenient collection of isotopic data throughout the troposphere.

Water-stable isotopic compositions of snow or ice (here δ18O and δD) represent the main way to reconstruct past temperature in Antarctica, and one way to interpret these isotopic signals is through the use of isotope-enabled atmospheric general circulation models. In this study, we combine isotopic observations from surface snow samples, daily precipitation and water vapour to evaluate the LMDZ6iso model in Antarctica from climatic to seasonal and sub-daily time scale. Time-averaged δ18O in precipitation from LMDZ6iso for the period 1980-2022 is in excellent agreement with δ18O of surface snow samples across the continent, but there is a strong disagreement for d-excess at cold temperature sites. For sub-annual time scale analyses, we focus on two sites in East Antarctica: the coastal station Dumont d'Urville and the continental station Concordia. The model accurately reproduces the seasonal isotopic cycle of daily precipitation at both stations, with better performances at Concordia. Moving from statistical evaluation to process analyses, we use water vapour isotopes to study water exchanges in the boundary layer. LMDZ6iso performs well in representing the observed diurnal isotope cycle at both sites. However, the model simulates a larger vapour δ18O depletion than observed during the night at Concordia. We analyse the contribution of each physical process affecting isotope concentrations in LMDZ6iso to show what controls the vapour isotope signal. At Concordia, surface sublimation during the day is the main driver of the diurnal cycle of vapour isotopes, whereas at Dumont d'Urville, daily isotope variations are driven by surface sublimation and turbulence during the day and by air advection from the katabatic flow during the night.

Observations of water stable isotopes in Antarctic surface snow, precipitation and water vapor are key for improving our understanding of the atmospheric water cycle and past climate reconstructions from ice cores. In this study, we use isotopic observations in Antarctica to assess the skill of the isotope‐enabled atmospheric general circulation model LMDZ6, nudged to ERA5 above the boundary layer (1980–2023 period). The model has no significant bias for time‐mean temperature and snow accumulation over the ice sheet. Sensitivity test on parameterized supersaturation strength highlights its opposite effect on precipitation O and d‐excess. Selecting an intermediate supersaturation strength resulted in a minimal bias for surface snow O across the continent, with a reduced but systematic positive bias in surface snow d‐excess (5‰). We then assessed seasonal and diurnal isotope variability with daily precipitation and continuous vapor isotopes at Dumont d’Urville (DDU, coastal station) and Concordia (inland station). On a seasonal scale, LMDZ6iso accurately reproduces the seasonal cycle of precipitation O and d‐excess at both stations. Moving from statistical evaluation to physical analysis, we use the individual process contributions to boundary‐layer water vapor isotopes to identify the main drivers controlling the clear‐sky isotopic daily cycles. At Concordia, daily isotope variations are mainly driven by surface sublimation, whereas at DDU they are driven by surface sublimation and advection by the katabatic flow. Our results suggest that to further improve water isotopes in LMDZ6iso, fractionation during surface sublimation should be included and fractionation at condensation for low temperature should be better constrained.

Isotopic analysis serves as a critical tool in understanding the complexities of the water cycle and quantifying the influence of distinct atmospheric processes.This research focuses on the spatio-temporal distribution of the HDO/H2O ratio in water vapor on Earth and Mars, identifying the processes that control these variations.Utilizing isotopic data from General Circulation Model LMDZ simulations for Earth and Planetary Climate Model (PCM) simulations for Mars, we investigate the similarities and differences in water vapor transport and phase changes within each planet's atmosphere. Key findings include a marked isotopic enrichment from ice sublimation in both planets, with a stronger effect observed on Mars due to longer crystal residence times. In contrast, Earth exhibits a buffering effect by the near-surface ocean not present on Mars. Our hypothesis that a unified conceptual framework can interpret isotopic distributions on both planets is supported, suggesting shared fundamental processes with adaptations to each planet's unique conditions.This comparative analysis not only highlights the similarities and differences in the water cycles of Earth and Mars, but also demonstates the adaptability of our conceptual framework to various planetary environments. These insights enhance our comprehension of planetary hydrological cycles and contribute to a deeper understanding of their underlying microphysical mechanisms.

Mesoscale Convective Systems (MCSs) that become large or have long lifespans contribute disproportionately to extreme rainfall. Gaining a better understanding of the factors that determine whether a system will become large could improve our understanding of extreme weather phenomena. The recent emergence of high-resolution global simulations from the DYAMOND project, coupled with a storm tracking algorithm called TOOCAN, provides a groundbreaking opportunity to study the factors controlling the maximum area of MCSs. In this study we use machine learning algorithms to predict the maximum area of convective systems based on their early development stages and initial environmental conditions. The results reveal that the initial evolution of the system anticipates its maximum area. Factors such as the presence of ice in the system's environment, proximity to surrounding systems, intensity of vertical velocity at 500 hPa, and the migration distance, have been identified as significant factors in improving the accuracy of the prediction. Using a linear model, we investigate the relative role of the environment and of the system itself, in the growth of the system.

Water isotopes are one of the most widely used proxies in ocean climate research. However, there are still gaps in our understanding of the processes that control their composition. Compared to other large ocean basins, the Mediterranean is ideally suited to improve our understanding of the processes influencing and driving oxygen isotopic variability, and to refine the current modelling approach. For the first time in a high-resolution Mediterranean dynamical model (NEMO-MED12), stable water isotopes (δ18O and δD) were successfully implemented and simulated in the whole basin. The well-known east-west gradient of δ18O in Mediterranean water masses is successfully simulated by the model. Results also show good agreement between simulated and observed δD. δD shows a strong linear relationship with δ18O (r2 = 0.98) and salinity (r2 = 0.94) for the entire Mediterranean basin. Furthermore, the modelled δ18O/salinity relationships are in good agreement with observations, with a weaker gradient simulated in the eastern basins than in the western basins. We investigate the relationship of the isotopic signature of the CaCO3 shell (δ18Oc) with temperature and the influence of seasonality. Our results suggest a more quantitative use of δ18O records, combining reconstruction with modelling approaches. This opens up broad perspectives for paleoclimate-related applications.

Stable water isotopes (δ18Ow and δDw) have been successfully implemented for the first time in a high-resolution model of the Mediterranean Sea (NEMO-MED12). In this numerical study, model results are compared with available in situ observations to evaluate the model performance of the present-day distribution of stable water isotopes and their relationship with salinity on a sub-basin scale. There is good agreement between the modelled and observed distributions of δ18Ow in the surface water. The model successfully simulates the observed east–west gradient of δ18Ow characterising surface, intermediate, and deep waters. The results also show good agreement between the simulated δDw and the in situ data. The δDw shows a strong linear relationship with δ18Ow (r2=0.98) and salinity (r2=0.94) for the whole Mediterranean Sea. Moreover, the modelled relationships between δ18Ow and salinity agree well with observations, with a weaker slope in the eastern basin than in the western basin. We investigate the relationship of the isotopic signature of the planktonic foraminifera shells (δ18Oc) with temperature and the influence of seasonality. Our results suggest a more quantitative use of δ18O records, combining reconstruction with modelling approaches.

In order to complement the picture of the atmospheric water cycle in the Southern Ocean, we have continuously monitored water vapor isotopes since January 2020 on Amsterdam Island in the Indian Ocean. We present here the first 2-year long water vapor isotopic record at this site. We show that the water vapor isotopic composition largely follows the water vapor mixing ratio, as expected in marine boundary layers. However, we detect 11 periods of a few days where there is a strong loss of correlation between water vapor δ18O and water vapor mixing ratio as well as abrupt negative excursions of water vapor δ18O. These excursions often occur toward the end of precipitation events. Six of these events show a decrease in gaseous elemental mercury, suggesting subsidence of air from a higher altitude. Our study aims to further explore the mechanism driving these negative excursions in water vapor δ18O. We used two different models to provide a data–model comparison over this 2-year period. While the European Centre Hamburg model (ECHAM6-wiso) at 0.9° was able to reproduce most of the sharp negative water vapor δ18O excursions, hence validating the physics process and isotopic implementation in this model, the Laboratoire de Météorologie Dynamique Zoom model (LMDZ-iso) at 2° (3°) resolution was only able to reproduce seven (one) of the negative excursions, highlighting the possible influence of the model resolution for the study of such abrupt isotopic events. Based on our detailed model–data comparison, we conclude that the most plausible explanations for such isotopic excursions are rain–vapor interactions associated with subsidence at the rear of a precipitation event.

Stable isotopic signals preserved in natural precipitation archives, such as ice cores, provide information on past climatic changes. When measured in the water vapor, water isotopes bear information on large-scale transport, convective and cloud processes. To document spatial and seasonal variations across China, we made in-situ observations of near-surface vapor isotopes over a large region (over 10000 km) across China in both pre-monsoon and monsoon seasons, using a newly-designed vehicle-based vapor isotope monitoring system. We found that the observed spatial variations in both periods represent mainly seasonal-mean spatial variations, but are also influenced by synoptic-scale variations during the monsoon period. The spatial variations of vapor δ18O are mainly controlled by Rayleigh distillation along air mass trajectories during the pre-monsoon period, but are significantly influenced by different moisture sources, continental recycling processes and convection along moisture transport during the monsoon period. Thus, the North-South gradient observed during the pre-monsoon period is counteracted during the monsoon period. These results provide an overview of the spatial distribution and seasonal variability of water isotopic composition in East Asia and their controlling factors, and emphasize the need to interpret proxy records in the context of the regional system. To better understand the physical processes that control the vertical distribution of vapor isotopes and the added value of vapor isotopic measurements to infer deep convective processes, we made observation of the vertical profiles of atmosphere vapor isotopes up to the upper troposphere (from the ground surface at 3856m up to 11000m a.s.l.) from June to October in the southeastern Tibetan Plateau using a specially-designed unmanned-aerial-vehicle system. The vertical distribution of atmospheric water vapor isotopes across the entire monsoon period up to the upper troposphere are derived for the first time. We find that the vertical profiles of water vapor isotopic composition reflects a combination of large-scale processes, in particular deep convection along trajectories, and local convective processes, in particular convective detrainment and sublimation of ice crystals. The observed vapor δ18O decreases and its d-excess increases with altitude up to 8-10km, consistent with the progressive condensation of water vapor and precipitation. Beyond this altitude, where the maximum convective detrainment occurs, the vapor δ18O re-increases with altitude, reflecting the sublimation of ice crystals detrained from convective clouds. The d-excess is maximum during the monsoon period, due to more depleted initial vapor. The observed seasonal and intra-seasonal variations are generally vertically coherent, due to the strong vertical convective mixing and local convective detrainment of vapor originating from the low levels. The vapor is more depleted during the monsoon season and in October, due to deep convection along trajectories.

The isotope-enabled general circulation models (GCM) have been widely applied to simulate the variability of stable isotopes in meteoric water at various time scales. The in-situ observations of water vapour isotopes are an important basis for assessing the performance of isotope-enabled GCMs, although they are still limited. Here we compiled the observations of near-surface water vapour isotopes on a daily scale at 17 stations in East Asia, and assessed the skill and the association between isotope error and meteorological errors on a daily scale. Generally, the spatial pattern and seasonal variability can be well simulated in the isotope-enabled GCMs. The models show better skill for warm and humid backgrounds, which also corresponds to the monsoonal regions with lower latitudes in East Asia. As spatial resolution is finer, the skill of models is better, which can be seen from the two GCMs. According to the correlation coefficient, the improvement of resolution is more obvious in summer than in winter, especially for IsoGSM. In addition, the correlation coefficient in winter is usually larger than that in summer. The daily modelling has good potential to investigate the daily or synoptic climate information in water isotopes. The findings are useful for understanding the applicability of isotope-enabled models in East Asia and the climate factors influencing the skill of isotope-enabled models on a daily basis.

Dans le passé, le climat a varié de manière naturelle à différentes échelles de temps. Dans le cadre du changement climatique en cours, les climats passés, ou paléo-climats, sont une occasion unique de tester notre compréhension du climat et notre capacité à modéliser son évolution future. Dans cet article, nous présentons deux marqueurs des paléo-climats : la composition isotopique de l’eau et les lœss. Le Laboratoire de Météorologie Dynamique a été pionnier dans la modélisation numérique de ces marqueurs, et continue à jouer un rôle moteur dans leur utilisation pour mieux documenter les paléo-climats.

Based on a 6-year long record (2014-2020) of the isotopic composition of rain (delta 18Op) at Reunion Island (55 degrees E, 22 degrees S), in the South-West Indian Ocean, this study shows that the annual isotopic composition of precipitation in this region is strongly controlled by the number of cyclones, the number of best-track days, and the proportion of cyclonic rain during the year. Our results support the use of delta 18Op in annual-resolved tropical climate archives as a reliable proxy of past cyclone frequency. The influence of the proportion of cyclonic rain on the annual isotopic composition arises from the systematically more depleted precipitation and water vapor during cyclonic events than during less organized convective systems. The analysis of the daily to hourly isotopic composition of water vapor (delta 18Ov) during low-pressure systems and the reproduction of daily delta 18Ov observations by AGCMs with a global medium to coarse resolution (LMDZ-iso and ECHAM6-wiso) suggest that during cyclonic periods the stronger depletion mainly arises from both enhanced large-scale precipitation and water vapor-rain interactions under humid conditions. Water molecules have different forms: 2 atoms of hydrogen and 1 atom of oxygen 16 (majority form), or 1 atom of hydrogen, 1 atom of deuterium (2H) and 1 atom of oxygen 16, or 2 atoms of hydrogen and 1 atom of oxygen 18. The relative proportion of these different molecules in precipitation and water vapor is called isotopic composition. We have analyzed the isotopic composition of rain and water vapor at Reunion Island (55 degrees E, 22 degrees S), in the South-West Indian Ocean, for 6 years (2014-2020). We show that the annual isotopic composition of precipitation in this region is a reliable indicator of the number of tropical cyclones (TCs) during the year. This opens the possibility to use annual-resolved tropical climate archives of the isotopic composition of precipitation, like speleothems, to study how cyclone frequency in the Indian Ocean has varied in the past under different mean climates (warmer or colder). This study also seeks to understand why the isotopic composition of precipitation behaves in this way during TCs. We show that this is due to increased precipitation at the regional scale and enhanced exchanges of water molecules between rain and water vapor in the vicinity of TCs. The annual isotopic composition of precipitation on Reunion Island is controlled by the number of cyclones per year Both precipitation and water vapor are systematically more depleted during cyclonic periods More large-scale precipitation and rain-vapor interactions are responsible for the stronger isotopic depletion during tropical cyclones (TCs)

Water isotopes are one of the most widely used proxies in ocean climate research. However, there are still gaps in our understanding of the processes that control their composition.  Compared to other large ocean basins, the Mediterranean is ideally suited to improve our understanding of the processes influencing and driving oxygen isotopic variability, and to refine the current modelling approach. For the first time in a high-resolution Mediterranean dynamical model (NEMO-MED12), stable water isotopes (δ18O and δD) were successfully implemented and simulated in the whole basin. The well-known east-west gradient of δ18O in Mediterranean water masses is successfully simulated by the model. Results also show good agreement between simulated and observed δD. δD shows a strong linear relationship with δ18O (r2 = 0.98) and salinity (r2 = 0.94) for the entire Mediterranean basin. Furthermore, the modelled δ18O/salinity relationships are in good agreement with observations, with a weaker gradient simulated in the eastern basins than in the western basins. We investigate the relationship of the isotopic signature of the CaCO3 shell (δ18Oc) with temperature and the influence of seasonality. Our results suggest a more quantitative use of δ18O records, combining reconstruction with modelling approaches. This opens up broad perspectives for paleoclimate-related applications.

ECHAM6wiso and LMDZ6iso simulations, and python script analyzing model outputs, associated with the article : Amaelle Landais, Cécile Agosta, Françoise Vimeux, Olivier Magand, Cyrielle Solis, Alexandre Cauquoin, Niels Dutrievoz, Camille Risi, Christophe Leroy Dos Santos, Elise Fourré, Olivier Cattani, Bénédicte Minster, Frédéric Prié, Mathieu Casado, Aurélien Dommergue, Yann Bertrand, and Martin Werner (submitted to Atmospheric Chemistry and Physics, 2023) Modeling abrupt excursions in water vapor isotopic variability during cold fronts at the Pointe Benedicte observatory in Amsterdam Island. If you use the data or the python script, please cite the last version of this article available on https://www.egusphere.net/ or https://acp.copernicus.org/. Please also cite the articles related to the model simulations: Risi, C., Bony, S., Vimeux, F., and Jouzel, J.: Water-stable isotopes in the LMDZ4 general circulation model: Model evaluation for present-day and past climates and applications to climatic interpretations of tropical isotopic records, Journal of Geophysical Research Atmospheres, 115, https://doi.org/10.1029/2009JD013255, 2010. Cauquoin, A. and Werner, M.: High-Resolution Nudged Isotope Modeling With ECHAM6-Wiso: Impacts of Updated Model Physics and ERA5 Reanalysis Data, Journal of Advances in Modeling Earth Systems, 13, e2021MS002532, https://doi.org/10.1029/2021MS002532, 2021. Cauquoin, A., Werner, M., and Lohmann, G.: Water isotopes -- climate relationships for the mid-Holocene and preindustrial period simulated with an isotope-enabled version of MPI-ESM, Climate of the Past, 15, 1913–1937, https://doi.org/10.5194/cp-15-1913-2019, 2019.

Abstract Squall lines are substantially influenced by the interaction of low‐level shear with cold pools associated with convective downdrafts. Beyond an optimal shear amplitude, squall lines tend to orient themselves at an angle with respect to the low‐level shear. While the mechanisms behind squall line orientation seem to be increasingly well understood, uncertainties remain on the implications of this orientation. Roca and Fiolleau (2020, https://doi.org/10.1038/s43247-020-00015-4) show that long lived mesoscale convective systems, including squall lines, are disproportionately involved in rainfall extremes in the tropics. This article investigates the influence of the interaction between low‐level shear and squall line outflow on squall line generated precipitation extrema in the tropics. Using a cloud resolving model, simulated squall lines in radiative convective equilibrium amid a shear‐dominated regime (super optimal), a balanced regime (optimal), and an outflow dominated regime (suboptimal). Our results show that precipitation extremes in squall lines are 40% more intense in the case of optimal shear and remain 30% superior in the superoptimal regime relative to a disorganized case. With a theoretical scaling of precipitation extremes (C. Muller & Takayabu, 2020, https://doi.org/10.1088/1748-9326/ab7130), we show that the condensation rates control the amplification of precipitation extremes in tropical squall lines, mainly due to its change in vertical mass flux (dynamic component). The reduction of dilution by entrainment explains half of this change, consistent with Mulholland et al. (2021, https://doi.org/10.1175/jas-d-20-0299.1). The other half is explained by increased cloud‐base velocity intensity in optimal and superoptimal squall lines.

Investigating tropical squall lines with a cloud resolving model Using a cloud resolving model, we attempt to clarify the physical processes responsible for the organization of deep clouds into squall lines in the tropics. To do so, we impose a vertical wind shear, and investigate the response of deep convection to different shear strengths in radiative convective equilibrium. As the magnitude of the shear increases, the convection becomes more and more organized into a line, perpendicular to the shear. It is due to the interaction of the low-level shear with the cold pools associated with convective downdrafts. Beyond a certain shear, called optimal shear, the line tends to orient at an angle to the shear. The existing literature suggests that this angle conserves the projection of the shear on the direction perpendicular to the squall line near the optimal value, a hypothesis that we further investigate here. In this work, we propose a systematic method, based on image auto-correlation, to determine the angle of the squall line with respect to the shear. We highlight the existence of the sub-critical and super-critical regime, as predicted by earlier studies. In the sub-critical regime, squall lines are indeed perpendicular to the shear. Yet, angles of squall lines in the super-critical regime do not clearly correspond to the conservation of the projected component of the shear near the optimal value. In particular, squall lines often remain more perpendicular to the shear than expected. We thus investigate the balance between shear and cold pool winds to explain this difference. Using statistical methods on extreme events, we find that this difference is due to an intensification of cold pool potential energy with shear. Cold pool intensification allows the squall line to better resist to the shear, and thus reduces its angle of orientation. This new feature leads us to conclude that two mechanisms maintain a squall line in wind shear : the orientation of clouds and the intensification of cold pools.