Dissertations / Theses on the topic 'Paleoclimate modelling'

Hourly winter weather conditions of the Last Glacial Maximum (LGM) are simulated using the Community Climate Model version 3 (CCM3) on a globally resolved T170 (~75 km) grid. This simulation has been run in-tandem with a lower temporally resolved six-year climatological run. The purpose of the study is to determine: (1) whether examination of higher-resolution simulations, on both spatial and temporal scales, can enhance paleometeorological inferences based previously on monthly statistics of model output and (2) whether certain synoptic-scale events, which may have only a modest impact on seasonal statistics, might exert a disproportionate impact on geological climate records. Analysis is focused on changes in wind flow, no analogue climate “states”, synoptic scale events including Northern Hemisphere cyclogenesis, and gust events over glacial dust source regions. Results show a decrease in North Atlantic and increase in North Pacific cyclogenesis during the LGM. Storm trajectories react to the mechanical forcing of the Laurentide Ice Sheet, with Pacific storms tracking over middle Alaska and northern Canada and terminate in the Labrador Sea. The latter result supports observations and other model runs showing a significant reduction in Greenland winter precipitation. The modified Pacific track results in increased precipitation and the delivery of warmer air along the west coast of North America. This could explain “early” glacial warming inferred in this region from proxy climate records, potentially representing instead a natural regional response to ice age boundary conditions. Results also indicate a low variability, “no analogue” region just south of the Laurentide Ice Sheet margin which has appropriate conditions to harbour temperature-sensitive trees west of the Appalachian Mountains. Combined with pollen data, this lends valuable insight into the known disagreement between modern seed dispersal experiments and calculated migration rates. Finally, hourly-scale gust events over dust source regions during the LGM are two to five times greater than the modern, providing a mechanism to help explain the increased glacial dust load seen in the ice cores. Backwards air-parcel trajectories from Antarctic ice core locations show air sources over Patagonia and the Altiplano with some inputs from South Africa agreeing with recent isotopic tracer analyses. Results demonstrate that high temporal and spatial resolution simulations can provide valuable insight to add to the cornucopia of information already available from lower-resolution runs. They can also enhance our interpretation of geological records, which have been previously assumed to record longer time-scale climatological mean-states and thus ignoring any extreme synoptic events which may actually have had a disproportionate impact on their preservation.

In the recent past it has become evident that the Earth's climate is changing, and that human activity play a significant role in these changes. One of the regions where the ongoing climate change has been most evident is in the Arctic: the surface temperature has increased twice as much in this region as compared to the global average, in addition, a significant decline in the Arctic sea-ice extent has been observed in the past decades. Climate model studies of past climates are important tools to understand the ongoing climate change and how the Earth's climate may respond to changes in the forcing. This thesis includes studies of the Arctic climate in simulations of the early and mid-Holocene, 9 000 and 6 000 years before present. Changes in the Earth's orbital parameters resulted in increased summer insolation as compared to present day, especially at high northern latitudes. Geological data imply that the surface temperatures in the early to mid Holocene were similar to those projected for the near future. In addition, the geological data implies that the Arctic sea ice cover was significantly reduced in this period. This makes the early to mid-Holocene an interesting period to study with respect to the changes observed in the region at present. Several model studies of the mid-Holocene have been performed through the Paleoclimate Modeling Intercomparison Project (PMIP1 to PMIP3). The simulations have been performed with climate models of varying complexity, from atmosphere-only models in the first phase to fully coupled models with the same resolution as used for future climate simulations in the third phase. The first part of this thesis investigates the simulated sea ice in the pre-industrial and mid-Holocene simulations included in the PMIP2 and PMIP3 ensemble. As the complexity of the models increases, the models simulate smaller extents and thinner sea ice in the Arctic; the sea-ice extent suggested by the proxy data for the mid-Holocene is however not reproduced by the majority of the models. One possible explanation for the discrepancy between the simulated and reconstructed Arctic sea ice extent is missing or inadequate representations of important processes. The representation of atmospheric aerosol direct and indirect effects in past climates is a candidate process. Previous studies of deeper time periods have concluded that the representation of the direct and indirect effects of the atmospheric aerosols can influence the simulated climates, and reduce the equator to pole temperature gradient in past warm climates, in better agreement with reconstructions. The second part of the thesis investigates the influence of aerosol on the early Holocene climate. The indirect effect of reduced aerosol concentrations as compared to the present day is found to cause an amplification of the warming, especially in the Arctic region. A better agreement with reconstructed Arctic sea ice extent is thus achieved.
Under senare tid har det blivit uppenbart att jordens klimat håller på att förändras, och att mänsklig aktivitet spelar en viktig roll för dessa ändringar. Ett av de områden där den pägäende klimatfärändringen har varit tydligast är Arktis: temperaturen vid ytan har ökat dubbelt så mycket här jämfört med det globala genomsnittet. Dessutom har man observerat en betydande nedgång i havsisens utbredning i Arktis de senaste decennierna. Simuleringar gjorda med klimatmodeller av forntida klimat är viktiga verktyg för att förstå de pågående klimatförändringarna och hur jordens klimat påverkas av ändringar i klimatsystemets drivningar. Denna avhandling består av studier av det arktiska klimatet i modellsimuleringar av tidig och mid-holocen, ca. 9 000 och 6 000 år före nutid. Förändringar i jordens bana kring solen resulterade i en ökad sommar-solinstrålning jämfört med nutid, särskilt vid höga nordliga breddgrader. Geologiska data antyder att jordens temperatur vid ytan under denna period kan jämföras med dem vi förväntar för den närmaste framtiden. Vidare indikerar geologiska data att havsisen i Arktisk var kraftigt reducerad under denna period. Detta gör tidig till mid-holocen till en intressant period att studera, med avseende på de förändringar som för närvarande har observerats i området. Flera modellstudier av mitt-holocen har utförts i de olika faserna av Paleoclimate Modeling Intercomparison Project (PMIP1 till PMIP3). Simuleringarna har utförts med klimatmodeller av varierande komplexitet, från atmosfärsmodeller i den första fasen, till fullt kopplade modeller med hög rumslig upplösning i den tredje fasen. I den första delen av denna avhandling undersöks den simulerade havsisen i de förindustriella och mid-holocen simuleringar som ingår i PMIP2 och PMIP3 ensemblerna. Modellerna simulerar mindre utbredning och tunnare havsis i Arktis i den senare PMIP ensemblen, men fortfarande återskapar inte modellerna generelt den havsisutbredning som de geologiska data indikerar. En möjlig förklaring till skillnaderna mellan den simulerade och rekonstruerade havsisutsträckningen kan vara att viktiga processer i klimatsystemet saknas eller inte är tillräckligt väl beskrivna i modellerna. Beskrivningen av atmosfäriska aerosoler och dess effekter på klimatet är en möjlig kandidatprocess. Från studier av forntida varma tidsperioder har man dragit slutsatsen att beskrivningen av aerosoleffekterna påverkar det simulerade klimatet. Bland annat kan man minska temperaturgradienten mellan ekvator och polerna i tidigare varma klimat, vilket bättre överensstämmer med temperaturrekonstruktioner. Den andra delen av avhandlingen undersöker påverkan av aerosoler på klimatet under tidig holocen. Den indirekta effekten som följer av lägre aerosolkoncentrationer i tidig holocen jämfört med i dag, visar sig orsaka en förstärkning av uppvärmningen, särskilt i det arktiska områet, vilket stämmer bättre med havsisrekonstruktioner från denna period.

QC 20130910

Cette étude visait à développer un modèle dynamique du pergélisol-carbone à intégrer dans le modèle CLIMBER-2 et d'effectuer des simulations en vue de contribuer à la connaissance du cycle du carbone. Ce travail pourrait, pour la première fois, permettre une étude de modélisation avec un modèle de système terrestre qui comprendrait l'atmosphère dynamique, l'océan dynamique, la végétation dynamique et les composantes de la cryosphere, y compris les terres gelées, afin d'étudier le paléoclimat. La disponibilité des données récentes du CO2 et de δ13C de CO2 dans l'atmosphère fourni un moyen de valider les résultats du modèle pour déterminer si une dynamique pergélisol-carbone pourrait avoir joué un rôle important au cours des climats changeants.Un mécanisme pergélisol-carbone simplifié a été développé et validé et à été réglé en utilisant les données de la terminaison 1 (T1). Il a été constaté que, pour reproduire des données de CO2 et δ13C atmosphériques (pour l'atmosphère et l'océan) au cours de T1, une combinaison des mécanismes océaniques-glaciaires et pergélisol-carbone ont été nécessaires. Suite à cette constatation, plusieurs cycles glaciaires ont été modélisés pour étudier la sensibilité du mécanisme pergélisol-carbone aux forçages de CO2, les calottes glaciaires et l'insolation. l'étendue des calottes glaciaires a été jugée particulièrement importante pour le contrôle de la superficie des terres disponibles pour le pergélisol, et donc aussi pour la dynamique du carbone du pergélisol-carbone. La libération du carbone dans les sols de dégel en réponse à l'augmentation de l'insolation d'été dans les hautes latitudes, a été jugée très probable comme la source des hausses initiales de CO2 dans l'atmosphère au cours des terminaisons glaciaires.Les données CO2 de terminaison 1 peuvent être bien reproduits, y compris le plateau de CO2 BA / YD, quand le forçage de l'eau douce est appliqué à l'Atlantique nord. Expériences avec forçage de l'eau douce ont souligné l'importance du mécanisme du pergélisol-carbone dans l'évolution rapide des climats. Les augmentations très rapides des niveaux de CO2 dans l'atmosphère peuvent être expliqués par la libération rapide des sols en carbone en réponse à l'augmentation du transport de chaleur vers l'hémisphère nord. C'est en réponse à la reprise de l'AMOC suite d'un événement AMOC arrêt/réduction, tels que des événements D/O vu dans les données du δ18O Groenland. Les projections de changement climatique représentent des événements de réchauffement rapide. La conduite du modèle par des projections d'émissions (base de données RCP) a prédit l'augmentation du CO2 de pic et une plus longue période a des niveaux élevées de CO2 par rapport aux sorties du modèle qui ne comprennent pas les évaluations du pergélisol-carbone.L'analyse de δ13C de l'océan doit tenir compte de la dynamique du pergélisol et du carbone de la terre en général et de son effet sur les niveaux de δ13C atmosphériques. Si ce n'est pas pris en compte alors la circulation océanique peut être trop invoquée pour tenter d'expliquer les changements de δ13C de l'océan et du CO2 atmosphérique.Le mécanisme pergélisol-carbone réagit aux changements de température et amplifie la réponse du cycle du carbone. Il est fortement dépendant non seulement de l'apport d'énergie (qui détermine la température du sol et de l'emplacement du pergélisol), mais également de la surface de terres disponible à l'échelle mondiale sur laquelle le pergélisol peut exister. Afin de modéliser et de comprendre correctement la réponse du système terrestre dans les climats futurs et passés, le mécanisme de rétroaction pergélisol-carbone est un élément important du système. Ce travail a été une première étape pour aborder le rôle que la cryosphere terrestre joue dans le cycle du carbone et du système climatique sur de longues échelles de temps, et que d'autres études sont essentielles
This study aimed to develop a permafrost-carbon dynamic model to incorporate into the CLIMBER-2 Earth system model and to carry out simulations with a view to contributing to the knowledge of the carbon cycle. The work would, for the first time, allow a fully coupled modelling study with an earth system model which included dynamic atmosphere, ocean, vegetation and cryosphere components including frozen land to study paleoclimates. The availability of recent ice core data for CO2 and δ13C of atmospheric CO2 was to provide a means of validating model findings to identify whether a permafrost-carbon dynamic could have played a significant role in past changing climates.The deep Southern Ocean is an area of particular interest for glacial-interglacial CO2 variability, and current modelling efforts aim to recreate the observed CO2 changes using ocean mechanisms. These are often related to deep southern ocean carbon storage and release. So far the terrestrial biosphere has not been well-considered in transient simulations of the carbon cycle in Earth system models.A simplified permafrost-carbon mechanism was developed and validated and tuned using data from termination 1. It was found that in order to reproduce atmospheric CO2 and δ13C data (for atmosphere and ocean) during the termination, a combination of glacial ocean mechanisms and the permafrost-carbon mechanism was required. Following this finding, several glacial cycles were modelled to study the sensitivity of the permafrost-carbon mechanisms to CO2, ice sheets and insolation. Ice sheet extent was found to be particularly important in controlling the land area available for permafrost and therefore the carbon dynamics of permafrost-carbon. The permafrost-carbon mechanism, via carbon release from thawing soils responding to increasing summer insolation in higher northern latitudes, was found to very likely be the source of initial rises in CO2 on glacial terminations.Termination 1 CO2 data could be well reproduced, including the B-A/YD CO2 plateau, when fresh water forcing was applied to the north Atlantic. Fresh water forcing experiments pointed to the importance of the permafrost-carbon mechanism in fast changing climates. Very fast increases in atmospheric CO2 levels may be explained by fast soil-carbon release responding to increased heat transport to the northern hemisphere on AMOC resumption following an AMOC switch-off/reduction event, such as D/O events seen in the Greenland δ18O record. Future climate change projections represent fast warming events. Driving the model by emissions projections (RCP database) predicted increased peak CO2 and much longer term elevated CO2 levels relative to model outputs which did not include the permafrost carbon feedback.Analysis of ocean δ13C must take into account the dynamics of permafrost and land carbon in general and its effect on atmospheric δ13C levels. If this is not taken into account then ocean circulation may be over-invoked in attempting to explain changes in ocean δ13C and atmospheric CO2. The Earth system is not simply atmosphere and ocean. The findings in this work highlight that it is essential to consider land carbon dynamics when interpreting paleo-indicators for the carbon cycle.The permafrost-carbon mechanism reacts to temperature changes and amplifies the carbon cycle's response. It is stongly dependent not only on energy input (that determines soil temperature and permafrost location), but also on the area of land available globally on which it can exist. In order to properly model and understand the Earth system response to forcing in both future and past climates, the permafrost-carbon feedback mechanism is an important system component. This work has been a first step to address the role that the land cryosphere plays in the carbon cycle and climate system on long timescales, and further studies are essential

In this work, we want to investigate the influences of water masses on the basal melting under the RIS at present and in the past. In particular, the research aimed at understanding the influences of Ross Sea water masses variability on the RIS basal melting both at present and in the past. A regional adaptation of the Massachusetts Institute of Technology general circulation model (MITgcm) was implemented on the Ross Sea to simulate ocean circulation on the continental shelf and under the RIS. A present-day transient run, forced by ocean (GLORYS12V1) and atmospheric (ERA5) reanalysis over the period 1993-2018, shows that: [1] simulated water masses present different timescales of variability in their properties: Circumpolar Deep Water and Antarctic Surface Waters show a strong seasonal cycle, modulated by strong interannual variability. High Salinity Shelf Water and Low Salinity Shelf Water, on the other hand, show a weaker seasonal cycle and a decadal oscillation in their salinity. Variability of CDW and AASW is probably related to wind variability associated with the Southern Annular Mode, the Amundsen Sea Low, and El-Niño Southern Oscillation, mediated by sea ice. Variability of HSSW and LSSW is probably related to variability of the sea ice and meltwater input, and katabatic wind strength, in turn associated with the Polar Cell. The same variability is observed for the water masses beneath the RIS. [2] Basal melting presents a distinct pattern, related to the current at draft level, and variability related to the changing water masses properties. A new method based on mixing of water masses was developed to disentangle the effect of mixing, and highlight the melting variability associated to each water mass. Results show basal melting of ∼78 Gt/yr, in line with the observations, and presenting variability at the seasonal, interannual and decadal scale indicative of changing water masses properties or volume expansion inside the cavity. Then, we run 21 snapshots at intervals of 1000 years, over the Last Deglaciation (∼21-0 kyears BP): each snapshot was 26 years long and branched on a separate 120 years-long spinup. Simulations are forced by the outputs from an existent transient global paleoclimate experiment TraCE-21ka. The purpose of the paleo experiment was: 1) to analyse the evolution of the water masses with varying deglacial climatic conditions, and 2) how circulation resumed on the continental shelf, starting from a condition restricted by a grounded ice sheet at LGM (∼21 ka), and retreating during the deglaciation. Results show that: [1] initially, circulation was limited to three sub-ice shelf cavities in the Western Ross Sea. In Pennel trough warm CDW water reached the cavity, whereas in the Drygaslki and Joides troughs, HSSW filled the bottom level. [2] During the millenium following the Meltwater Pulse 1-A (14.6-14.3 ka), deep ocean warming and sub-surface ocean freshening caused a weakening of the Antarctic Slope Front, fostered CDW flow in Pennel and the Whales Deep cavity, which experienced high rates of basal melting. HSSW production in the Drygaslki and Joides stopped during this event. [3] In the Early Holocene (∼11.8 ka) grounding line retreat uncovered growingly portions of the continental shelf, allowing stronger atmospheric cooling and resumption of HSSW production. At ∼10ka the RIS cavity began to form, and was melted on the Westward side by HSSW, and on the Eastward side by advected mCDW; therefore, the stronger melting role shifted to the HSSW at that time.
In this work, we want to investigate the influences of water masses on the basal melting under the RIS at present and in the past. In particular, the research aimed at understanding the influences of Ross Sea water masses variability on the RIS basal melting both at present and in the past. A regional adaptation of the Massachusetts Institute of Technology general circulation model (MITgcm) was implemented on the Ross Sea to simulate ocean circulation on the continental shelf and under the RIS. A present-day transient run, forced by ocean (GLORYS12V1) and atmospheric (ERA5) reanalysis over the period 1993-2018, shows that: [1] simulated water masses present different timescales of variability in their properties: Circumpolar Deep Water and Antarctic Surface Waters show a strong seasonal cycle, modulated by strong interannual variability. High Salinity Shelf Water and Low Salinity Shelf Water, on the other hand, show a weaker seasonal cycle and a decadal oscillation in their salinity. Variability of CDW and AASW is probably related to wind variability associated with the Southern Annular Mode, the Amundsen Sea Low, and El-Niño Southern Oscillation, mediated by sea ice. Variability of HSSW and LSSW is probably related to variability of the sea ice and meltwater input, and katabatic wind strength, in turn associated with the Polar Cell. The same variability is observed for the water masses beneath the RIS. [2] Basal melting presents a distinct pattern, related to the current at draft level, and variability related to the changing water masses properties. A new method based on mixing of water masses was developed to disentangle the effect of mixing, and highlight the melting variability associated to each water mass. Results show basal melting of ∼78 Gt/yr, in line with the observations, and presenting variability at the seasonal, interannual and decadal scale indicative of changing water masses properties or volume expansion inside the cavity. Then, we run 21 snapshots at intervals of 1000 years, over the Last Deglaciation (∼21-0 kyears BP): each snapshot was 26 years long and branched on a separate 120 years-long spinup. Simulations are forced by the outputs from an existent transient global paleoclimate experiment TraCE-21ka. The purpose of the paleo experiment was: 1) to analyse the evolution of the water masses with varying deglacial climatic conditions, and 2) how circulation resumed on the continental shelf, starting from a condition restricted by a grounded ice sheet at LGM (∼21 ka), and retreating during the deglaciation. Results show that: [1] initially, circulation was limited to three sub-ice shelf cavities in the Western Ross Sea. In Pennel trough warm CDW water reached the cavity, whereas in the Drygaslki and Joides troughs, HSSW filled the bottom level. [2] During the millenium following the Meltwater Pulse 1-A (14.6-14.3 ka), deep ocean warming and sub-surface ocean freshening caused a weakening of the Antarctic Slope Front, fostered CDW flow in Pennel and the Whales Deep cavity, which experienced high rates of basal melting. HSSW production in the Drygaslki and Joides stopped during this event. [3] In the Early Holocene (∼11.8 ka) grounding line retreat uncovered growingly portions of the continental shelf, allowing stronger atmospheric cooling and resumption of HSSW production. At ∼10ka the RIS cavity began to form, and was melted on the Westward side by HSSW, and on the Eastward side by advected mCDW; therefore, the stronger melting role shifted to the HSSW at that time.

Les enregistrements climatiques globaux à l’échelle géologique entre le Crétacé et le début du Cénozoïque indiquent des variations de grande amplitude. Sur le long terme, celles-ci sont déterminées par l’équilibre entre la composition atmosphérique en gaz à effet de serre, principalement le CO2, issus du dégazage volcanique et l’altération continentale, modulée par les mouvements tectoniques des continents. Dans cette thèse, les liens entre paléogéographie et CO2 ont été étudiés dans le contexte des interactions entre climat et calottes de glace au cours d’un intervalle de temps dit de « greenhouse », entre 120 et 34 Ma. L’utilisation d’une suite de modèles impliquant un modèle couplé moyenne résolution, un modèle atmosphérique haute résolution et un modèle de calotte de glace, a permis de montrer que les changements paléogéographiques survenant au Crétacé ont régulé la présence de glace en Antarctique. Dans un second temps, une nouvelle méthode de couplage climat-calotte a été développée pour étudier la glaciation Eocène-Oligocène. Ces développements ont permis de reconstruire une évolution fidèle de celle-ci, en bon accord avec les données. Deux rétroactions liées à cette glaciation et à la chute concomitante du CO2 atmosphérique sont étudiées. En premier lieu, l’impact de la glaciation sur le Courant Circumpolaire Antarctique est abordé, montrant que celle-ci génère une intensification de ce courant. Ensuite, au sein d’une étude mêlant données et modèles pour documenter la présence de moussons en Asie dès l’Eocène moyen, il est montré que le changement climatique de la fin de l’Eocène induit une baisse d’intensité de la mousson asiatique. Enfin, dans la perspective d’analyser les conséquences des changements paléogéographiques du Cénozoïque sur la biogéochimie marine, des tests de sensibilité aux passages océaniques de Panama et de Drake ont été réalisés
On geological timescales, global climate proxies indicate that variations of large magnitude occur between the Cretaceous and the Cenozoic. On the long term, these variations are mostly determined by the equilibrium between the greenhouse gases composition of the atmosphere, primarily the CO2, and continental weathering set up by the spatial location of Earth’s landmasses. Here, the links between paleogeography and CO2 are looked upon in a climate-ice sheet interactions framework during a greenhouse period of Earth history (120 – 34 Ma). A suite of models involving both coupled and ice sheet models have been used to demonstrate that paleogeographic reorganizations have regulated the presence of ice over Antarctica during the Cretaceous. In a second time and using a similar setup, a new method for climate-ice sheet coupling have been developed and applied to the Eocene-Oligocene (EO) glaciation to yield a new scenario of ice evolution, in good agreement with data. Two feedbacks related to this glaciation and the coeval atmospheric CO2 fall are investigated. First, it is shown that the EO glaciation generates an intensification of the Antarctic Circumpolar Current. Second, within a data-model study demonstrating active Asian monsoons as old as the mid-Eocene, it is shown that the climatic change at the end of the Eocene is responsible for a reduction in the intensity of the Asian monsoon. Finally, with the aim of analysing the effect of paleogeographic changes on marine biogeochemistry during the Cenozoic, sensitivity tests to Drake Passage and Panama Seaway have been carried out

Au Crétacé, l'évolution foliaire des plantes à fleurs, ou Angiospermes, vers de fortes densités de nervures et de stomates, suggère une augmentation de la conductance stomatique et des flux d’évapotranspiration sans précédent. Cependant, ces paléo-traits ne sont pas pris en compte dans les modèles de végétation qui visent justement à déterminer les effets de l’évapotranspiration sur le climat. L’objectif de ma thèse est donc de modéliser l’évolution de la conductance stomatique des plantes à fleurs au cours du Crétacé et d’en évaluer ses effets sur les interactions et rétroactions climat-végétation. En combinant des données fossiles et des modèles écophysiologiques, je développe une paramétrisation innovante de la végétation proto-angiosperme dans le modèle de végétation ORCHIDEE qui considère une réduction conjointe de leurs capacités hydrauliques et photosynthétiques. Avec le modèle couplé atmosphère-végétation LMDZOR, je montre que la radiation des Angiospermes génère un renforcement du cycle hydrologique et une baisse de la température de surface, dont les intensités sont modulées par la teneur en CO2 atmosphérique. En activant le modèle de végétation dynamique, je montre que la radiation des plantes à fleurs génère des boucles de rétroactions positives dans un contexte de baisse de la teneur en CO2 atmosphérique au cours du Crétacé : l’augmentation des capacités hydrauliques et photosynthétiques des plantes à fleurs constitue un avantage sélectif par rapport aux autres types de plantes qui leur permet de (i) maintenir leur productivité, (ii) développer des forêts tropicales et remplacer les conifères dans les forêts tempérées et boréales et (iii) renforcer les précipitations, limitant ainsi les effets du stress hydrique sur leur propre essor
During the Cretaceous period, the leaf evolution of flowering plants, or angiosperms, towards higher vein and stomata densities, suggests an unprecedented increase in stomatal conductance as well as evapotranspiration fluxes. However, these paleo-traits are not accounted for in vegetation models whose the aim is to evaluate the effects of evapotranspiration fluxes on the climate system. The purpose of this study is to simulate the stomatal conductance evolution of flowering plants through the Cretaceous period and assess their effects on interaction and feedback between climate and vegetation. By combining fossil data and empirical ecophysiological models, I develop an innovative parameterization of proto-angiosperm vegetation in the ORCHIDEE vegetation model which considers a reduction of both hydraulic and photosynthetic capacities. Thanks to the coupled atmosphere-vegetation model LMDZOR, I demonstrate that the radiation of flowering plants drives a strengthening of the hydrologic cycle and a decrease in surface temperature, the intensities of which are modulated by the atmospheric concentration of CO2. By activating the dynamic vegetation model, I show that flowering plant radiation triggers positive feedback loops in a context of decreasing atmospheric concentration of CO2 during the Cretaceous period: the increase in hydraulic and photosynthetic capacities of flowering plants constitutes a selective advantage compared to other types of plants by allowing them to (i) sustain their productivity, (ii) develop tropical forests and replace conifers in temperate and boreal forests and (iii) enhance precipitations, thus preventing water stress effects on their own development

The Earth's climate system is driven by varying insolation from the Sun. The dominant variations in insolation are at 23 and 40 thousand year periods, yet for the past million years the Earth's climate has glacial cycles at approximately 100 kyr periodicity. These cycles are a coupled variation in temperature, ice volume, and atmospheric CO₂. Somehow the Earth system's collective response to 23 and 40 kyr insolation forcing produces 100 kyr glacial-interglacial cycles. Generally it has been assumed that the causative mechanisms are a combination of ice dynamics (high ice reflectivity controlling temperature) and ocean circulation (changing carbon partitioning between the deep ocean and the atmosphere, and heat transport to the poles). However, these proposed mechanisms have not yet resulted in a compelling theory for all three variations, particularly CO₂. This thesis explores the role of volcanic CO₂ emissions in glacial cycles. I calculate that glacial-driven sea level change alters the pressure on mid-ocean ridges (MORs), changing their CO₂ emissions by approximately 10%. This occurs because pressure affects the thermodynamics of melt generation. The delay between sea level change and the consequent change in MOR CO₂ emissions is several tens-of-thousands-of-years, conceptually consistent with a coupled non-linear oscillation that could disrupt glacial cycles from a 40 kyr mode to a multiple of that period. I develop an Earth system model to investigate this possibility, running for approximately one million years and explicitly calculating global temperatures, ice sheet configuration, and CO₂ concentration in the atmosphere. The model is driven by insolation, with all other components varying in response (and according to their own interactions). This model calculates that volcanism is capable of causing a transition to ̃100 kyr glacial cycles, however the required average volcanic CO₂ emissions are barely within the 95% confidence interval. Therefore it is possible for volcanic systems and glacial cycles to form a 100 kyr coupled oscillation.

Les enregistrements climatiques globaux à l’échelle géologique entre le Crétacé et le début du Cénozoïque indiquent des variations de grande amplitude. Sur le long terme, celles-ci sont déterminées par l’équilibre entre la composition atmosphérique en gaz à effet de serre, principalement le CO2, issus du dégazage volcanique et l’altération continentale, modulée par les mouvements tectoniques des continents. Dans cette thèse, les liens entre paléogéographie et CO2 ont été étudiés dans le contexte des interactions entre climat et calottes de glace au cours d’un intervalle de temps dit de « greenhouse », entre 120 et 34 Ma. L’utilisation d’une suite de modèles impliquant un modèle couplé moyenne résolution, un modèle atmosphérique haute résolution et un modèle de calotte de glace, a permis de montrer que les changements paléogéographiques survenant au Crétacé ont régulé la présence de glace en Antarctique. Dans un second temps, une nouvelle méthode de couplage climat-calotte a été développée pour étudier la glaciation Eocène-Oligocène. Ces développements ont permis de reconstruire une évolution fidèle de celle-ci, en bon accord avec les données. Deux rétroactions liées à cette glaciation et à la chute concomitante du CO2 atmosphérique sont étudiées. En premier lieu, l’impact de la glaciation sur le Courant Circumpolaire Antarctique est abordé, montrant que celle-ci génère une intensification de ce courant. Ensuite, au sein d’une étude mêlant données et modèles pour documenter la présence de moussons en Asie dès l’Eocène moyen, il est montré que le changement climatique de la fin de l’Eocène induit une baisse d’intensité de la mousson asiatique. Enfin, dans la perspective d’analyser les conséquences des changements paléogéographiques du Cénozoïque sur la biogéochimie marine, des tests de sensibilité aux passages océaniques de Panama et de Drake ont été réalisés
On geological timescales, global climate proxies indicate that variations of large magnitude occur between the Cretaceous and the Cenozoic. On the long term, these variations are mostly determined by the equilibrium between the greenhouse gases composition of the atmosphere, primarily the CO2, and continental weathering set up by the spatial location of Earth’s landmasses. Here, the links between paleogeography and CO2 are looked upon in a climate-ice sheet interactions framework during a greenhouse period of Earth history (120 – 34 Ma). A suite of models involving both coupled and ice sheet models have been used to demonstrate that paleogeographic reorganizations have regulated the presence of ice over Antarctica during the Cretaceous. In a second time and using a similar setup, a new method for climate-ice sheet coupling have been developed and applied to the Eocene-Oligocene (EO) glaciation to yield a new scenario of ice evolution, in good agreement with data. Two feedbacks related to this glaciation and the coeval atmospheric CO2 fall are investigated. First, it is shown that the EO glaciation generates an intensification of the Antarctic Circumpolar Current. Second, within a data-model study demonstrating active Asian monsoons as old as the mid-Eocene, it is shown that the climatic change at the end of the Eocene is responsible for a reduction in the intensity of the Asian monsoon. Finally, with the aim of analysing the effect of paleogeographic changes on marine biogeochemistry during the Cenozoic, sensitivity tests to Drake Passage and Panama Seaway have been carried out

La période froide du Dernier Maximum Glaciaire était caractérisée, en regard de notre climat moderne, par une couverture de glace de mer australe accrue, une circulation profonde Atlantique moins profonde et une plus faible concentration en CO2 dans l’atmosphère. Ces différences sont bien connues grâce aux observations indirectes mais difficiles à représenter dans les simulations issues des modèles de climat. En effet, ces modèles simulent fréquemment une concentration en CO2 atmosphérique trop élevée, une circulation océanique trop profonde dans l’Atlantique et une banquise présentant une distribution trop circulaire dans l’océan austral ainsi qu’une étendue hivernale et une amplitude saisonnière trop faibles. Ces désaccords modèle-données observés au Dernier Maximum Glaciaire remettent en cause la représentation numérique de certains processus climatiques essentiels. Plusieurs études soulignent le rôle majeur de la glace de mer australe sur la capacité de stockage de carbone de l’océan et la circulation océanique profonde. Je me suis donc focalisée sur cette région pour mieux com-prendre les processus associés à ce stockage. Grâce aux simulations réalisées avec le modèle système terre iLOVECLIM, j’ai pu démontrer que les incertitudes liées à la représentation des calottes polaires ont un impact limité sur les variables examinées ici. En revanche, d’autres choix de conditions aux limites (affectant le volume de l’océan, l’ajustement de l’alcalinité) peuvent entraîner des modifications importantes du contenu total en carbone de l’océan. Je montre également que l’utilisation d’une paramétrisation simple de la plongée des saumures résultant de la formation de glace de mer permet d’améliorer significativement la simulation de la glace de mer australe, de la circulation océanique profonde et de la concentration en CO2 atmosphérique. Un ensemble de simulations incluant l’impact de différentes paramétrisations océaniques est utilisé pour montrer que la circulation océanique très profonde simulée par notre modèle ne peut être attribuée à une glace de mer australe insuffisante. En revanche, les processus de convection dans l’océan austral semblent clefs pour améliorer à la fois la glace de mer australe, la circulation océanique profonde et la concentration en CO2 atmosphérique auDernier Maximum Glaciaire
Compared to the present-day climate, the cold period of the Last Glacial Maximum was characterized by an expanded sea-ice cover in the Southern Ocean, a shoaled Atlantic deep ocean circulation and a lower atmospheric CO2 concentration. These changes are well-documented by indirect observations but difficult to represent in simulations of climate models. Indeed, these models tend to simulate a too high atmospheric CO2 concentration, a too deep Atlantic deep ocean circulation, and a sea-ice cover with a too circular distribution in the Southern Ocean and a too small winter extent and seasonal amplitude. The model-data discrepancies observed at the Last Glacial Maximum call into question the model representation of some important climate processes. Several studies have underlined the crucial role of the Southern Ocean sea ice on ocean carbon storage capacity and deep circulation. I have therefore focussed on this region to improve our understanding of the processes associated with this storage. Thanks to simulations performed with the Earth System Model iLOVECLIM, I have demonstrated thatthe uncertainties related to ice sheet reconstructions have a limited impact on the variables examined in this study. In contrast, other choices of boundary conditions (influencing the ocean volume and alkalinity adjustment) can yield large changes of carbon sequestration in the ocean. I also show that a simple parameterization of the sinking of brines consequent to sea-ice formation significantly improves the simulated Southern Ocean sea ice, deep ocean circulation and atmospheric CO2 concentration. A set of simulations including the effects of diverse ocean parameterizations is used to show that the too deep ocean circulation simulated by our model cannot be attributed to an insufficient sea-ice cover, whereas convection processes in the Southern Ocean seem crucial to improve both the Southern Ocean sea ice, the deep ocean circulation and the atmospheric CO2 concentration at the Last Glacial Maximum

Ce travail traite de la simulation numérique du climat des grandes calottes de glace, en particulier des calottes de l'Antarctique et du Groenland, toujours existantes, dans des conditions climatiques différentes, à l'aide de modèles de circulation générale de l'atmosphère (MCGA). Le MCGA à grille variable LMDz a été adapté aux spécificités du climat polaire et validé pour le climat actuel. L'approche d'une grille variable, qui permet d'utiliser le MCGA à haute résolution spatiale (autour de 100 km) sur la région d'intérêt à un coût numérique raisonnable, a été validée en analysant la dynamique atmosphérique au bord de la région ciblée à l'aide d'un schéma de suivi des cyclones individuels. Des simulations du climat du Dernier Maximum Glaciaire (DMG) ont été faites pour le Groenland et l'Antarctique et analysées en tenant compte des archives glaciaires disponibles. Une explication possible des différences entre les deux méthodes principales de reconstruction des paléotempératures - l'analyse des isotopes de l'eau et la mesure directe de la température de la glace dans le trou de forage - au centre du Groenland a pu être proposée. Cette explication est basée sur des changements de paramètres climatiques locaux. C'est la première fois que l'approche de grille variable a été utilisée dans un MCGA pour des simulations du climat polaire à l'échelle de quelques années. Les simulations paléoclimatiques faites avec LMDz sont à une résolution spatiale inégalée à ce jour. Finalement, le climat du DMG, simulé par plusieurs MCGA dans le cadre du projet international PMIP (Paleoclimate Modelling Intercomparison Programme), a été analysé, et des implications des résultats pour l'interprétation des enregistrements glaciaires ont été discutées.

The complex demographic changes that underlie the expansion of anatomically modern humans out of Africa have important consequences on the dynamics of natural selection and our ability to detect it. In this thesis, I aimed to refine our knowledge on human population history using ancient genomes, and then used a climate-informed, spatially explicit framework to explore the interplay between complex demographies and selection. I first analysed a high-coverage genome from Upper Palaeolithic Romania from ~37.8 kya, and demonstrated an early diversification of multiple lineages shortly after the out-of-Africa expansion (Chapter 2). I then investigated Late Upper Palaeolithic (~13.3ky old) and Mesolithic (~9.7 ky old) samples from the Caucasus and a Late Upper Palaeolithic (~13.7ky old) sample from Western Europe, and found that these two groups belong to distinct lineages that also diverged shortly after the out of Africa, ~45-60 ky ago (Chapter 3). Finally, I used East Asian samples from ~7.7ky ago to show that there has been a greater degree of genetic continuity in this region compared to Europe (Chapter 4). In the second part of my thesis, I used a climate-informed, spatially explicit demographic model that captures the out-of-Africa expansion to explore natural selection. I first investigated whether the model can represent the confounding effect of demography on selection statistics, when applied to neutral part of the genome (Chapter 5). Whilst the overlap between different selection statistics was somewhat underestimated by the model, the relationship between signals from different populations is generally well-captured. I then modelled natural selection in the same framework and investigated the spatial distribution of two genetic variants associated with a protective effect against malaria, sickle-cell anaemia and β⁰ thalassemia (Chapter 6). I found that although this model can reproduce the disjoint ranges of different variants typical of the former, it is incompatible with overlapping distributions characteristic of the latter. Furthermore, our model is compatible with the inferred single origin of sickle-cell disease in most regions, but it can not reproduce the presence of this disorder in India without long-distance migrations.

Le système turbiditique du Zambèze (Canal du Mozambique, Océan Indien occidental) est l'un des plus grands systèmes turbiditiques au monde et reste encore mal compris. L'acquisition récente de données bathymétriques multifaisceaux à haute résolution, de données de sismique réflexion haute et très haute résolution et de données sédimentologiques a permis d'étudier l'évolution de l'architecture et l'organisation des dépôts depuis l'Oligocène afin de comprendre les principaux facteurs de forçage qui contrôlent la sédimentation en eau profonde dans le Canal du Mozambique. Le système turbiditique du Zambèze est composé de deux systèmes de dépôt adjacents : l'éventail du Zambèze ("Zambezi Fan") et un éventail semi-confiné ("ponded fan") dans un bassin intermédiaire face à l'embouchure du Zambèze. Les résultats et les interprétations indiquent : (1) un important contrôle tectonique depuis le Miocène responsable d'une sur-incision profonde de la vallée du Zambèze et de débordements limités des courants turbiditiques ; (2) une influence importante des courants de fond qui induisent la rareté des turbidites fines, l'érosion des flancs des vallées et l'apparition généralisée de "sediment waves" ; (3) une faible activité turbiditique au cours des 700 derniers kyr qui ne montre, en outre, aucune relation avec les changements du niveau de la mer, l'activité turbiditique s'observant indépendamment des périodes glaciaires et interglaciaires ; (4) des pics de flux terrigènes corrélés aux maxima d'ensoleillement estival local, indiquant que la mousson est le contrôle majeur des apports de sédiments vers le système de dépôt marin profond ; (5) une évolution "on-off" du l'éventail du Zambèze qui démontre un déplacement du dépocentre de la partie distale de l'éventail vers le bassin intermédiaire proximal. Ces résultats soulignent la grande complexité du système turbiditique du Zambèze en raison de l'impact de facteurs de contrôles multiples
The Zambezi turbidite system (Mozambique Channel, Western Indian Ocean) is one of the largest turbidite systems in the world and yet still remains poorly understood. Newly acquired high-resolution multibeam bathymetry, seismic reflection and sedimentological data allowed to investigate the architecture evolution and depositional patterns since the Oligocene in order to understand the main forcing factors that control the deep sea sedimentation in the Mozambique Channel. It was found that the Zambezi turbidite system is composed of two adjacent depositional systems: the channelized Zambezi Fan and a semiconfined fan in the lntermediate Basin. Moreover, results and interpretations indicate: (1) important tectonic control since the Miocene that caused deep incision of the Zambezi Valley and limited overflow of turbidite currents; (2) an important influence of bottom-currents that induces scarcity of fine-grained turbidites, valley flanks erosion and widespread occurrence of sediment waves; (3) low turbidite activity for the last 700 kyr that shows no relationship with sea-level changes as turbidite activity occurred irrespective of glacial or interglacial periods; (4) peaks in terrigenous flux with maxima in local summer insolation, reflecting that monsoon controls the sediment inputs towards the deep marine depositional system; (5) an on-off evolution of the Zambezi Fan that demonstrates a depocenter shift from the distal Zambezi Fan to the proximal Intermediate Basin. All our findings underline the high complexity in depositional environments of the Zambezi turbidite system

Le dernier interglaciaire, qui s'étend de 129 000 à 116 000 ans avant notre ère, est l’une des périodes les plus chaudes de ces 800 000 dernières années. Cette période se caractérise par une distribution saisonnière et latitudinale de l’insolation différente de l’actuel, se manifestant par une hausse des températures dans les hautes latitudes de l’hémisphère nord par rapport à la période pré-industriel (1850). Durant cette période, l’élévation du niveau marin (+ 6 à 9 m par rapport au niveau actuel) indique que les calottes polaires étaient moins volumineuses qu'aujourd'hui. Le dernier interglaciaire est donc un sujet d'étude important étant donnés les risques de fonte des calottes glaciaires sous l'influence du réchauffement actuel et à venir. C’est aussi un bon cas d’étude pour identifier et quantifier les mécanismes à l’origine de l’amplification polaire dans un contexte de climat chaud.Dans le cadre de l’exercice d’intercomparaison de modèles CMIP6-PMIP4, j’ai analysé la simulation à climat constant lig127k réalisée au LSCE, à partir du modèle climatique IPSL-CM6A-LR. En Arctique (60-90°N), les variations d’insolation induisent un réchauffement annuel de 0,9°C par rapport à la période pré-industrielle, pouvant atteindre jusqu’à 4,0°C en automne. L’étude du bilan énergétique de la région Arctique a mis en évidence les rôles fondamentaux des variations de la couverture de glace de mer, du stockage de chaleur dans l’océan, ainsi que des changements des propriétés optiques des nuages sur le réchauffement de l’Arctique il y a 127 000 ans.En réponse au changement climatique du dernier interglaciaire, le modèle de calottes GRISLI simule une perte de 10,7 à 57,1% du volume de glace au Groenland. Ce recul de la calotte groenlandaise se traduit par 1) une hausse maximale du niveau marin estimée entre 0,83 et 4,35 m et 2) des interactions climat-calotte engendrant un réchauffement additionnel maximal de 0,2°C à l’échelle de la région arctique. Ces estimations illustrent bien le rôle important des calottes dans le système climatique et rappellent l’importance de coupler les modèles climatiques aux modèles de calottes. Dans ce cadre, une étude préliminaire a été menée à l’aide du modèle atmosphérique ICOLMDZOR v7, utilisant le nouveau cœur dynamique DYNAMICO développé à l’IPSL. Elle a montré que l’utilisation de champs atmosphériques à haute résolution améliorait le calcul du bilan de masse à la surface des calottes polaires. Elle encourage également le couplage asynchrone entre le modèle climatique de l’IPSL, le modèle ICOLMDZOR v7 et le modèle de calotte GRISLI.Enfin, l’analyse du bilan énergétique de la région Arctique à partir d’une expérience idéalisée, dans laquelle la concentration en CO2 atmosphérique augmente de 1% par an, a révélé que des processus similaires sont l'origine du réchauffement de l’Arctique au dernier interglaciaire et celui d’un futur proche
The Last Interglacial (129 -116 ka BP) is one of the warmest periods in the last 800 ka at many locations. This period is characterized by a strong orbital forcing leading to a different seasonal and latitudinal distribution of insolation compared to today. These changes in insolation result in a temperature increase in the high latitudes of the Northern Hemisphere and a rise in sea level of 6 to 9 m above present. Therefore, the Last Interglacial represents a good case study given the risks of melting ice sheets under the influence of current and future warming. It is also an opportunity to identify and quantify the mechanisms causing polar amplification in a warmer climate than today.Within the framework of the CMIP6-PMIP4 model intercomparison project, I analyzed the lig127k snapshot run with the IPSL-CM6A-LR climate model. In the Arctic region (60-90°N), the insolation variations induce an annual warming of 0.9°C compared to the pre-industrial period (1850) reaching up to 4.0°C in autumn. Investigate changes in the Arctic energy budget relative to the pre-industrial period highlights the crucial roles of changes in the sea ice cover, ocean heat storage and clouds optical properties in the Last Interglacial Arctic warming.As a result of climate change over the Last Interglacial, the GRISLI ice sheet model simulates a Greenland ice loss of 10.7-57.1%, corresponding to a sea level rise of 0.83-4.35 m and a 0.2°C additional warming in the Arctic region. These estimates illustrate the crucial role of polar ice sheets in the climate system. To better assess ice sheet-climate feedbacks in the Arctic, I have therefore carried out a preliminary study using the ICOLMDZOR model that includes the new dynamical core DYNAMICO developed at the IPSL. This study shows that the use of high-resolution atmospheric fields improves the calculation of the surface mass balance in Greenland.Finally, the comparison between past and future Arctic energy budget reveals that the processes causing Arctic warming during the Last Interglacial and the near future are similar

Deep atmospheric convection over the western equatorial Pacific occurs at the junction of the rising limbs of the meridional Hadley cells and the Pacific Walker circulation, making it one of the most atmospherically dynamic regions on Earth. Here, interactions between the Australasian monsoon and atmospheric convection result in highly variable regional precipitation patterns across different latitudes. The dynamics of the Australasian monsoon over the past ~40,000 years are relatively well understood at its northern limit (the East Asian Summer Monsoon), but less well known for its southern limit (the Indo-Australian Summer Monsoon). In the equatorial region however, even less is known about the past behaviour and dynamics of this major system. Here we present a new, continuous, absolutely dated speleothem record from southwest Sulawesi, Indonesia that spans the past 40,000 years. Isotopic ratios of oxygen (δ18O) and carbon (δ13C) in the speleothem calcite were analysed at ~50-yr resolution to reconstruct rainfall amount and vegetation productivity. The records show that the strength of regional deep atmospheric convection is primarily controlled by sea level via the exposure and inundation of the Sunda Shelf. This sea-level control results in a relatively dry last glacial period that was terminated by the onset of deglaciation and the inundation of the Sunda Shelf, which abruptly increased the intensity of deep atmospheric convection. The Sulawesi speleothem δ18O record does not capture millennial-scale variability in response to North Atlantic Heinrich events, in contrast to nearby speleothem records from Borneo and Flores. To explore this observation, the climatic impact of Heinrich events in the western equatorial Pacific region was simulated using idealised North Atlantic freshwater hosing experiments performed with the HadCM3 and CSIRO Mk3L general circulation models. Precessional forcing is shown to influence the manifestation of Heinrich events, particularly across the Southern Hemisphere via the varying response of the Intertropical Convergence Zone. Additionally, high atmospheric carbon dioxide levels increase the duration of the Heinrich climate anomaly, compared to pre-industrial levels. Sulawesi speleothem δ13C is interpreted as a record of changing vegetation productivity. Comparison of the speleothem carbon isotopes with ice core atmospheric methane concentrations reveals a significant relationship during the glacial and early-deglacial intervals. It is hypothesised that changing vegetation productivity as recorded by Sulawesi speleothems is indicative of broader tropical methane emissions, which are thought dominate the glacial methane budget. This idea is explored using the Sheffield Dynamic Global Vegetation Model to simulate global climate and methane emissions over the past 40,000 years. The data-model comparisons confirm that temporal changes in the Sulawesi δ13C record are in good agreement with modelled methane emissions over much of the tropics, lending weight to the likelihood that the tropics dominated total methane emissions during the glacial period when boreal sites were perennially frozen. Together this work demonstrates the importance of the western equatorial Pacific in influencing regional climate and global climate signals. It is vital therefore, to continue to explore the past dynamics of this region as a potential driver of global climate changes.