Atmospheric and Oceanic Sciences

Sea salt aerosols (SSA) play a vital role in Earth’s climate system by influencing radiative balance and cloud formation. However, discrepancies between observed and modeled SSA concentrations create uncertainties in climate projections. This dissertation examines SSA emissions, radiative effects, and interactions with the global climate, focusing on their role in atmospheric processes, climate modeling, and geoengineering. Chapter 1 identifies emission errors, vertical transport errors, and subgrid variability as key sources of discrepancies between modeled and observed SSA concentrations. By combining in situ measurements from NOAA PMEL cruise expeditions with GEOS-Chem model simulations, this chapter highlights the need for integrated approaches to improve SSA representation in global climate models (GCMs). Chapter 2 evaluates the performance of various sea spray generation functions in representing SSA emissions and aerosol optical depth (AOD). Comparisons of four sea spray generation functions in a NOAA GFDL atmosphere model reveal that simulations using the mechanistic formulation by \citet{deike2022mechanistic} achieve the highest correlation with global AOD observations. This chapter emphasizes the regional variability in SSA emission simulations and the challenges of reconciling AOD retrieval methodologies with model outputs. Chapter 3 explores the climatic impacts of geoengineering scenarios by examining the rapid adjustments and atmospheric circulation responses to increased tropical SSA emissions under the G4sea-salt protocol in a NOAA GFDL coupled model. Enhanced tropical SSA emissions induce cooling, reducing equator-to-midlatitude temperature gradients and weakening surface westerlies, which in turn partially offset increased emissions. The analysis reveals significant differences in climate responses to zonally uniform versus regionally targeted SSA geoengineering scenarios, providing critical insights for marine cloud brightening strategies. Chapter 4 assesses radiative forcing and climate feedback parameters in the context of SSA geoengineering, comparing G4sea-salt simulations with the SSP2-4.5 baseline. By isolating the effects of SSA geoengineering, this chapter demonstrates that geoengineered forcings and feedbacks can be analyzed similarly to traditional climate forcings, though nonlinearities in Earth's energy imbalance (EEI) complicate attribution. Collectively, this dissertation advances the understanding of SSA dynamics, their representation in climate models, and their potential role in geoengineering, providing critical insights for future climate mitigation and adaptation strategies.

Oxygen minimum zones (OMZs) are naturally occurring ocean regions with extremely low oxygen levels, maintained by a combination of sluggish physical oxygen supply and high biological demand. Chapters 1 and 2 introduce the physical and biogeochemical mechanisms that form OMZs and explore how these processes are being affected by climate change. As the ocean deoxygenates with warming, uncertainty remains over whether OMZs will intensify or diminish. This thesis examines the future evolution of OMZs and their ecological consequences. Chapter 3 presents a novel water mass framework to analyze global oxygen dynamics using an ensemble of Earth System Model (ESM) projections. The framework identifies consistent patterns across the ensemble and enables improved constraints on future oxygen levels by evaluating model skill. Models that better match present-day observations project weaker global deoxygenation and increased oxygen supply to OMZs under warming compared to more biased models. Chapter 4 focuses on regional projections for the tropical Pacific and Indian Ocean OMZs. Across models, OMZs show a consistent three-regime response: core regions contract and peripheral layers expand, while an intermediate transition zone between exhibits little change in volume. This interpretive framework reconciles discrepancies between models, observational data, and paleoceanographic records. In the Indian Ocean, oxygen transport via the Indonesian Throughflow and marginal seas emerges as a key driver of the OMZ response. Chapter 5 explores ecological impacts in the tropical Pacific using an ecosystem model applied to ESM projections. Although oxygen increases near OMZ cores, warming raises fish metabolic demand and offsets potential habitat gains. Finally in Chapter 6, I develop and apply a high-resolution regional model of the Indian Ocean to assess the role of tidal processes on biogeochemical cycling. While tides enhance productivity and respiration, they also promote vertical mixing that supports oxygen supply to the OMZ, likely sustaining higher thermocline oxygen concentrations overall. I conclude the thesis by synthesizing these findings to reassess the robustness of our understanding for how OMZs will respond in a warming world.

This thesis describes some theoretical and numerical modeling work with the aim of elucidating the dynamics of the Madden-Julian Oscillation (MJO) and convectively-coupled equatorial waves (CCEWs). The first component of the thesis is a linear stability analysis of a simple system of equations with relevance for the MJO and CCEWs. Through the application of formal asymptotic methods, it is shown that certain parameter regimes of the linear system are absolutely unstable, and will grow along a stationary characteristic. It is then shown analytically that in zonally-varying basic states, exponentially localized regions of instability may be formed around a global minimum of the gross moist stability. The resulting local instability has a zonal structure which shares some similarities with the manner in which the MJO develops as it transits the Indo-Pacific warm pool. The remaining chapters are centered around experiments with an idealized general circulation model. The model consists of a spectral dynamical core with a basic treatment of the hydrological cycle coupled to a slab ocean layer. When this model is augmented with a zonally asymmetric oceanic energy flux, so as to force zonal temperature gradients in the tropics, strong MJO events emerge as a result. The dynamical structure of these MJO events bears considerable resemblance to observations, and the events display sensitivities consistent with past modeling studies. Next, the CCEW variability in the zonally symmetric configuration of the model is considered. It is shown that the westward-propagating disturbances present in the model owe their existence to a coupling between tropical convection and the free, external Rossby modes of the atmosphere at planetary wavenumbers 5 and 6. The final chapter shows that the MJO events produced by the model are sensitive to the magnitude of surface friction in the model. This sensitivity reveals many new insights into the nature of the MJO; a connection between the MJO and the Rossby modes discussed in the previous chapter is explored in great detail.It is suggested that dynamical nonlinearity plays a key role in the genesis of MJO events in the model.

The hydroxyl radical (OH) is an atmospheric detergent, removing air pollutants and near-term climate forcers (NTCFs) such as ozone (O3) precursors as well as greenhouse gases like methane (CH4). Thus, understanding how it is changing and responding to its various drivers is important for air quality and climate. In the historical (1980-2014) CMIP6 simulation, we find that OH has increased in the Geophysical Fluid Dynamics Laboratory (GFDL) AM4.1 CCM mostly driven by increasing nitrogen oxide (NOx) emissions. In future CMIP6 scenarios, across multiple CCMS, we find that, in a high emission scenario (SSP3-7.0), while [OH] declines and CH4 lifetime increases due to the dominant chemical effects of changing NTCF and CH4, future climate change has a positive impact on OH and therefore a negative impact of CH4 lifetime, leading to a negative CH4-related climate feedback. Also, while air pollution controls alone further exacerbates the OH decline in SSP3-7.0, CH4 mitigation can effectively reverse the OH decline. These results suggest potential climate trade-offs from solely targeting air pollution and highlight the importance of CH4 mitigation to maximize future climate and air quality benefits. One potential OH driver is increasing H2 emissions from a future H2 economy. H2 and CH4 have similar composition and climate impacts, primarily because they share the same OH chemical sink. Hence, we find that CH4 mitigation can offset the impact of H2 leakage. In particular, the H2 warming impact is roughly a third that of CH4 per mass of emissions, and this stays relatively constant under a wide range of background CH4 emissions. Furthermore, when considering simultaneous changes in other NTCFs in a H2 economy, CH4 mitigation plays an important role in ensuring that there are net benefits to climate and air quality. We also find that, similar to CH4 emissions, H2 leakage increases surface O3, thus increasing O3-attributable deaths from cardiovascular and respiratory diseases, leading to increased economic cost. On a per-mass emission basis, H2 causes approximately half the health-related damages of CH4. These findings emphasize the key role of CH4 mitigation in offsetting unintended negative climate and health consequences of future H2 leakage.

The observed large-scale atmospheric circulation and precipitation pattern over North America consists of substantial zonal (east-west) asymmetries. Despite such a pattern's relevance to regional climate, a thorough explanation behind how such features arise remains somewhat lacking in existing literature. Here we further investigate the dynamical causes of zonal asymmetries in atmospheric circulation and precipitation over North America by performing and exploring a series of global climate model experiments with altered geography and land surface characteristics over North America and analyzing observations. In Chapter 2, we explore the role of the Gulf of California in the North American monsoon by comparing a realistic climate simulation with a water-filled Gulf of California to one where the Gulf of California is instead replaced with flat land. We find that the Gulf of California enhances North American monsoon precipitation, and the results provide numerous additional insights on the nature of the North American monsoon. In Chapter 3, we examine how North America's topography shapes the atmospheric circulation and precipitation pattern over North America and beyond during the cold season (December-March). Here, we perform and analyze a series of climate model simulations with altered topography over North America and show how the simulated impact of topography is consistent with observations. We find that North America's topography mechanically diverts the large-scale midlatitude westerly flow and subtropical low-level easterly flow meridionally, contributing substantially to the observed high-amplitude atmospheric wave pattern and zonal asymmetries in mean precipitation over North America. In Chapter 4, we return to the warm season and investigate how topography and land surface heating shape the summer atmospheric circulation and precipitation pattern over North America. We use the topography experiments performed in Chapter 3 and perform additional land albedo perturbation experiments. We find that both topography and land surface heating are essential. Whereas the northern Rockies primarily mechanically divert the large-scale midlatitude westerly flow so as to contribute to the observed ridge-trough pattern over northern North America, we explain how the circulation over the southern Rockies is largely driven by strong surface sensible heating over the region's sloping terrain. As such, in contrast to some recent work, we argue that well-known features such as the Great Plains low-level jet and North American monsoon can be largely characterized as thermally-driven.

The accurate prediction of future warming depends on understanding key processes like climate feedback, aerosol interactions, and radiative transfer, all of which are explored in this thesis. First, we explore the 6th Phase of the Coupled Model Intercomparison Project (CMIP6), where models exhibit higher estimates of effective climate sensitivity (ECS) due to more positive cloud feedback. Our analysis shows that models with more positive cloud feedback also feature stronger aerosol-cloud interactions, offsetting warming during the historical period. We find that the observed interhemispheric asymmetry in warming is better aligned with models that have weaker aerosol-cloud interactions and lower ECS, helping to reduce uncertainty in future warming projections. Chapter 3 focuses on the variability of the climate feedback parameter, driven by sea surface temperature (SST) patterns and atmospheric model physics. Through a series of targeted atmospheric global climate model (AGCM) experiments, this chapter quantifies the roles of warming pattern of SST and model physics in the intermodel spread of climate feedback and cloud feedback. The findings demonstrate that atmospheric model physics, particularly cloud-related processes, contribute significantly to the variability of climate feedback, emphasizing the importance of model accuracy in future climate predictions. Chapter 4 introduces a novel approach to address the computational challenges in climate modeling. A hybrid radiative transfer framework (HRadNN) that integrates machine learning with traditional physics-based models is developed, achieving a four- to five-fold increase in speed without sacrificing accuracy. The HRadNN model performs well across various climate scenarios, demonstrating its robustness and potential to improve computational efficiency in climate modeling. Together, these chapters provide insights into the complexities of climate sensitivity, feedback mechanisms, and modeling improvements, contributing to a more refined understanding of future climate change projections.

This thesis examines the complex dynamics between environmental conditions, seed genesis, and tropical cyclone (TC) genesis in the Eastern North Pacific (ENP) and North Atlantic (NA). Basin- and timescale-specific Genesis Potential Indices (GPIs) are constructed using Poisson regression, highlighting the importance of local upwelling and wind shear for seed genesis and TC development, respectively. To capture the influence of non-local factors, a novel trajectory-based GPI (traj-GPI) is introduced, merging adjacent non-local environments based on observed seed trajectory densities. The traj-GPI outperforms the original GPIs by simultaneously capturing seed activity, driven mainly by upward motion, and the transition to TCs, controlled primarily by vertical wind shear. The thesis further explores the differential responses of TC genesis to seed activity through a Hadley Circulation (HC) perspective. Reanalysis data reveals that the interannual variation in the Northern and Southern cells of the HC is more pronounced in the ENP and NA, respectively. These variations amplify seed frequency via enhanced upward motion within the ITCZ. In the ENP, however, an intensified Northern cell leads to increased shear over the TC genesis areas, hindering seed maturation and resulting in a trade-off balance. Finally, idealized aquaplanet simulations are conducted to investigate the impact of tropical SST warming patterns on the seed-TC relationship. We perform two sets of experiments with prescribed SST profiles replicating the equatorial and off-equatorial tropical warming associated with enhanced seed activity in the ENP and NA, respectively. The results not only show consistency with observed seed and TC activity but also align with HC-related upwelling and wind shear as evidenced by reanalysis data. This thesis thus offers an integrated framework that combines statistical modeling and dynamics, providing a explanation of how large-scale environmental conditions influence seed and TC genesis across different basins.

This thesis describes three novel mechanisms of air-sea interaction. The first two chapters focus on the amplification of the hydrological (water) cycle as the climate system warms in response to a CO2 increase, which is realized as an amplification of freshwater flux (precipitation - evaporation) patterns. We find that this freshwater flux pattern leads to a redistribution of oxygen and carbon in the ocean, which modifies the previously recognized changes due to the atmospheric CO2 increase, warming, and circulation changes. The change in oxygen concentrations results from the change in sea surface salinity patterns which modifies the ocean circulation and heat uptake. The change in carbon concentrations results from the dilution or concentration of carbonate species (in parallel to salinity), and the same changes in heat uptake. The redistributions of oxygen and carbonate species in response to hydrological cycle amplification are comparable to the effect of global warming, which decreases both oxygen and carbon concentrations throughout the ocean. The third chapter provides a causal mechanism, for the first time, to link the change in sea surface temperature that develops during El Niño events to the change in temperature of the tropical free troposphere. The temperature of the free troposphere is primarily determined by the temperature profile followed by moist convection over the ocean (rising air above high sea surface temperatures). We isolate the part of the sea surface associated with convection (which is determined on thermodynamic rather than geographic grounds) and demonstrate that the El Niño surface temperature increase is driven by a decrease in surface wind speed which damps the evaporation rate. This result suggests a possible relationship between the zonal symmetry of the tropical atmospheric circulation, temperature of the free troposphere, and top-of-atmosphere energy budget.

In this thesis, I study the impacts of convective heating, clouds and cloud-radiativeeffects on the tropical climate. Broadly speaking, these impacts are studied in frame- works where the distribution of heating or clouds is taken as given, and their effects on the resulting circulation are analyzed. In Chapter 2, the distribution of diabatic heating is prescribed in a linear equation for the Walker circulation streamfunction. Through tractable closed-form solutions, I conclude that the strength of the Indo- Pacific zonal overturning is not simply a function of the total convective heating, but that spatially more concentrated heating leads to more vigorous circulations. In Chapter 3, I use a cloud-locking technique in the ESM2G abrupt-4xCO 2 experiments to elucidate the role of the cloud-radiative effects (CREs). Hadley and Walker cir- culation weakening and Hadley expansion are analyzed, and the changing CREs are shown to have considerable impacts depending on the variable and timescale in ques- tion. In Chapter 4, we show that the interaction between longwave CREs, convection and circulation alter the spectrum of the tropical waves. The presence of this inter- action reduces gross-moist stability and leads to stronger moisture modes (like the Madden-Julian Oscillation) at the expense of the gravity-like waves (e.g., Kelvin). These effects are robust in both the AMIP and aquaplanet configurations of the AM4 model. Given the results in this thesis, I argue that analyzing systems where clouds and heating are uncoupled from circulation can yield useful insights of the coupled system.

Quasigeostrophic flows are induced by spatial variations in interior potential vorticity and boundary buoyancy. In the first part of this dissertation, we develop the geostrophic turbulence theory of boundary buoyancy anomalies in a quasigeostrophic fluid with vanishing potential vorticity. We find that the vertical stratification controls both the interaction range of boundary buoyancy anomalies and the dispersion of boundary-trapped Rossby waves. Buoyancy anomalies generate longer range velocity fields and more dispersive Rossby waves over decreasing stratification [$dN(z)/dz\le 0$, where $N(z)$ is the buoyancy frequency] than over increasing stratification [$dN(z)/dz\ge 0$].
Consequently, the surface kinetic energy spectrum is steeper over decreasing (mixed-layer like) stratification than in the classical uniformly stratified model. We therefore suggest that this steepening of the spectrum over mixed-layer like stratification accounts for the $k-2$ spectrum found in the wintertime upper ocean. This suggestion is consistent with numerical and observational evidence indicating that surface geostrophic velocities over wintertime extratropical currents are largely induced by surface buoyancy anomalies. We also find that, under certain conditions, the nonlinear interplay of boundary-trapped Rossby waves with the turbulence spontaneously reorganizes the flow into homogenized zones of surface buoyancy separated by surface buoyancy discontinuities, with sharp eastward jets centered at the discontinuities and weaker westward flows in between. Jet dynamics then depend on the vertical stratification. Over decreasing stratification, we obtain straight jets perturbed by dispersive eastward propagating waves. Over increasing stratification, we obtain meandering jets whose shape evolves in time due to westward propagating weakly dispersive waves. In the second part of this dissertation, we investigate normal modes in the presence of boundary-confined restoring forces, with the ultimate aim of obtaining an energy-conserving modal truncation of the quasigeostrophic equations. Such a modal truncation would generalize classical $N$-layer models to account for non-isentropic boundaries. Although we obtain orthogonal sets of vertical modes that diagonalize the energy and potential enstrophy in the presence of non-isentropic boundaries, we find that the loss of a crucial symmetry in the vertical coupling between the modes prevents modal truncations from conserving energy. Consequently, energy conserving modal truncations are not possible in the presence of non-isentropic boundaries.

Earth’s radiation budget is a fundamental indicator of the state of the climate. Satellite observations over 2000-2020 display significant monthly-to-decadal changes in the clear-sky greenhouse effect and outgoing longwave (LW) radiation (OLR). Radiative transfer experiments show that these OLR changes are driven by surface and atmospheric temperature increases (decreases) that are only partially offset by the accompanying water vapor increases (decreases). Water vapor in the middle troposphere, in addition to the upper troposphere, substantially contributes to the OLR changes. In addition to OLR changes, the observed net radiative imbalance displays a significant positive trend. Using climate model simulations, we show that it is exceptionally unlikely (< 1% probability) that this trend can be explained by internal variability alone. The trend is driven by a large decrease in reflected shortwave (SW) radiation and a small OLR increase. Model experiments show that this is because recent changes in effective radiative forcing and feedbacks are additive in the SW, while being nearly offset by each other in the LW. The SW forcing, dominated by cloud changes induced by rapid adjustment to greenhouse gas increases and aerosol decreases, helps explain the positive trend (heating) in the SW cloud radiative effect (CRE). LW greenhouse gas forcing decreases OLR more in clear skies than all-sky conditions because clouds mask the impact of greenhouse gas increases. This cloud masking helps explain the negative trend (cooling) in LWCRE. In contrast to the well-identifiable forcing contributions to CRE, cloud feedback contributions have a wide range. In fact, their signs are uncertain. To further understand the LW changes over this period, we investigate the changes across the LW spectrum. The observed spectrum displays significant OLR decreases in the CO2, O3, CH4, and N2O absorption bands as well as OLR increases in the H2O rotational, vibrational, and window bands. Line-by-line radiative transfer calculations and model simulations show that increases in well-mixed greenhouse gases caused an instantaneous radiative forcing and stratospheric cooling that decreased OLR, while the climate response, comprising temperature and humidity changes, increased OLR. We conclude that the satellite radiation record provides clear evidence of a human-influenced climate system.

Despite recent advances, the extratropical cloud feedback is at the heart of uncertainty and intermodel spread in climate sensitivity among comprehensive General Circulation Models (GCMs). Here, to shed light on the underlying mechanisms, we use an idealized GCM---the Held-Suarez dynamical core with the addition of passive water vapor and cloud tracers and full cloud microphysics---to cleanly decompose the effects of individual cloud processes. A set of process-level experiments are designed to elucidate the mechanisms of an increase in extratropical liquid water path (LWP) in mixed-phase clouds with warming: higher adiabatic cloud water content, weaker liquid-to-ice conversion through the Bergeron-Findeisen process, and faster melting of ice and snow to rain. Perturbed parameter experiments demonstrate a strong dependence of the LWP feedback on the climatological LWP and independence from the climatological susceptible ice and supercooled liquid fraction. To examine the broader physical controls of extratropical humidity and clouds, separate saturation adjustment and full cloud scheme controls are used to consider the strength of advection-condensation theory. It is found that sub-grid-scale relative humidity distribution assumed within the cloud macrophysics limits isentropic transport of tropical moisture to the polar troposphere. Also, within the model's explicit treatment of cloud microphysics, re-evaporation of hydrometeors moistens and increases clouds in the lower troposphere. To investigate the direct influence of dynamical changes on humidity and clouds, we explore the impacts of altered meridional temperature gradients. We show that predicted extratropical jet shifts with warming have little effect on shortwave cloud radiative effects because they mainly affect cloud fraction, not cloud thickness which is principally connected to absolute temperature change. Overall, the results substantiate the utility of such idealized models for elucidating cloud processes in a systematic manner and highlight key cloud processes to constrain. Additionally, a set of realistic experiments are designed in a comprehensive GCM to study geoengineering using stratospheric aerosols and compare efficacy and impacts of reflecting and absorbing aerosols (sulfate) versus only reflecting aerosols (calcite). We evaluate which regions are better or worse off with geoengineering in a 2xCO2 world in terms of various climate metrics and consider the policy implications.

Monsoons supply the majority of annual rainfall to land regions across the tropics and play a prominent role in global climate. Despite this, we lack a complete mechanistic understanding of monsoon characteristics and how they might change with warming. This dissertation draws on theory and a range of climate models to elucidate the physics governing monsoon circulations. First, we investigate the seasonally dependent responses of the South American monsoon to warming. In the latest Geophysical Fluid Dynamics Laboratory Atmospheric Model (AM4), uniform 2-K sea surface temperature warming reduces precipitation by 11%, and precipitation minus evaporation by 40%, in the monsoon onset season, producing a more severe dry season. A column-integrated moist static energy budget analysis helps elucidate the drying mechanism. By comparing the response across seasons we identify two necessary conditions for a region to exhibit a substantial drying response to warming. To better understand the seasonal and global variations in monsoon circulations, Chapter 3 utilizes an idealized modeling framework. We systematically vary the moisture and albedo parameters over an isolated South American continent with reduced heat capacity, but no topography. Irrespective of the local moisture availability, the seasonal cycles of precipitation and circulation over the monsoon sector are distinctly monsoonal with the default surface albedo, but with differing characteristics and dynamical mechanisms. The analysis supports that when the land-ocean thermal contrast is strong enough, inertial instability is sufficient for producing a shallow but vigorous circulation and converging a large amount of moisture from the ocean. Once the land is sufficiently moist, convective instability takes hold and the shallow circulation deepens. Chapter 4 investigates the monsoonal precipitation over a dry continent on synoptic timescales, and provides additional evidence for the importance of inertial instability. In Chapter 5, we utilize an Earth System Model (ESM4.1) to investigate changes in precipitation and aridity over the South American monsoon sector in future emissions scenarios. The projected runoff reductions have implications for energy system planning in Brazil, where 60-80% of monthly power generation comes from hydropower. The possibility of seasonal power generation deficits underscores the importance of climate risk management in the energy sector.

This dissertation synthesizes long-standing theoretical ideas to develop quantitative constraints on several aspects of tropical climate, namely convection, heat stress, precipitation, and outgoing longwave radiation. A zeroth-order picture of the tropical troposphere is that deep convection maintains a moist adiabatic temperature profile in the vertical and gravity waves quickly smooth any temperature gradients in the horizontal, which is formally known as convective quasi-equilibrium (QE) and the weak-temperature-gradient (WTG) assumption. We expect that strict QE-WTG should yield a uniform moist static energy (MSE) threshold for deep convection. Consistent with the theoretical expectation, we find that deep convection only occurs over the highest subcloud MSEs and that the convective subcloud MSE (daily-mean subcloud MSE weighted by precipitation) is roughly uniform between 20°S and 20°N. QE-WTG forces the highest subcloud MSEs to be equal over land and ocean, not only in the present climate but also in much colder and warmer climates. The annual-maximum wet-bulb temperature, a metric for extreme heat stress, is also controlled by the QE-WTG dynamics due to the functional relationship of wet-bulb temperature with MSE. We provide a theoretical projection that the annual-maximum wet-bulb temperature will increase roughly uniformly by about 1°C for each 1°C of tropical mean warming. This result suggests that limiting the mean surface warming also limits heat stress extremes in the tropics. QE-WTG controls the occurrence of deep convection and thus precipitation. Global climate models consistently predict that tropical precipitation will be distributed more unevenly in space with global warming. We show that the unevenness of precipitation can be traced back to the unevenness of subcloud MSE distribution. We then explain the change in sublcoud MSE distribution with a simple scaling accounting for the Clausius-Clapeyron increases of boundary-layer specific humidity under invariant relative humidities. The invariance of relative humidities also has implications for the linearity of outgoing longwave radiation with surface temperature.

Tropical stratospheric dynamics are dominated by the Quasi-Biennial Oscillation (QBO), an alternating, descending pattern of zonal winds with a period averaging 28 months. The QBO is driven by dissipation of vertically propagating atmospheric waves. Although its internal dynamics are driven by waves, the QBO vanishes several kilometers above the wave source, forming a buffer zone. Based on foundational 1D model experiments, it was previously accepted that the buffer zone formed where upwelling opposed the descent of the QBO. This dissertation overturns the upwelling hypothesis by demonstrating that those foundational experiments implicitly prescribed the buffer zone through the model boundary conditions. New 1D model experiments show that the only general way to form the buffer zone is to damp QBO angular momentum towards zero. Angular momentum damping of sufficient magnitude and vertical structure to form the buffer zone is diagnosed in atmospheric reanalyses, and is found to result from horizontal momentum flux divergence, which exports QBO angular momentum anomalies to the far field atmosphere. Exported QBO angular momentum anomalies are known to reach the solid Earth, leading to a correlation between the QBO and O(0.1 millisecond) variations in the length of day.

The upwelling hypothesis has previously been used to attribute decreasing trends in QBO amplitude in the lower stratosphere to global warming-induced increases in residual upwelling. These trends depended on record-low QBO amplitudes in the mid-2000s, from which the QBO has since recovered. An updated analysis reveals no significant trends in QBO amplitude in the lower stratosphere. Rather, decadal variability in QBO amplitude is reinterpreted as large internal variability of QBO amplitude in the buffer zone.

In 2016, the QBO exhibited an unprecedented disruption, during which easterly winds formed unexpectedly around 40 hPa, and wind anomalies stagnated and ascended above 40 hPa. Previously, each feature of the disruption had been explained by a different dynamical mechanism, without accounting for how these proposed mechanisms might be linked through the internal dynamics of the QBO. The 1D model of the QBO is used to demonstrate that all features of the disruption follow from the internal dynamics of the QBO in response to a localized incursion of momentum from the far field.

Dense gravity currents forced by surface buoyancy loss over continental shelf regions are important contributors to subsurface and abyssal ventilation throughout the World Ocean, yet remain challenging to model accurately. In this thesis, we present idealized experiments of rotating terrain-following gravity currents employing the nonhydrostatic MITgcm in z coordinates and the hydrostatic GFDL-MOM6 in z* and isopycnal coordinates. In the highest-resolution simulations, the dense flow undergoes geostrophic adjustment and forms bottom- and surface-intensified jets. The density front along the topography combined with geostrophic shear initiates submesoscale symmetric instability (SI), which leads to onset of secondary shear instability, dissipation of geostrophic energy, and irreversible mixing. We explore the impact of vertical coordinate, resolution, and parameterization of shear-driven mixing on water mass transformation. In isopycnal coordinates, limited vertical resolution in weakly stratified abyssal regions leads to inadequate representation of mixing. We develop and implement a parameterization for SI-driven turbulence to mediate this issue. The scheme is based on identifying unstable regions through a balanced Richardson number criterion and slumping the isopycnals towards a balanced state. A fraction of the released potential energy is passed to the shear mixing parameterization, so that potential energy extracted from the geostrophic flow by SI is converted to kinetic energy and used for vertical mixing. Such a scheme becomes crucial as ocean models move towards resolving mesoscale eddies and fronts but not the submesoscale phenomena they host. In the final thesis component, we examine how state-of-the-art global ocean models currently represent water transformation processes in the Arctic. We consider a 1/4 and analogous 1/8-degree model and find that ventilation by overflows as well as transformation of the warm, salty Atlantic inflow both contribute to creating highly dense waters in the Eurasian shelves. The 1/8-degree model performs better in capturing transient dense flows emanating from polynyas around coastal islands such as Novaya Zemlya; the 1/4-degree model marginally resolves overflows but has an overly diffuse vertical structure. As a next step in bridging our idealized process studies with the global simulations, we plan to perform regional mesoscale-resolving modeling with our SI parameterization to further constrain Arctic ventilation pathways.

Aspects of tropical and extratropical cyclone dynamics and their relationships with the climate are investigated. Theoretical advances are made in the development of analytical approximations verified in model experiments designed with hierarchical levels of process representation. Starting with the quasi-geostrophic theory, a new regime of solutions is investigated, which challenges the classical theories of wave-mean flow interaction. The results show the emergence of surface easterlies in the baroclinic zone, as opposed to the Earth-like surface westerlies, in the parameter space where the quasi-linear approximation of the large-scale midlatitude circulation breaks down.

With an increased level of complexity, a theory is developed to connect the convection-scale moist dynamics with the synoptic-scale circulation of a mature baroclinic cyclone. Verified in a quasi-steady simulation of a moist baroclinic cyclone, the theory establishes limits within which a rescaling of nonhydrostatic dynamics improves the representation of convection in global climate models.

A diagnostic theory is developed to explain the frequency of tropical cyclones simulated in global climate models. The complex relationship between the cyclone frequency and the large-scale circulation is simplified by decomposing the cyclone development process. A hierarchy of model experiments are conducted to isolate specific processes and to evaluate the corresponding components in the theory. The theory is applied to realistic simulations of historical and future climates, as well as observational tropical cyclone records. Physical interpretations are established regarding the diversity of tropical cyclone simulations across models and the projected frequency in response to various global warming perturbations.

The equilibration of the extratropical troposphere is controlled in large part by the transient eddy heat fluxes. This thesis analyzes some idealized models that capture the essential sources of complexity which have hindered our understanding of these fluxes. In particular, it attempts to eliminate some knowledge gaps between proposed turbulence diffusivity theories for these fluxes.

The thesis begins by comparing two of these theories, one motivated by homogeneous quasigeostrophic (QG) simulations and one by parameter variations in a comprehensive atmospheric model. The two theories share the same turbulent inverse energy cascade argument but differ in the assumptions concerning the mixing slope used in estimating the entropy budget. The importance of this mixing slope is verified using an idealized dry general circulation model (GCM).

We find that the inverse cascade part of these theories fails to explain some of the GCM simulations. Hence, we return to the homogeneous two-layer QG model. In the limit of vanishing environmental vorticity gradient 𝛽=0, we find that the diffusivity does not depend on the friction in the same way as the passive tracer diffusivity in barotropic turbulence. Unlike what is typically assumed, there is no classical inverse cascade inertial range in the two-layer model.

We next show that the 𝛽 dependence of heat flux diffusivity can be combined with the dependence on friction in a single nondimensional parameter. Based on this result, we discuss how the heat flux diffusivity is suppressed by 𝛽 due to wave propagation in addition to its effect on the extent of the inverse cascade.

Finally, we return to the idealized GCM to evaluate the maximum entropy production (MEP) argument, which suggests that the eddy heat fluxes adjust to maintain the MEP state. We solve for the MEP climate and show that it is unlikely to be physically relevant even if one constrains the global mean vertical temperature profile. We argue against the use of the MEP argument and reaffirm the need to thoroughly understand the eddy dynamics.

Through latent heat released by condensation and feedbacks involving radiation, physical processes related to water vapor play key roles in tropical dynamics. While some understanding of aspects of tropical climate can be obtained using dry general circulation models (GCMs), e.g. theories for what controls the width of the Hadley circulation, understanding of other aspects requires, or at the very least can be refined by, the inclusion of the influence of moist processes. A natural intermediate step in the model hierarchy between a dry GCM and a comprehensive GCM is the idealized moist model, which includes simplified treatments of moist processes. In this dissertation, we make use of this model, in this case coupled to a full radiative transfer code to additionally include water-vapor-radiation feedbacks, to study aspects of tropical mean state and transient precipitation in simplified setups.

We begin by quantifying the impact of including the interaction between water vapor and radiation on the sensitivity of the latitude of the intertropical convergence zone (ITCZ) to a cooling in the Northern Hemisphere (NH). We do so by applying varying magnitude cooling forcings in the NH in two model configurations: one with interactive water vapor and radiation, and one with the water vapor field seen by the radiation code prescribed as that from a hemispherically-symmetric control simulation. We find that the ITCZ shifts roughly twice as far for a given forcing in the interactive cases versus the prescribed cases.

We then move on to studying transient disturbances, starting with monsoon low pressure systems (MLPSs). To study these we add a simplified treatment of land to the model. With realistic continental geometry we show using regression and tracer budget analysis that the model supports MLPS-like storm systems in the South Asian monsoon region, whose dynamical properties share some commonalities with monsoon depressions found in reanalysis datasets. Finally we turn to equatorial waves. Through aquaplanet experiments in which heating perturbations with varying magnitude and vertical structures are added, we show that MJO-like variability can be preferentially enhanced if the net heating of the deep tropical atmosphere is increased.

The provision of accurate weather and climate predictions at timescales of days to decades has been a major research goal in atmospheric and oceanic sciences. This dissertation explores the issue of 'seasonal predictability', the potential of the climate system being predicted a season in advance. The studies in this dissertation examine the seasonal predictability of two aspects of the climate system: (1) rainfall, including its extremes; and (2) tropical cyclones (TCs), particularly those that undergo rapid intensification (RI).

The first study examines the response of rainfall in East Asia to the El Nino-Southern Oscillation (ENSO) phenomenon, and demonstrates an asymmetric response of rainfall to ENSO along the southeastern coast of China during boreal fall/winter. Anomalous rainfall is observed during both El Nino and La Nina compared to the ENSO-Neutral phase. We argue that precipitation anomalies during El Nino arise from anomalous onshore moisture fluxes, while those during La Nina are driven by the persistence of terrestrial moisture anomalies from earlier excess rainfall in this region, highlighting the role of land-atmosphere interactions in maintaining ENSO-climate teleconnections.

The second study explores the observational connections between the large-scale environment and the seasonal statistics of rapidly intensifying North Atlantic TCs. For TCs in the Central/Eastern tropical North Atlantic, the interannual variability of their probability to experience RI is influenced by the seasonal large-scale environment, but not for TCs over the Gulf of Mexico and Western Caribbean Seas. We suggest that this differentiated response is due to the former region exhibiting negatively correlated seasonal anomalies of vertical wind shear and potential intensity. This motivates a subsequent chapter, which examines the physical mechanisms behind the negative correlation, and applies the findings to global TC basins.

The final chapter extends the environmental controls on RI to numerical models, and explores: (1) the simulation of the seasonal large-scale environment in climate models, as an indirect means of RI seasonal predictability; (2) the role of large-scale environmental biases in TC intensity biases in weather forecast models. Assessment of RI predictability through weather and climate models will contribute to the long-term research effort in TC modeling and prediction.