Geosciences, 1929-2026

Single-crystal elastic constants and pressure derivatives are compiled for B1 rocksalt oxides, halides, and sulfides alongside B2, B3, garnet, and spinel analogues to characterize the elastic systematics of the (Mg,Fe)O solid solution and empirically test the Anderson and Liebermann (1970) prediction that dCij/dP is controlled by crystal structure rather than composition. At ambient conditions, C11 and C44 decrease monotonically with increasing Fe content across the (Mg,Fe)O solid solution while C12 increases, driving AZ toward the FeO endmember; compression is demonstrated to be a far more efficient mechanism for reducing anisotropy than Fe substitution. All B1 rocksalt compositions, regardless of mineral class, satisfy dC11/dP > 2.5 × dC12/dP and dCs/dP > 2 × dC44/dP, forming a ratio-bounded region in dCij/dP space that no B2, B3, garnet, or spinel composition satisfies, constituting the first multi-class empirical confirmation of Anderson and Liebermann (1970). (Mg0.92Fe0.08)O undergoes an isotropic crossover (AZ = 1) at ρ ≈ 4.2 g/cm3 and reaches AVSmax ≈ 37% and AVP ≈ 16% at the highest sampled density, far exceeding Pyrope and Ringwoodite, neither of which undergoes an analogous inversion. Within the alkaline-earth oxide series, B1–B2 transition pressure increases systematically with dC44/dP, while FeO departs from this trend, consistent with d-electron bonding modifying its lattice dynamics. These results confirm that dCij/dP ratio bounds provide a structure-diagnostic tool applicable to high-pressure phases of uncertain crystallography and carry direct implications for the sign and lateral variation of seismic anisotropy in the lower mantle.

Understanding how rhizosphere microbial communities influence plant phenotype is central to developing strategies for sustainable agriculture, but systematic study is difficult because microbes function within complex, interacting communities. Machine learning methods offer a way to navigate the large combinatorial space of possible microbial communities and support future efforts to model how community composition relates to plant phenotype. This thesis addresses technical hurdles required for future active-learning experiments pairing five-member synthetic microbial communities with \textit{Arabidopsis thaliana} plant phenotypes. First, we adapted an existing plant image-segmentation model and root-analysis software to efficiently extract phenotypic descriptors and root graph objects from plant image data, making scalable plant phenotyping more feasible for large-scale experiments. Second, I developed and evaluated descriptor-based and machine-learned representations of plant phenotype using 1,645 time-lapse images from 291 plants across 85 agar plates. Biological, topological, and shape-based descriptors were screened for variation and redundancy to create an interpretable validation set. I then trained a GINE graph neural network with contrastive learning to generate unsupervised plant-graph embeddings. UMAP visualizations and K-nearest neighbor agreement analyses showed that the learned embeddings organized plants along established biological descriptors, graph-topological features, and additional shape-based traits. Biological and topological descriptors showed strong neighborhood agreement, while shape-based descriptors were preserved more moderately but still meaningfully. Together, these results demonstrate that unsupervised graph-based plant representations can recover standard biological measures of phenotype while also capturing additional morphological structure. This work lays the foundation for active-learning experiments that efficiently explore how microbial community composition shapes plant phenotype.

Stable water isotope measurements from polar ice cores are used as a paleothermometer, providing high-resolution information about past precipitation as a function of climate. Traditionally, ice cores’ paleoclimate records have been obtained from high accumulation ice divides, where the assessment of climate signals and effects of glacial deformation is easily understood. Sites with complex stratigraphy offer snap- shots of ice that are older, albeit the detangling of the climate signal is further complicated. The Allan Hills blue ice area has been a site of interest for discovering the oldest ice on Earth, with the most recent work by COLDEX discovering 6-million-year-old ice. However, the meteorological and glacial origins of Allan Hills ice are not well understood, hindering our ability to truly parse the climate record from the water isotope record. Here, I build on earlier work by Davidge (2025) to create a new database of water isotope measurements generated by COLDEX, establish a local meteoric water line for the Allan Hills area, and suggest at least two separate origins of ice flow into the region. This thesis establishes a framework for interpreting the Allan Hills water isotope record from a theoretical perspective.

The long-standing fossil record of sharks has driven interest in reconstructing the diets and ecology of both modern and ancient species. Stable carbon isotope analyses of inorganic carbonate (δ¹³CCO₃) and tooth dentin collagen (δ13Ccollagen) are commonly used to investigate dietary information and environmental conditions. However, preservation of collagen does not exceed tens of thousands of years and δ¹³CCO₃ may be confounded by environmental signals. Enameloid is resistant to diagenetic processes and protects enameloid-bound organic matter (EBOM) over million-year timescales. This study ground-truths the carbon isotope signature of EBOM (δ13CEBOM) in shark teeth as a novel proxy for reconstructing diet. These measurements are made possible with a modified elemental analyzer–isotope ratio mass spectrometry (EA-IRMS) system (“nanoEA”) that can analyze small quantities of enameloid (2-4 mg containing ~100 nmol carbon). In this thesis, I demonstrate a strong correlation between δ13CEBOM and δ13Ccollagen in modern shark teeth, supporting its potential as a dietary proxy. I additionally measure δ13C across various carbon pools within the tooth (enameloid carbonate, dentin collagen, dentin-bound organic matter, and bulk dentin) to better understand the organic matter protected in the mineral structure and the observed offset of ~7‰ between δ13CEBOM and δ13Ccollagen. Even with a δ13C offset, the strong correlation between δ13CEBOM and δ13Ccollagen highlights the promise of δ13CEBOM for reconstructing shark diets and trophic dynamics in Earth history, which can provide new insight into shark ecology and their resilience throughout past climate and marine environmental changes.

A growing body of evidence supports the hypothesis that aqueous sedimentary processes were active on the Martian surface in the distant past. Constraints on the scale and persistence of these water bodies, however, remain limited by the lack of subsurface data beyond the weathered layer. Appropriating seismic migration algorithms to invert Ground-Penetrating Radar (GPR) data from martian rovers has proven a reliable and practical approach to imaging the structure of subsurface sedimentary deposits beyond the depth reasonable for, or attainable by, sample excavation and return. In study, we begin by implementing, from first principles, three of the most commonly employed migration algorithms, and rigorously assessing their reliability under increasingly non-ideal subsurface conditions. By benchmarking algorithm behavior against a range of conditions where ground truth is fully controlled by a known GPR model and synthetic data, we establish a basis for constraining the confidence in and qualitative interpretation of results obtained from real planetary data, where these conditions are largely unknown. We then apply these methods to data collected by RoPeR, the low-frequency GPR module of the Zhurong Rover along a 1.8 km traverse in Utopia Planitia—the site of a putative ancient coastline. Our interpretation of the migrated section reveals North–West-Northwest prograding clinoforms comprised of sigmoidal reflectors on a scale of tens of meters and with apparent dips of 5–20 degrees. The range of dips encompasses that of previous findings, but significantly more detail as to the distribution of these angles, and the spatial variation of these distributions along the traverse, is afforded with our multi-method approach. The lack of a common strike association between imaged reflectors suggests that coastal sedimentary processes were actively deforming the beach planform as a large body of water receded into Utopia basin. Moreover, diminished wave and current action in a large Martian ocean leaves these processes best explained by a glacio-marine setting, consistent with the cold, wet Hesperian hypothesis supported by other sites along and within the highland-lowland boundary.

Fraser fir (Abies fraseri) is the most popular species of live Christmas tree sold in the United States. It is endemic to southern Appalachia, with much of its commercial production concentrated in western North Carolina. As an agricultural commodity, the Fraser fir industry is susceptible to long-term climate trends. This thesis explores the viability of western North Carolina for growing commercial Fraser firs over the next century using climate reanalysis data and high-resolution atmospheric model projections. A suitability index was developed: the combination of vapor pressure deficit, growing degree days (Baskerville-Emin method), cold hardiness, and pH was found to be a proxy for Fraser fir potential. These variables were projected forward, and suitability recalculated with (1) no bias correction, (2) a climatological mean adjustment, or (3) a climatological mean with variance adjustment. Results agree on declined suitability across western North Carolina by 2080-2100; this finding is consistent with the hypothesis that rising temperatures will force Fraser fir production northward. Indeed, application of the suitability index to the continental USA showed a northward migration of optimal growing regions. No notable increase in precipitation during 2080-2100 relative to the historical period was identified, but current biotic threats could still be exacerbated by increased incidence of extreme rainfall events. These findings underscore the sensitivity of Fraser fir production to shifts in climate, and provide a basis for anticipating future changes in the geographic distribution of suitable regions for cultivation.

"Human land use, primarily through farming and its resulting land disturbances impacts the soil sink of atmospheric hydrogen maintained by soil microbial communities; this thesis’ aim is to assess, both through measurements of soil H2 uptake rate and bacterial genomics, what governs this rate. Genetic assessment shows that soils grown under different crops at the same site retain similarity in community diversity and hydrogen uptake, but undisturbed soils even as little as 20 meters away show much greater uptake of hydrogen, which appears to be attributable to soil organic carbon (SOC) loss through disturbance. Some preliminary data points to soil nitrogen, in particualr ammonium modulated through N fixers like soybeans and other legumes, as an equally significant contributor to hydrogen uptake as SOC, but the sample size of this was small. Questions of human impacts on atmospheric cycling remain to be answered in this case of trace gas sources and sinks, which this thesis only scratches the surface of."

Potassium (K), magnesium (Mg), and calcium (Ca) are vital macronutrients in plant biochemistry, regulating plant growth, nutrient and metabolite transport, responses to environmental stresses, etc. K and N fertilizers are frequently applied to crops while Mg and Ca are usually applied when deficiency is expected. This thesis investigates how varying K concentrations affect nutrient uptake in plants, particularly in Arabidopsis thaliana, using a controlled hydroponic experiment setup. Plants regulate potassium uptake through specialized transport systems. When potassium is limited (<0.1 mM), plants tend to depend on high-affinity transport systems (HATS) to uptake potassium, whereas low-affinity transport systems (LATS) dominate when potassium levels are plentiful (>0.1 mM). Previous experiments revealed that at 0.10 mM external potassium, plants exhibited δ⁴¹K isotopic signatures in line with HATS, employed under stressed conditions without showing significant reductions in plant K acquisition relative to plants grown in more replete potassium environments. To address the gaps from the previous experiments, we examined intermediate potassium concentrations to evaluate whether a gradual transition or a step change would occur between HATS and LATS. Our results demonstrate that δ⁴¹K values become more negative with rising potassium availability. Plants grown at 0.33 mM potassium displayed intermediate δ⁴¹K, suggesting the concurrent operation of both transport systems. In contrast, δ⁴⁴/⁴⁰Ca and δ²⁶Mg remained stable across varying potassium conditions, indicating that their uptake is not dependent on external potassium availability. Additionally, based on the observed relationship between δ⁴¹K and δ15N, potassium availability may have an influence on nitrogen source utilization in plants (NO3 or NH+4). These findings demonstrate how nutrient availability affects uptake mechanisms and isotopic fractionation, providing new insights to achieve more sustainable agriculture practices.

This study examines the spatial correlation between low-level rotation fields from the Multi-Radar Multi-Sensor (MRMS) system and tornado reports from the Storm Prediction Center (SPC) between 2021 and 2023. Tornadoes were analyzed by overlaying reported paths onto MRMS RotationTrack60min fields, which capture maximum low-level azimuthal shear over a rolling one-hour window. Two spatial search methods were used: a radial search centered on each tornado’s reported start point and an interpolated path-based search along the full track. Results show that MRMS-detected rotation signatures were found within 5 kilometers of approximately 95% of tornado reports using the radial method, with marginal improvement from path interpolation. Regional differences were evident; tornadoes in Tornado Alley exhibited stronger rotation and higher detection rates than those in Dixie Alley. Tornadoes associated with stronger radar-indicated rotation were also more likely to align with reports, revealing a detection bias toward more intense systems. These findings underscore the strengths and limitations of MRMS azimuthal shear for tornado detection. While MRMS enhances identification of well-organized tornadoes, weaker or short-lived events remain harder to detect. This research contributes to improving tornado climatology and real-time warning systems by cross-validating radar data with observational records. Future work expanding the MRMS archive and incorporating environmental classifications will further refine understanding of tornado detection efficiency.

Stable isotopes are useful tools to describe how elements move through plants, as transport mechanisms often favor certain isotopes. In the past, stable isotope research on plants has focused primarily on agricultural plants and the transport of elements with high concentrations. Only recently have improvements in mass spectrometry made natural-abundance measurements of heavier metal isotope systems, including magnesium (Mg) and calcium (Ca), possible. Work to characterize these systems in plants is ongoing, and no study measures both Ca and Mg in the same plants. The new potential to study Ca and Mg in plants motivates this thesis, as does the application of stable isotopes to paleodietary study. Ca and Mg isotopes within dinosaur tooth enamel have been used to hypothesize trophic level differences, spatial niche partitioning, feeding height stratification, and preferential feeding on plant groups and plant parts. Without a more complete understanding of isotopic fractionation in plants, however, it is difficult to support these conclusions.

In this thesis, I aim to advance understanding of Mg and Ca isotope systems through a survey of plant materials gathered from two field sites in the California redwoods. I observe Ca and Mg isotope signals (1) at ecosystem scale (from water to soil to plants and across plant species), (2) within plants, across different plant parts, and (3) across leaves up the height of a single 107-meter redwood tree. When considering Ca and Mg isotopes separately, differences in isotopic ratios occur across multiple scales (location, species, plant parts, leaf height), complicating paleodietary interpretations. A model for Mg and Ca transport is proposed as a framework to quantify some of these differences. When considering Ca and Mg isotope measurements together, species form distinct clusters, which may be explained by the proportion of different pools of nutrients, mediated by the presence of potassium (K). Thus, the dual measurement of the two isotope systems in plants and tooth enamel emerges as a promising tool to inform paleodietary reconstructions, for dinosaurs as well as other animals, and to improve understanding of Ca and Mg biogeochemical cycling. Future work should clarify the role of K in fractionation and further classify plant material in Mg-Ca isospace.

Oxygen minimum zones (OMZs) are regions of the world’s ocean that are permanently depleted in dissolved oxygen. Nitrite oxidation, traditionally considered an obligately aerobic process, has been observed in the anoxic core of these OMZs challenging longstanding assumptions about nitrogen cycling in low-oxygen environments. This paper investigates nitrite oxidation within the Eastern Tropical South Pacific (ETSP) oxygen minimum zone through ¹⁵NO₂⁻ tracer incubations on samples collected at three stations spanning from the oxygen minimum zone’s anoxic core to oxygenated waters outside of the OMZ. Experiments tested nitrite oxidation rates under anoxic, low oxygen (1 μM O2) and oxygen-spike treatments designed to mimic transient lateral oxygen intrusions. Results revealed unexpectedly high nitrite oxidation rates under anoxic treatment in samples from the anoxic core of the OMZ (92.52 ± 19.32 nM/day at station PS3 depth 230 m), with little to no activity under oxic or spike conditions. These findings contradict the hypothesis that episodic oxygen availability sustains nitrite oxidation in OMZs, suggesting instead that some nitrite-oxidizing bacteria (NOB) possess a previously unrecognized anaerobic metabolism. This discovery supports growing evidence of niche-adapted NOB within OMZs, distinct from their aerobic counterparts. Potential mechanisms include nitrite dismutation or alternate oxidants, though further work is needed to test these theories. As OMZs expand under climate change, understanding these alternative pathways is critical to predicting nitrogen cycle dynamics and fixed nitrogen loss in the ocean.

The coastal provinces of Ecuador are some of the country’s most climate-vulnerable and socioeconomically vulnerable places in the country. These areas also brace most climate impacts related to El Niño–Southern Oscillation (ENSO) through precipitation and flooding extremes. There is a significant literature gap on local and coast-specific ENSO impacts in all of Latin America, but Ecuador in particular is under-studied. This research characterizes the relation- ship between ENSO indices in the Central Pacific and Eastern Pacific (Niño 3.4 and Niño 1+2) and ENSO meteorological and social impacts on the coast of Latin America (Niño 1+2), to understand further the local meteorological variables that affect the livelihood of communities living in this under-studied and vulnerable area. Using locally collected meteorological and precipitation-linked weather event casualty data to visualize the formation of extreme El Niño events, we assess ENSO climate impacts at multiple scales. At the source-to-hazard level, Niño 1+2 showed stronger and more consistent correlations with coastal precipitation anomalies than Niño 3.4, particularly during the peak rainy season. At the hazard-to-exposure level, precipitation anomalies aligned with increases in people and homes affected, with stronger coupling in southern coastal regions. At the community level, household surveys revealed that limited hazard knowledge, weak institutional trust, and economic dependence on climate-sensitive sectors collectively amplify vulnerability to ENSO events. These findings shed light on potential local policy measures that would help risk mitigation and preparedness for an El Niño event for small coastal communities.

Climate change impacts hydrological systems differently in different parts of the world. In this study I assess the ability of various global climate models (GCMs) to accurately reproduce observed historical trends in the frequency of flooding at the global scale. I use a peak-over-threshold (POT) method and focus on the period 1985-2014. First, I compare trends in regions belonging to different climate zone types and find that tropical and arid areas adhere to the “wet-get-wetter, dry-get-drier” pattern, with significant increasing trends in tropical areas and significant decreasing trends in arid areas; the results are less definite in temperate, boreal, and polar regions. Second, I assess whether the outputs from 12 different GCMs are consistent with the results from the reference data, considering both annual and seasonalized time scales. I find that, in general, the GCMs are not able to accurately model the observed regional trends. Northern Australia (NAU), however, is an exception to this, and I use a subset of ten GCMs that perform well in this region to determine the projected changes in the frequency of flood events by the end of the 21st century. My results indicate that flooding is projected to decrease in frequency under all four emissions scenarios considered.

This thesis seeks to identify a relationship between the phase of the Madden Julian Oscillation (MJO) and the probability of rapid intensification (RI) of tropical cyclones (TCs) in the Gulf of Mexico and the Caribbean Sea. The following methods were used to investigate the effect of the MJO on the likelihood of RI in the Gulf and Caribbean: the probabilities of TCs undergoing RI in the Gulf of Mexico and Caribbean (and in the entire North Atlantic) were calculated for each phase of the MJO (using the eight-phase index developed by Wheeler and Hendon 2004). Phases 2 and 8 exhibited enhanced probabilities of RI in the Gulf of Mexico and Caribbean (and the entire North Atlantic). The probabilities of TC genesis were also calculated for each MJO phase in the Gulf of Mexico and Caribbean (and the North Atlantic) in order to determine whether the increased probability of RI in certain phases is mostly attributable to an increased probability of TC genesis or whether other environmental conditions largely independent of genesis are responsible for enhanced RI in certain phases. Additionally, vertical wind shear anomaly maps, outgoing longwave radiation (OLR) anomaly maps, sea surface temperature (SST) anomaly maps, and maps of the temperature anomaly at 250 hPa were made for each phase of the MJO in order to determine which large-scale conditions affected RI the most and whether there was a constructive relationship between wind shear and OLR for certain phases. None of these variables explains the MJO’s modulation of RI, although there is a constructive relationship between negative OLR anomalies, negative wind shear anomalies, and positive SST anomalies in phase 2 that (at least partially) explain its high probability of RI. Finally, relationship between the Madden Julian Oscillation and the El Niño Southern Oscillation (ENSO) was briefly discussed. Diverging theories about how ENSO is expected to change with warming were introduced, and future projections of the MJO were made (using bandpass-filtered variances of OLR and monthly SST climatology maps) and analyzed in the context of the relationship between the MJO and ENSO.

Past literature has identified several factors that limit deep convection, including sea surface temperature (SST) and the humidity within and around an atmospheric column. These limiting factors often have a nonlinear relationship to deep convection, imposing thresholds below which convection is rare. Empirical values for these thresholds are well known, but there is not a well-known theoretical basis for the values themselves. Why is deep convection rare below SSTs of 26°C rather than 24°C? Why is it rare for water vapor content below 35 kg/m2? This paper uses reanalysis data to compute moist static energy (MSE), a metric which accounts for temperature and humidity simultaneously. We test the hypothesis that deep convection begins when local MSE surpasses the rainy tropical maximum MSE (i.e. “lid” MSE) and find that in many regions, a local MSE within 5 kJ/kg of the lid MSE is enough for deep convection onset to occur. However, it is hard to find universal support for the hypothesis, as behavior varies widely between regions. Factors including latitude, topography, land-sea ratio, distance to lid origin location, and dry air entrainment may play a role in this variability.