Undergraduate Research Symposium 2025

Friday, October 17, 2025 Hartley Conference Center and PatioMitchell Earth Sciences Building

Research presented by students of the SDSS undergraduate research programs:

Program Schedule

2:00-3:00 pm - Select oral presentations in Hartley Conference Center 
3:00-5:00 pm - Poster presentations on the Patio 
5:00 pm - SDSS Alumni Awards Reception

Oral Presentations

Where Dendrites Begin: Spatial and Temporal Solid Electrolyte Interphase Evolution as a Mechanism for Non-Uniform Lithium Plating
Steven Liu, SUPER

Would My Home Survive An Earthquake? How Uncertainty in Building Inventories Impacts Our Understanding of Seismic Risk
AnnaElisa Huynh, SESUR

I'm Still Standing: Optimizing Ecosystem Services Through Tree Restoration in the Brazilian Atlantic Forest
Alice Heiman, MUIR

Sustainable Battery Manufacturing: Dry-coated Electrodes For Sodium-Ion Batteries
Alexi Lindeman, SUPER

Modeling Impact-Generated Tsunamis over a Late Hesperian to Early Amazonian Ocean on Mars
Jackson Moyer, SESUR

Abstracts of Posters and Oral Presentations

Assessing Distributed Acoustic Sensing Ground Motion Estimation with Co-located Seismometer Network for Earthquake Early Warning
Lela Hanson, Geophyics

This research expands on work in the field of Distributed Acoustic Sensing (DAS), a new technology that exploits unused optical fiber infrastructure for seismic strain sensing. When deployed on telecommunication cables, particularly offshore, the method has the potential to significantly enhance the coverage of Earthquake Early Warning (EEW) systems. However, while current EEW systems are designed to receive measurements of peak ground acceleration (PGA) from conventional seismic stations, the correlation between DAS amplitudes and ground motion is still not fully understood. There exists a need for more observations to validate this relationship before DAS technology can be included in EEW operations.

In this study, DAS peak strain-rate (PSR) values are converted into ground motion predictions by estimating the apparent wave speed using a slant-stack approach. These predictions are compared with real measurements of PGA from co-located seismic stations from the Community Seismic Network to observe correlation and evaluate the accuracy of the method. Results provide new insight into the relationship between DAS amplitudes and ground motion and demonstrate the potential for DAS measurements to augment existing EEW systems.

Building Data Foundations for Understanding Surface-to-Bed Meltwater Pathways on the Greenland Ice Sheet
Anagha Ramaswamy, SESUR

The Greenland Ice Sheet experiences accelerated mass loss through increased surface melt and glacier calving, yet the role of surface-to-bed meltwater pathways in completely understanding ice dynamics is poorly constrained in many aspects. Moulins, which route supraglacial meltwater to the bed, form a very important component of this process (can extend up to 1000 meters in depth) but are difficult to detect at scale since they are spatially sparse and small in size (0.5 - 10 meters in size). Current manual mapping approaches are limited to local domains. The development of annotated datasets for machine-learning detection is an intelligent step toward scaled mapping for bigger regions. This project focused on generating moulin training data from remote-sensing WorldView imagery. Moulins were manually labelled across nine multiband and higher resolution monochrome images in QGIS, and calculated fracture-cut and water-cut masks were used to guide annotation.

Annotation was focused not only on stream-ending moulins but also on surrounding features such as bath-ring formations, geometries of previously drained moulins, and ice burst-like structures radiating from entry points. Contextual cues provided added confidence in distinguishing moulins from other dark surface depressions or debris. The labels were lastly converted to files optimized for preprocessing and tiling pipelines, thereby helping produce image and mask arrays optimized for machine-learning workflows. To validate consistency across datasets, a visualization routine that plots image data, moulin masks, and ice fractures in parallel was implemented.

The resulting deliverables include moulin annotation masks, preprocessing workflows, and visualization scripts. Together, these outputs form an introductory basis for future model-based prediction, answer questions regarding the use of segmentation over classification for feature detection, and improve parameterizations of surface-to-bed meltwater routing in ice-sheet models. By improving the detection and mapping of moulins, this work acknowledges an important bottleneck in climate policy, long term adaptation and coastal infrastructure planning.  

Carbon Capture and Direct Air Capture Implementation in California
Miki Yang, SUPER

California has set a goal to achieve carbon-neutrality by the year 2045. Within that goal, California aims to sequester 100 million metric tons of CO2 through carbon capture and storage processes. My research examined three questions: 1) What regions are most suitable for carbon storage in line for California's geologic storage? 2) How could curtailed solar—excess solar energy at certain times of the day—be utilized to supplement direct air capture units? 3) How much do carbon capture technologies cost in different models and contexts? To answer these research questions, I utilized arcGIS to geospatially analyze and visualize where solar curtailment occurs by county in California. Additionally, I performed techno-economic analyses based upon research papers to simulate different case scenarios of price ranges.

Causal Impact of Flooding Events on Acute Gastrointestinal Illness Risk in California: A Case-Crossover and Causal Machine Learning Analysis
Emma Beharry, SESUR

Floods are the most common natural disaster, and due to climate change, annual global flood exposure is expected to increase drastically. However, the causal impacts of flooding events on human health remain understudied, viewed through the lens of correlation rather than causation, and do not leverage machine learning methods. This project conducts the first large-scale preliminary causal analysis of the relationship between flooding events and acute gastrointestinal illness (AGI) in California between 2016 and 2019. Using a dataset of approximately 6 million emergency room hospital visits, the project applied two methods: (1) a case-crossover study with conditional logistic regression and (2) double/debiased machine learning (DML) with random forest. Flood exposure was defined at the three-digit zip code level and measured across time lags of 0 to 4 days, 5 to 9 days, and 10 to 14 days after each flooding event. Preliminary results from the case-crossover analysis suggest short-term associations, with the strongest effects observed within 0–4 days post-flood; however, several zip codes had negative associations across all time lags. DML estimates also showed negative average treatment effects, though results varied depending on the covariates included. Together, these findings suggest that flooding may increase AGI risk in the short term, while also highlighting the importance of controlling for confounding factors, such as temperature, precipitation, and demographic data. Future work will refine causal inference through improved exposure measures, expanded covariate sets, and testing of alternative machine learning models.

Colorimetric Detection Of Fecal Contamination In Wastewater Samples
Tejaswi Poudel, MUIR

This project focused on developing a simple, low-cost colorimetric assay to detect fecal contamination in water, using E. coli, coliphage, and a colorimetric substrate as the core components. A variety of diseases can be spread through fecal contamination including Typhoid fever which continues to pose a major public health challenge in low and middle-income countries, where access to diagnostic testing such as blood culture is limited. Environmental surveillance through wastewater testing offers a promising alternative. The project used the existing EPA protocol to test field water samples from the Philippines, and five out of fourteen were positive for fecal contamination. In our next steps we tried to adjust this to a colorimetric assay. This colorimetric assay was guided by the goal of creating a field-deployable assay that could act as a screening tool for fecal contamination. This method is cheaper, and much quicker and easier to use. To optimize the assay performance for this, a range of experimental conditions were systematically tested. These included adjusting the initial concentrations of E. coli and coliphage to identify combinations that would reliably trigger detectable phage infection and produce a visible color change. Various growth conditions were also explored, including incubation at different temperatures (room temperature, 30°C, and the optimal 37°C), and using different growth conditions, ratios, etc. Incubation times were varied between 2 to 6 hours to assess the effect of bacterial growth phase on assay sensitivity. Although the assay is still under development, these foundational experiments provide a roadmap for continued refinement and highlight key variables for future testing. Further work will aim to improve reliability, sensitivity, and ease of interpretation for low-resource settings.

Counter Current Heat Exchanger for Sustainable Wood Heating
Chloe Hei, MUIR

Wood is the most commonly used biomass energy source, particularly for heating and cooking, contributing to air pollution that can have serious health and planetary consequences. This project aims to develop a stove with a counter-current heat exchanger containing extremely thin rectangular channels that act as pathways for burning pollutants contained in the air influx. The clean air output will then be transported between the annular spaces of the tubes. After much prototyping, the optimal design of the rectangular channels has been created. Assembling the channels into a more compact and durable stove prototype will be the next step. After more prototyping and testing, we hope that we can bring this sustainable wood stove heating device to improve the environment of communities that must rely heavily on wood heating.

Designing Your and Your Community's Energy Lifestyle (DYCEL): Energy Resiliency, Electrification, And AI Applications
Charlie Abowd, SUPER

The Designing Your and Your Community's Energy Lifestyle (DYCEL) program is a 10-week summer internship that engages high school students in energy data science, resiliency, and electrification at Stanford University. As the energy system undergoes transformation toward renewables and electrification, and artificial intelligence increasingly shapes society, household energy consumers gain access to data streams from smart meters, solar panels, electric vehicles, and smart appliances. However, research demonstrates that despite this data availability, households continue to have misconceptions about their energy consumption, largely due to limited ability to understand and respond to these data streams. The program addresses this challenge through a two-phase approach. In Phase 1 (weeks 1-4), students analyze their own household's smart-metered electricity data using tools including Tableau, Python, and AI assistants (ChatGPT, Claude, Perplexity) to develop data stories and interactive dashboards. Students learn core energy concepts such as load shapes, base and peak loads, and Time of Use pricing. In Phase 2 (weeks 5-10), participants use the Stanford Sustainability Data Commons platform and other open-source datasets to create data narratives focused on energy resiliency or electrification, culminating in presentations to expert jurors. This project-based program includes weekly lectures by Stanford experts, self-guided Canvas modules, virtual meetings, and mentorship from teaching assistants. By teaching youth data thinking, visualization, and storytelling capabilities, DYCEL enables students to become agents of change in promoting sustainable futures, while contributing to research on data science pedagogy and online learning engagement.

Drought Impacts in Louisiana Coastal Freshwater Forests
Tung Nguyen, SESUR

Climate change is driving ecological shifts in coastal freshwater forests (CFFs) through increasing the frequency and severity of disturbance events, such as saltwater intrusion, storm surges, and severe droughts and flooding. These ecosystems are of significant importance to the local communities of Louisiana, as they provide integral ecosystem services, habitat for fisheries, protection from storm surges, and facilitate ecotourism. In 2023, an extreme drought occurred in Louisiana, encompassing most of the state. Drought is an uncommon occurrence in CFFs, and the impact of droughts in these ecosystems is poorly understood. It is important to quantify the timing and severity of changes in drought indices to further understand the response and recovery of the vegetation in these ecosystems. We applied the Breaks For Additive Season and Trend (BFAST) method to the time series from 1995 to 2025 of various indices (PDSI, SPI, etc.) from the GRIDMET/DROUGHT dataset. The BFAST method decomposes the time series into seasonal, residual, and trend components and identifies significant changes within the trend. Based on the data produced by the BFAST method, we concluded that further investigation of both the data and the method is necessary to accurately determine the impact and recovery from the drought. Future research will utilize higher spatial resolution data (PRISM), as well as vegetation indices and analysis of vegetation samples to investigate the effects on vegetation and recovery. These results can inform predictive models and management practices for coastal freshwater forests to protect ecosystem services for local communities in Louisiana.

Effective Communication Strategies for Oceans and Food Research
Sophia Sanders, MUIR

The world’s oceans play a critical role in feeding billions of people, yet aquatic food systems remain under-resourced and often left out of international food system conversations. Oceans are a vital part of global food security, nutrition, and livelihoods, and this must be communicated with key actors like policymakers, practitioners, and local communities in an efficient and effective way. This research, conducted in collaboration with the Stanford Center for Ocean Solutions, examines how to effectively communicate research findings in ways that foster understanding, engagement, and action. We found that different stakeholders and platforms require different communication strategies, and that engagement, clarity, and resonance across audiences depends on specific message framing. During the research, I developed recommendations for how the Center can frame messages regarding topics like microplastics and seafood, and evaluated the tools best suited to reach key audiences including interactive media, digital storytelling, and even journalism pitches. The findings highlight that effective communication is not just about simplifying data or stating facts, but about connecting research to values, decisions, and lived experiences. Clear visuals, narrative framing, and audience-specific messaging were found to be particularly powerful. This project contributes to ongoing efforts at the Center for Ocean Solutions to catalyze research and innovation into action, improving the health of the oceans and the people who depend on them most. Through effective science communication, we can better equip decision makers and communities with the knowledge needed to respond to the urgent challenges facing our oceans and global food systems.

 

Electrochemical Dechlorination of Chlorinated Aromatics for The Regeneration of Activated Carbon-Based Filters in Wastewater Treatment,
Aki Yuasa, MUIR

Chlorinated aromatic compounds, including chlorobenzenes, are persistent organic pollutants (POPs) that pose risks to human and environmental health. Many were banned in the 1970s due to toxicity, yet their persistence and bioaccumulation continue to pose risks, as evidenced by ongoing detection in human blood and ecosystems. Activated carbon (AC) is the current standard for removing POPs from water due to its high sorptive capacity. However, once saturated, spent granular activated carbon (GAC) must be replaced or regenerated. Existing regeneration methods are costly and either release contaminants or relocate them, rather than destroy them. Electrochemical reduction using AC based cathodes represents a promising alternative. Prior studies have demonstrated potential in other halogenated compounds, but not for chlorinated aromatics.

This work evaluates AC-based cathodes for (a) filtration of chlorinated aromatics in repeated column experiments and (b) subsequent electrochemical reductive dehalogenation. Using chlorobenzene-spiked water, we sought to demonstrate contaminant removal through AC cathode columns and electro-regeneration, introducing an alternative pathway for sustainable contaminant destruction and GAC reuse.

Electrochemical Removal of Ammonia from Wastewater: Testing Novel Flow Regimes
Bryan Nguyen, SESUR

Excess ammonia discharged from wastewater treatment plants into aquatic ecosystems is a cause of eutrophication and leads to algal blooms and fish kills. Conventional biological treatments can reduce ammonia levels, but alternative electrochemical ion-exchange methods are being investigated to convert ammonia into nitrogen gas (N_). This often involves a recycling brine model with separate catholyte and anolyte reservoirs flowing through an electrochemical cell with a cation exchange membrane, but this study focuses on testing alternative types of flow regimes through an electrochemical cell. The cell’s cathode raises the pH of the solution to convert the majority of the ammonium (NH4+) into ammonia (NH3), and the anode performs the breakpoint chlorination reaction to convert the ammonia into nitrogen gas. Instead of maintaining separate catholyte and anolyte reservoirs, this study evaluated configurations in which the brine flowed sequentially between electrodes. Three modified flow regimes were tested with varied flow rates: two involved an uncycled flow-through from the anode to an intermediate reservoir, then through the cathode into a final reservoir, and one was similar but cycled the catholyte back into the system with the initial reservoir. Results showed that the cycled regime achieved the highest efficiency in terms of cell potential and nitrogen removal, likely due to increased residence time and favorable pH gradients reducing energy demand via the Nernst effect. However, nitrogen removal reached only 17.89% over three hours, indicating limited overall performance. Future work should explore higher current densities and longer experimental durations to enhance nitrogen conversion efficiency.

Electrochemical Technique For Removing Ammonia from Wastewater?
Tuong Phung, MUIR

The current treatment technologies for ammonia removal are based on biological methods like nitrification-denitrification, which can require a lot of energy and food for the bacteria. With electrochemical system, we can that can replace the current ammonia removal system with less energy and cost-saving.

Engineering Open-Sourced Glaciological Software Defined Radar Equipment
Jhonny Almeida, SESUR

Software Defined Radars (SDRs) provide versatility for glaciological research as they are able to be reconfigured based on the research need of different glaciological targets. The Stanford Radio Glaciology Lab’s Open-Source Code Architecture (ORCA) provides open-source software for Ettus SDRs but lacks depth in physical building instructions for a complete SDR system. This summer I built the EYAS and PEREGRINE radars to assess how ORCA can incorporate building instructions and make these systems more accessible for labs and researchers seeking to use them in their own research projects.

Exploring The Effect of Mantle Composition on Historical Outgassing Rates From Earth and Venus
Amara Orth, SESUR

Volcanic outgassing is a key process that shapes planetary atmospheres and long-term climate. To investigate how mantle composition influences this process, the Schaefer & Sasselov (2015) mantle-atmosphere model was modified in MATLAB. The model was updated with new solidus parameterizations to account for the effects of water, CO_, and Fe. Water was incorporated through a concentration-dependent solidus, CO_ was represented using a carbonate solidus from Duncan et al., and Fe effects were shown by updating the viscosity law to link Fe content with mantle flow. Simulations were performed for Earth and Venus-like cases to track mantle temperature, mantle melting, and volatile release over geologic time. Results show that water lowers the solidus temperature, causing melting to start earlier and leading to stronger volatile releases compared to dry cases. This drives rapid mantle water loss and suppresses long-term water recycling. CO_ in contrast, continues to outgas at lower temperatures, producing a more extended release. Incorporating Fe-dependent viscosity further reduces volatile recycling and reduces outgassing. Together, these effects highlight how small differences in mantle composition can reshape volatile evolution and may help explain why Earth retained surface H_O while Venus became H_O-poor and CO_-rich.

How Does Social Capital Shape Health and Climate Justice in Bay Area Communities?
Anju Griffeth, SESUR

Climate hazards like extreme heat, wildfire smoke, and flooding have an outsized impact on frontline communities in the Bay Area, including Belle Haven, East Palo Alto, North Fair Oaks, and Redwood City. These neighborhoods, already facing deep-rooted inequities, encounter added obstacles in building resilience. While climate adaptation often focuses on technical fixes, this research sheds light on the crucial role of social capital—bonding, bridging, and linking—in shaping health outcomes and advancing climate justice. This study, part of a long-term community-based project with 315 households in San Mateo County, used a mixed-methods approach: reviewing social capital literature, analyzing structural pressures like gentrification, and conducting four diverse focus groups across San Mateo and Santa Clara Counties to capture intersectional insights on resilience strategies. Our findings point to widespread concerns about displacement, visible in changes like tree removal and the loss of green space. Residents reported deep mistrust of government agencies but strong reliance on local nonprofits and grassroots networks they trust. Participants also described the health harms of climate change: respiratory problems tied to wildfire smoke, unbearable heat that turned homes into ovens, and uneven access to cooling relief. At the same time, they emphasized everyday strengths such as cultural pride, mutual aid traditions, and intergenerational knowledge-sharing as vital factors of resilience. These insights carry important policy implications. Supporting community leadership through platforms that are community-driven and culturally responsive, investing in local grassroots organizations, and recognizing cultural assets in adaptation planning are key steps forward. Relying on numerical data alone risks overlooking lived experiences, which is why ground-truthing with communities is essential. By acknowledging both risks and strengths, this study shows how social capital can play a central role in building health resilience and advancing equitable climate adaptation in vulnerable Bay Area neighborhoods. 

I'm Still Standing: Optimizing Ecosystem Services Through Tree Restoration in the Brazilian Atlantic Forest
Alice Heiman, MUIR

The Brazilian non-governmental organization SOS Mata Atlântica is working prioritize areas in the Atlantic Forest biome for reforestation to maximize conservation return on investment. They have identified a priority area that encompasses parts of the Tiête and Paraíba do Sul watersheds, including 170 municipalities across the states of São Paulo, Rio de Janeiro, and Minas Gerais. This study modelled the potential impacts of reforesting the riparian buffers that are protected by the Brazilian Forest Code (permanent protected areas, APPs) in that priority area. Reforesting APPs to achieve compliance with the Forest Code would increase forest cover in APPs in the priority area from 46.3% to 100% and would increase forest cover across the whole area from 28.1% to 44.3%. The Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) software Nutrient Delivery Ratio, Sediment Delivery Ratio, and Carbon Sequestration and Storage models, developed by The Natural Capital Project at Stanford University, were used to model the impact of land use on nitrogen, phosphorus, and sediment export to waterways and carbon storage. The models predict that restoring all APPs would decrease nitrogen export to surface freshwater by 11.86%, reduce phosphorous export by 20.84%, and reduce sediment export by 16.46%, thus improving river water quality. The study area with restored APPs would store 437 million metric tons of aboveground carbon, an 11.07% increase over the current land use. We identified the municipalities where restoring riparian buffers would have the greatest reductions in nitrogen, phosphorous, and sediment export and the greatest increases in carbon storage, and there was generally agreement in hotspots across these models. Consequently, focusing restoration efforts on these hotspot municipalities would increase river water quality and climate mitigation benefits.

Impact of Flooding on Key Health Impacts in California, 2016-2022
Shreya Ramachandran, SESUR

In California, flooding presents a major risk to people and livelihoods. More than 7.3 million people and structures valued at nearly $600 billion statewide are located in areas that have at least a 1-in-500 probability of flooding in any given year1. Flooding results not only in fatalities, but also in increased risk of adverse health outcomes, including respiratory and waterborne disease. However, the underlying causal links between key health metrics and flooding has not been well studied in California. 

Therefore, we utilized four years of emergency visit data between 2016-2022 and a poisson fixed effects regression model to study the impacts of flood occurrence on six outcomes of interest: cardiovascular, genitourinary, infectious, injury, mental, and respiratory. Multi-source data was used to identify floods and other relevant covariates; key confounders were controlled for, including air quality, temperature and population density. We observed a significant increase in respiratory, genitourinary and injury cases following flooding events. Flooding events in the top 10% of intensity resulted in an IRR of 1.20, a 20% increase in respiratory cases relative to no flooding. These results could be used to inform public health preparedness plans and resource allocation by identifying which clinical services will face the greatest surge demand after floods. 

Impact of mesopores on the structural stability of high-capacity electrodes particles for lithium-ion batteries
Natalia Armenta, SUPER

The core objective of this research is to build on the results of the paper, Mitigation of Volumetric Expansion in Silicon Anodes via Engineered Porosity: Electrochemical Performances and Stress Distribution Implication and explore the effects of having large pores in battery materials. Building on state-of-the-art modeling techniques, our work aims to enhance the microstructural optimization of silicon electrodes to extend their cycle life and improve overall battery performance. By incorporating mesoporosity into chemo-mechanical models, we seek to provide deeper insights into the design of high-capacity electrodes with superior stability. This research contributes to the development of longer-lasting and more reliable lithium-ion batteries, paving the way for advancements in renewable energy, electric vehicles, and portable electronics.

Implications of Varied Land Tenure Models for Tribal Cultural and Climate Resiliency in California’s Sierra Nevada
Hannah Dimock, SESUR

Recent die-off of pinyon-juniper woodlands and decreases in pine nut production across Southern and Eastern California have emphasized the critical need for resilient management practices and elevated Tribal leadership and stewardship in these ecosystems. In collaboration with the Bishop-Paiute and Washoe Tribes of California’s Eastern Sierra, this summer’s work consisted of interviews with Tribal members and Tribal staff to identify needs and opportunities for community climate resiliency. Each interview, ranging from 45 minutes to 2 hours, followed a semi-structured format. Broadly, interview questions regarded social and cultural connections to the local pinyon-juniper woodlands, land stewardship practices, relationships across governmental entities, and impacts of environmental change. I have chosen to focus my analysis on the implicit relationship between cultural and climate resiliency and the ways in which varied models of land tenure—reservation lands, public lands, trust lands, and allotment lands—dually impact the two. Using the interviews I conducted (n=20) as my data set, I reviewed interview transcriptions and summative memos to identify central themes and patterns, analyzing similarities and differences between the unique circumstances of the Bishop-Paiute and Washoe Tribes. As researchers and scientists continue to build scholarly knowledge about the uses, cultural values, and impacts of climate change on pinyon-juniper woodlands, the hope of this work is to inform Tribal environmental policies, projects, and programs regarding land management and restore Indigenous stewardship to the lands that Tribes have tended for thousands of years.

Investigating Deep-Sea Chemoautotrophy Powered by Carbon Monoxide Oxidation Using Genetic and Cell Density Evidence
Abby Nakamoto, SESUR

The marine carbon cycle is a key component of the global carbon budget, yet major gaps remain in our understanding of organic carbon production and consumption in the ocean. Microorganisms play a critical role in long-term carbon sequestration by fixing carbon dioxide via chemoautotrophy. Chemoautotrophy is the process of carbon fixation using chemical rather than light energy. Deep-sea chemoautotrophs, often excluded from carbon cycle models, may help explain discrepancies in estimates of oceanic carbon sinks. This study investigates the potential for deep-sea chemoautotrophy powered by carbon monoxide (CO) oxidation through two complementary approaches: (1) incubation experiments assessing microbial cell density responses to CO across depths (5m to 4000 m) in the Northeastern Pacific, and (2) genetic analyses of a metagenome-assembled genome from the Orca Basin, a deep-sea hypersaline brine pool. Cell counts revealed exponential declines in microbial abundance with depth. CO incubation cell density trials indicate potential for simultaneous microbial growth and mortality. Genetic analysis revealed the presence of CO-oxidation genes and pathways for autotrophic carbon fixation, though activity may be constrained by environmental stressors. Together, these findings provide both physiological and genomic evidence for CO-oxidizing chemoautotrophs in the deep ocean. This work highlights CO oxidation as a potentially overlooked energy source for microbes with implications for powering carbon sequestration, which can be further studied to improve the accuracy of global climate models. Future research may focus on measuring gene expression in less stressed conditions and conducting extended incubation trials to classify specific chemoautotroph metabolisms and CO effects on growth rates.

Lowering the Price of Floating Offshore Wind Turbines (FOWTs) Utilizing Engineered Wood Techniques
Paolo Reitz, SESUR

Floating Offshore Wind Turbines (FOWTs) face increasing economic challenges due to declining government subsidies and the high cost of conventional composite materials. The vacuum molding of carbon and glass fibers, while effective, is time-consuming and expensive, limiting the scalability of these systems. This research explores the use of engineered wood products as an alternative material for FOWT towers and blades. Specifically, OpenFAST—the wind turbine simulation software developed by the National Renewable Energy Laboratory (NREL)—was configured to model structures built from Douglas fir marine wood in glulam and veneer forms. Material properties were incorporated into the simulator to assess structural feasibility and cost efficiency.

Preliminary results indicate that engineered wood could serve as a cost-effective alternative to composites, particularly for small- to mid-scale turbines. The findings suggest that community-scale FOWTs constructed from wood products may reduce material costs, lower carbon emissions, and expand access to renewable energy in distributed settings. While further testing and validation are needed, this work demonstrates the potential for engineered wood to contribute to more affordable and sustainable offshore wind deployment. 

Managing Co-Cultures: Exploring Antibiotic Effects on the Growth of Marine Ammonia-Oxidizing Archaea
Istaara Amjad, SESUR

Ammonia-oxidising archaea (AOA) are chemosynthetic microorganisms that play a central role in the marine nitrogen cycle, as the predominant species of ammonia-oxidizers in the ocean ecosystem. Considering their abundance in the marine microbial population, it is pertinent to investigate the structure and function of these organisms in the marine ecosystem, including their nutrient sources, metabolic products, relationship with other microorganisms, and role in global nutrient cycles. We examined the ammonia-oxidising archaeon Nitrosarchaeum limnium SFB1, sourced from low-salinity sediments in the San Francisco Bay and coastal sand from the Stinson Beach. The enrichment cultures we have grown are in co-culture with heterotrophic bacteria, which is a common obstacle in the study of archaea that complicates efforts to study archaeal function in isolation. Our primary objective at this stage was to establish effective methods for isolating and cultivating the archaea in laboratory conditions. To this end, we experimented with the effects of multiple antibiotic treatments on enrichment cultures. We monitored the growth dynamics of the archaea by measuring the nitrite accumulation in each sample, as it reflects archaeal activity, alongside direct measurements of cell count and cell density data. Collecting this data allowed us to evaluate whether the growth rate differed across various antibiotic treatments, and to distinguish archaeal nitrification activity from potential contributions of co-cultured bacteria. This data provides a foundation for future experiments involving N. limnium and its physiological and ecological interactions. For instance, we plan to study the metabolic interactions of the archaea. 

Mid-Paleozoic Redox Shift and Associated Zooplankton Changes
Anneka Steen, SESUR

The evolution of land plants in the Silurian-Devonian fundamentally transformed the Earth system. One possible change was a shift in marine redox conditions, as enhanced weathering and increased atmospheric oxygen affected the transportation of ions into the marine water column. Based on data from stratigraphic sections in Yukon, Canada, it is hypothesized that the ocean changed from dominantly ferruginous (iron-rich) to more euxinic (sulfur-rich) in the Early Devonian, while remaining largely anoxic. We tested the regional extent of this redox change using Ordovician-Devonian shale from Bathurst Island in the Canadian Arctic Archipelago, deposited along an open-ocean margin. Using geochemical proxies including iron speciation and trace metals such as molybdenum, we reconstructed redox states at the time of deposition. Our results indicate largely ferruginous conditions throughout the Ordovician-Silurian with a shift towards euxinia during the Emsian (Early Devonian). We also investigated whether this redox change coincides with ecological change in marine zooplankton. The extinction of graptolites, which has been globally observed to occur around the Pragian/early Emsian, and the rapid rise of tentaculites around the same time, may reflect the increasing prevalence of sulfidic conditions. We postulate that increasing euxinia may have been a driver of extinction in graptolites while favoring tentaculites.

Modeling Impact-Generated Tsunamis over a Late Hesperian to Early Amazonian Ocean on Mars
Jackson Moyer, SESUR

Large volumes of liquid water once flowed on the Martian surface. In the late Noachian, precipitation carved channels that formed river networks. In turn, rivers transported sediment downslope, depositing deltas in lakes and, possibly, a northern ocean. Confirming the existence of a paleo-ocean would fundamentally reshape our understanding of Mars’ climate history, water cycle, and its potential for past life. As the planet’s atmosphere thinned, precipitation ceased, and rivers disappeared by the Early Hesperian. However, an ocean could have subsisted well into the Late Hesperian to Early Amazonian. Evidence for a Late-Hesperian to Early-Amazonian ocean includes boulder deposits emplaced by putative meteor-impact-generated tsunamis. Here, we seek to better understand the dynamics of tsunamis in a putative northern ocean on Mars through modeling, with the ultimate goal of testing the tsunami hypothesis from boulder deposits. Specifically, we set out to better understand the phenomenon of wave refraction in the Martian environment and on real topography. Wave refraction is not affected by gravity; however, the slope of the ocean bottom and the frequency of the wave form exert a strong control on propagation trajectories. Using ANUGA, a Python-based shallow-water-wave hydrodynamic model, we simulate the dynamics of wave propagation and inundation of the shoreline. My project was to derive and implement a simplified model initial condition to mimic the ocean bottom and water surface elevation immediately after impact using impact-cratering physical scaling relationships and mass conservation. Currently, I am evaluating its ability to produce realistic tsunami wave heights and frequencies.

Modeling the Effects of Anode Binders on Silicon Particles in Lithium Batteries
Diego Romero, SUPER

Amidst a global energy transition to renewable and clean energy sources, batteries remain one of the forefront challenges in making this transition possible. While receiving widespread adoption across many different applications, the industry-standard lithium-ion batteries are limited by their energy density, which prevents their adoption in aviation and fully outperforms gasoline in automobiles. An emerging solution is using Silicon in the battery anode rather than graphene. As an anode particle, silicon can promise a far greater theoretical specific energy and energy density. Along with this potential, there are also significant challenges associated with creating lithium-silicon batteries, including a 300% increase in volume during charging and discharging.

Throughout the summer, I worked on creating mathematical models for lithium diffusion into Silicon anode particles using MATLAB and COMSOL. Specifically, I worked with numerical methods for solving the diffusion equation and incorporating relevant parameters to model lithium diffusion at the scale of a single silicon anode particle. After creating the baseline models, my focus turned to understanding how anode binders affect lithium diffusion at this scale. Binder materials are used to help maintain the structural integrity of the anode and handle the stress and strain of the aforementioned volume increase. Using mathematical modeling gives valuable insight into the effects different binder materials have on silicon particles, and what materials are promising for use in physical batteries.

Modeling the Impact of Home Vehicle-to-Grid Charging on Distribution Grid Stability
Molly Maloney, SUPER

The increasing adoption of electric vehicles (EVs) poses challenges to distribution grid voltage stability. While prior studies suggest that siting charging infrastructure closer to the grid’s electricity source improves stability, most EV charging occurs at homes, which are typically located toward the ends of grid branches. This spatial mismatch can cause voltage instability. Vehicle-to-grid (V2G) charging technology can mitigate these instabilities to mediate this tension between optimal siting and user charging preferences. This study models varying EV penetrations and V2G on the IEEE 34-bus Test Feeder, assigning domestic and industrial loads and incorporating EV charging profiles from 748 California drivers. A V2G control algorithm reduces voltage violations by 11%, often improving conditions at nodes not directly participating in V2G, though daily variability sometimes increases violations. These findings highlight how interactions between charging behavior, distribution grid layout, and voltage regulation affect stability. These results are important for grid planners considering where V2G charging stations might complement home charging to reinforce voltage stability at vulnerable nodes under increasing EV adoption.

Monodominant Eucalyptus Seeds Germinate Faster and at Higher Levels when grown with Live Microbial Communities
Hailey Zhang, SESUR

Plants and soil microbial communities form complex relationships that have vast implications on their surrounding environments, specially on the local plant species population, nutrient cycling, decomposition, and more. Plant-microbe interactions, particularly plant-soil feedback (PSF), are a major factor in the formation of monodominant patches, where a single plant species makes up the majority of the community. Here, research has shown that soil characteristics, specifically microbial communities, favor plant species by generating positive plant-soil feedback. However, few studies have compared the effect of microbial communities on generating positive plant-soil feedback versus the effect of chemical exudates present in soils. Here, we germinated Tasmanian blue gum eucalyptus (Eucalyptus globulus) seeds, an invasive plant well-known for establishing monodominant patches, in two different soil treatments: live soils collected from three sites at Point Reyes National Seashore and sterile soils that were autoclaved to remove microbial communities. Eucalyptus trees in Point Reyes form monodominant stands, areas in which Eucalyptus is the dominant plant species and biodiversity is vastly reduced due to the suppression of other plant species. We expected that due to positive plant-soil feedback, the seeds would have greater germination timing and rates in live microbial soils compared to sterile soils. Statistical analysis revealed that live microbial soils led to faster germination and higher germination rates among Eucalyptus seeds. Our main objective was to find out the ability of soil microbes to alter the germination timing and rates of Eucalyptus to protect native plant biodiversity in California ecosystems. As climate change alters resource availability and leads to shifts in plant ranges, dismantling the mechanisms behind plant-soil feedback has become more vital than ever.

Near-Real-Time Monitoring of Livestock Methane Emissions Using Multi-Source Data Integration
Kim Nguyen, SESUR

Methane is a potent greenhouse gas with a high global warming potential and short atmospheric lifetime. Unlike carbon dioxide, methane reductions can yield rapid climate benefits within decades—within our lifetime—making it a critical target for near-term mitigation, especially given that roughly two-thirds of emissions are anthropogenic. Livestock is one of its most significant anthropogenic sources, yet official livestock population data—essential for emissions estimates—lag and are incomplete, limiting timely mitigation.

This project develops a machine learning model to predict near-real-time livestock methane population data and calculate its methane emissions. We focus on the top 50 livestock ultra-emitters, training the model on livestock trade prices, GDP, and population data from multiple sources to infer national-level livestock inventories for the period 2016-2024. We calculate emissions using the Tier 1 approach from the 2019 Refinement to the 2006 IPCC Guidelines for enteric fermentation. The approach produces monthly and annual estimates through 2024, extending beyond FAO’s data, which ends in 2022.

Comparisons with FAO-reported data reveal similar enteric fermentation emission levels and patterns at the annual level, although country-level divergences persist. In most cases, our estimates exceed those of the FAO, while in some instances, they are lower, despite the FAO data serving as inputs during overlapping years. In contrast, monthly estimates are less consistent and often fail to reproduce the same temporal trends as the annual inventory. The difficulty of capturing seasonal variability results from constrained model performance due to limited high-temporal-resolution livestock data. This work highlights both the potential and limitations of near-real-time livestock methane monitoring, as well as the critical need for higher-resolution livestock data to enable more accurate emissions inventories.

NeighborDrive: Community-driven Neighborhood Sensing through Vehicle On-board Sensors
Rahul Rejeev, SESUR

Homelessness has risen significantly in many U.S. cities since the COVID-19 pandemic, and San José is no exception. This surge has intensified demand for essential services while straining the limited resources available within the city, making it increasingly difficult to ensure that aid reaches those who need it most. Effective resource allocation such as food, sanitation, and water depends on knowing where and when needs are most acute. However, traditional tools, such as field surveys and fixed sensors (e.g., surveillance cameras), provide only partial insights about the location and time of the required needs, with limited scope, coverage, and timeliness. Specifically, the dynamic movement of unhoused populations makes this task especially challenging. To address the challenge of dynamic movement, we introduce NeighborDrive, which equips vehicles with mobile sensors including gas, thermal camera, RGB camera and temperature modules to collect continuous, granular data on neighborhood environments. This crowdsensing approach enables the creation of dynamic spatiotemporal maps of environmental stressors such as litter, air quality, and temperature. By detecting changing conditions in real time, NeighborDrive enhances the efficiency of place-based interventions and helps service providers, such City Authorities, and Non-Government Organizations (NGO), optimize limited resources. In the long term, NeighborDrive aims to connect environmental observations with well-being outcomes, offering cities a scalable tool to improve the lives of all residents.

 

Pinyon Community Climate Action Project: Climate, Health, And Social-Ecological Resilience of California Dryland Forests
Allison Burwell, MUIR

Pinyon-juniper woodlands serve as a critical component to the ecological, cultural, and social well-being of Tribal Nations in the Eastern Sierra. Pinyon pine trees and the edible pine nuts they produce hold deep cultural significance and are a nutritional staple to these communities and wildlife in the area. However, climate change, drought, wildfires, and land stewardship challenges are reducing the availability of pine nuts and limiting the resiliency of these ecosystems. These changes not only carry ecological consequences, but also implications for human health as a whole. This summer, I conducted semi-structured interviews under the Pinyon Community Climate Action Project with Bishop Paiute and Washoe Tribal members. This project seeks to understand how Tribes use and value these ecosystems, their pine nut picking practices, and environmental changes they have observed over time. From this we learned about the effects that climate change has already had on the physical, mental, and social health and well-being of these communities. Interviewees emphasized the importance of harvesting pine nuts in maintaining cultural practices, generational knowledge, and a spiritual connection to the land. They also described the emotional distress many have experienced in response to environmental change and a sense of fearful hope for what their reservations and Tribes will look like in the future. This underscores the importance of preparing for future impacts of climate change to mitigate the further impacts of environmental change on community health in vulnerable populations. The goal of our project is to advance Tribal stewardship over lands and create a plan that synthesizes community and Tribal goals for improving climate resilience.

Preliminary Quantification of the Economic Impacts of Subsidence in Madera County
Alexandra Beyret, SESUR

The San Joaquin Valley of California faces a historic struggle with subsidence driven by groundwater overdraft. The region’s primary economic sector, agriculture, relies heavily on groundwater usage, and excessive pumping over time has caused the valley to sink at rates exceeding 30 cm per year since 2006 (Faunt et al., 2016). Such severe rates of subsidence pose serious threats to agricultural productivity and regional infrastructure and are costly to mitigate, yet the true cost of subsidence has not been fully determined. This project examines Madera County, a region experiencing some of the highest subsidence rates in the San Joaquin Valley's alongside areas of minimal change, to develop a preliminary understanding of the economic impacts of subsidence from 2008–2023. GIS-based floodplain mapping, property value data, and estimates of infrastructure repair and policy expenditures suggest that subsidence imposes significant losses across multiple sectors, including agriculture, housing, and public works. However, limited data and inconsistent reporting on damages to infrastructure that may be affected by subsidence produce a challenge to confidently quantify county-level economic costs. This study therefore provides both a preliminary economic assessment of subsidence in Madera County and an identification of key data gaps. Ultimately, this project contributes to ongoing discussions of effective groundwater management strategies in order to ensure long-term economic and environmental prosperity in the Central Valley.

Recognizing Mantle Earthquakes Using Seismic Waveforms: Extending the Sn/Lg Method
Emmanuel Zheng, SESUR

The existence of cratonic mantle earthquakes has recently gained broad acceptance, but their global distribution and prevalence are still unknown. Numerical modeling using the Sn/Lg method can improve our ability to recognize such events. The Sn and Lg seismic phases travel in the upper mantle and in the lower crust, respectively. High-frequency (1-4 Hz) measurements of Sn and Lg amplitudes on the T (transverse-horizontal) component of seismic waveforms have demonstrated the presence of mantle earthquakes beneath the Tibetan Plateau (the “T/High” method).

Here we extend the current “T/High” Sn/Lg method to lower frequencies and to the vertical component of seismic waveforms because: (1) low-frequency measurements are less sensitive to scattering, attenuation, and site response; (2) low signal-to-noise ratios on the T component can restrict the application of the Sn/Lg method; (3) some seismic stations only record the Z component.

We tested the Sn/Lg method at broadband frequencies down to 0.02 Hz and on all three waveform components (Z, R, and T). The key feature of the Sn/Lg method is the step increase in Sn/Lg amplitude ratios for earthquakes immediately below the boundary between Earth’s crust and mantle (the Moho) as compared to those immediately above the Moho. Using synthetically generated waveforms, we showed that this effect remains prominent at low frequencies (down to 0.25-1 Hz) on the T component but is only prominent at high frequencies (1-4 Hz) on the Z component. We therefore propose two extensions of the Sn/Lg method based on our synthetics: “T/Low” and “Z/High.” We ran the existing “T/High” method alongside our two new methods on South Tibetan earthquakes and confirmed that both are effective on real data. We are now working to combine all three methods into a single, more robust metric.

Removal of Ammonia with Less Salt
Tuong Phung, MUIR

This project seeks to develop an electrochemical system capable of destroying ammonia with lower energy intensity than nitrification–denitrification, the current biological method. By using NaCl concentrations of 6%–10% in my reactors, I aim to achieve lower cell potentials and faster removal rates. I will then compare salt concentrations of 6%, 8%, and 10% to determine which level removes the most ammonia with the least energy.

Soil Moisture Dynamics Across Microclimates at Jasper Ridge Biological Preserve and Implications for Soil Respiration
Hailey Demars, MUIR

Climate models seek to project climate change by simulating the interactions between the atmosphere and other Earth systems. However, their limited spatial resolution can oversimplify local environments, ignoring microclimates: small areas with distinct climatic conditions created by topography and vegetation. These microclimates create temperature and moisture gradients that influence ecosystem processes such as soil respiration, a key pathway for CO_ emissions from soil to the atmosphere. At Jasper Ridge Biological Preserve, we investigated how microclimatic variability affects soil moisture and respiration dynamics through two complementary projects. In the first, we used TOMST TMS-4 dataloggers to monitor soil moisture and temperature across ecologically diverse sites, aiming to capture seasonal patterns and site-specific responses to rainfall events. Following storms, rehydrated microbial communities typically shift from dormancy to activity, triggering growth and CO_ release. In the second project, we aimed to quantify temperature response curves for autotrophic (root) and heterotrophic (microbial) respiration across microclimates. Total soil respiration would be measured with a Li-Cor 6400 system, and soil cores were incubated to isolate heterotrophic fluxes. While equipment issues limited our ability to complete the respiration measurements, we were able to characterize seasonal and post-storm soil moisture patterns across microclimates. Sites with drier soils are expected to limit microbial activity, leading to lower and more variable CO_ fluxes. In contrast, soils with prolonged moisture retention may sustain larger, more active microbial populations, resulting in higher and more stable respiration rates. This study emphasizes the significance of microclimates in the carbon cycle and recommends that they be more effectively integrated into global climate models.

Sustainable Battery Manufacturing: Dry-coated Electrodes For Sodium-Ion Batteries
Alexi Lindeman, SUPER

With sodium being far more abundant and widely available than lithium, sodium-ion batteries (SIBs) present a promising, more sustainable alternative to traditional lithium-ion batteries (LIBs). However, due to the larger ionic radius of sodium (Na_), SIBs typically suffer from lower energy density. To address this, our work explores the development of thicker, denser electrodes via dry processing techniques—an environmentally friendly and scalable approach that eliminates the need for toxic solvents. Using dry-coating specific Avolt equipment—including a powder mixer, kneader, three-roll mill, and calendaring system—we systematically studied the effects of material composition, morphology, and processing parameters on electrode formation. Our goal was to establish a repeatable, consistent process for producing dense, thick electrodes suitable for high-performance SIBs. We successfully fabricated dry-coated electrodes measuring approximately 11 _ 15 cm, composed of 5 µm hard carbon shards, achieving thicknesses of ~90–100 µm (excluding current collector) and densities exceeding 1.0 g/cm_ after calendaring. In comparison, spherical hard carbon enabled better fibrillation of the PTFE binder, but resulted in lower maximum densities (~0.97 g/cm_). These findings demonstrate the potential of dry processing and equipment-assisted optimization to advance the scalability and performance of next-generation sodium-ion batteries.

Techno-Economic Assessment of Conventional Lithium Extraction
Clara Drysdale, SESUR

By 2030, current mining capacity for Lithium (Li) is expected to fall 30% short of forecasted demand in the International Energy Agency net-zero scenario. Moreover, the United States currently depends on Li imports from China, Australia, and South America. In response, researchers have developed novel Li extraction techniques. To develop and commercialize new technologies, we must establish a baseline cost of conventional Li extraction for comparison and targeted innovation. We conducted a techno-economic assessment (TEA) of Li extraction from solar evaporation ponds. Using the Water Treatment Techno-economic Assessment Platform (WaterTAP), we modeled Sociedad Química y Minera’s (SQM) solar evaporation ponds in northern Chile. We estimate a baseline levelized cost of $6,458.61 /t Li in concentrated brine. The parameters with the greatest impact on the levelized cost are inlet Li concentration, evaporation rates, recovered solids processing costs, and government rights costs.

Techno-Economic Assessment of Conventional Lithium Extraction, ClaraDrysdale, SESUR, By 2030, current mining capacity for Lithium (Li) is expected to fall 30% short of forecasted demand in the International Energy Agency net-zero scenario. Moreover, the United States currently depends on Li imports from China, Australia, and South America. In response, researchers have developed novel Li extraction techniques. To develop and commercialize new technologies, we must establish a baseline cost of conventional Li extraction for comparison and targeted innovation. We conducted a techno-economic assessment (TEA) of Li extraction from solar evaporation ponds. Using the Water Treatment Techno-economic Assessment Platform (WaterTAP), we modeled Sociedad Química y Minera’s (SQM) solar evaporation ponds in northern Chile. We estimate a baseline levelized cost of $6,458.61 /t Li in concentrated brine. The parameters with the greatest impact on the levelized cost are inlet Li concentration, evaporation rates, recovered solids processing costs, and government rights costs.

The Impact of Climate Shocks on Food Prices and Urbanization Across African Countries
Aria Grossman, MUIR

In order to adapt to the impacts of climate change, which in many regions include an increased prevalence of climate shocks such as droughts and floods, we must understand how these climate shocks affect market activity and human behavior. This project seeks to illuminate the impact of climate shocks on food prices and rural to urban migration in various African regions. This two-part project occupies a unique position in the literature surrounding climate effects. The food prices component investigates how the prices of specific food items in markets across Eastern Africa change in response to climate shocks at a much more granular level than previous research has explored. The migration component uses census, satellite, and weather data and a two-way fixed effects model to pinpoint rural to urban migration patterns and informal settlement expansion in countries across the African continent over decades. These papers will support policymakers in considerations of how to best adapt infrastructure to meet the shifting demands brought about by our changing climate.

Uncovering Viral Diversity and Host Interactions Associated with Microbial Autotrophs from the Deep Sea
Jasmine Lewitton, SESUR

Viruses are abundant and dynamic components of deep-sea ecosystems, yet their interactions with chemoautotrophic microbes – organisms that produce their own food using chemical energy rather than sunlight – remain largely uncharacterized. This knowledge gap limits our understanding of how viral infection influences microbial metabolism, community structure, and carbon cycling in the deep ocean. Here, we leverage metagenomic data and viral-host linkage approaches to investigate the diversity and functional potential of viruses infecting chemoautotrophic microbes in the deep sea. We used bioinformatic tools to identify viral genomes and predict their microbial hosts using CRISPR spacer matches, based on metagenomes collected from 150 meters depth in Monterey Bay. We recovered a robust set of 5858 viral sequences with confidence threshold of 0.7 or higher (4999 had a confidence threshold of 0.9 or higher) and sequence identity to known viruses, confirming confident viral genome recovery. Viral genomes ranged from 304 to 464,851 bp and most were predicted to be lytic (3,644), with fewer lysogenic viruses (326). CheckV classified 318 as high quality, 520 as medium, and 6,963 as low quality. Only two genomes were circular, suggesting near-complete recovery in rare cases. Host-virus associations inferred through CRISPR spacer matching suggest viral infections in local microbial communities; however, further research is required to confirm whether these viruses specifically target chemoautotrophic hosts involved in carbon fixation at this depth. Our findings provide initial insights into the diversity and ecological roles of viruses linked to carbon-fixing microbes, highlighting their potential impact on the marine carbon cycle and the need to include them in future biogeochemical models. This analysis sheds light on how viral infections may affect carbon fixation and microbial interactions in coastal deep-sea environments. 

Where Dendrites Begin: Spatial and Temporal Solid Electrolyte Interphase Evolution as a Mechanism for Non-Uniform Lithium Plating
Steven Liu, SUPER

Lithium metal is a promising anode material for next-generation batteries due to its high theoretical energy density compared to current graphite and silicon-based anodes. A key challenge lies in stabilizing the solid electrolyte interphase (SEI), which plays a crucial role in performance and can suppress dendritic growth that leads to capacity decay and safety concerns. While prior studies have analyzed SEI preformation before Li plating and its compositional changes during post-cycling rest, little is known about how the SEI evolves in real time during active electrochemical cycling. In this study, we investigate the spatial and time dependent evolution of the SEI during active Li metal anode cycling, which we propose is a major factor in uneven Li growth during plating and dissolution during stripping. Through electrochemical tests, optical cell imaging, and electron microscopy, we observe that SEI composition is highly non-uniform across individual Li dendrites. Our results suggest that, unlike current understanding, the SEI at the base of Li dendrites is temporally aged with a more resistive composition that decreases Li ion conductivity, while the newly formed SEI at the tip exhibits high ionic conductivity. This heterogeneity creates gradients in Li ion transport that promote uneven plating and stripping. These findings shed light on a new mechanism of dendrite growth driven by asynchronous SEI aging during active cycling, with implications for the uniformity of Li stripping and the formation of isolated Li in rested cells. Our findings identify the electrolyte reaction stabilization rate as a key metric governing SEI homogeneity across the surface of a Li dendrite. Ultimately, this study provides new insight into SEI behavior in a previously unobserved state, offering pathways to improve the cyclability and capacity retention of lithium metal anodes.

Would My Home Survive An Earthquake? How Uncertainty in Building Inventories Impacts Our Understanding of Seismic Risk
AnnaElisa Huynh, SESUR

Natural hazards such as earthquakes can pose large risks to communities, and understanding their potential impacts is critical for planning mitigation strategies. Regional natural hazard simulations help us do so by exploring such impacts on a single city or region. These locally-specific risk assessments rely on high-resolution building inventory data as an input; however, underlying data sources often contain gaps, inaccuracies, and inconsistencies that hinder usage in structural risk analysis. Recent work proposed a methodology for synthesizing data to develop regional building inventories, and this project extends that work by presenting a quantitative evaluation of how decisions involving data sources and inventory development impact city-scale seismic risk estimates. Using Hayward, California as a case study, we analyzed widely-accessible national sources and region-specific local sources, such as tax data. We found that building inventory inconsistencies are more prevalent in national data than local data, exhibit geospatial clustering, and are linked to taller structures, mobile/manufactured homes, and multi-family housing. To demonstrate that decisions in inventory development create significant, non-uniform differences in building inventory composition and quantified seismic risk, we generated a comprehensive set of plausible inventories for Hayward from open-source national data and ran a regional risk scenario for a magnitude 7.0 earthquake on each one. Differences in generated inventories result in every structure possessing a range of possible values for specific building features and impacts of the seismic event. These spreads represent substantial uncertainty concerning Hayward’s building stock and the degree to which it is affected by an earthquake. Among residential occupancy classes, we found that uncertainty is concentrated in multi-family housing. Our study captures the uncertainty that stems from inventory development and is clustered around specific building features. This identifies limitations in current regional risk analysis and priorities to improve the exposure data necessary for disaster risk management.

What Events Can We See From Space With This Next Generation Satellite
Emily Jones, MUIR 

With NASA’s TEMPO instrument in geostationary orbit, launched in 2023 to monitor North America, we have been tracking air pollutants from large-scale events such as NASCAR races and NFL tailgates to evaluate what can be detected from space. TEMPO measures pollutants like nitrogen dioxide (NO₂), ozone (O₃), and formaldehyde (HCHO) before, during, and after these events, allowing us to assess whether spikes in emissions are captured by the satellite.

Beyond sporting events, we also examined larger-scale pollution sources including wildfires, port activity, and major metropolitan areas. Using TEMPO’s continent-wide coverage, we trimmed and visualized the satellite data to focus on these specific case studies, asking: what kinds of events, and at what scales, can we see from space with this next-generation mission?

Native and invasive plant functional composition affects late-season soil moisture in California serpentine grasslands
Atessa Anoshiravani, MUIR

Serpentine grasslands are critical habitats for California native forbs and grasses. Over eighty percent of California grasslands are non-native — but the harsher conditions of serpentine soils make them less favorable for non-natives. This project studied serpentine annuals’ water consumption under different climate and precipitation treatments, specifically comparing soil moisture consumption between native and invasive plants, and between early and late season plants.

This project was conducted in a mesocosm set up, with a total of 216 mesocosms. The mesocosms were divided into blocks of 12 sharing a climate and precipitation treatment. To measure soil moisture, we used time domain reflectometry (TDR) probes that send electrical currents through the soil to calculate water content. Each mesocosm had one probe at 35 centimeters and another at 60 centimeters depth. Three total rounds of TDR measurements were taken in June, July, and August. Statistical analysis was conducted in RStudio with the lme4 package and a generalized linear mixed model. The outcome was soil moisture and the independent variables were functional group, plant origin, precipitation and temperature treatments, month of measurement, block, and mesocosm.

Shallow (35 cm) soil moisture decreased between June and the later summer months (F(2, 587) = 33.910, p < 0.001), but there was no change in deep (60 cm) soil moisture (F(2, 538) = 2.640, p = 0.07). Uninvaded native plant community composition influenced shallow soil moisture by functional group (F(3, 592.573) = 23.654, p < 0.001) and plant origin (F(2, 593.283) = 30.205, p < 0.001). For deep measurements, the functional group (F(3, 542.872) = 34.004, p < 0.001) and plant origin (F(2, 541.446) = 24.107, p < 0.001) influenced soil moisture across native and invasive plant origin treatments. Precipitation and temperature treatments had no notable impact on soil moisture.

Plant water consumption helps make sense of plant resilience and invasion dynamics, so understanding these systems is crucial. Soil moisture declined in the shallow soil layer but not at depth throughout the summer — a relationship likely relating to evaporation. Invaders depleted available water in shallow soils, independent of their functional group. Without invaders, shallow soil moisture depended on native functional composition. Deep soil moisture depended on both native and invader functional group composition. Late-season plants without natives may draw down more water due to competition with early-season plants. Future studies should investigate early-season dynamics and per-unit-biomass differences in functional group impact.

Assessing Spatial Distribution of Nitrogen Concentration in Vegetation on Alcatraz Island to Understand Seabird Behavior
Quyen Vo, MUIR

Seabirds are important drivers of nutrient transfer between marine and terrestrial ecosystems since they spend most of their lives at sea and return to land to breed. Their nitrogen-rich guano has been observed to enrich surrounding vegetation through isotopic tracing experiments. Alcatraz Island has long stood as a breeding ground for bird colonies in the San Francisco Bay and is home to many thriving sea bird species, such as Brandt's Cormorants. Up to 30,000 birds are counted by the National Park Service (NPS) per year on Alcatraz, thereby providing the potential for a significant nutrient pulse to the island’s vegetation. Our study aims to investigate the influence of seabird colonies on the spatial distribution and concentration of elemental and marine nitrogen throughout various plant communities on Alcatraz Island.

Enhancing Regional Economic Resilience to Earthquakes Through Infrastructure Improvements
Margil Sanchez Carmona, SESUR

Earthquakes pose threats to individuals, infrastructure, and regional economic activities, either directly or indirectly. Damage to the built environment caused by earthquakes can trigger cascading effects through interdependency between physical and economic systems as well as the interconnectedness within economic sectors. Because of this, the role of resilience as it applies to these interdependencies has been increasingly studied. The aim of this project was to identify and evaluate the efficacy of mitigation strategies for seismic events on the regional economy by implementing certain strategies in predictive economic-sector and port-sector impact models. The identification process included extensive literature review of academic papers, government reports, and newspaper articles to contextualize what has already been explored in the field of earthquake and disaster resilience mitigations within the infrastructure and economic sectors. We identified 20 total strategies across three different sectors of interest (the business, port, and transportation sectors) and classified them as either adaptive (implemented after the disaster) or inherent (built into the system) mitigation strategies (Rose). The implementation process included a sector-specific prioritization analysis to identify critical sectoral mitigations, and testing a construction-sector mitigation into the economic model. We decided to test the impact of implementing an idealized overtime performance adaptation ratio of 1.5, which was found based on the overtime safety efficiency criteria outlined from a source in the literature (Goldenhar). This meant accounting for the “recovery” of half the lost labor hours by either increasing the production capacity ratio or the overproduction adaptation. Results showed that testing the strategy by changing the production capacity ratio shifted the recovery curve significantly (cutting recovery times), while testing the strategy by changing the overproduction adaptation had a much smaller impact on the original recovery curve, with a much more gradual recovery process.

Using GIS to plan and track sampling activities, ~150 vegetation samples were collected across 30x30m quadrant plots, to get representative samples of the plant types on the island. The samples were dried and grounded to a homogenous powder with a ball mill, and then processed with an Isotope Ratio Mass Spectrometer to determine elemental nitrogen concentration and isotope ratios. Since Nitrogen-15 is derived from marine environments, we are able to isolate marine-derived nitrogen and observe its impact on vegetation at individual sample and scaled plot-levels. Nitrogen concentrations were overlaid on a map across the island to reveal whether associations could be found against known seabird nesting sites. While this project is still in the data analysis stage, similar studies indicate a strong relationship exists between colony presence and abundance and enriched nitrogen signals in proximate vegetation. More broadly, this project aims to support NPS efforts in seabird vitality monitoring and to provide baseline nitrogen values across Alcatraz vegetation for remote sensing studies that may reduce or eliminate the need for fieldwork in sensitive coastal environments.

Gentle Winds Resuspend Toxic Lead Nanoparticles Into Air after Wildland-Urban Interface Fire
Dan Kubota, SESUR

The 2025 Wildland Urban Interface (WUI) fires in Los Angeles released substantial amounts of toxic metal particles, including lead (Pb), into soils and debris left behind after the fires. Postfire materials containing Pb can be resuspended into the air by wind, posing serious health risks through inhalation. We investigated how wind speed influences the resuspension of Pb particles after WUI fires. Samples were collected after both the Eaton and Palisade fires, and their Pb concentrations were measured. Selected samples with Pb concentrations above 500 ppm were introduced into a windbox, where they were resuspended under gentle wind speeds (6–10 mph), and total particulate matter with a size less than 2.5 µm (PM2.5) was measured using a PM2.5 monitor. The results showed that total PM2.5 concentrations in the windbox increased from 20 μg/m³ at 6 mph to >999 μg/m³ at 10 mph. Airborne particles were also collected with an active air sampler equipped with an impactor that collects particles with a size less than 10 µm, and particle mass measurements confirmed that the total mass on filters increased with wind speed. The presence of Pb particles on the filters was confirmed using scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectroscopy (EDS). SEM images revealed spherical and rhombohedral Pb particles ranging from <0.250 µm to 2 µm. These findings highlight the potential health risks posed by the wind-driven resuspension of Pb-containing postfire materials, which have implications for assessing long-term Pb exposure risks following WUI fires.