UC Davis

Anthropogenic climate change is a global issue that manifests in impacts of disproportionate severity across communities ranging from the regional scale all the way to local neighborhood differences. To address local and regional impacts, policy and planning decision-making at the regional, national, and local scales is required. This planning may facilitate adaptation to climate change impacts, as well as mitigation efforts that prevent further detriment, or have a reversing effect. Climate change impacts are experienced as a combination of regional conditions (e.g., biome, corresponding climatic and weather patterns) and social circumstances, which vary within a given region (e.g., income level and resources to adapt to changing conditions).While anthropogenic climate change impacts such as drought and sea-level rise occur irrespective of social circumstances, environmental degradation as a whole is not always as unbiased. Environmental degradation often persists in poor communities of color, versus affluent white communities. These groups disproportionately bear the brunt of burdens associated with pollution, waste, and poor air and water quality, as a result of issues such as the historical political power dynamic, and resource disparity to either zone polluting industries out of their communities or move away from the degraded environment.Environmental justice, as a concept, aims to rectify these undue burdens in terms of environmental degradation that disproportionately affect vulnerable communities. Evolutions of the movement have sprouted focused efforts, such as that of just transitions. Just transitions have most notable applications in the energy, climate, and environment spaces, revolving around procedural and representative justice. For purposes of this dissertation, just transitions is defined as the intentionality to include those who have been historically impacted, underrepresented, or excluded in decision-making and formation of environmental- and climate-related adaptation and mitigation plans, regulations, and policies, so as to foster equitable planning with minimized tradeoffs.Core to just transitions is inclusion, which can be operationalized through participation. Participation in decision-making processes not only increases representational equity in terms of embodiment of the full range of perspectives of those affected by a decision, but also increases procedural equity through the involvement of those perspectives within each step of the process. However, it can be difficult for decision-making bodies to determine when participation engagement is necessary, and how to facilitate engagement.Participatory scenario planning is a decision-making support tool that facilitates such engagement between the decision-makers and those affected by decision outcomes, often rendering these groups as one in the same. Participatory scenario planning approaches decision options as scenarios - narratives of plausible futures. Consideration of different scenarios allows fruitful thought surrounding a) decision-making options, b) pathways to achieve each option, c) benefits of the options, and d) consequences of the options. Employing participatory scenario planning therefore has the potential to produce informed, well-rounded decisions that, if facilitated and executed correctly, consider all relevant perspectives to maximize benefits and minimize trade-offs in decision-making. Participatory scenario planning offers flexibility and customization to employ engagement strategies most relevant and useful for each individual process. One limitation of participatory scenario planning, however, is its lack of defined methodologies for modeling the scenarios under consideration; literature varies in the extent of modeling undertaken, if at all, and type of modeling. Therefore, the process of deciding upon a given scenario is highly variable across exercises.Another decision-making support tool, life cycle assessment (LCA), is a standardized modeling tool that is often employed in the energy, climate, and environmental fields to gain insight on the impacts a given decision has on human health, the environment, and natural resources. As outlined by the International Organization for Standardization (ISO), LCA practitioners follow a formal methodology for modeling impacts. However, this methodology does not specify any participatory components; as such, there lacks a true participatory-LCA (p-LCA) framework in the literature, and p-LCA attempts to date have been inconsistent in their engagement methods.Combining participatory scenario planning with LCA to create a formal p-LCA framework produces a decision-making support tool that offers flexibility and customization while maintaining a standardized methodology, incorporating robust modeling, and combines qualitative insights with quantitative data to inform decisions. In this framework, individuals with a vested interest in the outcome of decisions have the opportunity to voice their perspectives and maintain ownership over the decision-making process, while still abiding to a clear order of operations to produce quantitative insights. In this way, the p-LCA framework will increase just transitions in the climate change decision-making space, while providing valuable decision-making data.This dissertation explores the fusion of participatory scenario planning and LCA to produce a formal p-LCA framework. The six chapters contribute to this exploration as follows: Chapter 1. Introduction. The first chapter introduces the overarching concepts of environmental justice, just transitions, and climate management strategies. The goal of this chapter is to level-set the contextual information as the foundation the dissertation relies on. The context is both historical and definitional. Concept history is explored to establish an understanding of the origins and evolution of central topics. Definitions are set for potentially ambiguous terms, or terms with varying definitions per field of application (for example, just transitions, or climate management decision-making) to ensure the intended interpretation is carried forward throughout the dissertation.Chapter 2. Participatory Scenario Planning as Applied to Climate Management Strategies. The second chapter in this dissertation employs a systematic review of the literature surrounding participatory scenario planning as applied to decision-making related to climate change. The intent of this chapter is to explore common practices from the literature, in order to suggest elements of participatory scenario planning methodology for implementation into the overall p-LCA framework.Chapter 3. Perceptions of Just Transitions in Participatory Scenario Planning of Salinity Management in the Delta. After establishing common practices in participatory scenario planning as applied to climate management decision-making, perception of actual climate management participatory scenario planning participants is explored through a case study, employing insights from the Just Transitions in the Delta: Drought, Salinity, and Sea Level Rise (hereafter "Just Transitions in the Delta") project. Analyzing survey responses to comments on perceptions of inclusion, equity, and justice related to the participatory scenario planning process and decision-making as a whole serve to determine the utility of employing participatory scenario planning processes through the lens of real-world participants.Chapter 4. Fusing Qualitative and Quantitative Practices for Climate Management Strategies. Chapter 4 establishes a proposed p-LCA framework through combination of the ISO LCA standardized methodology, and the common practices of participatory scenario planning gleaned from Chapter 2. Chapter 5. P-LCA of Salinity Management Strategies in the Delta.Chapter 5 serves as a proof-of-concept of the p-LCA framework established in Chapter 4, again through a case study involving the Just Transitions in the Delta research project. The p-LCA executed explores two alternative scenarios for salinity management in the face of climate change. Chapter 6. Conclusions & Recommendations. Chapter 6 concludes the dissertation, highlighting the pertinent results of each chapter, and provides further recommendations.The findings of Chapter 2 highlight common practices in participatory scenario planning processes, and insights into useful engagement strategies. Chapter 3 results confirm public acceptance and perceived utility and even enjoyment of the participatory scenario planning process, echoing findings from Chapter 2. Chapter 4 provides a pathway to fuse participatory scenario planning with LCA for a new p-LCA framework, to both bring a standardized modeling methodology to participatory scenario planning processes, and incorporate participation in the LCA process, for purposes of increasing representational and procedural equity in climate management decision-making for a just transition. Results of the proof-of-concept executed in Chapter 5 shows the p-LCA framework defined in Chapter 4 is both feasible to implement and useful in climate adaptation decision-making.

Many cultivars of almond (Prunus dulcis Mill. D. A. Webb) suffer from a disorder known as non-infectious bud failure in which preformed meristematic tissue becomes necrotic during dormancy, then fails to emerge in the following growing season. Non-infectious bud failure is a noteworthy disorder in perennial crops due to its miotic inheritance through asexual propagation and variance in severity between genetic clones. While the cause of the disorder is unknown, previous research has shown possible connections between DNA methylation and symptoms of the disorder. Non-infectious bud failure has been mitigated through propagation from basal epicormic buds, which show reduced non-infectious bud failure in clonal progeny relative to the axillary canopy buds typically used for propagation. In this dissertation, we explored several possibilities of detecting differences in DNA methylation and expression to develop potential biomarkers for non-infectious bud failure in almonds. First, we explored variation in DNA methylation between epicormic and axillary shoots that may provide insight into the spread and origins of non-infectious bud failure. We hypothesized that shoots derived from different tissues (epicormic and axillary buds) and growing environments would have distinct genome-wide patterns of DNA methylation. We sampled leaf tissue from epicormic and axillary canopy shoots from two almond varieties at four locations, then used whole-genome enzymatic methyl-sequencing to determine global methylation levels and differentially methylated regions between tissues. We found no significant differences in genome-wide methylation levels between different shoot types, locations, and varieties. Epicormic and axillary shoots displayed 23 differentially-methylated regions in one variety and 45 in the other, with no overlap in these regions between varieties. In addition, we identified 4546 regions that appeared differentially methylated between varieties; however, DNA polymorphism may account for some of these differences. In pairwise comparisons between locations, we identified 312 regions, with overlapping regions between some, but not all, locations. The remarkable stability of DNA methylation patterns across tissues and locations motivates further investigation into possible perturbations of these patterns by non-infectious bud failure and related genetic disorders in perennial plants. Next, we explored differences in DNA methylation and RNA expression between trees displaying different severities of non-infectious bud failure. To do so, we developed a new haplotype-phased reference assembly for the variety “Carmel”. One parent of Carmel, “Texas” has no history of bud failure symptoms, while its other parent, “Nonpareil” does. We then scored Carmel trees in a commercial orchard for non-infectious bud failure severity, then sampled leaf tissue from those trees. We used enzymatic methyl-seq to measure whole-genome DNA methylation and perform differential methylation analysis and we used 3` TAG RNA sequencing to measure gene expression between the levels. Our results show 20 regions of differential methylation between the two haplotypes. Additionally, 2083 genes were differentially expressed between the two haplotypes, with the expression of certain genes highly correlated with bud failure level. Three of these DMRs fall within 5000 base pairs of the differentially expressed genes. These areas of variation in methylation and expression data may indicate candidate genes for non-infectious bud failure, and may aid in the development of future diagnostic tools. Finally, we describe a successful program for zinnia breeding. Zinnias (Z. Elegans) have emerged as an important crop for many California organic growers. Due to their cold intolerance, zinnias cannot be refrigerated for extended periods of time. Therefore, production cannot be easily moved to locations that require long transportation times, which has resulted in localized production. However, most zinnia seed currently in production has been bred for conventional systems. Additionally, very few resources exist on the methods and strategies for improving zinnias for cut flower production. After evaluating available germplasm, we made 463 crosses, then selected 85 to evaluate in a single field season. Using a modified synthetic breeding strategy, we performed selection on color, petal doubling, stem length and flower size for five years. We introduce two new cultivars, ZP.23.225 and ZC.23.227, that have been bred specifically for organic systems and addressing new and emerging market demands. ZP.23.225 is a zinnia with a light pink color and a small flower head but maintains agronomic vigor. ZC.23.227 is an orange “cactus” type zinnia with good field performance and an exceptionally large flower head. These varieties were developed through the Student Organic Plant Breeding and Education program at the University of California, Davis.

Worldwide, over 15% of couples have difficulty conceiving a child. Long-term fertility in males is governed by spermatogenesis, the developmental program responsible for producing mature sperm. Sperm were once viewed primarily as carriers of paternal DNA, but they also carry epigenetic information that can influence fertilization and early embryonic development. This information is established through the remodeling of chromatin, the combination of DNA and associated proteins, as germ cells develop. However, why and how molecular regulators and environmental conditions shape the epigenome remains incompletely understood.To address this gap in knowledge, I investigated how intrinsic and environmental factors influence the epigenetic trajectory of the male germline. I approached this biological question using two complementary mouse models: genetic deletion of a chromatin regulator, ATF7IP2/MCAF2, and dietary folate deficiency. By independently assessing each model using a combination of imaging and genome-wide analysis of gene expression and chromatin, I found that disruption of either factor resulted in distinct, but converging cellular and epigenetic changes in spermatogenesis. Genetic deletion of ATF7IP2/MCAF2 revealed an unexpected role in gene regulation and epigenetic reprograming in male meiosis. These findings also refined our understanding of how meiotic checkpoints respond to disrupted chromatin regulation. In the second model, paternal folate deficiency altered germ cell development, meiotic gene expression and various chromatin features in meiosis. Together, these findings identify meiosis as a period of both epigenetic vulnerability and responsiveness. I propose that meiosis functions as a metabolic sensing window during which nutritional conditions can influence the chromatin landscape as it is being extensively reprogrammed. This conceptual framework helps explain how molecular regulation and environmental exposures interact to shape sperm development and potentially influence the information transmitted to the next generation.

Democracies across the globe, including large and significant democracies such as the United States, have experienced democratic backsliding over the last few decades, or even reversion to electoral authoritarianism. These trends have in turn prompted a literature that focuses on the correlation of democratic backsliding to various political, international, or societal factors. However, a relatively newer literature has focused on the process of democratic backsliding itself, or how prospective autocrats (backsliders) attempt to undermine their democracy. This latter literature has already provided important insights into the process of backsliding in the aggregate, but I find existing theories struggle to reliably predict individual backsliding sequences. The problem, I contend, is that the current literature focuses on identifying a narrow range backsliding sequences (or even a specific sequence) across all cases, which overlooks the influence of variations in (for example) institutions and society between one democratic context and another. Accordingly, I provide a theory that views individual backsliding sequences as idiosyncratic, but largely explainable and significantly shaped by three common and mutually-interacting, key elements: 1) backsliders' motivation for backsliding; 2) the institutions, societal actors, or laws that a backslider identifies as relevant for backsliding given their context ("targets"), and 3) the tools backsliders have available to undermine their democracy. To test my theory, drawing from a variety of sources (reports, news articles, scholarly journals), I collected data on 14 different cases where backsliding occurred, then subsequently conducted in-depth case studies on 8 of them, describing the backsliding sequence of latter cases in detail. I find that the logic of my theory largely holds, with backsliders visibly undermining targets relevant to their context, and episodes largely following my general predictions; however, my theory of how backsliders use their tools received relatively more mixed support. Ultimately, I contribute to the literature a new theory of the democratic backsliding process that is better at explaining individual backsliding cases, and an emphatic reminder that political context cannot be overlooked when theorizing about the process of democratic backsliding.

Faces are among the most important sources of social information in early childhood, providing cues about others' emotions, intentions, and mental states that support children's developing social understanding and behavior. The ability to perceive emotional facial expressions has been proposed to play a foundational role in the development of theory of mind (ToM) (Mitchell & Phillips, 2015; O’Brien et al., 2011), yet empirical evidence directly linking emotional face perception and ToM remains limited, particularly during early childhood when both abilities undergo rapid development. Moreover, although theoretical models of prosocial development posit that successful prosocial behavior depends on both children's capacity to reason about others' mental states and their ability to perceive socially relevant cues (Eisenberg & Mussen, 1989; Rahal & Fiedler, 2021), little research has examined how these processes work together to predict children's social behavior.To address these gaps, the present dissertation examined the role of emotional face perception in social-cognitive development using event-related potentials (ERPs), which provide a temporally precise measure of neural responses to emotional facial expressions (Luck, 2014). Across two studies of preschool-aged children, neural indices of early, automatic attentional orienting (P1) (Batty & Taylor, 2006), structural face processing (N170) (Blau et al, 2007), and attentional allocation to faces (Nc) (de Haan et al., 2002) were used to characterize individual differences in emotional face perception and their relations with behavioral measures of ToM and prosocial tendencies.The first study investigated whether individual differences in preschool children's neural responses to emotional facial expressions (as indexed by N170 and Nc) were associated with individual differences in behavioral ToM. Findings revealed a specific role for the Nc component, such that a greater Nc amplitude to negative (fear and anger) relative to neutral facial expressions were associated with more advanced mental-state reasoning, suggesting that individual differences in the allocation of attention to emotional faces are meaningfully related to emerging ToM abilities.The second study examined whether emotional face perception and ToM jointly predicted preschool children's prosocial tendencies. Results indicated that ToM was positively associated with prosocial behavior only among children who exhibited enhanced P1 amplitudes to emotional facial expressions, suggesting that early attention to emotional cues may facilitate the translation of social-cognitive inferencing into prosocial action.Together, these findings provide novel evidence that individual differences in neural responses to emotional faces contribute to variability in preschoolers' social cognition and prosocial behavior. By integrating developmental neuroscience with existing empirical and behavioral theories, this dissertation advances understanding of the mechanisms linking emotional face perception, mental-state inferencing, and prosocial development during a critical period of early childhood.

Lipid supplements are commonly used to increase the energy density of finishing diets, but their effects depend on how fatty acids interact with the basal diet and rumen environment. This thesis reviews the effects of lipid source, physical form, inclusion level, and basal diet on ruminant performance, carcass traits, meat quality, and meat fatty acid profile. It also evaluated relationships between trans(t)-10-shifted ruminal biohydrogenation and carcass fatness in feedlot-finished lambs and the effects of dietary induction of a t-10 shift. In the first experiment, 232 commercial lamb carcasses were evaluated, with backfat fatty acid profiles determined in 80 carcasses representing USDA Yield Grades 2-5. Backfat t10-18:1 (r = −0.50), trans-10,cis-12 CLA (t10,c12-CLA; r = −0.30), and the t10-18:1/trans-11 18:1 ratio (t10/t11; r = −0.54) were negatively correlated with Yield Grade, whereas t11-18:1 (r = 0.33) and cis-9,trans-11 CLA (r = 0.34) were positively correlated (P < 0.01). Based on these associations, a subsequent feeding trial was designed to experimentally induce a t-10 shift and examine potential cause-and-effect relationships between increased t-10 biohydrogenation intermediates, lipid metabolism, and carcass fatness. Twenty-four lambs were fed grain-based diets containing no added fat (CTL), 6% corn oil (COR), or 6% high-palmitic acid fat (PAL) for 35 days. The COR diet induced a pronounced t-10 shift, increasing t10-18:1 by ~400% in backfat and longissimus lumborum muscle and t10,c12-CLA by ~300% in backfat (P < 0.05). This treatment also reduced fatty acid synthase (FAS) mRNA expression in backfat (P < 0.05), indicating an effect on adipose lipogenic regulation. The COR diet reduced dry matter intake compared with CTL and PAL and improved feed efficiency compared with PAL (P < 0.05), without affecting weight gain or final body weight. Hot carcass weight was also increased in COR compared with CTL and dressing percentage compared with CTL and PAL (P < 0.01) but did not affect backfat thickness or Yield Grade. Dietary treatments also did not affect instrumental color, cooking loss, Warner-Bratzler shear force, or lipid oxidation (P > 0.10). Collectively, the survey identified an association between a greater t-10 shift and lower carcass fatness, whereas the controlled feeding trial demonstrated that dietary induction of a t-10 shift markedly altered tissue fatty acid composition and some aspects of lipid metabolism without reducing carcass fatness. The relatively short 35-d treatment period may have limited the time available for changes in lipid metabolism to translate into measurable differences in carcass fatness. Longer-term studies are warranted to determine whether sustained exposure to a t-10 shift and its biohydrogenation intermediates can ultimately influence carcass fat deposition.

The flat-tailed horned lizard (Phrynosoma mcallii) is a Colorado Desert endemic species with a long history of conservation and active management. Ongoing monitoring has relied on traditional mark-recapture data to inform management action. I used reduced representation genotype-by-sequencing techniques to generate a large single nucleotide polymorphism data set to assess fine-scale P. mcallii population structure, dispersal, landscape barriers to gene flow, and genetic diversity in California’s Imperial Valley with the overall goal to inform ongoing management and monitoring of the species.In my first chapter, over 500 individual P. mcallii DNA samples were genotyped to assess fine-scale population structure across the Imperial Valley. Using sparse non-negative matrix factorization paired with geographic capture data and discriminant analysis of principal components, I detected population structure across the east-west axis of the Imperial Valley, which had been described using mitochondrial DNA markers. I observed less pronounced but significant population structure between West Mesa and Yuha Basin populations which may indicate contemporary barriers to migration. I detailed estimates of contemporary genetic diversity, finding generally lower than expected heterozygosity, low inbreeding, and little variation in nucleotide diversity. These results indicate management-relevant population structure that aligns with existing management units.In my second chapter I developed genetic pedigrees to identify sex-biased patterns of dispersal, isolation-by-distance and landscape barriers to gene flow. I calculated pairwise relatedness between individuals and measured the decay of relatedness by distance to investigate sex-biased dispersal; with these methods we detected evidence of male-biased dispersal. I also used this data to build pedigrees that revealed the first documented instance of multiple paternity in P. mcallii. I tested for isolation-by-distance using the inverse proportion of shared alleles and found patterns that indicate high gene flow, and little effect of isolation-by-distance. We used Estimated Effective Migration Surfaces to identify barriers to gene flow associated with CA-98 and I-8 as well as Superstition Mountain and two BLM OHV areas. These contemporary barriers seem to reduce gene flow, but it is unclear to what extent. We recommend further landscape analyses to understand how landscape arrangement and abiotic factors may influence gene flow in the Western Imperial Valley.In my third chapter I used samples collected from ongoing P. mcallii long-term monitoring plots to estimate baseline measures of genetic diversity for three management areas, estimate effective population size on each demographic monitoring plot, and test for heterozygosity fitness correlations. I found no evidence of inbreeding through estimation of genetic diversity or heterozygosity fitness correlations. Effective population sizes differed by plot but were limited in their applicability beyond the scale of sampling. I estimated an effective population size of 1,048 individuals across the western Imperial Valley. I then synthesized the results of these chapters into recommendations for ongoing management of the species. I recommend the integration of genetic monitoring into ongoing monitoring efforts including decadal reassessments of genetic diversity and effective population sizes, with additional recommendations to assemble a high-quality P. mcallii reference genome, to research genetic variation near range edges, to research climate change adaptation, and to investigate hybrid zone dynamics with P. platyrhinos and P. goodei.

Methods for structure determination in single particle cryogenic electron microscopy upended imaging sciences by demonstrating that crystallization is not necessary to resolve biological molecules at the atomic level. As imaging techniques have matured, so has the mathematical theory used to analyze these data. In particular, the method of moments and algebraic techniques show promise for the challenges posed by a low signal-to-noise ratio, the unknown orientation of particles in the imaging plane, and different classes of particles in a mixture. Cryo-EM remains a challenging problem approached from this direction and has motivated the study of simpler models capturing its core challenges. In particular, multireference alignment (MRA) is a well-studied model in which one observes copies of a signal corrupted by unknown cyclic shifts and isotropic Gaussian noise. In this work, I study two variants of multireference alignment, one introducing a step modeling tomographic projection (projected MRA) and another allowing reflections of the signal as well as cyclic shifts to appear (dihedral MRA). For signals of length a power of two, I develop an efficient recursive method for recovering signals from the third moment tensor of the samples. Beyond these specific orbit recovery problems, a subroutine using in my algorithm solves a linear system with a solution constrained to be a symmetric rank-1 matrix. Such problems arise in algebraic compressive sensing. Even when a given linear system is under-determined, the rank-1 constraint provides a ``sparsity'' assumption which may ensure a unique solution. To prove that my algorithm succeeds, I develop a construction which can be used to certify that specific linear systems with a rank-1 constraint have unique solutions. We show the algorithm succeeds in recovering a generic signal on the condition that a particular symbolic matrix has full column-rank. This condition can be easily checked on the computer in polynomial time, and we give strong evidence the matrices have full column-rank for a couple of dimensions. The foray into algebraic compressive sensing also reveals a close connection to the problem of finding a set of rank-1 matrices ``planted'' in a subspace of matrices. An algorithm of Johnston, Lovitz and Vijayaraghavan for finding the intersection of a linear subspace with a conic variety can recover such rank-1 matrices as a particular application. However, the upper bound on generic subspace dimension for which these authors prove their algorithm will succeed proves pessimistic in this specific case. We conjecture that a simple dimension counting bound gives the sharp threshold for the maximum dimension of generic subspace from which the JLV intersection algorithm can successfully recover a set of planted rank-1 matrices. To give evidence for this conjecture, I implemented a method of computer-proof which certifies the sharp boundary in several thousand specific cases. We give an analogous conjecture for the variety of symmetric rank-1 matrices and similar computer-proof of a similar sharp boundary in this case. Moreover, I give a floating point implementation of the JLV intersection algorithm in Python which can compute the intersection of any linear subspace with any Segre-Veronese variety (the varieties of rank-1 tensor with any set of partial symmetries and of any shape). Using this implementation, I give further evidence for our conjecture, showing that the algorithm does indeed recover planted matrices beyond the bound proven in the original paper up to our newly conjectured threshold. As an immediate consequence, we more than double the maximum rank of generic third- and fourth-order tensors (with or without partial symmetry) for which the JLV intersection algorithm will give a provably correct decomposition.

Coastal carbon cycling is highly dynamic and varies across spatial and temporal scales because of interacting physical and biogeochemical processes. However, direct measurements of the carbon balance of coastal ecosystems remain limited, and their contributions to regional and global carbon budgets are still uncertain. This dissertation investigated carbon cycling in the nearshore coastal ocean and San Diego Bay using measurements of air-sea carbon dioxide (CO2) flux and carbon-related water-quality indicators. Air-sea CO2 flux was measured using eddy covariance methodology with a stationary pier-mounted tower and a mobile boat-based tower.Observations collected at the Ellen Browning Scripps Memorial Pier (hereafter Scripps Pier) from 2014-2025 indicated predominant net CO2 uptake, with a median annual flux of -85.3 mg C m-2 hr-1 (95% CI: -121.3 to -28.6). CO2 flux differed significantly among seasons, with mean uptake in spring and fall 28% greater than in summer and winter; however, the seasonal effect was small, and substantial variability occurred within seasons. No consistent diel pattern was detected, with CO2 uptake occurring throughout both day and night. No single environmental variable consistently explained CO2 flux. Instead, associations with wind speed, chlorophyll, sea surface temperature, photosynthetically active radiation (PAR), and salinity were generally weak or dependent on the temporal scale of analysis, indicating complex, scale-dependent controls on air-sea exchange. Fixed-interval median analyses reduced the influence of uneven sampling density and strengthened several apparent environmental relationships, although the resulting associations remained scale-dependent. Spatial variability in air-sea CO2 flux was evaluated using a mobile boat-based eddy covariance tower deployed throughout San Diego Bay from 2023-2025. Mean CO2 flux across the bay was -30 mg C m-2 hr-1 (95% CI: -89 to 29), indicating weak sink tendencies but no significant difference from neutral conditions during the study period. No overall difference was detected between North and South Bay when all seasons were combined because regional differences depended on season. South Bay had significantly more negative fluxes than North Bay during summer and fall, whereas no significant regional differences occurred during winter or spring. Accordingly, CO2 flux was significantly influenced by the interaction between season and bay region. Exploratory analyses identified significant associations of CO2 flux with PAR and turbidity across the bay and with pH in South Bay. These results demonstrate that whole-system averages can obscure localized and season-dependent patterns in estuarine CO2 exchange.Water quality indicators–including temperature, PAR, colored dissolved organic matter (CDOM), chlorophyll a (blue and red), pH, specific conductivity, turbidity, and dissolved oxygen–were also evaluated throughout San Diego Bay using a Manta+35 multiparameter sonde. Multivariate water-quality profiles did not differ significantly between North and South Bay, indicating that the bay was structured by continuous spatial gradients rather than distinct regional divisions. Seasonal variability was the dominant source of multivariate variation, with CDOM, temperature, and dissolved oxygen contributing most strongly. In contrast, individual variables exhibited significant spatial differences, with the strongest regional effects occurring for pH, CDOM, and temperature. Overall, water-quality conditions were generally comparable to those reported for other temperate estuaries, despite clear gradients in organic matter and indicators related to productivity. Collectively, these studies demonstrate that resolving the role of dynamic coastal ecosystems in carbon cycling requires long-term observations together with spatially and seasonally resolved measurements of air-sea CO2 exchange and associated environmental variables.

Sri Lanka’s 2021 ban on agrochemicals – and subsequent country-wide rice yield loss – offers a rare opportunity to study the components of yield resiliency during system-wide transitions away from synthetic agricultural inputs. We combine a six-year Sentinel-2 satellite based remote sensing analysis with semi-structured farmer interviews to assess heterogeneous yield impacts, recovery rates and post-ban adaptation strategies among rice farms in Sri Lanka’s Anuradhapura District. Using Normalized Difference Red Edge (NDRE) and a panel of 100,000 pixels, we developed three tools for measuring rice farm productivity changes before, during and after the ban. In parallel, we analyzed farmer interview responses through partitioning-around-medoids cluster analysis to identify farm typologies linked to farm attributes, reported yield loss, and yield recovery times. Both methods found that farms with the lowest pre-ban productivity experienced the smallest relative yield losses and fastest recoveries. The two methods identified different categories of most impacted farmers: our remote sensing analysis pointed to the highest productivity farmers experiencing the largest yield loss and largest recovery, while our cluster analysis identified these impacts in the middle-resource farmer cluster. Interview data further indicated that crop rotation, access to large-scale irrigation resources, proximity to government extension, previous usage of organic farming practices, and participation in post-ban trainings were associated with greater yield resiliency. Our findings highlight the power of pairing land-scape level satellite remote sensing monitoring with targeted, resource-sensitive farmer interventions for supporting any future large-scale transitions away from agrochemicals.