http://repository.arizona.edu:80 The DSpace digital repository system captures, stores, indexes, preserves, and distributes digital research material. Wed, 20 May 2026 02:21:23 GMT 2026-05-20T02:21:23Z http://hdl.handle.net/10150/680195 Climate’s Influence on the Light Reactions of Photosynthesis: Missing Representations in the Climate-Carbon Cycle Feedback Neri, Patrick The emission, transfer, and absorption of carbon across the ocean, atmosphere, and land is core in addressing the ongoing climate crisis. Being able to better inform and constrain carbon cycle modeling under extreme and varied conditions is a critical goal to prepare for climate change in the coming century. At the core of that cycle is the assimilation of carbon by terrestrial plants, a dynamic process dictated by the physiological, phenological, and morphological dynamics of diverse plant species and their interaction with the spatiotemporal environmental change. Estimating and projecting carbon assimilation of terrestrial plants under current and future climate conditions is a complex challenge. A core issue lies in the lack of systematic understanding and quantification of how the environment regulates physiological processes, particularly photosynthetic light reactions. This critical process, which determines the plant’s photosynthetic carbon assimilation capacity and efficiency under various environmental stresses, has been largely overlooked in long-term carbon cycle modeling efforts. Capturing the environmental influence on photosynthetic light reactions would increase the physical accuracy of photosynthesis models and provide a precursory foundation to assess the impacts of bioengineering efforts (e.g., genetically increased light use efficiency for photosynthetic carbon assimilation) on climate-carbon feedback. This dissertation seeks to advance photosynthesis modeling by demonstrating a novel modeling approach that incorporates environmental constraints on the light reactions of photosynthesis to improve the prediction accuracy of carbon assimilation by terrestrial plants. It also provides additional efforts in the development of machine-learning (ML)-based surrogate model systems specifically designed to tackle the challenges of simultaneously capturing PFT-specific photosynthetic carbon assimilation and leaf area index at a location with heterogeneous plant cover and canopy structure. Based on this approach, a comparison of how the addition of a photosynthetic light reaction can affect the prediction of gross primary productivity (GPP) across sites and climates was performed. Moreover, an expanded parameterization of environmental constraints on electron transport rate for photochemical carbon assimilation was implemented using ML-based symbolic regression, demonstrating its ability to systematically capture light reactions of photosynthesis. This dissertation includes five chapters. A review of photosynthesis modeling and several recent efforts in applying surrogate systems to more complex earth system models is given (Chapter 1). A parameterization of the temperature dependence on the maximum quantum yield of PSII is generated (Chapter 2). A PFT-specific surrogate model that replicates modeled daily GPP and LAI of the Community Land Model 5.0 (CLM5.0) across several sites was developed and used to optimize photosynthetic and phenology parameters compared to observations (Chapter 3). The implication of the PFT-specific parameterization for advancing environmental controls on photosynthesis modeling in CLM5.0 optimized using the results of Chapter 3 is assessed (Chapter 4). Further expansion on modeling of environmental control on light reactions, specifically the fraction of available reaction centers and non-photochemical quenching, was generated using symbolic regression with respect to temperature and light. These results were joined to the results of Chapter 2 to directly model the electron transport rate of photosynthesis (Chapter 5). This work outlines future use cases for this framework of light reaction-informed photosynthesis modeling and the next steps to improve physical interpretability. By demonstrating the power of this framework and its implementation and implications, this thesis offers a meaningful contribution to the field of earth system modeling, atmosphere-biosphere interactions, and photosynthesis modeling. By providing valuable and flexible methods and implementation of environmental controls on the light reactions of photosynthesis.  Thu, 01 Jan 2026 00:00:00 GMT http://hdl.handle.net/10150/680195 2026-01-01T00:00:00Z http://hdl.handle.net/10150/680194 Contesting Orthodoxy: The Kong Lineage’s Scholarship and Classical Studies During the Han and Wei-Jin Periods Wong, Kai Sum From the Han (202 BCE-220) through the Wei-Jin (220-420) periods, Confucius’s descendants occupied a distinctive position at the intersection of imperial privilege and cultural authority, yet they have received remarkably little academic attention. This dissertation places the Kong family at the center of inquiry, reassessing their role in the development of classical learning and exegetical traditions during this period. In the body chapters, through an examination of their political, social, and cultural engagements, I argue that an intellectual anxiety arose from the disparity between their genealogical prestige and their limited production of classical scholarship. In light of this, I revisit two key texts, the Kongzi jiayu and Kongcongzi. Moving beyond long-standing debates over authenticity, the study reframes them as compilations that functioned as strategic interventions in classical studies. I demonstrate how these compilations appropriated existing commentarial resources to become the sage’s cultural mantle, while simultaneously constructing and legitimizing the Kong family’s transmission of ritual and ancient-script classical traditions. The final chapter further examines the Shengzheng lun, in which these texts were first extensively mobilized by Wang Su in his debates with Zheng Xuan. I show how texts attributed to the sage’s lineage could be deployed as a form of hermeneutic capital, lending authority to interpretations of the classics. By tracing the production, redaction, and citation of the Kongzi jiayu and Kongcongzi, this study uncovers the long-obscured voice of the Kongs as a lineage-based scholarly tradition. In doing so, it introduces genealogy as a new analytical lens for understanding the development of classical learning. It also foregrounds a contesting claim to orthodoxy within this tradition. Tue, 01 Jan 2019 00:00:00 GMT http://hdl.handle.net/10150/680194 2019-01-01T00:00:00Z http://hdl.handle.net/10150/680190 Improving Comparability in Network Meta-Analysis: Dose, Heterogeneity, and Mixed Treatments in Depression Research Herder, Katherine Elise Background: When treatments differ in dose or intensity, network meta-analysis (NMA) faces a tradeoff: lumping doses together obscures a critical effect modifier, while splitting them fragments the evidence base and sacrifices power. Model-based NMA (MBNMA) resolves this by directly modeling dose-response relationships, yet its assumptions remain underscrutinized in real-world networks. Using depression treatment research, a domain with diverse dose definitions, complex network topology, and a rich evidence base, this dissertation systematically evaluates MBNMA’s assumptions, capabilities, and limitations. Methods: Four complementary analyses were conducted: (1) an empirical review of dose-aware depression NMAs; (2) a reanalysis of a 45-study exercise-depression network evaluating four MBNMA functional forms against lumping and splitting benchmarks across two dose operationalizations; (3) simulation and empirical analyses of single-class (SSRI, 60 trials) and multi-class (GRISELDA, 133 trials) antidepressant networks examining pooling assumptions, functional form misspecification, and model selection performance; and (4) exploratory boundary-condition simulations varying heterogeneity, evidence density, and connectivity to identify when treatment rankings become most reliable. Results: MBNMA expanded inference for nonpharmacologic intervention, enabling estimation at unobserved doses and extraction of minimum effective doses, but offered no meaningful gains in model fit or heterogeneity reduction over simpler approaches. In pharmacologic networks, dose-response models achieved moderate but consistent ranking accuracy regardless of specification, while the lumped NMA introduced systematic pairwise reversals driven by dose-averaging, showing that accounting for dose, even imperfectly, improves comparisons. The default shared ED50 assumption produced modest within-class bias but structural distortion across classes; class-effect specifications partially addressed this but with convergence limitations. Boundary-condition simulations identified evidence density as the most effective lever for improving ranking accuracy, and demonstrated that the lumped NMA’s bias worsened as evidence accumulated,the clearest illustration that misspecified models do not self-correct with more data. Conclusions: MBNMA is a valuable but assumption-sensitive tool whose reliability depends on rarely examined modeling choices. Dose-response models consistently outperform lumping, but default pooling specifications can silently distort estimates in multi-class networks, and standard model selection criteria are insufficient safeguards. Sparse networks impose structural ceilings on ranking accuracy that model complexity cannot overcome. This dissertation delivers concrete decision frameworks, practical recommendations, and reporting standards to help researchers deploy dose-response NMA with the rigor these high-stakes clinical comparisons demand. Thu, 01 Jan 2026 00:00:00 GMT http://hdl.handle.net/10150/680190 2026-01-01T00:00:00Z http://hdl.handle.net/10150/680188 Birds of a Feather: Cultural and Ecological Relationships between the Unangax̂ and Seabirds in the Aleutian Islands, AK LaZar, Miranda This dissertation investigates how marine ecosystems and ancestral Unangax̂ communities in the Aleutian Islands of Alaska adapted to Holocene climate variability. Today, rapid warming in the Arctic/Subarctic is causing disruptions in marine ecosystems, threatening food webs and the economic and cultural livelihoods of northern Indigenous communities. Identifying the ways ecosystems and people have adapted (or failed to adapt) to long-term climate change in the past can help us anticipate and plan for future environmental disturbances. Zooarchaeological data sets can contribute to this effort by evaluating past ecosystem resilience, establishing ecological baselines, and documenting Indigenous stewardship in the past. I integrate oral/ethnohistories with analyses of archaeological seabird bones from Sanak and Agattu Islands to track changes in local marine ecosystems and ancestral Unangax̂–seabird relations. Seabirds are important sentient beings in Unangax̂ ontologies and were used in ancestral times for everyday parkas, magical guises, and bone tools. Ecologically, seabirds are important bellwethers in marine ecosystems, responding quickly to changes in prey and ocean conditions. I show that environmental disturbances and ancestral Unangax̂ bird hunting practices varied across the archipelago, prompting ancestral Unangax̂ in the eastern and western Aleutians to hunt different species of birds for raw materials and food. Results from my bulk stable isotope analyses suggest that seabirds stayed relatively stable in marine food webs over the last ~4,500 years. Noting record murre (Uria spp.) die-offs after the 2014-16 marine heatwave, my use of a novel compound-specific stable isotope application establishes ecological baselines that underscore their sensitivity to shifts in marine fish productivity. Taken together, this dissertation provides evidence that Subarctic marine ecosystems and ancestral Unangax̂ communities in the Aleutian Archipelago were quite resilient to long-term climate change throughout the Holocene. Thu, 01 Jan 2026 00:00:00 GMT http://hdl.handle.net/10150/680188 2026-01-01T00:00:00Z