The research arrives at a precarious moment for the American nursing workforce. In 2021, the number of nurses under 35 in the workforce plummeted by more than 100,000 in a single year, a decline significantly higher than any seen in the previous four decades, although updated forecasts suggest the profession is recovering following the COVID-19 pandemic. The National Sample Survey of Registered Nurses projects a national deficit of 78,610 full-time equivalent registered nurses in 2025, easing slightly to 63,720 by 2030. Globally, shortages fueled by an aging, retiring workforce and persistent retention problems compound the crisis. And while the shortage is nationwide, it bites hardest in rural America, where approximately 46 million people live with disproportionately poor access to care.

The health consequences of that scarcity are stark. Rural residents are more likely than urban dwellers to die prematurely from the five leading causes of death: cardiovascular disease, cancer, unintentional injury, chronic lower respiratory disease, and stroke. Long distances to facilities, inadequate transportation, low insurance rates, limited health literacy, and chronic provider shortages all conspire to worsen outcomes. Between 2012 and 2019, rural populations recorded a 45 percent higher mortality rate from opioid overdose compared with urban areas. The pandemic deepened these divides, bringing lower vaccination rates and higher hospitalization and death rates to rural communities. For farmworkers, the picture is even grimmer: the average life expectancy of a United States farmworker is just 49 years, and chronic disease risk factors such as overweight and obesity, affecting 81 percent of surveyed California agricultural workers, exceed those of the general population.

Against this backdrop, researchers turned to a rural community college nursing program in a region federally designated as both a Medically Underserved Area and a Health Professional Shortage Area, where fewer than 10 percent of residents hold a bachelor’s degree compared with 23 percent nationally. In April 2022, the team invited 87 students from the first- and second-year cohorts of the associate degree registered nursing program to complete a cross-sectional survey, developed through multiple stages of pilot testing and built on questions from validated national surveys. The response was remarkable: 82 students returned surveys, a 91 percent response rate. The study was guided by an intersectionality framework examining how race, gender, and class hierarchies shape experiences within the nursing profession, and a single open-ended question invited students to describe what being a first-generation college nursing student meant to them.

The findings were striking. Approximately 70 percent of participants were first-generation college students, defined under the Higher Education Act as individuals whose parents did not complete a baccalaureate degree. The cohort was overwhelmingly female at 81.7 percent, ethnically diverse, and multilingual: half identified as Hispanic or Latino, one-quarter were fluent in Spanish, and others spoke Tagalog, Cebuano, and other languages. The mean age was 31.54 years, with no significant difference between first-generation and non-first-generation students. Most students were enrolled in a two-year Associate Degree in Nursing program, a pathway researchers note may be more feasible for first-generation students because of its affordability, shorter completion time, and flexibility for those with family caregiving responsibilities.

The statistical comparisons revealed a pattern that defies some conventional assumptions about first-generation disadvantage. While roughly 73 percent of participants were unsure of receiving mentoring support or received none during the admissions process, and while barriers such as the application process, tuition costs, accessing counselors, lack of information, and school distance were common, nearly all of these burdens were shared equally across both groups. Only one dimension separated them: first-generation college students were far more likely to report that high school guidance counselors had questioned their future potential, with 94.4 percent of first-generation students reporting this experience compared with just 5.6 percent of their peers, a statistically significant difference. Even more surprising, first-generation students carried significantly less student loan debt, averaging $4,192 compared with $21,404 for non-first-generation students, though the authors caution that the study did not capture the proportions of student income, living expenses, and tuition that might explain this gap.

The financial portrait of these students is sobering nonetheless. Nearly half reported that their families depended on them as a source of financial support while attending school, and 75.6 percent said their families relied on them for emotional support. Students funded their education through federal or state loans, grants, scholarships, savings, credit cards, and income from work, with nearly a third working part-time and others full-time or on-call. A substantial 31.6 percent reported financial concerns serious enough that they considered dropping out of college. Meanwhile, more than a third struggled to access health care services themselves, and comparable numbers lacked adequate study space, educational materials, transportation, or childcare. Experiences of bias or discrimination related to race, ethnicity, and national origin were reported about 7 percent of the time.

Yet the qualitative responses revealed a powerful undercurrent of purpose. Content analysis of the open-ended question identified five themes: breaking the cycle, financial challenges, empowerment and pride, pressure and uncertainty, and gratitude and motivation. One student captured the duality of the experience: being a first-generation college student is empowering and scary, empowering because she is able to break the chain and scary because there are many unknowns about the process. Others described breaking barriers and stigma, setting an example for younger siblings, and pursuing opportunities their parents never had. Many spoke of honoring parental sacrifice, with one student writing that accomplishing a college degree means the whole world to her, and another hoping to motivate her son, nieces, and nephews to continue their education.

Perhaps the most consequential finding for workforce policy is the students’ commitment to the communities that produced them. A significantly greater proportion of first-generation students grew up in rural or designated low-income areas, at 88 percent compared with 12 percent of non-first-generation peers. And while 40.2 percent of all respondents planned to work in under-resourced communities after finishing their schooling, first-generation students were dramatically more likely to express that intention, comprising 87.9 percent of those affirmatively interested versus 12.1 percent of non-first-generation students. The students also expressed overwhelming interest in continuing their education beyond the associate degree: 89.2 percent planned to complete a Bachelor of Science in Nursing within one to three years, 59.1 percent intended to pursue master’s-level education, and 31.5 percent showed interest in doctoral study.

The implications ripple outward from this single rural campus. Research consistently shows that culturally, socially, and racially concordant care, in which providers mirror the patients they serve, produces significantly better health outcomes and patient satisfaction. Yet only about 10 percent of registered nurses in the United States are Latinx, even as Latinos constitute California’s largest ethnic group at 15 million people and the state holds the most diverse older population in the nation. First-generation college healthcare providers may bring a unique perspective on the experiences of patients from similar backgrounds, strengthening patient-provider relationships and outcomes. The authors acknowledge limitations, including the cross-sectional design, the single rural site, and the focus on community college students, which may limit generalizability to baccalaureate and graduate programs or to first-generation students in other regions. Still, their conclusion is clear: recruiting and retaining first-generation college nurses represents a tangible, evidence-supported strategy for tackling nursing workforce shortages in underserved rural areas, one that requires concerted effort from policymakers, healthcare providers, and community members alike. For the students themselves, the mission is already personal. They are not merely filling workforce gaps; they are breaking a generational cycle, one degree at a time, in the communities that need them most.

Subject of Research: Experiences and workforce potential of first-generation college nursing students at a rural community college

Article Title: Breaking the Cycle: First‐Generation College Nursing Students Attending a Rural Community College

Article References: Arriaga, A. T., Castillo, E., Chan, G. K., & Wagner, L. M. (2026). Breaking the Cycle: First‐Generation College Nursing Students Attending a Rural Community College. Nursing Open, 13(10), Article e70878. https://doi.org/10.1002/nop2.70878

Image Credits: AI Generated

DOI: 10.1002/nop2.70878

Keywords: nursing workforce, first-generation college students, rural health, community college, health disparities, nursing education, farmworker health, California, registered nurses, health professional shortage areas, student loan debt, workforce diversity

The grasshoppers in question belong to the genus Trimerotropis, a speciose group of band-winged grasshoppers found throughout the American continent, from Canada down to Chile and Argentina, with their greatest abundance in desert and semi-desert regions. The genus currently comprises 49 recognized species, yet its members are notoriously difficult to tell apart morphologically. Diagnostic features include a pronotal crest always cut by two sulci, forewing bands crossing the dorsum in species that do not inhabit sandy terrain, relatively large eyes, hindwings usually marked with a clearly defined outer dark band, and a basal hindwing area that is yellow, yellow-green, or blue but never orange or red. Despite this ecological and taxonomic prominence, published records of parasitoids attacking Trimerotropis have been astonishingly scarce.

Until now, only a single confirmed parasitoid record existed for the entire genus: the flesh fly Blaesoxipha (Acridiophaga) caridei, reared from Trimerotropis ochraceipennis in Chile back in 1943. No parasitoid wasps have ever been documented in Trimerotropis species, although the grasshoppers do fall prey to sand wasps of the genus Tachysphex, which paralyze them as provisions for their nests. Some kleptoparasitic flies in the subfamily Miltogramminae have also been reported from paralyzed Trimerotropis individuals in sand wasp nests in the United States. Against this sparse backdrop, the new findings represent a significant expansion of knowledge.

The story began somewhat accidentally. Specimens of two Trimerotropis species were collected in 2023 as part of a research project investigating the evolutionary, environmental, and biotic mechanisms driving niche diversification in Oedipodinae grasshoppers, conducted by Dr. Tereza Jezkova’s team. Trimerotropis pallidipennis individuals were collected near the Mono Lake Navy Beach viewpoint in California on 16 August 2023, while T. melanoptera were gathered near the National Scouting Museum in New Mexico on 16 September 2023. The grasshoppers were captured with entomological nets and euthanized in the field with acetone-charged kill jars, since the project aimed to obtain material for genetic analyses.

What happened next transformed a routine collecting trip into a parasitological discovery. One femur from each grasshopper was removed and preserved in 100% ethanol for DNA work, and the remaining bodies were eviscerated through a small incision in the cervical membrane before being placed in newspaper envelopes lined with cotton. Approximately one week after fieldwork, when the specimens were being prepared for drying in an oven at 50 degrees Celsius, the researchers discovered fly puparia inside the entomological envelopes. Larvae had emerged from the grasshoppers after collection but before the drying process, and remarkably, they had survived despite their hosts having been submerged in acetone and subsequently eviscerated. The envelopes containing the puparia were then sealed in zip lock bags to allow adult fly development.

A single puparium was found among the T. pallidipennis specimens, from which an adult male fly successfully emerged. Two puparia were recovered from the T. melanoptera envelopes, though only one adult male emerged from those. The exact timing of larval emergence and adult fly development could not be determined because the puparia were found in envelopes, with separate envelopes used for each grasshopper species. The adult flies were preserved in 70% ethanol and sent to the Laboratório de Diptera of the Museu de Zoologia da Universidade de São Paulo in Brazil for identification, where specialists dissected the male terminalia in potassium hydroxide solution and compared the specimens against the holotype male of Sarcophaga angustifrons Aldrich, 1916, held at the National Museum of Natural History in Washington, DC.

The identification confirmed the flies as Blaesoxipha (Acridiophaga) angustifrons, a Nearctic species widely distributed from the southern United States to Canada. The genus Blaesoxipha, which comprises more than 240 species in 10 subgenera worldwide, is characterized by a posteromedian row of spines on the trochanter of the hind leg in both sexes, along with distinctive male terminalia features including a cercus bent posteriorly with short erect spines on its distal dorsal surface. Species in the subgenus Acridiophaga are distinguished by a setose proepisternum and male terminalia with the apical part of the juxtal lobes curving laterally or outwards. Externally, B. angustifrons closely resembles B. caridei, and the two were even considered synonymous in earlier works, but they can be separated by the length of the median stylus in the male terminalia, which is completely covered by the juxtal lobe in lateral view in B. angustifrons.

The host records themselves are scientifically valuable. Trimerotropis pallidipennis is the most widely distributed Oedipodinae grasshopper in the New World, ranging from southwestern Canada to at least central Argentina, and inhabits dry grasslands including western prairies, desert grasslands, and desert scrub. Trimerotropis melanoptera, recognizable by a very broad black band on the hindwings, a reddish abdomen, and a single pale band on the inner face of the hind femora, occupies open grasslands and desert grasslands. Beyond these two new hosts, confirmed host records for B. angustifrons are scarce and sometimes unclear. The Nearctic grasshopper Conozoa carinata was reported as a host by Pape in 1994, though without a specified locality. Other published associations, including probable parasitism of several Melanoplus species and Camnula pellucida in Canada, rest on questionable or unpublished data and remain pending revision.

The biology of these parasitoid flies is as striking as their taxonomy. Blaesoxipha females rely on visual cues when hunting, darting at flying or jumping grasshoppers or stalking them on the ground. They typically deposit a single, well-developed first-instar larva on or in the host, either on the ventral wing surface or injected between the abdominal intersegmental membranes. These flies have been recorded from many grasshopper groups and are considered important agents for biological control of pest grasshoppers. Yet attempts to use other sarcophagid parasitoids against Nearctic grasshoppers have met with mixed success; tests involving four Palaearctic species on numerous North American grasshoppers, including Spharagemon campestris, formerly classified as Trimerotropis, proved unsuccessful.

The practical stakes of this gap in knowledge are considerable. Some Trimerotropis species, including T. pallidipennis, can significantly impact crops during outbreaks by feeding on a wide variety of plants. Understanding which natural enemies attack these grasshoppers, and under what circumstances, could prove essential for managing outbreaks and appreciating the ecological checks that keep grasshopper populations in balance. The new records from California and New Mexico significantly expand the known host range of B. angustifrons and contribute to the limited understanding of host-parasitoid dynamics in this ecologically important genus. They also serve as a reminder that even routine fieldwork, conducted for entirely different purposes such as population genetics, can yield unexpected discoveries when researchers pay close attention to what emerges from their collecting envelopes.

Subject of Research: Parasitoid flesh fly host associations with Nearctic Trimerotropis grasshoppers

Article Title: New records of parasitoid flies ( Diptera , Sarcophagidae ) in Nearctic species of Trimerotropis Stål ( Orthoptera , Acrididae , Oedipodinae , Trimerotropini )

Article References: Gudin, F. M., Martins Silva, D. S., Otto, A., Bracken, J., & Jeskova, T. (2026). New records of parasitoid flies (Diptera, Sarcophagidae) in Nearctic species of Trimerotropis Stål (Orthoptera, Acrididae, Oedipodinae, Trimerotropini). Journal of Orthoptera Research, 35(1), 197-202. https://doi.org/10.3897/jor.35.175567

Image Credits: AI Generated

DOI: 10.3897/jor.35.175567

Keywords: Trimerotropis, Blaesoxipha, Sarcophagidae, parasitoid, grasshoppers, Acrididae, Oedipodinae, host-parasitoid, New Mexico, California, biological control, entomology

The research, conducted by Patrick Wickham, Timothy Bayley, and Cameron Tana of Montgomery and Associates Water Resource Consulting, was carried out to support the Santa Cruz Mid-County Regional Water Optimization Study. Rather than replacing the physics-based numerical model with a faster machine learning surrogate trained on a preexisting library of simulations, the team developed what they call Machine Learning Guided Optimization, or MLGO. The workflow uses machine learning algorithms and automated pre- and post-processing scripts to iteratively run a coupled groundwater and surface-water flow model, known as GSFLOW, building its own training compendium of scenarios as it progresses toward optimization. This self-building surrogate approach combines the speed of machine learning with the robustness of physical model predictions, without requiring a preexisting compendium of groundwater model runs or large observational datasets.

The stakes are high because the basin’s aquifer system, dominated by the Purisima Formation and overlain in the southeast by the Aromas Red Sands, is hydrologically connected to the Pacific Ocean. When pumping exceeds natural and artificial recharge, seawater intrudes into the freshwater aquifers. The two largest water agencies in the basin have contrasting supply portfolios: the Soquel Creek Water District relies exclusively on groundwater, while the City of Santa Cruz primarily uses surface water and depends on its basin wellfield for roughly 5 percent of normal-year supply, far more during droughts. To restore balance, the City is planning an aquifer storage and recovery project that converts select wells into injection and extraction wells, storing surplus surface water in wet periods for use in droughts, while the District is implementing Pure Water Soquel, an indirect potable reuse project that stores highly purified recycled water in the basin to raise groundwater levels.

Planning these projects involves evaluating numerous implementation parameters: the number and location of new ASR wells, the number of purified water injection sites, the aquifer layers into which wells are screened, maximum pumping and injection volumes, preferred distribution percentages among wells, and the timing and direction of bi-directional interagency transfers. In total, the MLGO implementation used 118 individual project parameters. Each configuration had to be simulated 53 years into the future, from water year 2023 to 2075, under a climate change projection significantly drier and warmer than historical conditions. Because a single GSFLOW simulation can take hours or days to run, and adequate optimization requires many iterative simulations, the number of potential project configurations is practically endless and manual exploration alone cannot cover it.

The MLGO framework operates as a coupled, automated loop. It begins with a small initial compendium of five manually created scenarios, each described by numerical project parameters as inputs and project goals as outputs. Machine learning algorithms are then trained on this compendium to perform inverse modeling: given a set of user-defined desired project goals, the algorithms suggest project parameters expected to meet them. An automated post-processing script transforms those suggestions into physical model input files, a Python-based water allocation model logically allocates ASR injection, ASR extraction, purified water injection, District extraction, and operational pumping changes based on each scenario’s unique infrastructure and monthly climate timeseries, and the GSFLOW model runs the assembled simulation. Results are post-processed with respect to project goals, the new scenario is added to the compendium, and the algorithms retrain on the growing dataset. As the compendium expands, the algorithms improve their understanding of how projects affect sustainability outcomes and predict progressively better configurations.

For this study, MLGO was tasked with achieving 58 individual project goals clustered into three groups. Sustainability goals require simulated groundwater elevations to remain above minimum thresholds at representative monitoring points spread across the basin and multiple aquifers, evaluated monthly or as five-year moving averages, to prevent seawater intrusion, depletion of interconnected surface water, and chronic lowering of groundwater levels. Feasibility goals maintain daily simulated elevations below infeasible levels at injection wells and keep pumping realistic by targeting zero seepage face losses, which the MODFLOW Multi-Node Well package calculates when wells are overpumped. Supply goals, quantified through the water allocation model, track both cumulative unmet demand over the 50-year simulation and maximum monthly unmet demand. Many different machine learning architectures were deployed concurrently, grouped into three categories: neural networks built with PyTorch, random forest ensembles built with Scikit-learn, and combined workflows in which a neural network performs feature extraction before a random forest makes predictions. The loss function driving learning is a weighted mean squared error in the model’s prediction of project input parameters, plus an out-of-bounds penalty whose magnitude increases with the square of a prediction’s distance from possible ranges, using a penalty coefficient of 0.1.

Over months of operation, the MLGO algorithms generated 3,407 scenarios. Performance was assessed with a combined metric summing the normalized sustainability, feasibility, and supply goal groups, where lower numbers indicate better adherence to desired goals. The best-performing algorithm type displayed a clear ability to learn over time, evidenced by a reduction in the combined evaluation metric. Optimization was organized into successive Tracks, starting with the smallest allowable infrastructure expansion and ending with the largest parameter space. At the start of each track, an exploration phase typically produced worse results as MLGO tested previously unavailable options and generated data; following this, results converged toward better solutions, a pattern visible in the 50-run moving average. Notably, the most optimized scenarios were not necessarily produced at the end of a given track, and staff modelers continuously reviewed high-performing runs for strategic insights throughout the process.

The process was not without challenges, and one episode illustrates how the workflow adapts. During Track 2 through 4 optimization, MLGO began producing highly optimized scenarios incorporating large pumping volumes. Manual review uncovered that these scenarios were exploiting seepage face losses, features of GSFLOW that limit prescribed pumping when a well is overpumped, meaning the model was not reflecting the prescribed inputs. The scenarios appeared optimized because the groundwater model silently reduced pumping, an error in this context. The team modified the MLGO framework to record seepage face losses and target zero as a project goal, artificially added these losses to previously conducted runs in the compendium, and MLGO subsequently avoided them quite well. Later, a water rights change reducing dry-period surface water volumes available for ASR injection altered the water allocation model’s output and the numerical parameter space; no compendium modifications were made, and MLGO continued to perform well despite the form of error this introduced into its training data. The authors acknowledge that results inherit any uncertainty in the sole numerical flow model and the single climate realization and aquifer parameter realization used, and suggest future efforts could employ multiple realizations of each.

Principal component analysis of the high-dimensional dataset, which contains 58 project goals and 118 project parameters, revealed that the first five principal components explain over 60 percent of all variability in the compendium, with PC1 alone accounting for just over 40 percent. These components elucidated practical trade-offs: high volumes of District-to-City transfers originating from pumping near existing purified water injection wells correlate with a lower supply gap; reducing injection at existing Pure Water Soquel wells correlates with sustainability problems; new wells screened solely in one aquifer correlate with feasibility issues, supporting recommendations to screen all new wells across two aquifers; and infrastructure expansion follows patterns of adding paired ASR and purified water wells screened in the same aquifers. From the thousands of scenarios, four project configurations were selected representing highly optimized, progressively expensive options for improving regional water supply.

Perhaps the most striking outcome is a strategy that had not been envisioned by the modeling team: symbiotically using ASR and purified water recharge by injecting surface water or recycled water in the basin’s highlands, along the border between the City of Santa Cruz and Soquel Creek Water District, upland of ASR facilities. This allows more pumping by the City while reducing the need for transfers from the District, mutually benefiting both agencies and reducing the need for large hydraulic intertie upgrades. The authors emphasize that the machine learning agents are heuristic in nature, attempting to solve the entire optimization problem at once by the shortest path possible, which risks missing alternative solutions that more systematic perturbation software such as PEST might capture; this risk was reduced by manual scenarios in the initial training data and by multiple algorithms exploring the parameter space differently. They also note that periodic metalearning, adjusting algorithm architecture and hyperparameters as the dataset expands, was largely user-imposed in this study, and future work could automate it. More broadly, the approach demonstrated here is not specific to this basin and should transfer to any planning problem in which a slow physical model must be interrogated across a large configuration space.

Subject of Research: Machine learning guided optimization of water management options with a physical groundwater model for the Mid-County Basin optimization study

Article Title: Machine learning guided optimization of water management options with a physical groundwater model for the Mid-County Basin optimization study

Article References: Wickham, P., Bayley, T., & Tana, C. (2026). Machine learning guided optimization of water management options with a physical groundwater model for the Mid-County Basin optimization study. Frontiers in Water, 8, Article 1930007. https://doi.org/10.3389/frwa.2026.1930007

Image Credits: AI Generated

DOI: 10.3389/frwa.2026.1930007

Keywords: machine learning, groundwater, water management, optimization, seawater intrusion, sustainability, California, GSFLOW, Machine, learning, guided, water

The study, led by Celeste L. Schonberg of the University of Miami’s Rosenstiel School of Marine, Atmospheric, and Earth Science, together with Holli C. Eskelinen of Dolphins Plus, Kelley A. Winship of the National Marine Mammal Foundation, and Mark J. Xitco Jr. of the Navy’s Marine Mammal Program, exploited a piece of technology few laboratories possess. The Enclosure Video Enrichment, or EVE, system is a computerized cognitive testing platform installed at the Navy facility in San Diego. It allows marine mammals voluntarily participating in husbandry and research sessions to interact with on-screen tasks, generating standardized, automated data of the kind that has transformed primate cognition research but has remained scarce for marine mammals.

The task itself was a maze. Each trial presented the sea lion with a layout of barriers on the screen, and success required maneuvering a cursor-like element from a start position to a target while avoiding obstacles. Critically, the researchers designed the layouts so that successful performance demanded short-term forward planning: an animal that simply lunged at the target or followed the most immediate path would run into a trap. Avoiding those traps required the animal to anticipate the consequences of its movements at least a step or two ahead, the cognitive equivalent of checking a route before committing to it.

Seven adult male sea lions completed the full protocol. The experiment consisted of 200 testing trials built from 100 unique maze layouts, which were separated into four trial blocks and repeated randomly across eight testing sessions. This design allowed the team to distinguish between performance on novel layouts and performance on layouts the animals had already seen, a distinction that matters enormously when testing for planning. If an animal is genuinely planning, it should perform well on brand-new configurations; if it is memorizing, its advantage should appear only on repeated exposures.

The researchers scored not just success and failure but the pattern of responses, categorizing trials by layout type, including direct layouts, reversal layouts, away layouts, and layouts requiring an approach-and-return strategy. Response types included clean successes, failures in which the animal fell into a trap, and failures without a trap. This fine-grained scoring was intended to reveal cognitive flexibility, problem-solving strategies, and any evidence that animals adjusted their approach as difficulty increased. A supplementary video released with the paper shows the full range of responses, from a sea lion sailing through a direct layout to one blundering into a trap on a reversal layout.

The results were moderate rather than spectacular. Sea lions achieved moderate success rates across the task, showing that they could use the EVE system and complete maze trials at levels well above pure chance on many layouts. But the more telling result was what did not appear: improvement. Across repeated exposures to the same layouts, success rates did not climb, and response patterns did not shift in ways that would indicate the animals were learning from experience or refining their strategies. In a task where learning should be easy to detect, the absence of a learning curve is itself a finding.

That absence led the authors to conclude that there was limited evidence of planning behavior in this paradigm. If the sea lions were employing short-term forward planning, one would expect them to avoid traps more effectively over time and to perform better on novel layouts as they gained general familiarity with the task structure. Instead, performance stayed flat. The researchers interpret this as suggesting that the sea lions may have relied on strategies other than planning, such as responding to immediate perceptual cues, or that the task’s demands did not align with the species’ natural cognitive strengths.

This is where the study’s comparative framing becomes important. Planning behavior, the ability to act now in anticipation of future states, has been documented in primates, corvids, and a handful of other species, typically using paradigms adapted to each animal’s ecology. Computerized maze tasks have been a staple of primate cognition research precisely because they can be standardized, repeated, and automated. Extending that paradigm to marine mammals is technically ambitious, and the authors emphasize that their goal was as much methodological as cognitive: to expand computerized testing beyond primates and to generate data that can be compared across species. The sea lion result, on its face a negative finding, still delivers value by marking the boundary of where this paradigm succeeds.

The authors are careful about interpretation. A flat learning curve in a computerized maze does not prove that sea lions cannot plan; it shows that this particular task, in this particular context, did not elicit measurable planning or learning. Species-specific cognition is shaped by adaptations and environmental conditions, and a pinniped that navigates three-dimensional underwater environments, hunts cooperatively, and relies on vibrissae and hearing rather than vision may simply not engage with a two-dimensional visual maze the way a primate does. The sensory and ecological mismatch between the test and the test-taker is a central theme of the paper, which argues that comparative cognition research must account for how adaptations influence performance on human-designed tasks.

There is also a welfare and husbandry dimension. The EVE system was developed as video enrichment, and voluntary participation in cognitive testing is a form of enrichment in itself for animals in managed care. The fact that seven sea lions completed 200 trials across eight sessions demonstrates that the platform works, that the animals engage with it, and that it can be deployed repeatedly without coercion. For the Navy Marine Mammal Program, whose animals participate in research and operational tasks, tools that measure cognition while providing enrichment serve dual purposes. For the broader research community, the study establishes a working pipeline for computerized marine mammal testing that future studies can refine, perhaps with designs that better match pinniped sensory worlds.

For now, the headline result is a study in cognitive humility. Sea lions, animals celebrated for trainability and social intelligence in managed care, confronted a maze task engineered to reward forward planning and did not show the improvement or response patterns that planning would produce. Whether that reflects a genuine absence of planning capacity, a mismatch between the paradigm and the species, or motivational and contextual factors unique to this setting remains an open question, and one the authors frame explicitly as a matter of species-specific cognition shaped by evolutionary history. What the study does establish is that the question can now be asked with standardized, automated tools, and that the answers, positive or negative, will accumulate in a form that can be compared across species, laboratories, and lines of evolutionary descent.

Subject of Research: Investigating planning behavior in California sea lions (Zalophus californianus) using computerized MAZE testing

Article Title: Investigating planning behavior in California sea lions (Zalophus californianus) using computerized MAZE testing

Article References: Schonberg, C. L., Eskelinen, H. C., Winship, K. A., & Xitco, M. J., Jr. (2026). Investigating planning behavior in California sea lions (Zalophus californianus) using computerized MAZE testing. Animal Cognition. https://doi.org/10.1007/s10071-026-02115-8

Image Credits: AI Generated

DOI: 10.1007/s10071-026-02115-8

Keywords: Investigating, planning, behavior, California, lions, Zalophus, californianus, computerized, MAZE, testing, scientific research

The research team, led by Halley E. Froehlich of the University of California, Santa Barbara, together with colleagues at UC Santa Cruz, UC Berkeley, Cal Poly San Luis Obispo, California Sea Grant, and the Hog Island Oyster Company, introduces and formalizes a concept they call secondary extensive aquaculture, or SEA. The idea is deceptively simple: seaweeds such as Ulva, Gracilaria, and Pyropia naturally recruit and grow on infrastructure built for a different primary species, typically oysters or clams. Instead of treating that growth purely as a maintenance problem, farmers could cultivate, harvest, and sell it as a co-product. Because the seaweed arrives on its own, the approach requires no seeding, no hatcheries, and no additional gear beyond what the farm already deploys.

The authors are careful to argue that SEA genuinely qualifies as aquaculture under the definitions used by the Food and Agriculture Organization of the United Nations and the European Union, both of which require some intervention in the rearing process and ownership of the stock. The placement and maintenance of longlines and bags, they contend, constitute the intervention: the infrastructure provides artificial hard substrate and alters local water flow, structuring where and how much seaweed recruits compared with an unmanaged wild population. Ownership is satisfied because the recruiting seaweed spends its entire rearing period on privately leased and actively managed infrastructure. The researchers draw an analogy to traditional Hawaiian fishponds, or loko i’a, which have long been recognized as aquaculture despite relying on tidal recruitment of wild fish through sluice gates rather than active stocking.

To test whether the concept could survive contact with economic reality, the team built a scenario-based bioeconomic model coupling logistic seaweed growth and sustainable harvest with a simple profit function. The framework borrows from fisheries science: because extensive aquaculture resembles harvesting from the wild, the researchers applied a surplus-production model of the kind routinely used to estimate maximum sustainable yield. Biomass dynamics were simulated at a daily time step over a 20-year horizon, with harvest rates linked to a profit function accounting for market price, variable costs, labor inputs, and fixed operating costs. Oyster production was treated as a fixed baseline — roughly $100,000 in revenue and $80,000 in costs per hectare per year — so that seaweed profitability could be read as an incremental gain or loss from diversification. Crucially, the model assumes that the labor once spent clearing biofouling is redirected into harvesting it, rather than disappearing.

The model was parameterized with literature-derived ranges for seaweed growth rates and carrying capacities, and grounded in real-world data from a pilot harvest conducted in April 2025 on the roughly 65-hectare Hog Island Oyster farm in Tomales Bay, Northern California. During two harvest events, crews pulled Gracilaria, known commercially as ogo, from bottom bags and Pyropia, the nori used in sushi restaurants, from suspended gear and lines. The first harvest, performed by three people, produced about 22.7 kilograms of wet nori in 19 minutes; the second, with six people, yielded roughly 104 kilograms of total wet biomass. The material was processed for Hog Island’s restaurant operations, giving the team an unusually direct measure of what SEA actually costs in labor.

The simulations revealed a clear hierarchy of what matters. Profitability increased with farm size, growing season length, and the intrinsic growth rate of the seaweed, with viable scenarios typically requiring more than 50 to 100 hectares and more than 60 growing days under conservative market assumptions of $400 per ton of wet weight. Strikingly, the intrinsic growth rate exerted a stronger influence on profitability than carrying capacity: the maximum possible profit under the fastest growth scenario was more than 500 percent higher than under the highest carrying capacity. That finding has practical consequences, because it points to fast-growing fouling species — Ulva, Gracilaria, and Pyropia among them — as the most suitable SEA candidates, and it means a farmer’s first diagnostic step is simply identifying which algae are colonizing the farm. Under very slow growth and low carrying capacity, profits were negative regardless of farm size, showing that some ecological conditions preclude feasibility altogether.

Labor emerged as the decisive bottleneck. The pilot harvests required roughly 50 to 60 labor-hours per ton for harvesting and between roughly 140 and 200 labor-hours per ton once processing was included — figures that dwarf the approximately 10 labor-hours per ton needed to break even at California’s minimum wage of $16.90 per hour. Doubling the crew from three to six people cut labor hours by about 30 percent, but that was nowhere near enough. The team’s modeling of efficiency improvements showed that technological and knowledge-driven gains matter far more than adding workers: with an 80 percent reduction in labor-hours per ton, breakeven became achievable with a crew of roughly eight people, while a 50 percent improvement required around 15. Those improvement targets are conservative by historical standards — Maine’s seaweed sector documented a roughly 14-fold increase in labor efficiency between 2017 and 2022, equivalent to about a 41 percent annual reduction in labor-hours per unit harvested.

Despite the labor challenge, the yield numbers were encouraging. The pilot SEA harvests produced between roughly 0.66 and 1.10 tons per hectare per year, and the model’s estimated maximum sustainable yield came in at about 0.94 tons per hectare per year — figures that sit near the lower bound of global seaweed aquaculture production but remain within plausible commercial ranges. The pilot yields were two to three times higher than the single lowest yield documented in a global systematic review of dedicated seaweed producers, a Gracilaria monoculture in China. Under modal ecological conditions on a 65-hectare farm, or at a higher seaweed price of $1,200 per ton typical of restaurant-grade nori, SEA could account for 4 to 11 percent of total farm profit — enough to recoup the 5 to 10 percent losses that unharvested biofouling typically inflicts.

The authors are candid about the caveats. Many parameters rest on limited pilot observations and model assumptions; seaweed markets remain emerging and volatile; and the analysis treated SEA as a standalone revenue stream, leaving out indirect benefits such as improved primary crop production from more thorough fouling removal. Because SEA depends on natural recruitment rather than seeding, farmers have limited control over which species colonize their gear, meaning farm suitability rather than species selection is the relevant management lever — and species composition may shift over time, complicating market strategy. A biostimulant market, where marine soil additives sell for roughly $0.22 to $0.66 per kilogram, may be more resilient to that variation than a nori-only market, and the researchers suggest smaller-scale, local-regional supply chains may be the ideal pathway for SEA harvests.

What emerges is a concept that is ecologically feasible but economically constrained — a promising complementary strategy for diversifying revenue and enhancing circularity in aquaculture systems, contingent on sufficient scale, favorable biological productivity, labor efficiency gains, and supportive policy frameworks that clarify permitting for co-harvested species. For the small farms that dominate the U.S. seaweed landscape, the message is sobering but precise: simply removing the fouling seaweed is likely the better financial option, unless family or voluntary labor changes the calculus. For larger operations in productive waters, the nuisance growing on the oyster lines may be quietly waiting to become a second harvest — provided the industry can learn to pick it fast enough to make the arithmetic work.

Subject of Research: Bioeconomic viability of harvesting naturally recruiting seaweed as a secondary crop on shellfish aquaculture farms

Article Title: Defining and finding the profit for seaweed secondary extensive aquaculture

Article References: Froehlich, H. E., Campbell, R. S., Miller, K. A., Johnson, K. M., Fleener, G., & Kroeker, K. J. (2026). Defining and finding the profit for seaweed secondary extensive aquaculture. Frontiers in Aquaculture, 5, Article 1844314. https://doi.org/10.3389/faquc.2026.1844314

Image Credits: AI Generated

DOI: 10.3389/faquc.2026.1844314

Keywords: seaweed aquaculture, secondary extensive aquaculture, biofouling, oyster farming, bioeconomic modeling, maximum sustainable yield, Ulva, Gracilaria, Pyropia, labor efficiency, blue economy, California

The study, led by Aniket Gupta with co-authors Ali Behrangi, Mohammad Farmani, Patrick Broxton, and Guo-Yue Niu, focused on the Sacramento and San Joaquin River Basins, the two great watersheds that drain the Sierra Nevada and the southern Cascades. Mountain snowpack supplies roughly 70 to 80 percent of the annual runoff in the western United States, making accurate snow estimates a matter of water security rather than academic curiosity. Previous work had shown that the National Water Model, which runs the Noah-MP land surface model at 1-kilometer resolution, underestimates snow water equivalent — the amount of water held in the snowpack — by 20 to 30 percent compared with ground observations from the SNOTEL network. Even sophisticated global reanalyses such as ERA, MERRA-2, and GLDAS miss the mark, largely because their coarse grids blur the sharp elevation gradients that govern where snow falls and lingers.

The Arizona team suspected that the problem lay in two intertwined weaknesses: the precipitation forcing fed into the model and the representation of snow–vegetation interactions. Forests are not passive backdrops to snowpack. Canopies intercept falling snow, some of which sublimates straight back into the atmosphere or drips to the ground as throughfall. Canopies also cast shade that alters the energy budget of the snow beneath, reducing the solar radiation reaching the surface while enhancing longwave emission from the warmer vegetation. These biophysical feedbacks are notoriously difficult to capture at basin scale, and few studies had quantified their influence across an entire mountainous watershed.

To disentangle these effects, the researchers ran eight model experiments using Noah-MP version 5.0 over the two basins at 1-kilometer resolution for the years 2015 through 2019. They paired two forcing datasets — the Analysis of Record for Calibration (AORC), the operational forcing of the National Water Model, and the North American Land Data Assimilation System phase 2 (NLDAS-2), statistically downscaled from roughly 12 kilometers to 1 kilometer using the WRF-Hydro Meteorological Forcing Engine — with two vegetation schemes. One scheme prescribed monthly leaf area index (LAI) and green vegetation fraction from a lookup table, as the National Water Model does. The other employed Noah-MP’s dynamic vegetation module, which prognostically simulates LAI and vegetation cover from a coupled photosynthesis and phenology model, allowing leaves to grow and senesce in response to simulated carbon budgets, temperature, and water stress.

The downscaling step proved decisive. The Mountain Mapper algorithm blends PRISM monthly precipitation climatology with real-time hourly rain gauge observations, computing site-specific bias ratios that are interpolated to the model grid and multiplied by the NLDAS-2 hourly fields. Temperature was downscaled using the NCAR lapse rate, while radiation, pressure, and humidity were adjusted for elevation. The result: the downscaled NLDAS-2 data showed systematically higher precipitation and colder temperatures in the mountains than AORC, producing more snowfall. In the Sacramento Basin, AORC underestimated daily-mean precipitation by about 0.1 millimeters per day relative to PRISM, with the deficit concentrated at the highest elevations where snowfall is greatest. NLDAS-2, by contrast, aligned closely with PRISM between 1,000 and 3,000 meters and matched available gauge observations from the GHCN and California Department of Water Resources networks.

When the researchers compared simulated snow water equivalent against independent estimates from SNODAS, the University of Arizona snowpack dataset, and the operational National Water Model, the pattern was unmistakable. Simulations driven by downscaled NLDAS-2 produced up to 100 millimeters more peak SWE than AORC-driven runs, bringing modeled snow amounts into far better agreement with ground truth. AORC-driven simulations, mirroring the operational model, fell short of SNODAS under every vegetation configuration. Validation at 23 CDWR snow stations across both basins confirmed the advantage: at 11 stations in the Sacramento Basin and 12 in the San Joaquin, the model consistently performed better with NLDAS-2 forcing. At higher-elevation stations such as CAP, GNL, TUM, and HHM, the combination of downscaled NLDAS-2 and dynamic vegetation produced optimal snow amounts, whereas AORC-driven runs underestimated peak March SWE by roughly 500 millimeters.

But the vegetation experiments revealed something more subtle — and more surprising. The prescribed LAI from the lookup table turned out to be substantially higher than both the dynamically predicted LAI and satellite-derived LAI from MODIS, especially in winter and spring above 1,000 meters. That excess foliage cast too much shade over the snowpack. Counterintuitively, the strongly shaded runs accumulated excessive snow on the ground: by suppressing the episodic mid-winter melt events that punctuate accumulation in these subtropical mountains, the heavy canopy allowed snow to pile up unrealistically. The dynamic vegetation module, producing lower LAI closer to MODIS observations, let more shortwave radiation reach the snow surface, enhancing those episodic melts during the accumulation season and thinning the overall snowpack. Validation against in-situ snow data showed that the dynamic vegetation runs, when driven by downscaled NLDAS-2, performed best of all.

The energy-budget analysis explained why. Noah-MP computes the energy available for snowmelt as the sum of net shortwave and net longwave radiation minus latent and sensible heat fluxes. Across all experiments, net solar radiation dominated the balance. During the accumulation period from December through March, the higher prescribed LAI reduced the shortwave radiation reaching the snow surface through canopy shading, which depends on vegetation density and solar angle. During the melt season from April through July, however, the dynamic vegetation runs showed higher green vegetation fraction than the prescribed runs, so shading became more sensitive to cover fraction than to LAI as the sun climbed higher. The net effect: the shading influence of the canopy on snowmelt and sublimation proved stronger than the interception influence on ground snow beneath the canopy, meaning sparser canopies yield thinner, shorter-lived snowpacks overall.

The findings carry direct implications for operational hydrology. The National Water Model, which uses AORC forcing and prescribed vegetation, inherits both weaknesses identified here. The researchers note that the forthcoming NLDAS-3 dataset, dynamically downscaled to 1-kilometer resolution, could substantially improve snow simulations across the western United States and merits careful evaluation within the National Water Model framework. They also caution that even with dynamic vegetation, NLDAS-2-driven simulations still produced higher SWE than SNODAS, likely because the modeled LAI remains larger than MODIS-derived values — suggesting that further calibration of vegetation parameters, such as type-dependent specific leaf area, could push simulations even closer to observations.

Beyond the operational realm, the study underscores a broader lesson about subgrid heterogeneity in Earth system modeling. Grid-averaged vegetation parameterizations, the authors argue, can obscure critical nonlinear responses in snow accumulation and melt, altering diurnal energy partitioning, runoff generation, and seasonal snow persistence. Because LAI controls snow duration through shading, longwave enhancement, and interception, it also modulates surface albedo and land–atmosphere energy exchanges, potentially feeding back into regional climate. As climate change shifts precipitation patterns and forest structure across the West, capturing these fine-scale vegetation processes may prove essential not just for forecasting next spring’s water supply, but for building physically consistent coupled models of land, water, and atmosphere in snow-dominated basins worldwide.

Subject of Research: Improving snow simulations through improved representations of vegetation conditions: insights from high resolution simulations over California

Article Title: Improving snow simulations through improved representations of vegetation conditions: insights from high resolution simulations over California

Article References: Gupta, A., Behrangi, A., Farmani, M., Broxton, P., & Niu, G.-Y. (2026). Improving snow simulations through improved representations of vegetation conditions: insights from high resolution simulations over California. The Cryosphere, 20(10), 5973-5992. https://doi.org/10.5194/tc-20-5973-2026

Image Credits: AI Generated

DOI: 10.5194/tc-20-5973-2026

Keywords: Improving, snow, simulations, improved, representations, vegetation, conditions, insights, high, resolution, over, California

The research, titled “Role of Language in Shaping Climate Futures: A Frame Analysis of Carbon Governance in China and the US,” was conducted by a team working at the intersection of ecolinguistics, computational text analysis, and environmental policy. Rather than treating policy language as a neutral container for technical decisions, the authors propose that discourse actively intermediates between institutional behavior, epistemic culture, and measurable environmental outcomes. In other words, the way a government talks about carbon reflects how it thinks about carbon, and that thinking constrains what its policies can realistically achieve.

To test this idea, the researchers assembled a corpus of 209 official policy documents covering carbon governance in China and in the United States—specifically California—between 2007 and 2022. The fifteen-year window captures a critical period in global climate politics, spanning the aftermath of the Copenhagen negotiations, the Paris Agreement, China’s pledge to peak carbon dioxide emissions before 2030, and California’s increasingly aggressive legislative and regulatory agenda. The comparison is deliberately asymmetric: a centralized state pursuing carbon targets through planning and industrial policy, and a subnational jurisdiction operating through courts, statutes, and regulatory agencies.

Methodologically, the study is notable for combining quantitative and qualitative techniques within a single analytical architecture the authors call the language–policy–outcome (LPO) framework. On the quantitative side, large-scale topic modeling and network analysis were applied to the document corpora to identify recurring themes and the structural relationships among them. On the qualitative side, the team performed ecolinguistic frame analysis, examining the metaphors, narratives, and value assumptions embedded in the texts. Finally, correlation analysis linked the discursive patterns to carbon dioxide emissions data, allowing the researchers to ask whether distinctive ways of talking about climate were associated with distinctive trajectories of carbon intensity and absolute emissions.

The results reveal sharply contrasting discursive cultures. Chinese policy discourse, the analysis found, exhibits a technocratic and development-oriented frame in which climate mitigation is tightly interwoven with modernization, innovation, and economic upgrading. Carbon reduction appears in these documents not as a constraint on growth but as a driver of it, articulated through the vocabulary of green technology, industrial transformation, and national development strategy. California’s policies, by contrast, are characterized by a legalistic and institutional tone. Climate action is organized around legislation, government agencies, and regulatory procedures, with the state’s discourse emphasizing mandates, standards, enforcement mechanisms, and the procedural machinery of environmental governance.

Yet the similarities proved as revealing as the differences. Despite their divergent rhetorical styles, both corpora share what the researchers describe as a weak sustainability paradigm. In both cases, the dominant frames emphasize short-term compliance, incremental efficiency improvements, and techno-economic optimization rather than systemic socioecological transformation. Neither discourse, in other words, fundamentally questions the growth-oriented structures that generate emissions in the first place. The shared paradigm suggests that even politically opposite systems can converge on a similar conceptual horizon, one in which climate policy is framed as a problem of managing and optimizing existing systems rather than reimagining them.

The comparison with emissions data added a sobering empirical layer. In China, several policy themes were correlated with improvements in carbon intensity—the amount of carbon dioxide emitted per unit of economic output. But the study found that these efficiency gains coexisted with continued growth in absolute emissions, a pattern the authors interpret as a potential rebound effect. When economic or infrastructural activity expands rapidly enough, the emissions saved through efficiency can be offset, or more than offset, by the sheer scale of new construction, energy demand, and industrial output. Efficiency, in this framing, is a necessary but insufficient condition for decarbonization.

California presented a more complicated picture. Policies there were associated with emissions outcomes, but the strength of those associations varied considerably across policy themes. Financial and technology-focused policy instruments—subsidies, investment programs, innovation incentives—showed weaker or more tangled relationships with environmental results than more direct regulatory approaches. This suggests that market-oriented and innovation-framed climate policies, however prominent in official discourse, may be less reliably connected to measurable emission reductions than their rhetorical weight in policy documents would imply.

From these findings the researchers derive what they call the discursive coherence–emission fragmentation paradox: a government can construct a highly coherent, internally consistent climate-policy narrative without producing equally coherent changes in actual emissions. Discourse, the study stresses, does not stand in a direct causal relationship with carbon outputs. Instead, policy language mediates and conditions decarbonization—it enables certain actions, constrains others, and orients institutional attention toward particular kinds of solutions—while economic dynamics and governance structures do much of the heavy lifting in determining final outcomes. The practical implication is that evaluating climate governance solely through targets, regulations, and emission statistics misses a crucial layer of the story: how words frame the data, and how those frames shape what policymakers perceive as possible.

The authors are careful about the limits of their analysis and outline an agenda for future work. They call for extending the comparison to additional countries and economic systems, and for combining policy-text analysis with stakeholder interviews, ethnographic research, case studies, and a broader set of economic and institutional indicators. Such triangulation, they argue, would provide a fuller account of how language operates within the wider socio-ecological system. The study was supported by the Key Project of the National Natural Science Foundation of China and by the Italian Ministry of University and Research. As nations race toward mid-century net-zero commitments, the research offers a reminder that the climate future is being written twice—first in the language of policy documents, and only afterward in the atmosphere itself—and that the first draft deserves far more scrutiny than it typically receives.

Subject of Research: Ecolinguistic frame analysis of carbon governance discourse and its relationship to carbon dioxide emissions in China and California from 2007 to 2022

Article Title: Role of language in shaping climate futures: a frame analysis of carbon governance in China and the US

Article References: Role of language in shaping climate futures: a frame analysis of carbon governance in China and the US. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: carbon governance, ecolinguistics, climate policy, frame analysis, China, California, language-policy-outcome framework, topic modeling, carbon emissions, rebound effect, sustainability discourse, decarbonization

Percutaneous coronary intervention, commonly known as PCI, is the definitive emergency treatment for many heart attacks. The procedure involves threading a catheter, usually through an artery in the wrist or groin, up to the blocked coronary artery that is starving heart muscle of oxygen. Cardiologists then inflate a small balloon and typically deploy a stent, a tiny metal scaffold, to prop the vessel open and restore blood flow. The intervention is exquisitely time-sensitive: the longer the artery stays occluded, the more heart muscle dies, and the greater the risk of fatal arrhythmias, heart failure, and death. That time dependence is why emergency medicine guidelines emphasize rapid door-to-balloon intervals, and why the geographic distance between a patient and a PCI-capable hospital can be a matter of life and death. Establishing a PCI program is also expensive and demanding, requiring specialized catheterization suites, imaging equipment, and a round-the-clock team of interventional cardiologists, nurses, and technicians.

The research team, led by Yu-Chu Shen of the Naval Postgraduate School and the National Bureau of Economic Research and Renee Y. Hsia, professor of emergency medicine at UCSF, took advantage of a natural experiment embedded in California’s health care landscape. Between 2011 and 2022, various hospitals across the state opened new PCI-capable facilities at different times and in different places. The researchers identified 590,282 hospitalizations for acute myocardial infarction, the medical term for heart attack, and asked a deceptively simple question: what happened to patients in a community after a PCI-capable facility opened within a 30-minute drive? Crucially, they separated communities into two groups based on their starting point, comparing places that had never had timely PCI access with places that already had a capable hospital within the same 30-minute radius.

The contrast between the two groups was stark. In communities that previously lacked timely access, the opening of a new facility was associated with a 53 percent relative increase in the likelihood that a heart attack patient would be admitted to a hospital capable of performing PCI, and a 16 percent increase in the likelihood that the patient would actually receive the procedure. Those changes in treatment were accompanied by substantial changes in survival: thirty-day mortality among heart attack patients fell by 20 percent, and one-year mortality fell by 15 percent. In communities where PCI was already available within 30 minutes, by contrast, the opening of yet another facility increased admission to a PCI-capable hospital by only 2 percent and use of the procedure by only 3 percent, and the study detected no significant change in mortality at either time horizon.

The pattern reveals a fundamental asymmetry in how health care capacity behaves. Where a service is absent, adding it changes clinical behavior profoundly, rerouting patients who would otherwise have received less definitive therapy toward the lifesaving procedure. Where the service already exists, an additional facility mostly reshuffles existing patterns of care, drawing some patients from one capable hospital to another without meaningfully changing who gets treated or how quickly. In economic terms, the marginal benefit of the first facility within reach of a population is enormous, while the marginal benefit of the second or third approaches zero. Hsia illustrated the point with a deliberately mundane analogy: adding another gas station to a neighborhood that already has five may make refueling slightly more convenient, but building one where residents previously had to drive 30 miles meets a genuine need. With emergency cardiac care, she noted, filling that kind of gap can save lives.

The geographic dimension of the problem was most acute in rural California. Among the 52,756 heart attack patients in the study who came from rural communities, 81 percent lived in places without a PCI-capable facility within a 30-minute drive at the beginning of the study period. That figure captures a structural vulnerability of rural health systems: specialized services that require high volumes, expensive equipment, and scarce subspecialists tend to concentrate in urban and suburban population centers, leaving rural patients with long transport times precisely when minutes matter most. For these communities, the study’s findings suggest that even a single new PCI-capable facility could produce mortality reductions on the order of 15 to 20 percent, an effect size that few other health system interventions can match.

The new analysis builds directly on earlier work by Hsia and colleagues, which had examined the supply side of the same question. That previous study found that the total number of PCI procedures performed in a community rose by 7.5 percent after a new facility opened, but the increase was far from uniform: it approached 20 percent in communities that had previously lacked timely access, compared with roughly 8 percent in communities where PCI was already available. In other words, new capacity reliably generates more procedures, but the increase partly reflects redistribution of existing patients among competing hospitals rather than genuinely new treatment for people who would otherwise have gone without. The current study extends that logic to the outcome that ultimately matters, asking whether the additional procedures improved survival, and finding that the answer depends entirely on whether the capacity filled a gap or duplicated an existing service.

That distinction carries weight because of how decisions about specialized services are actually made in the American health system. The researchers point out that the placement of new facilities is frequently driven by market demand, expected procedure volumes, and hospitals’ anticipated return on investment rather than by systematic assessments of population need. A hospital system may find it financially attractive to open a second PCI laboratory in a well-served suburban corridor where it can capture market share from a competitor, while no institution finds it profitable to establish one in a rural county with a small and dispersed population. The result, the study implies, is a systematic misallocation: capital flows toward communities where the clinical benefit of additional capacity is smallest, while the places where a new facility would save the most lives remain unserved. Market forces, in this domain, do not naturally point resources toward maximal population health benefit.

For policymakers, the findings suggest a concrete principle for allocating scarce health care infrastructure investment: measure access before adding capacity. Rather than asking whether a region can support more specialized services, planners could ask whether residents can already reach those services within the critical treatment window, and direct new facilities and support programs toward the communities that cannot. Hsia framed the principle as a question of priorities. When millions of dollars are invested in specialized health care infrastructure, she argued, the question should not simply be whether more services can be provided, but where those services will do the most good. Expanding access, she said, needs to be done wisely. The study’s mortality figures give that advice quantitative teeth: the same investment that yields no measurable survival benefit in a saturated market can cut heart attack deaths by a fifth in a community that has been going without.

The research also carries a broader lesson for how the public and the press evaluate health care expansion announcements. A ribbon-cutting for a new specialty facility is often reported as an unambiguous improvement in local care, and in a community with no existing access, that framing is accurate. But the California data show that in well-served areas, additional capacity may increase treatment rates only marginally while leaving survival untouched, meaning the true value of such investments is far smaller than the rhetoric suggests. As health systems, insurers, and governments confront rising costs and finite resources, the study offers a reminder that in medicine, as in geography, position matters. The lifesaving power of a procedure like PCI is not just a property of the catheter and the stent; it is a property of the distance between the patient and the door.

Subject of Research: Geographic access to percutaneous coronary intervention facilities and heart attack survival outcomes

Article Title: When it comes to health care access, more isn't always better

Article References: When it comes to health care access, more isn't always better. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: heart attack, percutaneous coronary intervention, health care access, rural health, mortality, emergency care, health policy, hospital capacity, California, JAMA Network Open, cardiovascular disease, health care infrastructure

The scale of the dataset is what makes the analysis unusual. Most studies of heat and health have focused on single cities, limited age groups, or narrow sets of diagnoses. This study, by contrast, drew on statewide emergency department records from the California Department of Health Care Access and Information, covering every visit in the state over a 13-year span. The researchers sorted the visits into three clinical groupings: heat-related physical illness, which includes septicemia, fluid and electrolyte disorders, congestive heart failure, renal failure, urinary tract infections, and heat stroke; heat-related mental health visits, spanning diagnoses from anxiety and mood disorders to schizophrenia, substance use, and intentional self-injury; and all-cause visits, meaning every emergency department encounter regardless of diagnosis. In total, the physical illness grouping contained more than 13.8 million visits and the mental health grouping more than 15.7 million.

To connect those visits to temperature, the team used daily gridded mean temperature data at 4-kilometer resolution from the PRISM Climate Group, a statistical mapping system that blends observations from federal, state, and local weather station networks with a digital elevation model to account for California’s notoriously complex topography, including rain shadows, temperature inversions, slope aspect, and coastal proximity. Using Google Earth Engine, the researchers performed a spatial join that computed the area-weighted mean of all grid cells intersecting each ZIP code polygon on each day. Over the study period, the statewide average temperature was 16.2 degrees Celsius, with ZIP code averages ranging from 12.6 to 18.1 degrees across the state’s six Health Officer Regions, which span everything from the cool, foggy Bay Area to the scorching agricultural interior of Central California.

The statistical engine of the study was the distributed lag nonlinear model, or DLNM, a framework widely used in environmental epidemiology that can simultaneously capture two features of the temperature-health relationship: its nonlinearity and its timing. In the first stage, the researchers fitted a quasi-Poisson time-series regression for each individual ZIP code, modeling the exposure-response curve with a quadratic B-spline whose three internal knots were placed at the 10th, 75th, and 90th percentiles of each location’s temperature distribution. The lag-response curve, which describes how risk evolves after exposure, was modeled with a natural cubic spline over a three-day window, reflecting evidence from prior heat-health studies that most excess risk occurs within zero to three days of exposure. Alternative specifications with longer lags and different spline complexity were compared using the Akaike Information Criterion, and the final model produced smooth, epidemiologically plausible curves, with risk peaking within a day of exposure and declining toward zero by day three.

Because small ZIP codes with few emergency visits produce noisy estimates, the second stage pooled the location-specific results with a multivariate meta-regression and derived best linear unbiased predictions for each ZIP code, allowing sparsely populated areas to borrow statistical strength from larger, climatically similar neighbors. Local average temperature and temperature range were included as meta-predictors, since local climate is known to modify how temperature affects health. From these fitted curves, the team calculated each ZIP code’s minimum morbidity temperature, the point at which heat-related illness is minimized, and used it as the reference for computing attributable fractions: the proportion of emergency visits that would not have occurred had temperatures stayed at that optimum.

The headline numbers are striking. Statewide, 5.1 percent of heat-related physical illness visits, 5.4 percent of heat-related mental health visits, and just under 1 percent of all-cause emergency visits were attributable to heat. Regionally, Central California and Los Angeles carried the largest heat-attributable fractions for physical illness, at 6.5 percent and 6.1 percent respectively, while Los Angeles, Southern California, and Central California led for mental health visits. The Rural North and Greater Sierra-Sacramento regions, despite their cooler average temperatures, had the largest fractions of all-cause heat-attributable visits, a pattern the authors suggest may reflect lower acclimatization and weaker adaptive capacity in communities unaccustomed to extreme heat.

The most consequential findings, however, concern social inequality. To measure it, the researchers turned to the California Healthy Places Index, or HPI, a composite metric developed by health departments and community organizations that ranks every census tract on a 0-to-100 scale based on 23 social determinants of health most strongly associated with life expectancy at birth. These indicators are grouped into eight policy-action domains covering the economy, education, transportation, housing, social resources, the built and clean environment, and health-care access. Low-scoring communities are marked by poverty, unemployment, overcrowded and cost-burdened housing, limited car ownership, long commutes, sparse tree canopy, higher pollution, and low insurance coverage. When the researchers divided ZIP codes into HPI quartiles, a clear gradient emerged: as community conditions improved, emergency visit rates and heat-attributable rates fell for all three outcome categories. ZIP codes in the lowest quartile held 27 percent of the state’s population but contributed 33 percent of physical illness visits, 32 percent of mental health visits, and 35 percent of all-cause visits, and had 2.2 times more total emergency visits than the healthiest quartile.

Expressed as rates, the disparity becomes even sharper. Statewide, heat-attributable visits averaged 141 per 100,000 people annually for physical illness, 165 for mental health, and 317 for all causes. In the lowest HPI quartile, the rate ratios relative to the healthiest quartile were 4.3 for physical illness, 2.1 for mental health, and 2.6 for all-cause visits, with confidence intervals that did not overlap across quartiles. Central California recorded the highest physical illness rate at 218 visits per 100,000, a rate ratio of 3.4 relative to the Bay Area, while the Rural North posted the highest mental health and all-cause rates at 234 and 852 per 100,000. The authors point to mechanisms embedded in the HPI itself: limited access to parks, beaches, shade, and transportation, combined with less air conditioning and fewer green spaces, can raise both heat exposure and vulnerability, while higher stress and reduced access to health services may amplify the physiological and psychological toll.

The study also exposed a systematic blind spot in official statistics. When the researchers compared their population-weighted county estimates with crude heat-associated emergency visit rates in Tracking California, the state’s environmental public health tracking network, their estimates were on average seven times higher, with a range of 2.3 to 18 times. Tracking California relies on heat-specific diagnosis codes appearing in medical records, which the authors argue captures only direct attribution and misses the many heat-related illnesses that present as cardiac, renal, or psychiatric emergencies. Correlations between the two data sources were strong for physical outcomes but weak and statistically insignificant for mental health, suggesting that the psychological burden of heat is the most undercounted of all.

With climate projections pointing to more frequent, intense, and longer-lasting heat waves, the authors frame their results as a conservative baseline rather than a worst case, noting that the study period ends before California’s recent record-breaking heat. They argue that extreme heat action plans, such as California’s 2022 state resilience strategy, could be made more effective by directing investments toward the ZIP codes this analysis identifies as having both unhealthy conditions and high heat-attributable rates: more trees and shaded areas, public air-conditioned spaces, cooling stations, cool roofs and pavement, and housing-focused measures such as weatherization incentives, subsidies for cooling technologies, and utility assistance. Because the study assigned visits to residential ZIP codes, it excluded people experiencing homelessness, who are among the most heat-vulnerable, and the authors caution that humidity and air pollution were not controlled. Even so, the message is unambiguous: heat is not an equal-opportunity hazard, and protecting the communities with the fewest resources for healthy living is both an equity imperative and the most efficient way to reduce the strain that rising temperatures place on emergency care.

Subject of Research: Inequities in heat-attributable emergency department visits across California communities

Article Title: Inequities in heat-related emergency department visits in California from 2005 – 2017

Article References: Donnelly, M., Lappe, B., Milet, M., & Vargo, J. (2026). Inequities in heat-related emergency department visits in California from 2005 – 2017. PLOS Climate, 5(10), e0000586. https://doi.org/10.1371/journal.pclm.0000586

Image Credits: AI Generated

DOI: 10.1371/journal.pclm.0000586

Keywords: extreme heat, emergency department visits, California, health equity, Healthy Places Index, distributed lag nonlinear model, climate change, mental health, social determinants of health, heat-attributable fraction, public health surveillance, acclimatization

The research team, led by Abid Sarwar and including Josué Medellín-Azuara, John T. Abatzoglou, and Joshua H. Viers, focused on a concept known as consumptive use. In water resources engineering, consumptive use refers to water that is taken up by plants and evaporated into the atmosphere and is therefore no longer available for downstream users or aquifer recharge. It is not the same as the water diverted from a canal or pumped from a well, because some of that diverted water returns to the system through runoff or deep percolation. Consumptive use is best approximated by actual evapotranspiration, the sum of evaporation from soil and transpiration through plant leaves, which is the quantity that ultimately determines how much water a farming operation truly removes from the landscape.

To measure evapotranspiration at the scale of individual fields, the researchers turned to OpenET, a publicly available platform that produces gridded estimates of actual evapotranspiration by combining multiple satellite-driven energy balance models into an ensemble product. They paired these evapotranspiration estimates with vegetation index data from the Sentinel-2 satellites, specifically the normalized difference vegetation index, or NDVI, which tracks the greenness and vigor of plant canopies through time. Because NDVI responds to canopy development, crop type, planting density, and stress, its seasonal trajectory carries a fingerprint of how each parcel is being managed. The team also incorporated precipitation data from the PRISM climate mapping system to place each parcel’s water use in its hydroclimatic context.

The study area covered 2,189 agricultural parcels spanning roughly 7,483 hectares of the Santa Clara Valley, an agriculturally productive and remarkably diverse region south of the San Francisco Bay. The analysis focused on three major crop categories: truck crops, which include vegetables and berries; vineyards; and hay crops. Data from 2019 through 2023 were assembled into time series for every parcel, creating a rich record of how vegetation greenness, water consumption, and rainfall varied seasonally and from year to year across the landscape.

Rather than assuming that all fields of a given crop behave alike, the researchers applied unsupervised machine learning, a family of algorithms that discovers natural groupings in data without any predefined labels. Two clustering approaches were tested: one using NDVI time series alone, and a second combining NDVI with the magnitude and trend of actual evapotranspiration. Both approaches identified 13 distinct consumptive-use clusters across the valley, but the composition was telling. Truck crops split into six clusters, vineyards into four, and hay crops into three. In other words, what appears on paper as a single crop category actually encompasses several operationally distinct water-use regimes.

The quantitative comparison between the two clustering strategies revealed why adding evapotranspiration information matters. The researchers evaluated cluster quality using the coefficient of variation, a normalized measure of dispersion, both within clusters and between cluster means. Under the NDVI-only approach, the ratio of between-cluster separation to average within-cluster variation was modest: 0.30 for truck crops, 0.46 for vineyards, and 0.74 for hay crops. When evapotranspiration magnitude and trend were added, those ratios jumped to 1.35, 1.06, and 1.07 respectively, meaning the clusters became far more distinct from one another relative to the scatter within them. Adding the water-use dimension also generally reduced the average within-cluster coefficient of variation. For context, the crop-wide coefficients of variation, the very heterogeneity that conventional averages conceal, stood at 17 percent for truck crops, 24 percent for vineyards, and 20 percent for hay crops.

Each crop category told its own story. Truck crops displayed the strongest seasonal and interannual heterogeneity of the three groups, which makes sense for a category encompassing vegetables and berries with different planting calendars, growth cycles, and irrigation demands. Vineyards showed clearer spatial differentiation and, notably, a modest decline in consumptive use over the study period, a pattern consistent with the widespread adoption of efficient irrigation technology and deliberate deficit irrigation strategies that premium winegrape growers use to control fruit quality. Hay crops separated into two fundamentally different systems: pasture-based operations and grain-forage systems, which have contrasting water delivery needs and different sensitivities to climate variability. A hay field that is actually irrigated pasture behaves very differently from one planted to a cereal forage crop, and the clustering algorithm detected that distinction without ever being told to look for it.

Crucially, the observed consumptive-use patterns aligned with known hydroclimatic stresses and management practices in the valley, including cover cropping, variation in irrigation technology, deficit irrigation, and seasonal fallowing. The authors are careful to note that directly attributing a specific cluster’s behavior to a specific practice would require independent evidence from field surveys, water permits, or metering data, which the remote sensing analysis alone cannot provide. Still, the correspondence between the statistical structure of the clusters and the agronomic realities on the ground lends credibility to the approach. The machine learning is not inventing patterns; it is surfacing real differences in how water moves through these agricultural systems.

The practical implications extend well beyond Santa Clara Valley. In regions where agriculture depends heavily on groundwater, regulators and irrigation districts need to know not just how much water a crop category consumes on average, but which individual parcels are consuming more or less than their peers and why. Parcel-scale consumptive-use clusters can support customized water-management plans, targeted irrigation monitoring, and conservation programs that reward growers who demonstrably reduce consumptive use rather than those who merely plant a low-use crop. During droughts, when fallowing decisions and curtailments are negotiated field by field, having an objective, remotely sensed baseline of who uses what, and how those patterns shift year to year, provides a common factual foundation for difficult allocation conversations.

The study also demonstrates a broader methodological lesson for the era of open environmental data. By fusing freely available Sentinel-2 imagery, the OpenET ensemble, and PRISM climate records with accessible machine learning tools, the researchers produced actionable water intelligence without deploying a single ground sensor across the study area. As satellite revisit times shorten and evapotranspiration products grow more accurate, similar analyses could be replicated in irrigated valleys worldwide, giving water managers in groundwater-dependent basins the fine-grained, adaptive picture of agricultural water use that legacy crop maps were never designed to deliver. For a state where every acre-foot counts, seeing water use field by field may prove as important as measuring it at the wellhead.

Subject of Research: Remote sensing and machine learning analysis of parcel-scale agricultural consumptive water use in California's Santa Clara Valley

Article Title: Identifying agricultural consumptive-use patterns to support adaptive water management in California’s Santa Clara Valley via remote sensing and machine learning

Article References: Sarwar, A., Medellín-Azuara, J., Abatzoglou, J. T., & Viers, J. H. (2026). Identifying agricultural consumptive-use patterns to support adaptive water management in California’s Santa Clara Valley via remote sensing and machine learning. PLOS Water, 5(7), e0000416. https://doi.org/10.1371/journal.pwat.0000416

Image Credits: AI Generated

DOI: 10.1371/journal.pwat.0000416

Keywords: remote sensing, machine learning, evapotranspiration, consumptive use, groundwater management, Santa Clara Valley, Sentinel-2, OpenET, NDVI, agriculture, water management, California