Gut microbiome community composition and diversity often shift in response to host diet, but this responsiveness may depend on microbial transmission mode. While environmentally acquired microbes are well documented to respond to dietary change, the ecological flexibility of vertically inherited microbiomes remains poorly understood. The wood-feeding cockroach Cryptocercus punctulatus harbors three phylogenetically distinct and taxonomically diverse microbial groups in its hindgut: two vertically transmitted protist lineages (Parabasalia and Oxymonadida) and a bacterial community acquired via mixed-mode transmission. We used feeding experiments and microbiome profiling to assess the responsiveness of these three communities to diet in late-instar C. punctulatus. The Parabasalia community did not differ with host diet changes, in line with the expectation of stability in vertically inherited microbial communities. In contrast, Oxymonadida exhibited differences in community structure with respect to changes in relative abundance but not composition of taxa. The bacterial community differed across a broader range of diversity and composition measures, including richness and community structure based on both presence/absence and relative abundance. These results indicate that some vertically inherited microbial communities retain ecological flexibility, but the level of responsiveness to environmental change varies.

Across a variety of ecosystems, taxonomically distinct primary producers have been shown to support communities that differ in diversity and/or community structure. Additionally, ecosystem engineers and other abundant species have been shown to change the structure and diversity of communities across different ecosystems. Deadwood-dependent (saproxylic) arthropod communities play a crucial role in forest ecosystems, facilitating decomposition of wood and nutrient cycling in forests Decaying deadwood habitats are highly heterogeneous and are able to support large and diverse communities of these deadwood-dependent arthropods. However, our understanding of how these communities differ across different ecosystem factors, such as community composition and characteristics of the wood itself, is limited. We investigated the impacts of two factors on saproxylic arthropod diversity: deadwood type (hardwood vs. softwood) and abundance of Cryptocercus punctulatus, an abundant, social, wood feeding insect. We sampled arthropod communities from 26 logs in the western Appalachian Mountains and identified arthropods to order- and family-level. We observed a large amount of variation in the abundance and diversity of the arthropod taxa sampled, but no effects of log type or C. punctulatus abundance on the arthropod community diversity or community structure. Our study suggests that broad scale deadwood arthropod diversity may not be structured by variation in log type or colonization history of social wood feeding insects, providing clarity about the relationship between two underexplored drivers of diversity within these essential arthropod communities.

A document by Therese Madeleine Tankam. Click on the document to view its contents.

In this paper, we consider the distributed Kalman filtering problem over sensor networks with a finite number of sensor nodes where each node can communicate only with its neighboring nodes. We first introduce the concept of gradual information fusion estimation (GIFE) and propose an algorithm for computing the GIFE that can obtain performance improvement via reducing estimation error covariance. Also, we prove that the GIFE can be expressed as a weighted sum of local estimates. Then, based on the results of GIFE and some results proposed in this paper, two distributed Kalman filters are developed where, at time step k, each node is allowed to communicate with its neighboring nodes at most once in the first filter and each node is permitted to communicate with its neighboring nodes twice in the second filter. In addition, we prove that either of the two proposed distributed Kalman filters is unbiased and the estimation error covariance of the first distributed Kalman filter is less than or equal to that without using information fusion estimation. We prove that the estimation error covariance of the second distributed Kalman filter is less than or equal to that of any local estimate belonging to a set. An example of unmanned ground vehicle is provided to illustrate the performance of the two proposed distributed Kalman filters.

This study presents an extended formulation of the observational validation and falsifiability conditions of the collective curvature and rotational memory framework. The purpose of this work is not to further generalize or mathematically formalize the theory, but to define, with greater clarity and structural rigor, the precise observational conditions under which the framework can be maintained or must be rejected. In this theory, collective curvature is not interpreted as the curvature produced by a single dominant mass, but as a layered structural curvature field formed through the superposition of stars, gas, primordial material, compression layers, and rotational memory established during formation epochs. This curvature is not static but evolves over time through processes such as material redistribution, shear accumulation, and structural compression. Rotational memory is defined as a persistent structural imprint of the initial conditions under which matter formed. It includes curvature width, curvature gradient, compression state, and rotational environment, all of which continue to influence the motion and spatial distribution of matter over long timescales. Thus, present-day kinematic states cannot be understood solely as a function of current mass distribution, but must be interpreted as the outcome of both present structure and formation history. Within this framework, outer high rotational velocities, inner slow old stellar populations, radial star formation patterns, and gravitational lensing structures are not independent phenomena. They are unified structural consequences of curvature width contraction and the persistence of rotational memory. This study distinguishes between observational patterns that must repeatedly appear if the framework is valid and those that would directly falsify it. Particular emphasis is placed on conditions that would lead to immediate structural collapse of the theory, such as the repeated observation of large arcs originating from internal collective curvature in mature spiral galaxies. The conceptual foundation of this framework is formally introduced in: https://doi.org/10.22541/au.177023319.90275686/v1

This paper presents a topological structural analysis of the Missouri Snap geodetic gate (Node 18) utilizing the 27-Dimensional Chronetic Topology (27DCT). By applying the 14th Holmes Law (Vertical Columnar Summation), we model the mechanical distribution of the 12.2 ZettaJoule (ZJ) energy imbalance across the Alpha, Beta, and Gamma rotations of the 1,080-node Straight-Screw Manifold. This analysis provides a unified reconciliation of the 3.3h Grid Lag and the 1.33ms LOD Jitter, establishing a 2.0nm-scale theoretical baseline within the model for planetary geodetic stabilization.

1. Savannas, spanning 20% of the Earth’s surface, are characterized by a continuous grass matrix interspersed with woody patches, supporting high biodiversity and providing ecological and economic services. Coexistence is maintained by interactions among climate, soil nutrients, and disturbance, but can be destabilized by invasion, land-use changes, and climate shifts. Although climatic and edaphic controls on savanna structure are well-studied, the contribution of soil microbial communities in maintaining spatial heterogeneity and coexistence remains poorly understood. We investigated whether savanna heterogeneity is mirrored belowground and how disturbance and invasion by Megathyrsis maximus modify these relationships. 2. We used a factorial field sampling design in a mesquite savanna to compare woody patches and adjacent grasslands with and without mechanical disturbance and invasion. We quantified soil physiochemical properties, plant community composition, and bacterial and fungal communities to evaluate linkages among vegetation and soils. 3. Grasslands and woody patches supported distinct soil and microbial assemblages, consistent with differences in vegetation inputs and nutrient regimes. Grasslands exhibited relatively simple assembly patterns, with plant diversity closely associated with soil nutrients and comparatively homogeneous microbial communities. In contrast, woody patches displayed more complex assembly dynamics, with microbial communities structured by plant diversity, soil nutrients, precipitation, and spatial distance, indicating greater environmental heterogeneity and stronger dispersal limitation. 4. Invasion by M. maximus increased soil nutrient availability and altered microbial community composition in grasslands, while mechanical disturbance produced similar but weaker effects. These impacts reduced grassland microbial distinctiveness and disrupted linkages between vegetation, soils, and microbes. 5. Synthesis. Aboveground patch structure in savannas is tightly coupled to belowground microbial assembly and nutrient dynamics. By demonstrating that invasion and disturbance weaken these spatial linkages, this study indicates that soil microbial communities contribute to savanna coexistence and resilience and reveals how belowground community assembly underpins landscape-scale heterogeneity and its destabilization under multiple stressors.

Sassafras tzumu, an important tree species in subtropical forests of China, provides a valuable model for understanding regional biodiversity patterns and informing conservation strategies. Here, we integrated nuclear microsatellites (nSSRs), chloroplast DNA (cpDNA) sequences, and ecological niche modeling (ENM) to investigate 29 populations spanning the species’ entire natural range. nSSR analyses revealed a moderate level of genetic diversity (mean HE = 0.308), and analysis of molecular variance showed that 33% of the total genetic variation occurred among populations (FST = 0.333), indicating substantial population differentiation. Bayesian clustering, principal coordinate analysis, and phylogenetic reconstruction consistently supported two major genetic lineages: Lineage I from central and southeastern China and Lineage II from southwestern China. Their geographic boundary closely corresponded to the Daba Mountains–Xuefeng Mountains topographic barrier. Six cpDNA haplotypes (H1–H6) were identified, and both haplotype network and phylogenetic analyses supported two major clades largely congruent with the nSSR-based lineage pattern. A significant phylogeographic structure was detected (NST > GST, P < 0.05). Divergence time estimation suggested that the split between the two lineages began in the Miocene (ca. 9.49 Mya), whereas major haplotype diversification occurred during the Pleistocene. Mismatch distribution and neutrality tests indicated historical population expansion. Ecological niche modeling indicated that the suitable range of S. tzumu during the Last Glacial Maximum was markedly reduced relative to the present, and that areas of high habitat suitability overlapped closely with several genetically inferred glacial refugia, including the Daba, Dabie, Wuling, and Wuyi Mountains, supporting a multiple-refugia scenario. Together, these results suggest that the current pattern of low diversity yet high differentiation in S. tzumu reflects the combined effects of Miocene geological events, Quaternary climatic oscillations, and strong contemporary geographic isolation, and provide a scientific basis for lineage-based conservation and climate-adaptive management.

This memorandum places into the public scientific record a compact, independently testable empirical finding: a hypothesis-testing cohort of biological period-doubling systems exhibits systematic, mechanism-dependent deviations from the classical Feigenbaum constant (δ ≈ 4.669), while physical and mathematical control systems remain tightly concentrated near δ ≈ 4.669. The dataset consists of 16 systems (10 biological + 6 controls) with canonical δ values extracted from the literature or digitizable bifurcation evidence and recorded under fixed selection rules. Two non-parametric tests show complete separation (i) between biological deviations and control deviations (Mann-Whitney U = 60/60, exact p ≈ 1.25 × 10⁻⁴) and (ii) between two biological mechanism-defined tiers (Tier 1 vs Tier 2: U = 0/21, exact p = 0.00833), with a 2.479 δ-unit tier gap and zero overlap. This document is not a final manuscript. The cohort and statistics may change as additional sources are incorporated and the extraction ledger is finalized. A full protocol-forward manuscript including the complete dataset, extraction images, and full reference ledger will follow.This work is a preprint. The original data record and timestamped version are archived at Zenodo: https://doi.org/10.5281/zenodo.18825703GitHub repository: https://github.com/devinromberger123-prog/rgl-scaling Software DOI: https://doi.org/10.5281/zenodo.19212391

Objectives: To examine geographic variation, temporal trajectories, and equity dimensions of quality in Australian residential aged care using the Star Ratings system introduced in late 2022. Design: Repeated cross-sectional and longitudinal panel analysis of publicly available administrative data, reported in accordance with the STROBE guidelines [1]. Setting: All rated residential aged care facilities in Australia (census of the rated population, not a sample). Participants: 3,002 unique facilities observed across 11 quarterly Star Ratings extracts (May 2023 to October 2025), yielding 28,708 facility-quarter records. Main outcome measures: Overall Star Rating (1-5 stars) and four sub-category ratings (residents' experience, compliance, staffing, quality measures). Secondary outcomes: spatial clustering (Global and Local Moran's I), quality trajectories by provider type (mixed-effects models), and quality desert identification (distance to nearest adequately rated facility). Results: Mean overall rating improved from 3.38 to 3.87 stars over 11 quarters, with improvement decelerating over time (quadratic term p < 0.001). In random-slopes mixed models, for-profit facilities rated 0.10 stars lower than not-for-profit (p < 0.001) and government facilities rated 0.43 stars higher (p < 0.001); these associations were robust to adjustment for facility size, state, remoteness, and area-level disadvantage. For-profit ownership was independently associated with persistent low performance (OR 2.75; 95% CI 1.13-6.70). Quality ratings clustered spatially (Global Moran's I = 0.24, p < 0.001), with 17 statistically significant low-quality clusters identified in Western Australia and South Australia. Eleven SA3 areas were classified as quality deserts (>50 km to a 3+ star facility), overwhelmingly in remote locations. Area-level socioeconomic disadvantage did not predict quality desert status after adjusting for remoteness (p = 0.60). Conclusions: Australia's Star Ratings system is associated with substantial sector-wide quality improvement, but a persistent for-profit quality gap and geographic clustering of low quality present opportunities for targeted regulatory intervention.

Background and Purpose Dexamethasone as the first-line treatment for neonatal respiratory distress syndrome (NRDS) has been found to simultaneously promote fetal lung mature and induce offspring neurotoxicity. This model informed study is aim to set the therapeutic window of dexamethasone for NRDS treatment. Experimental Approach Pregnant rats were injected with 0.1 or 0.4 mg/kg dexamethasone to simulate clinically equivalent prenatal-dexamethasone-exposure (PDE). Half rats were anesthetized. Surfactant protein A/B of fetal lung were quantified and pharmacokinetic/pharmacodynamics model was developed to determine the minimal effective concentration. The other half rats delivered naturally and the offspring were underwent open field testing. mRNA-sequencing of fetal hippocampus was performed and concentrations-toxicity assay was conducted in H19-7/IGR-IR cells. These data were integrated with maternal-fetal physiologically-based-pharmacokinetic model to determine the maximum tolerated concentration. Population-pharmacokinetic-model was used to propose optimal clinical dosing regimen. Key Results Pharmacokinetic/pharmacodynamics model well characterized dexamethasone-induced fetal lung maturation, indicating a minimal effective concentration as 1.63 ng/mL. Offspring of PDE rats exhibited significant behavioral alterations. mRNA-sequencing analysis revealed that neurotoxicity may be mediated through the Wnt pathway and renin-angiotensin system, with BDNF and CMA1 identified as potential neurotoxic biomarkers. Cell assays suggested a neurotoxic cut-off concentration of 17 ng/mL. Based on therapeutic window of 1.63-17 ng/mL, a half-dose clinical regimen was proposed as the optimal therapeutic strategy. Conclusion and Implications This model-informed precision dosing strategy offers a rational approach to optimize clinical dexamethasone dosing regimen.

This forensic audit utilizes the 27-Dimensional Chronetic Topology (27DCT) to map the 12.2 ZettaJoule (ZJ) Earth Energy Imbalance. By applying the 14 Holmes Laws, including the Holmes Law of Vertical Columnar Summation, we identify the 1,080-node Straight-Screw Manifold as the mechanical driver for the 1.33ms LOD Jitter. This document provides the Forever Data registry, locking the vertical geodetic threads between Node 32 (Bad Essen) and the 1,470-year Bond Harmonic.

Anthropogenic habitat modification can reshape dispersal dynamics, social organisation, and ultimately the genetic structure of populations. In cooperatively breeding species, where philopatry and helping are common, dispersal events are rare but critical for maintaining gene flow and reducing inbreeding. Using microsatellite data (N = 211 individuals) and long-term (1985 - 2023) social kinship data from the cooperatively breeding Arabian babblers (Argya squamiceps) living in natural and modified habitats in the hyper-arid Arava Valley (Israel), we examined how habitat modification influences the social and genetic structure. We combined spatial autocorrelation analyses, within-group relatedness estimates, and social kinship to assess patterns of dispersal, connectivity, and kinship across habitats. We found high gene flow across habitats and only weak differentiation among subpopulations, consistent with recent habitat changes that do not produce strong divergence. However, subtle but significant differences emerged between natural and modified habitats. Groups inhabiting modified areas showed shorter dispersal distances and lower within-group relatedness, suggesting higher turnover and increased immigration of unrelated individuals. In contrast, groups in natural habitats maintained stronger kin-based cohesion and broader positive spatial genetic structure. Our results indicate that habitat modification does not reduce connectivity in this system but reshapes the balance between philopatry and dispersal, favouring earlier dispersal and reduced kinship within groups. Modified habitats attract and concentrate individuals from multiple origins, while natural habitats function as demographic sources. These findings corroborate previous behavioural evidence of an accelerated pace of life in modified habitats and demonstrate that such life-history adjustments are detectable in the genetic and social structure. More broadly, our study highlights how anthropogenic changes in arid ecosystems can leave rapid genetic signatures in cooperative breeders, underscoring the importance of integrating behavioural, demographic, and genetic perspectives to understand population responses to environmental change.

Aims To quantify adherence to renal dosing recommendations in adults with advanced chronic kidney disease (CKD; stages 3–5, including dialysis) and to evaluate the clinical consequences of non-adherence. Methods This systematic review and meta-analysis was conducted in accordance with PRISMA 2020 and registered in PROSPERO (CRD42025620883). Eight electronic databases were searched from inception to 31 December 2025. Quantitative studies assessing renal dose adjustment practices in adults with advanced CKD were included. Two reviewers independently performed study selection, data extraction, and quality assessment using the Newcastle–Ottawa Scale and ROBINS-I. Random-effects meta-analyses (REML) were used to pool adherence proportions at prescription/order and patient levels. Results Thirty studies were included. Observational studies were of moderate to high quality (NOS 6–9/9), while two interventional studies had moderate risk of bias due to confounding. Adherence ranged from 33% to 83%. Pooled prescription/order-level adherence (n=16) was 48% (95% CI 0.42–0.55; I 2≈98%). Patient-level adherence (n=7) was higher at 70% (95% CI 0.57–0.83; I 2≈99.9%). Non-adherence was associated with adverse drug events, including hypoglycemia and nephrotoxicity, and with increased emergency department visits, hospitalizations, and mortality. Conclusions Adherence to renal dosing recommendations in advanced CKD is suboptimal and highly variable across settings. Interventions targeting prescribing systems, including clinical decision support tools, are warranted to improve guideline adherence and medication safety in this high-risk population.

We present a concrete ultraviolet-infrared (UV-IR) bridge connecting the Unified Fractal Quantum Field Theory (UFQFT) to observable cosmological structure. In UFQFT, spacetime is intrinsically fractal with an effective dimension D≈2.7 at microscopic (UV) scales, where particles emerge as resonance configurations of coupled energy (Φ) and charge (Ψ) fields. We show that, under coarsegraining, the theory exhibits a scale-dependent dimensional flow D(k)→3 toward macroscopic (IR) regimes, recovering an effective quasi-classical geometry. Within this framework, we derive a modified propagator of the form G(k)∼1/k 2+ϵ(k), where ϵ(k)=D(k)−3 encodes the fractal deviation. This directly leads to a suppression of the matter power spectrum, P(k)=PΛCDM(k) [1+(k/ktr) 2 ] −ϵ , providing a natural geometric origin for small-scale power suppression without invoking particle dark matter. The resulting predictions are quantitatively testable with upcoming large-scale structure observations, particularly the Euclid DR1 data release, offering a falsifiable pathway to connect fundamental fractal field dynamics with cosmological measurements.

INTRODUCTION Cybersecurity is crucial for organizations to contain and reduce cyber threats. However, modern teams are OVERWHELMED BY A HIGH VOLUME AND COMPLEXITY OF SECURITY ALERTS , which can lead to analyst fatigue, slower response times, and missed threats . On average, enterprise security teams analyze hundreds of thousands of events each day, leading to tens of thousands of hours wasted on false positives . This data overload exceeds human capacity, causing delays in responses and unresolved incidents  . Compounding the challenge is the CYBERSECURITY TALENT SHORTAGE. Industry studies predict that by 2025, a lack of skilled analysts (or human error due to overload) will account for over half of major security incidents . Meanwhile, attackers are employing advanced tactics; for example, threat actors are now leveraging generative to craft large-scale, convincing phishing campaigns . These trends emphasize the urgent need to IMPROVE AND AUTOMATE the incident response process. In this context, () and ) techniques offer promising solutions for IR automation. Many security workflows involve unstructured text data – from log messages and alerts to threat intelligence reports and incident tickets. Indeed, a large portion of cybersecurity information is encoded in textual or semi-structured form, making an invaluable tool for parsing and understanding security events . The rise of transformer-based language models has greatly enhanced capabilities in recent years. Transformers, such as and , possess a deep understanding of language, enabling them to perform tasks like classification, summarization, and anomaly detection on security-related texts. Previous research suggests that employing ( and transformers on cyber data can significantly enhance threat detection and incident analysis . For instance, IBM’s Watson for Cyber Security was trained on over one million security documents, assisting analysts in interpreting threat reports that were often inaccessible to traditional tools . Initial deployments indicated that cognitive technologies could reduce investigation times from weeks or days to just minutes by automating the correlation of threat intelligence with incident data . , an open-source library from Hugging Face, provides a gateway to the operationalization of such advanced models in practice. Hugging Face has built a vibrant platform with over 350,000 pre-trained models and 75,000 datasets, enabling practitioners to share and utilize state-of-the-art models  . Within the Hugging Face ecosystem, the _Transformers_ library focuses on text and sequence tasks. In contrast, the _Diffusers_ library specializes in TRANSFORMER-BASED DIFFUSION MODELS for generative tasks  . Diffusers simplify the development and inference of diffusion models (commonly used for generating images from text  ). Although primarily geared towards generative , leverages the same underlying transformer architecture that has revolutionized . In this paper, we adapt and fine-tune transformer models accessible through Hugging Face for cybersecurity incident response automation. We justify the use of (and related tools) due to their ease of use, access to pre-trained models, and strong community support, which together lower the barrier to applying cutting-edge in the security domain. Our goal is to demonstrate that transformer-based automation can transform incident response, reducing mean time to respond, improving detection accuracy, and enabling analysts to focus on complex decision-making rather than repetitive tasks. We organized the remainder of this paper as follows. - (SECTION 1) : Provides background on incident response processes and the motivation for automation. - (SECTION 2) : Offers an overview of Hugging Face Diffusers and explains why this framework is suitable for IR tasks. - (SECTION 3) : Reviews related work, including applications of NLP in cybersecurity, the use of transformers in threat detection, and real-world case studies of AI-driven IR. - (SECTION 4) : Provides background on incident response processes and the motivation for automation. - (SECTION 5) : Presents a case study that applies our approach to a phishing incident scenario, providing results with actual data and a performance evaluation. - (SECTION 6) Discusses the impact of such automation on IR effectiveness, its current limitations, and the associated ethical considerations. - (SECTION 7) Outlines future directions, including enhancing transformer-based solutions, incorporating multi-modal data, and expanding AI’s role in cybersecurity in a responsible manner. We conclude in - (SECTION 8) With a summary of findings and a call to action for adopting AI-driven incident response. Finally, - (SECTION 9) Acknowledges the collaborative and research support that enabled this work.

Maintaining natural streamflow variability along the drainage network is crucial for preserving ecological integrity, species diversity, and their abundance. Unplanned reservoir releases and climate change can alter the natural streamflow regime. A systematic assessment of streamflow alterations under reservoir operations and climate change scenarios along the river, across varying spatial scales, can offer crucial insights into the ecological vulnerability of the region. In this study, we analysed alterations to the natural streamflow regime driven by reservoir releases and an extreme future climate scenario (SSP5-8.5) in three nested basins of the Western Ghats, a recognised biodiversity hotspot in India. The objective was to assess whether reservoir operations or extreme climate conditions have a greater influence on natural flows and how the impacts vary across spatial scales. Our results indicate that reservoir operations after 2015 have substantially disrupted the natural streamflow regime, with effects that were largely consistent across all study basins. In contrast, under the projected extreme climate change scenario, alterations were comparatively lower and exhibited greater variability across basin scales. The largest basin showed minimal streamflow alteration, suggesting a higher resilience of larger basins to future climate change. Overall, our findings highlight that the reservoir operations in the study region need to be immediately regulated as they pose a threat to the natural streamflow regime, higher than that under the extreme climate change scenario.

Current treatment of patients with asthma or chronic obstructive pulmonary disease (COPD) predominantly involves the use of inhaled bronchodilators (B2 agonists and muscarinic receptor antagonists)(1) and anti-inflammatory corticosteroids (2). Often these drugs are used in fixed dose combination inhalers, either as dual inhibitors (1) or more recently so called “triple inhalers” (1,3). Whilst in many patients this approach is effective in relieving symptoms and providing maintenance treatment, it is also recognised that many patients remain symptomatic despite such treatment (4). Furthermore, there is ongoing concern about the side effect profile of inhaled corticosteroids, particularly when used in patients with COPD where there is an increased risk of infection (5) and in paediatric patients with asthma (6). Moreover, there are now multiple combinations of bronchodilators and inhaled corticosteroids licenced as medicines and they are delivered in a wide range of inhaler devices which can be bewildering to patients and affect adherence to treatment (1,4).

The rapid rate of climate change is shifting environmental conditions towards species thermal limits, at a pace that many species cannot match via genetic adaptation. Phenotypic plasticity can provide species with a rapid response to buffer the negative impacts of warming temperatures, however, not all populations may have the same potential for plasticity. How thermal sensitivity and plastic capacity differs across populations will determine resilience to future warming and heatwave events. Here, we investigate intraspecific variation in phenotypic plasticity in a tropical reef fish, the spiny chromis damselfish (Acanthochromis polyacanthus), to understand how low and high latitude populations differ in their response to ocean warming. To test plastic potential in low and high latitude populations, juveniles were exposed to developmental temperatures of 28.5 °C, 30 °C, and 31.5 °C, and investigated for differences in morphology (standard length and mass), critical thermal maximum, and oxygen uptake. At warmer temperatures fish grew to smaller sizes and displayed only marginal increases in critical thermal maximums (+0.16 °C). No differences were observed in critical upper temperatures, that represent a transition where cellular damage begins to outpace cellular repair, between developmental treatments or latitudes. Critical temperature thresholds (~33.92 °C) were identified as being ~3 °C below critical thermal maximums, and only ~2.19 °C above mean sea surface temperatures during recent heatwave events on the Great Barrier Reef. Thermal death time models found a > 17-day survival advantage at 33 °C, and ~1.5-day advantage at 34 °C, for fish that developed at 31.5 °C from both latitudes, revealing some potential beneficial plasticity. These results demonstrate that low and high latitude populations are expected to have similar capacities to respond to future warming that are limited in their ability to match the pace of climate change through phenotypic plasticity.

In times of species decline and biodiversity loss, large-scale, multi-species surveys are essential for effective conservation planning. Camera traps have become indispensable for wildlife monitoring, providing detailed insights into species abundance and distribution. However, manual processing of large camera trap datasets remains time-consuming and resource-intensive. Although machine learning algorithms offer a scalable solution, their performance on data collected in densely vegetated tropical forests remains to be validated. Here, we evaluated machine learning algorithms for species classification, using 5,886 videos and 1,369 images from 55 camera traps deployed in the rainforest near Salonga National Park, Democratic Republic of the Congo, between September 2023 and March 2024. We compared performance metrics across species and algorithms and assessed how factors such as model confidence, distance, and body size influenced accuracies. Overall classification accuracies were lower than officially reported but comparable across region-specific algorithms. Specialist algorithms outperformed a generalist algorithm for classifications on the species level. Performance varied strongly between species, particularly for rare and small-bodied animals. Labelling confidence-based thresholding improved accuracies but resulted in substantial loss of data. Algorithmic confidence score was the strongest predictor of classification accuracy in region-specific algorithms, while distance negatively affected detection and correct classifications. Our findings underscore the importance of critical validation when applying machine learning algorithms, especially in challenging environments such as tropical forests. Machine learning performance is highly location- and species-dependent, with errors potentially propagating to derived ecological metrics. We outline practical steps to integrate manual validation into machine learning workflows to help practitioners avoid pitfalls and ensure robust and rigorous wildlife monitoring that strengthens conservation efforts.

Amit Bansal1,*, Camilla Tvedt Ekanger1, Elin Hoffmann Dahl1, Bjørn Blomberg1,2,*,1Department of Clinical Science, Faculty of Medicine, University of Bergen, Bergen, Norway;2National Centre for Tropical Infectious Diseases, Haukeland University Hospital, Bergen, Norway;*Corresponding authors:Dr. Amit Bansal (e-mail: amit.bansal@uib.no, ORCID ID: https://orcid.org/0000-0002-0681-932X),Prof. Bjørn Blomberg (e-mail: Bjorn.Blomberg@uib.no, ORCID ID: https://orcid.org/0000-0001-5647-4297 ),Department of Clinical Science,Faculty of Medicine,University of Bergen,Bergen, Norway

We describe a synergetic photoredox and CO 2 •− catalysis for a regioselective hydroarylation of 1-alkenyl carbonyls with cyano(hetero)aromatics via radical anion sorting. Using the carbon dioxide radical anion (CO 2 •−) as a powerful single electron reductant cooperated with the photoreductive catalysis, this protocol enables the simultaneous formation of dual radical anions through single electron reduction of alkenes and cyano(hetero)aromatics, respectively, followed by radical anion-radical anion coupling to achieve hydroarylation leading to β-(hetero)aryl alkyl carbonyl compounds in a highly regioselective manner. This hydroarylation reaction proceeds under mild conditions, exhibits broad functional-group compatibility and accommodates a wide range of alkenes and cyano(hetero)aromatics with exquisite regioselectivity.

Addressing the demand for clean heating transition in northern rural areas, this study tackles the challenges of coordinating source-load bilateral uncertainty from high-proportion electric heating and low-carbon operation. It constructs an optimal scheduling model for a rural IES incorporating P2G and CCS coupling. Environmental costs are quantified via a tiered carbon trading mechanism, while carbon resource recycling and methane production are achieved through CCS-P2G coupling. On this basis, a two-stage "day-ahead robust optimization - intra-day rolling adjustment" strategy is proposed. The day-ahead layer employs robust optimization to handle uncertainties in wind and PV power fluctuations, while the intra-day layer performs minute-level rolling corrections based on MPC. Finally, case simulations demonstrate that the proposed electric heating model incorporating tiered carbon trading and carbon capture effectively guides low-carbon operation, providing theoretical support and methodological reference for the planning and operation of rural clean energy systems.

Sustainable hygroelectricity generation has emerged as a revolutionary technology capable of harvesting the latent heat from ambient environments and converting it into direct-current electricity. Recently, microbial biofilms have been regarded as promising hydrovoltaic materials for their low-cost, facile fabrication and environmentally friendly nature. However, external moisture dependence and ultra-high internal resistance have greatly limited the output performance of microbial biofilm-based hydrovoltaic electricity generators (BioHEGs). Herein, a core@shell structured biohybrid system (S. o@PPy) was constructed by in situ polymerization of pyrrole monomers on the surface of Shewanella oneidensis (S. o) cells. The presence of polymer-biofilm interfaces resulted in an exceptional performance of hygroelectricity generation originated from the built-in conductivity of hole-doped polypyrrole (PPy), as well the reduced internal resistance and improved water adsorption capacity of S. o biofilms. Hence, the S. o@PPy BioHEG generated a stable short-circuit current of ca. 33.5 μA at the optimal condition, which is significantly higher than all BioHEGs documented hitherto. Moreover, a novel mechanism of asymmetric charge redistribution was confirmed by density function theory (DFT) calculations. These results support a new perspective of hydrovoltaic effects and provide a viable strategy for advancing BioHEGs towards more practical scenarios.