Multi-point joint prediction of zone-wise arch dam displacements based on spatial partition and deep ensemble selective regression chain

Understanding how the remarkable phenotypic diversity observed in organisms arises through shifts in macroevolutionary patterns and tempos is a fundamental challenge in evolutionary biology. Phenotypes often evolve in a mosaic pattern during adaptive transitions. For organisms that have invaded highly specialized habitats, such as the unique two-phase interface habitats (water surfaces), the macroevolutionary history of their phenotypic diversification remains only superficially understood. Semiaquatic bugs (Insecta: Heteroptera: Gerromorpha), which exhibit extensive habitat diversification and unparalleled phenotypic innovation, represent one of the most successful extant adaptive groups at this interface and provide an excellent system for study. By analyzing their adaptive transitions and phenotypic macroevolutionary history, we demonstrate that Gerromorpha experienced a single major adaptive transition from land to the water surface. Subsequently, semiaquatic bugs successfully colonized a wide range of distinct water-surface habitats. During this process, phenotypic space was explored under strong constraints and along pronounced mosaic trajectories; i.e., different body regions exhibited markedly distinct patterns of phenotypic space occupation and partitioning, as well as markedly different evolutionary rates. Furthermore, we showed that this mosaic pattern of phenotypic space occupation and rates of exploration had complex and critical effects on the invasion and colonization of water-surface habitats. Our study provides a typical case of macroevolutionary dynamics in species adapted to specialized air-water interface habitats, emphasizing the complex and significant roles of environmental context and functional demands in shaping patterns of phenotypic evolution.

Extreme acidic environments represent natural laboratories for investigating the mechanisms of microbial community assembly, yet the ecological processes structuring these communities remain incompletely understood. Here, we investigate how spatial partitioning, hydrodynamics, and colonization history shape microbial succession in a unique sulfur-rich, acidic river of volcanic origin in northern Patagonia. We combined 16S rRNA gene profiling and shotgun metagenomics with a multi-scale experimental framework encompassing water column fractionation and colonization assays under native and controlled conditions. Microbial diversity was strongly influenced by spatial fractionation, with free-living communities exhibiting higher richness and temporal variability than particle-associated assemblages. Water flow modulated community structure, increasing evenness in free-living fractions under high-flow conditions, but had limited impact on particle-attached communities. Colonization of sulfur-beads followed a structured successional trajectory, with autotrophic sulfur oxidizers dominating early stages and heterotrophs adapted to biofilm lifestyles increasing over time. Ex situ recolonization assays revealed strong priority effects, with initial colonizers determining successional trajectories. Turnover analyses revealed that the balance among stochastic and deterministic assembly processes shifted across communities with pronounced stochasticity in the water column and flow-dependent effects in free-living communities, while biofilm associated communities on sulfur-beads exhibited stronger contribution of deterministic selection. These ecological patterns were mirrored by functional differentiation, with gene enrichment analyses revealing adaptive signatures of substrate attachment and resource acquisition. By integrating fine-scale environmental variation with colonization dynamics, this study reveals how microscale habitat structure and temporal fluxes jointly modulate microbial community assembly rules, offering a nuanced framework to dissect ecological processes in extreme systems.

The efficient removal of ethane (C2H6) and propane (C3H8) from natural gas is vital for purification. A synergistic pore engineering integrating pore space partition and fluorine functionalization in metal-organic frameworks (MOFs), which may effectively promote the C-H···π and C-H···F interactions for effective methane separation. This strategy was validated using two fluorine-functionalized pore-space-partitioned MOFs (SNNU-707/-708) constructed by introducing varying numbers of -CF3 groups on the pore surface. Single-component adsorption isotherms show high adsorption of SNNU-707/-708 for C2H6 and C3H8 were 94.9/63.6 cm3 g-1 and 96.4/68.9 cm3 g-1, significantly exceeding that of CH4 (18.9/13.4 cm3 g-1). Ideal adsorbed solution theory (IAST) indicated high selectivity values of 85.2/116.6 for C3H8/CH4 (50/50) and 16.7/17.0 for C2H6/CH4 (50/50). Notably, the actual breakthrough interval times of SNNU-707 for C3H8/CH4 (5/95) and C2H6/CH4 (10/90) can reach 502 and 78 min·g-1 and yield high-purity CH4 (>99.5%) at 5.89 mmol g-1 from ternary mixtures. Grand Canonical Monte Carlo (GCMC) simulations attribute this performance to synergistic weak interactions (C-H···π, C-H···F, C-H···O/N) between MOF and alkane. Specially, thanks to the fluorine-functionalized pore environments, both MOFs maintain structural integrity and separation performance under harsh conditions up to 98% relative humidity, which is crucial for practical wet natural gas separation.

Double-strand breaks (DSBs) threaten genomic stability and need immediate attention from DNA damage response (DDR) machinery involved in homologous recombination (HR) or nonhomologous end joining (NHEJ). DDR in heterochromatin is challenging owing to the distinct chromatin organization. Heterochromatin protein 1 (HP1) isoforms are central to heterochromatin structure and have been implicated in DDR. Mammalian HP1 has three isoforms, HP1α, HP1β, and HP1γ, which possess significant homology and yet have distinct functions. HP1α is the only isoform known to undergo liquid-liquid phase separation mediated by phosphorylation on the N-terminal extension (NTE). We show that the minute-scale dynamics of HP1α and HP1β differ dramatically and differentially influence the recruitment of HR vs. NHEJ factors at sites of laser-induced clustered DSBs. Perturbing HP1α phosphorylation impairs HR factor recruitment and reduces HR efficiency. Our study provides a potential link between phase separation and DDR-centric roles of HP1α and hints at spatial partitioning of repair pathways in response to damage in heterochromatin.

A hybrid membrane bioreactor (HMBR) enhances treatment performance by simultaneously utilizing organisms on both suspended and attached sludge, yet the microbial mechanisms underpinning their efficiency remain poorly understood. In this study, we investigate spatial variability within microbial communities in HMBRs and correlate this factor with pollutant removal capacity. High-throughput sequencing results revealed significant differences in community structure between suspended sludge, suspended media surfaces, and membrane module surfaces. Suspended sludge exhibited the highest species richness, whereas microbial communities on suspended media resembled those within the sludge, contrasting markedly with membrane surface communities. Key functional groups were enriched at specific locations: Pseudomonas and Comamonas dominate the surface of the suspension culture medium and participate in nitrification; phosphorus-accumulating organisms (PAOs), primarily from the Flavobacteriales and Planctomycetaceae phyla, were most abundant on suspended media surfaces. This spatial partitioning of functional microbes indicates cooperative division of labor. Media surfaces serve as primary sites for nitrification and phosphorus removal, whilst suspended sludge flocs and membrane module surfaces are the principal contributors to denitrification. The results of this study provide microbiological evidence for optimizing HMBR design and operation, confirming that spatial community structure is a key factor influencing performance.

Abstract Let $$V=V(d,q)$$ V = V ( d , q ) denote the vector space of dimension d over $${\mathbb F}_q$$ F q . A subspace partition $$\mathcal {P}$$ P of V , also known as a vector space partition , is a collection of nonempty subspaces of V such that each nonzero vector of V is in exactly one subspace of $$\mathcal {P}$$ P . Motivated by applications of minimum blocking sets and maximal partial t-spreads , Beutelspacher (Geom Dedic 9:425–449, 1980) determined in a lemma the minimum possible size $$\delta (d)$$ δ ( d ) over all (nontrivial) subspace partitions of V . In Heden et al. (Des Codes Cryptogr 64:265–274, 2012) and Năstase and Sissokho (Linear Algebra Appl 435:1213–1221, 2011), we extended Beutelspacher’s Lemma by determining the (first) minimum size $$\sigma _q(d,t)$$ σ q ( d , t ) of any subspace partition of V for which the largest subspace has dimension t , with $$1\le t<d$$ 1 ≤ t < d . In this paper, we build on the previous results and unveil additional structural information of subspace partitions. We determine the second minimum size $$\delta '(d)$$ δ ′ ( d ) over all (nontrivial) subspace partitions of V and furthermore, for $$d\equiv r \pmod {t}$$ d ≡ r ( mod t ) and $$0\le r<t<d$$ 0 ≤ r < t < d , we prove the exact value of the second minimum size $$\sigma _q'(d,t)$$ σ q ′ ( d , t ) of any subspace partition of V for which the largest subspace has dimension t and when at least one of the following holds: (i) $$r=0$$ r = 0 , (ii) $$t+r$$ t + r is even, (iii) $$d<2t$$ d < 2 t or (iv) the partition has only subspaces of two different dimensions. Finally, applications to the supertail of a subspace partition and the size of maximal partial spreads are given.

ABSTRACTThe subiculum is a main output structure of the hippocampus, transmitting information from the CA1 to the entorhinal cortex in a spatially structured manner. Prior studies revealed a high electrophysiological and molecular heterogeneity of subicular pyramidal neurons (PYNs) with evidence for further spatial subdivisions of the subiculum. In this study we focused on the cellular organization of the proximal to mid‐distal part of the subiculum, designated as subregion 2 (Sub2) and performed a comprehensive electrophysiological and morphological characterization of PYNs by whole‐cell patch‐clamp recordings combined with intracellular labeling in acute rat hippocampal brain slices. Principal component analysis based on discharge pattern‐related parameters and subsequent unsupervised, hierarchical clustering classified the PYNs into three subtypes: regular firing (RF), weak‐bursting (WB), and strong‐bursting (SB) neurons. Electrophysiological analysis revealed further differences between RF neurons and the two subtypes of bursting neurons in their active and passive properties. The three subtypes also showed differences in their morphometric features, including the apical and basal dendritic spread and branching pattern. Additionally, we identified a divergent morphological subset among RF neurons, bearing two apical dendrites. Mapping the three PYN subtypes onto the subiculum revealed specific spatial distributions along the superficial‐deep and proximo‐distal axes. Thus, this work maps the heterogeneity of subicular PYN onto differentially distributed subclasses in the Sub2 region with distinct physiological and morphological features. These findings together with prior observations of divergent anatomical projections from subicular subregions are pivotal for the understanding of how subicular morpho‐electric neuron types relate to each other and contribute to processing and distribution of information to cortical regions from this hippocampal output region.

In this research, an innovative scheme to generate heterogeneous acoustofluidic distributions in various pseudo-Sierpiński-carpet-shaped chambers with different filling fractions and cross-sectional configurations has been proposed and calculated for topographical manipulation of large-scale micro-particles. All of the structural components positioned in the pseudo-fractal chambers are symmetrically distributed in space, and all ultrasonic radiation surfaces hold the unified settings of input frequency point, oscillation amplitude, and initial phase distribution along their respective normal directions. A large number of fascinating acoustofluidic patterns can be generated in the originally-static pseudo-Sierpiński-carpet-shaped chambers at different recursion levels without complicated vibration parameter modulation. The simulation results of acoustofluidic distributions and particle motion trajectories under different radiation surface arrangements further demonstrate the manipulation performance of these specially designed devices, and indicate that controllable spatial partitioning and intensity modulation of the acoustofluidic field can be achieved by adjusting the hierarchical order, cross-sectional configuration and combination mode of the radiation surfaces. Unlike the existing device construction method of miniaturized microfluidic systems, the artificial introduction of fractal elements like Sierpiński carpet/triangle, Koch snowflake, Mandelbrot set, Pythagoras tree, etc., can provide extraordinary perspectives and expand the application range of the acoustofluidic effect, which also makes ultrasonic micro/nano-scale manipulation technology more abundant and diversified. This exploratory research indicates the potential possibility of applying fractal structures as alternative component parts to purposefully customize acoustofluidic distributions for the further research of patterned manipulation of bio-organisms and navigation of micro-robot swarms in brand new ways that cannot be achieved through traditional methods.

ABSTRACT Anomaly detection on the Earth’s surface plays a pivotal role in understanding dynamic environmental changes, detecting natural disasters, and supporting timely decision-making. While remote sensing imagery has been widely adopted for such tasks, existing methods often suffer from limited adaptability to heterogeneous landscapes and poor robustness to temporal noise or data gaps. To address these challenges, we propose the Spatio-temporal Adaptive Grid for Surface Anomaly Detection (STAGA), a novel two-stage framework that combines adaptive spatial partitioning, unsupervised temporal modeling, and spatially aware inference to improve both accuracy and operational utility of surface anomaly detection. In the offline stage, STAGA first performs geo-adaptive partitioning, segmenting the Earth’s surface into grid units, and represents each with compact multi-temporal features from spectral bands, remote sensing indices, spatial gradients, and land cover compositions. To model the normal surface evolution patterns, we train lightweight autoencoders per grid unit to learn smooth and denoised temporal trends, forming a Geoscientific Knowledge Grid (GKG). In the online stage, new observations are compared against GKG to compute standardized residuals, which are then refined via a Conditional Random Field (CRF) that fuses spatial context and residual similarity for anomaly inference. We validate STAGA across four representative regions: Los Angeles, Mandalay Poyang Lake, and Gaza, using over five years of Sentinel-2 imagery. The results show that STAGA achieves high detection accuracy (recall up to 93%). Its adaptive partitioning and unsupervised temporal modeling significantly outperform traditional fixed-grid or threshold-based methods. STAGA provides a robust and scalable solution for real-world surface anomaly monitoring.

Turbulence prediction is essential in engineering disciplines such as aerospace and energy, where precise modelling of anisotropic boundary layers and multiscale vortex structures improves design efficiency and safety. Conventional neural networks typically employ data-driven methods, which lack physical interpretability and perform poorly in complex flows. This paper presents Anisotropic and Multiscale Cross-Attention Physics-Inspired Neural Networks(AMSPINN), a novel architecture that incorporates turbulence physics as prior knowledge to enhance accuracy and scalability. The architecture integrates domain-specific features, including spatial partitioning for computational efficiency, anisotropic multi-head attention to capture directional dependencies, and a multiscale information block for hierarchical feature fusion, thereby bridging data-driven machine learning with fundamental turbulence physics. In evaluations on standard benchmarks and Johns Hopkins Turbulence Database(JHTDB), AMSPINN consistently outperforms state-of-the-art models such as Transolver. Across four benchmark datasets, it achieves error reductions ranging from 3.8% to 64%. On the JHTDB, AMSPINN demonstrates relative error reductions of 27.8%, 18.8%, and 37.2% for boundary-layer flow prediction, multiscale vortex prediction, and realistic turbulent flow prediction, respectively. This physics-inspired approach bridges data-driven and theoretical methods, offering promising advancements in turbulent flow simulations.

ABSTRACT Timely, detailed information on forest composition is essential for effective management, biodiversity protection, and understanding ecosystem dynamics. This study maps the distribution of seven dominant tree species in Swedish forests and produces spatially explicit, pixel-level estimates of classification uncertainty. The mapping framework integrates multitemporal Sentinel-1 radar and Sentinel-2 optical observations with field data from the Swedish National Forest Inventory and auxiliary predictors describing topography and canopy height. We trained a Bayesian-optimized extreme gradient boosting model on spatiotemporal metrics derived from these datasets and quantified classification confidence through entropy computed from the class-probability outputs. We applied a spatial block partitioning approach to limit the effects of spatial autocorrelation between optimization and validation data and ensure a more realistic assessment of the model’s generalization capacity. Model overall accuracy reached 85% (F1 = 0.82) using a 60 m spatial block validation. Under a more conservative 200 m block configuration, performance decreased to F1 = 0.63, reflecting reduced training data availability. The county-level species coverage derived from the classification aligned closely with published figures from the Swedish Forest Agency (Spearman’s ρ = 0.94, 95% CI: 0.89 – 0.96, p < 0.001). Variable importance analysis showed that Sentinel-2 spectral bands, particularly shortwave-infrared and red-edge captured during spring and summer, contributed most to species discrimination, while Sentinel-1 backscatter provided complementary structural information. The integration of forest inventory data, Earth observation, and machine learning to produce tree species maps and a spatially explicit measure of prediction uncertainty yields a robust and reproducible framework for large-area forest mapping. The results provide detailed, spatially continuous information on species composition along with an accompanying confidence surface. This offers practical value for ecological assessments, regional planning, and emerging legislative and environmental goals. The data are freely available for download and the maps can be interactively visualized using this link: https://ee-treespec.projects.earthengine.app/view/treespec.

Data-parallel ray tracing is a crucial technique for rendering large-scale scenes that exceed the memory capacity of a single compute node. It partitions scene data across multiple nodes and accesses remote data through inter-node communication. However, the resulting communication overhead remains a significant bottleneck for practical performance. Existing approaches mitigate this bottleneck by enhancing data locality through dynamic scheduling during rendering, typically employing spatial partitioning to enable access prediction. Although effective in some scenarios, these methods incur significant redundancy in base geometry when applied to large-scale instanced scenes. In this paper, we introduce the first object-space-based dynamic scheduling algorithm, which uses object groups as the scheduling units to eliminate redundant storage of base data in instanced scenes. Additionally, we propose two data access frequency prediction methods to guide asynchronous data prefetching, enhancing rendering efficiency. Compared to the state-of-the-art method, our approach achieves an average rendering speedup of 77.6%, with a maximum improvement of up to 146.1%, while incurring only a 5% increase in scene memory consumption.

A partition method for bounding continuous-time Markov chain models of general reaction network.

ABSTRACT Urban Underground Space (UUS) has been recognized as significantly important for the future sustainability of metropolitan areas. A fine, accurate, and efficient suitability evaluation has become a prerequisite and necessary means for the effective and rational development of UUS. As the evaluation methods transition from two‐dimensional to three‐dimensional and evolve towards greater refinement in three‐dimensional scales, the existing evaluation methods based on Grid + Voxel cells struggle to capture the scale differences of various geological factors and lack the requisite precision to represent the constraints of terrain, water bodies, buildings, faults, and karst features. In response to this challenge, this paper proposes a spatial partitioning method based on an improved Generalized Tri‐prism model for UUS Evaluation (GTP‐UE), which is utilized as the basic cell. The GTP‐UE model improves upon the limitations of the general GTP model in data structure while retaining its wide geometric applicability and strong maintenance scalability, allowing for simpler multi‐element recording, more efficient topology calculations, and faster collision detection, making it more suitable for UUS evaluation. This study first proposes the data organization format of GTP‐UE. Next, an optimisation algorithm for 3D Boolean operations based on the GTP‐UE model is presented. Finally, a case study is conducted in a specific area of Nanjing to verify and discuss the feasibility and advantages of the proposed technical approach. The results indicate that, in comparison to Grid + Voxel cells, the new evaluation method for UUS development based on the GTP‐UE model effectively considers the true representation of geological bodies and environmental characteristics, showing significant advantages in both computational efficiency and storage space during model generation. Evaluating the applicability of other irregular volumetric models in UUS evaluation and enhancing the Boolean operation efficiency in geometric models during multi‐indicator fusion are two worthwhile directions for future research.

ABSTRACT Understanding how gas mixtures diffuse and distribute within porous frameworks is central to designing advanced separation and storage materials. Here, the transport and spatial distribution of binary gas mixtures in a porous metal‐organic framework, viz., ZIF‐90, using molecular simulations is investigated. We performed grand canonical Monte Carlo (GCMC) simulations to examine the competitive adsorption of carbon dioxide (CO 2 ) and nitrogen (N 2 ) from a binary gas mixture in ZIF‐90, while molecular dynamics (MD) simulations are conducted to investigate the transport behavior of the adsorbed molecules within the framework. These integrated simulations reveal that the framework topology and pore chemistry jointly dictate diffusion pathways and preferential occupancy of gas species, underscoring their intrinsic interdependence. Competitive adsorption leads to distinct spatial partitioning within the pores, which in turn modulates mixture diffusivity inside the porous medium compared to their bulk properties. These results provide molecular‐level insight into how ZIF‐90 accommodates and separates gas mixtures, offering design principles for optimizing metal‐organic frameworks in energy and environmental applications.

Microbial generalists and specialists in permafrost demonstrate disparate responses to extreme environments.

A more appropriate framework for graphical attribution of hydrological change in the water-energy partitioning space

Species coexistence is shaped by how individuals share limiting resources such as space, food, and shelter. Theory predicts that niche differentiation promotes coexistence, depending on habitat characteristics, behavioral traits, and the intensity of competition. Niche usemay vary across life stages, as ontogenetic shifts alter habitat use and species interactions. We investigated habitat segregation, niche overlap, and the role of non-crop vegetation in affecting coccinellid coexistence across life stages through semi-controlledexperiments and field sampling at 42 sites. Species showed consistent differences in habitat and microhabitat use across developmental stages, influenced by innate behaviors and plastic responses to interspecific interactions. Superior competitors (Hippodamia convergens and Harmonia axyridis) generally dominated prey-rich crop areas during egg, larval, and adult stages, but shifted to sheltered sites outside the plants during pupation. In contrast, Eriopis connexa used soil microhabitats throughout its life cycle, reducing niche overlap with other species. Cycloneda sanguinea, a competitively inferior species, persisted by exploiting non-crop plants, which increased spatial heterogeneity and resource availability. Non-crop vegetation thus promoted coexistence by enabling spatial segregation and reducing presumed competitive asymmetries across life stages. These patterns highlight how the strength of species interactions and spatial partitioning changes ontogenetically, reflecting both behavioral flexibility and the influence of habitat features. We propose a plastic functional classification of species based on their behavioral responses to potential competition-risk scenarios across life stages, ranging from risk-tolerant generalists to conditional risk-avoiders and niche-fidelity strategists, that could be broadened and adapted to other study systems.

The Impact of Label Space Partitioning in Multi-label Code Smell Detection