Monitoring cetacean populations is essential for understanding the status of marine communities. In this regard, environmental DNA (eDNA) metabarcoding is an increasingly popular innovative and non-invasive genetic tool for marine biodiversity monitoring, particularly useful for assessing the populations of highly mobile species with a wide distribution range, such as cetaceans. In this study, we compared cetacean species detection and distribution patterns obtained through eDNA metabarcoding with the sighting records obtained during simultaneous traditional visual monitoring. A total of 258 seawater samples was analysed, collected at 129 sampling stations during the SCANS-IV survey carried out in oceanic waters of the North Atlantic Ocean. Nine cetacean species were identified by eDNA metabarcoding and ten by visual monitoring; three species were detected solely by eDNA, while four were exclusively detected by visual observation, demonstrating the complementarity of the two methods. The four most frequently sighted species were also the most frequently detected species in eDNA samples, showing similar detection frequencies and distribution patterns at broad and intermediate spatial scales. At a finer scale, however, species occurrence did not overlap between methods. Additionally, eDNA provided new insights into striped dolphin (Stenella coeruleoalba), which may be under-represented in traditional monitoring data. Overall, the results support the use of eDNA metabarcoding as a valuable complementary monitoring tool for assessing cetacean diversity and distribution, providing additional information on species presence throughout the study area. This is particularly relevant for species that are rarely recorded visually and for which there is insufficient information to design and implement appropriate management strategies.

Male mammals are of particular interest for molecular systematics as their cells contain two non-recombinant markers, the mitochondrial genome and the male-specific Y chromosome (MSY), which provide information on maternal and paternal evolutionary histories, respectively. Here, we assembled four single-copy MSY genes (AMELY, DDX3Y, SRY and ZFY), three homologs on theXchromosome, the mitogenome, and 21 autosomal introns from whole genome sequencing (WGS) data available for 123 male giraffes. We detected several instances of introgression between giraffe subspecies involving the mitogenome, MSY, and X-linked genes. The analysis of MSY haplotypes supports a deep separation in Africa between northern giraffes (subspecies antiquorum, peralta, rothschildi, and reticulata) and southern giraffes, with a large gap between intragroup and intergroup DNA distances (referred to as the 'MSY barcoding gap'). At a finer scale, southern giraffes can be divided into two geographic MSY groups that are distributed in East Africa (comprising the subspecies tippelskirchi and thornicrofti) and southern Africa (comprising the subspecies giraffa and wardi). These relationships are all supported by several exclusive synapomorphies in most DNA datasets. Our results provide strong support for two species of Giraffa, i.e., G. camelopardalis (northern giraffes) and G. giraffa (southern giraffes), but with ZFX alleles showing evidence of ancient introgression between the two taxa. The delimitation of Giraffa in two species is consistent with skull morphology and the evolution of highly distinctive phenotypes (reticulated versus Masai) in the hybrid zone between northern and southern species in southern Kenya which may have promoted the reinforcement of prezygotic isolation, thus limiting gene flow between them.

Microbial eukaryotes play a vital role in biogeochemical cycles and aquatic food webs. Over past decades, their taxonomic diversity has been investigated using short-read DNA metabarcoding, which, while effective, is hampered by the limited resolution of the targeted regions. In this study, we utilised both short-read (Illumina) and long-read (PacBio) metabarcoding approaches to analyse eukaryotic diversity in a lake ecosystem over one year, allowing for a direct comparison of the results obtained from each method. Our metabarcoding analysis revealed a high degree of congruence between long-read and short-read data at broad taxonomic levels. However, at finer taxonomic scales, such as the genus level, long-read sequencing achieved higher resolution, enabling more precise identification of microbial eukaryotes. This enhanced taxonomic resolution proved especially valuable for tracking seasonal dynamics within key groups, including ciliates and chytrid fungi. The resolution offered by long-read sequencing provides a more detailed picture of the microbial eukaryotic community, thereby facilitating the exploration of ecological interactions at a finer scale. For example, this approach allowed us to monitor the dynamics of the chytrid genus Zygorhizidium in relation to various diatom genera, which are primary targets for chytrid parasitism.

Agricultural intensification and rising insecticide use are known threats to biodiversity in agroecosystems. Using a series of agricultural and landscape metrics, we tested the hypothesis that barn swallows (Hirundo rustica), as an avian wildlife indicator species, are at greater risk of insecticide exposure in more agriculturally intensive landscapes. We used spatial datasets for insecticide use density, landscape simplification, and relative cover of row crops and aquatic habitats (wetlands and other water bodies) at multiple spatial scales (500m-20km) as proxies of agro-intensity. In 2022 and 2023, we captured and blood sampled 173 barn swallows from 54 farms across a large agricultural gradient in southern Saskatchewan, Canada, and collected invertebrate sweep samples from three habitat types (cropland, grass or pastureland, and wetland margins) on 37 farms for chemical analysis to determine detection prevalence and sum concentrations of seven systemic insecticides (imidacloprid, clothianidin, thiamethoxam, acetamiprid, thiacloprid, flonicamid, and dinotefuran). The neonicotinoids thiamethoxam, clothianidin, and imidacloprid were detected in 54%, 47%, and 27% of the 173 barn swallow plasma samples and 98% of the 120 invertebrate samples. Insecticide concentration and detection in swallow plasma declined over the season but showed no clear or strong relationship with the measured metrics of agricultural intensity. However, the detection or concentration of insecticides in invertebrate samples positively correlated with most agricultural metrics. These results are consistent with the foraging behavior of barn swallows, whose movement across several kilometers likely integrates variation in prey contamination across farmland. The lack of association of agricultural intensity with barn swallow plasma neonicotinoid levels and positive relationship with invertebrates indicates that either rapid metabolism of these compounds in birds prevents accurate exposure characterization using blood, or land-use intensity maps are insufficiently precise at the relevant spatial scales for a wide-ranging species, and/or lower detection frequency of insecticides precludes characterization of fine-scale insecticide risk. We advocate for finer scale spatial data on agricultural intensity by crop type and pesticide use and recommend broader invertebrate sampling to more accurately reflect pesticide exposure risk for wildlife.

Abstract Background Weather and fuels are among the critical, interacting factors that drive wildland fire behavior, and thus are primary factors in fire operations planning and decision support tools. In mesic forests, variation in stand structure may lead to heterogeneous microclimate and fuel moisture conditions in the understory where fires often ignite and spread. However, such variation in fuel availability is often overlooked by fire behavior models that assume spatially uniform weather and fuel conditions. In this study, we analyze a combination of field-based understory meteorology and fuel moisture data with terrestrial laser scanning (TLS) and traditional forest inventory data to develop new knowledge about relationships between forest structure, microclimate, and dead fuel moisture in USA northern conifer forests that can inform fire operations planning and decision support. Results We found that open canopy plots were significantly warmer (+ 7.82 °C average daily maximum air temperature) and drier (-24.1% average daily relative humidity) and with drier fuels (+ 8.03% fuel moisture) at midday in summer compared to closed canopy plots. Directly using microclimate variables (i.e., air temperature and relative humidity) resulted in better predictions of dead fuel moisture content (mean R 2 = 0.88) than using forest structure variables such as canopy openness (R 2 = 0.60). Furthermore, forest structure variables derived from TLS were better predictors of dead fuel moisture content (R 2 = 0.74) than traditional forest inventory metrics. Conclusions Our study used multi-modal measurements to demonstrate that dense forest cover linearly reduces fuel availability by buffering microclimate and maintaining fuel moisture. This research can be used to develop thinning prescriptions to achieve certain thresholds of understory temperature, relative humidity, and dead fuel moisture. Moreover, our results highlight the microclimate buffering effect of shaded fuel breaks used in fire suppression and containment tactics. Finally, our work suggests that tools like TLS can be used to fine tune fuel-weather relationships in fire behavior models that use spatially explicit fuels data to inform planning and predict fuels treatment effectiveness. This research enhances fire managers’ ability to plan and implement fuel treatments by highlighting how changes in forest stand structure drive fine scale heterogeneity in fuel availability.

ABSTRACT This study presents the first continent‐wide assessment of long‐term (1974–2023) summer heat stress in Africa using the Universal Thermal Climate Index (UTCI) derived from ERA5‐HEAT reanalysis datasets. Employing a grid‐specific definition of summer, we analysed UTCI trends across spatial (continental to city scale) and temporal (decadal to hourly) dimensions. While annual UTCI anomalies range 0.2°C–1.6°C, substantial intensification emerges at finer scales. The frequency, duration and spatial extent of ‘very strong’ (38°C–46°C) and ‘extreme’ (≥ 46°C) heat stress have increased markedly, particularly, in the Sahel. Heat‐stressed hours (UTCI > 32°C) rose by 2%–25% across most countries, with extreme events doubling in 3–5 and 6–8 days episodes. City‐level analyses reveal escalating risks in large urban centres like Cairo, Lagos and Kano. As climate projections indicate further intensification, the findings underscore the urgent need for targeted heat‐health adaptation and early warning systems to protect vulnerable populations across the continent.

This study investigates stochastic income convergence across Italian regions at two spatial scales—NUTS-2 (regions) and NUTS-3 (provinces)—in a context of an affluent North and the less-developed South. Using dynamic panel data models, both non-spatial and spatial, for 2001–2021, the analysis explores the roles of persistence, unobserved regional heterogeneity, and spatial dependence in income gap dynamics. Models without fixed effects show near-unit-root behavior, indicating high persistence or divergence, whereas including fixed effects substantially reduces persistence, providing evidence of conditional stochastic convergence at both spatial scales. Spatial dependence is limited at the NUTS-2 level, with coefficients generally small and insignificant, but at the NUTS-3 level, spatial interactions are economically meaningful and significant, highlighting localized spillovers in provincial income disparities. Ignoring these interactions at the finer scale can overstate persistence. To correct for Nickell bias in dynamic fixed-effects models, System GMM estimators are employed. Spatial System GMM results show strong but stable persistence, with coefficients around 0.86–0.87. The implied half-life of income shocks is roughly five years at the regional level and four to four-and-a-half years at the provincial level, indicating faster adjustments locally. Overall, the study finds weak but robust conditional stochastic convergence, emphasizing the importance of accounting for both spatial dependence and unobserved heterogeneity in understanding regional income dynamics in Italy.Unlike previous studies, our results quantify the persistence and half-life of income shocks, highlighting the difference between NUTS-2 and NUTS-3

Air pollution remains a critical environmental and public health threat, particularly in highly populated urban areas such as the Lisbon Metropolitan Area (LMA). This study provides a refined and detailed assessment of the spatial distribution of air pollution and associated attributable mortality across the LMA. High-resolution (1 km2) annual mean concentrations of key pollutants (PM2.5, PM10 and NO2) for 2022 and 2023 were estimated by integrating outputs from the URBAIR dispersion model with ground-based monitoring observations using advanced geostatistical data-fusion techniques. Air pollutant concentrations were combined with gridded population data and age-stratified baseline mortality rates within a Geographic Information System framework to quantify spatial variations in health impacts. Using the World Health Organization AirQ+ framework and established concentration–response functions, we estimated a total of 3195 air-pollution-attributable deaths across the Lisbon Metropolitan Area (LMA) in 2022, increasing to 4010 deaths in 2023. Fine particulate matter (PM2.5) was identified as the dominant contributor, accounting for more than 40% of the total health burden. At a high spatial resolution (1 km2 grid), estimated mortality exhibited substantial variability, ranging from 0 to 29 deaths per cell in 2022 and from 0 to 36 deaths per cell in 2023. These results highlight the importance of fine-scale spatial analysis, revealing intra-urban disparities that are not captured by aggregated estimates of total attributable mortality. The proposed methodological framework, integrating dispersion modelling, data fusion, and spatially explicit health impact assessment at fine spatial scales, provides a robust and transferable approach to support evidence-based air quality management and urban health policy development in European metropolitan contexts. This integrated approach enhances comparability, improves exposure assessment accuracy, and strengthens the scientific basis for designing targeted mitigation strategies that could prevent hundreds of premature deaths annually while addressing documented spatial inequalities in pollution exposure.

The symbiosis between clownfishes (or anemonefishes) and their host sea anemones ranks among the most recognizable animal interactions on the planet. Found on coral reef habitats across the Indian and Pacific Oceans, 28 recognized species of clownfishes adaptively radiated from a common ancestor to live obligately with only 10 nominal species of host sea anemones. Are the host sea anemones truly less diverse than clownfishes? Did the symbiosis with clownfishes trigger a reciprocal co-evolutionary response to the mutualism? To address these questions, we combined fine- and broad-scale biogeographic sampling with multiple independent genomic datasets for the bubble-tip sea anemone, Entacmaea quadricolor-the most common clownfish host anemone throughout the Indo-West Pacific. Fine-scale sampling and restriction site associated DNA sequencing (RADseq) throughout the Japanese Archipelago revealed three highly divergent cryptic species: two of which co-occur throughout the Ryukyu Islands and can be differentiated by the clownfish species they host. Remarkably, broader biogeographic sampling and bait-capture sequencing reveals that this pattern is not simply the result of local ecological processes unique to Japan, but part of a deeper evolutionary signal where some species of E. quadricolor serve as host to the generalist clownfish species Amphiprion clarkii and others serve as host to the specialist clownfish A. frenatus. In total, we delimit six cryptic species in E. quadricolor that have diversified within the last five million years. The rapid speciation of E. quadricolor combined with functional ecological and phenotypic differentiation supports the hypothesis that this diversification is an evolutionary response to mutualism with clownfishes. Clownfishes are not merely settling in locally available hosts but recruiting to specialized host lineages with which they have co-evolved. These findings have important implications for understanding how the clownfish-sea anemone symbiosis has evolved and will shape future research agendas on this iconic model system.

How does the human brain respond to novelty? Here, we address this question using functional magnetic resonance imaging data wherein human participants watch the same movie scene four times. On the first viewing, this movie scene is novel, and on later viewings it is not. We find that brain activity is lower-dimensional in response to novelty. At a finer scale, we find that this reduction in the dimensionality of brain activity is the result of increased coupling in specific brain systems, most specifically within and between the control and dorsal attention systems. Additionally, we find that novelty induced an increase in between-subject synchronization of brain activity in the same brain systems. We also find evidence that adaptation to novelty, herein operationalized as the difference between baseline coupling and novelty-response coupling, is related to fluid intelligence. Finally, using separately collected out-of-sample data, we find that the above results may be linked to psychological arousal.

Fire can profoundly affect ecosystem dynamics, species distribution and plant traits, especially in open biomes. Post-fire strategies, namely, resprouters and reseeders, offer a useful framework to examine eco-evolutionary relationships between plants and fire. However, whether resprouter and reseeder plants are consistently formed by distinct trait coordination (syndromes) and responses to fire at the intraspecific level and when considering the role of ontogeny, remain underexplored. This is a relevant lack as, within-species, plants can adjust their functioning and trait coordination can vary considerably along ontogeny. To address this gap, we analysed intraspecific trait coordination and post-fire responses, accounting for the effect of ontogeny in three widely distributed and locally abundant Mediterranean woody species: two resprouters (Erica arborea, Quercus ilex) and one reseeder (Cistus salviifolius). We collected 12 plant functional and architectural traits, including intraspecific variability, well related to fire and drought from three sites in Italy. We ran pairwise correlation and multivariate analyses to explore trait syndromes. We conducted linear regressions to examine relationships between fire regime (time since last fire) and trait responses. We then inspected whether fire regime affects key bivariate trait coordination and if ontogeny influences some trait-fire links. Findings are highly species-specific and generally do not align with a priori classification into post-fire strategies. In most instances, we reveal how either one of the resprouter species exhibits trait patterns more similar to those of the reseeder than to the other resprouter species. Fire can strongly affect trait coordination shaping plant functioning, whereas ontogeny influences a few trait-firelinks for the reseeder species while it has a weak effect on the two resprouter species. Our study, while limited to three species and three sites, emphasizes the importance of looking at plant life through a continuous and multidimensional lens which contemplates the inclusion of various sources of within-species variability. We acknowledge that a category-based or dichotomous view on plant functional strategies, including post-fire ones, remains valid and justified when working at coarse scales, whereas it can be much less so for trait-based analyses at fine scales.

ABSTRACTAsh dieback disease, caused by the pathogenic fungus Hymenoscyphus fraxineus, is now widespread in the UK. The pathogen has caused substantial loss of European ash over recent decades, the long‐term consequence of which is complex to assess. While no higher plant species are exclusively associated with ash, widespread ash mortality is predicted to affect woodland floral biodiversity because of the functional traits uniquely associated with this species: namely the greater light penetration through the canopy, rapid foliar decomposition and nutrient cycling abilities. In woodlands where ash is frequent, vascular plant communities may undergo considerable compositional changes after loss of ash trees. At present little is known regarding how ash trees contribute to plant community and functional diversity at a fine scale across different woodland types. Using data from 1274 survey plots characterized in the Woodland Survey of Great Britain alongside plant functional trait data, we examined the local relationship between ash trees and plant community diversity and functional trait distribution. We show that ash trees are significantly associated with key plant functional traits and diversity indices and that this functional association is not typical of other dominant canopy species in the UK. Specific leaf area, nutrient and pH requirements and community diversity show significant correlations with ash basal area across various woodland types. Additionally, effects of ash appear to interact with soil pH resulting in a greater structural effect of ash upon plant community composition in lower pH soils. These findings support previous suggestions that ash functions as a keystone species with respect to nutrient cycling and plant community composition with potentially stronger influence on lower pH soils. Consideration should be given to these ecological roles when monitoring and addressing the impacts of widespread ash mortality from dieback disease across Europe.

Abstract Morphing skin panels capable of undergoing large deformations are being investigated as a means of reducing fuel burn in next-generation aircraft through drag reduction. One approach to creating these panels is to 3D print Zero Poisson’s ratio cellular cores onto elastomeric facesheets, as proposed by the Geometrically Anisotropic ThermOplastic Rubber (GATOR) skins concept. Modelling these panels remains challenging due to multiple factors, including small feature sizes, hyperelastic material properties, geometric nonlinearities under large deformations, and complex interactions between the core and facesheets. While high-fidelity finite element analysis works well for unit cells and small components, it becomes prohibitively expensive for large-scale aircraft components. Hence, this work examines various approaches for mitigating the computational cost of analysing these sandwich panels, focusing on a morphing joint fairing for a folding wingtip. The analysis of these panels is simplified through scale separation, where the mechanical responses of the sandwich panel are homogenised into equivalent stiffness properties at the fine scale, allowing the structure to be modelled as a shell surface with equivalent stiffness properties at the coarse scale. This multiscale modelling approach can use homogenisation methods based on analytical formulations and finite element analysis, as well as shell formulations such as Kirchhoff-Love and Reisner-Mindlin plate theory, each with simplifying assumptions that affect the accuracy of the results. This paper examines the effects of these approaches and their underlying assumptions on solution accuracy and computational cost, providing a comprehensive understanding of the modelling approaches available. The results indicate that the analytical homogenisation underpredicts the stiffness of the panel, the effects of transverse shear deformation become negligible for long panels, and the geometric nonlinearity in the panel deformation becomes significant for large rotations of the wingtip. The paper further presents an approach based on data-driven, second-order, multiscale modelling that captures the geometric nonlinearities in GATOR panels’ responses accurately at a reduced computational cost.

Deforestation reshuffles communities across landscapes with myriad consequences for ecosystem function. Following deforestation, rapid exposure to novel microclimates can act as a strong environmental filter, favoring warm-adapted species and decoupling trophic interactions. Forest restoration may partly reverse this process through increased food resources, structural complexity of habitat, and buffering of microclimates-each potentially modified by tree diversity. Despite growing evidence that tree diversity and cool microclimates help maintain animal diversity in natural forests, less is known about how these factors shape species assemblages or multi-trophic dynamics in restoration areas. Here, using surveys and two field experiments within a long-term tree planting experiment, we assessed the relative effects of tree diversity, forest structure, and associated microclimate on fine-scale space use by birds and their top-down impacts on insects. Surveys showed that the probability of occurrences of birds increased in cooler plots, which were associated with higher tree diversity and vertical complexity. The strength of microclimate effects on bird occurrences was strongest for species that are forest specialists. To assess risk to insect herbivores from avian predation, we used a sentinel prey experiment and found that predation risk increased in warmer plots, counter to our expectations based on bird surveys. Last, we examined top-down effects of bird exclusion on leaf herbivory, finding that skeletonizing patterns of herbivory increased in exclosures and in cooler plots. Taken together, these results suggest that microclimate resulting from variation in forest structure shapes the space use of birds at fine scales with complex outcomes for bird-herbivore-tree interactions in planted forests. Active restoration methods that enhance below-canopy cooling may improve biodiversity outcomes and help maintain species interactions that underlie many ecosystem functions.

• Guidelines for parameterising new land covers into process-based ecological models • We tested proxy land covers, new field data, expert opinion, Bayesian calibration • Of these, new floral field data improved our pollinator model predictions most • Bee abundance is reliably predicted in agroforestry systems at fine scale (3 m) • First integration of novel silvoarable alley-cropping land cover into pollinator model As the need for sustainable agroecosystems gains recognition, new land cover classes are increasingly emerging in temperate landscapes. Process-based ecological models are often the most suitable initial option for predicting the biodiversity outcomes of such novel systems, particularly when implementation and large-scale baseline data remain scarce. However, there are no accepted guidelines for integrating new land covers into these models. Using UK silvoarable alley-cropping as a case study, we explore how to introduce this emerging land cover into the established process-based pollinator model, poll4pop. We demonstrate several parameterisation approaches, including proxy land covers, field data, expert opinion and Bayesian calibration. We also provide the first field-scale and seasonally-resolved evaluation of poll4pop, using pollinator abundance data collected at three UK silvoarable sites. Our results show that models using proxy land cover parameters can capture spatial trends in observed bee abundance where suitable proxies exist, but that predictions are improved by integrating field-derived floral cover. Neither bespoke, expert-derived, land cover attractiveness scores nor Bayesian-calibrated scores improved our model fit, although they did reveal valuable insights into model parameter sensitivity. Overall, poll4pop effectively reproduced observed fine-scale spatial variation in bumblebee and spring-flying solitary bee foraging activity in silvoarable systems. However, seasonal differences between communities resulted in reduced model-predictive performance for summer-flying solitary bees. We demonstrate that poll4pop is suitable for modelling fine-scale pollinator abundance in complex mixed-cropping systems. We also present a practical framework for integrating new land cover classes into process-based models which can guide future modelling of emerging land use systems.

Application of a burned area Spatio-Temporal Stratification–Based positive sample augmentation strategy in Fine-Scale wildfire susceptibility mapping

Detecting farmland green production efficiency considering Sustainable Development Goals (SDGs) in the Yellow River Basin: Dynamics, drivers, and challenges

ABSTRACT Aim To conduct a biogeographical regionalization of El Salvador, based on the potential distribution patterns of the superfamily Scarabaeoidea and using Parsimony Analysis of Endemicity. Location El Salvador, Central America. Taxon Superfamily Scarabaeoidea (Insecta: Coleoptera). Methods Potential distribution models were developed for 160 species using Maxent. Based on the resulting potential richness patterns, a Parsimony Analysis of Endemicity was performed using hexagonal grids of 50 km 2 across El Salvador. Clades of cells sharing species compositions (geographical synapomorphies) were identified, and the regions corresponding to each clade were delineated. Results Five clades were identified, corresponding to the following biogeographic regions: (1) Pacific Lowlands; (2) Eastern Interior Valley; (3) Chortis Volcanic Front (Central American Volcanic Arc); (4) Western Interior Valley; (5) Northern Mountains. Main Conclusions Species distribution models, combined with parsimony analyses, strengthen the versatility of both methods for identifying biogeographical patterns at finer spatial scales, facilitate the first species‐based biogeographic regionalization of El Salvador, and demonstrate that micro‐scale analyses can reveal internal structure within broadly defined Central American provinces.

• Small greenspaces can buffer extreme heat stress by up to 16 °C mPET on hot days • Vegetation structural complexity is a key driver of thermal comfort in small parks • Vegetation structure affects mPET directly and indirectly via sky view factor • Mature trees with tall canopies provide the strongest mPET buffering in small parks • Shading drives fine-scale thermal variability Urban greenspaces play an increasingly important role in urban planning and public health, particularly in providing thermal comfort during hot summer days. While the cooling potential of greenspaces generally increases with size, it also strongly depends on vegetation structure, particularly in small greenspaces. Optimizing these spaces for thermal comfort requires a clear understanding of how their vegetation structure shapes local cooling. Based on field measurements in 2024, we modeled the modified Physiologically Equivalent Temperature (mPET) across 12 structurally diverse small greenspaces (< 2 ha) in central Munich, Germany, and in their built surroundings. High-resolution vegetation structure derived from mobile terrestrial laser scanning was combined with sky view factor estimates from hemispherical photographs and micro-meteorological measurements. Using mixed-effects models and structural equation modeling, we assessed both direct and indirect pathways linking vegetation structure, sky openness, shading, and thermal comfort. Our results show that small greenspaces can reduce mPET by up to 16 °C compared to built surroundings. Vegetation structure emerged as a key determinant of this buffering capacity: greenspace plots with tall, multi-layered vegetation provided strong mPET buffering, whereas sparsely vegetated plots offered little to no thermal relief. Mean canopy height was the strongest predictor of mPET buffering, and vegetation structure influenced thermal buffering both directly and indirectly through reductions in sky view factor. At finer spatial scales, shading dominated local thermal variability, with shading effects being strongest in structurally complex plots. These findings highlight that small but structurally complex greenspaces can play a vital role in climate-adaptive urban planning.

Comparing two scales of watershed ecosystem services and their influencing factors: A case study of the Dongjiang River Basin.