Journal of Ecology

Coflowering, the temporal overlap in flowering among plant species, can influence plant fitness through its effects on heterospecific pollen transfer and competition for pollinators and resources. This temporal overlap is changing as plant species respond differently to climate change. However, difficulties in quantifying and comparing patterns of coflowering across diverse, multispecies communities have limited progress in understanding climate-driven shifts in coflowering. Network-based approaches offer a promising solution to this limitation. Here, we investigate how warming and altered precipitation influence coflowering network structure, flowering phenology, and seed production in 12 annual California serpentine grassland plant species in a mesocosm experiment, using coflowering network analysis and structural equation modeling. Communities consistently grouped into two phenological modules corresponding to early- and late-season annuals. Warming reduced the overall amount of coflowering in the community, driven primarily by weakened and less diverse coflowering among late-season species. Soil moisture moderated these effects: late-season coflowering was strongest under cool-wet conditions and weakest under warm conditions regardless of soil moisture. In contrast, early-season species had a more fixed phenological window, and maintained stable coflowering relationships across climates. Growth form (grass or forb) and origin (native vs. non-native) did not predict coflowering responses. We additionally found that temperature and soil moisture influenced seed production largely through their effects on flowering phenology, with early- and late-season species exhibiting distinct phenological responses and sensitivities. Our results demonstrate that warming and precipitation change can spread out flowering times within functional groups, reshaping community-wide coflowering networks through species-specific responses to altered climate conditions.

Streambank erosion degrades both terrestrial and aquatic ecosystems, and riparian buffers are among the most widely used tools to control it. Yet most of what is known about how riparian vegetation controls erosion comes from one or two-species systems, which overlook the diversity of riparian communities and species interactions (competition and facilitation) that may shape the root systems that bind soil. We used a streambank mesocosm experiment to test whether plant species richness alters root traits, and whether root traits and species interactions reduce erosion. We established 85 mesocosms ranging from unvegetated controls to six-species polycultures, measured root biomass, length, volume, and diameter, quantified species interactions with relative yield totals (RYT), and measured erosion rate with a jet erosion test. Vegetated mesocosms eroded significantly less than unvegetated controls (t = 4.05, p = 0.023). Richness did not affect most root traits, although mean root diameter decreased as richness increased ({beta} = -0.0225, p = 0.047). RYT ranged from competition (RYT < 1) to facilitation (RYT >1), yet neither interactions nor any root trait was directly associated with erosion rate. However, RYT was negatively associated with soil moisture ({beta} = -0.196, p = 0.010). Piecewise structural equation models showed that erosion was best explained by a pathway linking species presence, root length, and soil moisture (AIC = 39.27; Fisher's C = 11.27, df = 12, p = 0.506). Panicum virgatum increased root length, whereas Lolium multiflorum and Echinochloa crus-galli reduced it; longer roots were associated with drier soil (standardized {beta} = -0.31, p = 0.005), and wetter soils eroded more slowly (standardized {beta} = -0.39, p < 0.001). Which species were present mattered more than how many, and plants were linked to erosion through soil moisture rather than root traits alone. Species selection for riparian buffers may need to consider effects on soil moisture as well as rooting extent, particularly as droughts followed by intense storms become more common under a changing climate.

Forests are threatened by a multitude of stressors, including anthropogenic disturbances and climate change. Assessing how forests will respond to these stressors requires a comprehensive understanding of net primary productivity (Npp), environmental constraints on growth, and adaptive capacity. A parameter of significant uncertainty is belowground Npp (bNpp), which can account for up to 80% of total Npp but is poorly estimated and rarely measured directly. We used a cross-biome dataset of direct, field-based measurements of aboveground and belowground primary productivity and 21 climatic and soil variables to identify potential constraints on bNpp and belowground carbon allocation in boreal and cold temperate forests. Soil variables, rather than climate variables, were the main drivers of bNpp and belowground allocation across biomes. The importance of soil variables suggests that soil nutrient dynamics, especially soil nutrient pool and flux variables, must be explicitly modeled to more accurately predict feedbacks between climate, productivity, and within-tree carbon allocation. Within biomes, environmental drivers of belowground allocation varied between low versus high allocation forests, indicating that environmental drivers are site-specific and the development of within-biome, site-scale classifications for forest ecosystems could be useful. Changes in soil variables, such as increasing soil nitrogen pools, caused abrupt and large decreases in bNpp for boreal, but not cold temperate forests. Threshold-like shifts indicate that boreal forests might have lower adaptive capacity and higher sensitivity to disturbances than cold temperate forests. With 70% of boreal forests characterized by low bNpp, disturbances such as anthropogenic nitrogen deposition could cause large-scale decreases in bNpp that could push these forests beyond their adaptive capacity.

The increasing frequency and intensity of heatwaves under climate change highlight the need to understand how biodiversity regulates forest canopy thermal dynamics. Although tree diversity can buffer the microclimate, its effects on canopy temperature and the role of mycorrhizal symbioses remain unclear. We addressed this question in the MyDiv tree diversity experiment in Germany, where tree species richness (1-, 2-, and 4-species mixtures) and mycorrhizal types (arbuscular, ectomycorrhizal, and mixed) are factorially manipulated. During the 2024 growing season, we conducted nine uncrewed aerial vehicle (UAV) surveys using integrated thermal and LiDAR sensors to quantify canopy temperature and structural complexity, together with measurements of leaf water content, specific leaf area, soil moisture, and vapour pressure deficit (VPD). Increasing tree diversity generally reduced canopy temperature, although the strength of this relationship varied seasonally and among mycorrhizal types. Cooling effects were strongest during peak summer heat and were more pronounced in arbuscular mycorrhizal (AM) communities than in ectomycorrhizal (EM) and mixed (AM+EM) communities. In contrast, EM communities exhibited greater canopy structural complexity, whereas AM communities maintained higher soil and leaf water content. Structural complexity increased with tree diversity but did not necessarily result in greater canopy cooling. Structural equation modelling revealed that forest thermal buffering emerged through complementary structural and hydraulic pathways, whose relative importance shifted seasonally, with hydraulic regulation becoming increasingly important under hotter and drier conditions. By linking canopy temperature, canopy structure, and plant water relations, our study provides mechanistic insights for understanding how multiple facets of biodiversity regulate forest thermal buffering under climate warming.

Temporal stability of plant populations is a key determinant of species coexistence and ecosystem functioning, yet its trait-based explanation has remained largely aboveground. Whether root functional traits contribute to long-term population stability remains poorly understood, despite roots mediating water and nutrient acquisition. We combined 17 years of vegetation monitoring from 150 temperate grassland plots across three regions in Germany with root functional trait data for 72 common grassland species. Temporal stability was computed as coefficient of variation (CV), with lower CV indicating higher stability. We tested whether root traits predict stability, and specifically which belowground dimensions from resource acquisition/conservation or soil exploration dominate, and whether differences in land use intensity modifies these relationships. Root traits were stronger and more consistent predictors of temporal stability than the classical aboveground traits specific leaf area (SLA) and leaf dry matter content (LDMC). Higher average root diameter (AD), root hair incidence (RHI), specific root length (SRL), and root tissue density (RTD) were generally associated with higher temporal stability, although the strength of these relationships varied among regions. Land use modified selected trait and stability relationships, with the strongest shifts occurring along the mowing gradient. Increasing mowing strengthened the associations of higher RHI and RTD with temporal stability, whereas the associations of AD and root N shifted towards lower stability. Our results identify belowground functional traits as an important missing component of trait-based temporal stability research and advance our understanding of the functional strategies underlying population temporal stability in managed semi-natural grasslands.

O_LISpecies interactions are increasingly recognised as temporally dynamic. For plant-pollinator networks, evidence shows that interactions vary not only seasonally but also over the diel cycle. However, we still know little about what determines their diel structure, whether this structure is important for robustness to species loss, and which plant traits are associated with the roles plants play in this structure. These gaps are fundamental, because knowing what shapes networks over the diel cycle is required to predict the effects of global change drivers, such as light pollution, that may shift the timing of interactions. C_LIO_LIUsing 22 plant-pollinator networks sampled over morning, afternoon, and night, we addressed these gaps by applying a multilayer framework to characterise their diel structure, link it to robustness, and test which plant traits are associated with plant roles across diel periods (participation, versatility) and within them (centrality). C_LIO_LIDiel structure was non-random: interactions were segregated among diel periods yet integrated through plants visited across the diel cycle. Networks were more robust to simulated species loss when interactions were on average more evenly distributed across periods and when plants were more strongly interconnected among periods, although this benefit diminished when both properties were high simultaneously. The association between plant traits and their roles shifted with the temporal scale. Across diel periods, structural traits were the stronger predictors: taller plants were visited more evenly across the diel cycle (higher participation), whereas shorter plants shared pollinators with plants from multiple periods (higher versatility), potentially mediating indirect effects among them. Within diel periods, floral visual cues became more influential, with achromatic contrast the most consistent predictor: at night, plants with brighter flowers were well visited within the period, sharing pollinators with other plants of that period (higher centrality). C_LIO_LIThese findings identify diel structure as a functional axis of network organisation and indicate that plant-pollinator networks are assembled hierarchically: structural traits set a baseline across the diel cycle, whereas light conditions determine which traits matter within periods, ultimately defining distinct temporal pathways vulnerable to global change. C_LI

Aim Regional climate-change refugia - areas where abiotic conditions related to hydrology and topography sustain unique plant communities despite an otherwise unsuitable regional macroclimate - have been essential for maintaining biodiversity through past climatic transitions. At fine spatial scales, regional refugia can also contain microrefugia, where canopy structure and topography strongly buffer understory microclimate. We tested whether such microrefugia structure understory plant communities within a regional climate-change refugium, where sheltered, groundwater-fed canyons have sustained a biogeographically distinctive flora since the Last Glacial Maximum. Location Temperate forest in the North American Great Plains, Nebraska, USA. Methods In a 20.2-ha forest inventory plot spanning a sharp topographic gradient, we quantified diversity, composition, and distributions and of understory woody and herbaceous plants across seven functional groups along fine-scale topographic, light, and soil moisture gradients. Results Woody species density and richness were significantly higher within sheltered microrefugia in canyon bottoms and on forested slopes and strongly decreased with increasing canopy openness, whereas C4 grass richness was highest along the exposed, high-light forest-prairie ecotone, reflecting divergent functional group responses to microclimate buffering. Soil moisture structured richness across functional groups: woody plant and C3 grass richness peaked at moisture extremes, sedge richness at intermediate moisture variability, and C4 grass richness declined with increasing moisture variability. Both woody and herbaceous understory plants showed strong variation in composition with topography; in particular, woody seedlings and graminoids exhibited strong species turnover with increasing elevation. Main Conclusions These results demonstrate that topographically driven microrefugia support disproportionate understory biodiversity and suggest that protecting both regional refugia and finer-scale microrefugia will be critical for conserving plant communities, including species documented exclusively in microrefugia, as macroclimate continues to change.

The density-yield relationship of wild plant populations, which in agricultural systems is typically consistent with the constant final yield hypothesis, has remained largely unexplored in a restoration context despite its potential importance. In degraded California shrublands, non-native annual grasses are associated with reduced native shrub recruitment. By studying their density-yield relationship, we can potentially optimize current restoration practices by identifying the level of grass control needed to reduce the competitive pressure they create for woody plants. This in turn could enhance woody plant establishment. Here, we conducted field experiments at a degraded chaparral site in Santa Barbara County, California. We manipulated the density of Bromus diandrus, a dominant invasive annual grass, through 0-100% removal treatments to test its density-yield relationship. Simultaneously, we outplanted Ceanothus megacarpus seedlings as phytometers while also monitoring soil moisture and light availability throughout the growing season. Our results demonstrate that the Bromus diandrus density-yield relationship follows the constant final yield hypothesis, with strong intraspecific competition observed from moderate to high grass density contributing to a constant population biomass and a density dependence allometric exponent comparable to previous agricultural studies. A reduction in grass biomass (yield) was observed at 75% grass removal, yet an increase in resource availability (i.e., light and soil moisture) was only observed at 100% grass removal, as indicated by both direct measurements and enhanced phytometer performance. Applied to restoration in California Chaparral and other Mediterranean ecosystems, where the reduction of "weed" biomass is a common management goal, our findings suggest full removal of non-native annual grasses is likely needed to aid in the restoration of native woody seedlings.

1. Grassland ecosystems are pivotal for global biodiversity, carbon (C) storage and food security. This biome, however, together with the ecosystem services it provides, is threatened by global change perturbations. These perturbations rarely occur in isolation and most often affect environments simultaneously, which is a pressing issue in ecology. Nevertheless, most of the research conducted so far has focussed on single simultaneous perturbations or pairs thereof. The few examples considering the effect of more than two simultaneous perturbations focussed only on plants or soil separately, which ignores the potential buffering or feedback effects on ecosystem functioning induced by plant-soils interactions. 2. To tackle this knowledge gap, we used an experimental setup involving mesocosms (soil and plants intact) collected in an extensively managed permanent grassland. Mesocosms were arranged in a greenhouse experiment where we applied individual and combined global change perturbations (up to ten in total) including application of fertilisers, pesticides and antibiotics, simulated trampling and grazing, heat stress and drought. Measured ecosystem functions and properties included a large set of biological, physical and chemical soil and plant associated variables. 3. Results showed that multiple perturbations in most cases had negative effects irrespective of the direction found for individual perturbations. However, in contrast to previous bare soil experiments, we only found scarce evidence for substantial synergistic effects of simultaneous perturbations in our experiment conducted with mesocosms that contained vegetation. This suggests a mitigating impact of plants on soil responses to global change. Moreover, we found a strong dominative effect of drought on ecosystem functioning, driving most of the effects observed. 4. Synthesis: Our results highlight the importance of evaluating global change impacts under more realistic settings, i.e., applying multiple perturbations simultaneously and including plant-soil interactions. Mitigation of global change in terrestrial ecosystems will only be possible once we holistically understand how ecosystems respond to altered conditions.

Fire-dependent or fire-influenced ecosystems harbor species with diverse response strategies, such as smoke-stimulated germination and dormancy release through heat shocks. Although these cues influence germination in many fire-prone ecosystems, their effects on seeds of woody species from tropical dry forests (TDFs) remain poorly understood. We evaluated the germination response of twelve TDF woody species to three heat-shock temperatures (80, 120, and 160C), two exposure times (3 and 5 min), and two smoke conditions (with and without smoke water), plus dry and wet controls. We also assessed the relationship between germination and functional seed traits, including permeability, seed-coat thickness, mass, and shape index. Of the eleven species that germinated, four showed broad tolerance, with no significant reduction in germination even at 160C for 5 min, which was associated with lower permeability, thicker seed coats, and greater mass. Four other species showed restricted tolerance to heat, with significant declines in germination at higher temperatures or longer exposure times, and were associated with higher permeability, thinner seed coats, and lower seed mass. The remaining three species showed heat-stimulated germination between 80 and 120C and tended to have lower shape-index values. Smoke treatments significantly increased germination in three species, Ceiba pentandra, Guazuma ulmifolia, and Samanea saman. In the remaining species, significant differences were either absent or attributable to moisture or imbibition effects rather than to smoke itself. Overall, germination responses to heat were associated with variation in functional seed traits, whereas responses to smoke were less widespread and showed a less consistent functional pattern.

Forest canopy height is sensitive to shifts in demography (which affect tree size distributions and abundances of short-vs. tall-statured species) and allometry. Demographic and allometric effects on canopy height have not previously been studied across a large region. We analyzed more than two million individual tree measurements from about 60,000 remeasured inventory plots in non-plantation forests across 19 ecoregions of the conterminous United States to quantify changes in aboveground live tree biomass (AGB), canopy height, and its demographic and allometric components from ~2003-2025. We tested the hydraulic limitation hypothesis, which predicts decreasing tree height with increasing hydraulic stress. AGB was strongly correlated with canopy height across ecoregions, and AGB change was strongly correlated with canopy height change over time. Canopy height change was positive in nearly all eastern ecoregions but varied in sign across western ecoregions. Demographic shifts led to increased canopy height in most eastern ecoregions but decreased canopy height in most western ecoregions. In contrast, allometric shifts, averaged across species, caused either positive or no significant change in ecoregion-mean canopy height. Our results were consistent with several predictions of the hydraulic limitation hypothesis: (i) intraspecific allometric shifts were more positive for understory than for canopy trees, (ii) across ecoregions, canopy height was positively correlated with precipitation and negatively correlated with vapor pressure deficit, and (iii) canopy height change over time was positively correlated with precipitation change. However, our results suggest modifications to the hydraulic limitation hypothesis, because allometric shifts tended to be positive, even in ecoregions with decreasing precipitation and increasing vapor pressure deficits. Such allometric shifts, which seem counter to the hydraulic limitation hypothesis, may be caused by CO2-driven increases in water-use efficiency. Whatever their cause, allometric shifts should be accounted for to avoid biased estimates of canopy height change and biomass change.

Climate change is accelerating species losses in ecosystems across the world. Seed germination is a critical, climate-dependent phase of the plant life cycle; however, the ecological determinants of germination climate niches within diverse landscapes and across functional types (FTs) are still not well understood. In this study, we characterized seed germination temperature and water availability niches for 28 species that represent different FTs (tree, shrub, grass, forb) and vegetation types (grassy woodland, dry and wet forests) within a temperate bioregion (Sydney, Australia). We tested whether ecological determinants, specifically species climate of origin, seed traits, FT and vegetation type explain germination niches and predicted spatial and temporal patterns of germination potential across the landscape under high and low emission scenarios. We found wide variation in thermal and hydric germination niches among species. Optimal germination temperature (thermal niche) was predicted by FT, climate of origin and seed traits, such that shrubs, cool-origin species, and species with large seeds had significantly cooler optimal temperatures for germination. We also quantified spatial and temporal changes in germination potential to identify vulnerable areas and FTs. We found strong species-specific seasonal patterns in germination potential with future climate shifts affecting FTs differently; germination of woody species declined more than forbs. Future germination potential was predicted by historical climatic conditions, with warmer and drier localities being more vulnerable. Overall, our findings demonstrate that species germination responses to climate change depend on FT, seed traits, and species climate of origin, with woody species and warmer, drier parts of the landscape emerging as being particularly vulnerable to declines in recruitment. Our study provides a mechanistic understanding of germination responses to temperature and water availability, enabling predictions of vulnerable species and areas for conservation under climate change, and inform large-scale ecosystem restoration approaches through improved species selection and sowing times.

Balancing sexual reproduction with vegetative growth poses challenges for many plants. Additionally, reproductive effort, such as production of pollen or flowers, is not always correlated with numbers of viable and germinable seeds. Energetic tradeoffs like these can be particularly fraught for imperiled species and might be especially complicated in landscapes with a possibility of interspecific hybridization. In this multi-year study, we track the reproductive effort, flowering phenology, and reproductive output of two interfertile Sarracenia (pitcher plants). We show that reproductive effort varies among years and sites, and that production of pollen and flowers does not always result in higher seed output. We demonstrate that phenological overlap in the timing of flower production and pollen viability could enable hybridization, and that the more imperiled taxon produces fewer seeds per flower. Results have implications for conservation of these pitcher plants, management of hybridization in these and other systems, and general principles of reproductive allocation.

O_LIForests are vertically structured ecosystems where light attenuation, microclimatic buffering and resource availability occur along continuous gradients from the canopy to the understorey layers. Despite this complexity, studies rarely integrate trophic interactions across vertical layers, overlooking how vertical forest structure shapes consumer communities through abiotic and biotic pathways. C_LIO_LIIn this study, we combined layer-specific measurements of plant diversity, structure and microclimate with arthropod sampling in the canopy and understorey, as well as bird survey data, in a temperate forest. Applying Bayesian structural equation models with explicitly defined directional pathways, we modelled both consumer biomass and abundance across trophic levels and vertical layers. C_LIO_LIAbundance measures were predominantly filtered by local layer conditions, while biomass responded to conditions across layers, reflecting stand-level energy flow. This suggests that these two metrics capture fundamentally different ecological processes. Canopy conditions consistently predicted understorey arthropod abundance across trophic levels, while the reverse was not observed, suggesting a strong asymmetric downward propagation of canopy-driven effects. Furthermore, trophic interactions between arthropod primary and secondary consumers remained largely stratified within vertical layers, suggesting a vertical food web compartmentalisation rarely shown in structurally complex aboveground systems. C_LIO_LIPlant diversity, structure and microclimate shaped consumer communities mainly through the modulation of resource availability and plant apparency, with effects varying across vertical layers, trophic levels and taxonomic groups. Through complementary mechanisms, plant diversity likely determined the variability of resources available to consumers at different trophic levels. Structural properties, in contrast, potentially drove the spatial redistribution of these resources through light attenuation and microclimatic buffering, which in turn influenced the physiological capacity of consumers to access and exploit available resources. C_LIO_LIThese findings demonstrate that vertical stratification mediates trophic pathways in a highly directional manner, with canopy characteristics playing a disproportionate role in structuring the forest community across layers. Integrating layer-specific structural and trophic indicators into forest biodiversity assessments and management strategies is therefore essential to fully evaluate biodiversity dynamics and multifunctionality in structurally complex forest ecosystems. C_LI

Arboreal camera trapping is an emerging remote sensing tool for building plant-frugivore interaction networks with enormous potential to provide new insights into seed dispersal as an eco-evolutionary process. Interaction data derived from camera trap videos broadly falls into two categories: visitation, indicating the presence of an animal in a tree, and intake, indicating true fruit consumption. These linked approximations of seed dispersal may or may not provide distinct sets of information, shaping our understanding of network dynamics, yet the extent to which emergent network structure depends on variable choice is unknown. To address this gap, we deployed camera traps in the canopy of 21 plant species in the Costa Rican tropical dry forest, an endangered biome with sparse community-wide plant-frugivore interaction data. Our objectives were: 1) Quantify plant-frugivore network structure in the tropical dry forest via network, group, and species-level metrics; and 2) Compare metrics between visitation and intake networks. We find strong structural patterns of low connectance, low nestedness, high modularity, and low niche overlap that differ significantly from null expectations in the same direction in both networks. The visitation network exhibited significantly greater modularity & species degree than the intake network, while the intake network exhibited significantly greater plant niche overlap than the visitation network. These findings suggest that the likelihood of consuming fruit on camera differs across species, which may have downstream consequences on other emerging metric estimations. Researchers seeking to build quantitative plant-frugivore networks should therefore consider visitation and intake variables in concert when relying on camera trap data. This work provides foundational plant-frugivore interaction data for a critically endangered biome and enhances our understanding of camera trapping as a remote sensing tool in frugivory and seed dispersal research.

Lianas are an integral component of tropical forests, and their increasing abundance may alter forest structure and function. Whether infested host trees are shorter and smaller-crowned because lianas suppress them, or because lianas preferentially colonise already small trees, has remained unresolved. We used a longitudinal dataset of individual liana infestation histories spanning fourteen years (2011-2025), combined with terrestrial laser scanning (TLS)-derived structural metrics for 251 trees across 67 species on Barro Colorado Island, Panama. Bayesian analyses showed that persistently infested trees were shorter (11.0%), with smaller crown projected area (22.3%) and crown volume (25.9%) than never-infested trees; the 95% credible intervals for these effects excluded zero. Pre-infestation crown architecture did not credibly predict subsequent colonisation. These patterns are more consistent with persistent infestation imposing lasting architectural legacies than with strong pre-existing architectural susceptibility, suggesting rising liana abundance may progressively reduce forest stature and canopy density, with implications for tropical carbon storage projections.

Elevational gradients provide a framework for understanding how environmental filtering reorganises communities and interactions, but plant-pollinator interactions along temperate forest elevational gradients remain overlooked. We studied early-spring understorey communities at four forest sites spanning the foothills towards the timberline (450-1,000 m a.s.l.) in the Krkono[s]e Mountains, Czechia. Across six transects per elevation, we quantified flowering plant species richness, floral resources and traits, and video-recorded flowers, yielding 4,003 pollinator visits. We analysed elevational patterns in species richness, community composition, floral traits, and quantitative network characteristics. Visitation frequency and flowering plant and pollinator species richness peaked at intermediate elevations. The contribution of dipteran relative to hymenopteran pollinators increased towards higher elevations, principally because of non-hoverfly flies, whereas individual bee groups showed no uniform response. Floral resources and traits showed no uniform elevational responses, although total nectar sugar availability peaked at the highest site because of the dominant Vaccinium myrtillus. Most notably, both network-level specialisation and mean species-level specialisation were generally greater at the two higher elevations, whereas nestedness was lower and other network characteristics showed no consistent patterns. These findings suggest that shifts in pollinator composition and dominant floral resources potentially shaped interactions along the gradient. The increasing specialisation with elevation contrasts with the generalisation often expected under reduced partner availability and indicates that forest networks may follow elevational patterns not predicted from open habitats. Despite limited site-level replication, this study provides, to our knowledge, the first community-wide characterisation of plant-pollinator interactions along a temperate forest elevational gradient and identifies patterns requiring evaluation across replicated gradients.

Forest productivity and stability underpin ecosystem functioning and services, yet whether the drivers and mechanisms regulating these properties differ between natural and planted forests at large spatial scales remains unclear. Using data from 2,389 forest plots across China and a novel analytical framework, we systematically compared productivity and stability between natural and planted forests and identified their dominant drivers and underlying mechanisms across 38 biotic, structural, climatic, and edaphic variables. We found that natural forests were 14% more productive and 34% more stable than planted forests, although these differences varied across ecoregions and stand development stages. Productivity in natural forests was primarily driven by functional diversity and functional traits, supporting the niche complementarity and mass-ratio hypotheses, whereas nitrogen availability emerged as the dominant regulator of productivity in planted forests. Likewise, stability shifted from diversity-related regulation in natural forests to stronger abiotic control in planted forests, while functional diversity remained a key stabilizing mechanism in both forest types, supporting the insurance hypothesis. We also found that planted forests were more sensitive to hydroclimatic variability, indicating greater vulnerability to drought. Our findings reveal a fundamental shift in the mechanisms regulating forest productivity and stability from diversity-mediated control in natural forests to stronger abiotic regulation in planted forests, highlighting the importance of integrating biodiversity conservation and nitrogen management to sustain forest productivity and stability under future climate change.

Rapid urbanization is altering the seasonal life cycles, or phenology, of urban trees. Diverse species composition and heterogeneous urban thermal environments can generate pronounced phenological variation within cities, yet such fine-scale seasonality remains poorly understood. This limits the ability to anticipate localized ecosystem services and disservices, including canopy cooling and pollen exposure. Here, we analyzed individual-tree phenology and their response to temperature across 76 cities in the contiguous United States by integrating municipal street-tree inventories, PlanetScope satellite time series, and 1-km near-surface urban air-temperature data. We quantified intra-city and cross-city phenological variation, estimated species-level associations between phenology and fine-scale urban temperature, and tested whether these associations vary with regional climatic context. Intra-city phenological variation rivaled cross-city variation for 33.3% of species in spring and 53.7% in fall. Temperature-phenology associations were widespread: warmer local urban environments were associated with earlier spring onset in 51.5% of cases and delayed fall senescence in 66.5%, with approximately two-thirds of these responses cascading to longer growing seasons. These associations varied with regional climate, generally showing stronger spring advancement and fall delay in cooler and wetter cities, but weaker or reversed responses in hotter or drier cities. Together, these results suggest that urban tree phenology is shaped by interactions among fine-scale urban temperature, species identity, and regional climate. Accounting for this heterogeneity can improve predictions of urban forest function and inform climate-resilient species selection and management. Plain Language SummaryStreet trees within the same city do not all leaf out in spring or lose their leaves in fall at the same time. This seasonal timing, called phenology, affects services such as shade and cooling, as well as disservices such as pollen exposure. We studied individual tree phenology and their temperature response in 76 U.S. cities using street-tree records, satellite observations, and local air-temperature data. We asked how much phenology varies within cities, how it responds to neighborhood-scale temperature differences, and whether these responses depend on regional climate. Variation within a city was often as large as variation among cities, especially in fall. Trees in warmer parts of cities generally leafed out earlier and lost their leaves later, often extending the growing season. However, these responses were stronger in cooler and wetter cities and weaker or sometimes reversed in hotter or drier cities. These findings show that urban tree seasonality depends on local temperature, species identity, and regional climate, which should be considered when selecting and managing trees for future cities. Key PointsO_LISubstantial within-city phenological variation is not captured by city-level averages. C_LIO_LIFine-scale temperature shaped phenology and growing-season length, but responses varied across species, cities, and phases. C_LIO_LITemperature sensitivity depended on regional climate, weakening or reversing in hotter or drier cities. C_LI

Identifying how species interact in ecological networks is key to understanding species' contributions to ecosystem functioning, but can be difficult to establish in species-rich communities. Functional niches, which capture broader ecological strategies, are often easier to assess and could help approximate species' interaction roles. We tested this by integrating species interaction and leaf functional trait data of tree species and leaf-associated interaction types from a large-scale forest biodiversity experiment. We found that identity, variability, and similarities of functional niches shaped corresponding aspects of interaction roles. Interestingly, however, this association inversed in the case of functional niche variability, which decreased instead of increased the variability of species interactions. This suggests a potential trade-off between processes that determine variability in functional and interaction space. Furthermore, effects were not constrained within aspects, with similarities of interactions being particularly responsive to most tested niche aspects, indicating that interaction-based competition among plants is particularly sensitive to functional niche differences. Moreover, many effects of functional niches depended on tree species biomass and tree species richness, highlighting the importance of the larger ecological context for determining species interaction roles. Together, our findings show that species' interaction roles, while clearly altered by, are not simply a consequence of functional niches. Instead, interaction roles integrate multiple aspects of functional niches depending on species performances and community composition and can therefore be crucial for anticipating ecosystem change.