Biodiversity-oriented broadleaf forestry needs practical indicators that translate local forest-habitat conditions into information useful for silvicultural and operational decisions. Here, we used the Female-attracted Flight Interception Trap (F-FIT) to quantify spatial variation in male captures of Platycerus takakuwai Fujita, 1987 (Coleoptera: Lucanidae) as a local habitat indicator in a broadleaf forestry landscape in Hida, Japan. We sampled 27 trap locations spanning unharvested forest, a regenerating clear-cut opening, and an area thinned for temporary cable-yarding tower installation. Each location was sampled in the morning and afternoon, yielding 54 1-h F-FIT deployments. Capture counts were analyzed using negative-binomial spatial models. After evaluating contextual variables, including air temperature, wind speed, elevation, and potential solar radiation, the main model examined whether the association between stand density and captures varied spatially while accounting for edge distance, sampling period, spatial structure, and repeated sampling. A complementary edge-focused model summarized the overall association between forest edge position and captures. The association between stand density and captures was better represented as spatially varying than by a single coefficient shared across the study area, although local coefficients remained uncertain. The most negative point estimates occurred in the thinned area. The edge-focused model indicated that expected capture rates increased toward forest interior positions and decreased toward open areas. F-FIT capture data can, therefore, translate point-level Platycerus responses into spatially explicit information on broad edge-related gradients and locally varying stand-structural associations. This approach can complement stand- or plot-level assessments by informing discussions of retention placement, opening size, and operational elements.
In North America, many canopy ash trees (Fraxinus spp.) are being eliminated from forests by the Emerald Ash Borer (Agrilus planipennis). This is yet another case of a native tree species being functionally eliminated by non-native pests. Lasting effects on forests might be minimal if the species that replace ash are ecologically similar (functionally redundant). Alternatively, effects could be lasting if the vulnerable tree species have unique ecological attributes. As one of few ring-porous trees in north-temperate hardwood forests, ash leaves emerge and expand 1–2 weeks later than co-occurring deciduous trees our study region. This allows for higher light transmission into the subcanopy and could thereby benefit the understory plant community. In advance of their elimination, we tested for floristic differences beneath groves of canopy white ash compared to other canopy tree species. We studied understory flora with a factorial study that included four canopy tree species (white ash, American beech, yellow birch, sugar maple) on each of two soil types (Bh podzol and typical podzol). The subcanopy flora beneath white ash contained more plants of more species. Several species were more abundant on Bh podzol soil than typical podzol, and several species displayed effects of canopy tree species that depended upon soil type. There was limited evidence that the understory flora beneath ash was functionally unique. The flora beneath ash is apparently enriched due to greater radiation reaching the forest floor. The loss of canopy ash trees seems likely to reduce and homogenize the understory flora.
Pollinator communities in temperate forest mosaics are influenced by local habitat conditions, but within upland oak forests of the southeast U.S., these relationships remain poorly understood. This study assessed how differing land-cover types influence wild bee diversity and community composition across upland oak forest habitat. We evaluated how four land-cover types – mixed hardwood forest, pine/mixed pine forest, scattered hardwood, and open/agricultural fields – shape wild bee diversity, community composition, and functional structure in privately owned mixed-use forests in northern Mississippi. Using standardized blue vane and colored pan traps, we sampled bees monthly from May to September 2024. After excluding managed Apis mellifera, we recorded 1315 wild individuals representing 67 species and 26 genera. Shannon diversity differed among land-cover types (F(3, 8) = 5.749, p = 0.021), with open/agricultural fields showing higher diversity than mixed hardwood forest, although evenness did not differ. Community composition also varied by land cover (NMDS stress=0.099; PERMANOVA F(3, 8) = 1.964, R² = 0.424, p = 0.004), and seven species were significant indicators, primarily of open/agricultural fields and pine/mixed pine forest. Canopy cover was the strongest environmental correlate of assemblage structure (global p = 0.009). Open/agricultural sites supported the highest richness and diversity, pine/mixed pine forest and scattered hardwood were intermediate and often functioned as ecotones and closed-canopy hardwood forests had the lowest diversity. These results highlight the value of maintaining a heterogeneous land-cover mosaic and promoting canopy openness and ground-nesting resources to conserve wild bee communities in upland oak forests. Our results contribute to understanding how forest management practices such as thinning and selective logging can balance ecological integrity and enhance pollinator diversity in temperate ecosystems.
The Cerrado is a Neotropical savanna hotspot, crucial for biodiversity and carbon storage. In this seasonally dry environment, many woody species affected by anthropogenic disturbances can resprout, playing a central role in the onset of secondary succession and in the recovery of ecosystem functioning. Our objective was to evaluate the relative contribution of resprouting and non-resprouting functional groups to the recovery of taxonomic diversity and aboveground biomass stocks across secondary successional stages in Cerrado woodland in north-eastern Brazil. We sampled trees and shrubs (woody community) with diameter at breast height ≥ 10 cm in 5 fragments of Cerrado woodland at different successional stages after shifting cultivation. At each site, we established three randomly located 50 × 50 m plots (2500 m²), each subdivided into 25 subplots of 10 × 10 m (100 m²). Taxonomic diversity increased with stand age. Approximately 70% of total species occurred in 18- and 35-year-old fragments and mature reference forests. Across all successional stages, resprouting species contributed to 55% of woody plant species (richness) and 93% of total aboveground biomass (AGB), while non-resprouting species contributed to 45% of richness and 7% of AGB. The recovery of taxonomic diversity and AGB in Cerrado woodland is strongly driven by the functional complementarity between non-resprouting and resprouting plants. Resprouting species are central to the resilience of Cerrado woodland and to the early recovery of community structure and carbon stocks, whereas non-resprouting species contribute to the maintenance of diversity, resulting in joint biodiversity-carbon cobenefits.
Recurrent fire and vegetation recovery are the dominant ecological processes governing long-term aboveground carbon dynamics in tropical savannas and wetlands, yet their simultaneous multi-decadal quantification across multi-biome landscapes remains rare. Here, we reconstructed four decades (1985–2024) of aboveground carbon dynamics across Mato Grosso do Sul, Brazil, integrating 2662 field-calibrated environmental licensing inventories — an underexplored but empirically rich source of vegetation data — with annual land-cover maps from MapBiomas Collection 10 and burned-area data from MapBiomas Fire Collection 4. Biome- and physiognomy-specific carbon stocks were assigned to forest, savanna, wetland, and grassland formations, allowing annual reconstruction of carbon stocks, vegetation transitions, fire-induced losses, and net carbon balance across the Cerrado, Pantanal, and Atlantic Forest. Substantial changes in aboveground carbon storage occurred across all biomes between 1985 and 2024. In the Cerrado, forest and savanna carbon stocks declined by approximately 50%, while in the Pantanal, carbon stored in wetlands decreased from 27.2 to 3.3 million Mg C and grassland carbon increased from 7.0 to 33.9 million Mg C, becoming the largest aboveground carbon reservoir in the biome by 2024. Although woody encroachment, vegetation succession, and regeneration generated cumulative carbon gains of 34.4 million Mg C, these gains were overwhelmed by recurrent fire, which caused cumulative losses of 162.9 million Mg C — nearly fivefold greater. Consequently, reconstructed statewide carbon stocks declined from 248.2 to 162.8 million Mg C. The Pantanal revealed a striking carbon paradox: despite a relatively modest net stock decline of 9.7 million Mg C, cumulative fire-induced losses exceeded 145.7 million Mg C — almost twice the biome's initial carbon stock — indicating intense cycles of biomass combustion and recovery that are entirely concealed by conventional stock comparisons. Mann–Kendall analyses confirmed significant negative carbon trajectories across all biomes (τ = −0.997 to −0.455). Bootstrap analyses and a sensitivity analysis using IPCC et al., (2006) default combustion factors — which yielded even larger estimated fire-induced losses (+12.8% statewide) — collectively demonstrated that these conclusions are robust to parameter uncertainty. These results show that long-term carbon dynamics across this tropical savanna–wetland mosaic are governed primarily by the balance between fire disturbance and vegetation recovery, not by vegetation transitions alone. Fire management therefore emerges as a central strategy for maintaining ecosystem resilience and long-term aboveground carbon storage across tropical multi-biome landscapes.
Soil acidification, nutrient imbalances, and declining carbon stocks threaten the sustainability of subtropical coniferous plantations. Although the co-application of biochar and organic fertilizer shows promise for addressing these issues, its impacts on bacterial community assembly and carbon sequestration remain poorly understood. Here, we conducted a field experiment in Masson pine (Pinus massoniana) and Chinese fir (Cunninghamia lanceolata) plantations with four treatments in a randomized complete block design: control (CK), biochar (BC, 10 Mg ha⁻¹), organic fertilizer (OF, 10 Mg ha⁻¹), and combined application (BOF, 5 Mg ha⁻¹ each). Using 16S rRNA gene sequencing, null model analysis, network analysis, enzyme assays, and structural equation modeling (SEM), we evaluated soil properties, bacterial communities, and carbon dynamics. Our results showed that BOF significantly increased soil organic carbon by 58.5–88.6%, shifting bacterial community assembly from stochastic (NST ≈ 86 −93%) toward deterministic processes (NST ≈ 35 −44%), with homogeneous selection rising from ~15% to ~53% in P. massoniana and from ~5% to ~61% in C. lanceolata. BOF was also associated with greater bacterial network complexity and enhanced carbon-cycling enzyme activities, with β-glucosidase activity increasing by 1.87- to 2.11-fold. Tree species modulated the responses, with greater soil carbon stock gains in P. massoniana (93.8%) than in C. lanceolata (72.8%) relative to the control. SEM showed strong amendment-related effects on soil chemistry and bacterial community attributes. Accordingly, the shift toward deterministic assembly is interpreted as an ecological response associated with improved soil conditions under BOF treatment, rather than as a demonstrated mediator of carbon sequestration. These findings show that BOF simultaneously increased soil carbon storage and shifted bacterial community assembly toward deterministic processes, with stronger overall responses in Masson pine plantations.
Mangroves are important for carbon sequestration in coastal ecosystems, yet no study has investigated carbon accumulation patterns through growth models across mangrove tree classes in Bangladesh. A comprehensive field study was conducted to evaluate the carbon dynamics of different tree classes according to Kraft system–dominant (>8 m height, >15 cm DBH, upper canopy), average (4–8 m height, 8–15 cm DBH, middle canopy), and overtopped (<4 m height, <8 cm DBH, squeezed canopy)–across a 35-year chronosequence in mangrove plantations. Three growth models – Gompertz, Logistic and Richards, were applied to express the carbon accumulation trajectories, and the best-fit model was selected based on high R2 and low AICc values. The results indicated that the Gompertz model was the best fit for the dominant mangrove class, following an asymmetric trail (R2=0.988, AICc=39.05, carbon accumulation=2.46 Mg C ha−1 yr−1). A symmetric Logistic trajectory was followed by biomass carbon of the average tree class (R2=0.949, AICc=17.46, asymptote=11.5 Mg C ha−1). However, the biomass carbon of overtopped trees and soil carbon were not dependent on plantation age (R2≤0.05). Ecosystem carbon also followed an asymmetric Gompertz trajectory, similar to the dominant tree class (R2=0.963, AICc=47.91, asymptote=169.1 Mg C ha−1). Spatially, ecosystem carbon varied among mangrove plantation zones, with the highest carbon stored by the eastern coast (279.9 Mg C ha−1) followed by the central (175.2 Mg C ha−1) and present study sites (western side; 65.9 Mg C ha−1). The growth models performed well at the present study sites (western Sundarbans) but indicated a poor fit for carbon growth in central and eastern mangrove plantations (R2≤0.65). Thus, the Gompertz model should be applied cautiously and with site-specific calibration. We advocate class-stratified carbon accounting, adaptive thinning, and zone-specific modeling as essential measures to improve plantation carbon performance across the heterogeneous Sundarbans landscape and to ensure robust projections under climate mitigation frameworks.
Three-dimensional forest environments support diverse and unevenly distributed assemblages of flower-visiting insects, including many non-bee taxa, yet the mechanisms shaping these patterns remain poorly resolved. Although decades of research have documented vertical stratification in pollinator communities, most studies emphasize where pollinators occur rather than why, leaving the roles of unevenly distributed forest resources, microclimate, and species-specific constraints insufficiently integrated. Here, I develop a mechanistic synthesis showing that forest pollinator diversity is a function not only of resource heterogeneity but also of the number of alternative states each resource assumes as determined by microclimatic context (e.g., sunny vs. shaded). Forests characterized by both high resource diversity and strong microclimatic heterogeneity are predicted to support richer assemblages and more complex interaction networks than forests with comparatively homogeneous microclimates. I also find that the vertical distributions of most flower-visiting insects are dynamic rather than fixed, reflecting the tracking of preferred resource–microclimate combinations across space, time, and life stages. The existing literature points to three research priorities aimed at fully integrating forests and canopy resources into conservation plans. These are (1) distinguishing between forest-dependent and opportunistic pollinator species; (2) clarifying the value of old-growth attributes (tree diversity, large trees, dead wood, and structural heterogeneity) along both vertical and horizontal axes; and (3) optimizing the amount and spatial continuity of forests across the landscape.
Forest growth responses to climate change are influenced by interactions between local hydroclimatic conditions and species traits. However, the climatic drivers and temporal windows governing growth sensitivity and drought legacy effects remain poorly understood. We used tree-ring data from four conifer species across contrasting moisture regimes in interior British Columbia, Canada, to (1) identify seasonal climate drivers of radial growth and (2) quantify drought effects and drought legacy effects following an extreme drought in 2017. Climate window analyses revealed distinct site-specific sensitivities where trees growth at the xeric site integrated moisture over a longer spring–summer period. In contrast, growth at the mesic site responded primarily to shorter summer moisture windows. Species-specific analyses showed that growth was jointly influenced by tree size, competition, and climate, with symmetric (linear) responses to temperature and precipitation at both sites. The analyses also revealed that the 2017 drought reduced growth across all species and sites. However, the magnitude of the impact and persistence of the legacy effects differed markedly. At the mesic site, trees experienced stronger initial growth reductions but recovered rapidly, whereas those at the xeric site showed smaller initial impacts but prolonged legacy effects lasting up to three years. Species-specific recovery trajectories were evident, with Larix occidentalis exhibiting rapid compensatory growth and Pinus species showing persistent growth suppression. This pattern suggests trade-offs between drought sensitivity and recovery capacity. Overall, these results demonstrate that drought effect cannot be inferred from short-term growth responses alone and that post-drought climate and site conditions may influence recovery dynamics. By analysing the climate sensitivity and drought legacies together, this study provides a basis for understanding forest resilience under increasing climate variability and highlights the importance of hydroclimatic context in shaping forest responses to global change.
Juglans mandshurica is a key constructive species in secondary forests of the temperate mountainous regions of East Asia. Its natural regeneration relies largely on rodents dispersing seeds away from parent trees, thereby potentially reducing the negative effects of allelopathy near parent trees. Rodent-mediated seed dispersal is an important component of nature-based solutions (NbS), and promoting rodent-mediated dispersal of more J. mandshurica seeds over greater distances is critical for secondary forest restoration. Seed density, which varies temporally with seed production among years and spatially with parent-tree density, may strongly influence this process. However, few studies have considered the combined effects of these temporal and spatial factors, and it therefore remains unclear how temporal and spatial variation in seed density jointly influences rodent-mediated seed dispersal. To address this question, we conducted seed dispersal experiments in 14 plots spanning a gradient of parent-tree densities over two years, while monitoring seed fate and rodent visitation. The proportion of dispersed seeds was higher in the low seed-production year, and seed production was negatively associated with dispersal distance. Although parent-tree density had no significant overall effect on dispersal distance, rodents tended to move seeds toward areas with lower parent-tree density during the low seed-production year, indicating directed dispersal. Supplementary path analysis suggested that rodent activity may link seed-resource variation to seed-dispersal outcomes. Overall, temporal and spatial variation in seed resources influenced different components of rodent-mediated seed dispersal, with implications for NbS-based secondary forest restoration.
Forest size structure reflects the cumulative effects of climatic, edaphic, biotic, and topographic factors and strongly influences ecosystem functioning. However, whether climatic variability alters the rate of structural convergence remains poorly understood. Using four independent tree-ring datasets spanning regional plantations and global forest networks, we quantified temporal changes in tree size inequality and evaluated how drought influences structural development across age sequences. Across all datasets, tree size inequality declined consistently with age sequences, indicating a general trajectory of structural convergence. Breakpoint analyses identified a transition from a climate-sensitive developmental phase to a relatively stable phase at stand ages of 20–38 years. Drought significantly slowed structural convergence during the early stages, with drought effects on the annual increment of cumulative variability (AICV) being 3.5–5.3 times stronger before than after the developmental breakpoint. In contrast, drought effects became negligible in mature stands. Independent competition analyses further showed that drought intensified the negative effects of size-asymmetric competition on tree growth in young stands, where competition effects on growth were substantially stronger under drought than under normal or wet conditions. Correspondingly, AICV showed greater sensitivity to SPEI, indicating drought delayed structural convergence. These findings demonstrate that climatic variability regulates the rate rather than the direction of forest structural development and reveal an early developmental window during which drought exerts disproportionate influence on long-term stand structure. The results highlight the importance of developmental stage in shaping forest responses to increasing drought under climate change.
Commercial thinning can alter understory plant communities and ecosystem processes, but the short-term links between community reassembly and ecosystem responses remain unclear. In a commercially managed lodgepole pine (Pinus contorta) forest in British Columbia, Canada, we evaluated how strip thinning influenced understory community composition, richness structure, aboveground biomass (AGB), soil C:N ratio, and soil CO₂ flux. We partitioned beta diversity into turnover and nestedness components and quantified total richness, core richness (CR), transient richness (TR), and transient proportion (TP) across one pre-thinning year and two post-thinning years. Community change was dominated by species turnover rather than nestedness, indicating strong temporal replacement in understory composition. Strip thinning also altered richness structure, with total richness, CR, and TR varying significantly among years and thinning treatments, while TP showed a strong temporal response and a weaker treatment effect. AGB differed between years and among thinning treatments, and was associated with TR rather than CR, with the relationship between TR and AGB depending on strip width. Soil C:N ratio showed strong temporal and depth-specific responses, with TR negatively associated with C:N in the upper soil layer and year-dependent relationships in the lower layer. Soil CO₂ flux varied strongly with year and growing-season stage, and its relationship with TR depended on season rather than showing a consistent response across the growing season. Overall, these results show that early ecosystem responses following strip thinning were linked to transient components of the understory community, but that these relationships were context dependent across biomass, soil depth, and seasonal carbon flux dynamics.
Mediterranean open woodlands are increasingly affected by management change and climatic drying, yet the long-term consequences for soil carbon storage and stand condition remain poorly understood. We reconstructed soil organic carbon (SOC) trajectories in Iberian Quercus ilex dehesas by integrating CHELSA climate data, Spanish National Forest Inventory vegetation data, and regional soil observations from a European forest-monitoring network within the RothC model. Simulations covered the upper 20 cm soil layer from 1991 to 2017 and interpreted as model-based reconstructions of relative SOC trajectories. We compared three management trajectories inferred from shrub-cover dynamics: traditional management (TM), land-use abandonment (LA) and shrub clearance (SC). RothC simulations and additive mixed models indicated divergent modelled SOC trajectories associated with contrasting management legacies. LA plots showed the largest net SOC gain (5.12 Mg ha⁻¹) but also the highest tree mortality and shrub encroachment; SC plots showed net SOC losses (-0.43 Mg ha⁻¹) despite improved stand structure; and TM plots showed small net SOC gains (1.93 Mg ha⁻¹). These results suggest a management trade-off: pathways associated with higher modelled SOC accumulation may coincide with structural and demographic degradation that could compromise open-woodland persistence. Within TM plots, mixed‑effects models indicated that adult tree density was positively associated with SOC balance, whereas mortality became negatively associated during the later period; regeneration and shrub cover also showed positive later-period associations. We conclude that adaptive traditional management supporting tree continuity, regeneration, and controlled understory retention may contribute to maintaining both SOC stocks and dehesa functioning under increasing climatic stress.
Natural forests are major carbon sinks, but chronic herbivory by over-abundant ungulates (hereafter overbrowsing) and resulting soil erosion can compromise this function. We quantified the net ecosystem carbon balance (NECB) in southern Kyushu, Japan. Here, approximately 40 years of overbrowsing by sika deer (Cervus nippon) altered mixed broadleaf–conifer stands with an understory (PU) into stands lacking an understory (NU), and further into stands dominated by unpalatable shrublands (SR) or stands with canopy gaps (CG). PU stands maintained a positive NECB (four plot mean; 317.6 g C m−2 yr−1) because high net primary production (Pn; 721.9 g C m−2 yr−1) exceeded the sum of heterotrophic respiration (Rh; 164.8 g C m−2 yr−1) plus lateral carbon export via soil erosion (Se; 239.5 g C m−2 yr−1). Alteration from PU to NU resulted in a negative NECB (−80.1), because the Pn (400.2) did not offset the sum of Rh (152.5) plus Se (327.7). Further degradation to CG caused a profoundly negative NECB (−749.3), where Pn (88.2) offset only 11% of Rh (335.9) plus Se (501.5). Alteration to SR led to a partially recovered NECB (122.6) driven by shrub growth (Pn 554.5; Rh 143.9; Se 287.9), but recovery was limited given the lowered shrub biomass and topsoil loss via erosion. The results are consistent with previous findings that stand alteration from PU toward SR or CG through NU leads to as much as 49% loss of ecosystem carbon stocks. Preventing stand alteration and soil erosion are critical countermeasures against overbrowsing and subsequent erosion.
