Global Change Biology

Extreme climate events are rapidly reshaping ecosystems and contributing to biodiversity loss, yet their long-term impacts remain poorly understood, particularly in marine systems. In 2011, an unprecedented marine heatwave (MHW) on the west coast of Australia drove declines in habitat-forming plants and across taxa spanning multiple trophic levels. Shark Bay, home to an iconic Indo-Pacific bottlenose dolphin (Tursiops aduncus) population, endured temperatures 3-5{degrees}C above long-term means, triggering catastrophic seagrass loss. Using 20 years (2003-2023) of dolphin photo-identification data, we fitted a Bayesian Hidden Markov Model to estimate annual, age-class specific survival probabilities, abundance and recruitment rate, and their association with MHW-mediated seagrass loss in the western (WSB) and eastern (ESB) gulfs of Shark Bay. Adult survival declined markedly, from 0.99 (0.98-1.00) pre-MHW to 0.85 (0.79-0.90) post-MHW in WSB, and from 0.96 (0.92-0.98) to 0.89 (0.85-0.95) in ESB, remaining below pre-MHW levels for several years, while juvenile and calf survival declined in both gulfs to a lesser extent. Recruitment declined in both gulfs and, although recovery was evident in WSB from 2019 onward, continued to decline in ESB until the end of the study. These sustained demographic declines drove substantial reductions in population growth and abundance in both WSB (39%) and ESB (36%). Demographic changes were associated with seagrass loss, here a proxy for habitat and prey availability, with clearest support for adult survival in WSB and a two-year-lagged recruitment response in ESB. This pathway is likely to operate across diverse ecosystems as climate change drives increasing loss of habitat-forming species. The severity and duration of these demographic impacts illustrate that habitat change can cascade through trophic systems to overwhelm demographic resilience in upper-trophic species. This has negative implications for even long-lived, K-selected species, whose slow demographic recovery may not keep pace with the accelerating frequency and intensity of climate disturbance.

Land-use and climate change are dominant drivers of terrestrial biodiversity change, yet are usually analysed separately although they co-occur and may interact. We integrated global bird assemblage records from PREDICTS with temperature anomalies and species' ecological characteristics to test how land use modifies biodiversity responses to climate warming across three dimensions: local abundance and richness, compositional dissimilarity among assemblages, and species occurrence across ecological groups. Land use modified warming associations for both abundance and richness. Relative to primary vegetation, warming associations were more positive in low-intensity agriculture for both responses, but more negative for richness in high-intensity agriculture and urban land. Predicted richness remained below the reference, defined as primary vegetation at temperature anomaly 0, across all modified land uses, with the largest declines in high-intensity agriculture (-17.9%) and urban land (-21.8%). Within land-use classes, greater warming exposure was generally associated with higher compositional similarity, consistent with warming associated biotic homogenisation. Occurrence models revealed ecological filtering: broad-niched generalists were favoured in warmer modified habitats, whereas range-restricted habitat specialists were filtered out under intensive land use and warming. Conservation assessments should therefore integrate land-use-warming interactions with changes in assemblage composition and ecological strategies, particularly in agricultural and urban landscapes where warming further disadvantaged range-restricted specialists.

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.

Understanding how carbon-water coupling scales from leaves to ecosystems remains a major challenge because water-use efficiency (WUE) metrics integrate physiological, hydrological, and ecological processes. We integrated tree-ring width, triple stable isotopes ({delta}13C, {delta}18O and {delta}2H), eddy covariance (EC) observations, and a process-based model to investigate the drivers of intrinsic (iWUE) and ecosystem water-use efficiency (WUEeco) across biological scales in a mature Fagus sylvatica forest between 2004 and 2023. Modelled and isotope-derived iWUE were strongly correlated (r = 0.61, p < 0.05) and both indicated an approximately 20% long-term increase, consistent with increasing gross primary productivity. In contrast, WUEeco derived from EC was lower than model estimates and exhibited a positive trend not reproduced by the model. This divergence primarily reflected differences in biological representation and hydrological integration, as the single-species model underestimated ecosystem evapotranspiration by excluding co-occurring deciduous and evergreen species together with understory vegetation. Triple-isotope analyses showed that the long-term increase in iWUE was driven by enhanced photosynthetic assimilation rather than sustained stomatal closure, whereas {delta}18O and {delta}2H variability was mainly controlled by source-water dynamics. These findings provide a mechanistic framework for reconciling leaf- and ecosystem-scale WUE and improving predictions of forest carbon-water coupling under climate change.

Reforestation is a leading natural climate solution, but bulk soil organic carbon (C) often responds slowly, obscuring early belowground change. We tested whether soil aggregates reveal early structural reorganization that affects C retention pathways at a three-year-old experimental reforestation planting established on former pasture in Oregon, USA. We sampled soils at 0-20 and 20-40 cm beneath incense cedar (Calocedrus decurrens; arbuscular mycorrhizal [AM]), black cottonwood (Populus trichocarpa; AM and ectomycorrhizal [EcM]), ponderosa pine (Pinus ponderosa; EcM), and treeless controls. We measured bulk soil C concentration and C:N, aggregate size distribution and mean weight diameter (MWD), fraction-associated C, and, in a subset of surface aggregate fractions, natural-abundance {delta}13C to evaluate soil C pools, physical structure, C distribution, and C processing. After three years, reforestation did not produce significant differences in bulk soil C between planted trees and treeless controls. In surface soil, MWD averaged 36% higher under incense cedar and 58% higher under black cottonwood than under controls, whereas ponderosa pine remained similar to controls. The clearest treatment differences in C distribution occurred in macroaggregates. Incense cedar and black cottonwood had higher large (>2000 m) macroaggregate-associated C than controls, while black cottonwood combined a greater proportion of large macroaggregates with lower C concentrations, indicating that structural development and C accumulation were partly decoupled. Species patterns were broadly consistent with stronger early aggregate responses under AM-compatible species in former-pasture legacy conditions; by contrast, the EcM-associated ponderosa pine remained closer to treeless controls across multiple aggregate measures. Treatment effects weakened with depth and were limited in microaggregates (250-53 m) and silt-and-clay (<53 m) fractions. Smaller aggregate fractions were progressively enriched in {delta}13C, consistent with greater C processing in protected fractions, but treatment-level isotope differences had not yet emerged. Soil aggregates showed species-associated belowground reorganization, revealing how tree identity and former-pasture legacy may shape early soil C organization before consistent differences emerged in bulk soil C. Aggregate measures can complement bulk soil C measurements in restoration monitoring as early indicators of belowground soil C trajectories shaped by tree species identity and land-use legacy.

Ecological responses to environmental stressors are increasingly shaped by interactions among anthropogenic and climatic drivers, yet the influence of prior exposure history on subsequent stress responses remains poorly integrated into multiple stressor ecology. Using the blue mussel (Mytilus edulis) as a model system, we experimentally tested whether prior exposure to environmental nanoplastics modifies responses to subsequent stressors, including copper contamination, nutrient enrichment, and salinity fluctuations, individually and in combination. Nanoplastic pre-exposure reduced survival across subsequent treatments and altered stressor interaction outcomes at the community level, shifting responses toward synergistic effects that amplified metabolic costs. These results demonstrate that exposure history can modify both the magnitude and direction of multi-stressor effects, with consequences extending beyond individual performance to ecosystem functioning. By incorporating temporal exposure dynamics into multiple stressor research, our study highlights how ecological responses to global change depend not only on the identity and intensity of stressors, but also on prior environmental conditions shaping subsequent organismal and ecosystem health status and responses.

The increased frequency and intensity of coral bleaching events expose reefs to disturbance-recovery cycles that threaten their long-term dynamics. Understanding the impact of thermal stress requires evaluating coral mortality and recovery rates after disturbances. By integrating two long-term monitoring datasets, we examined coral mortality and recovery patterns at 13 sites in Mo'orea, French Polynesia, following the 2019 bleaching event, the most severe heat stress on record for the island. The maximum accumulated heat stress (AHSmax) varied from 5.76 to 6.26 {degrees} C weeks and enabled a direct analysis of heat-stress-related mortality. One year after bleaching, Pocillopora mortality increased with AHSmax, reducing mean coral cover from 35% to 17%. Regardless of AHSmax, Acropora mortality was high at all sites (median: 84%), reducing mean cover from 5% to 1% and indicating greater heat tolerance of Pocillopora than Acropora. Within four years after the bleaching event, no recovery occurred, and coral cover further declined relative to pre-bleaching levels. This decline was stronger for sites exposed to more intense heat stress. The impact of colony size was assessed by tracking 3090 individual Pocillopora colonies over six years and estimating size-dependent mortality and growth with a Bayesian mixed model. Immediately after the bleaching event, large Pocillopora corals exhibited a higher probability of complete and partial mortality than smaller colonies, highlighting size-specific consequences of bleaching. Even three years after the heat stress ended, large Pocillopora corals remained stressed, as evidenced by the continuously high probability of partial mortality. The lack of recovery observed here may be related to increased spatial competition with macroalgae. The disappearance of large fecund colonies and high mortality of coral recruits could have catastrophic consequences for the resilience of Mo'orea's reefs.

Mediterranean forests are emerging as one of the clearest hotspots of climate-change impacts, where rapid warming and increasingly severe droughts and heatwaves are pushing forest ecosystems toward their physiological and functional limits. These pressures threaten productivity, carbon dynamics, ecosystem stability, and the services forests provide. In this context, a process-based forest ecosystem model was used to assess climate-driven changes in carbon dynamics and productivity stability across Italian forests within the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) framework. 5135 plots were simulated, corresponding to ~80% of the field plots included in the second Italian National Forest Inventory (INFI). Model simulations were first evaluated against independent observations of gross primary production (GPP) from satellite-based products and forest carbon variables from the third INFI. Forest dynamics from 2015 to 2100 were simulated under three CMIP6-based emission scenarios (SSP1-2.6, SSP3-7.0, and SSP5-8.5) using two climate forcings (GFDL-ESM4 and UKESM1-0-LL). Future responses were assessed for GPP, aboveground biomass, growing stock volume, and carbon stocks, while variability in GPP was quantified using the coefficient of variation as an indicator of forest stability. Results show that the model reproduced the main spatial and temporal patterns of forest productivity and carbon stocks. Future projections revealed strong spatial heterogeneity and increasing divergence among scenarios after mid-century. GPP showed the earliest and strongest response to climate change, with mountain forests generally maintaining or increasing productivity and exhibiting greater stability, whereas coastal forests showed pronounced productivity declines and increasing variability, particularly under higher-emission scenarios and more extreme climate forcing. In contrast, carbon stocks generally continued to increase throughout the century, although carbon accumulation progressively slowed toward the end of the century, particularly under higher-emission scenarios.

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.

Disturbances alter ecosystem resilience by killing organisms and leaving behind material legacies. We tested how material legacies from marine heat waves and cyclones govern transitions between coral- and seaweed-rich states on tropical reefs. Heat waves kill coral tissue but leave their skeletons to erode gradually, whereas cyclones scour away emergent biota. Field experiments revealed that dead skeletons protect seaweeds from herbivory, catalyzing a transition to a seaweed-rich state that becomes self-sustaining after skeletons disappear. After four years, 73% of skeleton-added (heat wave) plots transitioned to seaweeds compared to 0% of scoured (cyclone) plots that lacked skeletons, while 87% of seaweed-rich plots scoured to simulate a cyclone stayed seaweed-free. Reef trajectories after natural disturbances were similar: four years after heat wave-induced coral mortality, 63% of time-series plots had transitioned to seaweeds with none recovering to coral; four years after a cyclone removed skeletons from a prior disturbance, none had transitioned to seaweeds and 60% had returned to coral. A disturbance not only can displace an ecosystem state, it can rapidly reshape the resilience landscape through the material legacies it creates or removes. Our findings also show that successive disturbances need not have cumulative negative effects: after a reef is driven into a seaweed-rich state, a cyclone can act as a rescue disturbance by removing established seaweeds and, when structural refugia are also removed, shrinking the seaweed stability domain. By reshaping resilience through contrasting material legacies, disturbance type and sequence can be as important as frequency and severity in determining ecosystem trajectories.

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.

Secondary forests are increasingly important for biodiversity recovery and ecosystem functioning after disturbance. As vegetation regenerates, accompanying shifts in soil conditions and microbial communities regulate climate-relevant processes such as atmospheric methane (CH$_4$) uptake. Whether these belowground dimensions recover synchronously, however, remains unclear. Here we show that belowground recovery followed asynchronous trajectories across a forest chronosequence in Metok, Eastern Himalayas. Soil inorganic properties and microbial community composition converged progressively on the primary-forest state as soils acidified. By contrast, soil organic carbon, microbial diversity and abundance appeared to recover by 25 years but diverged again at 40 years. Co-occurrence network topology resembled the primary forest by 40 years, yet the keystone taxa underlying it were entirely stage specific. Atmospheric CH$_4$ uptake fluctuated: uptake recovered to primary-forest levels in 25-year stands but weakened in 40-year stands to rates comparable with the recently disturbed 5-year forest. Rather than canonical methane-cycling taxa, four amplicon sequence variants---three assigned to \textit{Mycobacterium}---tracked CH$_4$ flux closely while acting as hubs or connectors in the co-occurrence networks, warranting genomic and experimental work to resolve their function. Apparent recovery within the first 25 years therefore did not persist, cautioning against inferring restoration success from early succession alone and highlighting the need for longer chronosequences and repeated observation to resolve the dynamics of secondary-forest recovery.

Mangrove reforestation is expanding across the Caribbean, but restoration success is still judged almost exclusively by tree survival and canopy cover. Whether the sediment microbiome, which governs carbon burial and nutrient cycling, follows the recovery of the vegetation remains largely unknown. We combined 16S rRNA gene metabarcoding (V4, Illumina MiSeq) with CHNS elemental analysis, pH and salinity measurements to compare sediment prokaryotic communities along 0-100 cm profiles in a reference, a reforested (Rhizophora mangle plantings since 2019) and an altered mangrove within the Las Dunas de Las Calderas Natural Monument, Dominican Republic (39 cores pooled into 28 depth-resolved composite samples). Reference sediments contained 2.5-5.5-fold more organic carbon (median 7.5% vs 3.0% in altered and 1.3% in reforested sediments), had higher C/N ratios and were less alkaline (pH 8.2 vs 8.8). Reference communities were more diverse and even, and differed in composition from both disturbed stands (pairwise PERMANOVA, R^2 ~ 0.09, adjusted P = 0.0015), whereas reforested and altered communities were statistically indistinguishable (R^2 = 0.04, P = 0.25). Disturbed sediments were dominated by Cyanobacteriota (~41-45% of reads), which tracked high pH, whereas reference sediments were enriched in Chloroflexota, Thermodesulfobacteriota and Thermoproteota, lineages linked to anaerobic degradation of complex organic matter and sulfate reduction. Depth structured communities by separating surface from subsurface layers, and sediment organic carbon and nitrogen were the only significant constraints in distance-based redundancy analysis. These results indicate that, within a few years of the first plantings (2019), reforestation had not restored the carbon-rich, anaerobic sediment habitat of the reference forest. We propose sediment organic carbon and microbial community structure as complementary indicators for mangrove restoration monitoring in the Caribbean.

Climate change poses a major threat to marine mammal populations, many of which remain vulnerable after years of overexploitation during the 17th to 20th centuries. Extreme climate events have grown in both frequency and severity in recent decades, and this trend is projected to accelerate under continued greenhouse gas emissions. Here, we assess how this intensification affects a top marine predator at the epicentre of extreme El Nino events - the Peruvian population of Otaria byronia (South American sea lion). We first apply a Bayesian population model that best predicts observed population dynamics to reconstruct female O. byronia population trends, estimate vital rates, and evaluate the effects of environmental drivers on survival and reproduction. We then use the same model to project extirpation risk under multiple future climate scenarios reflecting anticipated increases in both the intensity and frequency of extreme El Nino events. Our reconstruction reveals that the Peruvian O. byronia population has shown considerable resilience to episodic El Nino events over the past quarter-century, likely sustained by high female survival during normal and favourable years. However, under a scenario of moderate increases in El Nino frequency and intensity, the probability of extirpation over the next 50 years is 83.4% higher than under current conditions. As disturbance regimes intensify globally, we suggest that resilience-oriented management frameworks, grounded in species-specific demographic data and environmental context, as demonstrated here, will be essential for safeguarding marine mammal populations against an increasingly volatile climate.

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.

Marine heatwaves cause widespread coral bleaching and mortality, yet thermal responses often differ between adult and juvenile corals, with the latter frequently being more tolerant. Whether the observed enhanced tolerance of juveniles reflects intrinsic physiological differences across life stages or simply reduced environmental stress due to shaded microhabitats occupied by juveniles, or a combination of both factors, remains unclear. To test whether colony size determines thermal responses and if a threshold size exists for tolerance, we exposed 89 colonies of the reef-building coral Acropora aff. digitifera to a 6-week, 32.5 C simulated marine heatwave under controlled light conditions, with cumulative stress reaching 20 C-weeks. Contrary to expectations, ~2 C-weeks less heat stress was required to trigger bleaching and mortality in coral juveniles compared to adults. Mortality divergence across colony sizes was greatest at ~10 C-weeks, when the smallest corals exceeded 80% mortality, whereas the largest colonies showed <10% mortality. A breakpoint in survival odds was apparent at 13.4 cm diameter, corresponding to the size at onset of sexual maturity in the native wild population for this species. These results indicate an ontogenetic shift in physiological heat tolerance, with larger reproductively mature colonies showing enhanced tolerance. The apparent survival advantage of juveniles in the reef frequently reported during bleaching events could be more likely due to micro-environmental buffering than intrinsic physiological tolerance. Our findings suggest that juvenile corals may be more vulnerable to climate change than previously thought, particularly as declining habitat complexity could reduce microhabitat refugia on future reefs.

Palm oil is the worlds leading vegetable oil, with production continuing to expand, particularly across Southeast Asia. Within established plantations, conserving and restoring riparian margins can reduce erosion, improve connectivity, and support biodiversity. Simultaneously, climate change is increasing thermal stress within tropical agricultural landscapes, raising the potential importance of riparian buffers as microclimatic refugia, although this has not previously been evaluated. The Riparian Ecosystem Restoration in Tropical Agriculture (RERTA) Project is a long-term Before-After Control-Impact experiment established within oil palm plantations in Riau, Indonesia, to evaluate alternative riparian restoration strategies following plantation replanting. Treatments included passive restoration (retaining mature oil palms and allowing natural regeneration), active restoration (enrichment planting with six native tree species), and a control (replanting immature oil palms to the river edge). Six years after restoration, we assessed vegetation structure using terrestrial laser scanning and thermal stability using temperature loggers and thermal imaging. We also evaluated the long-term temperature buffering capacity of each treatment under current and projected climates using CHELSA climate data and downscaled outputs from multiple General Circulation Models under Shared Socioeconomic Pathway scenarios for 2045 and 2085. Active restoration produced greater structural complexity, a more closed canopy, and cooler, more thermally stable conditions than passive restoration and control treatments. Restored margins also contained significantly more cool surface microhabitats, reducing thermal extremes under both current and future climate scenarios. Our findings demonstrate that active restoration of riparian margins can enhance thermal buffering and create microclimatic refugia for biodiversity within tropical oil palm landscapes. As agricultural intensification and climate change continue to reshape tropical ecosystems, restoring structurally complex riparian vegetation will be critical for sustaining biodiversity and the ecosystem services upon which agricultural production depends.

Earth's biodiversity is distributed unevenly, typically peaking in warm, vegetated, and stable environments. This has frequently been linked to energy availability, which boosts productivity, bolsters populations, and facilitates coexistence. Here, we examine the relationship between energy and biodiversity across coastal oceans using particulate organic matter biogeochemistry and environmental DNA. We reveal that energy quantity (concentrations of carbon, nitrogen, and hydrolyzable amino acids) and salinity, two variables that co-vary with freshwater inflow, jointly predict species richness at regional and global scales. Biodiversity was lowest at sites and locations with high energy inflow and low salinity, suggesting that environmental filtering associated with the osmotic stress induced by freshwater inflow outweighs raw resource availability in governing species richness. Independent of this gradient, however, energy quality (defined via C:N ratios, {delta}C values, and amino acid profiles) was associated with higher biodiversity. Labile, protein-rich, marine-derived resources supported higher biodiversity across all taxa at the regional scale, and a larger number of planktonic consumer taxa at both regional and global scales. High energy quality sites were also enriched in key planktonic groups such as calanoid and cyclopoid copepods, suggesting that these sites act as hotspots of planktonic biodiversity across coastal seascapes. Our findings suggest that few species can directly harness the plentiful resources provided by terrestrial subsidies in coastal oceans because they arrive in low salinity waters. In turn, high-quality food attracts a diverse range of consumers, which may seed the patchy foraging hotspots that characterize open-water food webs. Altered coastal hydrodynamics and biogeochemistry may therefore affect nearshore biodiversity, food webs, and fisheries.

Forecasting infectious disease dynamics is important for ecological management and emerging disease preparedness, yet many systems lack the long-term datasets required to develop reliable predictions. Spatial replication may provide an alternative by allowing information to be shared among populations, but the extent to which this improves forecasting remains unclear. Here, I used epidemic and temperature data from 20 replicated semi-natural Daphnia-parasite pond populations monitored across four seasons (80 epidemics) to test whether information from multiple populations improves forecasts and whether environmental similarity identifies informative populations. I forecasted disease prevalence, infected host density, and healthy host density using a benchmark model, autoregressive integrated moving average (ARIMA), and time-series regression models. Models were trained using focal-population data alone, mean dynamics across other populations, or temperature-weighted mean dynamics from environmentally similar populations. Epidemics showed strong seasonal structure, with prevalence and infected host density generally peaking during the warmest period of the season. However, no forecasting approach consistently outperformed others across all variables. Temperature improved forecasts of prevalence and infected host density but not healthy host density. Multi-population training improved forecasts in some cases, particularly for ARIMA models, whereas temperature-weighted averaging provided no additional benefit over simple averaging. Regression models consistently produced the most accurate forecasts of healthy host density. These results show that borrowing information among populations can improve forecasts under specific conditions but is not universally beneficial, highlighting the importance of matching forecasting approaches and data sources to the ecological processes being predicted.

The consequences of maternal effects for offspring fitness can depend on the environmental conditions offspring encounter and the fitness traits considered. We used a transgenerational experiment in Daphnia magna to test how maternal exposure to warming and elevated salinity influenced offspring survival, reproduction and body size across matching, mismatching and multiple-stressor environments. Maternal exposure to salinity enhanced offspring survival under matching salinity, and benefits extended to mismatching environments. In contrast, maternal exposure to warming provided no survival benefit under matching warming and reduced survival in mismatching environments, while accelerating reproduction across contexts. Multiple-stressor conditions revealed pronounced maternal effects, while contrasting responses across survival, reproduction and body size showed that benefits to one fitness trait did not extend to others. Our findings demonstrate that environmental matching alone is insufficient to predict maternal effects, whose consequences depend on offspring environmental complexity and the fitness trait considered.