Critical slowing down of the Amazon forest after increased drought occurrence.

Dynamic ecosystems, such as the Amazon forest, are expected to show critical slowing down behavior, or slower recovery from recurrent small perturbations, as they approach an ecological threshold to a different ecosystem state. Drought occurrences are becoming more prevalent across the Amazon, with known negative effects on forest health and functioning, but their actual role in the critical slowing down patterns still remains elusive. In this study, we evaluate the effect of trends in extreme drought occurrences on temporal autocorrelation (TAC) patterns of satellite-derived indices of vegetation activity, an indicator of slowing down, between 2001 and 2019. Differentiating between extreme drought frequency, intensity, and duration, we investigate their respective effects on the slowing down response. Our results indicate that the intensity of extreme droughts is a more important driver of slowing down than their duration, although their impacts vary across the different Amazon regions. In addition, areas with more variable precipitation are already less ecologically stable and need fewer droughts to induce slowing down. We present findings indicating that most of the Amazon region does not show an increasing trend in TAC. However, the predicted increase in extreme drought intensity and frequency could potentially transition significant portions of this ecosystem into a state with altered functionality.

Predicting the responses of European grassland communities to climate and land cover change.

European grasslands are among the most species-rich ecosystems on small spatial scales. However, human-induced activities like land use and climate change pose significant threats to this diversity. To explore how climate and land cover change will affect biodiversity and community composition in grassland ecosystems, we conducted joint species distribution models (SDMs) on the extensive vegetation-plot database sPlotOpen to project distributions of 1178 grassland species across Europe under current conditions and three future scenarios. We further compared model accuracy and computational efficiency between joint SDMs (JSDMs) and stacked SDMs, especially for rare species. Our results show that: (i) grassland communities in the mountain ranges are expected to suffer high rates of species loss, while those in western, northern and eastern Europe will experience substantial turnover; (ii) scaling anomalies were observed in the predicted species richness, reflecting regional differences in the dominant drivers of assembly processes; (iii) JSDMs did not outperform stacked SDMs in predictive power but demonstrated superior efficiency in model fitting and predicting; and (iv) incorporating co-occurrence datasets improved the model performance in predicting the distribution of rare species. This article is part of the theme issue 'Ecological novelty and planetary stewardship: biodiversity dynamics in a transforming biosphere'.

Microclimate reveals the true thermal niche of forest plant species.

Species distributions are conventionally modelled using coarse-grained macroclimate data measured in open areas, potentially leading to biased predictions since most terrestrial species reside in the shade of trees. For forest plant species across Europe, we compared conventional macroclimate-based species distribution models (SDMs) with models corrected for forest microclimate buffering. We show that microclimate-based SDMs at high spatial resolution outperformed models using macroclimate and microclimate data at coarser resolution. Additionally, macroclimate-based models introduced a systematic bias in modelled species response curves, which could result in erroneous range shift predictions. Critically important for conservation science, these models were unable to identify warm and cold refugia at the range edges of species distributions. Our study emphasizes the crucial role of microclimate data when SDMs are used to gain insights into biodiversity conservation in the face of climate change, particularly given the growing policy and management focus on the conservation of refugia worldwide.

The essential role of biodiversity in the key axes of ecosystem function.

Biodiversity is essential for maintaining the terrestrial ecosystem multifunctionality (EMF). Recent studies have revealed that the variations in terrestrial ecosystem functions are captured by three key axes: the maximum productivity, water use efficiency, and carbon use efficiency of the ecosystem. However, the role of biodiversity in supporting these three key axes has not yet been explored. In this study, we combined the (i) data collected from more than 840 vegetation plots across a large climatic gradient in China using standard protocols, (ii) data on plant traits and phylogenetic information for more than 2,500 plant species, and (iii) soil nutrient data measured in each plot. These data were used to systematically assess the contribution of environmental factors, species richness, functional and phylogenetic diversity, and community-weighted mean (CWM) and ecosystem traits (i.e., traits intensity normalized per unit land area) to EMF via hierarchical partitioning and Bayesian structural equation modeling. Multiple biodiversity attributes accounted for 70% of the influence of all the variables on EMF, and ecosystems with high functional diversity had high resource use efficiency. Our study is the first to systematically explore the role of different biodiversity attributes, including species richness, phylogenetic and functional diversity, and CWM and ecosystem traits, in the key axes of ecosystem functions. Our findings underscore that biodiversity conservation is critical for sustaining EMF and ultimately ensuring human well-being.

Integrating multiple plant functional traits to predict ecosystem productivity.

Quantifying and predicting variation in gross primary productivity (GPP) is important for accurate assessment of the ecosystem carbon budget under global change. Scaling traits to community scales for predicting ecosystem functions (i.e., GPP) remain challenging, while it is promising and well appreciated with the rapid development of trait-based ecology. In this study, we aim to integrate multiple plant traits with the recently developed trait-based productivity (TBP) theory, verify it via Bayesian structural equation modeling (SEM) and complementary independent effect analysis. We further distinguish the relative importance of different traits in explaining the variation in GPP. We apply the TBP theory based on plant community traits to a multi-trait dataset containing more than 13,000 measurements of approximately 2,500 species in Chinese forest and grassland systems. Remarkably, our SEM accurately predicts variation in annual and monthly GPP across China (R2 values of 0.87 and 0.73, respectively). Plant community traits play a key role. This study shows that integrating multiple plant functional traits into the TBP theory strengthens the quantification of ecosystem primary productivity variability and further advances understanding of the trait-productivity relationship. Our findings facilitate integration of the growing plant trait data into future ecological models.

Disentangling drivers of litter decomposition in a multi-continent network of tree diversity experiments.

Litter decomposition is a key ecosystem function in forests and varies in response to a range of climatic, edaphic, and local stand characteristics. Disentangling the relative contribution of these factors is challenging, especially along large environmental gradients. In particular, knowledge of the effect of management options, such as tree planting density and species composition, on litter decomposition would be highly valuable in forestry. In this study, we made use of 15 tree diversity experiments spread over eight countries and three continents within the global TreeDivNet network. We evaluated the effects of overstory composition (tree identity, species/mixture composition and species richness), plantation conditions (density and age), and climate (temperature and precipitation) on mass loss (after 3 months and 1 year) of two standardized litters: high-quality green tea and low-quality rooibos tea. Across continents, we found that early-stage decomposition of the low-quality rooibos tea was influenced locally by overstory tree identity. Mass loss of rooibos litter was higher under young gymnosperm overstories compared to angiosperm overstories, but this trend reversed with age of the experiment. Tree species richness did not influence decomposition and explained almost no variation in our multi-continent dataset. Hence, in the young plantations of our study, overstory composition effects on decomposition were mainly driven by tree species identity on decomposer communities and forest microclimates. After 12 months of incubation, mass loss of the high-quality green tea litter was mainly influenced by temperature whereas the low-quality rooibos tea litter decomposition showed stronger relationships with overstory composition and stand age. Our findings highlight that decomposition dynamics are not only affected by climate but also by management options, via litter quality of the identity of planted trees but also by overstory composition and structure.

Predicting ecosystem productivity based on plant community traits.

With the rapid accumulation of plant trait data, major opportunities have arisen for the integration of these data into predicting ecosystem primary productivity across a range of spatial extents. Traditionally, traits have been used to explain physiological productivity at cell, organ, or plant scales, but scaling up to the ecosystem scale has remained challenging. Here, we show the need to combine measures of community-level traits and environmental factors to predict ecosystem productivity at landscape or biogeographic scales. We show how theory can extend the production ecology equation to enormous potential for integrating traits into ecological models that estimate productivity-related ecosystem functions across ecological scales and to anticipate the response of terrestrial ecosystems to global change.

Climatic legacy effects on the drought response of the Amazon rainforest.

Extreme precipitation and drought events are predicted to become more intense and more frequent over the Amazon rainforest. Because changes in forest dynamics could prompt strong feedback loops to the global climate, it is of crucial importance to gain insight into the response of tropical forests to these recurring extreme climatic events. Here, we evaluated the Amazon forest stability (resistance and resilience) to drought in the context of past dry and wet climatic events using MODIS EVI satellite imagery and cumulative water deficit anomalies. We observed large spatial differences in the occurrence of extreme climatic events from 1980 to 2019, with an increase in drought frequency in the central and northern Amazon and drought intensity in the southern Amazon basin. An increasing trend in the occurrence of wet events was found in the western, southern, and eastern Amazon. Furthermore, we found significant legacy effects of previous climatic events on the forest drought response. An extreme drought closely preceding another drought decreased forest resilience, whereas the occurrence of a recent drier-than-usual event also decreased the forest resistance to later droughts. Both wetter-than-usual and extreme wet events preceding an extreme drought increased the resistance of the forest, and with similar effects sizes as dry events, indicating that wet and dry events have similarly sized legacy effects on the drought response of tropical forests. Our results indicate that the predicted increase in drought frequency and intensity can have negative consequences for the functioning of the Amazon forest. However, more frequent wet periods in combination with these droughts could counteract their negative impact. Finally, we also found that more stable forests according to the alternative stable states theory are also more resistant and resilient to individual droughts, showing a positive relationship between the equilibrium and non-equilibrium stability dynamics.

Litter quality and stream physicochemical properties drive global invertebrate effects on instream litter decomposition.

Plant litter is the major source of energy and nutrients in stream ecosystems and its decomposition is vital for ecosystem nutrient cycling and functioning. Invertebrates are key contributors to instream litter decomposition, yet quantification of their effects and drivers at the global scale remains lacking. Here, we systematically synthesized data comprising 2707 observations from 141 studies of stream litter decomposition to assess the contribution and drivers of invertebrates to the decomposition process across the globe. We found that (1) the presence of invertebrates enhanced instream litter decomposition globally by an average of 74%; (2) initial litter quality and stream water physicochemical properties were equal drivers of invertebrate effects on litter decomposition, while invertebrate effects on litter decomposition were not affected by climatic region, mesh size of coarse-mesh bags or mycorrhizal association of plants providing leaf litter; and (3) the contribution of invertebrates to litter decomposition was greatest during the early stages of litter mass loss (0-20%). Our results, besides quantitatively synthesizing the global pattern of invertebrate contribution to instream litter decomposition, highlight the most significant effects of invertebrates on litter decomposition at early rather than middle or late decomposition stages, providing support for the inclusion of invertebrates in global dynamic models of litter decomposition in streams to explore mechanisms and impacts of terrestrial, aquatic, and atmospheric carbon fluxes.

Foliar optical traits capture physiological and phenological leaf plasticity in Tilia×euchlora in the urban environment.

Knowledge on the response of trees to the urban heat island (UHI) effect and soil sealing is currently limited, yet of vital importance in an era characterized by both climate change and urbanization. We investigated the physiological and phenological leaf plasticity of Tilia×euchlora trees to the UHI effect and soil sealing and explored the potential of leaf optical traits to quantify the magnitude of leaf plasticity. Temporal changes of leaf water content (LWC), specific leaf area (SLA), total chlorophyll (Chl) and carotenoids (Car) content, Car:Chl ratio and leaf reflectance for 46 Tilia×euchlora trees were measured along a soil sealing and urbanization gradient. The leaf functional traits displayed trait-specific temporal patterns during the growing season. We observed higher LWC and SLA but lower Chl and Car contents in the coolest zones. We found earlier autumn downregulation in Chl and Car content at paved sites compared to unsealed sites (maximum difference = 13 days). The magnitude of plasticity in relation to the UHI and soil sealing varied in leaf functional traits with largest variation observed in Chl (38%), followed by Car:Chl (31%), Car (29%), SLA (26%) and LWC (8%). The proposed spectral indices calculated using leaf reflectance measurements were able to track the spatiotemporal variations and phenology in the leaf functional traits. Our results clearly demonstrate the leaf plasticity of Tilia×euchlora trees, which provides Tilia×euchlora trees the necessary capacity to adapt to rapid changes in the urban environment. More importantly, we demonstrated the suitability of leaf optical traits to serve as a proxy of leaf functional traits for studying the spatiotemporal response of urban trees to environmental factors, which opens up new possibilities for large scale ecological studies using remote sensing.

Climatic conditions, not above- and belowground resource availability and uptake capacity, mediate tree diversity effects on productivity and stability.

Tree species diversity promotes multiple ecosystem functions and services. However, little is known about how above- and belowground resource availability (light, nutrients, and water) and resource uptake capacity mediate tree species diversity effects on aboveground wood productivity and temporal stability of productivity in European forests and whether the effects differ between humid and arid regions. We used the data from six major European forest types along a latitudinal gradient to address those two questions. We found that neither leaf area index (a proxy for light uptake capacity), nor fine root biomass (a proxy for soil nutrient and water uptake capacity) was related to tree species richness. Leaf area index did, however, enhance productivity, but negatively affected stability. Productivity was further promoted by soil nutrient availability, while stability was enhanced by fine root biomass. We only found a positive effect of tree species richness on productivity in arid regions and a positive effect on stability in humid regions. This indicates a possible disconnection between productivity and stability regarding tree species richness effects. In other words, the mechanisms that drive the positive effects of tree species richness on productivity do not per se benefit stability simultaneously. Our findings therefore suggest that tree species richness effects are largely mediated by differences in climatic conditions rather than by differences in above- and belowground resource availability and uptake capacity at the regional scales.

Global maps of soil temperature.

Research in global change ecology relies heavily on global climatic grids derived from estimates of air temperature in open areas at around 2 m above the ground. These climatic grids do not reflect conditions below vegetation canopies and near the ground surface, where critical ecosystem functions occur and most terrestrial species reside. Here, we provide global maps of soil temperature and bioclimatic variables at a 1-km2 resolution for 0-5 and 5-15 cm soil depth. These maps were created by calculating the difference (i.e. offset) between in situ soil temperature measurements, based on time series from over 1200 1-km2 pixels (summarized from 8519 unique temperature sensors) across all the world's major terrestrial biomes, and coarse-grained air temperature estimates from ERA5-Land (an atmospheric reanalysis by the European Centre for Medium-Range Weather Forecasts). We show that mean annual soil temperature differs markedly from the corresponding gridded air temperature, by up to 10°C (mean = 3.0 ± 2.1°C), with substantial variation across biomes and seasons. Over the year, soils in cold and/or dry biomes are substantially warmer (+3.6 ± 2.3°C) than gridded air temperature, whereas soils in warm and humid environments are on average slightly cooler (-0.7 ± 2.3°C). The observed substantial and biome-specific offsets emphasize that the projected impacts of climate and climate change on near-surface biodiversity and ecosystem functioning are inaccurately assessed when air rather than soil temperature is used, especially in cold environments. The global soil-related bioclimatic variables provided here are an important step forward for any application in ecology and related disciplines. Nevertheless, we highlight the need to fill remaining geographic gaps by collecting more in situ measurements of microclimate conditions to further enhance the spatiotemporal resolution of global soil temperature products for ecological applications.

Maintaining forest cover to enhance temperature buffering under future climate change.

Forest canopies buffer macroclimatic temperature fluctuations. However, we do not know if and how the capacity of canopies to buffer understorey temperature will change with accelerating climate change. Here we map the difference (offset) between temperatures inside and outside forests in the recent past and project these into the future in boreal, temperate and tropical forests. Using linear mixed-effect models, we combined a global database of 714 paired time series of temperatures (mean, minimum and maximum) measured inside forests vs. in nearby open habitats with maps of macroclimate, topography and forest cover to hindcast past (1970-2000) and to project future (2060-2080) temperature differences between free-air temperatures and sub-canopy microclimates. For all tested future climate scenarios, we project that the difference between maximum temperatures inside and outside forests across the globe will increase (i.e. result in stronger cooling in forests), on average during 2060-2080, by 0.27 ± 0.16 °C (RCP2.6) and 0.60 ± 0.14 °C (RCP8.5) due to macroclimate changes. This suggests that extremely hot temperatures under forest canopies will, on average, warm less than outside forests as macroclimate warms. This knowledge is of utmost importance as it suggests that forest microclimates will warm at a slower rate than non-forested areas, assuming that forest cover is maintained. Species adapted to colder growing conditions may thus find shelter and survive longer than anticipated at a given forest site. This highlights the potential role of forests as a whole as microrefugia for biodiversity under future climate change.

Global patterns and drivers of rainfall partitioning by trees and shrubs.

Spatiotemporal redistribution of incident rainfall in vegetated ecosystems results from the partitioning by plants into intercepted, stemflow, and throughfall fractions. However, variation in patterns and drivers of rainfall partitioning across global biomes remains poorly understood, which limited the ability of climate models to improve the predictions of biome hydrological cycle under global climate change scenario. Here, we synthesized and analyzed the partitioning of incident rainfall into interception, stemflow, and throughfall by trees and shrubs at the global scale using 2430 observations from 236 independent publications. We found that (1) globally, median levels of relative interception, stemflow, and throughfall accounted for 21.8%, 3.2%, and 73.0% of total incident rainfall, respectively; (2) rainfall partitioning varied among different biomes, due to variation in plant composition, canopy structure, and macroclimate; (3) relative stemflow tended to be driven by plant traits, such as crown height:width ratio, basal area, and height, while relative interception and throughfall tended to be driven by plant traits as well as meteorological variables. Our global assessment of patterns and drivers of rainfall partitioning underpins the role of meteorological factors and plant traits in biome-specific ecohydrological cycles. We suggest to include these factors in climate models to improve the predictions of local hydrological cycles and associated biodiversity and function responses to changing climate conditions.

Drivers of carbon stocks in forest edges across Europe.

Forests play a key role in global carbon cycling and sequestration. However, the potential for carbon drawdown is affected by forest fragmentation and resulting changes in microclimate, nutrient inputs, disturbance and productivity near edges. Up to 20% of the global forested area lies within 100 m of an edge and, even in temperate forests, knowledge on how edge conditions affect carbon stocks and how far this influence penetrates into forest interiors is scarce. Here we studied carbon stocks in the aboveground biomass, forest floor and the mineral topsoil in 225 plots in deciduous forest edges across Europe and tested the impact of macroclimate, nitrogen deposition and smaller-grained drivers (e.g. microclimate) on these stocks. Total carbon and carbon in the aboveground biomass stock were on average 39% and 95% higher at the forest edge than 100 m into the interior. The increase in the aboveground biomass stock close to the edge was mainly related to enhanced nitrogen deposition. No edge influence was found for stocks in the mineral topsoil. Edge-to-interior gradients in forest floor carbon changed across latitude: carbon stocks in the forest floor were higher near the edge in southern Europe. Forest floor carbon decreased with increasing litter quality (i.e. high decomposition rate) and decreasing plant area index, whereas higher soil temperatures negatively affected the mineral topsoil carbon. Based on high-resolution forest fragmentation maps, we estimate that the additional carbon stored in deciduous forest edges across Europe amounts to not less than 183 Tg carbon, which is equivalent to the storage capacity of 1 million ha of additional forest. This study underpins the importance of including edge influences when quantifying the carbon stocks in temperate forests and stresses the importance of preserving natural forest edges and small forest patches with a high edge-to-interior surface area.

Biogeographic historical legacies in the net primary productivity of Northern Hemisphere forests.

It has been suggested that biogeographic historical legacies in plant diversity may influence ecosystem functioning. This is expected because of known diversity effects on ecosystem functions, and impacts of historical events such as past climatic changes on plant diversity. However, empirical evidence for a link between biogeographic history and present-day ecosystem functioning is still limited. Here, we explored the relationships between Late-Quaternary climate instability, species-pool size, local species and functional diversity, and the net primary productivity (NPP) of Northern Hemisphere forests using structural equation modelling. Our study confirms that past climate instability has negative effects on plant functional diversity and through that on NPP, after controlling for present-day climate, soil conditions, stand biomass and age. We conclude that global models of terrestrial plant productivity need to consider the biogeographical context to improve predictions of plant productivity and feedbacks with the climate system.

Mapping tree species vulnerability to multiple threats as a guide to restoration and conservation of tropical dry forests.

Understanding the vulnerability of tree species to anthropogenic threats is important for the efficient planning of restoration and conservation efforts. We quantified and compared the effects of future climate change and four current threats (fire, habitat conversion, overgrazing and overexploitation) on the 50 most common tree species of the tropical dry forests of northwestern Peru and southern Ecuador. We used an ensemble modelling approach to predict species distribution ranges, employed freely accessible spatial datasets to map threat exposures, and developed a trait-based scoring approach to estimate species-specific sensitivities, using differentiated trait weights in accordance with their expected importance in determining species sensitivities to specific threats. Species-specific vulnerability maps were constructed from the product of the exposure maps and the sensitivity estimates. We found that all 50 species face considerable threats, with an average of 46% of species' distribution ranges displaying high or very high vulnerability to at least one of the five threats. Our results suggest that current levels of habitat conversion, overexploitation and overgrazing pose larger threats to most of the studied species than climate change. We present a spatially explicit planning strategy for species-specific restoration and conservation actions, proposing management interventions to focus on (a) in situ conservation of tree populations and seed collection for tree planting activities in areas with low vulnerability to climate change and current threats; (b) ex situ conservation or translocation of populations in areas with high climate change vulnerability; and (c) active planting or assisted regeneration in areas under high current threat vulnerability but low climate change vulnerability, provided that interventions are in place to lower threat pressure. We provide an online, user-friendly tool to visualize both the vulnerability maps and the maps indicating priority restoration and conservation actions.

Litter quality and the law of the most limiting: Opportunities for restoring nutrient cycles in acidified forest soils.

The adverse effects of soil acidification are extensive and may result in hampered ecosystem functioning. Admixture of tree species with nutrient rich litter has been proposed to restore acidified forest soils and improve forest vitality, productivity and resilience. However, it is common belief that litter effects are insufficiently functional for restoration of poorly buffered sandy soils. Therefore we examined the effect of leaf litter on the forest floor, soil chemistry and soil biota in temperate forest stands along a range of sandy soil types in Belgium, the Netherlands and Germany. Specifically, we address: i) Which tree litter properties contribute most to the mitigation of soil acidification effects and ii) Do rich litter species have the potential to improve the belowground nutrient status of poorly buffered, sandy soils? Our analysis using structural equation modelling shows that litter base cation concentration is the decisive trait for the dominating soil buffering mechanism in forests that are heavily influenced by atmospheric nitrogen (N) deposition. This is in contrast with studies in which leaf litter quality is summarized by C/N ratio. We suggest that the concept of rich litter is context dependent and should consider Liebig's law of the most limiting: if N is not limiting in the ecosystem, litter C/N becomes of low importance, while base cations (calcium, magnesium, potassium) become determining. We further find that on poorly buffered soils, tree species with rich litter induce fast nutrient cycling, sustain higher earthworm biomass and keep topsoil base saturation above a threshold of 30%. Hence, rich litter can trigger a regime shift to the exchange buffer domain in sandy soils. This highlights that admixing tree species with litter rich in base cations is a promising measure to remediate soil properties on acidified sandy soils that receive, or have received, high inputs of N via deposition.

Disentangling how management affects biomass stock and productivity of tropical secondary forests fallows.

A better understanding of biomass production in secondary forests after cultivation is essential for assessing the resilience of slash and burn systems and their capacity to deliver ecosystem services. Biomass production is influenced by management legacies, landscape configuration and soil, but these drivers are rarely studied simultaneously, nor is the role of changes in vegetation properties in linking them to biomass production. We assessed how management legacies affect biomass in secondary forests created by slash and burn in the central Congo Basin, and tested whether changes in productivity could be attributed to changes in stem density, functional diversity, functional identity or soil. Using data from 6452 trees in 96 fallow plots nested in 3 study sites, we looked for the main determinants of aboveground biomass (AGB) of woody vegetation in fallow systems. Next, using a subset of 58 plots in fallow fields aged 5 to 10 years, we used confirmatory path analysis to explore the relations between management history, soil, vegetation properties and biomass productivity. The sampled fallow fields had, on average, 58.4 (±46.2) Mg ha-1 AGB. AGB was positively related to both fallow age and to the proportion of remnant trees in AGB and negatively related to the number of previous cultivation cycles. Biomass productivity varied with the number of previous slash-and-burn cycles, with notable declines in the fourth cycle. The effect of management history was mainly through a reduction in the dominance of fast growing tree species and in the number of regenerating stems, which were also indirectly affected by an increase in C. odorata cover. Soil fertility status and the biomass of remnant trees also modified biomass productivity. Our findings suggest that under the current management intensity the capacity of the slash and burn system to provide important ecosystem functions, such as carbon sequestration, is declining.

Tree diversity mitigates defoliation after a drought-induced tipping point.

Understanding the processes that underlie drought-related tree vitality loss is essential for anticipating future forest dynamics, and for developing management plans aiming at increasing the resilience of forests to climate change. Forest vitality has been continuously monitored in Europe since the acid rain alert in the 1980s, and the intensive monitoring plots of ICP Forests offer the opportunity to investigate the effects of air pollution and climate change on forest condition. By making use of over 100 long-term monitoring plots, where crown defoliation has been assessed extensively since 1990, we discovered a progressive shift from a negative to a positive effect of species richness on forest health. The observed tipping point in the balance of net interactions, from competition to facilitation, has never been reported from real ecosystems outside experimental conditions; and the strong temporal consistency of our observations with increasing drought stress emphasizes its climate change relevance. Furthermore, we show that higher species diversity has reduced the severity of defoliation in the long term. Our results confirm the greater resilience of diverse forests to future climate change-induced stress. More generally, they add to an accumulating body of evidence on the large potential of tree species mixtures to face manifold disturbances in a changing world.