Search

Representations of interactions between the C and N cycles in terrestrial ecosystems are now implemented in a majority of state-of-the-art Dynamic Global Vegetation Models (C-N models). Standard models for simulating the response of individual processes to changes in N availability have not yet emerged and widely used models have not been tested against the full diversity of empirical data. Large remaining model structural uncertainty has important implications for projections and hindcasts of the land C uptake.Here, we summarise the current state of global land C balance simulations by comparing C-N models to C-only models; summarise data from field surveys and experiments to elucidate the role of soil N in controlling photosynthesis and its acclimation, stoichiometry, allocation, and growth; and demonstrate how optimality principles can guide the representation of acclimation and allocation for simulating ecosystem responses to experimental treatments of CO2 and soil N – consistent with observations. Promising model results are achieved by assuming that the atmospheric environment, including CO2, is the principal driver for photosynthetic capacities and leaf N following optimality theory of photosynthetic acclimation (Prentice et al., 2014). In turn, the functional balance hypothesis (Bloom et al., 1985) yields accurate predictions for how soil N availability and CO2 influence allocation and growth in different tissues.Our results show how confronting new theoretical approaches to simulating ecosystem C-N interactions against the collective constraints from diverse types of observations can guide model development and potentially reduce the large uncertainty in global carbon cycle projections.ReferencesBloom, Arnold J, F Stuart Chapin, and Harold A Mooney. “Resource Limitation in Plants--An Economic Analogy” Annual Review of Ecology and Systematics, 16, no. 1 (1985): 363–92. https://doi.org/10.1146/annurev.es.16.110185.002051.Prentice, I. Colin, Ning Dong, Sean M. Gleason, Vincent Maire, and Ian J. Wright. “Balancing the Costs of Carbon Gain and Water Transport: Testing a New Theoretical Framework for Plant Functional Ecology.” Ecology Letters 17, 1 (2014): 82–91. https://doi.org/10.1111/ele.12211.

Acclimation of photosynthesis allows plants to adjust to seasonal environmental changes and gradual long-term changes of similar magnitude. However, current global photosynthesis models diverge in their representation of temperature acclimation. Here, we applied fundamental principles of acclimation to identify key processes for predicting the response of photosynthesis to global warming. We investigated how the instantaneous temperature response of photosynthesis changes due to acclimation of the photosynthetic capacities (changes in base rates of carboxylation, electron transport and respiration), the stomatal response (sensitivity to changes in vapour pressure), and the enzymatic response (changes in the enzyme kinetics of carboxylation and electron transport). Using a dataset of gas exchange measurements from globally distributed sites, we compared predicted and observed relationships between prevalent growth temperature (Tgrowth) and the optimal temperature of photosynthesis (Topt), the photosynthesis rate at Topt (Aopt), and the temperature sensitivity (width of the temperature response curve, Tspan). The observational data showed a significant linear increase of Topt with Tgrowth (0.74 °C/°C) and no correlations between Tgrowth and Aopt, respectively Tspan. To accurately predict Topt, all acclimation processes were required (R2 = 0.74). Acclimation of the enzymatic response was a key driver but caused an underestimation of Topt in tropical climates. This underestimation was resolved through acclimation of the base rate of respiration and the stomatal sensitivity to vapour pressure changes. Both decreased with increasing Tgrowth, resulting in an upwards shift of Topt and accurate predictions in tropical climates. Additionally, acclimation of the photosynthetic capacities was necessary to avoid an otherwise falsely predicted increase of Aopt with Tgrowth. The model predicted a linear decrease of Tspan with Tgrowth, indicating an incomplete formulation for the acclimation of the enzymatic response. Our results demonstrate that the thermal acclimation of Topt and Aopt is predictable from the environment across species and that global photosynthesis models should adopt acclimation of the photosynthetic capacities, stomatal sensitivity and enzymatic response to predict the response of photosynthesis to global warming accurately.

Nitrous oxide (N2O) is a greenhouse gas that causes both global warming and ozone depletion in the stratosphere. Global N2O fluxes are likely to change rapidly with global environmental changes (increasing CO2, warming and changes in soil moisture) and are also influenced by nitrogen (N) fertilizer use in agriculture and managed grasslands, atmospheric N deposition, and soil organic carbon (SOC) levels. Environmental dependencies of N2O emissions have been investigated through both field measurements and experimental studies, but disparate results have been obtained. A confrontation of model-simulated N2O emissions against a diversity of observations, from flux measurements and experiments, has not yet been performed.We compiled data on annual total N2O emissions from published field (n = 214 sites, 835 observations) and experimental (n = 55 sites, 142 observations) studies, and used these data to develop statistical models to model the responses of N2O emission to local N fertilization, climate variables, green vegetation cover and SOC. Using field measurements, we found that N2O emissions from forest soils increase with growth temperature (Tg) and observed soil moisture, likely reflecting higher nitrification and denitrification rates in warmer and wetter soils. In grasslands, we found that N2O emissions increase with N fertilization, and with the seasonal minimum value of fraction of absorbed photosynthetically active radiation (min fAPAR). Lower min fAPAR is associated with grasslands in dry or cold climates that constrain both productivity and the rate of organic matter turnover. In croplands, N2O increases with N fertilization, temperature, and humidity, consistent with the above, but also increases with SOC, and with total incident photosynthetic photon flux density over the growing season (total PPFD) and max fAPAR – two important controls of total annual primary production. The partial response of cropland N2O emission to min fAPAR however is negative, likely reflecting enhanced N2O emission from soil during periods when the crop cover is absent.In the experimental database, N2O response to elevated CO2 (eCO2) varies strongly across experiments, with log response ratios ranging from ­–2.2 to 1.8. Overall, N2O tends to decrease with eCO2, which is likely due to mineralized N being taken up by more rapidly by faster-growing plants. Response ratios also increase with N fertilization and total PPFD. In warming experiments, response ratios increase with SOC and temperature, consistent with what we found in field measurements.We have constructed data-driven models that shows significant responses of N2O emission to climate, N fertilization and SOC. We plan to use these as benchmarks for the evaluation of emergent N2O responses to global environmental changes in Earth System models. We will use LPX-Bern, and other models participating in the Global N2O Model Intercomparison Project (NMIP), to compare simulated environmental dependencies of N2O emission with our data-driven models. The data-driven models will also allow us to independently quantify N2O emission factors in croplands, and to compare global N2O-climate and N2O-CO2 feedbacks with previously published values.

Energy partitioning between surface latent (LE) and sensible (H) heat fluxes is a key factor in the development of the boundary layer and the regulation of the hydrological cycle. Climate factors and surface cover are commonly considered the major controlling effects on energy partitioning. However, the influence of other drivers such as water table depth and groundwater convergence has rarely been considered.Here, we use an extensive dataset of eddy covariance and global remote-sensing data to show that not only climate, but also water table depth and plant functional type (PFT) play an important role in energy partitioning across different biomes. Our findings illuminate the understanding of plant water stress in terrestrial ecosystems.

The rooting-zone water-storage capacity—the amount of water accessible to plants—controls the sensitivity of land–atmosphere exchange of water and carbon during dry periods. How the rooting-zone water-storage capacity varies spatially is largely unknown and not directly observable. Here we estimate rooting-zone water-storage capacity globally from the relationship between remotely sensed vegetation activity, measured by combining evapotranspiration, sun-induced fuorescence and radiation estimates, and the cumulative water defcit calculated from daily time series of precipitation and evapotranspiration. Our fndings indicate plant-available water stores that exceed the storage capacity of 2-m-deep soils across 37% of Earth’s vegetated surface. We fnd that biome-level variations of rooting-zone water-storage capacities correlate with observed rooting-zone depth distributions and refect the infuence of hydroclimate, as measured by the magnitude of annual cumulative water-defcit extremes. Smaller-scale variations are linked to topography and land use. Our fndings document large spatial variations in the efective root-zone water-storage capacity and illustrate a tight link among the climatology of water defcits, rooting depth of vegetation and its sensitivity to water stress., Nature Geoscience, 16 (3), ISSN:1752-0908, ISSN:1752-0894

Leaf phenology is key for regulating total growing season mass and energy fluxes. Long-term temporal trends towards earlier leaf unfolding are observed across Northern Hemisphere forests. Phenological dates also vary between years, whereby end-of-season (EOS) dates correlate positively with start-of-season (SOS) dates and negatively with growing season total net CO2 assimilation (Anet). These associations have been interpreted as the effect of a constrained leaf longevity or of premature carbon (C) sink saturation - with far-reaching consequences for long-term phenology projections under climate change and rising CO2. Here, we use multi-decadal ground and remote-sensing observations to show that the relationships between Anet and EOS are opposite at the interannual and the decadal time scales. A decadal trend towards later EOS persists in parallel with a trend towards increasing Anet - in spite of the negative Anet-EOS relationship at the interannual scale. This indicates that acclimation of phenology has enabled plants to transcend a constrained leaf longevity or premature C sink saturation over the course of several decades, leading to a more effective use of available light and a sustained extension of the vegetation CO2 uptake season over time.

While enhanced tree growth over the last decades has been reported in forests across the globe, it remains unclear whether it drives persistent biomass increases of the stands, particularly in mature forests. Enhanced tree growth and stand-level biomass are often linked with a simultaneous increase in density-driven mortality and a reduction in tree longevity. Identifying empirical evidence regarding the balance between these processes is challenging due to the confounding effects of stand history, management, and environmental changes. Here, we investigate the link between growth and biomass via the shift in the negative relationship between average tree size and stand density (tree number). We find increasing stand density for a given tree size in unmanaged closed-canopy forests in Switzerland over the past six decades and a positive relationship between growth and stand density - qualitatively consistent with simulations by a mechanistic, cohort-resolving ecosystem model (LM3-PPA). Model simulations show that, in the absence of other disturbances, enhanced growth persistently increases biomass stocks despite simultaneous decreases in carbon residence time and tree longevity, independent of assumptions about the drivers of tree mortality. However, the magnitude of simulated changes critically depends on the shape of the mortality parameterizations. Our analyses reconcile reports of growth-induced reductions of tree longevity with model predictions of persistent biomass increases, and with our finding of a trend towards denser forests in response to growth - also in mature stands.

Plant biomass production (BP), nitrogen uptake (Nup) and their ratio, nitrogen use efficiency (NUE), must be quantified to understand how nitrogen (N) cycling constrains terrestrial carbon (C) uptake. But the controls of key plant processes determiningNupand NUE, including BP, C and N allocation, tissue C:N ratios and N resorption efficiency (NRE), remain poorly known. We compiled measurements from 804 forest and grassland sites and derived regression models for each of these processes with growth temperature, vapour pressure deficit, stand age, soil C:N ratio, fAPAR (remotely sensed fraction of photosynthetically active radiation absorbed by green vegetation) and growing-season average daily incident photosynthetic photon flux density (gPPFD) (effectively the seasonal concentration of light availability, which increases polewards) as predictors. An empirical model for leaf N was based on optimal photosynthetic capacity (a function of gPPFD and climate) and observed leaf mass-per-area. The models were used to produce global maps ofNupand NUE. Global BP was estimated as 72 Pg C/yr;Nupas 950 Tg N/yr; and NUE as 76 gC/gN. Forest BP was found to increase with growth temperature and fAPAR and to decrease with stand age, soil C:N ratio and gPPFD. Forest NUE is controlled primarily by climate through its effect on C allocation – especially to leaves, being richer in N than other tissues. NUE is greater in colder climates, where N is less readily available, because belowground allocation is increased. NUE is also greater in drier climates because leaf allocation is reduced. NRE is enhanced (further promoting NUE) in both cold and dry climates. These findings can provide observationally based benchmarks for model representations of C–N cycle coupling. State-of-the-art vegetation models in the TRENDY ensemble showed variable performance against these benchmarks, and models including coupled C–N cycling produced relatively poor simulations ofNupand NUE.

Droughts affect terrestrial ecosystems directly and concurrently and can additionally induce lagged effects in subsequent seasons and years. Such legacy effects of drought on vegetation growth and state have been widely studied in tree ring records and satellite-based vegetation greenness, while legacies on ecosystem carbon fluxes are still poorly quantified and understood. Here, we focus on two ecosystem monitoring sites in central Germany with a similar climate but characterised by different species and age structures. Using eddy covariance measurements, we detect legacies on gross primary productivity (GPP) by calculating the difference between random forest model estimates of potential GPP and observed GPP. Our results showed that, at both sites, droughts caused significant legacy effects on GPP at seasonal and annual timescales, which were partly explained by reduced leaf development. The GPP reduction due to drought legacy effects is of comparable magnitude to the concurrent drought effects but differed between two neighbouring forests with divergent species and age structures. The methodology proposed here allows the quantification of the temporal dynamics of legacy effects at the sub-seasonal scale and the separation of legacy effects from model uncertainties. The application of the methodology at a larger range of sites will help us to quantify whether the identified lag effects are general and on which factors they may depend., Biogeosciences, 19 (17), ISSN:1726-4170

The advent of abundant Earth observation data enables the development of novel predictive methods for forecasting climate impacts on the state and health of terrestrial ecosystems. Here, we target the spatial and temporal variations of land surface reflectance and vegetation greenness, measuring the density of green vegetation and active foliage area, conditioned on current and past climate and the local topography. We train two alternative recurrent deep learning models that rely on convolutional layers for forecasting the spatially resolved deviation of surface reflectance across a heterogeneous landscape from a specified initial state (Baseline Framework). We demonstrate efficiency of the Baseline Framework with respect to training convergence speed. Using data from diverse ecosystems and land cover types across Europe and following a standardized model evaluation framework (EarthNet2021 Challenge), results indicate increased performance in predicting surface greenness during drought events of the models presented here, compared to currently published benchmarks. Our results demonstrate how deep learning methods enable early-warning of vegetation responses to the impacts of climatic extreme events, such as the drought-related loss of green foliage.

A poor understanding of the fraction of global plant biomass occurring belowground as roots limits our understanding of present and future ecosystem function and carbon pools. Here we create a database of root-mass fractions (RMFs), an index of plant below- versus aboveground biomass distributions, and generate quantitative, spatially explicit global maps of RMFs in trees, shrubs and grasses. Our analyses reveal large gradients in RMFs both across and within vegetation types that can be attributed to resource availability. High RMFs occur in cold and dry ecosystems, while low RMFs dominate in warm and wet regions. Across all vegetation types, the directional effect of temperature on RMFs depends on water availability, suggesting feedbacks between heat, water and nutrient supply. By integrating our RMF maps with existing aboveground plant biomass information, we estimate that in forests, shrublands and grasslands, respectively, 22%, 47% and 67% of plant biomass exists belowground, with a total global belowground fraction of 24% (20–28%), that is, 113 (90–135) Gt carbon. By documenting the environmental correlates of root biomass allocation, our results can inform model projections of global vegetation dynamics under current and future climate scenarios. ISSN:2397-334X

Terrestrial ecosystems remove about 30 per cent of the carbon dioxide (CO2) emitted by human activities each year1, yet the persistence of this carbon sink depends partly on how plant biomass and soil organic carbon (SOC) stocks respond to future increases in atmospheric CO2 (refs. 2,3). Although plant biomass often increases in elevated CO2 (eCO2) experiments4–6, SOC has been observed to increase, remain unchanged or even decline7. The mechanisms that drive this variation across experiments remain poorly understood, creating uncertainty in climate projections8,9. Here we synthesized data from 108 eCO2 experiments and found that the effect of eCO2 on SOC stocks is best explained by a negative relationship with plant biomass: when plant biomass is strongly stimulated by eCO2, SOC storage declines; conversely, when biomass is weakly stimulated, SOC storage increases. This trade-off appears to be related to plant nutrient acquisition, in which plants increase their biomass by mining the soil for nutrients, which decreases SOC storage. We found that, overall, SOC stocks increase with eCO2 in grasslands (8 ± 2 per cent) but not in forests (0 ± 2 per cent), even though plant biomass in grasslands increase less (9 ± 3 per cent) than in forests (23 ± 2 per cent). Ecosystem models do not reproduce this trade-off, which implies that projections of SOC may need to be revised. A synthesis of elevated carbon dioxide experiments reveals that when plant biomass is strongly stimulated by elevated carbon dioxide levels, soil carbon storage declines, and where biomass is weakly stimulated, soil carbon accumulates.

Understanding how leaf autumn phenology varies at different spatio-temporal scales is key to accurately predicting phenological changes under future climate. Recent projections and observations of autumn phenology in deciduous temperate and boreal forests appear conflicting. At the interannual scale, autumn senescence correlates positively with spring leaf-out and negatively with growing season total photosynthesis. These links have been interpreted as the effect of premature carbon sink saturation with potentially far-reaching consequences for carbon cycle projections in a high-CO2 world. In this study, we use multi-decadal ground and remote-sensing observations to show that these relationships are opposite at the interannual and the decadal time scales. We found a decadal trend towards later autumn senescence in parallel with a trend towards increasing photosynthesis, despite their negative relationship at the interannual scale. Our findings reveal that the leaf longevity constraint has not remained stationary over longer time scales. These shifting relationships suggest a gradual acclimation of phenology, leading to a relief of effects that dominate year-to-year variations. This apparent acclimation implies that in the long run, trees may benefit from increased photosynthesis under rising CO2 and evade a direct effect by which increased photosynthesis induces an earlier leaf senescence. This apparent plasticity in phenology appears to have driven plants towards optimal functioning in a changing climate.

As an important aspect of the leaf economic spectrum, leaf nitrogen (N) and phosphorus (P) concentrations are key traits that shape plant growth and function and reflect evolutionary history. Their patterns across the globe have been interpreted as being reflective of controls by climate, soil fertility, and atmospheric N and P deposition. Yet, recent research has emphasized an overriding importance of species identity and phylogeny in determining stoichiometry. This limits the applicability of methods for predicting global patterns in foliar stoichiometry from environmental covariates and implies that stoichiometry responses to global environmental change are limited by the rate of species replacement.Here, we investigated these contrasting views. We established a comprehensive global data set of 36,413 paired observations of leaf N and P concentrations with their specific phylogenetic taxonomy and 46 environmental covariates. For leaf N, we identified the most important predictors being N deposition, irradiance, temperature of the coldest month, atmospheric CO2, elevation, mean annual vapor pressure deficit, and aluminum saturation of the soil solution. For leaf P, the predictors are N deposition, temperature of the coldest month, aridity index, atmospheric CO2, soil phosphorus concentration, annual mean ratio of actual over potential evapotranspiration, multi-day average stomatal conductance, available water storage capacity and precipitation of the driest month. Together, the predictors explain 46% and 33% of leaf N and P variations respectively in the data aggregated by sites, using a Random Forest model. Using linear models on the full non-aggregated data and not accounting for interactions between predictors, species identity explains the largest portion of observed variations in individual-level foliar N (50.4%) and P (33.9%), while environmental covariates explain only 3.8%, and 2.8%, respectively.A trait gradient analysis reconciles these contrasting results and suggests more within-species variations in response to the environment as would be expected from the results of the linear models. We found (i) that the genetic variation or environmentally-induced plasticity in foliar N and P within species is substantial for most species, (ii) that species’ distributions cover a wide range of environments as expressed by site-level mean leaf N and P, and (iii) that variations of leaf N and P stoichiometry across sites are not merely emerging as a result of distinct species present at different sites. This challenges the notion of a fixed biogeochemical niche occupied by individual species and suggests that plants mediate their stoichiometry in response to their environment, and potentially to global environmental change. This insight also provides the basis for a robust spatial upscaling of foliar N and P, using global fields of environmental covariates as predictors. Our global predictions also suggest that the highest foliar N concentration occurs in cold and dry climates and at high-elevation.

Using steady-state photosynthesis-intercellular CO

The rooting zone water storage capacity (S0) extends from the soil surface to the weathered bedrock (the Critical Zone) and determines land-atmosphere exchange during dry periods. Despite its importance to land-surface modeling, variations of S0 across space are largely unknown as they cannot be observed directly. We developed a method to diagnose global variations of S0 from the relationship between vegetation activity (measured by sun-induced fluorescence and by the evaporative fraction) and the cumulative water deficit (CWD). We then show that spatial variations in S0 can be predicted from the assumption that plants are adapted to sustain CWD extremes occurring with a return period that is related to the life form of dominant plants and the large-scale topographical setting. Predicted biome-level S0 distributions, translated to an apparent rooting depth (zr) by accounting for soil texture, are consistent with observations from a comprehensive zr dataset. Large spatial variations in S0 across the globe reflect adaptation of zr to the hydroclimate and topography and implies large heterogeneity in the sensitivity of vegetation activity to drought. The magnitude of S0 inferred for most of the Earth’s vegetated regions and particularly for those with a large seasonality in their hydroclimate indicates an important role for plant access to water stored at depth - beyond the soil layers commonly considered in land-surface models.

Future climate will be characterized by an increase in frequency and duration of drought and warming that exacerbates atmospheric evaporative demand. How trees acclimate to long-term soil moisture changes and whether these long-term changes alter trees' sensitivity to short-term (day to months) variations of vapor pressure deficit (VPD) and soil moisture is largely unknown. Leaf gas exchange measurements were performed within a long-term (17 years) irrigation experiment in a drought-prone Scots pine-dominated forest in one of Switzerland's driest areas on trees in naturally dry (control), irrigated, and 'irrigation-stop' (after 11 years of irrigation) conditions. Seventeen years of irrigation increased photosynthesis (A) and stomatal conductance (g(s)) and reduced g(s) sensitivity to increasing VPD and soil drying. Following irrigation-stop, gas exchange decreased only after 3 years. After 5 years, maximum carboxylation (V-cmax) and electron transport (J(max)) rates in irrigation-stop recovered to similar levels as to before the irrigation-stop. These results suggest that long-term release from soil drought reduces the sensitivity to VPD and that atmospheric constraints may play an increasingly important role in combination with soil drought. Moreover, our study indicates that structural adjustments lead to an attenuation of initially strong leaf-level acclimation to strong multiple-year drought. Acclimation to irrigation increased gas exchange in Pinus sylvestris, but reduced the sensitivity to short-term changes. In addition, structural adjustments led to an attenuation of initially strong leaf-level acclimation.

International audience; Global vegetation and land-surface models embody interdisciplinary scientific understanding of the behaviour of plants and ecosystems, and are indispensable to project the impacts of environmental change on vegetation and the interactions between vegetation and climate. However, systematic errors and persistently large differences among carbon and water cycle projections by different models highlight the limitations of current process formulations. In this review, focusing on core plant functions in the terrestrial carbon and water cycles, we show how unifying hypotheses derived from eco-evolutionary optimality (EEO) principles can provide novel, parameter-sparse representations of plant and vegetation processes. We present case studiesthat demonstrate how EEO generate parsimonious representations of core, leaf-level processes that are individually testable and supported by evidence. EEO approaches to photosynthesis and primary production, dark respiration, and stomatal behaviour are ripe for implementation in global models. EEO approaches to other important traits, including the leaf economics spectrum and applications of EEO at the community level are active research areas. Independently tested modules emerging from EEO studies could profitably be integrated into modelling frameworks that account for the multiple time scales on which plants and plant communities adjust toenvironmental change.

Accurate simulations of gross primary production (GPP) are vital for Earth System Models that must inform public policy decisions. The instantaneous controls of leaf-level photosynthesis, which can be measured in manipulative experiments, are well established. At the canopy scale, however, there is no consensus on how GPP depends on (a) light or (b) other aspects of the physical environment such as temperature and CO2. Models of GPP make a variety of different assumptions when ‘scaling-up’ the standard model of photosynthesis. As a troublesome consequence, they make a variety of different predictions about how GPP responds to projected environmental change.This problem can be tackled by theoretical modelling and by empirical analysis of GPP as reconstructed from eddy-covariance flux measurements. Theoretical modelling has provided an explanation for why ‘light-use efficiency’ (LUE) models work well at time scales of a week or longer. The same logic provides a justification for the use of LUE as a key metric in an empirical analysis. By focusing on LUE, we can isolate the drivers of GPP independent of its over-riding control by absorbed light. We have used open-access eddy covariance data from over 100 sites, collated over 20 years (the number of sites has grown with time). These sites, located in a wide range of biomes and climate zones, form part of the FLUXNET network. We have combined the flux data with a satellite product (EVI from MODIS) that allows spatial estimates of the fraction of incident light absorbed by green vegetation. Matching soil moisture data were estimated using the SPLASH model, with appropriate meteorological inputs, and soil water-holding capacity derived using SoilGrids. LUE was then calculated as the amount of carbon fixed per unit of absorbed light. We then explored additive models (incorporating multiple explanatory factors) that support non-linear responses. Recognising that our longitudinal data lack independence, we controlled for the hierarchical nature of the dataset through a variance structure that nests measurement year within site location.In arriving at a preferred parsimonious model, we show that daytime air temperature and vapour pressure deficit, and soil moisture content are all salient predictors of LUE. The same explanatory terms are retained in iterations of this analysis run at timescales from weeks to months. As a model-comparison exercise, we used that portion of our dataset which overlaps the North American Carbon Program to apply our empirical model structure to site-based estimates of GPP generated by 19 discrete Terrestrial Biosphere Models (TBMs). The comparative analysis reveals wide variation between the TBMs in the shape, strength and even sign of the environmental effects on modelled GPP.This empirical analysis suggests it is feasible to predict GPP using a single model structure, common across vegetation categories. And the appeal of such universal approaches is highlighted by inconsistent relationships with key environmental drivers within extant terrestrial models.

The future Earth is projected to experience elevated rainfall variability, with more frequent and intense droughts, as well as high-rainfall events. Increasing CO2 concentrations are expected to raise terrestrial gross primary productivity (GPP), whereas water stress is expected to lower GPP. Plant responses to water stress vary strongly with timescale, and plants adapted to different environmental conditions differ in their functional responses. Here, we embed a unified optimality-based theory of stomatal conductance and biochemical acclimation of leaves we have recently developed [Joshi, J. et al. (2020) Towards a unified theory of plant photosynthesis and hydraulics. bioRxiv 2020.12.17.423132] in an eco-evolutionary vegetation-modelling framework, with the goal to investigate emergent functional diversity and associated GPP impacts under different rainfall regimes.The model of photosynthesis used here simultaneously predicts the stomatal responses and biochemical acclimation of leaves to atmospheric and soil-moisture conditions. Using three hydraulic traits and two cost parameters, it successfully predicts the simultaneous declines in CO2 assimilation rate, stomatal conductance, and leaf photosynthetic capacity caused by drying soil. It also correctly predicts the responses of CO2 assimilation rate, stomatal conductance, leaf water potential, and leaf photosynthetic capacity to vapour pressure deficit, temperature, ambient CO2, light intensity, and elevation. Our model therefore captures the synergistic effects of atmospheric and soil drought, as well as of atmospheric CO2 changes, on plant photosynthesis and transpiration.We embed this model of photosynthesis and transpiration in a trait-height-patch structured eco-evolutionary vegetation model. This model accounts for allometric carbon allocation, height-structured competition for light, patch-structured successional dynamics, and coevolution of plant functional traits. It predicts functional species mixtures and emergent ecosystem properties under different environmental conditions. Using this model, we investigate the evolution of plant hydraulic strategies under different regimes of drought and rainfall variability. Our approach provides an eco-evolutionarily consistent framework to scale up the responses of plant communities from individual plants to ecosystems to provide ecosystem-level predictions of functional diversity, primary production, and plant water use, and could thus be used for reliable projections of the global carbon and water cycles under future climate scenarios.

Leaf nitrogen (N) and phosphorus (P) are essential nutrients constraining plant growth and ecosystem productivity. Stoichiometric relationships between N and P concentration have been explored at both regional and global levels in recent decades, especially for leaves, resulting in the increasing number of available data. For example, threshold values of leaf N:P ratios and a 2/3 power law of leaf N-P scaling relationship have been proposed, and are widely used to detect nutrient limitation in terrestrial ecosystems. However, our recent work has revealed that significant variation in the scaling exponent of leaf N against P concentration appeared not only across plant functional and taxonomic groups, but also along climatic and soil fertility gradients. This suggests the N vs. P scaling exponent to be a relatively plastic “trait” that is shaped by local climate environments, soil nutrient conditions, and phylogeny in conjunction, with important feedbacks to ecosystem function and crucial insights into traditional mechanistic ecosystem models, which have mostly modelled the controls on leaf N and P stoichiometric characteristics through the relationships of foliar nutrient concentrations with soil nutrient availability across plant functional types (PFT), ignoring the important differences within PFTs and species along climatic gradients.To further identify relationships of leaf N against P concentrations with multiple controls across the globe, we compiled a comprehensive leaf N and P dataset. Totally, this dataset contains 12,453 individual records spanning 2,158 sites worldwide, including 203 families, 1,291 genera and 3,361 species. The climatic, soil variables and phylogenetic information provided for each individual record included: (1) geographical location information; (2) topography indices; (3) sampling period during the field campaigns and experiments; (4) taxonomic information; (5) life form (i.e. angiosperm vs. gymnosperm, monocotyledonous vs. dicotyledonous, woody (including coniferous, deciduous broad-leaved and evergreen broad-leaved woody) vs. herbaceous plants; (6) climate indices, including mean annual temperature, mean annual precipitation, aridity index, maximum cumulative water deficit; (7) pairwise soil N and P concentrations for part of the whole records; (8) corresponding soil class information from the Harmonized World Soil Database; (9) corresponding atmospheric CO2 concentration and bulk atmospheric N deposition during the sampling period; and (10) specific leaf area (TRY Plant Trait Database).Moreover, we built empirical models of varying complexity, which provides critical insight into the relative importance of plant phylogeny, soil properties and climatic variables in shaping global-scale leaf N and P stoichiometry of terrestrial plants. Then, we assessed the power of the modelled optimal maximum Rubisco carboxylation rate standardized to 25 ℃ (Vcmax25), based on leaf N concentrations, as an independent predictor to investigate the power of the state-of-the-art optimality model driven by climate variables alone for photosynthetic traits. We further explored the use of the N~P scaling relationship for each specific species, life form and PFT in parameterizing current ecosystem models. Collectively, our work may provide an important benchmark for connecting plant elemental stoichiometry with next-generation earth system models, which can enhance our predictions of ecosystem functioning and vegetation dynamics, especially in the context of rapidly environmental changes with increasing CO2 and N deposition.

Plants not only acquire carbon to sustain biomass production, autotrophic respiration, and the production of non-structural compounds, but also require nitrogen to support carboxylation and growth. However, available observations have not fully been integrated and used for modelling growth, carbon allocation to different compartments, and how different compartments’ nitrogen-to-carbon ratio vary across large climatic and soil gradients. This leaves substantial uncertainty in estimates of the global distribution of growth and nitrogen uptake by plants. Here, we used the P-model, a first principles-derived and remote sensing-driven model for terrestrial gross primary production (GPP) to simulate the global distribution of GPP. Using comprehensive datasets with locally measured covariates for climatic and edaphic conditions and vegetation structure, we modelled the fractional allocation of GPP to biomass production (BP), aboveground net primary production (ANPP), and leaf NPP based on linear mixed-effects regression models. We defined BP as the sum of NPP in leaves, wood and roots. It thus does not include additional components such as exudates and labile carbon to mycorrhizae. Leaf nitrogen-to-carbon was modelled based on the maximum rate of carboxylation at 25 degrees Celsius (Vcmax25) and leaf mass per area (LMA). We then used global gridded data for the covariates that entered as predictors in site-level empirical models to simulate global C and N allocated to each component. We finally validated our global simulation results with an extended set of globally distributed GPP, BP and nitrogen-to-carbon ratio observations. GPP was well predicted (R2 = 0.61). In forests, ratios of BP/GPP and ANPP/GPP decreased with soil C/N and stand-age but increased with humidity and with the fraction of absorbed photosynthetically active radiation (fAPAR). The ratio of leaf NPP to ANPP, increased with light availability and growth temperature, but decreased with vapor pressure deficit. Leaf nitrogen-to-carbon ratio was positively related to the ratio of Vcmax25 to LMA. Leaf nitrogen resorption efficiency (NRE) was increased in drier and colder environments. Through our data validation at the end, we have shown a prediction for NPP (R2 = 0.26), ANPP (R2 = 0.28), leaf NPP (R2 = 0.39), NRE (R2 = 0.30), leaf N/C (R2 = 0.26) and leaf N flux (R2 = 0.35). Simulated global total GPP is 125 Pg C yr-1. Based on these statistical models, global mean carbon-use-efficiency (BP/GPP) was estimated to be 40%. The ratio of ANPP/BP was 72%, and ANPP was further split with 46% to leaf NPP and 54% to wood NPP. Simulated global total nitrogen acquisition (total of uptake from the soil and symbiotic N fixation) was 860 Tg N yr-1. Growth in the leaf, wood and root compartment accounted for 39%, 23% and 38% of global N acquisition, respectively. We suggest that plant adaptations result in higher ANPP, leaf NPP and finally leaf N flux under warmer, wetter, more abundant light and N-rich soil conditions, which aims to support higher rate of photosynthesis with greater nitrogen investment in the leaf.

The rooting zone water storage capacity (S) defines the total amount of water available to plants for transpiration during rain-free periods. Thereby, S determines the sensitivity of carbon and water exchanges between the land surface and the atmosphere, controls the sensitivity of ecosystem functioning to progressive drought conditions, and mediates feedbacks between soil moisture and near-surface air temperatures. While being a central quantity for water-carbon-climate coupling, S is inherently difficult to observe. Notwithstanding scarcity of observations, terrestrial biosphere and Earth system models rely on the specification of S either directly or indirectly through assuming plant rooting depth.Here, we model S based on the assumption that plants size their rooting depth to maintain function under the expected maximum cumulative water deficit (CWD), occurring with a return period of 40 years (CWDX40), following Gao et al. (2014). CWDX40 is “translated” into a rooting depth by accounting for the soil texture. CWD is defined as the cumulative evapotranspiration (ET) minus precipitation, where ET is estimated based on thermal infrared remote sensing (ALEXI-ET), and precipitation is from WATCH-WFDEI, modified by accounting for snow accumulation and melt. In contrast to other satellite remote sensing-based ET products, ALEXI-ET makes no a priori assumption about S and, as our evaluation shows, exhibits no systematic bias with increasing CWD. It thus provides a robust observation of surface water loss and enables estimation of S with global coverage at 0.05° (~5 km) resolution.Modelled S and its variations across biomes is largely consistent with observed rooting depth, provided as ecosystem-level maximum estimates by Schenk et al. (2002), and a recently compiled comprehensive plant-level dataset. In spite of the general agreement of modelled and observed rooting depth across large climatic gradients, comparisons between local observations and global model predictions are mired by a scale mismatch that is particularly relevant for plant rooting depth, for which the small-scale topographical setting and hydrological conditions, in particular the water table depth, pose strong controls.To resolve this limitation, we investigate the sensitivity of photosynthesis (estimated by sun-induced fluorescence, SIF), and of the evaporative fraction (EF, defined as ET over net radiation) to CWD. By employing first principles for the constraint of rooting zone water availability on ET and photosynthesis, it can be derived how their sensitivity to the increasing CWD relates to S. We make use of this relationship to provide an alternative and independent estimate of S (SdSIF and SdEF), informed by Earth observation data, to which S, modelled using CWDX40, can be compared. Our comparison reveals a strong correlation (R2=0.54) and tight consistency in magnitude between the two approaches for estimating S. Our analysis suggests adaptation of plant structure to prevailing climatic conditions and drought regimes across the globe and at catchment scale and demonstrates its implications for land-atmosphere exchange. Our global high-resolution mapping of S reveals contrasts between plant growth forms (grasslands vs. forests) and a discrepant importance across the landscape of plants’ access to water stored at depth, and enables an observation-informed specification of S in global models.

Forest demographic processes - growth, recruitment and mortality - are being altered by global change. The changing balance between growth and mortality strongly influences forest dynamics and the carbon balance. Elevated atmospheric carbon dioxide (eCO2) has been reported to enhance photosynthesis and tree growth rates by increasing both light-use efficiency (LUE) and water-use efficiency (WUE). Tree growth enhancement could be translated into an increase in biomass stocks or could be associated with a reduction in the longevity of trees, thus reducing the ability of forest ecosystems to act as carbon sinks over long timescales. These links between growth and mortality, and the implications for forest stand density and self-thinning relationships are still debated. Scarce empirical evidence exists for how changing drivers affect tree mortality due to existing data and modelling limitations. Understanding the causes of observed mortality trends and the mechanisms underlying these processes is critical for accurate projections of global terrestrial carbon storage and its feedbacks to anthropogenic climate change.Here, we combine a mechanistic model with empirical forest data to better understand the causes of changes in tree mortality and the implications for past and future trends in forest tree density. Specifically, we test the Grow-Fast-Die-Young hypothesis to investigate if a leaf-level CO2 fertilization effect may lead to an increase in the biomass stock in forest stands. We use a novel vegetation demography model (LM3-PPA) which includes vegetation dynamics with biogeochemical processes allowing for explicit representation of individuals and a mechanistic treatment of tree mortality. The key links between leaf-level assimilation and stand dynamics depend on the carbon turnover time. In this sense, we investigate alternative mortality assumptions about the functional dependence of mortality on tree size, tree carbon balance or growth rate. These formulations represent typical approaches to simulate mortality in mechanistic forest models. Model simulations show that increasing photosynthetic LUE leads to higher biomass stocks, with contrasting behavior among mortality assumptions. Empirical data from Swiss forest inventories support the results from the model simulations showing a shift upwards in the self-thinning relationships, with denser stands and bigger trees. This data-supported mortality-modelling helps to identify links between forest responses and environmental changes at the leaf, tree and stand levels and yields new insight into the causes of currently observed terrestrial carbon sinks and future responses.

Terrestrial evapotranspiration (ET) is a key factor in the global energy and water cycles. It is constrained by the transport of moisture from the soil and from vegetation to the atmosphere. The water storage capacity in the root zone (Sr) is an important parameter in land-atmosphere water exchanges, defining how long vegetation is able to transpire during drought. However, Sr is hard to measure directly, being associated with the depth of plant roots actively involved in water uptake and the potential of tap roots accessing deep water and enabling sustained transpiration during drought. In this study, we present a method to estimate Sr from flux measurements, based on a deep neural network approach trained on eddy covariance (EC) data, multiple soil moisture datasets and a remotely sensed index of vegetation greenness. We derive a soil moisture stress function (fET) that isolates the control of soil moisture on ET. We then use EC data to estimate Sr by investigating how it relates to the climatology of the maximum cumulative water deficit (CWD, defined as the cumulative difference between actual ET and precipitation) experienced by the vegetation across different sites. We hypothesize that plants exposed to high CWD develop higher Sr (acclimation to water stress) and that maximum CWD is thus a good estimator of Sr. To identify root zone water storage from flux measurements, we regress the output of fET against CWD and estimate the maximum CWD for stress by calculating the intersection of the regression line with the x-axis. The apparent sensitivity of fET to CWD and its correlation with the maximum CWD across sites are indicative of adaptation to the prevailing climate and drought regime. We find that for many sites, particularly in seasonally dry climates, fET does not exhibit a continuous decline with increasing CWD, but follows a step-change and levelling off. That is, at the high end of the CWD spectrum, fET no longer appears sensitive to further increases in CWD. This suggests that plants may have access to deep water reservoirs during the unfolding of a drought event, indicating that plant access to water becomes decoupled from water use. This study highlights the need to investigate the representation of plant access to deep water reservoirs during drought in terrestrial ecosystem models. These findings could improve our understanding of land-climate interactions, particularly under water-limited conditions.

SummaryThe future climate will be characterized by an increase in frequency and duration of drought and warming that exacerbates atmospheric evaporative demand. How trees acclimate to long-term soil moisture changes and whether these long-term changes alter trees’ sensitivity to short-term (day to months) variations of vapor pressure deficit (VPD) and soil moisture is largely unknown.Leaf gas exchange measurements were performed within a long-term (17 years) irrigation experiment in a Scots pine-dominated forest in one of Switzerland’s driest areas on trees in naturally dry (control), irrigated, and‘irrigation-stop’ (after 11 years of irrigation) conditions.Seventeen years of irrigation increased photosynthesis (A) and stomatal conductance (gs) and reduced the gs sensitivity to increasing VPD but not to soil drying. Following irrigation-stop, gas exchange did not decrease immediately, but after three years, had decreased significantly in irrigation-stop trees. Vcmax and Jmax recovered after five years.These results suggest that long-term release of soil drought reduces the sensitivity to atmospheric evaporative demand and that atmospheric constraints may play an increasingly important role in combination with soil drought. In addition, they suggest that structural adjustments lead to an attenuation of initially strong leaf-level acclimation to strong multiple-year drought.

Significance Forests remove about 30% of anthropogenic CO2 emissions through photosynthesis and return almost 40% of incident precipitation back to the atmosphere via transpiration. The trade-off between photosynthesis and transpiration through stomata, the water-use efficiency (WUE), is an important driver of plant evolution and ecosystem functioning, and has profound effects on climate. Using stable carbon and oxygen isotope ratios in tree rings, we found that WUE has increased by a magnitude consistent with estimates from atmospheric measurements and model predictions. Enhanced photosynthesis was widespread, while reductions in stomatal conductance were modest and restricted to moisture-limited forests. This result points to smaller reductions in transpiration in response to increasing atmospheric CO2, with important implications for forest–climate interactions, which remain to be explored., Multiple lines of evidence suggest that plant water-use efficiency (WUE)—the ratio of carbon assimilation to water loss—has increased in recent decades. Although rising atmospheric CO2 has been proposed as the principal cause, the underlying physiological mechanisms are still being debated, and implications for the global water cycle remain uncertain. Here, we addressed this gap using 30-y tree ring records of carbon and oxygen isotope measurements and basal area increment from 12 species in 8 North American mature temperate forests. Our goal was to separate the contributions of enhanced photosynthesis and reduced stomatal conductance to WUE trends and to assess consistency between multiple commonly used methods for estimating WUE. Our results show that tree ring-derived estimates of increases in WUE are consistent with estimates from atmospheric measurements and predictions based on an optimal balancing of carbon gains and water costs, but are lower than those based on ecosystem-scale flux observations. Although both physiological mechanisms contributed to rising WUE, enhanced photosynthesis was widespread, while reductions in stomatal conductance were modest and restricted to species that experienced moisture limitations. This finding challenges the hypothesis that rising WUE in forests is primarily the result of widespread, CO2-induced reductions in stomatal conductance.

Northern peatlands store large amounts of carbon (C) and have played an important role in the global carbon cycle since the Last Glacial Maximum. Most northern peatlands have established since the end of the deglaciation and accumulated C over the Holocene, leading to a total present-day stock of 500 ± 100 GtC. This is a consolidated estimate, emerging from a diversity of methods using observational data. Recently, Nichols and Peteet (2019 Nature Geoscience 12: 917-921) presented an estimate of the northern peat C stock of 1055 GtC—exceeding previous estimates by a factor of two. Here, we will review various approaches and estimates of northern peatlands C storage in the literature and consider peat C storage in the context of the Holocene global C budget. We argue that the estimate by Nichols and Peteet is an overestimate, caused by systematic bias introduced by their inclusion of data that are representative for the major peatland regions and of records that lack direct measurements of C density. In particular, some “peatland” sites and data that were included in their synthesis were likely from lacustrine sediments prior to the onset of peat deposits. Furthermore, we argue that their estimate cannot be reconciled within the constraints offered by ice-core and marine records of stable C isotopes and estimated contributions from other processes that affected the terrestrial C storage during the Holocene.

Global environmental changes are rapidly altering the functioning and structure of terrestrial ecosystems.Particularly, rising CO2 atmospheric concentrations have been reported to increase photosynthesis by increasing carbon assimilation and water-use efficiency. This leaf-level CO2 fertilization effect may lead to an increase in the biomass stock in forest stands. However, previous studies argued that an enhanced tree growth rate is associated with a reduction in the longevity of trees, thus reducing the ability of forest biomass to act as carbon sinks over long timescales. In addition, faster growth may lead to an acceleration of self-thinning whereby tree density in the stand is reduced due to progressive mutual shading as tree crowns expand and a resulting increase in shaded individuals’ mortality. Nevertheless, previous results relied on empirical relationships between tree growth rates and longevity, without considering any positive effects of elevated CO​2 on individual tree’s carbon balance. Individual-based forest datasets such as tree ring width data and forests inventories have been widely used to monitor long-term changes in forest demography. Yet, the mechanistic underlying these processes remains poorly understood and challenges persist in upscaling from individual measurements to higher level of organization.Here, we use a vegetation demography model (LM3-PPA) which simulates vegetation dynamics and biogeochemical processes by explicitly scaling from leaf up to ecosystem level by resolving leaf-level physiology, growth, and height-structured competition for light, using the perfect plasticity approximation (PPA). Using this simulation model, we investigate the links between individual trees’ carbon balance under rising CO2 levels, their longevity under alternative mortality parametrizations, and the implications for forest dynamics and self-thinning rates. Inventory data from long-term forest reserves is used to assess empirical support for these simulated links. Specifically, we test the hypothesis of faster growth-earlier death in order to assess forests’ capacity to store carbon under environmental changes. This provides key mechanistic insights to reveal whether increased CO2 fertilization on leaf-level photosynthesis positively affects tree’s C balance and thereby reduces the mortality under competition for light in the canopy.

The question of whether prehistoric human impacts on land cover (i.e. anthropogenic land cover change due to land use, LULC) were sufficiently large to have a major impact on regional cli-mates is still a matter of debate. Climate model simulations have shown that LULC datasets can have large regional impacts on climate in recent and prehistoric time (1). But there are major differences between the available LULC scenarios/datasets such as HYDE (History Database of the Global En-vironment) and Kaplan’s KK10 (2), and diagnoses of inferred carbon-cycle impacts show that none of the scenarios are realistic (3). The only way to provide a useful assessment of the potential for LULC changes to affect climate in the past, is to provide more realistic LULC data based on palaeovegetation and archaeological evidence to improve the LULC datasets used in climate modelling(4). We use the REVEALS model to reconstruct LC from pollen data at a regional scale, and archaeological data to map LU types and distribution, and estimate per capita LU. The archaeology-based LU maps and per-capita LU estimates are used to improve LULC datasets. Pollen-based REVEALS LC estimates are then used to evaluate/validate the new, improved LULC datasets. These new datasets will be used to implement past land use in palaeoclimate and carbon cycle model simulations. Such simulations are necessary to assess the impact of LULC changes in the past and understand the effect of ecosys-tem management on future climate. We present results from five years of PAGES LandCover6k activities. (1) Strandberg G, Kjellström E, Poska A, Wagner S, Gaillard M-J et al. (2014) Regional climate model sim-ulations for Europe at 6 and 0.2 k BP: sensitivity to changes in anthropogenic deforestation. Clim. Past 10, 661–680.(2) Gaillard M-J, Sugita S, Mazier F et al (2010) Holocene land-cover reconstructions for studies on land cover-climate feedbacks. Clim. Past 6, 483-499.(3) Stocker B, Yud Z, Massae C, Joos F (2017) Holocene peatland and ice-core data constraints on the tim-ing and magnitude of CO2 emissions from past land use. www.pnas.org/cgi/doi/10.1073/ pnas.1613889114.(4) Harrison S P, Gaillard M-J, Stocker B D, Vander Linden M, Klein Goldewijk K, Boles O, Braconnot P, Dawson A, Fluet-Chouinard E, Kaplan J O, Kastner T, Pausata F S R, Robinson E, Whitehouse N J, Madella M, and Morrison K D (2019) Development and testing of scenarios for implementing Holocene LULC in Earth Sys-tem Model Experiments, Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2019-125, in review, 2019.

Rooting depth governs plants' access to water during dry (rain-free) periods and is thus a key variable determining the sensitivity of transpiration and carbon assimilation to soil moisture drought and land-climate coupling during drought and heat events. Plant rooting depth variations of two orders of magnitude have been recorded across the globe and a substantial portion of this variation appears within species and functional groups. This suggests that plant rooting depth is a relatively plastic “trait” that may adjust to the local environment and soil hydrology and possibly also to temporally changing climatic conditions. Yet, plant rooting depth is commonly treated as a fixed parameter in global vegetation and climate models, specified for each plant functional type. How can respective trait flexibility be introduced in such models? Approaches to explain these large variations have focussed on the depth of the local water table as a constraint on the maximum rooting depth (Fan et al., 2017), on the mean seasonality and synchrony of precipitation and radiation (Gao et al., 2014; Schenk et al., 2005), or on optimality principles for balancing the trade-offs between the benefits of deep rooting and their associated costs (carbon used for root construction and respiration and maintaining water transport along the entire soil-root-leaf pathway) (Kleidon and Heimann, 1998; Schymanski et al., 2009).Here, we follow the approach by Gao et al. (2014), assuming that plants’ rooting depth is adjusted to sustain cumulative water deficits (precipitation minus evapotranspiration: P - ET) of a magnitude, corresponding to an event with a return period of N years, where N is a global constant. The water deficit is determined using daily reanalysis data of precipitation, while ET is estimated from remotely sensed vegetation cover and observed radiation, and from simulated stomatal responses to observed vapour pressure deficit across the globe for the past 35 years. Cumulative water deficits are translated into a “soil drying depth” using information on soil texture and its estimated water holding capacity.First results reveal a global pattern of plant rooting depth, adjusted to the apparent cumulative water deficit during rain-free periods, peaking at intermediate aridity, where dry periods are long, radiation is high, and vegetation cover is substantial. A comparison to biome-level distributions of observed rooting depths indicates that the control by the cumulative water deficit, embodied in our model, is a powerful driver of variations between biomes. Shallowest rooting depths are observed and modelled in boreal forests and alpine meadows, while deepest rooting is observed and modelled in xeric forests. Stomatal response to dry conditions appears to be an important mechanism that mitigates the necessity for deep rooting, particularly in broadleaved tropical and temperate forests. Our analysis also revealed challenges in bridging the scales of observations and our global-scale model, possibly due to small-scale heterogeneity in soil water availability driven by the topographic setting.

Plants and vegetation play a critical-but largely unpredictable-role in global environmental changes due to the multitude of contributing processes at widely different spatial and temporal scales. In this Perspective, we explore approaches to master this complexity and improve our ability to predict vegetation dynamics by explicitly taking account of principles that constrain plant and ecosystem behaviour: natural selection, self-organization and entropy maximization. These ideas are increasingly being used in vegetation models, but we argue that their full potential has yet to be realized. We demonstrate the power of natural selection-based optimality principles to predict photosynthetic and carbon allocation responses to multiple environmental drivers, as well as how individual plasticity leads to the predictable self-organization of forest canopies. We show how models of natural selection acting on a few key traits can generate realistic plant communities and how entropy maximization can identify the most probable outcomes of community dynamics in space- and time-varying environments. Finally, we present a roadmap indicating how these principles could be combined in a new generation of models with stronger theoretical foundations and an improved capacity to predict complex vegetation responses to environmental change.Integrating natural selection and other organizing principles into next-generation vegetation models could render them more theoretically sound and useful for earth system applications and modelling climate impacts.

Summary Terrestrial primary productivity and carbon cycle impacts of droughts are commonly quantified using vapour pressure deficit (VPD) data and remotely sensed greenness, without accounting for soil moisture. However, soil moisture limitation is known to strongly affect plant physiology.Here, we investigate light use efficiency, the ratio of gross primary productivity (GPP) to absorbed light. We derive its fractional reduction due to soil moisture (fLUE), separated from VPD and greenness changes, using artificial neural networks trained on eddy covariance data, multiple soil moisture datasets and remotely sensed greenness.This reveals substantial impacts of soil moisture alone that reduce GPP by up to 40% at sites located in sub‐humid, semi‐arid or arid regions. For sites in relatively moist climates, we find, paradoxically, a muted fLUE response to drying soil, but reduced fLUE under wet conditions. fLUE identifies substantial drought impacts that are not captured when relying solely on VPD and greenness changes and, when seasonally recurring, are missed by traditional, anomaly‐based drought indices. Counter to common assumptions, fLUE reductions are largest in drought‐deciduous vegetation, including grasslands. Our results highlight the necessity to account for soil moisture limitation in terrestrial primary productivity data products, especially for drought‐related assessments., See also the Commentary on this article by https://doi.org/10.1111/nph.15176.

Contents Summary 507 I. Introduction 507 II. The return on investment approach 508 III. CO2 response spectrum 510 IV. Discussion 516 Acknowledgements 518 References 518 SUMMARY: Land ecosystems sequester on average about a quarter of anthropogenic CO2 emissions. It has been proposed that nitrogen (N) availability will exert an increasingly limiting effect on plants' ability to store additional carbon (C) under rising CO2 , but these mechanisms are not well understood. Here, we review findings from elevated CO2 experiments using a plant economics framework, highlighting how ecosystem responses to elevated CO2 may depend on the costs and benefits of plant interactions with mycorrhizal fungi and symbiotic N-fixing microbes. We found that N-acquisition efficiency is positively correlated with leaf-level photosynthetic capacity and plant growth, and negatively with soil C storage. Plants that associate with ectomycorrhizal fungi and N-fixers may acquire N at a lower cost than plants associated with arbuscular mycorrhizal fungi. However, the additional growth in ectomycorrhizal plants is partly offset by decreases in soil C pools via priming. Collectively, our results indicate that predictive models aimed at quantifying C cycle feedbacks to global change may be improved by treating N as a resource that can be acquired by plants in exchange for energy, with different costs depending on plant interactions with microbial symbionts.

CO2 emissions from preindustrial land-use change (LUC) are subject to large uncertainties. Although atmospheric CO2 records suggest only a small land carbon (C) source since 5,000 y before present (5 kyBP), the concurrent C sink by peat buildup could mask large early LUC emissions. Here, we combine updated continuous peat C reconstructions with the land C balance inferred from double deconvolution analyses of atmospheric CO2 and [Formula: see text]C at different temporal scales to investigate the terrestrial C budget of the Holocene and the last millennium and constrain LUC emissions. LUC emissions are estimated with transient model simulations for diverging published scenarios of LU area change and shifting cultivation. Our results reveal a large terrestrial nonpeatland C source after the Mid-Holocene (66 [Formula: see text] 25 PgC at 7-5 kyBP and 115 [Formula: see text] 27 PgC at 5-3 kyBP). Despite high simulated per-capita CO2 emissions from LUC in early phases of agricultural development, humans emerge as a driver with dominant global C cycle impacts only in the most recent three millennia. Sole anthropogenic causes for particular variations in the CO2 record ([Formula: see text]20 ppm rise after 7 kyBP and [Formula: see text]10 ppm fall between 1500 CE and 1600 CE) are not supported. This analysis puts a strong constraint on preindustrial vs. industrial-era LUC emissions and suggests that upper-end scenarios for the extent of agricultural expansion before 1850 CE are not compatible with the C budget thereafter.