ARTICLE https://doi.org/10.1038/s42003-020-0839-y OPEN Increasing atmospheric CO2 concentrations correlate with declining nutritional status of European forests 1234567890():,; Josep Penuelas 1,2 ✉, Marcos Fernández-Martínez3, Helena Vallicrosa1,2, Joan Maspons 1,2, Paolo Zuccarini 1,2, Jofre Carnicer1,2, Tanja G.M. Sanders 4, Inken Krüger4, Michael Obersteiner5, Ivan A. Janssens 3, Philippe Ciais 6 & Jordi Sardans1,2 The drivers of global change, including increases in atmospheric CO2 concentrations, N and S deposition, and climate change, likely affect the nutritional status of forests. Here we show forest foliar concentrations of N, P, K, S and Mg decreased significantly in Europe by 5%, 11%, 8%, 6% and 7%, respectively during the last three decades. The decrease in nutritional status was especially large in Mediterranean and temperate forests. Increasing atmospheric CO2 concentration was well correlated with the decreases in N, P, K, Mg, S concentrations and the increase of N:P ratio. Regional analyses indicated that increases in some foliar nutrient concentrations such as N, S and Ca in northern Europe occurred associated with increasingly favourable conditions of mean annual precipitation and temperature. Crucial changes in forest health, structure, functioning and services, including negative feedbacks on C capture can be expected if these trends are not reversed. 1 CSIC, Global Ecology Unit CREAF-CSIC-UAB, 08913 Bellaterra, Catalonia, Spain. 2 CREAF, 08913 Cerdanyola del Vallès, Catalonia, Spain. 3 Research group PLECO (Plants and Ecosystems), Department of Biology, University of Antwerp, 2610 Wilrijk, Belgium. 4 Thünen Institute of Forest Ecosystems, Alfred- Möller-Straße 1, Haus 41/42, Eberswalde 16225, Germany. 5 International Institute for Applied Systems Analysis (IIASA), Ecosystems Services and Management, Schlossplatz 1, A-2361 Laxenburg, Austria. 6 Laboratoire des Sciences du Climat et de l’Environnement, IPSL, 91191 Gif-sur-Yvette, France. ✉email: josep.penuelas@uab.cat COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio 1 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y A tmospheric CO2 concentrations and nitrogen (N) and in Europe by 5%, 11%, 8%, 6% and 7%, respectively, and that sulfur (S) deposition, together with warming and drought, these decreases were especially large in Mediterranean and tem- likely affect the nutritional status of forests1–6 and perate forests and mainly related to the rising of atmospheric CO2 therefore their functioning, structure and ecosystem services7–9. concentration. In contrast, foliar N, S and Ca concentrations Elevated atmospheric CO2 concentrations, usually tested at increased associated with increasingly favourable conditions of 500–700 ppm, decrease the N and P concentrations of plants10–13. mean annual precipitation and temperature in boreal forests. Increases in atmospheric CO2 concentrations are frequently Crucial changes in forest health, structure, functioning and ser- correlated with higher growth10 and more efficient photosynth- vices, including negative feedbacks on C capture can be expected esis, and thus likely dilute leaf-level nutrient concentrations. if these trends in central and southern European forests are not Increases in atmospheric CO2 concentrations also reduce tran- reversed. spiration14 and stomatal conductance15, thus also hindering nutrient uptake16,17 that may even ultimately limit the initial Results increase in plant production under the rise of atmospheric Declining nutritional status. Foliar N, P and K concentrations CO2 concentrations18–20. N deposition also increases tree decreased in European forests during the last three decades, by 5%, productivity21,22 and foliar N concentrations but can decrease 11% and 8%, respectively (Fig. 1), especially in central and foliar P and Mg concentrations23–25. Warming tends to increase southern Europe (Fig. 2 and Supplementary Fig. 1). An exception is the mineralisation, cycling and availability of nutrients when northern Europe where foliar N concentration increased and foliar water is available12, but the consequent increase in growth P concentrations did not change. The foliar N:P ratio increased involves a dilution of nutrients that leads to frequent decreases everywhere by an average of 7% (Fig.1 and Supplementary Fig. 2). in foliar N concentrations26,27 and increases in C:nutrient Foliar S and Mg concentrations decreased in Europe, by 6% and ratios12,28. Plants at sites not limited by water respond by 7%, respectively (Fig. 3), although foliar S concentrations, as Ca increasing nutrient uptake29,30, but if warming persists or even concentrations, increased in northern Europe and decreased in increases in the long term, nutrients can become limiting31,32. central Europe (Fig. 4 and Supplementary Fig. 3). The trends were Warming in dry environments, though, can increase soil drought, not dominated by any extreme values; the effect of the anomalously exacerbating limitations of water and nutrients33,34. Plants under high or low years for Mg and S concentrations (Fig. 3) was these conditions respond by activating mechanisms for conser- minimised in the mixed model analyses. The results of the analyses ving and taking up water and nutrients but C:nutrient ratios still after removing years 1996 and 2012 (respectively anomalously high frequently increase in photosynthetic tissues7,35–38. and low) consistently showed that foliar Mg still decreased at a rate The concentrations of atmospheric CO2 have increased from of −0.0036 ± 0.0007 mg g−1 (P < 0.0001)—lower slope than in ~350–360 ppm in the 1990s to the current 410 ppm (in 2019)39. Fig. 3—and S decreased at a very similar rate as shown in Fig. 3 The deposition of oxidised N in some regions of the world such as (−0.0030 ± 0.0007 mg g−1, P < 0.0001). Europe peaked during the 1980s, up to 6–8-fold higher than in The foliar elemental concentrations followed similar trends in 1900, but has since decreased to half its highest value40,41. The all species (Fig. 2 and Supplementary Figs. 4–9), with few annual deposition of reduced Nin Europe is currently more than exceptions such as P and Mg increase in P. sylvestris at northern two-fold higher than in 190042. S deposition in Europe has latitudes (Figs. 2 and 4). Each species had a distinct foliar decreased to ~70% of the level in 190042. Europe has also warmed elemental composition (elementome), even though the individual faster than the global average of 0.27 °C per decade during the last trees grew under different environmental conditions and three decades, and this warming has varied throughout Europe. limitations at distinct latitudes (see Fig. 5 and Supplementary Temperatures have risen by 0.48, 0.44 and 0.34 °C per decade in Tables 3–10 for a DA and Supplementary Fig. 10 for similar northern, central and southern Europe, respectively, in the period results in a PCA; the DA and the PCA were applied to multi- 1979–201043,44. elemental data space to quantify the ‘elementome’), result These increasing CO2 concentrations, changes in N and S consistent with recent studies showing the strong species identity deposition and climate change have been accompanied by a in foliar elemental composition53. But despite the observed general decrease in foliar P concentrations and a consequent species identity in their foliar elemental composition, all studied increase in N:P ratios in recent decades in Fagus sylvatica24,45–47, seven species, i.e. F. sylvatica, P. sylvestris, A. alba, P. abies, Q. Picea abies and Pinus sylvestris24,48 and Quercus petraea24. Clear ilex, Q. petraea and Q. robur, changed their elemental composi- general patterns for foliar N concentrations, however, have not tion between 1990 and 2016. They shifted their foliar elemental been found, with decreases, increases or no changes, depending composition along the axis toward decreasing foliar P, K and Mg on species and foliar cohorts24,49–52.Likely local, regional or concentrations and increasing N:P ratio during 2005–2016 latitudinal differences have not been considered, so these changes relative to 1990–2004. The overall nutritional status of all species in foliar nutrient concentrations have not been consistently thus declined. The regional analysis, though, indicated that this attributed to a particular environmental driver or combination of decline did not extend to the northern forests. The foliar drivers. Furthermore, most reported nutritional changes in plants elemental composition of the trees at northern latitudes, mostly of refer to N and P concentrations, but other important nutrients P. sylvestris and P. abies, shifted toward increasing N but also to are key to the nutritional status of trees, such as K, S, Mg and Ca4. increasing N:P ratio (Fig. 5b) suggesting a softening of N We analysed (i) the changes in foliar elemental composition limitation. and stoichiometry during the last three decades for the main tree species in forests throughout Europe (Supplementary Fig. 1) at three different spatial scales: over the entire Europe, at different Possible drivers. CO2 concentrations during 1990–2016 latitudes, and locally, as well as (ii) the empirical relationships of increased everywhere by ca. 50 ppm, N and S deposition these changes with their possible drivers, i.e. increased atmo- decreased on average by ca. 25% and 65%,respectively, and spheric CO2 concentrations, N and S deposition and climate temperature increased everywhere, especially in the north by change, using statistical attribution analyses, data available from almost 1 °C, whereas precipitation increased by ca. 50 mm year−1 field experiments and models of the responses of nutrients to in the north and decreased by ca. 100 mm year−1 in the south these drivers of global change. The results showed that forest (Supplementary Fig. 11). The increase in atmospheric CO2 was foliar concentrations of N, P, K, S and Mg decreased significantly the only predictor systematically associated with the decreases in 2 COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y ARTICLE a ΔFoliar N: -0.038 ± 0.004 mg g−1 y−1; P<0.001 dioxide Enrichment) experiments to test for changes in foliar 20.0 nutrient concentrations in response to experimentally elevated 2 Foliar N (mg g−1) CO2. The decreases in N and P concentrations in response to Density 19.5 1 58% 42% 19.0 similar increases of 50 ppm CO2 were 3.0% and 2.3% respectively, 0 18.5 −1.5 −1.0 −0.5 0.0 0.5 1.0 with the consequent decrease in N:P ratios (Fig. 6b), i.e. both the ΔFoliar N (mg g−1 y−1) data from European forests and the data from elevated CO2 18.0 experiments showed a decreasing effect of increasing CO2 on N 17.5 and P concentration. The data from the experiments showed 17.0 though lower decreases than those observed in the European 1995 2000 2005 2010 2015 forests, especially for P (Fig. 1). b 1.5 ΔFoliar P: -0.006 ± 0.001 mg g y ; P<0.001 −1 −1 1.4 Foliar P (mg g−1) Discussion 1.3 The overall nutritional status of all species declined. The regional 20 analysis, though, indicated that this decline did not extend to the Density 1.2 10 65% 35% northern forests. A general decline in plant nutritional status has 1.1 0 −0.10 −0.08 −0.04 0.00 0.04 0.08 also been reported in other regions such as North America54. 1.0 ΔFoliar P (mg g−1 y−1) Herbarium analyses1,54–57 also indicated that declines in plant N 1995 2000 2005 2010 2015 have already occurred in the last century despite increases in N c ΔFoliar N:P: 0.037 ± 0.004 mg g−1 y−1; P<0.001 deposition. Our findings are also consistent with a recent study 15.5 suggesting a general global pattern of decreasing foliar N con- 15.0 centration ca. 9% over the last 40 years56 and with many exam- Foliar N:P 14.5 ples of local to regional decreases in foliar N and P 2.0 concentrations24,54. 14.0 Density 1.0 Consistently with our results, recent meta-analyses of elevated 42% 58% 13.5 0.0 CO2 experiments have found that rising CO2 concentrations have 13.0 −2 −1 0 ΔFoliar N:P (mg g−1 y−1) 1 2 led to decreases in N:P ratios in different plant tissues58 and 1995 2000 2005 2010 2015 woody plants59. Deng et al.58 hypothesise that the experimental 7.0 increases in atmospheric CO2 concentrations stimulate higher ΔFoliar K: -0.019 ± 0.002 mg g−1 y−1; P<0.001 plant uptakes of P than N. However, FACE and OTC experiments 6.8 are not fully comparable with a progressive increase of atmo- Foliar K (mg g−1) 6.6 spheric CO2 concentrations in natural conditions where not only 6.4 CO2 concentrations change but also many other factors at the 4 same time. For example, several European regions have become Density 3 6.2 2 1 58% 42% more arid in the last decades, especially in southern Europe60 and 6.0 0 rises of plant N:P ratios have been reported in response to −1.5 −1.0 −0.5 0.0 0.5 1.0 ΔFoliar K (mg g−1 y−1) increasing drought13,61. Consistently, we have also observed that 1995 2000 2005 2010 2015 MAT has contributed significantly to increase foliar N:P ratios. Year FACE and OTC, instead, estimate the CO2 effects by comparing treatment with control plots with all the other changes affecting Fig. 1 Decreasing tree foliar N, P and K concentrations and increasing N:P equally to both treatment and control plots. In ICP data for ratio. The black lines indicate average trends, and the shaded areas European forests some other not controlled factors may have also indicate the standard errors of the average trends. The inset shows the contributed to decrease more foliar P than N concentration. The percentages of forests with decreasing and increasing foliar nutrient frequency and intensity of some forest pests have increased62,63, concentrations. Red and blue indicate forests with increasing and so biotic factors not controlled in this study could have also decreasing trends, respectively. All values were adjusted to the same favoured the P-uptake drop64. Moreover, soil P availability tends mean to remove forest-specific variability. See Supplementary Table 1 for to decrease through time by natural processes65 whereas N detailed results of the model lme (foliar variable ∼year, random = ∼1| availability can continuously be maintained or increase by con- country/plot/species, data=dades, method = “REML”). Statistically tinuous loads of N deposition and N2-fixation. The short-term significant trends (in percentage of sites): N+: 5.68%, N−: 13.18%; P+: character of these experiments also does not allow the develop- 1.21%, P−: 17.30%; N:P+: 14.72%, N:P−: 4.29%; K+: 2.02%, K−: 9.90%. ment of long-term processes, such as the feed-backs due to lower litter quality and the decreases in N and S deposition in recent nutrient concentrations (Fig. 6). The mixed-model analyses at the decades5. The experimental decrease may have been smaller also level of individual trees produced similar results (Supplementary because the experiments test responses to increases in atmo- Table 2). The regional analyses indicated that the increases in spheric CO2 concentrations in a less sensitive range of higher CO2 some foliar nutrient concentrations in northern Europe were concentrations than the actual current range of 360–410 ppm. associated with increasingly favourable MATs at northern lati- The decrease in N deposition was slightly and positively related tudes (see interactions CO2 x MAT in Supplementary Table 1). with foliar Mg and negatively with K concentrations. The These mixed-model analyses also included tree growth, which decrease in S deposition was slightly and positively related with was not selected in any of the explanatory models. The decreases foliar K and negatively with Mg concentrations. The effects of the in nutrient concentrations were thus not due to a dilution as a shifts in N and S deposition on foliar elemental composition result of individual tree growth, because the increase in the dia- during the studied period were thus weak. Sulfur deposition has meter of the tree stems at breast height was never in the most dropped in general across Europe since 1980s, but N deposition parsimonious models for the driving factors. despite having decreased in general at European scale in the last We then conducted a meta-analysis of the available literature two decades, has not decreased in all sites, N loads, despite lower, on open-top chambers (OTC) and FACE (Free-Air Carbon continue being substantial, and, in general, no symptoms of COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio 3 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y a Entire Europe N P K 0.3 0.00 0.1 0.0 -0.04 0.0 -0.1 -0.3 -0.06 -0.2 -0.6 -0.12 b Pinus sylvestris N P K 0.025 0.0 0.2 0.000 -0.1 0.0 -0.025 -0.2 -0.2 -0.050 -0.3 -0.4 -0.4 c Picea abies N P K 0.4 0.00 0.3 0.00 0.2 -0.02 -0.05 0.1 -0.04 -0.10 0.0 -0.06 -0.15 -0.1 -0.08 d Fagus sylvaca N P K 0.2 0.00 0 0.1 -0.25 -1 -0.50 0.0 -2 -0.1 -0.75 -1.00 Fig. 2 Geographical distribution of the annual rate of change for N, P, and K foliar concentration in mg g−1. a the entire forests, b Pinus sylvestris, c Picea abies, and d Fagus sylvatica. The estimations are based on neural networks with MAP, MAT and nutrient deposition as inputs, using 80% of the trees with more than five measurements for training and 20% for validation. We replicated the process 1000 times and averaged the results. Map spatial resolution 0.1°. 4 COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y ARTICLE a 7.0 mineralomasses9,35,70–72. A widespread decline in crown condi- ΔFoliar Ca: -0.0026 ± 0.0029 mg g−1 y−1; P=0.37 tion, disruption of food webs and increased tree mortality with increased drought associated with climate change have also been Foliar Ca (mg g−1) 6.5 reported for these southern European forests60. 6.0 Other foliar chemistry factors and processes not-measured in 6 ICP Forest such as resorption or tree age could also be underlying Density 5.5 4 65 % 35 % the observed decrease of foliar nutritional condition. However, 2 5.0 the leaves selected for foliar analyses are mature non-senescent 0 −1.5 −1.0 −0.5 0.0 leaves and the trees selected for foliar analyses are adult non- ΔFoliar Ca (mg g−1 y−1) senescent trees, and given the frequent long life of trees, the 1995 2000 2005 2010 2015 b changes in tree age during the studied period, ~25 years as ΔFoliar Mg: -0.0055 ± 0.0009 mg g−1 y−1; P<0.001 1.6 maximum in the individual plots, should not affect much the foliar concentration. Furthermore73, reported that foliar N con- Foliar Mg (mg g−1) 1.4 tents and dry weight tended to slightly decrease with age but that 30 this was not the case for N concentrations in Fagus sylvatica and 1.2 Picea abies in European forests. Density 20 10 Nutrient impoverishment can have multiple effects on the 1.0 53 % 47 % 0 structure, function and ecosystem services of forests. For example, −0.10 −0.05 0.00 0.05 N and P are fundamental to C assimilation and protein synthesis, 0.8 ΔFoliar Mg (mg g−1 y−1) so their decreased concentrations could constrain the capacity to 1995 2000 2005 2010 2015 c take up carbon and the effect of CO2 fertilisation in forests8. ΔFoliar S: -0.0036 ± 0.0007 mg g−1 y−1; P<0.001 Foliar N:P ratios are negatively correlated with plant net photo- 1.4 synthesis and growth3, so the increasing foliar N:P ratios in Foliar S (mg g−1) European forests (Fig. 2) suggested a worsening nutrient imbal- 1.2 80 ance that may partly account for a lower effect of CO2 fertilisa- 60 tion8. The consequent changes in plant C:N and C:P Density 40 1.0 61 % 39 % stoichiometries can also drive ecosystem-level N availability by 20 0 the effects on litter quality decrease, microbial N immobilisation −0.02 0.00 0.02 0.04 0.06 0.08 and mineralisation19. The reduction in the availability of N may 0.8 ΔFoliar S (mg g−1 y−1) in turn affect the efficiency of sequestration of C. Increases in 1995 2000 2005 2010 2015 foliar C and decreases in N and P are associated with increases in Year non-structural carbohydrates and carbon-based secondary Fig. 3 Decreasing tree foliar Ca, Mg and S concentrations. The black lines metabolites41,74 and decreases in foliar protein content, thus indicate the average trends, and the shaded areas indicate the standard decreasing the nutritional quality of plants56,75,76 for wildlife and errors of the average trends. The inset shows the percentages of forests livestock. with decreasing and increasing foliar nutrient concentrations. Red and blue We thus conclude that foliar concentrations of N, P, K, S and indicate forests with increasing and decreasing trends, respectively. All Mg are generally decreasing in European forests. These decreases values were adjusted to the same mean to remove forest-specific variability. are generally larger for P than N, so the foliar N:P ratio has See Supplementary Table 1 for detailed results of the model lme (foliar increased in most European forests. These trends are mostly variable ∼year, random = ∼1|country/plot/species, data = dades, method = associated with increasing atmospheric CO2 concentrations that “REML”). Statistically significant trends (in percentage of sites):Ca+: 7.46%, have led to a higher nutrient demand by trees. The soil nutrient Ca−: 6.45%; Mg+: 7.11%, Mg−: 8.13%; S+: 6.09%, S−: 12.19%. supply was probably not always sufficient to meet the growing demands by trees, which could partly explain the deterioration of tree mineral nutrition. These decreasing trends are stronger in southern and central Europe than in northern Europe where the significant recovery of soil status have been observed (Schmitz concentrations of some elements are even increasing, all in con- et al.5, and the references therein). N deposition can cause defi- sonance with the increase in MAT favouring nutrient availability ciencies in other nutrients than N and nutrient imbalances due to and uptake in northern Europe while hindering them in a range of effects, including stimulation of plant growth (dilution increasingly dry southern Europe. These nutrient limitations for effect) and negative effects on tree nutrient acquisition by mod- forest growth should be taken into account by the scientific and ifying mycorrhizal associations24,66. Increasing mean annual environmental management communities to avoid over- temperature (MAT) also contributed to the decrease in Mg and estimations of forest productivity in response to elevated atmo- the increase in Ca and N:P. Ca Moreover, in general, increases in spheric CO2 when developing global climate models and soil pH translate into higher foliar cation concentrations67–69. projections. The consequences of such pervasive nutrient The higher temperatures at the northern European sites impoverishment can be key for forest structure, function, health favoured longer growing seasons, biological activity and nutrient and capacity to provide ecosystem services. up-take, thus accounting for the lower general decreases in foliar nutrient concentrations and even the increase in N. In contrast, the increases in temperature and consequently in drought Methods Data acquisition. We used foliar and growth data of the International Co- (decreases in mean annual precipitation (MAP) in southern operative Programme on Assessment and Monitoring of Air Pollution Effects on Europe could account for the decline in soil fertility and capacity Forests operating under the UNECE Convention on Long-range Transboundary of nutrient uptake, all of which contribute to a decrease in foliar Air Pollution (CLRTAP) (ICP Forests). Activities under ICP Forests are conducted nutrient concentration, as in experimental drought studies con- in 42 states using harmonised sampling and analysis following the manuals on ducted in Mediterranean forests where mineralisation, soil Sampling and Analysis of needles and leaves—Part XII: http://www.icp-forests.org/ pdf/manual/2016/ICP_Manual_2017_01_part12.pdf and Tree growth—Part V: enzymatic activity and plant growth also decreased, thus (http://www.icp-forests.org/pdf/manual/2016/ICP_Manual_2016_01_part05.pdf). leading to a large decrease in aboveground nutrient In this study, we have gathered all the available data of foliar N, P, K, S, Ca and Mg COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio 5 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y a Entire Europe Ca Mg S 0.6 0.050 0.025 0.02 0.4 0.2 0.000 0.00 0.0 -0.025 -0.02 - -0.2 -0.050 -0.04 - -0.4 b Pinus sylvestris Ca Mg S 0.04 0.2 0.06 0.02 0.1 0.04 0.00 0.02 -0.02 0.0 0.00 -0.04 -0.1 -0.02 c Picea abies Ca Mg S 0.06 0.050 0.2 0.04 0.02 0.025 0.1 0.000 0.00 0.0 -0.02 -0.025 -0.04 d Fagus sylvaca Ca Mg S 0 0.1 0.3 0.0 -1 0.2 -0.1 -2 0.1 -0.2 0.0 -0.3 -3 -0.4 Fig. 4 Geographical distribution of the trend in annual rate of variation for Ca, Mg, and S foliar concentration in mg g−1. a the entire forests, b Pinus sylvestris, c Picea abies, and d Fagus sylvatica. The estimations are based on neural networks using 80% of the trees with more than five measurements for training and 20% for validation. We replicated the process 1000 times and averaged the results. 6 COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y ARTICLE Fig. 5 Shifts of the elementome (elemental composition) of European tree species toward lower nutrient concentrations except for northern latitudes. Discriminant analyses of the foliar elemental concentrations and N:P ratios for the seven dominant species, Pinus sylvestris, Picea abies, Abies alba, Fagus sylvatica, Quercus robur, Quercus petraea and Quercus ilex, for the entire Europe and for northern, central and southern Europe. All plots compare the data for 1990–2004 with the data for 2005–2016. The circles/ellipses for each species and period depict the mean position and the space occupied by the 95% confidence interval for the scores of each species. concentrations in annual series. We have used data from 28 countries with a total v3.25 of the Climatic Research Unit77. Annual data for N (NO3− + NH4+) and S of 528 different plots and with 506 of these plots with the canopy clearly dominated (SO4−) atmospheric deposition were extracted from the European Monitoring and by one tree species, 21 co-dominated by two tree species, and 1 plot co-dominated Evaluation Programme (EMEP) with a spatial resolution of 0.1 × 0.1° for longitude by three tree species. and latitude, and the MSC-W chemical transport model developed to estimate Foliar samples were taken at least biannually at the intensive forest monitoring regional atmospheric dispersion and deposition of acidifying and eutrophying plots (Level II) from the living crown of the dominant canopy layer providing compounds of N and S over Europe. information on the nutrient status of one or more species per plot. The analysis was done for 1000 needles or 100 leaves covering a range of elements (for detailed information please refer to the ICP Forest Manual). Briefly, in each plot a Statistics and reproducibility. First, we used time series of observations of foliar minimum of five dominant trees were randomly selected avoiding the trees used in N, P, K, S, Ca and Mg concentrations at 410 European sites (Supplementary crown assessment, so to avoid crown damage of these trees. A composite sample by Fig. 12) for the last three decades (1990–2016) to investigate their temporal trends species was made by mixing equal quantities of each sample per plot after drying. using mixed models, where year was the fixed covariate and site-species was the Sampled leaves were mature current year leaves or needles. Only in the case of random factor. Second, we repeated the analyses separately for northern, central evergreens such as Q. ilex both mature (non-senescent) one- and second-year and southern European forests (separated by parallels at 46° and 58° North; leaves were sampled. The sample analyses procedure was based on homogenised Supplementary Fig. 1b). Third, we repeated the previous two analyses for each of methods, basically direct C and N determination by elemental-analyser (Kjeldhal the most abundant species: F. sylvatica, P. sylvestris, and P. abies. Fourth, we method was also allowed for N determination), whereas for the other elements the predicted the rate of changes in foliar elemental concentration across Europe as a most common procedure was based on acid digestion with nitric acid or nitric acid function of MAP, MAT and nutrient deposition rate of change using neural net- mixtures (43–46%) followed by wet ashing (40%) and posterior determination, works. We calculated the rate of changes of the foliar elements for each tree with mostly by inductively coupled plasma coupled to atomic emission spectrometry more than five measurements and the rate of change of the MAP, MAT and (ICP-AES). Quality assurance and assessment of the analytical process was nutrient deposition for the same period and location using Theil-Sen robust controlled by the regular organisation of Inter-laboratory Comparisons by the regressions implemented in mblm R package71. Then, we used 80% of the esti- Forest Foliar Co-ordinating Centre. The laboratory results were considered of mated rates to train the neural networks and 20% for validation using keras78 in sufficient quality when the laboratory receives a qualification for the concerning R79 with TensorFlow80 backend. The neural networks had two hidden layers with parameter(s) after participation in the inter-laboratory comparisons. The growth 128 units each with rectified linear activation functions. We repeated the process survey assessed the diameter at breast height (dbh) for the dominant trees at a five 1000 times making predicted maps and averaging the results. Finally we masked or ten yearly interval. Additionally, permanent girth bands provided annual data the pixels with no forest81 and for species specific models, we also applied a mask of dbh. with the distribution maps82–84. Fifth, we used multivariate analyses, including a For atmospheric CO2 concentration, we used monthly data from the Mauna discriminant analysis (DA) and a principal component analysis (PCA) of all Loa Observatory, available online (http://scrippsco2.ucsd.edu/data/ nutrient variables to analyse the shifts in the elementome53,85,86 for each species for atmospheric_co2/) and provided by the Scripps Institution of Oceanography the entire Europe and each of the three latitudes. (Scripps CO2 programme). We calculated the annual averages to use in our We explored which environmental factors could better explain the observed statistical analyses. We obtained the meteorological-climate data from CRU TS changes in foliar elemental concentrations. To do so, we estimated the temporal COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio 7 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y Fig. 6 a Factors contributing to the decreasing tree foliar N, P, K, S and Mg concentrations and b a meta-analysis of the effects of experimentally elevated atmospheric CO2 concentrations (using both FACE and OTC methodologies) on the N, P, K, Mg and Ca concentrations and the N:P ratio of green mature leaves for all types of vegetation. a Plot of the temporal contribution of the predictor variables on N, P, K, S and Mg concentrations and the N:P ratio (Ca concentration did not change significantly; see Fig. 2). The models (see Supplementary Table 1) suggest that increasing CO2 is the main and only contributor to the decreases in N, P, K and S concentrations. The temporal variations of the predictors are shown in Supplementary Fig. 12. Error bars indicate standard errors. Units are ppm for CO2, kg ha−1 y−1 for S and N deposition and °C for temperature. See Methods for information about the methodology used to calculate the contributions. (*), P < 0.1; *P < 0.05; **P < 0.01; ***P < 0.001. b Meta-analysis of published studies (353, 297 and 684 studies for N and P concentrations and the N:P ratio, respectively). See references in Supplementary material for Fig. 6b. contributions of the predictor variables to the trends of foliar N, P and K ARMA (p = 1, q = 0) structure to account for temporal autocorrelation. We then concentrations and the N:P ratio following the methodology established by calculated the trends of foliar nutrient concentrations predicted by the final model Fernandez-Martinez et al.41,87. We first fitted one model for each foliar nutrient (containing all significant predictors) and the trends of the foliar nutrient using a generalized linear mixed model (GLMM), with the species and concentrations predicted by the same model but sequentially maintaining the experimental plot as random factors, using the lme function from the nlme R temporally varying predictors constant (e.g. temporal anomalies in MAT were held package88 and an autoregressive and moving-average (ARMA) (p = 1, q = 0) constant using the median for each site, but all other predictors varied based on the temporal autocorrelation structure using the corCAR1 function. We then fitted the observations). The difference between the prediction of the final model and the full (saturated) models for each foliar nutrient as a function of atmospheric CO2, prediction of the model when one predictor was controlled was assumed to be the MAT and MAP and N (oxidised and reduced) and S deposition. We also included contribution of that predictor to the temporal change in foliar nutrient the interactions between mean site values and their temporal anomalies to account concentrations. The differences between the sum of all temporal contributions and for different effects of, for example, increasing temperature (annual anomalies) the observed trend of a foliar nutrient were considered as unknown temporal depending on the average temperature of the site (e.g. decreasing temperature may contributions. All errors were calculated using error propagation. have a positive effect in warm climates but a negative effect in cold climates). We We finally conducted a meta-analysis of published experimental data for the also included the first-order interactions between atmospheric CO2, climate and environmental factor that was best correlated with nutrient depletion: CO2 variables of atmospheric deposition. We then removed non-significant terms from fertilisation. We used the keywords: atmospheric CO2, C:N, CO2 fertilisation, C:P, the full models until obtaining the final model (containing only significant terms) decrease, dilution, FACE, foliar, increase, leaf, needle, nitrogen, N:P, phosphorus, using stepwise backward-forward selection. The amount of the variance explained photosynthetic tissues, rise, and stoichiometry in our web search in Web of Science by the models was assessed using the r.squared GLMM function in R and google scholar between 1988 and 2018. We gathered the available studies that (MuMInpackage: Barton89) following the method of Nakagawa and Schielzeth)90. tested for the effects of elevated atmospheric CO2 concentrations (both using FACE We next used the final models (lme models explained above) to identify the and OTC methodologies) on N and P concentrations and the N:P ratio of green drivers contributing to the changes in foliar nutrient concentrations using the mature leaves for all types of vegetation (353, 297 and 684 studies, respectively). TempCont R package. We first calculated the trend (mean ± standard error of the The list of the published articles considered appears into supplementary Material. mean) of a foliar nutrient concentration using raw data with GLMMs with an The R metafor (v. 2.0-0) and forest plot (v. 1.7.2) packages were used for the 8 COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y ARTICLE analyses as described in Hedges et al.91. We calculated the response ratio (lnRR) as 18. Murray, M. B., Smith, R. I., Friend, A. & Jarvis, P. G. Effect of elevated [CO2] ln (Xi/Xn) = lnXi - lnXn, where Xi and Xn are the values of each chemical and varying nutrient application rates on physiology and biomass compound in leaf tissue observation before and after the treatment respectively. accumulation of Sitka spruce (Picea sitchensis). Tree Physiol. 20, 421–434 The sampling variance for each lnRR was calculated as ln[(1/ni) × (Si/Xi)2 + (1/ (2000). nn) × (Sn/Xn)2], where ni, nn, Si, Sn, Xi and Xn are the post-treatment and control 19. Luo, Y. et al. Progressive nitrogen limitation of ecosystem responses to rising sample sizes, standard deviations, and mean response values, respectively. The atmospheric carbon dioxide. Bioscience 54, 731 (2004). sensitivity was evaluated per ppm of CO2 added in the treatment. We thereafter 20. Norby, R. J., Warren, J. M., Iversen, C. M., Medlyn, B. E. & McMurtrie, R. E. standardised the effect size to an increase of 50 ppm since atmospheric CO2 CO2 enhancement of forest productivity constrained by limited nitrogen concentrations increased everywhere by ca. 50 ppm during 1990–2016. availability. Proc. Natl Acad. Sci. 107, 19368–19373 (2010). 21. Hans-Peter Kahle et al. Causes and Consequences of Forest Growth Trends in Reporting summary. Further information on research design is available in Europe – Results of the RECOGNITION Project. (Brill, 2008). the Nature Research Reporting Summary linked to this article. 22. Schulte-Uebbing, L. & de Vries, W. Global-scale impacts of nitrogen deposition on tree carbon sequestration in tropical, temperate, and boreal forests: A meta-analysis. Glob. Chang. Biol. 24, e416–e431 (2018). Data availability 23. Sardans, J. et al. Foliar and soil concentrations and stoichiometry of nitrogen The datasets generated during and/or analysed during the current study are available and phosphorous across European Pinus sylvestris forests: relationships with from the corresponding author, Prof. Josep Penuelas. climate, N deposition and tree growth. Funct. Ecol. 30, 676–689 (2016). 24. Jonard, M. et al. Tree mineral nutrition is deteriorating in Europe. Glob. Code availability Chang. Biol. 21, 418–430 (2015). All codes are available upon request to the corresponding author, Professor Josep 25. Waldner, P. et al. Exceedance of critical loads and of critical limits impacts tree Penuelas. nutrition across Europe. Ann. Sci. 72, 929–939 (2015). 26. Suzuki, S. & Kudo, G. Responses of alpine shrubs to simulated environmental change during three years in the mid-latitude mountain, northern Japan. Received: 22 August 2019; Accepted: 20 February 2020; Ecography 23, 553–564 (2000). 27. Shen, H., Du, H., Wang, Z. & Huang, B. Differential responses of nutrients to heat stress in warm-season and cool-season Turfgrasses. HortScience 44, 2009–2014 (2009). 28. Hyvönen, R. et al. The likely impact of elevated [CO2], nitrogen deposition, increased \ntemperature and management on carbon sequestration in References temperate and \nboreal forest ecosystems: a literature review. N. Phytol. 173, 1. Peñuelas, J. & Matamala, R. Changes in N and S leaf content, stomatal density 463–480 (2007). and specific leaf area of 14 plant species during the last three centuries of CO2 29. Dijkstra, F. A., Morgan, J. A., Fischer, J. Cvon & Follett, R. F. Elevated CO 2 increase. J. Exp. Bot. 41, 1119–1124 (1990). and warming effects on CH 4 uptake in a semiarid grassland below optimum 2. Peñuelas, J. & Matamala, R. Variations in the mineral composition of soil moisture. J. Geophys. Res. 2, 426–437 (2011). Herbarium plant species collected during the last three centuries. J. Exp. Bot. 30. Olsrud, M., Carlsson, B. Å., Svensson, B. M., Michelsen, A. & Melillo, J. M. 44, 1523–1525 (1993). Responses of fungal root colonization, plant cover and leaf nutrients to long- 3. Peñuelas, J. et al. Human-induced nitrogen-phosphorus imbalances alter term exposure to elevated atmospheric CO2 and warming in a subarctic birch natural and managed ecosystems across the globe. Nat. Commun. 4, 2934 forest understory. Glob. Chang. Biol. 16, 1820–1829 (2010). (2013). 31. Jonasson, S., Castro, J. & Michelsen, A. Litter, warming and plants affect 4. Sardans, J. et al. Elemental composition of European forest tree species results respiration and allocation of soil microbial and plant C, N and P in arctic from evolutionary traits and the present environmental and competitive mesocosms. Soil Biol. Biochem. 36, 1129–1139 (2004). conditions. Glob. Ecol. Biogeogr. 24, 240–255 (2015). 32. Jarvis, P. & Linder, S. Constraints to growth of boreal forests. Nature 405, 5. Schmitz, A. et al. Responses of forest ecosystems in Europe to decreasing 904–905 (2000). nitrogen deposition. Environ. Pollut. 244, 980–994 (2019). 33. Link, S. O., Smith, J. L., Halvorson, J. J. & Bolton, H. A reciprocal transplant 6. Sardans, J. et al. Plant invasion is associated with higher plant-soil nutrient experiment within a climatic gradient in a semiarid shrub-steppe ecosystem: concentrations in nutrient poor-environments. Glob. Chang. Biol. https://doi. effects on bunchgrass growth and reproduction, soil carbon, and soil nitrogen. org/10.1111/gcb.13384 (2016). Glob. Chang. Biol. 9, 1097–1105 (2003). 7. Sardans, J. & Peñuelas, J. Plant-soil interactions in Mediterranean forest and 34. Allison, S. D. & Treseder, K. K. Warming and drying suppress microbial shrublands: Impacts of climatic change. Plant Soil 365, 1–33 (2013). activity and carbon cycling in boreal forest soils. Glob. Chang. Biol. 14, 8. Peñuelas, J. et al. Shifting from a fertilization-dominated to a warming- 2898–2909 (2008). dominated period. Nat. Ecol. Evol. 1 (2017). 35. Sardans, J., Peñuelas, J., Prieto, P. & Estiarte, M. Drought and warming 9. Peñuelas, J. et al. Assessment of the impacts of climate change on induced changes in P and K concentration and accumulation in plant biomass Mediterranean terrestrial ecosystems based on data from field experiments and soil in a Mediterranean shrubland. Plant Soil 306, 261–271 (2008). and long-term monitored field gradients in Catalonia. Environ. Exp. Bot. 152, 36. Milla, R., Castro-Díez, P., Maestro-Martínez, M. & Montserrat-Martí, G. 49–59 (2018). Relationships between phenology and the remobilization of nitrogen, 10. Reich, P. B., Hungate, B. A. & Luo, Y. Carbon-nitrogen interactions in phosphorus and potassium in branches of eight Mediterranean evergreens. terrestrial ecosystems in response to rising atmospheric carbon dioxide. Annu. N. Phytol. 168, 167–178 (2005). Rev. Ecol. Evol. Syst. 37, 611–636 (2006). 37. Yarie, J. & VanCleve, K. Effects of carbon, fertilizer, and drought on foliar 11. Tognetti, R. & Penuelas, J. Nitrogen and carbon concentrations, and stable chemistry of tree species in interior Alaska. Ecol. Appl. 6, 815–827 (1996). isotope ratios in Mediterranean shrubs growing in the proximity of a CO2 38. Limousin, J. M., Misson, L., Lavoir, A. V., Martin, N. K. & Rambal, S. Do spring. Biol. Plant. 46, 411–418 (2003). photosynthetic limitations of evergreen Quercus ilex leaves change with long- 12. Sardans, J. & Peñuelas, J. The role of plants in the effects of global change on term increased drought severity? Plant, Cell Environ. 33, 863–875 (2010). nutrient availability and stoichiometry in the plant-soil system. Plant Physiol. 39. NOAA. Carbon dioxide levels rose at record pace for 2nd straight year. http:// 160, 1741–1761 (2012). www.noaa.gov/news/carbon-dioxide-levels-rose-at-record-pace-for-2nd- 13. Yuan, Z. Y. & Chen, H. Y. H. Decoupling of nitrogen and phosphorus in straight-year.2018 (2018). terrestrial plants associated with global changes. Nat. Clim. Chang. 5, 465–469 40. Lajtha, K. & Jones, J. Trends in cation, nitrogen, sulfate and hydrogen ion (2015). concentrations in precipitation in the United States and Europe from 1978 to 14. Del Pozo, A. et al. Gas exchange acclimation to elevated CO2 in upper-sunlit and 2010: a new look at an old problem. Biogeochemistry 116, 303–334 (2013). lower-shaded canopy leaves in relation to nitrogen acquisition and partitioning 41. Fernández-Martínez, M. et al. Atmospheric deposition, CO2, and change in in wheat grown in field chambers. Environ. Exp. Bot. 59, 371–380 (2007). the land carbon sink. Sci. Rep. 7, 1–13 (2017). 15. Samarakoon, A. B. & Gifford, R. M. Soil water content under plants at high 42. Engardt, M., Simpson, D., Schwikowski, M. & Granat, L. Deposition of CO 2 concentration and interactions with the direct CO 2 effects: a species sulphur and nitrogen in Europe 1900–2050. Model calculations and comparison. J. Biogeogr. 22, 193–202 (1995). comparison to historical observations. Tellus, Ser. B Chem. Phys. Meteorol. 69, 16. Pritchard, S. & Rogers, H. Spatial and temporal deployment of crop. N. Phytol. 1–20 (2017). 147, 55–71 (2000). 43. Agency, E. E. Global and European temperature. www.eea.europa.eu/data- 17. BassiriRad, H., Gutschick, V. P. & Lussenhop, J. Root system adjustments: and-maps/indicators/global-and-european-temperature-8/assessment. (2018). Regulation of plant nutrient uptake and growth responses to elevated CO2. 44. IPCC. Climate Change 2013: The Physical Science Basis. Contribution of Oecologia 126, 305–320 (2001). Working Group I to the Fifth Assessment Report of the Intergovernmental Panel COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio 9 ARTICLE COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y on Climate Change (eds Stocker, T. F., D. Qin, G-K. Plattner, M. Tignor, S. K. 69. Collin, A., Messier, C., Côté, B., Fontana, M. & Bélanger, N. Contrasting Allen, J. Boschung, A. Nauels, Xia, Y). Cambridge University Press, 2013). nutritional acclimation of Sugar Maple (Acer saccharum Marsh.) and red https://doi.org/10.1017/CBO9781107415324.Summary. maple (Acer rubrum L.) to increasing conifers and soil acidity as 45. Merino, A., Real, C. & Rodríguez-Guitián, M. A. Nutrient status of managed demonstrated by foliar nutrient balances. Front. Ecol. Evol. 4, 1–12 (2016). and natural forest fragments of Fagus sylvatica in southern Europe. For. Ecol. 70. Sardans, J. & Peñuelas, J. Drought decreases soil enzyme activity in a Manag. 255, 3691–3699 (2008). Mediterranean Quercus ilex L. forest. Soil Biol. Biochem. 37, 455–461 46. Jonard, M., André, F., Dambrine, E., Ponette, Q. & Ulrich, E. Temporal trends (2005). in the foliar nutritional status of the French, Walloon and Luxembourg broad- 71. Sardans, J. & Peñuelas, J. Soil enzyme activity in a mediterranean forest after leaved plots of forest monitoring. Ann. For. Sci. 66, 412–421 (2009). six years of drought. Soil Sci. Soc. Am. J. 74, 838 (2010). 47. Talkner, U. et al. Phosphorus nutrition of beech (Fagus sylvatica L.) is 72. Sardans, J., Peñuelas, J., Estiarte, M. & Prieto, P. Warming and drought alter C decreasing in Europe. Ann. For. Sci. 72, 919–928 (2015). and N concentration, allocation and accumulation in a Mediterranean 48. Mellert, K. H., Prietzel, J., Straussberger, R. & Rehfuess, K. E. Long-term shrubland. Glob. Chang. Biol. 14, 2304–2316 (2008). nutritional trends of conifer stands in Europe: results from the 73. Bauer, G., Schulze, E.-D. & Mund, M. Nutrient contents and concentrations in RECOGNITION project. Eur. J. For. Res. 123, 305–319 (2004). relation to growth of Picea abies and Fagus sylvatica along a European 49. Jonard, M. et al. Deterioration of Norway spruce vitality despite a sharp transect. Tree Physiol. 17, 777–786 (1997). https://doi.org/10.1093/treephys/ decline in acid deposition: A long-term integrated perspective. Glob. Chang. 17.12.777. Biol. 18, 711–725 (2012). 74. Peñuelas, J. & Estiarte, M. Can elevated CO2 affect secondary metabolism and 50. Sanders, T. G. M., Schmitz, A. & Edinger, J. Trends in foliar nitrogen and ecosystem function? Trends Ecol. Evol. 13, 20–24 (1998). phosphorus concentrations and ratios since 2000. Forest Condition in Europe - 75. Leakey, A. D. B. et al. Elevated CO2 effects on plant carbon, nitrogen, and 2017 Technical Report of ICP Forests. Report under the UNECE Convention on water relations: Six important lessons from FACE. J. Exp. Bot. 60, 2859–2876 Long-range Transboundary Air Pollution (CLRTAP), BFW Dokumentation 24/ (2009). 2017. (eds Seidling, W., Michel, A.) (BFW Austrian Research Centre for 76. Loladze, I. Rising atmospheric CO2 and human nutrition: toward globally Forests, 2017). imbalanced plant stoichiometry? Trends Ecol. Evol. 17, 457–461 (2002). 51. Verstraeten, A. et al. Multiple nitrogen saturation indicators yield 77. Harris, I., Jones, P. D. D., Osborn, T. J. J., & Lister, D. H. H. Updated high- contradicting conclusions on improving nitrogen status of temperate forests. resolution grids of monthly climatic observations - the CRU TS3.10 Dataset. Ecol. Indic. 82, 451–462 (2017). Int. J. Climatol. 34, 623–642 (2013). https://doi.org/10.1002/joc.3711. 52. Wellbrock, N., Bolte, A., Flessa, H. & (Eds). Waldernährung. Dynamik und 78. Chollet, F. & Allaire, J. keras: R Interface to ‘Keras’. R package version 2.2.4.1. räumliche Muster forstlicher Standorte in Deutschland - Ergebnisse der https://CRAN.R-project.org/package=keras. https://github.com/rstudio/keras. Bodenzustandserhebung im Wald 2006 bis 2008, Thünen Report. Johann (2019). Heinrich von Thünen-Institut, Braunschweig, https://doi.org/10.3220/ 79. R Core Team. R: A language and environment for statistical computing. REP1473930232000 (2016). https://www.r-project.org/. (Vienna, 2018). 53. Peñuelas, J. et al. The bioelements, the elementome and the “biogeochemical 80. Abadi, M. et al. TensorFlow: Large-Scale Machine Learning on Heterogeneous niche”. Nat. Ecol. Evol. 100, e02652 (2019). Distributed Systems. Software available from tensorflow.org. http://arxiv.org/ 54. McLauchlan, K. K. et al. Centennial-scale reductions in nitrogen availability in abs/1603.04467. (2015). temperate forests of the United States. Sci. Rep. 7, 1–7 (2017). 81. Hansen, M. C. et al. High-resolution global maps of 21st-Century forest cover 55. McLauchlan, K. K., Ferguson, C. J., Wilson, I. E., Ocheltree, T. W. & Craine, J. change. Science 342, 850–853 (2013). M. Thirteen decades of foliar isotopes indicate declining nitrogen 82. Mátyás, C., Ackzell, L. & C.J.A. Samuel. EUFORGEN: Technical guidelines for availability in central North American grasslands. N. Phytol. 187, 1135–1145 genetic conservation and use for Scots pine (Pinus sylvestris). (International (2010). Plant Genetic Resources Institute, R. 2004). 56. Craine, J. M. et al. Isotopic evidence for oligotrophication of terrestrial 83. Koski, V., Skrøppa, T., Paule, L., Wolf, H. & Turok, J. EUFORGEN: Technical ecosystems. Nat. Ecol. Evol. 2, 1735–1744 (2018). guidelines for genetic conservation of Norway spruce (Picea abies (L.). 57. Peñuelas, J. & Estiarte, M. Trends in plant carbon concentration and plant (International Plant Genetic Resources Institute, R., 2003). demand for N throughout this century. Oecologia 109, 69–73 (1997). 84. Wühlisch, G. v. EUFORGEN:Technical Guidelines for genetic conservation and 58. Deng, Q. et al. Down-regulation of tissue N:P ratios in terrestrial plants by use for European beech (Fagus sylvatica). (Bioversity International, Rome, elevated CO2. Ecology 96, 3354–3362 (2015). 2008). 59. Yue, K. et al. Effects of three global change drivers on terrestrial C:N:P 85. Peñuelas, J., Sardans, J., Ogaya, R. & Estiarte, M. Nutrient stoichiometric stoichiometry: a global synthesis. Glob. Chang. Biol. 23, 2450–2463 (2017). relations and biogeochemical niche in coexisting plant species: effect of 60. Carnicer, J. et al. Widespread crown condition decline, food web disruption, simulated climate change. Pol. J. Ecol. 56, 613–622 (2008). and amplified tree mortality with increased climate change-type drought. Proc. 86. Peñuelas, J. et al. Faster returns on ‘leaf economics’ and different Natl Acad. Sci. USA 108, 1474–1478 (2011). biogeochemical niche in invasive compared with native plant species. Glob. 61. He, M. & Dijkstra, F. A. Drought effect on plant nitrogen and phosphorus: a Chang. Biol. 16, 2171–2185 (2010). meta-analysis. N. Phytol. 204, 924–931 (2014). 87. Fernández-Martínez, M. et al. Global trends in carbon sinks and its 62. Avila, J. M., Gallardo, A., Ibáñez, B. & Gómez-Aparicio, L. Quercus suber relationship with CO2 and temperature. Nature Climate Change. 9, 73–79 dieback alters soil respiration and nutrient availability in Mediterranean (2019). forests. J. Ecol. 104, 1441–1452 (2016). 88. R Core Team. R: A language and environment for statistical computing. (2016). 63. Mecheri, H., Kouidri, M., Boukheroufa-Sakraoui, F. & Adamou, A.-E. 89. Barton, K. MuMIn: Multi-model inference. R package versión 1.9.0. Available Variation du taux d’infestation par Thaumetopoea pityocampa du pin d’Alep : at: http://CRAN.R-project.org/package=MuMIn. (2013). effet sur les paramètres dendrométriques dans les forêts de la région de Djelfa 90. Nakagawa, S. & Schielzeth, H. A general and simple method for obtaining R2 (Atlas saharien, Algérie) . Variation in Thaumetopoea pityocampa infestation from generalized linear mixed-effects models. Methods Ecol. Evol. 4, 133–142 rate of Aleppo pine: effect on dendrometric parameters in the Djelfa (2013). region forests (Saharan Atlas, Algeria). Comptes Rendus Biol. 341, 380–386 91. Hedges, L. V., Gurevitch, J. & Curtis, P. S. The meta-analysis of response ratios (2018). in experimental ecology. Ecology 80, 1150–1156 (1999). 64. Fleischmann, F., Göttlein, A., Rodenkirchen, H., Lütz, C. & Oßwald, W. Biomass, nutrient and pigment content of beech (Fagus sylvatica) saplings infected with Phytophthora citricola, P. cambivora, P. pseudosyringae and P. undulata. For. Pathol. 34, 79–92 (2004). Acknowledgements 65. Vitousek, P. M., Porder, S., Houlton, B. Z. & Chadwick, O. A. Terrestrial Funding for this research was provided by the Spanish Government grants CGL2016- phosphorus limitation: Mechanisms, implications, and nitrogen-phosphorus 79835-P and CGL2016‐78093‐R, the European Research Council Synergy grant ERC- interactions. Ecol. Appl. 20, 5–15 (2010). SyG-2013-610028 IMBALANCE-P and the Catalan Government grant SGR 2017-1005. 66. de Witte, L. C., Rosenstock, N. P., van der Linde, S. & Braun, S. Nitrogen M.F-M. is a postdoctoral fellow of the Research Foundation—Flanders. deposition changes ectomycorrhizal communities in Swiss beech forests. Sci. Total Environ. 605–606, 1083–1096 (2017). 67. Miller, D. E. & Watmough, S. A. Air pollution, climate, soil acidity, and indicators of forest health in Ontario’s sugar maple forests. Can. J. For. Res. 39, Author contributions 2065–2079 (2009). J.P. and J.S. designed this study. J.P. H.V., M.F.M., P.Z., J.M. and J.S. conducted the 68. Miller, D. E. & Watmough, S. A. Soil acidification and foliar nutrient status of analyses with the help of the rest of coauthors. J.P. and J.S. wrote the paper with Ontario’s deciduous forest in 1986 and 2005. Environ. Pollut. 157, 664–672 contributions from all the co-authors, M.f.M., H.V., J.M, P.Z., J.C., T.G.M.S., I.B.K., M.O., (2009). I.A.J. and P.C. 10 COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio COMMUNICATIONS BIOLOGY | https://doi.org/10.1038/s42003-020-0839-y ARTICLE Competing interests Open Access This article is licensed under a Creative Commons The authors declare no competing interests. Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Additional information Commons license, and indicate if changes were made. The images or other third party Supplementary information is available for this paper at https://doi.org/10.1038/s42003- material in this article are included in the article’s Creative Commons license, unless 020-0839-y. indicated otherwise in a credit line to the material. If material is not included in the Correspondence and requests for materials should be addressed to J.P. article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from Reprints and permission information is available at http://www.nature.com/reprints the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. © The Author(s) 2020 COMMUNICATIONS BIOLOGY | (2020)3:125 | https://doi.org/10.1038/s42003-020-0839-y | www.nature.com/commsbio 11