This is the accepted version of the article: Sardans i Galobart, Jordi; Urbina Barreto, Ifigenia; Grau Fernández, Oriol; [et al.]. «Long-term drought decreases ecosystem C and nutrient storage in a Mediterranean holm oak forest». Environmental and Experimental Botany, Vol. 177 (Sep. 2020), art. 104135. DOI 10.1016/j.envexpbot.2020.104135 This version is avaible at https://ddd.uab.cat/record/232202 under the terms of the license 1 Long-term drought decreases ecosystem C and nutrient storage in a 2 Mediterranean holm oak forest. 3 4 Jordi Sardans1,2^, Ifigenia Urbina1,2,*^, Oriol Grau1,2,3, Dolores Asensio1,2, Romà Ogaya1,2, Josep 5 Peñuelas1,2 6 1 7 CSIC, Global Ecology CREAF- CSIC-UAB, Cerdanyola del Valles, 08193 Barcelona, Catalonia, 8 Spain; 2 9 CREAF, Cerdanyola del Valles, 08193 Barcelona, Catalonia, Spain; 3 10 Cirad, UMR EcoFoG (AgroParisTech, CNRS, Inra, Univ Antilles, Univ Guyane), Campus 11 Agronomique, 97310 Kourou, French Guiana 12 *Corresponding author 13 ^Equally contributed 14 15 16 Running title: Nutrient stocks decrease under drought 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 1 32 Abstract 33 Aridity has increased in recent decades in the Mediterranean Basin and is projected to continue 34 to increase in the coming decades. We studied the consequences of drought on the 35 concentrations, stoichiometries and stocks of carbon (C), nitrogen (N), phosphorus (P) and 36 potassium (K) in leaves, foliar litter of a three dominant woody species and soil of a 37 Mediterranean montane holm oak forest where soil-water content was experimentally reduced 38 (15% lower than the control plots) for 15 years. Nitrogen stocks were lower in the drought plots 39 than in the control plots (8.81±1.01 kg ha-1 in the forest canopy and 856±120 kg ha-1 in the 0-15 40 cm soil layer), thus representing 7 and 18% lower N stocks in the canopy and soil respectively. 41 15N was consistently higher under drought conditions in all samples, indicating a general loss 42 of N. Foliar C and K stocks were also lower but to a lesser extent than N. Decreases in biomass 43 and C and N stocks due to drought were smallest for the most dominant tall shrub, Phillyrea 44 latifolia, so our results suggest a lower capacity of this forest to store C and nutrients but also 45 substantial resulting changes in forest structure with increasing drought. 46 Key words: Climate change, aridity, nitrogen, phosphorus, potassium, stoichiometry, carbon 47 stocks. 48 49 50 51 52 53 54 55 56 57 58 59 60 2 61 1.Introduction 62 Mediterranean climates occur in five occidental coastal regions of the continents between 63 latitudes 34 and 45° in both hemispheres and are characterized by moderately wet and cold 64 winters and dry and hot summers. The intensity of the drier periods, increasing from high to low 65 latitudes in these areas, can vary widely (Di Castri and Mooney, 1973; Di Castri, 1981). Woody 66 plants with conservative traits, such as slow growth and high sclerophylly associated with low 67 availabilities of water and nutrients, typically dominate the Mediterranean communities 68 (Valiente-Baunet et al., 1998; Ching et al., 2000; Edwards et al., 2000; Barchuk and Valiente- 69 Banuet, 2006; Sardans et al., 2006). Most studies suggest that the impacts of global change 70 promote degradation and desertification in current Mediterranean areas (Sardans and 71 Peñuelas, 2013; Salvati et al., 2014; Loizidou et al., 2016; Peñuelas et al., 2017a; 2018). 72 Precipitation has exhibited either a long-term downward trend, principally in the dry season 73 (Kutiel et al., 1996; Esteban-Parra et al., 1998), or absence of any significant change, even though 74 accompanied by an increase in the potential evapotranspiration that has led to increased aridity 75 (Piñol et al., 1998; Peñuelas et al., 2002). Furthermore, there has been a rise in climate extreme 76 events such as severe droughts and heat waves (Goldreich, 2003; Sánchez et al., 2004; 77 Hernández-Ceballos et al., 2016).The decline in total rainfall and/or availability of soil water 78 expected for the coming decades (IPCC, 2013) may be even more severe under warmer 79 conditions, with a CO2-rich atmosphere and a higher demand for water (Zhao and Running, 80 2010). Defoliation has substantially increased in the Mediterranean forests of southern Europe 81 (Carnicer et al., 2011; De la Cruz et al., 2014; De Marco et al., 2014). 82 Several specific structural and functional characteristics of Mediterranean plants are 83 associated with conservative mechanisms linked to the avoidance of water stress but also 84 frequently to the scarcity of soil nutrients (Bussotti et al., 2000; Verdú et al., 2003; Kooilman et 85 al., 2005; Galmés et al., 2007).. Nutrient supplies are often an important factor in the growth, 86 structure and distribution of these ecosystems (Specht, 1979; Henkin et al. ,1998; Sardans and 87 Peñuelas, 2013), and Mediterranean plants frequently change and adapt foliar chemical and 88 physical traits in response to both water deficits and nutrient stress (Gratani et al., 2003; Sardans 89 et al., 2006, 2008a,b,c). Drought conditions can both shorten the growing period in spring and 90 increase water stress during the summer dry season (Peñuelas et al., 2013). Mediterranean soils 91 are often deficient in nutrients (Specht, 1979; Kruger, 1979; Sardans and Peñuelas, 2013), which 92 can be exacerbated by climate change. Moreover, drought can decrease the capacity to retain 93 nutrients in Mediterranean forests (Kreuzwieser and Gessler, 2010; Peñuelas et al., 2018). 94 Drought may negatively interact with nutrient limitation by favoring soil degradation and 3 95 nutrient losses thus impairing ecosystem water-use efficiency (WUE), which could lead to a point 96 of no return toward desertification (García-Fayos and Bochet, 2009; Matías et al., 2011; Sardans 97 and Peñuelas, 2013; Peñuelas et al., 2013). 98 Long-term field experiments, though, have not provided consistent evidence that 99 moderate drought can also imply a change in nutrient concentrations, stoichiometries and 100 stocks. The recent implementation of mechanistic nitrogen (N) and phosphorus (P) cycling in 101 models of the terrestrial carbon (C) cycle underscores the importance of nutrient feedbacks, 102 indicating potential reductions in the productivity of terrestrial ecosystems (Goll et al., 2012; 103 Peñuelas et al., 2013; Fernández-Martínez et al., 2017; Peñuelas et al., 2017b; Sun et al., 2017; 104 Peñuelas et al., 2020). The recent increase in aridity at a global continental scale together with 105 nutrient imbalances have demonstrated decreases in the continental capacity to store C 106 (Peñuelas et al., 2017b). Drought linked to increased nutrient limitation can thus feedback on C 107 balances, because limited P and N availability are likely to jointly reduce future C storage by 108 natural ecosystems during this century (Peñuelas et al., 2013). This feedback can be especially 109 important in semiarid regions such as the Mediterranean Basin and the Sahel where water 110 limitations can have strong negative feedbacks with increased nutrient deficits and thus with a 111 strongly diminished productive capacity and ecosystem biomass (IPCC 2007, Sardans et al., 112 2012; Sanaullah et al., 2012; Sardans and Peñuelas, 2013; Gessler et al., 2016). 113 Long-term experimental climatic-manipulation studies are rare in the Mediterranean 114 biome. C, N, P and potassium (K) cycles clearly changed after five years of continuous drought 115 in a field experiment in an old holm oak forest (Quercus ilex L. ssp ilex) established in 1999. C, N 116 and P mineralization tended to be slower and N accumulation was lower in some dominant plant 117 species in response to five years of continuous drought (Sardans et al., 2008a,b; Sardans and 118 Peñuelas 2007, 2010). The results after the five years of continuous drought, however, were not 119 conclusive, because the stand biomass stocks only decreased in some (Quercus ilex and Arbutus 120 unedo L.) but not all species, whereas C and N accumulation increased in the soil (Sardans et al., 121 2008a,b; Sardans and Peñuelas, 2007, 2010). We now aimed to identify the mid- to long-term 122 links of ten more years of prolonged moderate drought with the concentrations, stocks and 123 stoichiometries of C, N, P and K in this Mediterranean holm oak forest. 124 The potential shifts in plant elemental concentrations and stoichiometries can provide 125 information of the capacity of different species to adapt or disappear due to some biotic and 126 abiotic stresses (Peñuelas et al., 2019). The biogeochemical niche (BN) hypothesis is an emerging 127 tool to test the shifts in plant communities along spatial and temporal scales as functions of the 128 main environmental abiotic and biotic conditions (Peñuelas et al., 2019). The BN hypothesis is 4 129 defined as the multidimensional space defined by the concentrations of bio-elements in 130 individuals of a given species (Peñuelas et al., 2008, 2019). The hypothesis is thus based on the 131 species-specific needs and uses of bio-elements in different amounts and proportions from the 132 organismic structure and function in abiotic and biotic environmental spaces. BN distances 133 among species should thus be a function of taxonomic difference, sympatric coexistence and 134 homeostasis/plasticity. The hypothesis thus also suggests stoichiometric flexibility (the capacity 135 to change the BN) of species, being an important trait, because it involves their capacity to cope 136 with changes in environmental conditions. Each species conserved its own foliar elemental 137 composition, consistent with the BN hypothesis, after five years of continuous drought, despite 138 tending toward changing its overall foliar elemental composition in response to drought 139 (Peñuelas et al., 2019). We now aimed to determine if more intense changes under long-term 140 (15 years) continuous drought could affect the BNs of the dominant species. 141 We thus studied the effects of 15 years of continuous drought on ecosystem 142 concentrations and stocks of C, N, P and K. We also assessed the isotopic signals of C and N cycles 143 in this long-term field climatic-manipulation experiment in the old Mediterranean holm oak 144 forest. We hypothesized a loss of C and nutrient stocks in foliar biomass. We also hypothesized 145 a loss of capacity to retain C in plant, soil and litter and a decrease in soil stocks of some bio- 146 elements due to the lower capacity of plants to retain and store them under drought. Reductions 147 in plant uptake and activities of soil enzymes should imply a decrease in the rates of cycling of 148 bio-elements in the plant-soil system (Sardans and Peñuelas, 2005) and thus increasing the 149 probability of leaching and erosion in this climatic zone characterized by a high recurrence of 150 intense torrential rain (Llasat et al., 2016; Ziv et al., 2016). Finally, we hypothesized a general 151 shift in elemental composition and stoichiometry of the dominant woody species, but with 152 maintenance of their BNs as expected by the BN hypothesis. We tested these hypotheses by 153 investigating the effects of 15 continuous years of drought on: (i) the foliar, litter and soil 154 concentrations of C, N, P and K and their pairwise ratios in the three dominant woody species, 155 (ii) the effects on C and N cycles by analyzing foliar, litter and soil 13C and 15N, (iii) the shifts in 156 the BNs of the three species due to drought and (iv) the shifts in the C, N, P and K stocks in 157 leaves, litter and soil. 158 159 2.Materials and Methods 160 2.1.Study site 5 161 The study was carried out on a south-facing slope (25%) in a natural holm oak forest in 162 the Prades Mountains in southern Catalonia (northeastern Iberian Peninsula, 41°13′N, 0°55′ E). 163 The soil is a stony Xerochrept on a bedrock of metamorphic sandstone ranging between 35 and 164 90 cm in depth. The average annual temperature is 12 °C, and the average annual rainfall is 658 165 mm. The summer drought is pronounced and usually lasts for three months. The vegetation 166 consists of a dense forest dominated by Quercus ilex L. with abundant Phillyrea latifolia L., 167 Arbutus unedo L. and other evergreen species well-adapted to drought conditions (Erica arborea 168 L., Juniperus oxycedrus L. and Cistus albidus L.) and the occasional occurrence of deciduous 169 species (Sorbus torminalis L. Crantz and Acer monspessulanum L.). 170 2.2.Experimental design 171 Eight 15 × 10 m plots were delimited at the same altitude (950m) along the slope (25º) 172 (Ogaya et al. 2020). Half of the plots received the drought treatment, and the other half were 173 designated as control plots. The drought treatment consisted of the partial exclusion of rain by 174 suspending PVC strips and funnels 0.5–0.8 m above the soil surface. The strips and funnels 175 covered approximately 30% of the total plot surface. A ditch ca. 0.8 m deep was excavated along 176 the entire top edge of the treatment plots to intercept runoff, and all water intercepted by the 177 strips, funnels and ditches was conducted below the bottom edge of the plots. This drought 178 treatment has been applied since 1999. 179 2.3.Sampling 180 We sampled biomass (including leaves and foliar litter) and soil at the beginning of the 181 experiment (December 1999) and after six years of applying the drought treatment (January 182 2005) to estimate the total contents (Sardans et al., 2008a,b). Tereafter, in the present study, 183 we sampled leaves, foliar litter and soil in 2014 in mid-May (spring), the end of July (summer) 184 and mid-November (autumn) to determinate the changes in C, N, P and K stocks after 15 years 185 of constant chronic drought. Litterfall has been collected every three months during the two 186 pass campaigns (1999 and 2004) and the present campaign (2014) in 20 circular baskets (27 cm 187 in diameter with 1.5-mm mesh diameter) randomly distributed on the ground in each one of the 188 eight plots. Total litterfall was estimated by the proportion of the surface area of the plots 189 covered by the collection baskets. We have measured the sample biomass since the beginning 190 of the experiment several times, the methodology is described in (Ogaya and Peñuelas, 2007). 191 Briefly, the stem circumference was measured annually since 1999 each winter throughout the 192 study. The measurements were conducted with a metric tape at 50 cm height in all living stems 193 of all the species with a diameter larger than 2 cm at 50 cm height. Aboveground biomass in the 6 194 plots was estimated by allometric relationships between tree aboveground biomass (AB) and 195 the stem diameter at 50 cm height (D50). Outside the plots, 12 Q. ilex and 13 P. latifolia trees 196 were harvested, their perimeter at 50 cm height was measured and all their aboveground 197 biomass was weighed after drying them in an oven to reach constant weight. To estimate the 198 aboveground biomass of Q. ilex and P. latifolia we used the calculated allometric relationships 199 (lnAB = 4.9 + 2.3 lnD50; R2 = 0.98; n = 12: P < 0.001 for Q. ilex, and lnAB = 4.3 + 2.5 lnD50; R2 = 200 0.97; n = 13; P < 0.001 for P. latifolia), and to estimate the aboveground biomass of A. unedo we 201 used the allometric relationship calculated few years ago on the same area by Lledo´ (1990) 202 (lnAB = 3.8 + 2.6 lnD50; R2 = 0.99; n = 10; P < 0.001) (Ogaya & Peñuelas, 2007). The foliar 203 biomasses of the three studied species in 1999, 2004 and 2014 in control and drought plots are 204 shown in the Table S1. 205 We sampled leaves of the three dominant species of this ecosystem: Quercus ilex, 206 Phillyrea latifolia and Arbutus unedo (that represents on average the 97% of total aboveground 207 biomass) (Ogaya and Peñuelas, 2007). The sample collection was always standardized to avoid 208 biased results due to different tissue ages and orientation to sunlight, the sampled leaves were 209 sun-lit leaves oriented toward the south. The mean life span of the leaves of these species is 210 approximately 18 months (Ogaya and Peñuelas, 2006a), so for example most sampled foliar 211 biomass during 2014 corresponded to the leaves of the 2013 cohort. We randomly sampled five 212 trees of each species on each seasonal sampling day in each plot. Leaves were sampled at a 213 height of 2-3 m. We only sampled trees and shrubs with diameters between 2 and 12 cm (at 5 214 cm), which represented the majority (approximately 70%) of the community biomass (Ogaya 215 and Peñuelas, 2003, 2006b). We randomly collected five soil cores to a depth of 15 cm of horizon 216 A in each plot on each seasonal sampling day. 217 218 2.4.Chemical and physical analyses 219 220 All samples were taken to the laboratory and stored at 4 °C until analysis. The soil 221 samples were sieved, and the 2-mm fraction was analyzed. The leaves and foliar litter were 222 washed with distilled water as described by Porter (1986) for analyzing C, N, P and K. The washed 223 samples were dried in an oven at 60 °C to a constant weight and then ground in a CYCLOTEC 224 1093 (Foss Tecator, Höganäs, Sweden) for analyzing the biomasses or in a FRITSCH Pulverisette 225 (Rudolstadt, Germany) for analyzing the soils. The concentrations of C and N in the foliar, foliar 226 litter and soil samples were quantified by placing 1.4 mg of powder of each sample in a tin 227 microcapsule and using combustion coupled to gas chromatography in a CHNS-O Elemental 7 228 Analyser (EuroVector, Milan, Italy). The concentrations of P and K in the samples were 229 determined by first grinding the samples and then digesting them in acid in a MARSXpress 230 microwave (CEM, Mattheus, USA) at high pressure and temperature (see Peñuelas et al., 2010). 231 A total of 250 mg of leaf powder, 5 mL of nitric acid and 2 mL of H2O2 were placed into Teflon 232 tubes. The products of the digestions were placed in 50-mL flasks and diluted with Milli-Q water 233 to a volume of 50 mL. The concentrations of P and K were determined using Optic Emission 234 Spectrometry with Inductively Coupled Plasma. The accuracy of the digestions and the analytical 235 biomass procedures were assessed using a certified standard biomass (NIST 1573a, tomato leaf; 236 NIST, Gaitherburg, MD) as a reference. 237 The isotopic analyses of 13C and 15N in the foliar and soil samples (sampled in 2014) 238 were conducted in the Stable Isotope Facility at the University of California, Davis using using a 239 PDZ Europa ANCA-GSL elemental analyzer connected to a PDZ Europa 20-20 isotope ratio mass 240 spectrometer (Sercon Ltd., Cheshire, UK). 241 Soil moisture has been measured each month throughout the experiment by time 242 domain reflectometry (Tektronix 1502C, Tektronix, Beaverton, OR, USA) connecting the time 243 domain reflectometer to the ends of three stainless steel cylindrical rods, 25 cm long, fully driven 244 into the soil (Zegelin et al. , 1989). Four sites per plot were randomly selected to install the steel 245 cylindrical rods for soil moisture measurements (Ogaya et al., 2014). During these 15 years of 246 drought treatment soil moisture of drought plots compared to control plots (P < 0.01), but this 247 reduction was larger during rainy seasons and lower during dry seasons (Ogaya et al., 2014). 248 2.5.Statistical analyses 249 Significant differences in foliar, foliar litter and soil C, N, P and K concentrations and 250 ratios, 13C, 15N and C, N, P and K stocks in 2014 among treatments and seasons were tested 251 by general mixed models, with treatment and season as fixed factors and plot as a random 252 factor. We also used mixed models to analyze the treatment effects within each season, with 253 treatment as a fixed factor and plot as a random factor. We used the “lme” function of the 254 “nlme” R package (Pinheiro et al., 2016). Non-normally distributed variables were normalized 255 by log-transformation. We chose the best model for each dependent variable using the Akaike 256 information criterion (AIC). We used the MuMIn (Barton, 2012) R package in the mixed models 257 to estimate the percentage of the variance explained by the model. We conducted Tukey’s post 258 hoc tests to detect significant differences in the analyses for more than two communities using 259 the “glht” function of the “multcomp” (Hothorn et al., 2013) R package. 8 260 We also performed multivariate statistical analyses. We used a general discriminant 261 analysis (GDA) to determine the overall differences of the foliar variables: C, N, P and K 262 concentrations and their pairwise ratios among the three dominant species sampled at different 263 times during the experiment: Desember 1999 (before the start of the experiment), November 264 2005 (after six years of continuous drought) and November 2014 (after 15 years of continuous 265 drought). GDA uses squared Mahalanobis distances that depend on the Euclidean distance in 266 the model between two sets of samples. The squared Mahalanobis distance is low when the 267 samples are similar, indicating little difference between the two sets. In contrast, the distance is 268 higher and the difference between the sets increases the more the samples differ (De 269 Maesschalk et al., 2000). We conducted one-way ANOVAs with Bonferroni post hoc tests of the 270 scores of the first two GDA roots to identify differences among the dominant species sampled 271 over time. We used Statistica 8.0 (StatSoft, Inc. Tulsa, USA) for these analyses. 272 273 3.Results 274 3.1.Foliar, litter and soil C, N, P and K concentrations and ratios 275 The effects of the drought treatment on foliar and foliar litter C, N, P and K 276 concentrations depended strongly on season and species (Tables S2 and S3). A decrease in foliar 277 K concentrations in summer in the three species and a decrease in foliar P concentrations in P. 278 latifolia and A. unedo (Table S2) were the most general effects of the treatment. 279 The drought treatment increased foliar C and N concentrations and decreased foliar K 280 concentrations in summer and increased foliar N concentrations in autumn in Q. ilex, the 281 dominant species (Table S2). The treatment also decreased foliar C:N ratios in summer and 282 autumn in Q. ilex. Foliar N and P concentrations were higher and the N:P ratio was lower in 283 spring, and the foliar N and P concentrations were lowest in summer, in all species in both the 284 drought and control plots. Drought was associated with decreases in C and K concentrations and 285 increases in N and P concentrations in the Q. ilex foliar litter in summer (Table S3). 286 Phillyrea latifolia foliar P and K concentrations decreased and foliar C:P, C:K, N:P and N:K 287 ratios increased under drought in summer (Table S2). Foliar C, N and P concentrations were 288 lowest in summer. Drought was associated with a decrease in K concentrations and increases in 289 C:K, N:K and P:K ratios in the foliar litter in summer (Table S3). Arbutus unedo foliar N and K 290 concentrations and the foliar N:P ratio decreased under drought in summer (Table S2). Foliar C, 291 N and P concentrations were lowest in summer for both A. unedo and P. latifolia. Drought was 9 292 associated with decreases in C concentrations and C:N and C:P ratios and an increase in N 293 concentrations of the foliar litter in spring (Table S3). 294 Drought was associated with lower total soil C and N concentrations, especially in spring 295 and summer (Table S3). Total soil C:N and C:P ratios were lower in the drought plots, particularly 296 in summer and autumn, whereas total soil C:K, N:K and P:K ratios were also lower in the drought 297 plots, particularly in spring and summer. 298 3.2.13C and 15N 299 Drought had moderate effects on 13C of leaves and soil. 13C was lower in Q. ilex leaves 300 under drought, mainly in summer, whereas soil 13C was higher under drought in summer (Table 301 S4). Drought was associated with higher 15N in the leaves and foliar litter of Q. ilex, P. latifolia 302 (not in autumn) and A. unedo and in the soil in all seasons (Figures 1 and 2, Table S4). 303 3.3.Overall temporal shifts in foliar composition in the species under drought 304 The GDA hypervolume shifted in the sample sets of the three species sampled at 305 different times. All elemental concentrations and their pair-wise ratios contributes significantly 306 in the model (Table S4). The controls occupied a hypervolume distinct from that of the samples 307 of the three species sampled in 2005 and in P. latifolia sampled in 2014 under drought (Table 308 S5, Figure 3). This shift between the control and drought samples in A. unedo and Q. ilex was 309 along Root 1, with a displacement of the drought plots toward lower N concentrations and 310 N:Mg, N:P and N:Ca ratios and higher Mg concentrations and C:N ratios. The multivariate 311 analyses indicated that the effect of drought on foliar elemental concentrations was not 312 homogeneous. The squared Mahalanobis distances differed significantly between the species, 313 except for P. latifolia and Q. ilex between the control and drought plots sampled in 1999 (Table 314 S5). 315 3.4.Foliar, litter and soil C, N, P and K stocks during the 15 years of drought 316 Drought decreased the accumulation of foliar C stocks in Q. ilex (significantly only for 317 1999-2004) (Figure 4A), A. unedo (Figure 4B) and the sum of all three species (Figure 4d), mainly 318 associated with the smaller increase in foliar biomass under drought. Drought decreased the 319 accumulation of foliar N stocks in A. unedo and the sum of all three species (significantly only 320 for 1999-2014) (Figure 5), mainly associated with the lower N concentrations and smaller 321 increase in foliar biomass under drought. Drought did not affect the accumulation of foliar P 322 stocks relative to the control plots in the three species and the study periods (Figure 6). Drought 323 decreased the accumulation of foliar K stocks in A. unedo (significantly only for 1999-2014) 10 324 (Figure 7). The overall results indicated a significant decrease in foliar C, N and K stocks in A. 325 unedo and in C and N stocks in the three dominant species (representing >90% of the total 326 ecosystem aboveground biomass) after 15 years of moderate drought (a mean reduction in soil- 327 water content of 15%). C stocks decreased significantly in the 0-15 cm soil profile for 2004-2014 328 in the drought but not the control plots (Figure 8), but not significantly (P=0.87). The difference 329 in total soil N stock between the drought and control plots for 2005-2014 in the upper 15 cm of 330 soil was 856 ± 120 kg ha-1. Total soil P stocks in the upper 15 cm did not change significantly in 331 either the control or drought plots. Total soil K stocks for 2005-2014 generally decreased in the 332 upper 15 cm (3476 ± 444 and 3541 ± 200 kg ha-1 in the control and drought plots, respectively), 333 but not significantly. C, N, P and K stocks of leaf-litter did not change in 2015 relative to those 334 in 1999 (Table S6). 335 336 4.Discussion 337 This study demonstrated that a moderate increase in drought, in this case a mean 338 reduction in soil-water content of 15%, for 15 years strongly affected C and some nutrient stocks 339 at the ecosystem level. Drought decreased the stocks of N, the most limiting nutrient, in the 340 entire canopy and the upper soil layer. However, the drought effects on C and nutrient stocks 341 varied depending on the species. This asymmetrical effect of drought on species stocks agrees 342 with changes in other variables such as growth and mortality and thus species biomass: the 343 dominant evergreen shrub P. latifolia maintained its foliar C and nutrient stocks, whereas C and 344 nutrient stocks in the dominant evergreen tree Q. ilex and mainly the shrub A. unedo decreased. 345 These results indicate that a moderate increase in drought in the coming decades would 346 accelerate shifts in species composition and dominance, from a forest to a shrubland ecosystem 347 dominated by evergreen shrubs (Peñuelas et al., 2018). 348 4.1.Foliar C, N, P and K concentrations after 15 years of drought 349 The foliar elemental concentrations and ratios had a strong seasonal component 350 independent of the treatments, e.g. the concentrations of N, P and K were higher in spring, 351 coinciding with the period of growth in this Mediterranean ecosystem. The foliar N:P ratio in the 352 three species was noticeably lower in spring, thus coinciding with the higher foliar N and P 353 concentrations in the growing season, consistent with the growth-rate hypothesis, which states 354 that the period of growth coincides with a higher capacity for fast protein synthesis (Sterner and 355 Elser, 2002). In fact, this link between higher metabolic capacity associated with higher foliar N 356 and P concentrations and a lower N:P ratio was experimentally demonstrated in a previous study 11 357 (Rivas-Ubach et al., 2012), indicating that the plasticity of this phenological component was not 358 affected by the drought treatment. The maximum productive capacity, from a nutritional 359 perspective, thus coincides with the optimum annual climatic conditions in both the control and 360 drought plots. 361 Foliar concentrations tended to decrease in the drought plots or did not change relative 362 to the control plots. The differences between the foliar concentrations in the control and 363 drought plots were largest in summer, thus coinciding with the most arid conditions with water 364 deficits and stress. The effects of drought on the foliar concentrations also depended on the 365 species and element, similar to the elemental stocks. Global changes in foliar concentrations 366 associated with drought, however, were rare. Differences between the control and drought 367 treatments generally occurred in summer; leaves in the drought plots had lower concentrations 368 of C and N in Q. ilex, P and K in P. latifolia and N and K in A. unedo. Studies of gradients and/or 369 field experiments frequently find that N foliar concentrations in Mediterranean forest species 370 are very homeostatic in response to abiotic shifts (Martin-StPaul et al., 2013). From an ecological 371 stoichiometric perspective (Sterner and Elser, 2002), elemental composition is more 372 homeostatic in stress-tolerant plant species than competitor and ruderal species in response to 373 environmental changes (Peñuelas et al., 2019). Our results showing only small changes in foliar 374 C, N, P and K concentrations support the idea of the stress tolerance strategy in Mediterranean 375 plants. Mediterranean forest tree species such as Quercus ilex or Pinus halepensis have less 376 variability in their elemental foliar composition than wet temperate species such as Quercus 377 robur or Quercus petraea adapted to wetter climate and nutrient richer soils (Sardans et al. 378 2020). We had previously observed in 2011 that Q. ilex foliar K concentrations were higher in 379 the drought than the control plots (Sardans et al., 2013; Rivas-Ubach et al., 2014), whereas Q. 380 ilex leaves analyzed in 2014 had the opposite pattern, specifically in summer, perhaps due to 381 the loss of K in the upper 15 cm of soil in both the control and drought plots from 2005 to the 382 present. Some studies have reported that some elemental concentrations can increase and 383 others decrease, depending on the site and species (Luo et al., 2018; Urbina et al., 2015). We 384 thus cannot draw a clear conclusion about the impact of experimental drought on distinct foliar 385 elemental compositions in Mediterranean environments under field conditions. 386 The level of resistance to drought stress in Mediterranean plant species is highly 387 variable, so fitness decreases more in some species than others (Sardans and Peñuelas, 2013). 388 For example, foliar N concentrations were lower in summer and litter N concentrations were 389 higher in spring in A. unedo in the drought than the control plots. These results are consistent 390 with previous results that drought affects fitness and all the variables involved in growth (Ogaya 12 391 and Peñuelas, 2007; Sardans et al., 2008a,b; Barbeta et al., 2013), photosynthetic capacity 392 (Ogaya and Peñuelas, 2006) and flowering (Ogaya and Peñuelas, 2004; Barbeta et al., 2013) that 393 we measured more in A. unedo than the other dominant species. Some studies, however, have 394 found that P. latifolia was the most stress-tolerant species of our three dominant species, where 395 growth and mortality (Ogaya and Peñuelas, 2007; Barbeta et al., 2013) or ecophysiology (Ogaya 396 et al., 2011) were little or not affected. Our results are consistent with these studies, because 397 bio-elemental concentrations and foliar stocks were less affected by drought in P. latifolia than 398 the other species Our results are consistent with these studies, because bio-elemental 399 concentrations, foliar stocks and growth were less affected by drought in P. latifolia than the 400 other species, specially than A. unedo (Ogaya and Peñuelas, 2007; Barbeta et al., 2013;2015). 401 Our results therefore reinforce previous findings in this same study and in other studies, with 402 evidence of species specific responses to drought in Mediterranean ecosystems, which may 403 eventually lead to a partial community shift favoring more drought-resistant species (Sardans 404 and Peñuelas, 2013; Peñuelas et al., 2013; 2017a; 2018; Gazol et al., 2017).The adaptation of 405 Mediterranean plants to drought generally frequently varies with species; some species cope 406 with moderate drought without significant losses in fitness, but others are strongly affected 407 (Sardans and Peñuelas, 2013). Our experiment demonstrated that the three dominant species 408 responded to this persistent moderate drought, but to different extents. As observed in previous 409 studies in this research site, the studied species, despite at different extend, all they the plants 410 subjected to the long-term experimental drought shifted water uptake toward deeper (10–35 411 cm) soil layers during the wet season and reduced groundwater uptake in summer, indicating 412 plasticity in their functional distribution of fine roots that dampened the effect of our 413 experimental drought in the long term (Barbeta el at., 2015). 414 The reduction of foliage biomass under drought relative to the control plots, and thus 415 plant cover, may decrease soil quality, as in several Mediterranean areas (Gallardo et al., 2000; 416 Garcia et al., 2002), decreasing the capacity of soil to retain moisture and thus fertility (Sadaka 417 and Ponge 2003; Ruiz-Sinoga et al., 2011). In fact, reductions in canopy cover and N 418 concentrations have been reported in previous studies in various Mediterranean communities, 419 including forests (Goldstein et al., 2000; Inclan et al., 2005; Pérez-Ramos et al., 2010; Cotrufo et 420 al., 2011), consistent with the decrease in soil N stocks in the drought versus the control plots 421 for 2005-2014. 422 4.2.13C and 15N in the soil, foliar litter and leaves after 15 years of drought 423 13C has frequently been used as an alternative measurable variable to WUE, especially 424 in drought conditions (Jucker et al., 2017). Plants with limited capacity for water uptake 13 425 preferentially have higher 13C (Jucker et al., 2017). Soil 13C in our study was higher in the 426 drought than the control plots, potentially associated with a high rate of loss of C from the soil 427 and thus with a more open soil C cycle. In contrast, 13C in the summer leaves of the dominant 428 species Q. ilex was lower in the drought than the control plots, indicating better conditions for 429 fixing C in the drought plots. This apparently contradictory results are probably linked to the 430 large loss of biomass in the Q. ilex canopy during the period 1999-2014 in the drought plots (- 431 5000 ± 1600 kg ha-1), which did not occur in the control plots, and thus the drop in biomass due 432 to 15 consecutive years of chronic drought has decreased the competition for available water 433 counteracting the decrease in its availability. Less canopy biomass in Q. ilex, nearly 7% in the 434 drought plots, together with the general decrease in ecosystem aboveground biomass in the 435 drought than the control plots for this period (Barbeta et al., 2013), can improve WUE in the 436 remaining canopy leaves, thus decreasing 13C. Soil 13C was higher in the drought plots. Soil C 437 concentrations are negative correlated with 13C, such as demonstrated in several studies 438 (Farquar et al., 1989; Balesdent et al., 1993; Garten et al., 2000), indicating a decay of soil C cycle 439 rates, thus consistent with the slower soil enzyme activity observed in previous studies in 440 drought plots (Sardans and Peñuelas, 2005; Sardans et al., 2008a,b), and thus with slower 441 mineralization rates under drought conditions. 442 15N in all foliar, foliar-litter and soil samples was higher in the drought than the control 443 plots, highlighting the increase in N-cycle openness in the drought relative to the control plots, 444 consistent with the decreases in the accumulation of N stocks for 1999-2014 in leaves and for 445 2005-2014 in soil. Drought has thus been correlated with a loss of N in these ecosystems, which 446 is especially important due to the limiting role of N in this mountainous area (Sabaté and Gracia, 447 1994). The decreases in plant N uptake and activities of soil enzymes involved in N mineralization 448 reported in previous studies (Sardans et al., 2008a,b) could account for this reduction in N stocks 449 under drought. Lower rates of the plant-soil N cycle, with lower proportional stocks of N stored 450 in plants relative to the soil, can increase the loss of soil N associated with the large capacity for 451 leaching and erosion due to the frequent and intense torrential rains typical of Mediterranean 452 areas. Drier soil conditions can favor desertification by several processes affecting the sequential 453 loss of plant cover (Sardans and Peñuelas, 2013). 454 455 4.3. BN shifts along 15 year of drought 456 The GDA suggested the existence of species-specific BNs, consistent with the BN 457 hypothesis (Peñuelas et al., 2009, 2019); each species had a unique foliar elemental composition 14 458 but also an overall foliar elemental plasticity to shifts, depending on the season and 459 environmental circumstances but within its species-specific space in the hypervolume. The 460 spaces of multivariant analyses also shifted over time in response to drought (in some cases) 461 within the BN space of each species. Phillyrea latifolia and A. unedo have evolved over time 462 toward higher foliar K concentrations, which should be correlated with the decrease in soil K 463 concentrations and stocks during the study period as a mechanism to conserve K as much as 464 possible. The large differences in K concentration between the leaves and foliar litter, with 8.5, 465 7.3 and 4.0 (mg g-1 DW) less K in the litter than the leaves in Q. ilex, P. latifolia and A. unedo, 466 respectively, suggests a large effort to resorb K in this ecosystem (in both the control and 467 drought plots). Our data thus generally indicated that at least P. latifolia and A. unedo have 468 increased their foliar K concentrations, whereas the upper soil layer has lost K. 469 470 4.4.Changes in C, N, P and K stocks in leaves, foliar litter and soil 471 Drought decreased total soil N stocks in the upper 15 cm. Interestingly, C and K were 472 notably lost in both the control and drought plots for 1999-2015. Previous studies have reported 473 that past recent extreme droughts in the study area (such as in summer 2011) could have 474 decreased the reserves of ground water and decreased weathering from rocks and thus the 475 uptake by plants of ground water and minerals (Barbeta et al., 2015) such as K, which may have 476 contributed to the decrease in K concentrations in the upper soil layer and the leaves. We thus 477 observed how a long-term moderate reduction in the availability of soil water (averaging 15%) 478 could be sufficient to cause a general and continuous loss of nutrient stocks in a Mediterranean 479 forest. The effects of droughts are element-dependent and affect nutrient stoichiometry at the 480 ecosystem level, with the potential to generate nutrient imbalances. The most abundant shrub, 481 P. latifolia was able to conserve more the C and nutrient stocks than the dominant tree species 482 Q. ilex. Thus these new results are consistent with those of previous studies in this long-term 483 experiment of chronic drought in field conditions observing an slow but continuous increase in 484 dominance of the most drought-tolerant species, the shrub P. latifolia (Ogaya and Peñuelas, 485 2006,2007; Barbeta et al., 2015,2017; Ogaya et al., 2020). 486 487 Final remarks and conclusions 488 (i) The variation in foliar elemental composition was strongly seasonal but notably homeostatic 489 in response to the drought treatment. This result is consistent with the notable homeostatic 490 elemental composition expected for stress tolerant species biogeochemical niche (Peñuelas et 15 491 al., 2019). Moreover, N, P and K concentrations were higher and the N:P ratio was lower in 492 spring, coinciding with the period of growth in this Mediterranean ecosystem, consistent with 493 the growth-rate hypothesis. 494 (ii) The foliar concentrations of C and N in Q. ilex, P and K in P. latifolia and N and K in A. unedo 495 in summer were lower in the drought plots .The results thus showed that drier conditions have 496 a detrimental effect on nutrient capture mainly during the already dry periods in this 497 Mediterranean system. The largest differences in foliar concentrations between the control and 498 drought plots were found in summer, thus coinciding with the most arid annual conditions. 499 (iii) Soil 13C was higher in the drought plots, indicating a decay of soil C cycle and consistent 500 with lower soil C and could be related with slow soil enzyme activity observed in previous studies 501 and thus with slower mineralization rates. 15N was higher in the drought than the control plots 502 in all samples, highlighting the increase in openness of the N cycle in the drought relative to the 503 control plots. This difference is especially important given the limiting role of N in this 504 mountainous area in previous studies. It indicates a decrease in the capacity of the global 505 ecosystem to store C and an increase in the soil N losses associated with leaching and erosion 506 due to the frequent and intense torrential rains typical of Mediterranean areas.13C in the 507 summer leaves of the dominant Q. ilex, however, was lower in the drought than the control 508 plots, indicating better conditions for fixing C in the drought plots, probably due to the large loss 509 of aboveground biomass in the drought plots, which could improve WUE in the remaining leaves 510 of the canopy. 511 (iv) Each species had its own foliar elemental composition, consistent with the BN hypothesis. 512 The foliar elemental composition, however, shifted in response to drought in the BN space of 513 each species. Phillyrea latifolia and A. unedo have evolved over time toward higher foliar K 514 concentrations associated with the decrease in soil K concentrations and stocks during the study 515 period. The species were thus able to vary over time due to environmental changes but 516 maintaining their own “multi-variant space”. 517 (v) A long-term moderate reduction in the availability of soil water (averaging 15%) can cause a 518 general and continuous loss of nutrient stocks in this Mediterranean forest. This reduction 519 would have a negative synergistic effect of less water together with lower nutrient availability 520 on the capacity of the forest to resist these conditions. A gradual substitution of forest by 521 shrubland is thus expected. The most drought-tolerant species in this study, the shrub P. 522 latifolia, which retained more C and nutrients in its foliar stocks under drought, is slowly but 523 continuously increasing its dominance over time. 16 524 Acknowledgements 525 The authors would like to acknowledge the financial support from the European Research 526 Council Synergy grant ERC-SyG2013-610028 IMBALANCE-P, the Spanish Government grant 527 CGL2016-79835-P and the Catalan Government grant SGR 2017-1005. 528 References 529 530 Barbeta, A., Ogaya, R., Peñuelas, J., 2013. Dampening effects of long-term experimental drought 531 on growth and mortality rates of a holm oak forest. Global Change Biology, 19, 3133-3144. 532 Barbeta, A., Mejía-Chang, M., Ogaya, R., Voltas, J., Dawson, T.E., Peñuelas, J. 2015. The combined 533 effects of a long-term experimental drought and an extreme drought on the use of plant- 534 water sources in a Mediterranean forest. Glob. Change Biol. 21, 1213-1225. 535 Barchuk, A.H., Valiente-Banuet, A., 2006. Comparative analysis of leaf angle and sclerophylly of 536 Aspidosperma quebracho-blanco on a water déficit gradient. Austral Ecol. 3, 882-891 537 Barton, K., 2012. MuM. In: Multi-model inference. R package version 1.7.2. http://CRAN. R-283 538 project.org/package=MuMIn. at http://cran.r-project.org/ package=MuMIn> 539 Balesdent, .J, Girardin, C., Mariotti A., 1993. Site-related 13C of tree leaves and soil organic matter 540 in a temperate forest. Ecology 74, 1713–1721. 541 Bussotti, F., Borghini, F., Celesti, C., Leonzio, C., Bruschi, P., 2000. Leaf morphology and 542 macronutrients in broadleaved trees in central Italy. Trees, 14, 361–368. 543 Carnicer, J., Coll, M., Ninyerola, M., Pons, X., Sanchez, G., Peñuelas, J., 2011. Widespread crown 544 condition decline, food web disruption, and amplified tree mortality with increased climate 545 change-type drought. Proc. Natl. Acad. of Sci. 108, 1474-1478. 546 Ching, L.J., Shi, G.Y., Romane, F., Li, J.Q., Guo, Y.S., 2000. Environmental heterogeneity and 547 population variability of sclerophylly oaks (Quercus Sec. Suber) in East Himalayan 548 region. For. Stu China 2, 1–15. 549 Cotrufo, M.F., Alberti, G., Inglima, I., Marjanovic, H., LeCain, D., Zaldei, A., Peressotti, A., 550 Miglietta, F., 2011. Decreased summer drought affects plant productivity and soil carbon 551 dynamics in a Mediterranean woodland. Biogeosciences 8, 2729–2739. 552 De la Cruz, A.C., Gil, P.M., Fernandez-Cancio, A., Minaya, M., Navarro-Cerrillo, R.M., Sánchez- 553 Salguero, R., 2014. Defoliation triggered by climate induced effects in Spanish ICP forest 554 monitoring plots. For. Ecol. Manag. 331, 245-255. 555 De Marco, A., Proietti, C., Cionni, I., Fischer, R., Screpanti, A., Vitale, M. 2014. Future impacts of 556 nitrogen deposition and climate change scenarios on forest Crown defoliation. Environ. 557 Pollut. 194, 171-180 558 De Maesschalck, R., Jouan-Rimbaud, D., Massart, D.L., 2000. The Mahalanobis distance. 559 Chemometrics Intelligent Laboratory Systems 50, 1–18. 560 Edwards, C., Read, J., Sanson, G., 2000. Characterising sclerophylly: Some mechanical properties 561 of leaves from heath and forest. Oecologia 123, 158ecolo 562 Esteban-Parra, M.J., Rodrigo, F.S., Castro-Diez, Y., 1998. Spatial and temporal patterns of 563 precipitation in Spain for the period 1880– 1992. Inter. J. Climat. 18, 1557–1574. 564 Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989. Carbonisotope discriminationand 565 photosynthesis. An. Rev. Plant Biol. 40, 503–537. 566 Fernández-Martínez, M., Vicca, S., Janssens, I.A., Espelta, J.M., Peñuelas J., 2017. The role of 567 nutrients, productivity and climate in determining tree fruit production in European 568 forests. New Phytol. 213, 669-679. 569 Gallardo, A., Rodríguez-Saucedo, J.J., Covelo, F., Fernández-Alès, R., 2000. Soil nitrogen 570 heterogeneity in a Dehesa ecosystem. Plant Soil, 222, 71–82. 571 Galmés, J., Flexas, J., Savé, R., Medrano, H., 2007. Water relations and stomatal characteristics 572 of Mediterranean plants with different growth forms and leaf habits: responses to water 573 stress and recovery. Plant Soil, 290, 139-155. 17 574 Garcia, C., Hernandez, T., Roldan, A., Martin, A., 2002. Effect of plant cover decline on chemical 575 and microbiological parameters under Mediterranean climate. Soil Biol. Biochem. 34, 635– 576 642. 577 García-Fayos, P., Bochet, E., 2009. Indication of antagonistic interaction between climate change 578 and erosiono n plant species richness and soil properties in semiarid Mediterranean 579 ecosystems. Glob. Change Biol. 15, 306-318. 580 Gazol, A., Sangüesa-Barreda, G., Granda, E., Camarero, J.J., 2017. Tracking the impact of drougth 581 on functionality different Woody plants in a Mediterranean scrubland ecosystem. Plant 582 Ecol. 218, 1009-1020. 583 Garten, C.T., Cooper L.W., Post III, W.M., Hanson, P.J., 2000. Climate controls on forest soil C 584 isotope ratios in the southern Appalachian mountains. Ecology 81, 1108–1119. 585 Gessler, A., Schaub, M., McDowell, N.G., 2016. The role of nutrients in drought-induced tree 586 mortality and recovery. New Phytol. 214, 513-520. 587 Goldreich, Y., 2003. The climate of Israel. Observation, research and application. New York, NY, 588 USA: Kluwer Academic/Plenum Pub. 589 Goldstein, A.H., Hultman, N.E., Fracheboud, J.M., Bauer, M.R., Panek, J.A., Xu, M., Qi Y., 590 Guenther, A.B., Baugh, W., 2000. Effects of climate variability on the carbon dioxide, water, 591 and sensible heat fluxes above a ponderosa pine plantation in the Sierra Nevada (CA). Agric. 592 For. Metreol. 101, 113–129 593 Goll, D.S., Beovkin, V., Parida, B.R., Reick, C.H., Kattge, J., Reich, P.B., Van Bodegom, P., 594 Niinemetts, Ü., 2012. Nutrient limitation reduces land carbon uptake in simulations with a 595 model of combined carbon, nitrogen and phosphorus cycling. Biogeosciences 9, 3547–3569 596 Gratani, L., Meneghini, M., Pesoli, P., Cresence, M.F., 2003. Structural and functional plasticity 597 of Q. ilex seedlings of different provinances in Italy. Trees 17, 515–521. 598 Henkin, Z., Seligman, N.G., Kafkafi, U, Noy-Meir, I., 1998. Effective growing days: a simple model 599 of response of herbaceous plant growth in a Mediterranean ecosystem to variation in 600 rainfall and phosphorus predictive availability. Journal of Ecology 86, 137–148. 601 Hernández-Ceballos, M.A., Brattich, E., Cinelli, G., 2016. Heat-wave events in Spain: air mass 602 analysis and impacts on 7Be concentrations. Adv. Meterol. 8026018 603 Hothorn, T., Bretz, F., Wesrfall, P., 2013. Package “mulcomp” (WWW document). U.R.L. 604 http://cran.stat.sfu.ca/web/packages/mulcomp/mulcomp.pdf. (accessed 19.12.17). 605 Inclan, R., Gimeno, B.S., Dizengremel, P., Sanchez, M., 2005. Compensation processes of Aleppo 606 pine (Pinus halepensis Mill.) to ozone exposure and drought stress. Environ. Pollut. 137, 607 517–524. 608 IPCC., 2013. International Panel on Climate Change. Climate Change 2013: The Physical Science 609 Basis. IPCC Working Group I Contribution to AR5. Edited by: Stocker, T. F., Qin, D., Plattner, 610 G. K., Tignor,M. M. B., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P. M. 611 Jucker, T., Grossiord, C., Bonal, D., Bouriaud, O., Gessler, A., Coomes, D.A., 2017. Detecting the 612 fingerprint of drought across Europe’s forest: do carbon isotope ratios and stem growth 613 rates tell similar stories? Forest Ecosystems, 4, art. Number 24. 614 Kooijman, A.M., Jongejans, J., Sevink, J., 2005. Parent materials effects on Mediterranean 615 woodland ecosystems in NE Spain. Catena, 59, 55-68. 616 Kreuzwieser, J., Gessler, A., 2010. Global climate change and tree nutrition: influence of water 617 availability. Tree Physiol. 30, 1221-1234. 618 Kruger, F.J., 1979. South African Heathlands. In Ecosystems of the World. (Ed.) R L Specht. pp. 619 19–80. Heathlands and Related Shcrublands. Elservier, Amsterdam. 620 Kutiel, H., Maheras, P., Guika, S., 1996. Circulation and extreme rainfall conditions in the Eastern 621 Mediterranean during the last century. Inter. J. Climat 16, 72–92. 622 Llasat, M.C., Marcos, R., Turco, M., Gilabert, J., Llasat-Botja, M., 2016. Trends in flash flood 623 events versus convective precipitation in the Mediterranean region: the case of Catalonia. 624 Jo. Hydrol. 541, 24-37. 18 625 Lledó, M.J., 1990. Compartimentos y flujos biogeoquímicos en una cuenca de encinar del Monte 626 de Poblet. Ph.D. thesis, Universitat d’Alacant. 627 Loizidou, M., Giannakopoulos, C., Bindi, M., Moustakas, K., 2016 Climate change impacts and 628 adaptation options in the Mediterranean basin Reg. Environ.Change 16, 1859–1861. 629 Luo, W.T., Zuo, X.A., Ma, W., Xu, C., Li, A., Yu, Q., Knapp, A.K., Tognetti, R., Dijstra, F.A., Li, M.H., 630 Han, G., Wang, Z., Han, X.G., (2018). Differential response of canopy nutrients to 631 experimental drought along a natural gradient. Ecolog, 99, 2230-2239. 632 Martin-StPaul NK, Limousin, JM, Vogt-Schild H, Rodriguez-Calcerrada J, Rambal S, Longepierre 633 D, Misson L. 2013. The temporal response to drought in a Mediterranean evergreen: 634 comparing a regional precipitation gradient and athroughfall exclusion experiment. Glob. 635 Change Biol. 19, 2413-2426. 636 Matías, L., Castro, J., Zamora, R. 2011. Soil-nutrient availability under a global-change scenario 637 in a Mediterranean mountain ecosystem. Glob. Change Biol. 17, 1646–1657. 638 Ogaya, R., Peñuelas, J., 2003. Comparative field study of Quecus ilex and Phillyrea latifolia: 639 Photosynthetic response to experimental drought conditions. Environ. Exp, Bot. 50, 137– 640 148. 641 Ogaya, R., Peñuelas, J., 2004. Phenological patterns of Quercus ilex, Phillyrea latifolia, and 642 Arbutus unedo growing under a field experimental drought. Ecoscience 11, 263-270. 643 Ogaya, R., Peñuelas, J., 2006. Contrasting foliar responses to drought in Quercus ilex and 644 Phillyrea latifolia. Biol. Plant. 50, 373-382. 645 Ogaya, R., Peñuelas, J., 2007. Tree growth, mortality, and above-ground biomass accumulation 646 in a holm oak forest under a five-year experiment field drought. Plant Ecol. 189, 291-299. 647 Ogaya, R., Peñuelas, J., Asensio, D., Llusià, J., 2011. Chlorophyll fluorescence responses to 648 temperature and water availability in two co-dominant Mediterranean shrub and tree 649 species in a long-term field experiment simulating climate change. Environ. Exp. Bot. 73, 650 89-93. 651 Ogaya, R., Llusià, J., Barbeta, A., Asensio, D., Liu, D., Alessio, G.A., Peñuelas, J., 2014. Foliar CO2 652 in a holm oak forest subjected to 15 years of climate change simulation. Plant Sci. 226, 101– 653 107. 654 Ogaya, R., Liu, D., Barbeta, A., Peñuelas, J., 2020. Stem Mortality and Forest Dieback in a 20- 655 Years Experimental Drought in a Mediterranean Holm Oak Forest. Frontiers in Forest Global 656 Change, 2, 89. doi: 10.3389/ffgc.2019.00089. 657 Peñuelas, J., Sardans, J., Llusia, J., Owen, S., Carnicer, J., Giambeluca, T.W., Rezende, E.L., Waite 658 M., Niinemets Ü., 2010. Faster returns on ‘leaf economics’ and different biogeochemical 659 niche in invasive plant species. Glob. Change Biol. 16, 2171-2185. 660 Peñuelas, J., Sardans, J., Estiarte, M., Ogaya, R., Carnicer, J., Coll, M., Barbeta, A., Rivas-Ubach, 661 A., Llusia, J., Garbulski, M., Filella, I., Jump, A.S., 2013. Evidence of current impact of climate 662 change on life: a walk from genes to the biosphere. Glob. Change Biol. 19, 2303-2338. 663 Peñuelas, J., Sardans, J., Filella, I., Estiarte, M., Llusià, J., Ogaya, R., Carnicer, J., Bartrons, M., 664 Rivas-Ubach, A., Grau, O., Peguero, G., Margalef, O., Pla-Rabés, S., Stefanescu, C., Asensio, 665 D., Preece, C., Liu, L., Verger A., Rico, L., Barbeta, A., Achotegui-Castells, A., Gargallo- 666 Garriga, A., Sperlich, D., Farré-Armengol, G., Fernández-Martínez, M., Liu, D., Zhang, C., 667 Urbina, I., Camino, M., Vives, M., Nadal-Sala, D., Sabaté, S., Gracia, C., Terradas, J., 2018. 668 Assessment of the impacts of climate change on Mediterranean terrestrial ecosystems 669 based on data from field experiments and long-term monitored field gradients in Catalonia. 670 Environ. Exp. Bot. 152, 49-59. 671 Peñuelas, J., Sardans, J., Filella, I., Estiarte, M., Llusia, J., Ogaya, R., Carnicer, J., Bartrons, M., 672 Rivas-Ubach, A., Grau, O., Pegure, G., Margalef, O., Pla-Rasbés, S., Stefanescu, C., Asensio, 673 D., Preece, C., Liu, L., Verger, A., Barbeta, A., Achotegui-Castells, A., Gargallo-Gartriga, A., 674 Sperlich, D., Farré-Armengol, G., Fernandez-Martinez, M., Liu, D., Zhan, C., Urbina, I., 675 Camino-Serrano, M., Vives-Ingla, M., Stocker, B., Balzarolo, M., Guerrini, R., Paucelle, M., 676 Marañon-Jimenez, S., Bórnez-Mejías, K., Mu, Z., Desclas, A., Castellanos, A., Terradas, J., 19 677 2017a. Impacts of global change on Mediterranean forests and their services. Forests 8, 678 463. 679 Peñuelas, J., Ciais, P., Canadell, J.G., Janssens, .I, Fernandez-Martinez, M., Carnicer, J., 680 Obersteiner, M., Piao, C., Vautard, R., Sardans, J,. 2017b. Shifting from a fertilization- 681 dominated to a warming-dominated period. Nature Ecol. Evol. 1, 1438-1445. 682 Peñuelas, J.., Janssens, I., Obersteniner, M., Ciais, P., Sardans, J., 2019. Biogeochemical niche: 683 the link between bio-elements and ecosystems. Ecology 100, e02652. 684 Peñuelas, J., Poulter, B., Sardans, J., Ciais, P., van der Velde M.,, Bopp, L., Boucher, O., Godderis, 685 Y., Hinsinger, P., Llusia, J., Nardin, E., Vicca, S., Obersteiner, M., Janssens, I.A., 2013. Human- 686 induced imbalances alter natural and manages ecosystems across the globe. Nat. Com. 4, 687 2934. 688 Peñuelas, J., Filella, I., Comas, P., 2002. Changed plant and animal life cycles from 1952 to 2000 689 in the Mediterranean Region. Glob. Change Biol. 8, 531–544. 690 Peñuelas, J., Janssens, I.A., Ciais, P., Obersteiner, M., Sardans, J., 2020. Anthropogenic global 691 shifts in biospheric N and P concentration and ratios and their impacts on biodiversity, 692 ecosystem productivity, food security, and human health. Glob. Change Biol. In press. 693 Piñol, J., Terradas, J., Lloret, F., 1998. Climate warming, wildfire hazard, and wildfire occurrence 694 in coastal eastern Spain. Clim. Change 38, 347–357. 695 Pérez-Ramos, I.M., Ourcival, J.M., Limousin, J.M., Rambal S,. 2010. Mast seedling under 696 increasing drought: results from a long-term data set and from a rainfall exclusion 697 experiment. Ecology 91, 3057–3068 698 Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., Core TR. 2016. nlme: Linear and Nonlinear Mixed 699 Effects Models. R package version 3.1-126, http://CRAN. 700 Porter, J.R., 1986. Evaluation of washing procedures for pollution analysis of Ailanthus altissima 701 leaves. Environ. Pollut. 12, 195–202. 702 Rivas-Ubach, A., Gargallo-Garriga, A., Sardans, J., Ovarec, M., Mateu-Castell, L., Perez-Trujillo M., 703 et al. Peñuelas, J., 2014. Drought enhances folivory by shifting foliar metabolomes in 704 Quercus ilex trees. New Phytol. 202, 874-885. 705 Ruiz-Sinoga, J.D., Gabarón Galeote, M.A., Martínez Murillo, J.F., Garcia Marín, R., 2011. 706 Vegetation strategies for soil consumption along a pluviometric gradient in southern Spain. 707 Catena 84, 12–20. 708 Sabaté, S., Gracia, C., 1994. Canopy nutrient of a Quercus ilex L. forest: fertilization and irrigation 709 effects. For. Ecol. Manag. 68, 31–37. 710 Sadaka, N., Ponge, J.F., 2003. Climatic effects on soil trophic networks and the resulting humus 711 profiles in holm oak (Quercus rotundifolia) forests in the high Atlas of Morocco as revealed 712 by correspondence analysis. Eur. J. Soil Sci. 54, 767–777. 713 Salvati, L.,, Kosmas C.,, Kairis, O., Karavitis, C.A., Acikalin, S., Belgacem, A., Solé-Benet, A., 714 Chaker M.,, Fassouli, V., Gokceoglu C.,, Gungor, H., Hessel, R., Khatteli H.,, Kounalaki, A., 715 Laouina, A., Ocakoglu, F.,Ouessar, M., Ritsema, C. J., Sghaier, M., , Taamallah, H.M.H., 716 Tezcan,L.,, de Vente, J., 2014 Unveiling soil degradation and desertification risk in the 717 Mediterranean basin: a data mining analysis of the relationships between biophysical and 718 socioeconomic factors in agro-forest landscapes. J. Environ. Plan. Manag. 58, 1789-1803. 719 Sanaullah, M., Rumpel, C., Charrier, X., Chabbi, A., 2012. How does drought stress influence the 720 decomposition of plant litter with contrasting quality in a grassland ecosystem? Plant Soil 721 352, 277–288. 722 Sánchez, E., Gallardo, C., Gaertner, M.A., Arribas, A., Castro, M., 2004. Future climate extreme 723 events in the Mediterranean simulated by a regional climate model: a first approach. Glob. 724 Planet. Change 44, 163-180. 725 Sardans, J., Peñuelas, J., 2005. Drought decreases soil enzyme activity in a Mediterranean 726 Quercus ilex L. forest. Soil Biol. Biochem. 37, 455-461. 20 727 Sardans, J., Peñuelas, J., Rodà, F., 2006. Plasticity of leaf morphological traits, leaf nutrient 728 content, and water capture in the Mediterranean evergreen oak Quercus ilex subsp. Ballota 729 in response to fertilization and changes in competitive conditions. Ecoscience 13, 258–270. 730 Sardans, J., Peñuelas J., 2007. Drought changes phosphorus and potassium accumulation 731 patterns in an evergreen Mediterranean forest. Funct. Ecol. 21, 191-201. 732 Sardans, J., Peñuelas, J., Ogaya, R., 2008a. Drought-induced changes in C and N stoichiometry in 733 a Quercus ilex Mediterranean forest. For. Sci. 54, 513-522. 734 Sardans, J., Peñuelas, J., Ogaya, R., 2008b. Experimental drought reduced acid and alkaline 735 phosphatase activity and increased organic extractable P in soil in a Quercus ilex 736 Mediterranean forest. Eur. J. Soil Sci. 44, 509-520. 737 Sardans, J., Peñuelas, J., Estiarte, M., Prieto, P., 2008c. Warming and drought alter C and N 738 concentration, allocation and accumulation in a Mediterranean shrubland. Glob. Change 739 Biol. 14, 2304–2316. 740 Sardans, J., Peñuelas, J., 2010. Soil enzyme activity in a Mediterranean forest after six years of 741 drought. Soil Sci. Soc. Am. J. 74: 838-851. 742 Sardans, J., Rivas-Ubach, A., Peñuelas, J., 2012. The C:N:P stoichiometry of organisms and 743 ecosystems in a changing world: A review and perspectives. Perspectives in Plant Ecology, 744 Evolution and Systematics, 14, 33-47. 745 Sardans, J., Peñuelas J., 2013. Plant-soil interactions in Medierranean forest and shrublands: 746 impacts of climate change. Plant Soil 365, 1-33. 747 Sardans, J., Rivas-Ubach, A., Estiarte, M., Ogaya, R., Peñuelas, J., 2013. Field-simulated droughts 748 affect elemental leaf stoichiometry in Mediterranean forest and shrublands. Acta Oecol. 749 50, 20-31. 750 Sardans, J., Vallicrosa, H., Zuccarini, P. Farré-Armengol G., Fernández-Martínez, M., Guille 751 Peguero, Gargallo-Garriga, A., Ciais P., Janssens I.A., Obersteiner M., Richter, A., Peñuelas, 752 J. (2020) Empirical support for the Biogeochemical Niche Hypothesis in forest trees. Nature 753 Ecology and Evolution. Under final revision. 754 Specht, R.L., 1979. The sclerophyllous (heath) vegetation of Australia: The eastern and central 755 states. In Heathlands and Related Shrublands. Descriptive Studies. (Ed.) R L Specht. pp. 756 125–210. Elservier. Amsterdam 757 Sterner, R.W., Elser, J.J., 2002. Ecological Stoichiometry the Biology of Elements from Molecules 758 to the Biosphere. Princeton University Press, Princeton, NJ, USA. 759 Sun, Y., Peng, S., Goll, D.G., Ciais, P., Guenet, B., Guimbertea, M., Hinsinger, P., Jannssens, I.A., 760 Peñuelas J., Piao, S., Poulter, B., Violette, A., Yang, X., Yin, Y., Zeng, H., Zeng, H., 2017. 761 Diagnosing phospsorus limitations in natural terrestrial ecosystems in carbon cycle models. 762 Earth's Fut. 5, 730-749 763 Urbina, I., Sardans, J., Beierkuhnlein, C., Jenntsch, A., Backhaus, S., Grant, K.,,…et al. Peñuelas, 764 J., 2015. Shifts in the elemental composition of plants during a very severe drought. 765 Environ. Exp. Bot. 111, 63-73. 766 Valiente-Baunet, A., Flores-Hernández, N., Verdú, M., Dávila, P., 1998. The chaparral vegetation 767 in Mexico under non- Mediterranean climate: The convergence and Madrean-Tethyan 768 hypotheses reconsidered. Am. J. Bot. 85, 1398–1408. 769 Verdú, M., Dávila, P., García-Fayos, P., Flores-Hernández, N., Valiente-Banuet, A., 2003. 770 “Convergent” traits on Mediterranean Woody plants belong to pre-Mediterranean 771 lineages. Biol. J. Linenean Soc. 78, 415-427. 772 Zegelin, S.J. , White, D.R., Jenkins, D.R., 1989. Improved field probes for soil water content and 773 electrical conductivity measurement using time domain reflectometry. Water Resour. Res. 774 25, 2367 – 2376 775 Zhao, M.S., Running, S.W., 2010. Drought-Induced reduction in global terrestrial net primary 776 production from 2000 through 2009. Science 329, 940–943. 21 777 Ziv, B., Harats, N., Morin, E., Yai, Y., Dayan, U., 2016. Can serve rain events over the 778 Mediterranean region be detected through simple numerical indices? Natural Hazards 83, 779 1197-1212. 780 781 Figure Captions 782 783 Figure 1. Seasonal 15N (mean ± S.E.) in Q. ilex leaves (A) and foliar litter (B), P. latifolia leaves 784 (C) and foliar litter (D) and A. unedo leaves (E) and foliar litter (G) in the control and drought 785 plots 15 years after the beginning of the experiment. Different letters indicate significant 786 differences (P<0.05). 787 Figure 2. Seasonal soil 15N (mean ± S.E.) in the control and drought plots 15 years after the 788 beginning of the experiment. Different letters indicate significant differences (P<0.05). 789 Figure 3. GDA analysis of the foliar concentrations in the three dominant species and pairwise 790 ratios as variables. The samples of each species in the control and drought plots were 791 collected in 1999, 2004 and 2014 as grouping factors. 792 Figure 4. Differences in foliar carbon stocks (mean ± S.E.) for 1999-2004 and 1999-2014 793 between the control and drought plots for Q. ilex (A), P. latifolia (B), A. unedo (C) and all 794 three species (D). Different letters indicate significant differences (P<0.05). Positive and 795 negative values indicate increases and decreases, respectively, of the carbon stocks during 796 the corresponding time period. 797 Figure 5. Differences in foliar nitrogen stocks (mean ± S.E.) for 1999-2004 and 1999-2014 798 between the control and drought plots for Q. ilex (A), P. latifolia (B), A. unedo (C) and all 799 three species (D). Different letters indicate significant differences (P<0.005). Positive and 800 negative values represents increases and decreases, respectively, of the nitrogen stocks 801 during the corresponding time period. 802 Figure 6. Differences in foliar phosphorus stocks (mean ± S.E.) for 1999-2004 and 1999-2014 803 between the control and drought plots for Q. ilex (A), P. latifolia (B), A. unedo (C) and all 804 three species (D). Different letters indicate significant differences (P<0.005). Positive and 805 negative values represents increases and decreases, respectively, ofthe phosphorus stocks 806 during the corresponding time period. 807 Figure 7. Differences in foliar potassium stocks (mean ± S.E.) for 1999-2004 and 1999-2014 808 between the control and drought plots for Q. ilex (A), P. latifolia (B), A. unedo (C) and all 809 three species (D). Different letters indicate significant differences (P<0.005). Positive and 810 negative values represents increases and decreases, respectively, of potassium stocks 811 during the corresponding time period. 812 Figure 8. Differences in soil stocks (upper 15 cm, kg ha-1) for 2005-2014 between the control and 813 drought plots. Positive and negative values represents increases and decreases, respectively, 814 of the carbon, nitrogen, phosphorus stocks in soils during the corresponding time period. 815 816 817 818 819 820 821 822 823 824 825 826 827 828 22 829 830 831 832 833 LEAVES FOLIAR LITTER 1 Control Drought A B Quercus ilex  N (‰) 0 15 -1 -2 a -3 a a a a a b -4 b b b b b -5 1 C D 0 Phillyrea latifolia  N (‰) -1 15 -2 -3 -4 a a a a -5 b b b b 1 E F 0 Arbutus unedo  N (‰) -1 15 -2 a a -3 a a a a b b -4 b b b -5 b Spring Summer Autumn Spring Summer Autumn Season Season Figure 1 834 835 23 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 Figure 2 853 854 855 856 857 858 859 860 24 861 862 863 864 865 866 867 868 869 870 Quercus ilex control (A) Phillyrea latifolia control (C) Arbutus unedo control (E) Quercus ilex drought (B) Phillyrea latifolia drought (D) Arbutus unedo drought (F) 4 0.8 B (1999) A B C:K N:K B (2005) 2 0.4 Root 2 (24.3) C:N Root 2 (24.3) Ca A (2005) E (1999) F (2005) C:P Mg Loading A (1999) D (1999) Score 0 E (1999) 0 A (2014) D (2005) N:P P B (2014) C (1999) E (2005) C Mg:Ca C (2005) P:Mg F (2014) N:Mg P:Ca -0.4 N:Ca N -2 E (2014) D (2014) C (2014) K -4 -0.8 -4 -2 0 2 4 -1 -0.5 0 0.5 1 Root 1 (34.5) Root 1 (34.5) Score Loading 871 872 873 874 875 876 Figure 3 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 25 898 899 900 901 902 903 904 905 906 907 908 909 910 911 140 112.5 a Control A B Change in carbon stocks (kg ha-1) Change in carbon stocks (kg ha-1) Drought 120 100 Phillyrea latifolia 0 Quercus ilex 80 -112.5 60 40 b -225 20 0 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 200 C 300 a D Change in carbon stocks (kg ha-1) Change in carbon stocks (kg ha-1) Q. ilex + P. latifolia + A. unedo a 150 200 a Arbutus unedo 100 100 a 0 50 b -100 b b 0 -200 b 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 912 913 Figure 4 914 915 916 917 918 919 920 921 922 923 924 26 925 926 927 928 929 930 931 932 933 934 935 936 937 5 8 Control A B Change in nitrogen stocks (kg ha-1) Change in nitrogen stocks (kg ha-1) Drought 6 0 4 Phillyrea latifolia Quercus ilex -5 2 -10 0 -2 -15 -4 -20 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 10 10 C D Change in nitrogen stocks (kg ha-1) a Change in nitrogen stocks (kg ha-1) a Q. ilex + P. latifolia + A. unedo 5 8 0 Arbutus unedo 6 -5 4 b -10 b 2 -15 0 -20 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 938 939 940 Figure 5 941 942 943 944 945 946 947 948 949 950 27 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 70 A 0.6 B Change in phosphorus stocks (kg ha-1) Control Change in phosphorus stocks (kg ha-1) 60 Drought 0.5 50 0.4 Phillyrea latifolia Quercus ilex 40 0.3 30 0.2 20 10 0.1 0 0 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 0.8 Change in phosphorus stocks (kg ha-1) 70 C D Change in phosphorus stocks (kg ha-1) Q. ilex + P. latifolia + A. unedo 60 0.6 50 Arbutus unedo 40 0.4 30 0.2 20 10 0 0 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 966 967 Figure 6 968 969 970 971 972 973 974 975 976 28 977 978 979 980 981 982 983 984 985 986 987 988 12 Control A B Change Potassium stocks (kg ha-1) Change Potassium stocks (kg ha-1) 20 Drought 10 8 Phillyrea latifolia 15 Quercus ilex 6 10 4 5 2 0 0 1999-2004 1999-2014 1999-2004 1999-2014 Period Period 10 30 C D Change Potsasium stocks (kg ha-1) a Change Potassium stocks (kg ha-1) Q. ilex + P. latifolia + A. unedo 8 25 20 Arbutus unedo 6 15 4 b 10 2 5 0 1999-2004 1999-2014 0 1999-2004 1999-2014 Period Period 989 990 991 992 993 994 995 996 997 998 999 1000 Figure 7 1001 1002 29 1003 1004 1005 1006 1007 1008 1009 1010 1011 Control Drought Carbon Nitrogen b a 4 4 -1 10 -5000 0 5000 1 10 -800 -400 0 400 800 1200 Phosphorus Potassium -20 -10 0 10 20 -4000 -2000 0 2000 4000 -1 Increase of stocks (kg ha in the first 15 cm of soil) in the period 2005-2014 1012 1013 Figure 8 30