Water source partitioning among plant functional types in a semi-arid dune ecosystem

1 This is the peer-reviewed version of the article accepted for publication in JOURNAL OF VEGETATION 1 SCIENCE Volume 29, 671-683. 2018, which has been published in final form at 2 https://doi.org/10.1111/jvs.12647 3 4 Water source partitioning among plant functional types in a semi-arid 5 dune ecosystem 6 7 Cristina Antunes a,b*, Maria Cruz Díaz Barradas c, Maria Zunzunegui c, Simone Vieira d, 8 Cristina Máguas a 9 (a) Centro de Ecologia, Evolução e Alterações Ambientais, Faculdade de Ciências, Universidade de 10 Lisboa, Lisboa, Portugal 11 (b) PPG - Ecologia, Instituto de Biologia, Universidade Estadual de Campinas, São Paulo, Brazil 12 (c) Departamento de Biología Vegetal y Ecología, Universidad de Sevilla, Spain 13 (d) Núcleo de Estudos e Pesquisas Ambientais, Universidade Estadual de Campinas, São Paulo, 14 Brazil 15 * Corresponding author: [email protected] 16 17 Abstract 18 Question(s): The aim of this work was to characterize the main water sources used by the 19 woody plant community of a semi-arid coastal dune ecosystem and to enlighten the relevance 20 of soil water partitioning. We ask: Do the seasonal water sources used by plants differ 21 between functional types? Does the seasonal drought promote shifts towards deeper soil 22 layers in all plants? Does greater use of deep soil water enables the maintenance of a more 23 favorable plant water status? Do water-sources-use strategies mirror overall drought 24 strategies? 25 Location: Semi-arid coastal dune system in Doñana Biological Reserve, southwest Spain. 26 Methods: Fourteen coexisting woody species with contrasting functional traits were 27 considered. We analyzed each plant’s xylem water oxygen isotope composition and 28 compared it to water sources (soil water at 10, 30, 50 cm, rain and groundwater), both in 29 2 spring and dry summer. Bayesian isotope mixing models were used to estimate the proportion 30 of each water source used by plants. We tested the influence of different (categorical) traits 31 (growth form, water requirement, leaf longevity and leaf form) on the patterns of plants’ 32 water-source-use. We evaluated the relationships between proportions of deep soil water use 33 and ecophysiological parameters related to water economy (leaf δ13C and water-index). 34 Results: A greater similarity between species’ water-source-use was found in periods of 35 higher water availability (spring). Contrastingly, during the dry season, a high inter-specific 36 differentiation was observed. All four functional traits showed a significant effect on water 37 sources used in summer. Accordingly, species clustered in five water-use functional groups, 38 exploring water from shallower to deeper soil layers. Under strong drought conditions, the 39 highest use of shallow soil layers was observed in narrow-leaved xerophytic shrubs, while 40 evergreen xerophytic trees and evergreen hygrophytic species showed a higher use of deep 41 soil layers. Greater use of deep soil layers in the dry season was linked to a better seasonal 42 maintenance of plant water status. 43 Conclusions: Coexisting plant functional types segregated along a wide spectrum of 44 contrasting water-source use under extremely dry conditions, evidencing great soil water 45 partitioning. Relevant seasonal water-use shifts towards deeper soil layers were observed, 46 but not in all functional types. Furthermore, we confirmed that the water-source-use behavior 47 accompanied broader strategies of drought resistance. 48 49 key-words: soil water partitioning; water-sources segregation; coexistence; groundwater 50 use; Mediterranean; functional groups; sandy coastal ecosystem; dry conditions; stable 51 isotope mixing model; water-use shifts; drought 52 53 running head: segregation in water-sources used by woody plants 54 55 nomenclature: Castroviejo et al. 1986–2012 56 57 58 3 59 Introduction 60 The function, performance and, ultimately, survival of plant species living in dry ecosystems 61 depend significantly on water availability as well as on the ability to use the available water 62 (Austin et al. 2004; Grossiord et al. 2017; Palacio et al. 2017). In semi-arid systems, 63 interspecific (belowground) competition for water defines plant communities (Fowler 1986; 64 Eissenstat & Caldwell 1988; Casper & Jackson 1997; Schenk & Jackson 2002). Nevertheless, 65 different species can coexist by using different ranges and proportions of resources 66 (Silvertown et al. 1999, 2015; Filella & Peñuelas 2003; Valladares et al. 2015; Palacio et al. 67 2017). Hence, temporal and spatial (vertical) water partition is expected to represent an 68 important process in these seasonal dry ecosystems. This is especially relevant in sandy soils 69 with low water retention (Veneklaas & Poot 2003; Alessio et al. 2004). When exposed to an 70 extreme decrease in soil water availability and in the absence of rain, dune woody plants will 71 need to rely on another water source and/or readjust root system and water uptake to new 72 water levels. In fact, under drought, water adjustments can and must be made, otherwise the 73 impacts will be detected in plant performance and eventually in plant dieback (Nepstad et al. 74 2007; Padilla & Pugnaire 2007; Allen et al. 2010; Barbeta et al. 2015). 75 There is evidence of water sources use segregation between growth forms in several 76 environments (Silvertown et al. 1999, 2015; Nippert & Knapp 2007; Barbeta & Peñuelas 77 2017). In fact, an important differentiation in Mediterranean communities is established 78 between tree species with deep roots, with more constant water and nutrient resources, and 79 shrubby species with shallow roots, which use episodic rainwater and associated nutrient 80 uptake (Moreno-Gutiérrez et al. 2012). Water-use under drought will determine the degree 81 to which a plant experiences a given hydric variation, with deep-rooted species being 82 possibly buffered from drought to a greater extent than shallow-rooted species (Nepstad et 83 al. 1994, 2007; Oliveira et al. 2005; West et al. 2012). Besides growth form, other key 84 functional traits can play a role in water-sources use patterns (Leng et al. 2013). Leaf 85 phenology is one of them. Studies in Mediterranean scrub, desert savannas, and tropical dry 86 forests have shown that evergreen species shift water uptake from shallow water to deep 87 water in the dry season (Jackson et al. 1999; Hasselquist et al. 2010; Esquivias et al. 2014; 88 4 Ellsworth & Sternberg 2015; Grossiord et al. 2017). As an adaptation for water absorption 89 and for protection against water deficit, leaf shape can also affect water-use patterns of plants 90 (Wright et al. 2006; Antunes et al. 2018). Additionally, an important functional syndrome 91 related with water-use is plant water requirement. Species can explore different niches 92 according to their relative requirement of soil moisture; hygrophytes will grow in a moist 93 habitat and xerophytes are adapted to survive in an environment with little water (Ackerly 94 2004). This classification has been pointed out as key in differentiating semi-arid 95 Mediterranean coastal dune plants’ physiological patterns (Zunzunegui et al. 2005, 2009, 96 2011). 97 Although the basis for niche segregation by rooting depth in plant communities have been 98 studied (Silvertown et al. 1999, 2015; Filella & Peñuelas 2003; Araya et al. 2011; Peñuelas 99 et al. 2011; Palacio et al. 2017) there are water-sources use differences between plants that 100 still have been overlooked. Attempts to understand mechanisms underlying plant responses 101 to drought have led to the emergence of distinct hydraulic and water strategies among 102 coexisting species (McDowell et al. 2008; West et al. 2012; Skelton et al. 2015; Pivovaroff 103 et al. 2016). However, when exploring plant water strategies, specific quantification of water 104 sources used by plants is not usually used. This is an important piece of the puzzle since it 105 can be linked to the ability to explore different water sources (and in some cases buffering 106 the drought effect). Furthermore, it can clarify soil water partitioning (i.e. differentiation of 107 water-uptake depth) among species. Because partitioning of water based solely on 108 differences in rooting depth does not explain patterns of water uptake for many species 109 (Jackson et al. 1999; Oliveira et al. 2005; Nippert & Knapp 2007; Ellsworth & Sternberg 110 2015), quantifying the relative contribution of different water sources to the composition of 111 the xylem water is utmost relevant. One valuable approach to do so is through Bayesian stable 112 isotope mixing models (Parnell et al. 2013; Stock et al. 2016; Palacio et al. 2017). 113 The aim of this work was to characterize the woody community seasonal water source use 114 and enlighten the relevance of segregation in the sources of water used by woody species in 115 a semi-arid coastal dune ecosystem. Through an isotopic approach, and considering fourteen 116 species of widely contrasting functional traits, we explored: (i) which are the main water 117 sources used by plants under contrasting water availability conditions, (ii) the influence of 118 functional traits on patterns of water-sources use, (iii) the seasonal water-source-use variation 119 and possible shifts towards deeper soil layers in drier conditions, (iv) the effect of water- 120 5 sources used on plant water status, and (v) the link between water-use functional groups and 121 plants’ drought strategies. 122 143 Methods 144 Study site and plant species 145 The study was conducted at a semi-arid Mediterranean coastal dune ecosystem at Biological 146 Reserve of Doñana, southwest of Spain (AI=0.47) (UNEP 1992; http://www.cgiar- 147 csi.org/data/global-aridity-and-pet-database) (Fig. S1). It showed a typical western 148 Mediterranean climate, with a total annual rainfall (August 2012 – August 2013) of 566.1 149 mm (weather station at RBD: 36°59'19"N; 6°26'35"W, Fig. S1). During the dry season, the 150 well-drained sand, underlying the sandy plant community in Doñana scrub, exacerbates soil 151 drying and water limitation. Soil water content, measured at 10 cm depth, was reduced by 152 88% from spring to summer, reaching very low values in the dry season (Fig. S1). 153 The sand mantle is mostly covered by Mediterranean scrub vegetation with local patches of 154 the original Juniperus and Quercus woodlands (Muñoz-Reinoso & García Novo 2005). To 155 represent the dominant vegetation of the studied sandy ecosystem, we selected fourteen 156 woody plant species, with contrasting functional traits. The categorical functional traits 157 considered were: growth form (GF), water requirements (WR), leaf longevity (LL) and leaf 158 form (LF) (Table 1). The following classes within each trait were considered: GF – shrub or 159 tree; WR – xerophytic, mesophytic or hygrophytic; LL – summer semi-deciduous, evergreen, 160 aphyllous; LF – narrow, broadleaf, needle, scale, spike/aphyllous. The sampled species and 161 their functional classification (based on (Díaz Barradas et al. 1999; Zunzunegui et al. 2005; 162 Zunzunegui et al. 2011; Castroviejo 2012) are described in Table 1. To better characterize 163 the woody community at landscape level, and encompass eventual local heterogeneity, we 164 consider a study area covering the prevalent dune formations (of ~7 km2). Eighteen sampling 165 plots (20 m x 20 m), separated by at least 200 m, were distributed throughout the study area. 166 In each plot, we sampled 3 to 4 species, belonging to contrasting functional types. We assured 167 that, within each plot, species of: both GF classes, at least two WR classes (except in one 168 plot), at least two LL classes, and at least two LF classes were sampled (Table S1). When 169 6 possible, we sampled three individuals per species within plot (replicates), resulting in a total 170 of 180 (adult) individuals sampled in each sampling date (Table 1, Table S1). 171 172 Analysis of xylem and sources water δ18O 173 To access the water being used by plants, lignified woody stems of each plant individual 174 selected were collected in two contrasting water conditions: spring (humid) and summer (dry) 175 (Fig. S1). 176 To determine available water sources for the roots, water samples were collected from soil, 177 rain and groundwater. Rainwater was sampled on one pluviometer equipped with liquid 178 paraffin to prevent evaporation (at 36°59'19"N, 6°26'35"W). We considered the rainwater 179 collected over the 2 weeks prior to stem sampling. Groundwater samples were collected in 180 piezometers or wells reaching the water table on the two periods of stem sampling. Mean 181 depth to water table was 2.8 and 3.9 m in spring and summer sampling dates, respectively. 182 Soil samples were collected at 3 depths: 10 cm, 30 cm and 50 cm in each sampling plot with 183 3 replicates per depth, on the same days of stem sampling. Having in mind the extreme 184 conditions of Mediterranean dune sites, especially in summer, soil and twig samples were 185 immediately stored in airtight vials, sealed with a cap and parafilm and placed in a cool ice 186 chest until they could be stored. All samples were kept in the fridge until isotopic analysis. 187 Water from xylem and soil samples was extracted using a custom-made cryogenic vacuum 188 distillation system housed at SIIAF (Ehleringer & Dawson 1992; West et al. 2007). All 189 samples were kept at 4°C until isotopic analysis. When the isotopic fraction of the available 190 water sources (soil water, rain, and groundwater) is known, an analysis of the oxygen isotopic 191 composition of xylem water provides information about the water sources being used by the 192 plant at the time of study (Ehleringer & Dawson 1992; Dawson 1993; Dawson & Pate 1996). 193 We used stem water δ18O values (rather than δD) as a measure of plant water sources because 194 xerophytic plants might be able to discriminate against deuterium during soil water uptake 195 by roots (Ellsworth & Williams 2007). Oxygen stable isotope ratio analyses were performed 196 by headspace equilibration, on an Isoprime (Micromass, UK) SIRMS, coupled on continuous 197 flow mode to a Multiflow (Micromass, UK) auto-sampler and sample equilibration system. 198 Analytical uncertainty is <0.1‰. 199 7 Soil water at 10, 30 and 50 cm depth and groundwater are from now on named SW10, SW30, 200 SW50 and GW respectively. 201 202 Quantification of water sources used by plants 203 The relative contribution of different water sources to the composition of the xylem water 204 was estimated by Bayesian stable isotope mixing models using the graphical user interface 205 and model framework MixSIAR for R (Parnell et al. 2013; Stock et al. 2016). In our study, 206 the model used as input the individual isotope values of xylem water δ18O of each plant 207 (‘mixture’ raw data), all sources described in water sources (mean plus standard deviation) 208 and a matrix for potential fractionation (set to 0). We calculated the water use proportion in 209 the two seasons and in each plot separately, using an individual-based model to obtain 210 separate estimates for each of the plants. We ran models using the individual plant (‘ID’ 211 within the ‘plot’) values of xylem water δ18O, the sampling plot water sources mean (plus 212 standard deviation) and the ‘process only (n=1)’ error structure option. We set the Markov 213 Chain Monte Carlo to ‘long’ or ‘very long’ model run, until the diagnostics were satisfying. 214 Spring precipitation was (prior) aggregated with SW10 because their isotopic composition 215 was similar. Thus, spring mixing model ran with only 4 sources (Precipitation+SW10, SW30, 216 SW50 and GW). Summer mixing model ran with the 4 sources sampled (SW10, SW30, 217 SW50 and GW), since it didn’t rain in the two weeks prior to stem sampling (Fig. S2). Details 218 of extremely enriched xylem water samples are given in Appendix 1. 219 220 Physiological parameters 221 The following ecophysiologcal traits were measured in all plants (n=180) (Table 1, Table S1) 222 in both spring and summer sampling dates (Fig S1). 223 Leaf carbon isotope composition (  13C) 224 Leaf 13C is a good proxy of leaf-level intrinsic water use efficiency (WUEi), which is given 225 by the ratio between leaf net photosynthetic rate (A) and gs (Farquhar et al. 1989; Dawson et 226 al. 2002). Mature leaves (2-10, depending on leaf size) were collected from each plant. The 227 bulk leaf samples were dried at 60 ºC for at least 48 h, and milled to fine powder in a ball 228 mill (Retsch MM 2000, Germany) for isotopic analysis. Leaf 13C was determined by 229 continuous flow isotope ratio mass spectrometry (CF-IRMS) on a Sercon Hydra 20-22 230 8 (Sercon, UK) stable isotope ratio mass spectrometer, coupled to a EuroEA (EuroVector, 231 Italy) elemental analyser. Uncertainty of the isotope ratio analysis, calculated using values 232 from 6 to 9 replicates of secondary isotopic reference material interspersed among samples 233 in every batch analysis, was ≤ 0.1‰. 234 Reflectance water index 235 We used a nondestructive optical method, based on the reflectance of light by an intact leaf. 236 Spectral reflectance was measured using a UniSpec Spectral Analysis System (PP Systems, 237 Haverhill, Massachusetts, USA) and carried out in 6 full expanded leaves per plant. The mean 238 value of the 6 measurements per plant was considered to calculate: 239 Water Index (WI) = R900/R970, where R900 and R970 are the reflectance at wavelengths of 240 900 nm and 970 nm (Peñuelas et al. 1997). This index is highly related with plant water 241 content, so it can be used as a proxy for water status in the plant (Claudio et al. 2006). 242 Additionally, the plant water content (PWC) was calculated as: 𝑃𝑊𝐶(%)= −620 +684 ∗ 243 𝑊𝐼 (following Peñuelas et al. (1997)). 244 245 Statistical analysis 246 To explore the influence of different traits on the patterns of plants’ water-source use, firstly 247 we performed Principal Component Analyses (PCA) based on all individuals information 248 (n=180), using (i) the relative contribution of the water sources to xylem water per season 249 (Fig. S3ab), and (ii) the seasonal differences of the relative contribution of the water sources 250 (spring to summer variation, i.e. summer minus spring values) (Fig.S3c). Positive values 251 represent an increase of the proportions of water-source use from spring to summer, and 252 negative values a decrease from spring to summer. Secondly, we tested if these water-source- 253 use patterns were dependent on growth form (GF), water requirements (WR), leaf longevity 254 (LL) or leaf form (LF). For that, based on the individual scores of PCAs’ first axis (PC1), we 255 performed an analysis of variance (factorial ANOVA) to test the main effects of the four 256 categorical functional traits (GF, WR, LL and LF) on: (i) the water used by plants in spring 257 (PC1spring~GF+WR+LL+LF), (ii) the water used by plants in summer 258 (PC1summer~GF+WR+LL+LF), and (iii) the water-sources-use seasonal variation (summer 259 minus spring) (PC1diff~GF+WR+LL+LF). Lastly, we explored the differences of water- 260 sources used between plant functional groups. To do so, we used a priori functional 261 9 classification for each plant as the concatenation of the four trait variables 262 (GF&WR&LL&LS). We performed Tukey HSD post-hoc multiple comparisons of means to 263 test differences in water-sources-use between the a priori functional groups (PC1~FG4traits; 264 with PC1spring, PC1summer and PC1diff as dependent variables). Then, we used the grouping 265 resulting from the post-hoc information to define the new (a posteriori) functional groups of 266 water-source use. 267 We further tested the relationships between the proportions of deep soil use (i.e. the 268 proportion of water from deeper than 0.5 m soil layers used by plants, i.e. % of GW use) and 269 the physiological parameters (leaf 13C and WI in spring and summer, and seasonal variation 270 of leaf 13C and PWC) through Pearson correlations (for the community and per a posteriori 271 functional group). We also explored the differences in the physiological parameters (leaf 272 13C and WI) and in the seasonal variation of leaf 13C and PWC among the functional groups 273 by performing an analysis of variance (ANOVA) and post-hoc Tukey HSD. Preliminary 274 analyses were performed to ensure that there was no violation of statistical assumptions. 275 All statistical analyses were performed in R 3.2.3 (R Core Team, 2015). 276 277 Results 278 A greater similarity between species’ water-source use was found in periods of higher water 279 availability (spring). In this season, we only found a significant effect of growth form on 280 plants’ water-source use (F = 173.63, p<0.0001, Table 2). The composition of xylem water 281 of shrubs was better explained by precipitation mostly present at SW10, while water used by 282 trees was mostly from soil layers between 30 and 50 cm (Fig. 1a). At this period, some 283 individuals of the xerophytic tree J. phoenicea were using a high percentage of groundwater 284 (>60%) (Fig. 1a). During the dry season (summer), we observed a high range of water sources 285 used, from shallower to deeper soil layers, depending on plant functional traits. Growth form, 286 water requirements, leaf longevity, and leaf form showed a significant effect on the sources 287 of water used in summer (Table 2). After testing differences in water-source use, the 288 following new functional groups were defined: xerophytic shrubs with narrow leaves (NXs), 289 semi-deciduous mesophytic shrubs (sMs), aphyllous/spike or semi-deciduous broad-leaf 290 shrubs (Os), evergreen hygrophytes (eH) and evergreen xerophytic trees (eXt) (Fig. S4b). 291 These functional types acquired water from different soil depths, with their reliance on deeper 292 16 Araya, Y.N., Silvertown, J., Gowing, D.J., McConway, K.J., Peter Linder, H., & Midgley, G. 2011. 481 A fundamental, eco-hydrological basis for niche segregation in plant communities. New 482 Phytologist 189: 253–258. 483 Austin, A.T., Yahdjian, L., Stark, J.M., Belnap, J., Porporato, A., Norton, U., Ravetta, D.A., & 484 Schaeffer, S.M. 2004. Water pulses and biogeochemical cycles in arid and semiarid 485 ecosystems. Oecologia 141: 221–235. 486 Barbeta, A., Mejía-Chang, M., Ogaya, R., Voltas, J., Dawson, T.E., & Peñuelas, J. 2015. The 487 combined effects of a long-term experimental drought and an extreme drought on the use of 488 plant-water sources in a Mediterranean forest. Global Change Biology 21: 1213–1225. 489 Barbeta, A., & Peñuelas, J. 2017. Relative contribution of groundwater to plant transpiration 490 estimated with stable isotopes. Scientific Reports 7: 10580. 491 Casper, B.B., & Jackson, R.B. 1997. Plant Competition Underground. Annual Review of Ecology 492 and Systematics 28: 545–570. 493 Castroviejo, S. 2012. Flora Iberica (C. Castroviejo, S., Aedo, C., Cirujano, S., Laínz, M., 494 Montserrat, P., Morales, R., Muñoz Garmendia, F., Navarro, C., Paiva, J. & Soriano, Ed.). 495 Real Jardín Botánico, CSIC, Madrid. 496 Chapin, F.S., Autumn, K., & Pugnaire, F. 1993. Evolution of Suites of Traits in Response to 497 Environmental Stress. The American Naturalist 142: S78–S92. 498 Chesson, P., Gebauer, R.L.E., Schwinning, S., Huntly, N., Wiegand, K., Ernest, M.S.K., Sher, A., 499 Novoplansky, A., & Weltzin, J.F. 2004. Resource pulses, species interactions, and diversity 500 maintenance in arid and semi-arid environments. Oecologia 141: 236–253. 501 Claudio, H., Cheng, Y., Fuentes, D., Gamon, J., Luo, H., Oechel, W., Qiu, H.-L., Rahman, F., & 502 Sims, D. 2006. Monitoring drought effects on vegetation water content and fluxes in 503 chaparral with the 970 nm water band index. 504 Dawson, T. E. 1993. Water Sources of Plants as Determined from Xylem-Water Isotopic 505 Composition: Perspectives on Plant Competition, Distribution, and Water Relations. 506 Dawson, T.E., Mambelli, S., Plamboeck, A.H., Templer, P., & Tu, K.P. 2002. Stable Isotopes in 507 Plant Ecology. Annual Review of Ecology and Systematics 33: 507–559. 508 Dawson, T.E., & Pate, J.S. 1996. Seasonal water uptake and movement in root systems of 509 Australian phraeatophytic plants of dimorphic root morphology: a stable isotope investigation. 510 Oecologia 107: 13–20. 511 Díaz Barradas, M.C., Zunzunegui, M., Tirado, R., Ain-Lhout, F., & García Novo, F. 1999. Plant 512 functional types and ecosystem function in Mediterranean shrubland. Journal of Vegetation 513 Science 10: 709–716. 514 17 Ehleringer, J.R., & Dawson, T.E. 1992. Water uptake by plants: perspectives from stable isotope 515 composition. Plant, Cell & Environment 15: 1073–1082. 516 Eissenstat, D.M., & Caldwell, M.M. 1988. Competitive ability is linked to rates of water extraction. 517 Oecologia 75: 1–7. 518 Ellsworth, P.Z., & Sternberg, L.S.L. 2015. Seasonal water use by deciduous and evergreen woody 519 species in a scrub community is based on water availability and root distribution. 520 Ecohydrology 8: 538–551. 521 Ellsworth, P.Z., & Williams, D.G. 2007. Hydrogen isotope fractionation during water uptake by 522 woody xerophytes. Plant and Soil 291: 93–107. 523 Esquivias, M.P., Zunzunegui, M., Barradas, M.C.D., & Álvarez-Cansino, L. 2014. The role of water 524 use and uptake on two Mediterranean shrubs’ interaction in a brackish coastal dune ecosystem. 525 Ecohydrology 7: 783–793. 526 Farquhar, G.D., Ehleringer, J.R., & Hubick, K.T. 1989. Carbon Isotope Discrimination and 527 Photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology 40: 503– 528 537. 529 Filella, I., & Peñuelas, J. 2003. Partitioning of water and nitrogen in co-occurring Mediterranean 530 woody shrub species of different evolutionary history. Oecologia 137: 51–61. 531 Fowler, N. 1986. The Role of Competition in Plant Communities in Arid and Semiarid Regions. 532 Annual Review of Ecology and Systematics 17: 89–110. 533 Grossiord, C., Sevanto, S., Dawson, T.E., Adams, H.D., Collins, A.D., Dickman, L.T., Newman, 534 B.D., Stockton, E.A., & McDowell, N.G. 2017. Warming combined with more extreme 535 precipitation regimes modifies the water sources used by trees. New Phytologist 213: 584– 536 596. 537 Hasselquist, N.J., Allen, M.F., & Santiago, L.S. 2010. Water relations of evergreen and drought- 538 deciduous trees along a seasonally dry tropical forest chronosequence. Oecologia 164: 881– 539 890. 540 J. Wright, I., S. Falster, D., Pickup, M., & Westoby, M. 2006. Cross‐species patterns in the 541 coordination between leaf and stem traits, and their implications for plant hydraulics. 542 Jackson, P., C. Meinzer, F., Bustamante, M., Goldstein, G., Franco, A., Rundel, P., Caldas, L., Igler, 543 E., & Causin, F. 1999. Partitioning of soil water among tree species in a Brazilian Cerrado 544 ecosystem. 545 Leng, X., Cui, J., Zhang, S., Zhang, W., Liu, Y., Liu, S., & An, S. 2013. Differential water uptake 546 among plant species in humid alpine meadows. Journal of Vegetation Science 24: 138–147. 547 McDowell, N., Pockman, W.T., Allen, C.D., Breshears, D.D., Cobb, N., Kolb, T., Plaut, J., Sperry, 548 18 J., West, A., Williams, D.G., & Yepez, E.A. 2008. Mechanisms of plant survival and mortality 549 during drought: why do some plants survive while others succumb to drought? New 550 Phytologist 178: 719–739. 551 Merino, O., Villar, R., Martín, A., García, D., & Merino, J. 1995. Vegetation Response to Climatic 552 Change in a Dune Ecosystem in Southern Spain BT - Global Change and Mediterranean-Type 553 Ecosystems. In Moreno, J.M. & Oechel, W.C. (eds.), pp. 225–238. Springer New York, New 554 York, NY. 555 Moreno-Gutiérrez, C., Dawson, T.E., Nicolás, E., & Querejeta, J.I. 2012. Isotopes reveal 556 contrasting water use strategies among coexisting plant species in a Mediterranean ecosystem. 557 New Phytologist 196: 489–496. 558 Muñoz-Reinoso, J.C., & García Novo, F. 2005. Multiscale control of vegetation patterns: the case 559 of Doñana (SW Spain). Landscape Ecology 20: 51–61. 560 Nepstad, D.C., de Carvalho, C.R., Davidson, E.A., Jipp, P.H., Lefebvre, P.A., Negreiros, G.H., da 561 Silva, E.D., Stone, T.A., Trumbore, S.E., & Vieira, S. 1994. The role of deep roots in the 562 hydrological and carbon cycles of Amazonian forests and pastures. Nature 372: 666. 563 Nippert, J.B., & Knapp, A.K. 2007. Soil water partitioning contributes to species coexistence in 564 tallgrass prairie. Oikos 116: 1017–1029. 565 Oliveira, R.S., Bezerra, L., Davisdson, E.A., Pinto, F., Klink, C.A., Nepstad, D.C., & Moreira, A. 566 2005. Deep root function in soil water dynamics in cerrado savannas of central Brazil. 567 Functional Ecology 19: 574–581. 568 Padilla, F.M., & Pugmaire, F.I. 2007. Rooting depth and soil moisture control Mediterranean 569 woody seedling survival during drought. Functional Ecology 21: 489–495. 570 Palacio, S., Montserrat-Martí, G., & Ferrio, J.P. 2017. Water use segregation among plants with 571 contrasting root depth and distribution along gypsum hills. Journal of Vegetation Science 28: 572 1107–1117. 573 Parnell, A.C., Phillips, D.L., Bearhop, S., Semmens, B.X., Ward, E.J., Moore, J.W., Jackson, A.L., 574 Grey, J., Kelly, D.J., & Inger, R. 2013. Bayesian stable isotope mixing models. 575 Environmetrics 24: 387–399. 576 Peñuelas, J., Llusia, J., Pinol, J., & Filella, I. 1997. Photochemical reflectance index and leaf 577 photosynthetic radiation-use-efficiency assessment in Mediterranean trees. International 578 Journal of Remote Sensing 18: 2863–2868. 579 Pivovaroff, A.L., Pasquini, S.C., De Guzman, M.E., Alstad, K.P., Stemke, J.S., & Santiago, L.S. 580 2016. Multiple strategies for drought survival among woody plant species. Functional Ecology 581 30: 517–526. 582 19 Rossatto, D.R., de Carvalho Ramos Silva, L., Villalobos-Vega, R., Sternberg, L. da S.L., & Franco, 583 A.C. 2012. Depth of water uptake in woody plants relates to groundwater level and vegetation 584 structure along a topographic gradient in a neotropical savanna. Environmental and 585 Experimental Botany 77: 259–266. 586 Saha, A.K., Lobo O’Reilly Sternberg, L. da S., & Miralles-Wilhelm, F. 2009. Linking water sources 587 with foliar nutrient status in upland plant communities in the Everglades National Park, USA. 588 Ecohydrology 2: 42–54. 589 Sardans, J., & Peñuelas, J. 2013. Plant-soil interactions in Mediterranean forest and shrublands: 590 impacts of climatic change. Plant and Soil 365: 1–33. 591 Schenk, H.J. 2006. Root competition: beyond resource depletion. Journal of Ecology 94: 725–739. 592 Schenk, H.J., & Jackson, R.B. 2002. Rooting depths, lateral root spreads and below-ground/above- 593 ground allometries of plants in water-limited ecosystems. Journal of Ecology 90: 480–494. 594 Schwendenmann, L., Pendall, E., Sanchez-Bragado, R., Kunert, N., & Hölscher, D. 2015. Tree 595 water uptake in a tropical plantation varying in tree diversity: interspecific differences, 596 seasonal shifts and complementarity. Ecohydrology 8: 1–12. 597 Schwinning, S., & Ehleringer, J.R. 2001. Water use trade-offs and optimal adaptations to pulse- 598 driven arid ecosystems. Journal of Ecology 89: 464–480. 599 Silvertown, J., Araya, Y., & Gowing, D. 2015. Hydrological niches in terrestrial plant communities: 600 a review. Journal of Ecology 103: 93–108. 601 Silvertown, J., Dodd, M.E., Gowing, D.J.G., & Mountford, J.O. 1999. Hydrologically defined 602 niches reveal a basis for species richness in plant communities. Nature 400: 61. 603 Skelton, R.P., West, A.G., & Dawson, T.E. 2015. Predicting plant vulnerability to drought in 604 biodiverse regions using functional traits. Proceedings of the National Academy of Sciences 605 112: 5744–5749. 606 Stock, B., Jackson, A., Ward, E., & Venkiteswaran, J. 2016. MixSIAR: v3.1.7. 607 UNEP, (United Nations Environment Programme). 1992. World atlas of desertification / UNEP. 608 Edward Arnold, London ; Baltimore. 609 Valladares, F., & Sánchez-Gómez, D. 2006. Ecophysiological Traits Associated with Drought in 610 Mediterranean Tree Seedlings: Individual Responses versus Interspecific Trends in Eleven 611 Species. Plant Biology 8: 688–697. 612 Veneklaas, E.J., & Poot, P. 2003. Seasonal patterns in water use and leaf turnover of different plant 613 functional types in a species-rich woodland, south-western Australia. Plant and Soil 257: 295– 614 304. 615 Voltas, J., Lucabaugh, D., Chambel, M.R., & Ferrio, J.P. 2015. Intraspecific variation in the use of 616 20 water sources by the circum-Mediterranean conifer Pinus halepensis. New Phytologist 208: 617 1031–1041. 618 Wegener, F., Beyschlag, W., & Werner, C. 2015. High intraspecific ability to adjust both carbon 619 uptake and allocation under light and nutrient reduction in Halimium halimifolium L. 620 Frontiers in Plant Science 6: 609. 621 West, A.G., Dawson, T.E., February, E.C., Midgley, G.F., Bond, W.J., & Aston, T.L. 2012. Diverse 622 functional responses to drought in a Mediterranean-type shrubland in South Africa. New 623 Phytologist 195: 396–407. 624 West, A.G., Hultine, K.R., Burtch, K.G., & Ehleringer, J.R. 2007. Seasonal variations in moisture 625 use in a piñon–juniper woodland. Oecologia 153: 787–798. 626 West, A., Hultine, K., Jackson, T., & Ehleringer, J. 2008. Differential summer water use by Pinus 627 edulis and Juniperus osteosperma reflects contrasting hydraulic characteristics. Tree 628 physiology 27: 1711–1720. 629 Williams, D.G., & Ehleringer, J.R. 2000. Intra- and Interspecific Variation for Summer 630 Precipitation Use in Pinyon-Juniper Woodlands. Ecological Monographs 70: 517–537. 631 Zunzunegui, M., Ain-Lhout, F., Barradas, M.C.D., Álvarez-Cansino, L., Esquivias, M.P., & García 632 Novo, F. 2009. Physiological, morphological and allocation plasticity of a semi-deciduous 633 shrub. Acta Oecologica 35: 370–379. 634 Zunzunegui, M., Barradas, M.C.D., Ain-Lhout, F., Clavijo, A., & Novo, F.G. 2005. To live or to 635 survive in Doñana dunes: Adaptive responses of woody species under a Mediterranean 636 climate. Plant and Soil 273: 77–89. 637 Zunzunegui, M., Díaz Barradas, M.C., Aguilar, F., Ain Lhout, F., Clavijo, A., & García Novo, F. 638 2002. Growth response of Halimium halimifolium at four sites with different soil water 639 availability regimes in two contrasted hydrological cycles. Plant and Soil 247: 271–281. 640 Zunzunegui, M., Díaz Barradas, M.C., Ain-Lhout, F., Alvarez-Cansino, L., Esquivias, M.P., & 641 García Novo, F. 2011. Seasonal physiological plasticity and recovery capacity after summer 642 stress in Mediterranean scrub communities. Plant Ecology 212: 127–142. 643 644 645 SUPPORTING INFORMATION 646 Fig S1. Study site: sampling points and climatic conditions. 647 Fig. S2 Isotopic composition (δ18O) of xylem water of the studied species and of the water 648 sources considered at spring and summer (dry) season. 649 21 Fig S3. Principal Component Analyses considering: (a) spring water-use, (b) summer water- 650 use and (c) seasonal variation of water-use. 651 Fig S4. Functional groups classification, based on grouping of trait-categories, considering: 652 (a) spring water-use, (b) summer water-use and (c) seasonal variation of water-use. 653 Fig S5. Relationships between physiological parameters (δ13C and plant water index) and the 654 proportion of deep soil (>50 cm deep) water use (%). 655 Table S1. Plants, functional types and species sampled by plot. 656 Appendix 1. Xylem water δ18O isotopic enrichment of narrow-leaved xerophytic shrub 657 species in summer. 658 659 22 Table 1. Species, species’ family, code, number of sampling individuals in each sampling date (n) 660 and traits considered: growth form (GF), water requirements (WR), leaf longevity (LL) and leaf form 661 (LF). Classes considered were: GF – shrub (shrub) or tree (tree); WR – xerophytic (xero), mesophytic 662 (meso) or hygrophytic (hygr); LL – summer semi-deciduous (semd), evergreen (ever), aphyllous 663 (aphy); LF – narrow (narr), broadleaf (brle), needle (need), scale (scal), spike/aphyllous (aphy). 664 Classification based on Díaz Barradas et al. (1999), Muñoz-Reinoso & García Novo (2005), 665 Zunzunegui et al. (2005, 2011) and Castroviejo (2012). 666 Species Family code GF WR LL LF n Cistus libanotis Cistaceae Cl shru xero semd narr 3 Cistus salviifolius Cistaceae Cs shru hygr semd brle 3 Corema album Ericacea Ca shru xero ever narr 8 Erica scoparia Ericacea Es shru hygr ever narr 33 Halimium calycinum Cistaceae Hc shru xero semd narr 14 Halimium halimifolium Cistaceae Hh shru meso semd brle 42 Juniperus phoenicea Cupressaceae Jp tree xero ever scal 24 Lavandula stoechas Lamiaceae Ls shru xero semd brle 6 Phillyrea angustifolia Oleaceae Pa shru hygr ever brle 4 Pinus pinea Pinaceae Pp tree xero ever need 23 Quercus suber Fagaceae Qs tree hygr ever brle 5 Rosmarinus officinalis Lamiaceae Ro shru xero semd narr 9 Stauracanthus genistoides Fabaceae Sg shru xero aphy aphy 3 Ulex australis Fabaceae Ua shru xero aphy aphy 3 667 668 669 Table 2. Results of the variance analysis (ANOVA) conducted to test the effect of growth form (GF), 670 water requirement trait (WR), leaf longevity (LL) and leaf form (LF) on spring water-use, summer 671 water-use and seasonal variation of water-use. Bold represents significant effects (p<0.01). 672 Spring WU Summer WU Seasonal WU variation Effect df SS F p SS F p SS F p GF 1 215.55 173.63 <2e-16 66.68 44.50 3.4e-10 26.03 15.17 1.4e-4 WR 2 2.85 1.15 0.32 76.85 25.65 1.8e-10 38.37 11.18 2.7e-5 LL 2 6.96 2.80 0.06 27.68 9.24 1.5e-4 16.81 4.90 0.009 LF 3 8.01 2.15 0.10 29.57 6.58 3.1e-4 21.40 4.16 0.007 673 674 675 23 676 677 678 Table 3. Functional classification of the studied species based on water-use patterns. Species code as in Table 679 1. A priori functional groups (FG) defined as a concatenation of the four trait categories considered in this 680 study [Leaf form (LF), leaf longevity (LL), water requirements (WR) and growth form (GF)]. For classification 681 name, the first letter of each trait’ classes (Table 1) was used. A posteriori FGs were defined considering the 682 comparison analysis of water-use (WU) patterns between a priori FGs and respective grouping, for spring, 683 summer and WU seasonal variation (see Fig. S4). An overview of the mean water use and seasonal water 684 uptake shifts made by each a posteriori WU FG is also provided: the mean water sources used by the different 685 FGs ranged from shallow to deep soil layers as indicated by dashed arrow; seasonal water uptake shifts are 686 shown as up- or downward arrows representing the direction of water uptake changes (e.g. downward arrows 687 indicate seasonal changes to deeper soil layers). See Fig. 2 for functional groups’ quantitative water-use in 688 spring and summer and Fig. 3 for quantitative water-use seasonal variation. 689 a piori FG a posteriori FG Mean sources seasonal used WU shifts Species Spring WU Summer WU WU seasonal variation Description Cl NsXs shrub NXs NXs narrow-leaved xerophytic shrubs shallow ↑ Hc NsXs Ro NsXs Ca NeXs Cs BsHs Os Os aphyllous/spike or semi-deciduous broadleaf shrubs ↓ Ls BsXs Sg AaXs Ua AaXs Hh BsMs sMs sMs semi-deciduous mesophytic shrub ↔ Es NeHs eH eHs evergreen hygrophytic shrubs ↓↓↓ Pa BeHs Qs BeHt tree eHt evergreen hygrophytic tree ↓↓ Pp NEeXt eXt eXt evergreen xerophytic trees deep ↓↓ Jp SeXt 690 691 24 692 Fig. 1. Water sources proportion used by plants in spring and summer. Four main water sources 693 considered: soil water at 10cm, 30 and 50cm (SW10, SW30, SW50) and groundwater (GW). In 694 spring, SW10 is also representing rain water. SW30 was aggregated with SW50 (posteriorly to mixing 695 model calculation), and represented as SW30+SW50. Species grouping and functional classification 696 was done considering the significant effects of functional traits on water sources used and the 697 comparison between prior classified plant groups (see Table 3 and Fig. S4ab). (b) Spring individual 698 water source proportion used (%) by the water-use FGs identified in this season (see Table 3): trees 699 (red) and shrubs (green). Inner labels represent species name (see species code in Table 1). (c) 700 Summer individual water source proportion used by the FGs identified in this season (see Table 3). 701 Inner labels represent species name (see species code in Table 1). (c) Spring (blue) and summer (red) 702 mean of water source proportion used (%) by the FGs defined for the two seasons (unfilled circle, see 703 Table 3) and overall community (filled circle). 704 705 25 706 Fig. 2. Water-use seasonal variation (from spring to summer, %) considering: (a) species, (b) 707 functional groups and (c) overall woody community. Positive values represent higher values in 708 summer compared to spring, and negative values lower values in summer compared to spring. Water 709 sources considered were SW10, SW30, SW50 and GW (white to dark grey, as described in inner 710 legend). (a) Species represented as code name described in Table 1. For figure simplification, SW30 711 was aggregated with SW50, and represented as SW30+SW50 in this panel. Grey horizontal thick line 712 below species code indicates the species included in the grouping of (b). (b) Functional groups’ 713 definition was done considering the significant effects of functional traits on seasonal variation of 714 water use and the comparison between prior classified plant groups (see Table 3 and Fig. S4c): 715 xerophytic shrubs with narrow leaves (NXs), semi-deciduous broad-leaf or aphyllous/spiky shrubs 716 (Os), semi-deciduous mesophytic shrubs (sMs), evergreen xerophytic trees (eXt), evergreen 717 hygrophytic tree (eHt) and evergreen hygrophytic shrubs (eHs). Number of individuals considered 718 (n) represented by functional group. 719 720 721 722 723 724 725 726