This is the accepted version of the article: Ogaya Inurrigarro, Romà; Escolà, Anna; Liu, Daijun; [et al.]. «Effects of thinning in a water-limited holm oak forest». Journal of sustainable forestry, Vol. 39, Issue 4 (October 2019), p. 365-378. DOI 10.1080/10549811.2019.1673179 This version is avaible at https://ddd.uab.cat/record/224213 under the terms of the license 1 Effects of thinning in a water-limited holm oak forest. 2 3 Romà Ogaya1,2*, Anna Escolà1,2, Daijun Liu3, Adrià Barbeta4, Josep 4 Peñuelas1,2 5 6 7 1 CREAF, Cerdanyola del Vallès, E-08193, Catalonia, Spain 8 9 2 CSIC, Global Ecology Unit CREAF-CEAB-CSIC, UAB, E-08193, Cerdanyola del Vallès, 10 Catalonia, Spain. 11 12 3 School of Geography, Earth and Environmenttal Sciences, University of Birmingham 13 Edgbaston, Birmingham B15 2SQ, United Kingdom. 14 15 4 INRA, UMR ISPA, F-33140, Villenave d’Ornon, France. 16 17 18 19 20 21 *Corresponding author: Romà Ogaya, Global Ecology Unit CREAF-CEAB-CSIC, 22 Universitat Autònoma de Barcelona, E-08193, Cerdanyola del Vallès, Catalonia, Spain. 23 Tel.: +34-935814036, Fax: +34-935814151, E-mail: r.ogaya@creaf.uab.cat 24 25 1 26 Effects of thinning in a water-limited holm oak forest. 27 28 29 Abstract 30 31 A natural holm oak forest was selectively thinned to test thinning as a tool to reduce tree 32 mortality, increase productivity, and reverse the recent regression of the dominant species 33 (Quercus ilex) induced by climate change. Thinning increased aboveground productivity 34 and reduced stem mortality in this Mediterranean forest during four years just after 35 thinning, contributing to the maintenance of forest functioning under changing climatic 36 conditions. Q. ilex was the only species positively affected by the thinning: stem growth 37 increased for all stem sizes, and mortality was significantly lower in thinned plots. On the 38 contrary, mortality rates of Phillyrea latifolia and Arbutus unedo were not significantly 39 lower. Stem growth increased for P. latifolia only in the smallest stem-size class. Our 40 results highlight the suitability of selective thinning for improving the forest productivity 41 and ensuring the conservation of Mediterranean coppices. Other benefits of selective 42 thinning, such as a decrease in the risk of fire dispersion and an increase in the water 43 supply for human populations, are also discussed. 44 45 46 47 Keywords: carbon sink; climate change; forest dieback; forest management; holm oak; 48 Mediterranean forest; selective thinning; tree growth; tree mortality. 49 50 2 51 Introduction 52 53 Air temperatures have already increased globally and are projected to continue to increase 54 in the coming decades, and it has been predicted that the intensity and duration of extreme 55 weather events, such heat weaves, floods, and droughts spells will increase (IPCC, 2018; 56 Wang, Peng, Kneeshaw, Larocque, & Luo, 2012). Higher evapotranspiration rates 57 induced by warmer temperatures coupled to slight decreases in precipitation are expected 58 in many areas, such as the Mediterranean Basin, that are already subjected to seasonal 59 drought (IPCC, 2013). Robust climatic models have projected a continuous increase in 60 the severity of warming and drought in the Mediterranean Basin for the coming decades, 61 which could have severe impacts on carbon (C) sinks in forest ecosystems and alter 62 regional C budgets (Bates, Kundzewicz, Wu, & Palutikof, 2008; Dai, 2013; Reichstein et 63 al., 2013). Moreover, strong relationships between deficits in precipitation and 64 subsequent occurrences of hot extremes have been widely observed (Mueller & 65 Seneviratne, 2012), and higher frequencies of heat waves coinciding with summer 66 drought, higher evapotranspiration, and the subsequent low availability of water are also 67 expected for Mediterranean regions (Fischer & Schar, 2010). Some forested ecosystems 68 in semi-arid and Mediterranean areas are seasonally exposed to drought and may be 69 particularly vulnerable to even slight increases in water deficit, which can reduce tree 70 growth (Ogaya & Peñuelas, 2007; Barbeta, Ogaya, & Peñuelas, 2013) and increase tree 71 mortality (Breshears et al., 2005; Allen et al., 2010; Peng , Ma , Lei, & Peng, 2011; 72 Williams et al., 2013). 73 Forest ecosystems contain large stocks of C and represent important potential C 74 sinks (Pan et al., 2011). Forest ecosystems may mitigate climate change by sequestering 75 C, so the effects of forest management practices on ecosystemic C sinks need to be 3 76 assessed. Selective thinning is a common practice for improving forest health and 77 productivity (Roberts & Harrington, 2008) and accelerating forest succession (Sullivan 78 & Sullivan, 2016), and the trees remaining after selective thinning generally receive more 79 solar radiation, soil water, soil organic matter, and nutrients, which increase their 80 photosynthetic capacity (Peterson, Seiler, Nawak, Ginn, & Kreh, 1997; Tang, Chambers, 81 Guddanti , & Barmett, 1999; Tang, Qi, Xu, Misson, & Goldstein, 2005; Selig, Seiler, 82 Tyree, 2008; Chang et al., 2016), stem growth (Idol, Morales, Friday, & Scowcroft, 2018; 83 Jiménez, Navarro, Sánchez-Miranda, & Ripoll, 2019), and crown development (Sorg, 84 Urech, Mamadzhanov, & Rehnus, 2016; Valinger, Sjogren, Nord, & Cedergren, 2019). 85 Selective thinning is a common management strategy in Mediterranean forests to increase 86 tree growth and ecosystemic productivity, but selective thinning can also increase the 87 capacity of Mediterranean trees to tolerate drought conditions due to the reduction of stem 88 density and the consequent increase in water availability per stem (Cotillas, Sabaté, 89 Gracia, & Espelta, 2009; López, Gracia, Sabaté, & Keenan, 2009; Chang et al., 2016). 90 Several Mediterranean forests are mainly coppices with high stem densities due to ancient 91 management practices to obtain wood and charcoal, followed by a high number of sprouts 92 per cut stem (Ogaya & Peñuelas 2007). 93 Holm oak (Quercus ilex L.) is a widespread and dominant tree species in sub- 94 humid areas of the Mediterranean Basin. Many tall shrub species with lower growth rates 95 but a higher resistance to drought are associated with holm oak forests (Peñuelas, Filella, 96 Lloret, Siscart, & Piñol, 1998; Ogaya & Peñuelas 2003). Higher mortality rates and lower 97 seed production and seedling survival of Q. ilex under future drier conditions could 98 decrease the distribution of this dominant species and favour the associated species that 99 are more resistant to a low availability of water (Peñuelas, Lloret, & Montoya, 2001; 100 Lloret, Peñuelas, & Ogaya, 2004; Barbeta et al., 2013; Liu, Ogaya, Barbeta, Yang, & 4 101 Peñuelas, 2015). All small stems are cut during selective thinning in Mediterranean 102 forests, so only the largest trees remain, mainly Q. ilex trees. Shrub species can thus be 103 completely cut, conferring a competitive advantage to Q. ilex in addition to the decrease 104 in stem density. 105 The main objectives of this study were to test selective stem thinning as a tool to 106 compensate forest dieback and tree mortality in a typical Mediterranean forest and to 107 investigate the interaction between drivers of global change and recent practices of forest 108 management (Doblas-Miranda et al., 2015). We also compared the responses of Q. ilex 109 and coexisting species (more resistant to drought conditions) to the experimental selective 110 thinning. We tested these responses in the various species by cutting several small sprouts 111 of individual trees during the selective thinning, so the largest stems of all species 112 remained after thinning. 113 114 115 Materials and Methods 116 117 Study Site 118 The study area was on a south-facing slope (25% slope) in the Prades Mountains, 119 Catalonia, northeastern Spain (41°21’ N, 1°2’ E), at an altitude of 950 m a.s.l. The soil is 120 a Dystric Cambisol over Palaeozoic schist, ranging in depth from 35 to 100 cm. The 121 average annual temperature is 12.3 °C, and the average annual rainfall is 603 mm (data 122 from 1998-2015). Summer drought is pronounced and usually lasts for three months. The 123 vegetation is a very dense multi-stem forest (16 616 stems ha-1) dominated by Q. ilex 124 (8633 stems ha-1), Phillyrea latifolia L. (3600 stems ha-1), and Arbutus unedo L. (2200 125 stems ha-1), with an abundance of other evergreen species well adapted to dry conditions 5 126 such as Erica arborea L., Juniperus oxycedrus L., and Cistus albidus L. and occasional 127 individuals of deciduous species such as Sorbus torminalis (L.) Crantz and Acer 128 monspessulanum L. This forest has not been disturbed for 80 years, and the maximum 129 height of the dominant trees is approximately 6-10 m. An automated meteorological 130 station installed at the study site has monitored temperature, humidity, and precipitation 131 since late 1998. 132 133 Experimental Design and Tree-growth Data 134 Three 15 × 10 m plots were delimited at the same altitude along the slope. The plots were 135 thinned at the end of 2011, removing about 20% of the basal area (BA) (Table 1). Another 136 four 15 × 10 m plots from another experiment conducted at the same study site were not 137 thinned for use as controls (Ogaya & Peñuelas, 2007; Barbeta et al., 2013; Liu et al., 138 2015). Stem size was measured in the winters of 2012 and 2016 in both the thinned and 139 control plots and in late autumn 2011 in the thinned plots immediately before thinning 140 for calculating the percentage BA removed from the plots after thinning. Stem size was 141 calculated from the circumference at a height of 50 cm measured with a metric tape for 142 all living stems of all species with diameters >2 cm. Stem sizes were classified into three 143 categories: small stems (2-6 cm diameter), intermediate stems (6-10 cm diameter), and 144 large stems (>10 cm diameter). BA was calculated for each tree from its stem 145 circumference at 50 cm height, assuming a stem circumference section. The annual rate 146 of stem mortality (m) was calculated as (Sheil, Burslem, & Alder, 1995): m=1-(1-(No- 147 Nt)/No)1/t, where “No” is the number of stems at the start of the studied period, “Nt” is the 148 number of stems at the end of the studied period, and “t” is the number of years of the 149 studied period. 6 150 Aboveground biomass in the plots was estimated from the allometric relationships 151 between tree aboveground biomass (AB) and stem diameter at a height of 50 cm (D50). 152 Q. ilex and P. latifolia trees outside the plots were harvested, their circumferences at a 153 height of 50 cm were measured, and all aboveground biomass was weighed after drying 154 in an oven to a constant weight. The aboveground biomasses of Q. ilex and P. latifolia 155 were estimated from the calculated allometric relationships (lnAB=4.9+2.3 lnD50, 156 R2=0.98, n=12, P<0.001 for Q. ilex, and lnAB=4.3+2.5 lnD50, R2=0.97, n=13, P<0.001 157 for P. latifolia). The aboveground biomass of A. unedo was estimated from the allometric 158 relationship previously calculated for the same area by Lledó (1990) (lnAB=3.8+2.6 159 lnD50, R2=0.99, n=10, P<0.001). 160 161 Leaf Area Index Data 162 The biomass of leaves was also estimated using allometric relationships between the 163 biomass of leaves in each tree (BL) and its stem diameter at a height of 50 cm (D50), 164 from Q. ilex and P. latifolia trees outside the plots. Their circumferences at a height of 50 165 cm were measured, and the biomass of all leaves were weighed after drying in an oven to 166 a constant weight. The biomass of leaves in Q. ilex and P. latifolia were estimated from 167 the calculated allometric relationships (lnBL=3.5+1.7 lnD50, R2=0.91, n=12, P<0.001 for 168 Q. ilex, and lnBL=1.4+2.4 lnD50, R2=0.81, n=13, P<0.001 for P. latifolia). The biomass 169 of leaves in A. unedo was estimated from the allometric relationship previously calculated 170 for the same area in [32] (lnBL=1.9+2.2 lnD50, R2=0.95, n=10, P<0.001). For the other 171 species, their AB was estimated from Q. ilex allometric relationships for other tree species, 172 and from P. latifolia allometric relationships for other shrub species. 173 The total leaf surface was calculated assuming a mean leaf mass per area ratio of 174 20 mg cm-2 for Q. ilex leaves, and 15 mg cm-2 for P. latifolia and all other species, as it 7 175 was observed in a previous study conducted in the same site (Ogaya & Peñuelas, 2006). 176 LAI was calculated from total leaf biomass and total leaf surface data for each studied 177 species (Q. ilex, P. latifolia, and A. unedo), and a total LAI data for each plot (Table 2). 178 179 Data Analysis 180 The effect of thinning on BA during the study period (2012-2015) was tested by a general 181 linear model, with thinning, category of stem size, and species as independent factors, 182 and with the mean percentage increase in BA in each plot as a dependent variable. Post 183 hoc tests compared the effects of thinning on increases in BA for each species and stem- 184 size category. Analyses of variance (ANOVAs) were performed to test the effect of 185 thinning on stem mortality and biomass, with thinning and species as independent factors, 186 and with annual mortality rate or the percentage increase in aboveground biomass as 187 dependent variables. Data for the percentage increases in BA and aboveground biomass 188 (p) and annual mortality rate (m) were transformed to arcsin (p)1/2 and arcsin (m)1/2 to 189 satisfy the normality assumptions of the ANOVA. All analyses were performed with the 190 Statistica10 software package (StatSoft Inc., Palo Alto, CA, USA). 191 192 193 Results 194 195 Climatic Data 196 The climatic conditions during the study period (2012-2015) were typical of 197 Mediterranean areas, with dry and hot summers and moderately cold winters. The weather 198 was especially hot and dry in 2015 and especially humid and cool in 2013. Climatic 199 conditions were average in 2012 and 2014. Mean annual temperature and total annual 8 200 precipitation ranged from 11.9 °C and 783 mm in 2013 to 13.3 °C and 355 mm in 2015, 201 respectively (Fig. 1). 202 203 Basal Area and Mortality 204 Thinning significantly increased BA for Q. ilex relative to the control plots (P<0.01). In 205 contrast, thinning did not significantly affect BA for A. unedo and only tended to increase 206 BA for the smaller stems of P. latifolia (P=0.09) (Figure 2). BA increased more for small 207 than intermediate and large stems for A. unedo and P. latifolia but increased less for small 208 than intermediate and large stems for Q. ilex. BA increased most for A. unedo in the 209 control plots, whereas BA increased least for Q. ilex, especially for smaller stems, but 210 also increased more in the thinned plots for all stem sizes. BA increased more for P. 211 latifolia than Q. ilex for small stems (with also a large increase induced by thinning), but 212 the growth of intermediate stems was similar for P. latifolia and Q. ilex, with an increase 213 Annual rates of stem mortality were higher for Q. ilex than P. latifolia and for P. 214 latifolia than A. unedo in the control plots, but thinning strongly reduced the rate for Q. 215 ilex and had no significant effect on the other two species. The annual mortality rate in 216 the thinned plots was therefore even lower for Q. ilex than the other two species (Figure 217 3). Thinning significantly decreased the annual rate of stem mortality for all species 218 together relative to the control plots. 219 220 Aboveground Biomass and Leaf Area Index 221 Aboveground biomass increased more in the thinned than in the control plots (P=0.02) 222 (Figure 4). The relative aboveground biomass increased more for A. unedo than the other 223 two species in the control plots. The increase in relative aboveground biomass in the 224 thinned plots did not differ significantly between A. unedo and P. latifolia. The difference 9 225 in relative aboveground biomass increase between A. unedo and Q. ilex was lower than 226 in the control plots, because thinning increased the biomass for Q. ilex (P<0.01) and P. 227 latifolia (P<0.05) but not A. unedo (Figure 4). In terms of absolute values of aboveground 228 biomass increase, thinning treatment doubled the biomass increase of Q. ilex, while it was 229 not observed any aboveground biomass increase (in absolute values) in the two other 230 studied species induced by thinning (Figure 4). Because of that, in thinned plots the 231 aboveground biomass increment was higher in Q. ilex than in the other two species, while 232 this difference was not observed in control plots. 233 The percentage of LAI increment was higher in thinned plots than in control plots 234 (P<0.05). The thinning treatment increased LAI in Q. ilex (P<0.01) and P. latifolia 235 (P<0.05), but not in A. unedo (Figure 5). On the other hand, the percentage of LAI 236 increment in Q. ilex was smaller than in P. latifolia and A. unedo in both control and 237 thinned plots. 238 239 240 Discussion 241 242 Several studies have reported an increase in stem growth (Roberts & Harrington, 2008) 243 and tree C sequestration (Keyser, 2010) induced by selective thinning, in accordance with 244 our results of higher radial growth and biomass in selectively thinned plots, but the 245 response to thinning was species-specific. The ability of tall shrub species such as P. 246 latifolia to cope with water stress conferred a competitive advantage over Q. ilex at the 247 same study site (Ogaya & Peñuelas, 2007; Barbeta et al., 2013; Liu et al., 2015). Stem 248 growth and biomass increase strongly decreased, and stem mortality strongly increased, 249 for Q. ilex but not P. latifolia subjected to an experimental drought (Ogaya & Peñuelas, 10 250 2007; Barbeta et al., 2013; Liu et al., 2015). The decrease in water availability induced 251 by climate change (IPCC, 2013) is progressively substituting Q. ilex with tall shrub 252 species more resistant to drought, such as P. latifolia (Liu, Ogaya, Barbeta, Yang, & 253 Peñuelas, 2018). Selective thinning increased Q. ilex stem growth for all stem sizes but 254 only increased the growth of smaller P. latifolia stems and did not have a significant effect 255 on A. unedo. Selective thinning strongly decreased stem mortality for Q. ilex but not the 256 other two species. Selective thinning was thus a good strategy of forest management for 257 increasing Q. ilex tree growth but also for delaying the progressive replacement of this 258 dominant species by more drought-resistant shrub species. Similar results were observed 259 in other sites around the world, where selective thinning increased stem growth and 260 enlarged the growing season (Idol, Morales, Friday, & Scowcroft, 2018; Jiménez, 261 Navarro, Sánchez-Miranda, & Ripoll, 2019). 262 Water availability is very important in this water-limited environment, and the 263 selective thinning increased the amount of resources available to the remaining stems and 264 decreased the transpiration surface (LAI) of the forest, as observed in other thinning 265 managements conducted in other forests (Valinger, Sjogren, Nord, & Cedergren, 2019). 266 Furthermore, in another water-limited forest, it has been observed that thinning-induced 267 stem growth and crown area increases are strongly correlated (Sorg, Urech, 268 Mamadzhanov, & Rehnus, 2016). Despite of the higher LAI increase in thinned plots, at 269 the end of this experiment LAI values were still smaller in thinned plots than in control 270 plots, so the duration of the improvement of stem water availability induced by the 271 selective thinning is long. Higher net photosynthetic rates and stem growth in response to 272 thinning in semi-arid habitats may be mediated by higher water availability and stomatal 273 conductance (Moreno-Gutiérrez et al., 2011). These responses were observed in 274 temperate forests only in unusually dry years (Herbs, Mund, Tamrakar, & Knohl, 2015). 11 275 This increase in net photosynthetic rate has been associated with an increase in tree 276 growth but also with an increase in the capacity to mitigate climate change due to an 277 increase in forest atmospheric CO2 uptake. Increased warming and drought, however, will 278 likely alter the capacity of Mediterranean forests to absorb CO2, so forest management 279 may be a key factor determining the response of forest C balances to the changing climate 280 (Vayreda, Martínez-Vilalta, Gracia, & Retana, 2012). Mediterranean plants usually close 281 their stomata under severe drought conditions to avoid water loss from transpiration, but 282 stomatal closure also strongly decreases net photosynthetic rates (Mooney, Harrison, & 283 Morrow, 1975; Tenhunen et al., 1980; Llusià & Peñuelas, 2000). Moreover, the emission 284 of volatile organic compounds (VOCs), which are greenhouse gases and contribute to 285 climate change, did not decrease under drought (Llusià & Peñuelas, 2000; Peñuelas et al., 286 2013), so Mediterranean vegetation can lose the capacity to mitigate climate change under 287 severe drought conditions and can even exacerbate climate change when the effect of both 288 plant respiration and VOC emission exceed atmospheric CO2 uptake by photosynthesis 289 (Doblas-Miranda et al., 2015; Peñuelas et al., 2013). 290 Mediterranean ecosystems are currently water limited. Small mitigating 291 interventions could help to preserve their functioning under the impacts posed by 292 increased water stress (Gracia, Vanclay, Daly, Sabaté, & Gyengé, 2011) and the 293 functioning of other interconnected ecosystems (Gasith & Resh, 1999). High stem density 294 could also decrease the water supply for human populations, because vegetation 295 consumes at lot of rainwater (Birot, Gracia, & Palahí, 2011). High stem density could 296 favour the spread of forest fires due to the abundance of dead stems and branches as a 297 consequence of the strong competition for water and resources (Ogaya & Peñuelas, 2007). 298 In conclusion, our study highlights the appropriateness of selective thinning for 299 improving the growth and conservation of Mediterranean forests, better coping with the 12 300 environmental constraints posed by climate change, and increasing the functioning of 301 these ecosystems to mitigate climate change. Selective thinning can also ameliorate 302 several environmental problems typical in Mediterranean areas, such as decreasing the 303 risk and spread of forest fires and increasing the water supply for human populations. 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 13 325 Acknowledgments 326 327 We are grateful to DARP (Generalitat de Catalunya), Xavier Buqueras, and Toni Mestres 328 for their permission and assistance to conduct this research in the Poblet Holm Oak Forest 329 throughout the Fundació La Caixa project ORGEST. This research was also financially 330 supported by the Spanish government project CGL2016-79835-P-, the European 331 Research Council Synergy grant ERC-2013-SyG-2013-610028 IMBALANCE-P, and the 332 Catalan government grant SGR-2017-1005. 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 14 350 References 351 352 Allen, C.D., Macalady, A.K., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier, 353 M., Kitzberger, T., Rigling, A., Breshears, D.D., Hogg, E.H., González, P., Fensham, R., 354 Zhang, Z., Castro, J., Demidova, N., Lim, J.H., Allard, G., Running, S.W., Semerci, A., 355 & Cobb, N. (2010). A global overview of drought and heat-induced tree mortality reveals 356 emerging climate change risks for forests. Forest Ecology and Management, 259, 660- 357 684. 358 359 Barbeta, A., Ogaya, R., & Peñuelas, J. (2013). Dampening effects of long-term 360 experimental drought on growth and mortality rates of a Holm oak forest. Global Change 361 Biology, 19, 3133-3144. 362 363 Bates, B.C., Kundzewicz, Z.W., Wu, S., & Palutikof, J.P. (2008). Climate Change and 364 Water. Technical Paper of the Intergovernmental Panel on Climate Change. Geneva: 365 Intergovernmental Panel of Climate Change. 366 367 Birot, Y., Gracia, C., & Palahí, M. (2011). Water for forests and people in the 368 Mediterranean region. European Forest Institute. 369 370 Breshears, D.D., Cobb, N.S., Rich, P.M., Price, K.D., Allen, C.D., Balice, R.G., Romme, 371 W.H., Kastens, J.H., Floyd, M.L., Belnap, J., Anderson, J.J., Myers, O.B., & Meyer, C.W. 372 (2005). Regional vegetation die-off in response to global-change-type drought. 373 Proceedings of the National Academy of Sciences of the USA, 102, 15144-15148. 374 15 375 Chang, C.T., Sperlich, D., Sabaté, S., Sánchez-Costa, E., Cotillas, M., Espelta, J.M., 376 Gracia, C. (2016). Mitigating the stress of drought on soil respiration by selective 377 thinning: contrasting effects of drought on soil respiration of two oak species in a 378 Mediterranean forest. Forests, 7(263), doi:10.3390/f7110263. 379 380 Cotillas, M., Sabaté, S., Gracia, C., Espelta, J. (2009). Growth response of mixed 381 mediterranean oak coppices to rainfall reduction. Could selective thinning have any 382 influence on it? Forest Ecology and Management, 258, 1677-1683. 383 384 Dai, A.G. (2013). Increasing drought under global warming in observations and models. 385 Nature Climate Change, 3, 52-58. 386 387 Doblas-Miranda, E., Martínez-Vilalta, J., Lloret, F., Álvarez, A., Ávila, A., Bonet, F.J., 388 Brotons, L., Castro, J., Curiel Yuste, J., Díaz, M., Ferrandis, P., García-Hurtado, E., 389 Iriondo, J.M., Keenan, T.F., Latron, J., Llusià, J., Loepfe, L., Mayol, M., Moré, G., Moya, 390 D., Peñuelas, J., Pons, X., Poyatos, R., Sardans, J., Sus, O., Vallejo, V.R., Vayreda, J., & 391 Retana, J. (2015). Reassessing global change research priorities in Mediterranean 392 terrestrial ecosystems: how far have we come and where do we go from here? Global 393 Biology and Biogeography, 24, 25-43. 394 395 Fischer, E.M., Schar, C. (2010). Consistent geographical patterns of changes in high- 396 impact European heatwaves. Nature Geoscience 3, 398-403. 397 16 398 Gasith, A., Resh, V. (1999). Streams in Mediterranean climate regions: Abiotic influences 399 and biotic responses to predictable seasonal events. Annual Review of Ecology, Evolution, 400 and Systematics, 30, 51-81. 401 402 Gracia, C., Vanclay, J., Daly, H., Sabaté, S., & Gyengé, J. (2011). Agua para los bosques 403 y la sociedad en el mediterráneo - Un difícil equilibrio. [Water for forests and society in 404 the Mediterranean – A difficult equilibrium]. Barcelona: European Forest Institute. 405 406 Herbs, M., Mund, M., Tamrakar, R., & Knohl, A. (2015). Differences in carbon uptake 407 and water use between a managed and an unmanaged beech forest in central Germany. 408 Forest Ecology and Management, 355, 101-108. 409 410 Idol, T.W., Morales, R.M., Friday, J.B., & Scowcroft, P.G. (2018). Precommercial release 411 thinning of potential Acacia koa crop trees increases stem and crown growth in dense, 8- 412 year-old stands in Hawaii. Forest Ecology and Management, 392, 105-114. 413 414 IPCC. (2013). Climate change 2013: The physical Science Basis. Contribution of 415 Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on 416 Climate Change. Cambridge, UK, and New York, NY, USA: Cambridge University Press. 417 418 IPCC. (2018). IPCC 2018: Summary of Policymakers in: Global warming of 1.5ºC. An 419 IPCC Special Report on the impacts of global warming of 1.5ºC above pre-industrial 420 levels and related global greenhouse gas emission pathways, in the context of 421 strengthening the global response to the threat of climate change, sustainable 17 422 development, and efforts to eradicate poverty, Masson-Delmotte V. et al. eds, World 423 Meteorological Organization, Geneva, Switzerland, 32 pp. 424 425 Jiménez, M.N., Navarro, F.B., Sánchez-Miranda, A., Ripoll, M.A. (2019). Using stem 426 diameter variations to detect and quantify growth and relationships with climatic 427 variables on a gradient of thinned Aleppo pines. Forest Ecology and Management, 442, 428 53-62. 429 430 Keyser, T.L. (2010). Thinning and site quality influence aboveground tree carbon stocks 431 in yellow-poplar forests of the southern Appalachians. Canadian Journal of Forest 432 Research, 40, 659-667. 433 434 Liu, D., Ogaya, R., Barbeta, A., Yang X., & Peñuelas, J. (2015). Contrasting impacts of 435 continuous moderate drought and episodic severe droughts on the aboveground-biomass 436 increase and litterfall of three coexistint Mediterranean Woody species. Global Change 437 Biology, 21, 4196-4209. 438 439 Liu, D., Ogaya, R., Barbeta, A., Yang X., & Peñuelas, J. (2018). Long-term experimental 440 drought combined with natural extremes accelerate vegetation shift in a Mediterranean 441 holm oak forest. Environmental and Experimental Botany, 151, 1-11. 442 443 Lledó, M.J. (1990). Compartimentos y flujos biogeoquímicos en una cuenca de encinar 444 del Monte de Poblet. [Biogeochemical compartments and fluxes in a holm oak watershed 445 of Poblet Mountains] (Doctoral dissertation). Universitat d’Alacant, Alacant, Spain. 446 18 447 Lloret, F., Peñuelas, J., & Ogaya, R. (2004). Establishment of co-existing Mediterranean 448 tree species under a varying soil moisture regime. Journal of Vegetation Science, 15, 237- 449 244. 450 451 Llusià, J., & Peñuelas, J. (2000). Seasonal patterns of terpene content and emission from 452 seven Mediterranean woody species in field conditions. American Journal of Botany, 87, 453 133-140. 454 455 López, B., Gracia, C., Sabaté, S., & Keenan, T. (2009). Assessing the resilience of 456 Medietrranean holm oaks to disturbances using selective thinning. Acta Oecologica, 35, 457 849-854. 458 459 Mooney, H.A., Harrison, A., & Morrow, P. (1975). Environmental limitations of 460 photosynthesis on a California evergreen shrub. Oecologia, 19, 293-301. 461 462 Moreno-Gutiérrez, C., Barberà, G.G., Nicolás, E., de Luís, M., Castillo, V.M., Martínez- 463 Fernández, F., Querejeta, J.I. (2011). Leaf δ18O of remaining trees is affected by thinning 464 intensity in a semiarid pine forest. Plant Cell Environment, 34, 1009-1019. 465 466 Mueller, B., & Seneviratne, S. (2012). Hot days induced by precipitation deficits at the 467 global scale. Proceedings of the National Academy of Sciences of the USA, 109, 12398- 468 12403. 469 470 Ogaya, R., & Peñuelas, J. (2003). Comparative seasonal gas exchange and chlorophyll 471 fluorescence of two dominant woody species in a holm oak forest. Flora, 198, 132-141. 19 472 473 Ogaya, R., & Peñuelas, J. (2006). Contrasting foliar responses to drought in Quercus 474 ilex and Phillyrea latifolia. Biologia Plantarum, 50, 373-382. 475 476 Ogaya, R., & Peñuelas, J. (2007). Tree growth, mortality, and above-ground biomass 477 accumulation in a holm oak forest under a five-year experimental field drought. Plant 478 Ecology, 189, 291-299. 479 480 Pan, Y., Birdsey, R.A., Fang, J., Houghton, R., Kauppi, P.E., Kurz, W.A., Phillips, O.L., 481 Shvidenko, A., Lewis, S.L., Canadell, J.G., Ciais, P., Jackson, R.B., Pacala, S.W., 482 McGuire, A.D., Piao, S., Rautiainen, A., Sitch, S., & Hayes, D. (2011). A large and 483 persistent carbon sink in the world’s forests. Science, 333, 988-993. 484 485 Peng, C., Ma, Z., & Lei, X. (2011). A drought-induced pervasive increase in tree mortality 486 across Canada’s boreal forests. Nature Climate Change, DOI: 10.1038/nclimate1293. 487 488 Peñuelas, J., Filella, I., Lloret, F., Siscart, D., & Piñol, J. (1998). Comparative field study 489 of spring and summer leaf gas exchange and photobiology of the Mediterranean trees 490 Quercus ilex and Phillyrea latifolia. Journal of Experimental Botany, 49, 229-238. 491 492 Peñuelas, J., Lloret, F., & Montoya, R. (2001). Severe drought effects on Mediterranean 493 woody flora in Spain. Forest Science, 47, 214-218. 494 495 Peñuelas, J., Guenther, A., Rapparini, F., Llusià, J., Filella, I., Seco, R., Estiarte, M., 496 Mejía-Chang, M., Ogaya, R., Ibáñez, J., Sardans, J., Castaño, L.M., Turnipseed, A., Duhl, 20 497 T., Harley, P., Vila, J., Estavillo, J.M., Menéndez, S., Facini, O., Baraldi, R., Geron, C., 498 Mak, J., Patton, E.G., Jiang, X., & Greenberg, J. (2013). Intensive measurements of gas, 499 water, and energy exchange between vegetation and troposphere during the MONTES 500 campaign in a vegetation gradient from short semi-desertic shrublands to tall wet 501 temperate forests in the NW Mediterranean Basin. Atmospheric Environment, 75, 348- 502 364. 503 504 Peterson, J.A., Seiler, J.R., Nawak, J., Ginn, S.E., & Kreh, R.E. (1997). Growth and 505 physiological responses of young loblolly pine stands to thinning. Forest Science, 43, 506 529-534. 507 508 Reichstein, M., Bahn, M., Ciais, P., Frank, D., Mahecha, M.D., Seneviratne, S.I., 509 Zscheischler, J., Beer, C., Buchmann, N., Frank, D.C., Papale, D., Rammig, A., Smith, 510 P., Thonicke, K., van der Velde, M., Vicca, S., Walz, A., & Wattenbach, M. (2013). 511 Climate extremes and the carbon cycle. Nature, 500, 287-295. 512 513 Roberts, S.D., & Harrington, C.A. (2008). Individual tree growth response to variable 514 density-thinning in coastal Pacific Northwest forests. Forest Ecology and Management, 515 255, 2771-2781. 516 517 Sheil, D., Burslem, D.F.R.P., & Alder, D. (1995). The interpretation and misinterpretation 518 of mortality rate measurements. Journal of Ecology, 83, 331-333. 519 520 Selig, M.F., Seiler, J.R., & Tyree, M.C. (2008). Soil carbon and CO2 efflux as influenced 521 by the thinning of loblolly pine (Pinus taeda L.) plantations on the Piedmont of Virginia. 21 522 Forest Science, 54, 58-66. 523 524 Sorg, J.P., Urech, Z.L., Mamadzhanov, D., & Rehnus, M. (2016). Thinning effects on 525 walnut and crown diameter growth and fruiting in the walnut-fruit forests of Kyrgyzstan. 526 Journal of Mountain Science, 13, 1558-1566. 527 528 Sullivan, T.P., & Sullivan, D.S. (2016). Acceleration of old-growth structural attributes 529 in lodgepole pine forest: Tree growth and stand structure 25 years after thinning. Forest 530 Ecology and Management, 365, 96-106. 531 532 Tang, Z., Chambers, J.L., Guddanti, S., & Barmett, J.P. (1999). Thinning, fertilization, 533 and crown position interact to control physiological responses of loblolly pine. Tree 534 Physiology, 19, 87-94. 535 536 Tang, J., Qi, Y., Xu, M., Misson, L., & Goldstein, A.H. (2005). Forest thinning and soil 537 respiration in a ponderosa pine plantation in the Sierra Nevada. Tree Physiology, 25, 57- 538 66. 539 540 Tenhunen, J.D., Lange, O.L., Braun, M., Meyer, A., Losch, R., & Pereira, J.S. (1980). 541 Midday stomatal closure in Arbutus unedo leaves in a natural machia and under simulated 542 habitat conditions in an environmental chamber. Oecologia, 47, 365-367. 543 544 Valinger, E., Sjogren, H., Nord, G, & Cedergren, J. (2019). Effects on stem growth of 545 scots pine 33 years after thinning and/or fertilization in northern Sweden. Scandinavian 546 Journal of Forest Research, 34, 33-38. 22 547 548 Vayreda, J., Martínez-Vilalta, J., Gracia, M., & Retana, J. (2012). Recent climate changes 549 interact with stand structure and management to determine changes in tree carbon stocks 550 in Spanish forests. Global Change Biology, 18, 1028-1041. 551 552 Wang, W., Peng, C., Kneeshaw, D.D., Larocque, G.R., and Luo, Z. (2012). Drought- 553 induced tree mortality: ecological consequences, causes, and modeling. Environmental 554 Review 20, 109-121. 555 556 Williams, A.P., Allen, C.D., Macalady, A.K., Griffin, D., Woodhouse, C.A., Meko, D.M., 557 Swetnam, T.W., Rauscher, S.A., Seager, R., Grissino-Mayer, H.D., Dean, J.S., Cook, 558 E.R., Gangodagamage, C., Cai, M., & McDowell, N.G. (2013). Temperature as a potent 559 driver of regional forest drought stress and tree mortality. Nature Climate Change, 3, 292- 560 297. 561 562 563 564 565 566 567 568 569 570 571 23 572 Table 1. Basal area (BA) of the dominant species in each thinned plot immediately before 573 and after thinning. 574 Plot Species BA before thinning BA after thinning BA removed (%) (m2 ha-1) (m2 ha-1) 1 Q. ilex 43.57 38.41 11.84 P. latifolia 2.64 1.10 58.39 A. unedo 10.06 5.80 42.37 All species 56.27 45.31 19.48 2 Q. ilex 29.31 24.59 16.09 P. latifolia 2.57 1.77 31.04 A. unedo 9.75 6.43 34.02 All species 41.83 33.00 21.11 3 Q. ilex 42.14 33.86 19.64 P. latifolia 4.94 1.45 70.53 A. unedo 5.15 3.59 30.36 All species 53.71 39.53 26.39 575 576 577 578 579 580 581 582 24 583 Table 2. Leaf area index (LAI) of the dominant species in each thinned plot immediately 584 before and after thinning. 585 Plot Species LAI before thinning LAI after thinning LAI removed (%) (m2 of leaves per m2 (m2 of leaves per m2 of plot surface) of plot surface) 1 Q. ilex 4.52 3.86 14.54 P. latifolia 0.16 0.07 59.49 A. unedo 0.77 0.44 43.25 All species 5.45 4.36 19.95 2 Q. ilex 3.13 2.52 19.49 P. latifolia 0.18 0.13 28.74 A. unedo 0.75 0.48 36.55 All species 4.07 3.14 22.98 3 Q. ilex 4.69 3.54 24.54 P. latifolia 0.44 0.10 77.54 A. unedo 0.39 0.30 29.22 All species 5.64 3.97 29.61 586 587 588 589 590 591 592 25 593 Figure captions 594 595 Figure 1. Mean annual temperature and annual precipitation at the study site during the 596 study period. 597 598 Figure 2. Average percentage basal-area increase (BAI) in the thinned and control plots 599 (3 and 4 plots, respectively) for each species and stem-size category, during the overall 600 studied period. Error bars correspond to the standard error of the mean. 601 602 Figure 3. Average annual rates of stem mortality in the thinned and control plots (3 and 4 603 plots, respectively) for each species. Error bars correspond to the standard error of the 604 mean. *, P<0.05; **, P<0.01. 605 606 Figure 4. Average increases in aboveground biomass (in percentage and absolute values) 607 in the thinned and control plots (3 and 4 plots, respectively) for each species, during the 608 overall studied period. Error bars correspond to the standard error of the mean. *, P<0.05; 609 **, P<0.01. 610 611 Figure 5. Average percentage of leaf area index (LAI) increment in the thinned and 612 control plots (3 and 4 plots, respectively) for each species during the overall studied 613 period. Error bars correspond to the standard error of the mean. *, P<0.05; **, P<0.01. 614 615 616 617 26 618 619 620 621 622 623 Fig. 1. 27 624 625 626 627 628 629 Fig. 2. 28 630 631 632 633 634 635 636 637 638 639 Fig. 3. 29 640 641 642 643 644 645 Fig. 4. 30 646 647 648 649 650 651 652 653 654 655 Fig. 5. 31