Abstract

Severe droughts may increase physiological stress on long-lived woody vegetation, occasionally leading to rapid defoliation and progressive increase in mortality of overstorey trees. Over the last few years, episodes of drought-induced tree dieback have been documented in a variety of woodlands and forests around the world. However, the factors determining tree survival and subsequent recovery are still poorly understood, especially in resprouter species. We have studied the effects of a single drought episode on crown condition in a holm oak (Quercus ilex L.) forest located in NE Spain 7 years after the drought event. Generalized linear models were used to study the environmental correlates of forest crown condition 7 years after the drought event. Additionally, we evaluated the association between crown condition and the carbon and nutrient reserves stored in lignotubers 7 years after the drought. Our study reveals the multifactor nature of a drought-driven forest dieback in which soil depth and the characteristics of individual trees, particularly their number of stems, determined a complex spatial pattern of tree-level responses. This dieback was associated with a depletion of the carbon reserves in lignotubers 7 years after the episode, representing a reduction of up to 60% in highly drought-damaged trees. Interestingly, in the absence of new acute droughts, successive surveys in 2007–11 showed a direct association between carbon reserves depletion and further deterioration of crown condition. More frequent droughts, as predicted by climate change projections, may lead to a progressive depletion of carbon reserves and to a loss of resilience in Mediterranean resprouter species.

Introduction

Severe drought events associated with global climate change may increase physiological stress on long-lived woody vegetation, occasionally leading to rapid defoliation and mortality of overstorey trees (Bréda et al. 2006). Recent episodes of drought-induced tree dieback have been reported worldwide from a variety of woodland and forest communities (see review by Allen et al. 2010). Tree dieback is a complex process normally involving a wide range of potential causes (Franklin et al. 1987, Waring 1987) that some authors have framed in the context of the decline-disease theory (Manion 1991, see also Bigler et al. 2006, Galiano et al. 2010). The importance of these widespread mortality events is reinforced because they have the capacity to transform the stand structure and dynamics, and the ecosystem functioning of regional landscapes on a sub-decadal timescale (McDowell et al. 2008).

In Mediterranean-type climates, drought has been recognized as the main factor limiting plant species growth and distribution (Mooney 1983, Terradas and Savé 1992). In the Mediterranean Basin, forest ecosystems may be particularly challenged if drought periods become even more frequent and intense (Borghetti et al. 1998, Martínez-Vilalta and Piñol 2002, Ogaya et al. 2003, Sarris et al. 2007), as predicted by climate change projections (Christensen et al. 2007, IPCC 2007). In recent times, coupled to the negative consequences of increased drought, changes in human use of many Mediterranean forests have resulted in denser stands due to agricultural land abandonment (Poyatos et al. 2003, Romero-Calcerrada and Perry 2004), artificial afforestation (Martínez-García 1999, Navarro et al. 2010) and a decline in logging practices (Terradas 1999, Linares et al. 2009, 2010). Denser stands normally result in stronger plant competition for resources because of the corresponding reduction in soil water availability per tree, potentially exacerbating the vulnerability of forests to water stress (Bigler et al. 2006, Linares et al. 2009, Vilà-Cabrera et al. 2011).

Mediterranean-type ecosystems have been documented as being highly resilient to disturbances (Malanson and Trabaud 1987). The resprouting ability adopted by many Mediterranean trees and shrubs to recover after disturbance (Canadell and Zedler 1995) is a key element of the resilience of these communities (sensu Westman 1986). Many Mediterranean woody species have a large lignotuber—a woody swollen structure at the stem base (James 1984)—that acts as a reservoir of dormant buds, carbohydrates and nutrients to ensure rapid regrowth after disturbances (Mullette and Bamber 1978, Canadell and Zedler 1995). However, severe disturbances may produce loss of resilience by depletion of reserves in surviving organs, particularly when new disturbances occur before the reserves are fully recovered (Jones and Laude 1960, Rundel et al. 1987, Malanson and Trabaud 1988). In fact, a progressive loss of resilience produced by recurrent disturbances in Mediterranean ecosystems has already been reported (Díaz-Delgado et al. 2002, Lloret et al. 2004). Although several studies have examined the dependence of resilience on the replenishment rates of carbon reserves after experimental disturbances (e.g., logging; Canadell and López-Soria 1998, López et al. 2009), less is known about its response to drought under natural conditions.

The evergreen holm oak (Quercus ilex L.) is a classical example of a Mediterranean species regenerating by resprouting from lignotubers after disturbance. Although the largest populations of this species occur in the western part of the Basin (Barbéro et al. 1992), it is present over a large area extending 6000 × 1500 km, from Portugal to Syria and from Morocco and Algeria to France (Terradas 1999). Holm oak is considered a typical Mediterranean drought-resistant tree species (Canadell and Rodà 1991, Savé et al. 1999, Infante et al. 2003). Nevertheless, holm oak trees have been observed to experience relatively low leaf water potentials (Limousin et al. 2009), high losses of xylem conductivity (Tognetti et al. 1998, Martínez-Vilalta et al. 2003) and negative carbon balances (Gracia et al. 2001) under extreme water stress conditions. In fact, drought-related holm oak dieback has already been reported in the Mediterranean Basin (Camarero et al. 2004), showing that this species is more vulnerable to drought than other co-occurring species (Lloret and Siscart 1995, Peñuelas et al. 2000, 2001, Martínez-Vilalta et al. 2002, Lorenz et al. 2006).

In this study, we analyze the effects of a single drought episode on crown condition in a holm oak forest located in NE Spain 7 years after the drought event. The studied impacts may include prolonged effects that appeared during the drought episode but persisted for several years or delayed effects that appeared several years after the episode as a result of physiological disorders induced by the past drought (Bréda et al. 2006, Galiano et al. 2011). The main objectives were to: (i) examine the environmental correlates of forest crown condition emphasizing the role of forest intrinsic factors such as topographic conditions, stand structure and soil depth; (ii) evaluate whether the amount of carbon and nutrient reserves stored in lignotubers as well as the physiological performance of current leaves were related to the crown condition of trees; and (iii) explore whether the amount of carbon and nutrients stored in lignotubers may determine the changes of crown condition 4 years after the initial sampling. In relation to the two latter objectives we hypothesized that, 7 years after the drought episode, stored carbon reserves would not be completely recovered in highly damaged trees and so we predict that those trees would be less able to cope with new drought events.

Materials and methods

Study site

The study was carried out in a holm oak (Quercus ilex L.) forest located in NE Spain (SW Garrotxa region, 42°8′43″ N, 2°27′41″ E, ∼92 km2) mainly on eastern, southern and western slopes and distributed at altitudes from 500 to 900 m above sea level. Holm oak is the dominant tree species at the study site, but a number of other typically Mediterranean species occasionally appear in the understorey (Pistacia lentiscus L., Juniperus oxycedrus L., Arbutus unedo L., Viburnum tinus L., Phillyrea latifolia L., Erica arborea L.). The shrub layer is predominantly occupied by Buxus sempervirens L., Rosmarinus officinalis L. and Cistus spp. L. This forest was intensively coppiced for charcoal and firewood production until the late 1970s (Agelet and Montserrat 2002). According to López et al. (2009) we expect that the lapse since that date has been long enough to allow the recovery of carbon reserves by the year of the studied drought episode (2000). This type of forest shows extremely high stem densities due to natural regeneration from stump resprouting (Ibàñez et al. 1999), while natural seedling recruitment is typically very low (Espelta et al. 1995). Northern slopes and valley bottoms in the area are occupied by Euro-Siberian species such as Fagus sylvatica L., Corylus avellana L., Betula pendula Roth., Quercus pubescens Mill. and Quercus robur L. The main bedrock type is calcareous conglomerate with abundant rocky outcrops distributed all over the area.

The climate of the region is characterized by a mean annual temperature of 12.8 °C and a mean annual rainfall of around 1070 mm (climate data for the period 1951–99 from the Climatic Digital Atlas of Catalonia (CDAC; Pons 1996, Ninyerola et al. 2000)) corresponding to the temperate oceanic sub-Mediterranean bioclimatic region (Worldwide Bioclimatic Classification System 1996–2009). In summer 2000, the study area experienced a severe drought episode with an average rainfall from June to August of 125 mm (see Supplementary Figure S1 available as Supplementary Data at Tree Physiology Online). In August 2000, in addition to the rainfall scarcity and high temperatures, there was an intense regime of warm westerly winds (Agelet and Montserrat 2002). Although noticeable summer drought develops every year in the study area, climate records showed the absence of new extreme drought events since the 2000 drought episode until the sampling dates in November 2007 and March 2011, when the average rainfall from June to August was 205 and 295 (previous year 2010), respectively (see Supplementary Figure S1 available as Supplementary Data at Tree Physiology Online). In a preliminary analysis, the Catalan Forest Service associated the 2000 drought with increased canopy foliage dying and browning on around 20% of the study forest area (Agelet and Montserrat 2002). They distinguished four levels of canopy browning at the patch scale (see Figure 1 and Table 1): high (>60% of holm oaks damaged), medium (30–60% of holm oaks damaged), low (1 to <30% of holm oaks damaged) and no canopy browning (0% of holm oaks damaged). High, medium and low canopy browning patches represented 42, 40 and 18% of the damaged area, respectively.

Figure 1.

Location map of the study area showing sites of plots (white dots). Areas with high, medium, low and no canopy browning (see Materials and methods: Study site) are depicted with different gray shades within the distribution of holm oak in the area.

Figure 1.

Location map of the study area showing sites of plots (white dots). Areas with high, medium, low and no canopy browning (see Materials and methods: Study site) are depicted with different gray shades within the distribution of holm oak in the area.

Table 1.

Topographic and structural attributes of the studied forest at the patch and plot level, split by different levels of canopy browning.

 2000 Crown transparency
 
 Control <30% 30–60% >60% 
Local level  N = 26 N = 63 N = 33 
 Altitude (m) — 644 (16) 641 (11) 682 (16) 
 Aspect (°) — 193 (10) 188 (8) 163 (11) 
 Slope (°) — 22 (1) 22 (1) 21 (1) 
Plot level N = 5 N = 5  N = 6 
 Stand density (ind/ha) 1647 (113) 1037 (77) — 910 (192) 
 Basal area (m2/ha) 34.6 (5.0) 21.7 (2.1) — 17.5 (3.3) 
 Tree height (m) 7.2 (0.5) 5.9 (0.4) — 4.8 (0.3) 
 Stems per tree 2.5 (0.3) 3.2 (0.4) — 3.4 (0.2) 
 Soil depth (cm) 35.7 (1.2) 29.4 (2.7) — 17.7 (1.4) 
 Dead individuals (D, %) 0.9 (0.9) 5.9 (3.7) — 10.2 (5.6) 
 Crown sprouters (CS, %) 15.0 (6.3) 26.8 (3.7) — 48.1 (4.6) 
 Lignotuber sprouters (LS, %) 0.9 (0.9) — 15.5 (1.5) 
 Number of sampled trees 25 65 — 83 
 2000 Crown transparency
 
 Control <30% 30–60% >60% 
Local level  N = 26 N = 63 N = 33 
 Altitude (m) — 644 (16) 641 (11) 682 (16) 
 Aspect (°) — 193 (10) 188 (8) 163 (11) 
 Slope (°) — 22 (1) 22 (1) 21 (1) 
Plot level N = 5 N = 5  N = 6 
 Stand density (ind/ha) 1647 (113) 1037 (77) — 910 (192) 
 Basal area (m2/ha) 34.6 (5.0) 21.7 (2.1) — 17.5 (3.3) 
 Tree height (m) 7.2 (0.5) 5.9 (0.4) — 4.8 (0.3) 
 Stems per tree 2.5 (0.3) 3.2 (0.4) — 3.4 (0.2) 
 Soil depth (cm) 35.7 (1.2) 29.4 (2.7) — 17.7 (1.4) 
 Dead individuals (D, %) 0.9 (0.9) 5.9 (3.7) — 10.2 (5.6) 
 Crown sprouters (CS, %) 15.0 (6.3) 26.8 (3.7) — 48.1 (4.6) 
 Lignotuber sprouters (LS, %) 0.9 (0.9) — 15.5 (1.5) 
 Number of sampled trees 25 65 — 83 

Field sampling methods

In November 2007, 16 uneven-sized plots were surveyed along a transect using a nearest-neighbor method based on the random pairs technique (Cottam and Curtis 1949). In this technique, distances among the nearest neighbors located at the same and the opposite site of transect are recorded to define the sampling area. Eleven plots were surveyed within the damaged stands previously delimited by the Catalan Forest Service (see Study site section and Figure 1: n = 6, 5 and 5 plots in the bins of high canopy browning (>60%), low browning (1 to <30%) and no browning or control areas, respectively). Distances between plots were always >50 m, and sampled areas were selected so that they showed no signs of recent management or disturbances other than the studied drought. For all plots, total soil depth (measured with a metal stick from the beginning of the mineral soil to the bedrock with a precision of 1 cm) and topographic characteristics such as slope, aspect and altitude were recorded as correlates of water availability in the soil (Western et al. 2002).

As a result of coppicing for charcoal production, holm oak trees usually exhibit a multi-stemmed structure. Accordingly, in each plot we sampled all stems (at least 2 cm diameter) of all adult individuals. For each stem, we considered four categories according to the drought effects on crown condition: ND, no dieback or completely recovered; CS, branches' dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters; D, total crown mortality without signs of resprouting (see Figure 2). For each tree, we measured diameter at breast height (DBH), basal area (from DBH), number of stems, height of the tallest stem and we performed a whole characterization of crown condition at the tree level (ND, CS, LS or D categories as previously described for stems). In multi-stemmed individuals, tree crown condition was determined by the state of the size-dominant living stem (ND, CS or LS). A tree was considered dead (D) if all its stems were completely dry and without signs of resprouting. Note that due to the time lapsed since the drought episode this record does not allow assessing the immediate impact of the 2000 drought. We could not discern, for instance, whether ND stems had been damaged in 2000 and had recovered completely by 2007, or they simply had not been damaged. Stand density and total basal area were calculated at the plot level.

Figure 2.

Photographs of dead (D) and living holm oak individuals, including three categories of crown condition: ND, no dieback or completely recovered; CS, branches’ dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters (see more details in Materials and methods: Field sampling methods).

Figure 2.

Photographs of dead (D) and living holm oak individuals, including three categories of crown condition: ND, no dieback or completely recovered; CS, branches’ dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters (see more details in Materials and methods: Field sampling methods).

For all plots, three to five individuals of each tree crown condition category (only living trees: ND, CS and LS) were excavated by digging around their lignotuber to a depth of 20 cm, obtaining ∼1 cm-thick samples of the external layer of the lignotuber. Previous assays with Lugol solution indicated that most nonstructural carbohydrates (NSC) are stored in this external layer, which correspond to the sapwood area. These samples were used to analyze NSC and nutrients (see the ‘Lignotubers’ NSC and nutrients' section). Overall, we sampled 173 individuals, with an unbalanced design because the CS and LS categories were scarce in low- and no-canopy browning patches. We also collected 20–30 exposed and apparently healthy current-year leaves from two mid-canopy branches of each sampled individual to analyze their carbon isotope composition and their nitrogen content (see the ‘Foliar carbon isotope and nitrogen content’ section). Sampling of both lignotuber and leaves took place late at the end of the growing season (November) when pool reserves are expected to reflect the net balance between sinks and sources for the complete growing season, and have been shown to be relatively stable (Sabaté et al. 1995, Hoch et al. 2003). For the analysis of carbon isotope composition, leaves were collected on the same date to obtain integrated information of stomatal conductance and C assimilation for the whole growing season (Farquhar et al. 1989). In addition, in March 2011, tree crown condition was recorded again for all sampled individuals except for 15 individuals that could not be relocated.

Lignotubers' NSC and nutrients

Lignotuber samples were transported in a cooler over ice until sample processing in the laboratory on the same day. Once at the laboratory, the samples were microwaved for 90 s to stop enzymatic activity, oven-dried for 72 h at 65 °C and ground to fine powder. Nonstructural carbohydrates were defined as free sugars (glucose and fructose), low molecular weight sugars (free sugars and sucrose) plus starch, and were analyzed following the procedures described by Hoch et al. (2002), with some minor modifications. Sapwood powder (∼12–14 mg) was extracted with 1.6 ml distilled water at 100 °C for 60 min. After centrifugation, an aliquot of the extract was used for the determination of low molecular weight sugars after enzymatic (invertase from Saccharomyces cerevisiae and glucose hexokinase (GHK) assay reagent, I4504 and G3293, Sigma-Aldrich, Spain) conversion of sucrose and fructose into glucose. Another aliquot was incubated with an amyloglucosidase from Aspergillus niger (10115 Sigma-Aldrich) at 50 °C overnight, to break down all NSC (starch included) to glucose. The concentration of free glucose was determined photometrically in a 96-well microplate reader (Sunrise™ Basic Tecan, Männedorf, Switzerland) after enzymatic (GHK assay reagent) conversion of glucose to gluconate-6-phosphate. The dehydrogenation of glucose causes an increase in optical density at 340 nm. Starch was calculated as total NSC minus low-molecular-weight sugars. All NSC values are expressed as percent dry matter. Nutrient concentrations were analyzed at the Chemistry Laboratory (SAQ) at the Autonomous University of Barcelona, by combustion gas chromatography using a CHNS-O Elemental Analyzer Euro EA3000 (EuroVector SpA, Milan, Italy) for determining total nitrogen, and by inductive coupled plasma (ICP-OES) spectrometry Optima 4300 analysis (PerkinElmer Inc., Waltham, MA, USA) after pressure digestion with HNO3 for determining total phosphorus.

Foliar carbon isotope and nitrogen content

Current-year leaves were oven-dried for 72 h at 65 °C and ground to fine powder. Ground samples were analyzed for carbon stable isotope composition and nitrogen content at the Cornell Isotope Laboratory at Cornell University, using a Thermo Delta V isotope ratio mass spectrometer interfaced to an NC2500 elemental analyzer. The carbon stable isotope composition was expressed in delta notation: δ13C (‰) = (Rsample/Rstandard – 1) × 1000, where Rsample is the 13C/12C ratio of the sample and Rstandard is the 13C/12C ratio of the international Vienna Pee Dee Belemnite carbon standard. Precision (standard deviation = 0.10‰) and accuracy (0.05‰) were determined by analysis of an in-house isotope standard (‘BCBG’), which is calibrated against IAEA standards. Nitrogen concentrations are expressed as percent dry matter.

Statistical analyses

Generalized linear mixed models were used to study the environmental correlates of tree mortality, stem mortality and tree crown condition at the individual level (N = 368) with plot effects modeled as a random factor to account for the spatial autocorrelation among individuals within plots. Tree mortality was binomially distributed. For stem mortality and tree crown condition, the number of standing dead stems and the number of damaged stems (crown condition CS, LS or D) in each tree were considered as count response variables. The logarithm of the total number of stems per tree was introduced into both models as an explanatory variable, and errors were assumed to follow a Poisson distribution which is effectively similar to (but more flexible than) modeling the proportion of dead or damaged stems per tree (Faraway 2006). We conducted additional analyses using a more restrictive definition of damaged tree crown condition considering only those trees exhibiting stems with total crown mortality (D or LS categories), obtaining very similar results (not shown).

All models (tree mortality, stem mortality and tree crown condition) were identical in terms of predictor variables and included tree basal area (ln transformed), stems per tree (ln transformed), plot basal area, plot soil depth, plot aspect and plot altitude. All predictor variables were normally distributed or could be normalized using standard transformations except for stems per tree, which remained slightly normal after logarithmic transformation. Aspect was normally distributed, as the measured variable covered a relatively narrow range between 150 and 250° across all plots. We included the interaction between ln(stems per tree) and plot basal area because competition between stems within the same individual may be influenced by competition at the stand level, and the interaction between plot basal area and plot soil depth because competition at the stand level may be more intense on shallow soils. Some variables were not introduced into the models because they were highly correlated to other predictor variables: tree height was correlated to plot soil depth (r = 0.801, P < 0.001), and plot density of individuals was correlated to plot basal area (r = 0.872, P < 0.001). The residuals of all models were normally distributed. A Moran's test (Cliff and Ord 1981) of the residuals indicated that it was not necessary to further correct for the spatial structure of the data. Parameters (β) of all fitted models were estimated using maximum likelihood methods, and model selection was performed using Akaike's information criterion (AIC).

Additional analyses were carried out using analysis of variance (one-way ANOVA using post hoc multiple comparisons) to compare the amounts of carbon and nutrient reserves stored in lignotubers and the physiological performance of leaves among trees with different crown conditions. The relationship between NSC stored in lignotubers in 2007 and changes in crown condition of trees between 2007 and 2011 was analyzed using two-way ANOVA tests, with the tree crown condition in 2007 (ND, CS or LS) and the qualitative change of the tree crown condition between 2007 and 2011 (improved, equal or worsened) as the independent factors. Note that an improvement for CS and LS categories implies that new resprouts surpass the initial height of the tree. In contrast, a worsened change implies that ND individuals become CS, LS or D, CS individuals become LS or D, and LS individuals become D. We also used Pearson and Spearman correlation coefficients as a measure of association between pairs of variables. All statistical analyses were carried out with R version 2.12.0 (R Development Core Team 2010, Vienna, Austria).

Results

Determinants of crown condition

The proportion of trees exhibiting some degree of crown damage (CS, LS or D individuals) varied with the level of canopy browning in the stand, being 15.9, 33.6 and 73.8% in no, low and high canopy browning patches, respectively (Table 1; see also Figure 1). At the tree level, mortality was mostly influenced by the number of stems per tree and, to a lesser extent, by tree basal area: larger number of stems per tree and higher basal area were associated with higher probabilities to survive (Table 2a). Some plot attributes, such as soil depth and altitude, were also significant in the mortality model. Increases in soil depth were associated with higher survival probabilities, while location at higher altitude was associated with higher mortality. Interactions among explanatory variables were not significant and were removed from the final model because including them worsened model fit in terms of AIC.

Table 2.

Generalized linear mixed models for tree mortality (a), stem mortality (b) and tree crown condition (c) at the individual level (N = 368).

 β SE z P 
(a) Tree mortality 
 Intercept −11.737 6.369 −1.843 0.065 
 Ln[Tree basal area (cm2)] −0.659 0.305 −2.156 0.031 
 Ln[Stems per tree] −1.387 0.525 −2.644 0.008 
 Plot basal area (m2/ha) 0.003 0.045 0.074 0.940 
 Plot soil depth (cm) −0.172 0.060 −2.851 0.004 
 Plot aspect (°) 0.022 0.014 1.531 0.126 
 Plot altitude (m) 0.018 0.007 2.621 0.008 
(b) Stem mortality 
 Intercept −5.853 2.909 −2.012 0.044 
 Ln[Tree basal area (cm2)] −0.399 0.117 −3.390 <0.001 
 Ln[Stems per tree] 0.727 0.301 2.415 0.015 
 Plot basal area (m2/ha) −0.013 0.026 −0.501 0.616 
 Plot soil depth (cm) −0.090 0.026 −3.493 <0.001 
 Plot aspect (°) 0.014 0.006 2.470 0.013 
 Plot altitude (m) 0.007 0.003 2.070 0.038 
 Ln[Stems per tree] * plot basal area 0.030 0.012 2.471 0.013 
(c) Tree crown condition 
 Intercept −5.212 2.939 −1.773 0.076 
 Ln[Tree basal area (cm2)] −0.087 0.073 −1.188 0.235 
 Ln[Stems per tree] 1.063 0.097 10.872 <0.001 
 Plot basal area (m2/ha) 0.007 0.020 0.364 0.715 
 Plot soil depth (cm) −0.096 0.026 −3.619 <0.001 
 Plot aspect (°) 0.013 0.006 2.101 0.035 
 Plot altitude (m) 0.006 0.003 1.681 0.092 
 β SE z P 
(a) Tree mortality 
 Intercept −11.737 6.369 −1.843 0.065 
 Ln[Tree basal area (cm2)] −0.659 0.305 −2.156 0.031 
 Ln[Stems per tree] −1.387 0.525 −2.644 0.008 
 Plot basal area (m2/ha) 0.003 0.045 0.074 0.940 
 Plot soil depth (cm) −0.172 0.060 −2.851 0.004 
 Plot aspect (°) 0.022 0.014 1.531 0.126 
 Plot altitude (m) 0.018 0.007 2.621 0.008 
(b) Stem mortality 
 Intercept −5.853 2.909 −2.012 0.044 
 Ln[Tree basal area (cm2)] −0.399 0.117 −3.390 <0.001 
 Ln[Stems per tree] 0.727 0.301 2.415 0.015 
 Plot basal area (m2/ha) −0.013 0.026 −0.501 0.616 
 Plot soil depth (cm) −0.090 0.026 −3.493 <0.001 
 Plot aspect (°) 0.014 0.006 2.470 0.013 
 Plot altitude (m) 0.007 0.003 2.070 0.038 
 Ln[Stems per tree] * plot basal area 0.030 0.012 2.471 0.013 
(c) Tree crown condition 
 Intercept −5.212 2.939 −1.773 0.076 
 Ln[Tree basal area (cm2)] −0.087 0.073 −1.188 0.235 
 Ln[Stems per tree] 1.063 0.097 10.872 <0.001 
 Plot basal area (m2/ha) 0.007 0.020 0.364 0.715 
 Plot soil depth (cm) −0.096 0.026 −3.619 <0.001 
 Plot aspect (°) 0.013 0.006 2.101 0.035 
 Plot altitude (m) 0.006 0.003 1.681 0.092 

The number of damaged stems (tree crown condition model; Table 2c) and the number of dead stems (stem mortality model; Table 2b) were associated negatively with soil depth and positively with aspect (higher damage was found on western aspects). Locations at higher altitudes and smaller tree basal areas were also associated with higher mortality of stems. The number of damaged stems increased proportionally with the number of stems per tree (β ≈ 1, Table 2c). Interestingly, in the stem mortality model, the interaction between the number of stems per tree and plot basal area was significant, indicating that the effect of the total number of stems on stem mortality becomes larger as plot basal area increases (from β ≈ 1 at the lowest value of plot basal area, to β ≈ 2.5 at the highest basal area; see Table 2b and Supplementary Figure S2 available as Supplementary Data at Tree Physiology Online).

Carbon and nutrient reserves

The most damaged living trees (LS) presented lower current water-use efficiencies estimated from foliar δ13C than trees resprouting from the crown (CS), while control trees with no dieback (ND) showed intermediate values (Table 3 and Figure 3a). Leaf N content did not differ among the three crown condition categories (Table 3 and Figure 3b) and it was unrelated to leaf δ13C (r = 0.026, P = 0.731).

Figure 3.

Leaf δ13C (a) and leaf N (b) measured in 2007 as a function of tree crown condition: ND, no dieback or completely recovered; CS, branches' dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters (see more details in Materials and methods: Field sampling methods). Different letters indicate statistically significant differences (one-way ANOVA test using post hoc multiple comparisons). Error bars show standard errors. The number of trees (N) for each group is also shown.

Figure 3.

Leaf δ13C (a) and leaf N (b) measured in 2007 as a function of tree crown condition: ND, no dieback or completely recovered; CS, branches' dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters (see more details in Materials and methods: Field sampling methods). Different letters indicate statistically significant differences (one-way ANOVA test using post hoc multiple comparisons). Error bars show standard errors. The number of trees (N) for each group is also shown.

Table 3.

One-way ANOVA tests of the effects of tree crown condition (ND, no dieback or completely recovered; CS, branches' dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters; see more details in Materials and methods: Field sampling methods) on carbon isotopic composition and N concentration of leaves and the amounts of carbon and nutrient reserves (N and P) stored in lignotubers.

  Leaf δ13C (‰)
 
Leaf N (%)
 
Lignotuber NSC (%)
 
Lignotuber N (%)
 
Lignotuber P (%)
 
Crown condition d.f. F P F P F P F P F P 
 2, 170 5.378 0.005 1.491 0.228 65.915 <0.001 3.572 0.03 4.574 0.012 
  Leaf δ13C (‰)
 
Leaf N (%)
 
Lignotuber NSC (%)
 
Lignotuber N (%)
 
Lignotuber P (%)
 
Crown condition d.f. F P F P F P F P F P 
 2, 170 5.378 0.005 1.491 0.228 65.915 <0.001 3.572 0.03 4.574 0.012 

NSC concentrations stored in lignotubers were depleted by 26 and 60% in CS and LS individuals, respectively, relative to healthy (ND) trees, which were considered as controls (Table 3 and Figure 4a). The basal area of trees was not related to NSC concentrations (r = 0.091, P = 0.233) and, thus, differences in NSC concentrations among the three categories of crown condition were not caused by differences in the tree basal area. Interestingly, the amount of carbon reserves analyzed in 2007 was positively associated with the improvement of crown condition between years 2007 and 2011 (two-way ANOVA F = 14.845, d.f. = 2, P < 0.001; Figure 5). Regarding the nutrients stored in lignotubers, in contrast to the case of NSC concentrations, LS individuals contained larger amounts of N and P than ND and CS individuals, although significant differences only occur between CS and LS individuals for both nutrients (Table 3 and Figure 4b and c).

Figure 4.

Nonstructural carbohydrate concentrations (a), N content (b) and P content (c) stored in lignotubers of trees in 2007 as a function of tree crown condition: ND, no dieback or completely recovered; CS, branches' dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters (see more details in Materials and methods: Field sampling methods). Different letters indicate statistically significant differences (one-way ANOVA test using post hoc multiple comparisons). Error bars show standard errors. The number of trees (N) for each group is also shown.

Figure 4.

Nonstructural carbohydrate concentrations (a), N content (b) and P content (c) stored in lignotubers of trees in 2007 as a function of tree crown condition: ND, no dieback or completely recovered; CS, branches' dieback and crown sprouters; LS, total crown mortality and lignotuber sprouters (see more details in Materials and methods: Field sampling methods). Different letters indicate statistically significant differences (one-way ANOVA test using post hoc multiple comparisons). Error bars show standard errors. The number of trees (N) for each group is also shown.

Figure 5.

Nonstructural carbohydrate concentrations stored in lignotubers of trees in 2007 as a function of the qualitative change of crown condition occurring in trees between 2007 and 2010 (improved, equal or worsened; see Materials and methods: Statistical analyses). Broken lines segregate groups of trees according to the crown condition in 2007 (ND, CS or LS; see more details in Materials and methods: Field sampling methods). Different letters indicate statistically significant differences between groups of trees with different qualitative change of crown condition within the previous groups (two-way ANOVA test using post hoc multiple comparisons). Error bars show standard errors. The number of trees (N) for each group is also shown.

Figure 5.

Nonstructural carbohydrate concentrations stored in lignotubers of trees in 2007 as a function of the qualitative change of crown condition occurring in trees between 2007 and 2010 (improved, equal or worsened; see Materials and methods: Statistical analyses). Broken lines segregate groups of trees according to the crown condition in 2007 (ND, CS or LS; see more details in Materials and methods: Field sampling methods). Different letters indicate statistically significant differences between groups of trees with different qualitative change of crown condition within the previous groups (two-way ANOVA test using post hoc multiple comparisons). Error bars show standard errors. The number of trees (N) for each group is also shown.

Discussion

Determinants of crown condition

We report the effects of an extreme drought on a holm oak forest seven years after the episode. In some stands, up to 73.8% of trees still showed some level of crown damage as a result of the stem and branch mortality induced by the past drought. The patchy pattern of damage observed in the studied area suggests that other factors operating at a local or micro-local scale (in our case, soil properties, microtopography and stand structure likely related to past management) were involved in predisposing trees to be more susceptible to climatic drought (Manion 1991, see also Bigler et al. 2006, Galiano et al. 2010). Although precise records of soil water availability are not available, soil depth is particularly relevant controlling water availability at the plot level in the studied forests, with shallow soils that develop over conglomerates. This bedrock impedes the growth of roots into deep layers (Lloret et al. 2004) and provides low water supply to plants, particularly under low-rainfall conditions (cf. Peñuelas et al. 2000, also for Q. ilex; see also López-Soria and Castell 1992). Nevertheless, location at higher altitudes also determined the ultimate mortality of trees in our case (Table 2). In the study area, holm oak forests occupy a relatively narrow range of altitudes, and the effect of exposure to wind and direct solar radiation on the ridge tops apparently prevails over the effect of cooler temperatures at higher altitudes.

Mortality tends to be higher in small-sized individuals (smaller basal areas), likely due to their more superficial rooting system and to their smaller bud bank and resprouting ability (Trabaud 1987, Obón 1997, Pugnaire et al. 2000, cf. Lloret et al. 2004, also for Q. ilex). In contrast, increasing number of stems per tree implies higher drought survival. In general, disturbances trigger the growth of the meristematic tissue of lignotubers, which are further enlarged by the fusion of the stem bases of the new emerging shoots (Canadell et al. 1999). Enlarged lignotubers provide large amounts of resources that will ensure at least the survival of one of the stems and, thus, the perpetuation of the individual.

Water availability per unit of basal area is normally lower in areas with higher basal area (cf. Callaway and Walker 1997, Briones et al. 1998). The relatively high stem density in our particular holm oak forest is due to the reduction of coppicing since the late 1970s (Agelet and Montserrat 2002), as has been also observed in other areas (Barbéro et al. 1992, Terradas 1999). In our case, competition at stand level seems not to be intensified by lower soil water availability, as suggested by the absence of a significant interaction between the effects of plot basal area and soil depth. However, competition among stems within the same individual appears to be influenced by the total basal area of the stand, as has long been observed in other holm oak forests (Retana et al. 1992, Gracia et al. 1999). In these forests, intense competition normally causes very low growth rates that may lead to an almost permanent state of stagnation (Gracia et al. 1999). Extreme dry conditions that occurred in summer 2000 would have triggered self-thinning within individuals that resulted in a release of competition. Our results suggest that thinning could reduce competition and improve the resistance of holm oak forests to climate change-type droughts (see also Cotillas et al. 2009, Sánchez-Humanes and Espelta 2011, Rodríguez-Calcerrada et al. 2011), although the relationship between short-term stimulation of growth induced by thinning (López-Soria and Castell 1992, Retana et al. 1992) and carbon storage merits further investigation (cf. López et al. 2009).

Carbon and nutrient reserves

We found that individuals resprouting from lignotubers exhibited lower water-use efficiency and higher levels of nutrient reserves compared to those individuals that maintain a living crown (cf. Castell et al. 1994, also for Q. ilex). Resprouting individuals normally rely upon their pre-existing root system, which results in a greater root-to-shoot ratio and higher availability of soil resources per unit of leaf area (Savé et al. 1999).

Stored carbohydrates are the most important sources supporting new growth after disturbances that result in seriously reduced leaf area and carbon assimilation (Kays and Canham 1991, Canadell and López-Soria 1998, López et al. 2009). To our knowledge, this is the first study examining the replenishment of carbon reserves in resprouter trees after a natural drought episode. Other studies have documented return times of 20 years to completely recover carbon reserves in lignotubers of holm oaks after severe experimental thinning (López et al. 2009). In our case, stored carbon reserves were still depleted by 60% in highly drought-damaged trees 7 years after the episode. In contrast, nutrient pools are completely recovered 7 years after the drought, assuming some degree of depletion during tree resprouting after drought, as reported by previous studies on disturbed Mediterranean woody plants (cf. Canadell and López-Soria 1998, for A. unedo and E. arborea). It should be noted, however, that in our study measures of carbohydrate storage were more representative than those of nutrients, as the latter include both available and structural forms.

Successive surveys in 2007 and 2011 showed a direct association between carbon reserve depletion and further deterioration of crown condition, as has been referred to by earlier studies (Bréda et al. 2006, Galiano et al. 2011). Our interpretation is that current levels of carbon uptake still suffer from a reduced photosynthetic area and they are not enough to meet the demands—new growth, respiration—of an extensive root system and the replenishment of pool reserves (Canadell and López-Soria 1998). Low levels of stored carbohydrates have been associated with increased risk of drought-induced mortality by carbon starvation (McDowell et al. 2008, Galiano et al. 2011). However, we cannot discard that other mechanisms might have been in operation, as has been recently proposed in the current debate about the role of carbon depletion in tree mortality (cf. Sala et al. 2010, McDowell et al. 2011, Sala et al. 2012). In this sense, more research is needed to determine whether our results are generalizable to other species or situations. In any case, our results highlight the determinant role of carbon reserves in tree-level responses to drought and subsequent recovery, and suggest that an increase in the frequency of dry periods, as may occur according to climate change projections (IPCC 2007), may lead trees to a progressive depletion of carbon reserves and, consequently, to a loss of resilience in the face of future droughts.

In conclusion, our study documents a drought-driven forest dieback episode in which soil depth and characteristics of individual trees determined tree-level effects, illustrating how climatic trends interact with local drivers to produce complex patterns of responses at the landscape level. We show how this dieback results in a depletion of carbon reserves that lasts for several years after the episode, which, in turn, is likely to induce a progressive loss of forest resilience in the face of future droughts. Under climate change conditions, holm oak forests may be particularly challenged if drought events become more frequent and intense and return periods become shorter than the time period required for reserves to recover.

Supplementary data

Supplementary data for this article are available at Tree Physiology Online.

Conflict of interest

None declared.

Funding

This study was supported by the Spanish Ministry of Education and Sciences via competitive projects CGL2006-01293, CGL2007-60120, CSD2008-0004 and CGL2009-08101, and by the Government of Catalonia via AGAUR grant 2009 SGR 247. L.G. was supported by an FPI scholarship from the Spanish Ministry of Education and Sciences.

Acknowledgments

We thank the Catalan Forest Service from the Garrotxa Volcanic Zone Natural Park, and especially Joan Montserrat i Reig, for facilitating our field work and for their generous comments. We appreciate help from the undergraduate students (Albert Rivas, Albert Vilà, Josep Barba, Jose Donate and Teresa Salvadó) who were involved in this study. We are also indebted to the Anna Sala's plant physiology lab (The University of Montana, Missoula, USA) for teaching us the technique to analyze carbohydrates stored in wood, and to the Department of Genetics and Microbiology (Universitat Autònoma de Barcelona, Spain) for allowing us to use their labs to perform the analyses.

References

Agelet i Subirada
A.
Montserrat i Reig
J.
.
2002
.
Els alzinars afectats per la secada durant l'estiu de l'any 2000 (termes municipals de les Preses, la Vall de Bas, Sant Feliu de Pallerols, les Planes d'Hostoles i Sant Aniol de Finestres)
.
Delimitació, seguiment i estudi de les masses boscanes implicades. Avaluació de l'impacte i propostes de futur. Parc Natural Zona Volcànica de la Garrotxa. Departament de Medi Ambient, Generalitat de Catalunya
.
Allen
C.D.
Macalady
A.
Chenchouni
H.
et al
2010
.
A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests
.
For. Ecol. Manage
 .
259
:
660
684
.
Barbéro
M.
Loisel
R.
Quézel
P.
.
1992
.
Biogeography, ecology and history of Mediterranean Quercus ilex ecosystems
.
Vegetatio
 
99–100
:
19
34
.
Bigler
C.
Bräker
O.U.
Bugmann
H.
Dobbertin
M.
Rigling
A.
.
2006
.
Drought as an inciting mortality factor in Scots pine stands of the Valais, Switzerland
.
Ecosystems
 
9
:
330
343
.
Borghetti
M.
Cinnirella
S.
Magnani
F.
Saracino
A.
.
1998
.
Impact of long-term drought on xylem embolism and growth in Pinus halepensis Mill
.
Trees Struct. Funct
 .
12
:
187
195
.
Bréda
N.
Badeau
V.
.
2008
.
Forest tree responses to extreme drought and some biotic events: towards a selection according to hazard tolerance?
CR Geosci
 .
340
:
651
662
.
Bréda
N.
Huc
R.
Granier
A.
Dreyer
E.
.
2006
.
Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences
.
Ann. For. Sci
 .
63
:
625
644
.
Briones
O.
Montaña
C.
Ezcurra
E.
.
1998
.
Competition intensity as a function of resource availability in a semiarid ecosystem
.
Oecologia
 
116
:
365
372
.
Callaway
R.M.
Walker
L.R.
.
1997
.
Competition and facilitation: a synthetic approach to interactions in plant communities
.
Ecology
 
78
:
1958
1965
.
Camarero
J.J.
Lloret
F.
Corcuera
L.
Peñuelas
J.
Gil-Pelegrín
E.
.
2004
.
Cambio global y decaimiento del bosque. In
Ecología del Bosque Mediterráneo en un Mundo Cambiante. Ed
 .
Valladares
F.
.
Ministerio de Medio Ambiente
,
Madrid, Spain
, pp
397
423
.
Canadell
J.
Rodà
F.
.
1991
.
Root biomass of Quercus ilex in a montane Mediterranean forest
.
Can. J. For. Res
 .
21
:
1771
1778
.
Canadell
J.
Zedler
P.
.
1995
.
Underground structures of woody plants in Mediterranean ecosystems of Australia, California and Chile. In
Ecology and Biogeography of Mediterranean Ecosystems in Chile, California and Australia. Eds
 .
Fox
M.
Kalin
M.
Zedler
P.
.
Springer
,
New York
, pp
177
210
.
Canadell
J.
López-Soria
L.
.
1998
.
Lignotuber reserves support regrowth following clipping of two Mediterranean shrubs
.
Funct. Ecol
 .
12
:
31
38
.
Canadell
J.
Djema
A.
López
B.C.
Lloret
F.
Sabaté
S.
Siscart
D.
.
1999
.
Structure and dynamics of the root system. In
Ecology of Mediterranean Evergreen Oak Forests. Ecological Studies
 , Vol.
137
. Eds.
Rodà
F.
Retana
J.
Gracia
C.A.
Bellot
J.
.
Springer
,
Berlin, Heidelberg
, pp
47
59
.
Castell
C.
Terradas
J.
Tenhunen
J.D.
.
1994
.
Water relations, gas exchange, and growth of resprouts and mature plant shoots of Arbutus unedo L. and Quercus ilex L
.
Oecologia
 
98
:
201
211
.
Christensen
J.H.
Hewitson
B.
Busuioc
A.
et al
2007
.
Regional climate projections. In
Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the IPCC. Eds
 .
Solomon
S.
Qin
D.
Manning
M.
Chen
Z.
Marquis
M.
Averyt
K.B.
Tignor
M.
Miller
H.L.
.
Cambridge University Press
,
Cambridge, UK and New York, NY
, pp
847
943
.
Cliff
A.D.
Ord
J.K.
.
1981
.
Spatial processes: models and applications
.
Pion Limited
,
London
.
Cotillas
M.
Sabaté
S.
Gracia
C.
Espelta
J.M.
.
2009
.
Growth response of mixed Mediterranean oak coppices to rainfall reduction. Could selective thinning have any influence on it?
For. Ecol. Manage
 .
258
:
1677
1683
.
Cottam
G.
Curtis
J.T.
.
1949
.
A method for making rapid surveys of woodlands by means of pairs of randomly selected trees
.
Ecology
 
30
:
101
104
.
Díaz-Delgado
R.
Lloret
F.
Pons
X.
Terradas
J.
.
2002
.
Satellite evidence of decreasing resilience in Mediterranean plant communities after recurrent wildfires
.
Ecology
 
83
:
2293
2303
.
Espelta
J.M.
Riba
M.
Retana
J.
.
1995
.
Patterns of seedling recruitment in West-Mediterranean Quercus ilex forests influenced by canopy development
.
J. Veg. Sci
 .
6
:
465
472
.
Faraway
J.J.
2006
.
Extending the linear model with R: generalized linear, mixed effects and nonparametric regression models
.
Chapman and Hall
,
Boca Raton, FL
.
Farquhar
G.D.
Ehleringer
J.R.
Hubick
K.T.
.
1989
.
Carbon isotope discrimination and photosynthesis
.
Annu. Rev. Plant Physiol
 .
40
:
503
537
.
Franklin
J.F.
Shugart
H.H.
Harmon
M.E.
.
1987
.
Tree death as an ecological process: the causes, consequences and variability of tree mortality
.
BioScience
 
37
:
550
556
.
Galiano
L.
Martínez-Vilalta
J.
Lloret
F.
.
2010
.
Drought-induced multifactor decline of Scots pine in the Pyrenees and potential vegetation change by the expansion of co-occurring oak species
.
Ecosystems
 
13
:
978
991
.
Galiano
L.
Martínez-Vilalta
J.
Lloret
F.
.
2011
.
Carbon reserves and canopy defoliation determine the recovery of Scots pine 4 yr after a drought episode
.
New Phytol
 .
190
:
750
759
.
Gracia
C.A.
Sabaté
S.
Martinez
J.M.
Albeza
E.
.
1999
.
Functional responses to thinning. In
Ecology of Mediterranean Evergreen Oak Forests. Ecological Studies
 , Vol.
137. Eds
.
Rodà
F.
Retana
J.
Gracia
C.A.
Bellot
J.
.
Springer
,
Berlin
, pp
329
338
.
Gracia
C.
Sabaté
S.
López
B.C.
Sánchez
A.
.
2001
.
Presente y futuro del bosque mediterráneo: balance de carbono, gestión forestal y cambio global. In
Aspectos funcionales de los ecosistemas mediterráneos. Eds
 .
Zamora
R.
Pugnaire
F.I.
.
CSIC-AEET
,
Granada, Spain
, pp
351
372
.
Hoch
G.
Popp
M.
Körner
C.
.
2002
.
Altitudinal increase of mobile carbon pools in Pinus cembra suggests sink limitation of growth at the Swiss treeline
.
Oikos
 
98
:
361
374
.
Hoch
G.
Richter
A.
Körner
C.
.
2003
.
Non-structural carbon compounds in temperate forest trees
.
Plant Cell Environ
 .
26
:
1067
1081
.
Ibàñez
J.J.
Lledó
M.J.
Sánchez
J.R.
Rodà
F.
.
1999
.
Stand structure, aboveground biomass and production. In
Ecology of Mediterranean Evergreen Oak Forests. Ecological Studies
 , Vol.
137. Eds
.
Rodà
F.
Retana
J.
Gracia
C.A.
Bellot
J.
.
Springer
,
Berlin
, pp
31
45
.
Infante
J.M.
Domingo
F.
Fernández Alés
R.
Joffre
R.
Rambal
S.
.
2003
.
Quercus ilex transpiration as affected by a prolonged drought period
.
Biol. Plant
 .
46
:
49
55
.
IPCC
.
2007
.
Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the IPCC
.
Cambridge University Press
,
Cambridge, UK and New York, NY
.
James
S.
1984
.
Lignotubers and burls—their structure, function and ecological significance in Mediterranean ecosystems
.
Bot. Rev
 .
50
:
225
266
.
Jones
M.B.
Laude
H.M.
.
1960
.
Relationship between sprouting in chamise and the physiological condition of the plant
.
J. Range Manage
 .
13
:
210
214
.
Kays
J.S.
Canham
C.D.
.
1991
.
Effects of time and frequency of cutting on hardwood root reserves and sprout growth
.
For. Sci
 .
37
:
524
539
.
Limousin
J.M.
Rambal
S.
Ourcival
J.M.
Rocheteau
A.
Joffre
R.
Rodríguez-Cortina
R.
.
2009
.
Long-term transpiration change with rainfall decline in a Mediterranean Quercus ilex forest. Glob
.
Change Biol
 .
15
:
2163
2175
.
Linares
J.C.
Camarero
J.J.
Carreira
J.A.
.
2009
.
Interacting effects of changes in climate and forest cover on mortality and growth of the southernmost European fir forests
.
Glob. Ecol. Biogeogr
 .
18
:
485
497
.
Linares
J.C.
Camarero
J.J.
Carreira
J.A.
.
2010
.
Competition modulates the adaptation capacity of forests to climatic stress: insights from recent growth decline and death in relict stands of the Mediterranean fir Abies pinsapo
.
J. Ecol
 .
98
:
592
603
.
Lloret
F.
Siscart
D.
.
1995
.
Los efectos demográficos de la sequía en poblaciones de encina
.
Cuad. Soc. Esp. Cienc. For
 .
2
:
77
81
.
Lloret
F.
Siscart
D.
Dalmases
C.
.
2004
.
Canopy recovery after drought dieback in holm-oak Mediterranean forests of Catalonia. Glob
.
Change Biol
 .
10
:
2092
2099
.
López
B.C.
Gracia
C.A.
Sabaté
S.
Keenan
T.
.
2009
.
Assessing the resilience of Mediterranean holm oaks to disturbances using selective thinning
.
Acta Oecol
 .
35
:
849
854
.
López-Soria
L.
Castell
C.
.
1992
.
Comparative genet survival after fire in woody Mediterranean species
.
Oecologia
 
91
:
493
499
.
Lorenz
M.
Becher
G.
Mues
V.
et al
2006
.
Forest condition in Europe. 2006 Technical Report of ICP Forests
.
Institute for World Forestry, Hamburg, Germany. http://icp-forests.net/
.
Malanson
G.P.
Trabaud
L.
.
1987
.
Ordination analysis of the components of resilience of Quercus coccifera garrigue
.
Ecology
 
68
:
463
472
.
Malanson
G.P.
Trabaud
L.
.
1988
.
Vigour of post-fire resprouting by Quercus coccifera L
.
J. Ecol
 .
76
:
351
365
.
Manion
P.D.
1991
.
Tree disease concepts
.
Prentice Hall
,
Upper Saddle River, NJ
.
Martínez-García
F.
1999
.
Los bosques de Pinus sylvestris L. del Sistema Central español. Distribución, historia, composición florística y tipología
.
PhD dissertation, Universidad Complutense de Madrid, Madrid, Spain
.
Martínez-Vilalta
J.
Piñol
J.
.
2002
.
Drought-induced mortality and hydraulic architecture in pine populations of the NE Iberian Peninsula
.
For. Ecol. Manage
 .
161
:
247
256
.
Martínez-Vilalta
J.
Piñol
J.
Beven
K.
.
2002
.
A hydraulic model to predict drought-induced mortality in woody plants: an application to climate change in the Mediterranean
.
Ecol. Model
 .
155
:
127
147
.
Martínez-Vilalta
J.
Mangirón
M.
Ogaya
R.
Sauret
M.
Serrano
L.
Peñuelas
J.
Piñol
J.
.
2003
.
Sap flow of three co-occurring Mediterranean woody species under varying atmospheric and soil water conditions
.
Tree Physiol
 .
23
:
747
758
.
McDowell
N.
Pockman
W.T.
Allen
C.D.
et al
2008
.
Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought?
New Phytol
 .
178
:
719
739
.
McDowell
N.
Beerling
D.J.
Breshears
D.D.
Fisher
R.A.
Raffa
K.F.
Stitt
M.
.
2011
.
The interdependence of mechanisms underlying climate-driven vegetation mortality
.
Trends Ecol. Evol
 .
26
:
523
532
.
Mooney
H.A.
1983
.
Carbon-gaining capacity and allocation patterns of Mediterranean-climate plants
.
Ecol. Stud
 .
43
:
103
119
.
Mullette
K.J.
Bamber
R.K.
.
1978
.
Studies of the lignotubers of Eucalyptus gummifera. III Inheritance and chemical composition
.
Aust. J. Bot
 .
26
:
23
28
.
Navarro
F.B.
Jiménez
M.N.
Cañadas
E.M.
Gallego
E.
Terrón
L.
Ripoll
M.A.
.
2010
.
Effects of different intensities of overstory thinning on tree growth and understory plant-species productivity in a semi-arid Pinus halepensis Mill. afforestation
.
For. Syst
 .
19
:
410
417
.
Ninyerola
M.
Pons
X.
Roure
J.M.
.
2000
.
A methodological approach of climatological modelling of air temperature and precipitation through GIS techniques
.
Int. J. Climatol
 .
20
:
1823
1841
.
Obón
B.
1997
.
Recuperació de la vegetació després del gran incendi del estiu de 1994 a Gualba (Vallès Oriental)
.
MSc thesis, Universidad de Lleida, Lleida, Spain
.
Ogaya
R.
Peñuelas
J.
Martínez-Vilalta
J.
Mangirón
M.
.
2003
.
Effect of drought on diameter increment of Quercus ilex, Phillyrea latifolia, and Arbutus unedo in a holm oak forest of NE Spain
.
For. Ecol. Manage
 .
180
:
175
184
.
Peñuelas
J.
Filella
I.
Lloret
F.
Piñol
J.
Siscart
D.
.
2000
.
Effects of a severe drought on water and nitrogen use by Quercus ilex and Phillyrea latifolia
.
Biol. Plant
 .
43
:
47
53
.
Peñuelas
J.
Lloret
F.
Montoya
R.
.
2001
.
Severe drought effects on Mediterranean woody flora in Spain
.
For. Sci
 .
47
:
214
218
.
Pons
X.
1996
.
Estimación de la Radiación Solar a partir de modelos digitales de elevaciones. Propuesta metodológica. In
VII Coloquio de Geografía Cuantitativa, Sistemas de Información Geográfica y Tele­detección. Eds
 .
Juaristi
J.
Moro
I.
.
Vitoria-Gasteiz
,
Spain
, pp
87
97
.
Poyatos
R.
Latron
J.
Llorens
P.
.
2003
.
Land use and land cover change after agricultural abandonment. The case of a Mediterranean mountain area (Catalan Pre-Pyrenees)
.
MT Res. Dev
 .
23
:
362
368
.
Pugnaire
F.I.
Armas
C.
Tirado
R.
.
2000
.
Balance de las interacciones entre plantas en ambientes mediterráneos. In
En Ecosistemas Mediterráneos. Aspectos Funcionales. Eds
 .
Zamora
R.
Pugnaire
F.I.
.
CSIC-AEET
,
Granada, Spain
, pp
213
235
.
Retana
J.
Riba
M.
Castell
C.
Espelta
J.M.
.
1992
.
Regeneration by sprouting of holm-oak (Quercus ilex) stands exploited by selection thinning
.
Vegetatio
 
99–100
:
355
364
.
Rivas-Martinez
S.
Rivas-Saenz
S.
.
Worldwide Bioclimatic Classification System
.
1996
.
Phytosociological Research Center
,
Spain
. .
Rodríguez-Calcerrada
J.
Pérez-Ramos
I.M.
Ourcival
J.M.
Limousin
J.M.
Joffre
R.
Rambal
S.
.
2011
.
Is selective thinning an adequate practice for adapting Quercus ilex coppices to climate change?
Ann. For. Sci
 .
68
:
575
585
.
Romero-Calcerrada
R.
Perry
G.L.W.
.
2004
.
The role of land abandonment in landscape dynamics in the SPA ‘Encinares del río Alberche y Cofio, Central Spain, 1984–1999
.
Landscape Urban Plan
 
66
:
217
232
.
Rundel
P.W.
Baker
G.A.
Parsons
D.J.
Stohlgren
T.J.
.
1987
.
Post-fire demography of resprouting and seedling establishment by Adenostoma fasciculatum in the California chaparral. In
Plant Response to Stress—Functional Analysis in Mediterranean Ecosystems. NATO Advanced Science Institute Series. Eds
 .
Tenhunen
J.D.
Catarino
F.M.
Lange
O.L.
Oechel
W.C.
.
Springer
,
Berlin
, pp
575
596
.
Sabaté
S.
Sala
A.
Gracia
C.A.
.
1995
.
Nutrient content in Quercus ilex canopies. Seasonal and spatial variations within a catchment
.
Plant Soil
 
168/169
:
297
304
.
Sala
A.
Piper
F.
Hoch
G.
.
2010
.
Physiological mechanisms of drought-induced tree mortality are far from being resolved
.
New Phytol
 .
186
:
274
281
.
Sala
A.
Woodruff
D.R.
Meinzer
F.C.
.
2012
.
Carbon dynamics in trees: feast or famine?
Tree Physiol
 . .
Sánchez-Humanes
B.
Espelta
J.M.
.
2011
.
Increased drought reduces acorn production in Quercus ilex coppices: thinning mitigates this effect but only in the short term
.
Forestry
 
84
:
73
82
.
Sarris
D.
Christodoulakis
D.
Körner
C.
.
2007
.
Recent decline in precipitation and tree growth in the eastern Mediterranean
.
Glob. Change Biol
 .
13
:
1187
1200
.
Savé
R.
Castell
C.
Terradas
J.
.
1999
.
Gas exchange and water relations. In
Ecology of Mediterranean Evergreen Oak Forests, Ecological Studies
 , Vol.
137. Eds
.
Rodà
F.
Retana
J.
Gracia
C.A.
Bellot
J.
.
Springer
,
Berlin
, pp
135
147
.
Terradas
J.
1999
.
Holm oak and holm oak forests: an introduction. In
Ecology of Mediterranean Evergreen Oak Forests, Ecological Studies
 , Vol.
137. Eds
.
Rodà
F.
Retana
J.
Gracia
C.A.
Bellot
J.
.
Springer
,
Berlin
, pp
3
14
.
Terradas
J.
Savé
R.
.
1992
.
The influence of summer and winter stress and water relationships on the distribution of Quercus ilex L
.
Vegetatio
 
100
:
137
145
.
Tognetti
R.
Longobucco
A.
Raschi
A.
.
1998
.
Vulnerability of xylem to embolism in relation to plant hydraulic resistance in Quercus pubescens and Quercus ilex co-occurring in a Mediterranean coppice stand in central Italy
.
New Phytol
 .
139
:
437
447
.
Trabaud
L.
1987
.
Natural and prescribed fire: survival strategies of plants and equilibrium in Mediterranean ecosystems. In
Plant Response to Stress. Functional Analysis in Mediterranean Ecosystems. Eds
 .
Tenhunen
J.D.
Catarino
F.M.
Lange
O.L.
Oechel
W.C.
.
Springer
,
Berlin-Heidelberg-New York-Tokyo
, pp
607
621
.
Vilà-Cabrera
A.
Martínez-Vilalta
J.
Vayreda
J.
Retana
J.
.
2011
.
Structural and climatic determinants of demographic rates of Scots pine forests across the Iberian Peninsula
.
Ecol. Appl
 .
21
:
1162
1172
.
Waring
R.H.
1987
.
Characteristics of trees predisposed to die
.
Bioscience
 
37
:
569
574
.
Western
A.W.
Grayson
R.B.
Blöschl
G.
.
2002
.
Scaling of soil moisture: a hydraulic perspective
.
Annu. Rev. Earth Planet. Sci
 .
30
:
149
180
.
Westman
W.E.
1986
.
Resilience: concepts and measures. In
Resilience in the Mediterranean-type Ecosystems. Eds
 .
Dell
B.
Hopkins
A.J.M.
Lamont
B.B.
.
Junk Publishers
,
Dordrecht, Netherlands
, pp
5
19
.

Supplementary data