Intra- and Inter-Annual Growth Patterns of a Mixed Pine-Oak Forest under Mediterranean Climate

Abstract

Temperature and precipitation variability throughout the year control the intra-annual dynamics of tree-ring formation. Physiological adaptation of trees to climate change is among the key issues to better understand and predict future forest performance and composition. In this study, we investigated the species’ coexistence and performance of Scots pine and pubescent oak growing in a mixed sub-Mediterranean forest in the northeast of the Iberian Peninsula. We assessed intra-annual cumulative growth patterns derived from band dendrometers during four consecutive growing seasons and long-term changes in basal area increment for the period 1950–2014. Our results revealed that Scots pine followed an intra-annual bimodal growth pattern. Scots pine growth was mainly limited by water availability at intra-annual, interannual and decadal time scales, which resulted in a negative long-term growth trend. Conversely, oak displayed a unimodal growth pattern, which was less climatically constrained. A significant increase in basal area of oak denotes an overall better potential acclimation to prevailing climatic conditions at the expenses of a higher risk of physiological failure during extreme climate events.

1. Introduction

Climate change is altering forest ecosystems at various time and spatial scales, modifying their composition, structure and functioning [1]. The Mediterranean Basin is considered a biodiversity hotspot and one of the regions most affected globally by the negative effects of global warming and its large ecological impacts [2,3]. Drought is one of the most hazardous consequences of climate change in Mediterranean ecosystems, since water availability is already scarce, particularly during summer [4]. This situation could be further aggravated during the 21st century since projected changes in climate point to a mean temperature increase coupled to larger precipitation variability as well as more frequent, persistent climate extreme events in the area [5,6]. As a consequence, drought-induced tree mortality is expected to occur more frequently in the Mediterranean region, as has been previously reported [7,8,9]. In this context, vegetation shifts may occur in both latitudinal and altitudinal transitions between Mediterranean and temperate climate zones [10,11,12,13].

Summer drought has also been one of the main drivers shaping physiological and structural adjustments in life-history traits of woody plants growing in the Mediterranean region, either through genotypic evolution (adaptation) or species phenotypic plasticity (acclimation) [14]. In fact, the high-plant diversity observed in the Mediterranean region is likely the reflection of diverse adaptive strategies performed by the different taxa to enhance survival and viability under limiting summer conditions [15]. Evolutionary plant biology states that within a plant community, species presenting a competitive advantage will exclude others by reducing resource availability [16]. In contrast, different plant species from the same community can coexist when resources are used in a complementary way (i.e., facilitation; [17]) or if there are trade-offs in resource allocation (niche segregation, [18]). Resource availability under the Mediterranean climate is not temporally continuous, i.e., water availability is lower in summer than in spring. The resulting environmental heterogeneity may allow variable species-specific timing to exploit the available resources favouring coexistence (fluctuation niche, [19,20]). However, when climate conditions become stressful and limiting, plant species adopting less successful physiological strategies will become less abundant. In the case of long-living organisms such as trees, processes of competitive exclusion might be only noticeable in the long-term. Therefore, understanding coexistence processes between tree species growing at their tolerance limits to climate are crucial to better assess and predict future forest composition and functioning.

Tolerance to drought is an expression of multiple combinations of physiological, anatomical and morphological adjustments [21]. Most of the studies have focused on morphological and physiological adjustments on traits in relation to water availability [22], but little is known regarding interspecific differences in growth sensitivity and its prominent role on the overall tree response. Growth timing and rates might result in functional traits depicting tree-life strategies [23]. In fact, drought-induced inhibition of cell formation may occur at the gas exchange level long before carbon supply [24]. Such sensitivity of cambial phenology and intra-annual growth dynamics have been already successfully monitored using band dendrometers [25,26,27]. These analyses can infer growth rates as well as identify intra-annual growth variability patterns in response to stressful climatic conditions. In the Mediterranean region, previous studies have shown a certain degree of plasticity in secondary growth in some tree species [28,29,30]. The bimodal pattern tracks the optimum climatic conditions during the growing season, which are split into two periods of growth interrupted by summer drought. Bimodality could also be considered as an avoidance strategy to drought, which does not necessarily match with other strategies displayed by the tree at different compartments. For instance, holm oak has been described as a tolerant tree species to water restrictions based on photosynthetic and hydraulic parameters [31], but it may also display a bimodal growth pattern [29], pointing to dissimilar levels of tolerance to drought among traits [32]. Nevertheless, the study of intra-annual growth patterns allows monitoring the timing when trees might be taking advantage of favourable climatic conditions and when resources might be allocated [33]. Consequently, these growth patterns associated with tolerance to drought could explain niche differentiation, contributing to a better understanding of some of the factors ruling coexistence processes among tree species.

During the last four decades, dendroecological studies have shown the potential to assess long-term tree growth limitations [34]. This retrospective analysis allows to reconstruct long-term changes in forest productivity since radial growth is considered as a valuable proxy of tree performance [35]. Despite the relevance of the information encoded in intra-annual patterns derived from dendrometer data and in long-term growth patterns obtained from tree-ring series, the comparison between growth drivers at intra- and inter-annual scales has been rarely explored. Moreover, both approaches could be used to assess the coexistence among tree species through the evaluation of the different strategies to cope with drought conditions.

The aim of this study is to assess the coexistence between two functionally distinct tree species such as Scots pine and pubescent oak by means of stem dendrometer records and tree-ring width series. The two selected species have contrasting leaf phenology and wood anatomy, as well as distinct ecological requirements i.e., light requirements and successional stages. We evaluated the phenology of secondary growth during four consecutive growing periods and assess the intra-annual and inter-annual growth dynamics and climatic drivers. Specifically, our objectives were (1) to assess the phenological differences in annual tree-girth dynamics between both species, (2) to identify the specific-species intra-annual growth patterns aiming at providing insights of differential use of resources, (3) to characterise long-term growth trends as a proxy of tree performance, and (4) to identify growth climatic drivers at intra- and inter-annual level.

2. Materials and Methods

2.1. Study Site and Species

This study was conducted in Prades (41°20′0.87″ N, 1°0′46.12″ E; 929 m a.s.l.), which is located in a central Mediterranean coastal mountain range in the NE Iberian Peninsula (Figure 1a). The climate at the site is typically Mediterranean with an average total annual rainfall of 595 mm, spring and autumn being the wettest seasons, and with a marked dry period during summer in which only 88 mm are annually recorded on average for the study period 1950–2014 (Figure 1b). The annual mean temperature is 10.8 °C with frosts occasionally occurring during winter. The area has experienced a significant increase in mean annual temperature since 1950 whereas precipitation remained stable (Figure S1). The soil of the study area is mostly Xerochrepts with loam texture and high gravel content [30]. This area is dominated by a mixed forest where the canopy-dominant tree species is Scots pine (Pinus sylvestris L.), and the understory is mostly composed of Quercus pubescens Willd. (=Q. humilis Mill.), Quercus ilex L., and other Mediterranean species.

Scots pine is an evergreen conifer that exhibits a wide Euro-Siberian distribution. It is considered a moderately drought-sensitive species due to its relative isohydric strategy with an efficient stomatal control of transpiration avoiding water losses [36]. Despite such tight control, mortality events have been reported in the same study area [7,37], due to the effects of severe drought since the 1990s [37] achieving stand mortality of about 20% in some areas [30,36]. Pubescent oak is a deciduous or marcescent sub-Mediterranean broadleaved species, whose distribution is mainly restricted to southern Europe. As a result of its ability to withstand considerably low water potentials during dry spells and reach water from deep soil layers [38,39], it has been classified as a relatively drought-tolerant tree species [40].

2.2. Climate Data Processing

Daily and monthly precipitation and temperature data were obtained from different stations in the area. The original climate data were retrieved from different agencies (Meteocat, Meteoprades and Aemet, Spain) as well as gridded products such as CRU TS4.03 [41] to cover the study period 1950–2014 (Table S1). Temperature data is highly related to elevation and therefore the data of all the available stations to estimate the temperature values for our study site were corrected for the difference in elevation. Precipitation does not show an altitudinal gradient. Thus, we obtained monthly and daily precipitation data series as the average of the data of the nearest climate stations that surrounded our site; those stations were Monastir Poblet, Prades, Riudabella and Tossal Baltasana (Table S1). Before computing the average, the homogeneity of the data was checked, and no correction needed to be undertaken.

2.3. Intra-Annual Observations: Stem Girth Measurements

To characterise intra-annual radial growth patterns, manual band dendrometers (Agricultural Electronics Co., Tucson, AZ, USA) were used to monitor six and ten Scots pines and pubescent oaks, respectively. The dendrometers were installed after carefully smoothing and removing the outermost layer of bark at a height of 1.2–1.3 m from the trunk base in January 1997 and August 1993 in Scots pine and pubescent oak, respectively. Therefore, and taking in consideration the adaptation period, complete year measurements were obtained from 1997 to 2000 for Scots pines and from 1994 to 2000 for pubescent oaks. Stem girths were read to the nearest 0.01 mm every 21 days in both species during the vegetation period. Measurements were also recorded during the dormant period (i.e., winter season) at a lower frequency. All measurements were obtained in the morning to reduce diurnal bias, which is caused by stem shrinkage from transpiration. Girth increment data were corrected for temperature effect using the band dendrometer thermal expansion factor (11.2 × 10⁻⁶ mm⁻¹ °C⁻¹) and transformed into cumulative radial increment data assuming a cylindrical stem. Daily radial-increment rates were calculated by dividing the radial increment by the number of days between two consecutive observation dates.

2.4. Inter-Annual Observations: Tree-Ring Chronologies

Increment cores were obtained in 2014 and 2018 for Scots pine and pubescent oak, respectively. Two or three samples were taken at breast height per dominant or co-dominant tree (16 Scots pine and 9 pubescent oaks), including the ones used for dendrometer measurements, using 5−mm increment borers. Additionally, diameter at breast height (DBH) and total height of trees were measured (Table S2). Samples were prepared following standard dendrochronological procedures [34]. The 5-mm cores were air-dried and sanded with progressively finer sandpaper until rings were completely distinguishable. Samples were visually cross-dated under a binocular microscope, and afterwards, scanned at 2400 dpi using a high-resolution scanner (Epson Expression 10,000 XL, Seiko Epson 185 Corporation, Suwa, Japan). Tree-ring widths were measured for each year using ImageJ software [42]. The quality and correct dating of the series were checked with the COFECHA program [43].

To remove non-climatic variability inherent in tree-ring width series, each individual series was detrended using a cubic smoothing spline with 50% frequency cut-off at 25 years [44]. Tree-ring chronologies were built by averaging the resulting detrended series using a robust mean. Furthermore, we applied a second type of detrending to assess overall long-term tree performance by transforming tree-ring width series into basal area increments (BAI; mm²). Detrending procedures and calculation of BAI were performed using the R package dplR [45].

2.5. Statistical Analyses

Similar to previous studies [25,26,46], cumulative radial increments were modelled applying Gompertz functions by first calculating the average of the increment growth of all the trees per species, according to the following formula:

$$\mathrm{Growth}=\mathrm{A}{\mathrm{e}}^{-\mathrm{\lambda }{\mathrm{e}}^{-\mathrm{k} \mathrm{DOY}}}$$

(1)

where A is upper asymptote of the maximum growth, and k is the rate of change parameter. The mean maximum growth rate (rmax) per species and year were derived from the first derivative, as well as day of the year when it was achieved (DOYmax). The start and the end of the growing period (DOYi and DOYe) were considered when 5% and 95% of the maximum increment (A) was achieved, respectively. Model fitting was carried out using the nls function from R package stats. Differences in each parameter between species were assessed using Student’s t-tests.

Generalised additive mixed models (GAMM; [47]) were used to analyse intra-annual growth patterns derived from dendrometer data, similar to previous studies [30]. Such models allow to characterise observed nonlinear variability at different time scales between response variable and one or several explanatory variables by using smoothing functions. Growth rates were modelled per species as a function of annual pattern (day of year; DOY) and the duration of the complete experiment (time) as:

$$\mathrm{Growth} \mathrm{rate} \left(\mathrm{mm} {\mathrm{day}}^{-1}\right) = {\mathrm{s}}_1 \left({\mathrm{DOY}}_{\mathrm{i}}\right) + {\mathrm{s}}_2\left({\mathrm{time}}_{\mathrm{i}}\right) + {\mathsf{\epsilon }}_{\mathrm{i}}$$

(2)

where s₁ and s₂ represent smoothing functions for each explanatory variable. Tree identity was defined as a random variable to account for repeated measurements. A second GAMM was explored per species allowing s₁ to vary as a function of year. GAMM’s were built using the R package mgcv [48].

The relationships between mean increment growth rates (mm day⁻¹) and climate were explored calculating Pearson’s correlations. Input climate data were daily mean temperature and total precipitation for the 5, 10, 15, 20 and 30 days before each sampling date, thus, taking into consideration a potential delay in climate-growth responses. Correlations were run considering values from the entire year and for the different seasons: winter (previous December, January and February), spring (March, April and May), summer (June, July and August) and autumn (September, October and November).

The quality of the tree-ring width chronologies was assessed using descriptive statistics and parameters such as the mean sensitivity (Ms) to reflect the high-frequency variance; the mean first-order autocorrelation (AR), the correlation among series (Corr), and the expressed population signal (EPS). EPS is a parameter that indicates the agreement among all the series included in a chronology [49]. The influence of climate (monthly total precipitation and mean temperature) on tree growth for the study period (1950–2014) was assessed by means of bootstrapped Pearson’s correlations using the R package treeclim [50]. Correlations were run from September of the previous year to December of the current year. All correlation coefficients displaying p-values < 0.05 were considered statistically significant. The long-term tree performance was assessed by analysing patterns of BAI during the study period 1950–2014. Linear trends were calculated using least-squares regressions for BAI chronology of each species. All statistical analysis, computing and graphics were performed using R [51].

3. Results

3.1. Intra-Annual Growth Patterns and Climatic Drivers

According to the parameters of the Gompertz function fitted to the cumulative growth increments, we did not observe significant differences in average maximum increment (A) between species (

Table 1. Parameters of the Gompertz function fitted to cumulative radial growth increments per species and year. Different lowercase letters represent interspecific significant differences (p < 0.05). A, uppermost asymptote or maximum radial increment; rmax, maximum growth rate; DOYrmax, day of the year that the maximum growth rate occurs; DOYi, day of the year that occurs the 5% of A; DOYe; day of the year that occurs the 95% of A.

Pinus Sylvestris L.
Year	A	rmax	DOYrmax	DOYi	DOYe
1997	0.149	0.055	137	95	251
1998	0.154	0.057	119	89	201
1999	0.208	0.077	126	92	217
2000	0.307	0.113	129	104	195
mean ± SD	0.205 ± 0.07 a	0.076 ± 0.03 a	127 ± 7 a	95 ± 6 a	216 ± 25 a
Quercus Pubescens Willd.
Year	A	rmax	DOYrmax	DOYi	DOYe
1997	0.353	0.130	140	102	242
1998	0.284	0.105	141	105	236
1999	0.228	0.084	132	102	215
2000	0.228	0.083	144	112	228
mean ± SD	0.273 ± 0.06 a	0.101 ± 0.02 b	139 ± 5 b	105 ± 5 b	230 ± 12 a

, Figure 2). The onset of the growing period started 10 days earlier in the case of Scots pine than in pubescent oak (5 April and 15 April, Table 1). Although growth cessation seemed to occur later in pubescent oak than in Scots pine, the difference was non-significant.

Growth bimodality was clear for Scots pine in all studied years but not for pubescent oak, which only displayed a smoothed version of such pattern during the years 1998 and 2000 (Figure 3). Except for the year 2000, pubescent oak grew during early summer, but at slower rates than during the beginning of the growing season. Growth rates during the second growing peak were higher in Scots pine than in pubescent oak in all studied years.

Overall, temperature seemed to play a more prominent role on intra-annual growth dynamics of pubescent oak than precipitation, whereas both temperature and precipitation influenced Scots pine dynamics (Figure 4). At the intra-annual level, positive correlations were found between daily radial increment rates of Scots pine and lagged climatic variables for total precipitation averaged for the 10, 15, 20, and 30 days before the sampling date, showing the latter the highest correlation (r = 0.47). In contrast, mean daily temperature did not show any significant correlation with Scots pine growth rates. In the case of pubescent oak, temperature exerted a significant positive effect on growth rates, and the mean temperature of the 10 days before sampling showed the highest correlation (r = 0.64). A non-significant correlation was found with total precipitation.

At the seasonal level, winter climate did not exert a significant influence on intra-annual growth, except for the temperature from 5 days before the sampling date that had a negative effect on pubescent oak radial growth rates (Figure 4). Spring temperature showed a strong influence on intra-annual radial growth rates of both species, particularly noticeable in pubescent oak (15 days, r = 0.91). In contrast, spring precipitation only significantly affected Scots pine intra-annual radial growth rates, showing the maximum correlation with the accumulative precipitation from 30 days before sampling (r = 0.70). Summer temperature was negatively related to growth rates for both species. The cumulative effect of precipitation (30 days) in Scots pine intra-annual growth rates was also significant, whereas growth rates of pubescent oak were not significantly related to summer precipitation. Autumn temperature did not show any significant correlation with intra-annual radial growth rates in any of the studied species. Only autumn precipitation from 30 days before sampling had a significant effect on Scots pine radial growth rates.

3.2. Long-Term Growth Patterns and Climatic Drivers

Both species displayed similar synchrony among series as indicated by mean inter-series correlation (Cor) (

Table 2. Characteristics of the tree-ring width chronologies of Scots pine and pubescent oak. Corr, correlation with master series; Ms, mean sensitivity; EPS, expressed population signal.

ID	Cores/Trees	Time Span		Mean Length (Years)	Corr	Ms	Growth (SD) mm/yr	EPS > 0.85 Since
RW-Quercus	17/9	1925	2017	70.9	0.557	0.267	0.93 (0.44)	1943
RW-Pinus	31/15	1853	2014	120.3	0.647	0.362	1.31 (0.65)	1865

). The pubescent oak chronology showed lower mean sensitivity (Ms) in comparison to Scots pine. The Scots pine chronology spanned longer than the chronology from pubescent oak, and therefore, the reliable period of the Scots pine chronology (EPS > 0.85) started earlier. Nevertheless, both chronologies showed reliable EPS during the study period (1950–2014). In the case of Scots pine, we observed a significant decreasing long-term growth trend for the period 1950–2014 (slope = −3.28, p < 0.05; Figure 5). In contrast, pubescent oak displayed a significant increasing growth trend during the study period (slope = 0.88, p < 0.01).

In general, Scots pine seemed to be more sensitive to climate than pubescent oak as shown by the larger number of significant Pearson’s correlation coefficients (Figure 6). Precipitation seemed to exert a more prominent role in interannual growth variability of both tree species than temperature. Temperature from late spring to early summer (May, June, and July) exerted a significant negative effect on Scots pine growth. Conversely, precipitation from current year late spring to early summer (May, June, and July) had a positive effect on tree growth along with precipitation in winter and early spring (January, March, May, June, July, and September). However, precipitation from current November was negatively correlated with Scots pine chronology. Precipitation was also the main climatic driver of pubescent oak radial growth. Late spring (April and May) and early summer (June) precipitation, as well as January precipitation, exerted a significant positive effect on radial growth. In contrast, pubescent oak radial growth only correlated negatively with current July temperature (r = −0.27).

4. Discussion

4.1. Niche Complementarity: Phenology and Growth Patterns

Our results evidence different patterns in intra-annual growth and growth phenology between Scots pine and pubescent oak, which may indicate a potential niche complementarity; the different timing would mainly drive this complementarity on resource use displayed by the studied species. In our study site, radial growth of Scots pine started 10 days earlier than in the case of pubescent oak, but there were no significant differences in the cessation of stem growth, resulting in a longer growing period but at lower growth rates than oak. The opposite pattern has been previously observed in pine-oak comparisons under milder climate conditions [26,33]. As a conifer, Scots pine growth can begin independently of the current year needle unfolding [52], since this species relies on the photosynthetic capacity of the needles from previous cohorts instead of mobilizing old carbon reserves [26]. In contrast, pubescent oak as a ring-porous and deciduous tree species displays a different functional strategy. The pubescent oak growth onset, and particularly the earlywood vessel formation, occurs prior to photoassimilate synthesis and leaf budburst [53,54], but at expenses of carbon reserves accumulated during previous years [26,55]. Such differences in leaf phenology due to the physiological functioning of the studied species could explain the distinct growth rate patterns during the early growing season. Scots pine displayed higher growth rates than pubescent oak during spring. This capacity of Scots pine of restarting tree machinery after winter dormancy faster than pubescent oak provides certain advantages, such as greater spring water availability as a result of a reduced competition.

The delayed growth onset of pubescent oak with respect to Scots pine could be compensated by the higher capacity of this species to withstand growth during late spring and early summer in comparison to Scots pine. Indeed, pubescent oak displayed the highest growth rates during the second half of June and Scots pine interrupted growth earlier in summer (end July–beginning August). Similar results were found comparing xylem formation of Scots pine and the sub-Mediterranean oak Quercus pyrenaica under rainfall exclusion treatments [33]. Such distinct sensitivities to summer drought observed in both studied species were previously observed in stomatal-related strategies [56]. Gas exchange is tightly controlled via stomatal closure during dry spells in the case of Scots pine [33,39,57], which could modify the allocation of resources and reduce radial growth under drought. In contrast, drought-tolerant species, such as pubescent oak, perform a more relaxed stomatal control maintaining high transpiration rates during dry spells, partly due to the ability to extract water from deeper soil layers [33]. However, several studies indicated that trees might adjust to drought by first limiting secondary growth rather than carbon assimilation [58]. The differences in drought tolerance and physiological strategies uncouples growth and gas exchange responses, as previously observed in other oak species [32]. Nevertheless, further studies monitoring gas exchange and intra-annual growth patterns would be helpful to disentangle differences in drought sensitivity among tree compartments.

Our results are in line with previous studies using Mediterranean tree species and describing a bimodal growth pattern originated by a quiescent period during summer drought [27,29,30,59]. However, our results also indicate higher xylem plasticity in the case of the boreal Scots pine rather than in the sub-Mediterranean pubescent oak, since the former displayed a bimodal growth pattern in the studied years. A unimodal growth pattern has also been described in Scots pine growing under continental Mediterranean conditions [28], pointing to high phenotypical plasticity of the species under contrasting climatic conditions.

Except for holm oak, little has been explored regarding the capacity of angiosperm species to display a bimodal growth pattern under Mediterranean conditions (but see [60]). Our study highlighted the potential growth bimodality observed in pubescent oak during years with extremely low summer water availability such as 1998 and 2000 (see climatic conditions in Figure 1c). In fact, we hypothesize that such a pattern only appears when water availability is scarce in deeper soil layers. Other studies have suggested that bimodality is imposed in oak species due to the differences in growth rates between early-latewood [26]. Our results did not show bimodality in growth rates on a yearly basis for pubescent oak discarding such hypothesis, but further analyses such as cambial activity monitoring would be required to reinforce our results.

Late summer—early autumn radial increment was observed for both species. This pattern is common to other pine species. Results from Pinus halepensis Mill. growing in a nearby study site also showed a marked growth bimodality [30]. Such pattern might indicate a high plastic and opportunistic strategy displayed by the studied species to the intra-annual variability of the Mediterranean climate. The physiological basis and drivers of tree growth reactivation after a quiescent period induced by summer drought are not yet fully understood, but it might be a strategic adjustment to favourable environmental conditions (i.e., higher water availability). However, after a harsh summer drought, tree stems shrink and some studies have suggested that the growth peak observed in autumn is the result of the stem rehydration rather than cambial reactivation [61]. Indeed, stem girth variations record not only intra-annual changes in xylem but also in other tissues such as phloem and periderm; therefore, tree shrinking and swelling could introduce internal noise in the measurements [62]. Although previous studies have demonstrated the validity of using dendrometer measurements to record cambial activity under similar climatic conditions [28,30], further studies as monitoring xylogenesis during autumn could be useful to disentangle physiological mechanisms behind growth bimodality. However, it also needs to be acknowledged the fact that dendrometers integrate stem changes in volume over the whole perimeter of the trunk whereas cambial monitoring is usually limited at one point of the stem, which might partly explain the discrepancies between both methodologies. Nevertheless, other authors have even demonstrated that cambial activity can continue in conifers during mild winters [30,63], pointing to a mostly certain radial increment due to cambial activation in the case of Scots pine, but it remains uncertain in the case of pubescent oak.

4.2. Differential Sensitivity to Climate

Our analyses highlighted the species-specific sensitivity to climate at intra-annual to inter-annual time scales. Scots pine radial growth was generally more dependent on water availability than pubescent oak. Such results agree with other studies comparing climate growth drivers of both species at different locations [40,64]. Spring climate has strong effects on intra-annual increments. Warm spring temperature promotes the onset of xylem growth after winter dormancy as previously established by several authors [65,66]. However, Scots pine intra-annual growth was also highly influenced by intra-annual spring precipitation. This result was also observed when analysing inter-annual climate drivers of tree growth. In fact, the largest portion of tree rings in Scots pine are produced during spring before the growth cessation due to summer drought [67]. The negative effect of summer temperature was observed at intra- and inter-annual time scales in both species, but it was especially constraining Scots pine growth. In contrast, spring and summer precipitation were significant at inter-annual but not at intra-annual times scales for pubescent oak. Such a difference could be explained by the particularities of the studied years using dendrometers in comparison with the 64 years included in the climate-growth correlations. Spring precipitation has been defined as a key driver for earlywood vessels formation [68], which is consistent with the correlations found in our study. Although our oak chronology was statistically robust, we analysed only nine oak trees, which is below the desired sample depth for dendrochronological analyses and could limit the extent of our findings.

The positive relationship between intra-annual radial increments of Scots pine and 30-day accumulated precipitation during summer and autumn periods is remarkable. This might suggest that tree growth might be reactivated after a summer quiescent period only when water is available, reinforcing bimodality in growth recorded by the dendrometers. Such relationship was not observed in pubescent oak either at intra-annual- or inter-annual time scales, the potential growth reactivation of the species at the end of the summer remaining uncertain.

4.3. Long-Term Tree Performance

In the case of Scots pine, a long-term growth reduction has been observed since 1950. This negative growth trend agrees with other studies analysing Scots pine growth in the NE Iberian Peninsula during the last decades [32,37,40,69]. The long-lasting tree growth decline has been widely proposed as a potential predictor of mortality events. Negative growth trends are related to a gradual decline in hydraulic performance coupled with depletion in carbon reserves [34]. Mortality events have been reported in the same Scots pine forest (Gutiérrez’s personal observations) and in stands close by [7,37]. In contrast, pubescent oak showed a positive, but rather slow, increase in BAI since 1950. This suggests a potential better acclimation of the species to current climatic conditions in the area. However, it is important to highlight the age difference between the two species, which could partially explain some of the differences in growth trends.

Under projected changes in climate, summer aridity is expected to keep increasing, which could narrow the temporal windows of suitable climate that allows Scots pine radial growth leading to an impairment of water transport [64]. Despite the higher capacity of pubescent oak to withstand dry climatic conditions, high temperatures are also limiting pubescent oak radial growth in summer (July), and consequently, it might also be compromised with increasing aridity.

Some studies comparing tree performance and recruitment of Scots pine and co-occurring oak species such as pubescent oak growing in Mediterranean conditions found a better performance of the latter under projected changes in climate conditions, even pointing to vegetation shifts in some regions [11,13,70]. Although other factors such as forest structure, local site conditions and past human management also play an important role [12], Scots pine performance and recruitment are highly threatened by increased aridity [37,40]. Our study unveils that the drought-avoidance strategy performed by Scots pine might not be sufficient to counterbalance the negative impacts of climate change, even when competition is reduced during spring because of niche complementarity. In contrast, the drought-tolerance strategy displayed by pubescent oak allows growing at low rates but it might come along with higher risks of, for instance, xylem embolisms.

5. Conclusions

Our results highlight the capacity of boreal species, such as Scots pine, to stop secondary growth during summer and resume it after the unfavourable period. This was reinforced by the positive relationships between autumn climate conditions and radial increments at intra-annual and inter-annual time scales. We hypothesize that pubescent oak only displays a bimodal growth pattern during years with low water availability at deep soil layers. However, this would need to be confirmed by directly monitoring cambial activity, since no significant correlations were found between autumn climate and intra-annual radial growth. The analyses of the intra-annual dynamics of the coexisting tree species in the area will help to better assess the future dynamic of this mixed forest.