Five study sites were designated in Hungary (Figure 1, Table 1). These sites exhibit annual precipitation ranging from 562 mm to 705 mm and annual air temperatures between 9.3°C and 10.3°C. The southwestern regions typically have moister climates, while the northeastern areas are characterized by drier conditions. None of the selected sites have access to groundwater and they are situated at elevations between 215 m and 370 m above mean sea level (Table 1). Since 1971, annual precipitation at the sampling sites has not shown a significant trend, however, the drier sites showed greater interannual variability (Figure 2). In contrast, air temperature has exhibited a clearer overall trend and significant interdecadal variability. A notable cooler period occurred during the 1980s at all sites, but since 1990, warmer years have been recorded. Over the study period, all sites have experienced a warming trend, with particularly steep increases in annual mean temperature of 0.05°C to 0.06°C per year in the past 30 years (Figure 2). For this study, we designated one species-rich mixed forest stand at each selected site, featuring varying species composition and age, ranging from 42 to 82 years. Each tree species was represented in at least two of the mixed forest stands (Table 2). The five tree species investigated - Acer campestre L., Fraxinus ornus L., Quercus pubescens Willd., Tilia tomentosa Moench. and Quercus cerris L.—are commonly found in Central European broadleaf mixed forests. They typically occur in communities classified under the phytosociological alliances Carpinion betuli (oak-hornbeam forests), Quercetum petraeae-cerris (sessile-Turkey oak forests), Aceri campestri and tatarici-Quercetum (maple-oak forests) (Leuschner and Ellenberg, 2017). To minimize the impact of forest management practices—such as selective thinning and regeneration cutting - on growth-climate relationships, we selected forest stands with low management intensity over the past few decades. Additionally, to reduce competition, we avoided selecting suppressed trees, as they exhibit larger growth responses to various interventions compared to trees in the upper canopy (Nowacki and Abrams, 1997). The selected trees were primarily grown from seed, with a smaller portion being coppice regeneration.

Meteorological data were obtained from the HUCLIM daily gridded climate dataset of the Hungarian Meteorological Service (HMS, https://odp.met.hu/). This dataset has an approximate spatial resolution of 10 km and covers the period from 1971 to 2022. We assigned the nearest grid points to the study sites and aggregated daily mean temperatures and total precipitation into monthly averages. To account for the altitude differences between the study sites and the corresponding grid points, we adjusted the mean monthly temperature data using monthly elevation gradients (Péczely, 1979).

The monthly water balance (WB) was calculated as the difference between precipitation and potential evapotranspiration, following the method, described by McCabe and Markstrom (2007). WB serves as an ecologically relevant indicator of water availability for tree growth and has a strong correlation with the radial growth of various tree species (Stojanović et al., 2018; Vitasse et al., 2019). WB effectively reflects the negative impacts of increased temperatures on water availability (Vitasse et al., 2019). We calculated WB for different periods ranging from 1 to 12 months, starting in August of the current year of growth. In addition to WB, we used the Forestry Aridity Index (FAI) (Führer et al., 2011) which is defined as the ratio of the mean temperature of July and August and the precipitation sums of May to July plus the precipitation sum recorded from July to August. This index is particularly relevant for assessing tree growth (Führer et al., 2011).

For monthly soil water balance modelling, we used the Thornthwaite-type model (Thornthwaite, 1948) covering the period from 1971 to 2022. The input variables for the model included monthly mean air temperature, precipitation sum, the latitude of the sites and the plant-available soil-water storage capacity (AWC). To estimate the AWC of the soil, we collected 4–5 soil samples from individual soil horizons down to a depth of 100 cm. These samples were analysed for various chemical and physical parameters (Table 3). Field capacity and wilting point water content of the soil samples were estimated from the particle composition and base rock fraction content using pedo-transfer functions and the Rosetta3 model within the “soilDB” package of R software (Zhang and Schaap, 2017). We summed the differences in water content between field capacity (pF = 2.5) and permanent wilting point (pF = 4.2) over the entire rooting depth, which exceeded 1 meter at all sites, as indicated by visual observations of fine roots in the soil samples. Finally, we calculated the summer water stress index (Is) by dividing soil water deficit and the maximum extractable water for 120 cm soil depth or up to the bedrock depth, assuming uniform soil texture below 100 cm depth, as described by Granier et al. (1999).

We selected up to 24 trees of each species based on the available number of trees in each stand and extracted one core at breast height (1.3 m) from each tree (Fritts, 1976). A total of 375 cores were collected, using a Pressler increment borer (Haglöf, Långsele, Sweden), where the influence of tension wood and other anomalies were assumed to be smallest. The cores were then air-dried and glued to grooved wooden mounting boards (Speer, 2010). Afterwards, we sanded and scanned the cores at a resolution of 1,200 dpi (EPSON Expression 11000XL Model: J331A) (Supplementary Figure S1). We measured tree-ring widths (TRW) on the digital images with an accuracy of 0.01 mm using WinDENDRO software ver. 2014a (Regent Instruments Inc., Canada). The resulting TRW series were visually checked for characteristic rings and cross-dated using the software COFECHA with 50 years segments lagged successively by 25 years (Holmes, 1983). Cores that did not meet the default cross-dating correlation threshold were excluded, leaving a total of 346 cores available for further analysis. Tree age was estimated based on the number of tree rings counted from the bark to the pith (Table 2). The number of missing rings to the pith was approximated using the diameter and core length of each tree with the radius-length method (Norton et al., 1987).

To account for age and size-dependent trends, we applied rigorous detrending techniques since the stands varied in age and potential management intensity. As the focus was on single-year climate-growth interactions and short-term responses to drought events, flexible cubic smoothing splines with a 50% frequency cut-off at 25 years were used to detrend and standardize the raw ring-width series (Cook and Peters, 1981; Speer, 2010). We did not remove the first-order autocorrelation as it could significantly affect drought legacies (Yue et al., 2011). Detrended chronologies were built for all populations via Tukey’s biweight robust mean (Mosteller and Tukey, 1977) and truncated to the period of 1972–2021 to exclude younger tree life stages (Supplementary Figure S2). The signal strength of the final chronologies was assessed by the expressed population signal (EPS) and mean inter-series correlation (Rbar).

To assess long-term growth trends, we used the regional curve standardization (RCS) method (Briffa et al., 1983; Biondi and Qeadan, 2008) to age-detrend the ring-width series. We estimated the regional age trend for all investigated species by aligning the tree-ring series for each species based on cambial age. The estimated average growth curve was then used to detrend individual series. After this, we constructed RCS chronologies using Tukey’s bi-weight robust mean. We evaluated the significance of growth trends over the period from 1972 to 2021 using the Mann-Kendall trend test. All detrending procedures and the statistical analyses of the chronologies were performed using the software R (version 4.0, R Core Team, Vienna) with the “dplr” package (Bunn, 2008).

Large-scale pest outbreaks can significantly impact tree growth and confound the climate sensitivity analysis. Considering the occurrence of such events (Hirka, 2022), we assessed any potential biotic effects on radial growth by utilizing Cook’s distance in the linear regression between detrended growth and water balance (calculated from the previous September to the current August) for each population (Cook and Weisberg, 1982). This approach helped us identify any outlier years in the datasets that aligned with reports of biotic damage from local forest managers.

We assessed the relationship between climate variables (monthly temperature and precipitation, and derived indices FAI, WB, and Is) and standard chronologies using Pearson correlation coefficients. Monthly temperature and precipitation were examined throughthe response-function analysis taking into account the inter-correlations among the climatic variables using the “treeclim” package (Zang and Biondi, 2015). We analysed monthly meteorological data over the preceding 16 months, from June of the previous year until September of the actual year of ring formation. For FAI and Is, the correlation coefficient was computed on an annual basis, while for WB, we considered all seasons (March–May, June–August, September–November and December–February). Additionally, the correlation for WB was calculated for a 3-to-12-month window from August of the current year until September of the previous year. The significance of the correlations was tested using bootstrap resampling (Zang and Biondi, 2015). We also explored whether there were significant differences in the climate-growth correlations among different tree species within a specific stand. The Pearson r values were transformed to Fisher’s z to normalize the variance, followed by a pairwise t-test. This way we were able to determine if the differences among correlation coefficients were statistically significant. After identifying the climatic parameter with the highest correlations, we calculated the correlation coefficients for this parameter using a moving window approach (window size: 20 years, window offset: 1 year). This allowed us to search for temporal changes in growth sensitivity to this climatic driver over a period of 50 years.

We evaluated growth synchrony by calculating the mean inter-series correlation (Rbar), which reflects the average Pearson correlation among all tree-ring series within a specific chronology (Wigley et al., 1984). To obtain running synchrony values (Rbar), we used a 20-year moving window. Additionally, we assessed the significance of trends for the period from 1972 to 2021 using the Mann-Kendall trend test.

A certain year was considered a drought year when the standardized 12-month water balance (from August of the current year until September of the previous year) was lower than −0.84 (Fuchs et al., 2021b), regardless of any observed growth reductions, as suggested by Schwarz et al. (2020). The range of the standardized 12-month water balance varied from −2.2 to 3.4. We selected all drought events for conducting the superposed epoch analysis (SEA) of growth depressions. For multi-year droughts, we selected the year with the most negative water balance. An epoch of 11 years was chosen, encompassing five years before and after each drought year. SEA calculates the mean departure in growth performance for each year within the epoch from the mean of all analysed epochs per chronology (Lough and Fritts, 1987). To define 95% confidence intervals of the departures, we employed bootstrapping with 5,000 random draws from the respective chronology. SEAs were conducted using the “dplR” package in R (Bunn, 2008).

We calculated three indices of drought response for each population based on the detrended radial growth (Lloret et al., 2011). These indices assess how the trees withstand drought (resistance), recover from growth reduction during the drought (recovery) and their capacity to reach pre-drought growth levels (resilience). We analysed the effects of drought on these indices using reference periods of varying lengths: 1 year, 3 years and 5 years of mean radial growth before and after the drought event. This approach helped limit the influence of other factors, such as defoliation caused by insects (Schwarz et al., 2020). For multi-year droughts, the resistance and recovery indices were calculated by averaging the radial growth across successive drought years.

To test for differences in drought indices - resistance, recovery and resilience - among species and sites, a simple one-way ANOVA and for pair-wise comparison Tukey HSD test was used (Abdi and Williams, 2010). We also analysed the relationship between recovery and resistance for each tree species across all sites and available drought events using linear regressions by examining the ANOVA p-value from the interaction of resistance by species and comparing the slopes in the R package “lsmeans” (Lenth, 2016).

The mean annual tree-ring width (MRW) of the analysed species ranged from 1.3 to 3.0 mm, with no clear trend observed as we moved from moister to drier sites (Table 4). The standard deviation of ring widths for the two diffuse-porous species A. campestre and T. tomentosa was approximately 25% higher (ranging from 1.21 to 1.66 mm) compared to the ring-porous species Q. cerris, Q. pubescens and F. ornus, which ranged from 1.07 to 1.22 mm (Table 4). Among the sites, the lowest annual radial growth was observed in the Keszthely Mts. (1.88 mm) and the highest in the Somogy Hills (2.81 mm). The MRW did not show any significant relationship with tree age, suggesting that the populations experienced a stable growth rate during the period analyzed. Instead, the MRW of the tree populations was more closely associated with site conditions. Growth rates were lower in the Keszthely Mountains, which have low water-holding capacity, while areas with greater soil water capacity, such as Zselic, Vértes, and Szántód, exhibited higher growth rates (Tables 2, 4). The expressed population signal (EPS) was remarkably high, ranging from 0.85 to 0.99 for all species, consistently reaching or exceeding the generally accepted threshold of 0.85 (Wigley et al., 1984) (Table 4).

We observed only a few significant growth trends during the studied period. Specifically, the growth of A. campestre in Gödöllő Hills and Q. pubescens in Vértes Mts. has increased, while Q. cerris in Keszthely Mts. and T. tomentosa in the Zselic have shown a decline (Supplementary Figure S3).

The outlier analysis using Cook’s distance revealed that the growth of Q. cerris and Q. pubescens was significantly affected by the spongy moth outbreak (Lymantria dispar L.) in the Keszthely Mts. and the Somogy Hills sites in 2005 (Csóka and Hirka, 2009). This outbreak had a considerable effect on radial growth, obscuring the influence of climate during that year, leading us to exclude that year from further analysis in the affected populations.

Pearson’s correlation coefficients between monthly precipitation and ring widths were highest in May, June, and July of the current year, as well as in September of the previous year with r values reaching up to 0.52 (Figure 3A). Significant negative correlations with temperature were mainly observed from May to August of the current year with r values dropping to −0.48 (Figure 3B). Considering the intercorrelations among climatic variables considerably reduced the number of significant climatic variables (Supplementary Figure S4). Correlations of precipitation mostly remained significant during the May–July period and for September of the previous year (Supplementary Figure S4A). Furthermore, the temperature signals of May and June continued to show significance, particularly for A. campestre (Supplementary Figure S4B).

The derived indices demonstrated a strong correlation with radial growth. The mean correlation coefficient for FAI was r = 0.45, while the mean for Is was r = 0.55 when considering all populations. The water balance displayed the highest positive correlation (ranging from r = 0.36 to r = 0.63) with radial growth during the current summer across different populations. The water balances in the spring of the current year and the autumn of the prior year also positively influenced tree growth. The radial growth of A. campestre, Q. cerris and T. tomentosa showed a moderate correlation (ranging from r = 0.38–0.52) with summer water balance, but only Q. cerris and T. tomentosa showed a higher correlation with winter water balance (r = 0.29 for both species) compared to the other species.

The 12-month water balance between previous year September and current year August (WBsep-aug) exhibited the highest correlation among the monthly aggregated water balance variables with radial growth, achieving an average correlation coefficient of r = 0.6 across most populations (Figure 4).

This correlation was notably high for the climatically drier sites, with average coefficient values of r > =0.6, while it was comparatively lower for wetter sites (ranging from r = 0.42 to r = 0.55) (Figure 4). In the Keszthely Mts, the radial growth of Q. cerris and Q. pubescens showed the highest correlation with the four-month water balance from May to August of the current year. In contrast, A. campestre and T. tomentosa had stronger correlations with the summer water stress index (Is) in the Zselic (Figure 4).

The temporal change in correlations between the detrended chronologies and the seasonal water balance revealed the increasing (positive) effect of winter water balance on most sites and tree species, particularly over the past two decades (Figure 5). However, this trend was not apparent for A. campestre and T. tomentosa. Generally, the summer water balance played a crucial role for all populations, showing only a slight decline in the strength of its positive correlation in recent years. For the drier sites—GOD, VER, and SOM - the effect of spring water balance was significant, while it had a lesser impact on the wetter sites, KES and ZSE. Finally, the autumn water balance typically exhibited a negative correlation with radial growth, although some sites showed an increase in correlation (Figure 5).

Growth synchrony (Rbar) of the detrended chronologies was higher for Q. cerris, T. tomentosa and Q. pubescens than for A. campestre and F. ornus (Table 4). Generally, growth synchrony was lowest at the wettest site (Zselic) and highest at the driest site (Gödöllő) for most species. The lower growth synchrony of A. campestre at certain sites was due to the limited number of sampled trees and difficulties in accurately dating its tree rings, which often resulted from a high occurrence of uncertain ring boundaries. Notably, two-thirds of the populations exhibited a significant change in growth synchrony during the analysed period, most of which were increases (Figure 6). Recently, there has been a decrease in growth synchrony for most of populations except in Somogy Hills, where it has increased and remained at high levels for both Q. cerris and T. tomentosa over the past few decades (Figure 6).

Drought years were rare events between 1971 and 1990 but their frequency and severity increased significantly afterward (Supplementary Figure S5). We identified 5–7 drought years across different sites (1983, 1990, 1992–1993, 2000–2003, 2007, 2011–2012 and 2017). Superposed epoch analysis (SEA) revealed a significant (p < 0.05) reduction in growth for all populations in the drought years (Figure 7). Among the analysed tree species, Q. pubescens, Q. cerris and A. campestre experienced the strongest growth reductions. Most sites showed no significant differences between ring-porous and diffuse-porous species. However, T. tomentosa demonstrated a lag effect after drought years at both study sites, showing reduced growth in subsequent years (Figure 7). Moving further away from the drought year, the differences become increasingly difficult to interpret, particularly since droughts have occurred every 2–3 years in recent decades.

The combined data from all sites and drought events indicated comparable resistance levels among the five tree species (Figure 8). In terms of the recovery index, Q. cerris demonstrated the highest recovery value following drought, while F. ornus exhibited the least recovery in the 5 years after the drought event. Over the one-year reference period, a tendency of lower recovery can be seen for T. tomentosa (Supplementary Figure S6). The variability in recovery data was extensive for the two diffuse-porous species but more concentrated for the three ring-porous species. The growth resilience among the species was balanced (close to 1) with no significant differences noted during the five-year reference period, although a tendency for lower resilience in T. tomentosa was apparent in the distribution of index values (Figure 8).

Regression analysis of resistance and recovery for all species indicated that A. campestre, F. ornus and T. tomentosa generally have higher recovery values for lower resistance levels, than the two oak species (Figure 9). For the one-year reference period, we found no significant differences among the tree species, although the slope of regression was more gradual for T. tomentosa and F. ornus (Supplementary Figure S8). The slope of the linear regression between resistance and recovery values was significantly steeper for A. campestre compared to Q. cerris and Q. pubescens for both the three- and five-year reference periods (Supplementary Figure S9, Figure 9).

Under stressful climatic conditions, particularly at the distribution margins of tree species, strong environmental drivers such as drought would commonly affect entire populations, leading to high growth synchrony among individuals (Shestakova et al., 2016; del Río et al., 2021). In contrast, in optimal climatic conditions, individual competitive abilities can develop better, resulting in more diverse growth patterns. In our study sites, we observed a constantly rising summer water deficit over the study period (Supplementary Figure S10). However, most populations have shown a bell-shaped growth synchrony curve with decreasing tendencies in recent decades (Figure 6). Although the number of stands examined was small, this may suggest an increased within-population diversity in responses to drought, i.e., that some individuals have a better capacity to adapt to the stressor (Muffler et al., 2020). Therefore, decreasing growth synchrony could signify conditions conducive for natural selection leading to local adaptation under changing climatic conditions, whereas high synchrony may reflect a lower adaptive potential, even in the face of significant stress. Factors such as microenvironmental heterogeneity, differing availability of deep soil water pools, and variable intrinsic tree characteristics—like tree height—could also explain the observed decrease in growth synchrony (Vilà-Cabrera et al., 2019; Muffler et al., 2020; Ripullone et al., 2020; González de Andrés et al., 2021).

The water balance was more important for the radial growth of all species than precipitation or temperature alone. This indicates that soil water availability, in conjunction with atmospheric evaporative demand regulates the water status of leaves, stems, and roots, ultimately affecting cambial growth activity in the investigated sites (Trotsiuk et al., 2021). It is worth mentioning that beyond the available soil water, extreme heat events can impact a wide variety of tree functions. At the leaf level, photosynthesis may be reduced, photooxidative stress can increase, leaves may abscise and at the whole plant level the growth rates can decrease (Teskey et al., 2015). Of the seasons analysed, the summer water balance had the most substantial impact on growth, a finding that aligns with several studies conducted in Central Europe (Fuchs et al., 2021b; Kasper et al., 2022; Mészáros et al., 2022). In the preceding year, only the precipitation in September was notable for all tree species and there was no significant carry-over effect from the past summer, as observed with F. sylvatica (Di Filippo et al., 2007; Müller-Haubold et al., 2015). Most populations showed the strongest correlations between growth and the water balance of the preceding 12 months, between September of the previous year and August of the growth year, similar to the observations on Quercus species in Serbia (Stojanović et al., 2018). However, in the Keszthely Mts (KES) the growth of oaks (Q. cerris and Q. pubescens) showed the strongest correlation with the water balance between May and July. A very similar growth response was found for oak stands studied ~30–40 km westward from the KES site (Kern et al., 2009). This trend is likely due to the low soil water capacity at that site (Table 3), which might contribute significantly to the limiting summer conditions for growth.

We found that the dry populations were more sensitive to climate than those in wetter locations (Figures 4, 5). This aligns with regional and European-wide studies that reported stronger growth responses in areas with low water availability compared to wetter sites (Scharnweber et al., 2011; Bose et al., 2021; Bouwman et al., 2025). Following the year 2000, we observed an increasing positive correlation between radial growth and winter water balance (Figure 5). This upward trend was more pronounced for the studied deep-rooting oak species than for the other investigated species with a more horizontal root system (Crow, 2005; Taneda and Sperry, 2008). The likely reason for this change is the shift in the overall water balance in our study sites over recent decades, driven by increased evapotranspiration pressure from rising temperatures, while precipitation levels have largely remained unchanged (Figure 2) (Trotsiuk et al., 2021). Consequently, the tree species in the examined sites may increasingly rely on deep soil water sources due to worsening drought conditions during the growing season, as reported in other studies involving various tree species (Mészáros et al., 2022).

Lloret indices have revealed only a few significant differences in the drought response among the species. However, it was evident that A. campestre, F. ornus and T. tomentosa exhibited higher variability in their drought indices, particularly when compared to the two oak species studied (Figure 8, Supplementary Figure S6, S7). This observation aligns with previous studies, such as those by Kunz et al. (2018) and Schmucker et al. (2023), regarding A. campestre and Italiano et al. (2024) for F. ornus. In our study, A. campestre demonstrated a great capacity for recovery across all three reference periods, particularly during extreme droughts. In contrast, Q. pubescens showed the lowest variability of the drought indices and therefore the most balanced growth among the analysed tree species. This phenomenon is well-documented for temperate tree species that generally display lower resistance but tend to have higher recovery rates than their more resistant counterparts (Gazol et al., 2017; Schwarz et al., 2020). All examined tree species showed high resilience, with values around one, indicating their ability to maintain vitality and growth even after experiencing extreme droughts. Nonetheless, both the superposed epoch analysis and the Lloret indices revealed that T. tomentosa took longer to return to its pre-drought growth rate compared to the other species studied (Figure 7, Supplementary Figure S6). This finding is consistent with low resilience values for T. tomentosa in two- and five-year assessments following drought events in western Romania (Kasper et al., 2022). Leuschner et al. (2024) also found that Tilia (T. cordata Mill.) exhibits moderate drought resistance hardly withstanding extreme droughts, as noted by de Jaegere et al. (2016).

The seasonality and duration of drought events impact species resistance and resilience (D’Orangeville et al., 2018). However, due to the limited number of available drought events (ranging from 5 to 7 events per site), we could not take this into account. This is important because different tree species exhibit varying growth dynamics and wood anatomy, leading to different vulnerability to spring or summer droughts (Michelot et al., 2012). Additionally, in ring-porous species, the formation of earlywood is heavily influenced by the remobilization of stored carbon, making it less reflective of the actual weather conditions during that time (Michelot et al., 2012). Furthermore, we did not sample dead trees, as only a few were encountered during our study. Consequently, we were unable to evaluate any climatic factors that might have a fatal impact on tree growth.

In the absence of long-term mortality records for unmanaged stands of the studied species, it remains debatable whether tree species that are less responsive to droughts are more successful than those that are highly sensitive. Recently, Gessler et al. (2020) proposed that the delayed recovery of trees following a disturbance is not necessarily indicative of vitality loss due to the negative impacts of drought rather it may signify physiological acclimation processes. Drought-induced growth legacies and wood anatomical adjustments can lead to improved resistance to recurrent droughts (Tomasella et al., 2019). In this context, a reduction in growth following a drought can be viewed as a positive adjustment aimed at enhancing a tree’s long-term survival (Galiano et al., 2017). In contrast, fast-growing tree species often adopt a riskier strategy. After experiencing drought, damaged xylem conduits need to be rebuilt, which may decrease pest defence capacity and make these species more vulnerable to subsequent droughts (Beloiu et al., 2022).

The properties of the hydraulic system are also of central importance for the drought tolerance of tree species. The water-conducting channels in trees can withstand varying degrees of tension before embolisms—air bubbles forming in the xylem—occur. Hydraulic safety margins refer to the relationship between resistance at the xylem level and stomatal control (Meinzer et al., 2009). Specifically, these margins are defined as the difference between the minimum leaf water potential and a measure of xylem embolism resistance (e.g., P50). This represents a tree’s hydraulic strategy conservatism (Choat et al., 2012). Generally, Quercus species exhibit wider hydraulic safety margins and greater drought resistance (Cochard et al., 1992; Nardini and Pitt, 1999; Lobo et al., 2018). In contrast, Acer and Fraxinus species show variable but often moderate resistance (Schumann et al., 2019). Tilia, on the other hand, has narrower margins, making it more vulnerable to hydraulic dysfunction during severe drought conditions (Fuchs et al., 2021a). Conversely, the tree species, which display higher responsiveness to drought and narrower hydraulic safety margins may face considerable vulnerability to future climate change. In the current study, this is particularly true for A. campestre, F. ornus and T. tomentosa, consistent with findings by Kunz et al. (2018) in Germany. Additionally, other factors must be considered, such as the trunk’s water storage capacity, the rooting depth, and the competition among crowns and roots within a stand. Understanding the differences among tree species is essential for predicting how they will respond to climate change. This necessitates a more comprehensive approach to assessing the resilience of individual species and selecting the most appropriate species for planting in various environments.