Radial Increment of Beech (Fagus sylvatica L.) Is under a Strong Impact of Climate in the Continental Biogeographical Region of Croatia

European beech (Fagus sylvatica L.) is an important component of forests in the alpine and continental biogeographical regions of Croatia. This study aimed to (1) analyze the long-term response of beech to climate, (2) identify potentially critical climatic conditions that could negatively affect the radial increment (RI) and vitality of beech, and (3) evaluate differences in the response of beech between the two biogeographical regions in Croatia. We used the 16 × 16 km Croatian ICP Forests Level 1 network. On a total of 25 plots, we cored between 5 and 24 trees for dendrochronological analysis. Tree-ring widths (TRW) were measured and standardized using cubic spline. TRW chronologies for the two regions were calculated and correlated to the temperature and precipitation data and Standardized Precipitation and Evapotranspiration Index (SPEI) using bootstrapped correlations. Continental region precipitation from April to August and alpine region precipitation from June to August were significantly important for RI. Temperature was less important for RI than precipitation in both regions, but the importance of the negative impact of above-average temperatures in the continental region and the positive impact of above-average precipitation in the alpine region has increased over the last two decades. A comparison with the 3-month SPEI confirmed the significant influence of high temperatures and the lack of precipitation in August on the RI of beech trees in both regions.


Introduction
Meteorological and climatic conditions have direct and indirect effects on vegetation. The direct effects include responses to temperature, while the indirect effects primarily occur as soil-mediated phenomena, such as the influence of precipitation on soil moisture regimes [1][2][3][4]. Climate change likely represents the greatest threat to existing forest ecosystems in Europe [5]. However, there is considerable uncertainty regarding the magnitude and character of climate change, particularly at the regional level [6].
Extreme climatic events, such as droughts, are thought to be important in initiating changes in forest ecosystems [7]. Southeastern Europe is considered one of the most vulnerable hotspots, with an expected intensification of the severity and duration of droughts and heatwaves. This region has already experienced a high frequency of drought events. After the year 2000, significant droughts and heatwaves were observed in 2002, 2003, 2007-2008, Based on these ideas, we hypothesize that both temperature and precipitation significantly influence the radial growth of common beech, with the actual (micro) climatic conditions of the specific area playing a crucial role and drought having the most detrimental effect. Our main objectives were as follows: (1) to analyze the long-term growth response of beech to climate; (2) to identify extreme drought events that have negatively impacted the radial increment of beech; (3) to determine whether there are significant differences in the growth response of beech trees between the alpine and continental biogeographical regions of Croatia under changing climate conditions.

Radial Increment in the Alpine and Continental Biogeographical Regions of Croatia
The results ( Figure 1) show that radial increment is generally higher in the continental biogeographical region than in the alpine biogeographical region. The differences between the regions have remained stable over time, with radial increment mostly lower in the alpine region. However, there is a brief period between 1950 and 1970 when the two regions exhibit similar radial increments. Despite these differences, both tree-ring width chronologies show good agreement (t BP = 5.49, GLK% = 68.30%) and display a common response overall, particularly during extreme years such as 2000, 2003, and 2013. Dendroecological analyses typically utilize local data and provide valuable regional insights into growth responses to local habitat conditions and climate change. Successful upscaling of tree-ring data, however, requires dense networks covering the full range of bioclimatic and ecological conditions of the study species, as discussed by del Castillo et al. [37].
Based on these ideas, we hypothesize that both temperature and precipitation significantly influence the radial growth of common beech, with the actual (micro) climatic conditions of the specific area playing a crucial role and drought having the most detrimental effect. Our main objectives were as follows: (1) to analyze the long-term growth response of beech to climate; (2) to identify extreme drought events that have negatively impacted the radial increment of beech; (3) to determine whether there are significant differences in the growth response of beech trees between the alpine and continental biogeographical regions of Croatia under changing climate conditions.

Radial Increment in the Alpine and Continental Biogeographical Regions of Croatia
The results ( Figure 1) show that radial increment is generally higher in the continental biogeographical region than in the alpine biogeographical region. The differences between the regions have remained stable over time, with radial increment mostly lower in the alpine region. However, there is a brief period between 1950 and 1970 when the two regions exhibit similar radial increments. Despite these differences, both tree-ring width chronologies show good agreement (tBP = 5.49, GLK% = 68.30%) and display a common response overall, particularly during extreme years such as 2000, 2003, and 2013.  The expressed population signal [38] (EPS, Figure 1, lower panel) in both the alpine and continental biogeographical regions is high (>0.85) and stable from 1945 to 2020. This indicates that the climate-growth relationship calculation can be performed and that the results should have a meaningful statistical interpretation.

General Response of Radial Increment to Climate
Based on the high EPS value observed in the period 1945-2020 for both regions, we were able to perform a climate-growth analysis. The general pattern of the growth response of European beech to climate shows a smaller number of significant correlations between tree-ring widths and temperatures and precipitation for plots in the alpine region (Table 1). Correlations between tree-ring widths and temperatures and precipitation on the research plots in the continental region are significant, with higher absolute values compared to the alpine region (Table 1). The tree response pattern in both biogeographical regions is clear, indicating that precipitation has a positive influence on radial increment. In the alpine region, this influence is less pronounced than in the continental region, and the window where precipitation is significant is narrower compared to that in the continental region. In the alpine region, precipitation is important for radial growth during the summer months (June, July, and August), whereas in the continental region, this period extends from April to August. The correlations in the continental region also exhibit much higher absolute values compared to the alpine region (values shown in Table A1). Table 1. General pattern of climate-growth responses for each research plot for the period 1950-2020. Precipitation is indicated by the blue color, and temperatures are indicated by the light red color. The sign indicates the direction of the correlation, with "+" representing a positive correlation and "−" representing a negative correlation. Months are denoted by single capital letters from J (January) to S (September).

Alpine Biogeographical Region
No response detected 72 The influence of temperature on radial growth (Table 1) is negative in both biogeographical regions and less pronounced in the alpine region than in the continental region. In the alpine region, it is challenging to identify a clear pattern, as there are only indications that temperature has a generally negative influence on the radial growth of beech, while in the continental region, this negative influence is clearly present only in the summer months. The response of radial increment to climate differs between the two studied regions. An analysis of the correlation between tree-ring widths and the 3-month Standardized Precipitation and Evapotranspiration index (SPEI-3, Figure 2A) reveals a significant correlation with the summer months, with August SPEI-3 showing the highest correlation regardless of the biogeographical region. In terms of absolute values, the correlations in the continental region (r = 0.71) are higher than in the alpine region (r = 0.49). This indicates a strong dependency between radial increment and climatic conditions and indicates that aboveaverage summer temperatures and a lack of precipitation, in combination with increased evapotranspiration, reduce the radial growth of beech in both biogeographical regions.
biogeographical regions and less pronounced in the alpine region than in the continental region. In the alpine region, it is challenging to identify a clear pattern, as there are only indications that temperature has a generally negative influence on the radial growth of beech, while in the continental region, this negative influence is clearly present only in the summer months.
The response of radial increment to climate differs between the two studied regions. An analysis of the correlation between tree-ring widths and the 3-month Standardized Precipitation and Evapotranspiration index (SPEI-3, Figure 2A) reveals a significant correlation with the summer months, with August SPEI-3 showing the highest correlation regardless of the biogeographical region. In terms of absolute values, the correlations in the continental region (r = 0.71) are higher than in the alpine region (r = 0.49). This indicates a strong dependency between radial increment and climatic conditions and indicates that above-average summer temperatures and a lack of precipitation, in combination with increased evapotranspiration, reduce the radial growth of beech in both biogeographical regions. In the alpine region ( Figure 2B), correlations between tree-ring widths and average monthly temperature, monthly sum of precipitation, and their combinations are lower compared to the continental region. In the continental region ( Figure 2C), precipitation between April and August is the most important factor for radial increment (r = 0.65), meaning that more precipitation in that period results in a larger radial increment. A similar influence of precipitation was observed in the alpine region, but here precipitation within a shorter timespan (June to August, r = 0.48) was most important for radial growth. In the alpine region ( Figure 2B), correlations between tree-ring widths and average monthly temperature, monthly sum of precipitation, and their combinations are lower compared to the continental region. In the continental region ( Figure 2C), precipitation between April and August is the most important factor for radial increment (r = 0.65), meaning that more precipitation in that period results in a larger radial increment. A similar influence of precipitation was observed in the alpine region, but here precipitation within a shorter timespan (June to August, r = 0.48) was most important for radial growth.
Regardless of the biogeographical region, the response of radial increment to aboveaverage temperatures is generally negative. In the continental region, temperatures are important, but less so than precipitation ( Figure 2C). We found significant correlations between radial increment and temperatures in the summer months (May, r = −0.28; June, r = −0.36; and August, r = −0.33) and for the period from May to August (r = −0.39).
In the alpine region, temperatures are only important when considered as a combination of July and August temperatures (r = −0.28) or as August temperature alone (r = −0.26). Unlike the general response to climate (Figure 2), which illustrates how beech radial increment responds to climate parameters over a longer time span, moving window statistics provide insights into the temporal stability of the response and whether certain climate factors have gained importance as we approach the present time (e.g., the influence of a changing climate).
In the continental region, moving window correlation statistics (Figure 3, bottom) reveal a high, positive, significant, and stable relationship between radial increment and SPEI in June and July throughout the entire studied period. This indicates that aboveaverage precipitation, in combination with average temperature and evapotranspiration, positively impacts beech growth in the continental biogeographical region. Moreover, this influence has become increasingly important over time, as indicated by higher absolute values of the moving correlations as we approach the present. In contrast, the alpine biogeographical region shows a different response than the continental region. The moving window correlation shows a significant response only in July and August, but the response is not stable over time. Specifically, the period between 1960 and 2004 lacks any significant response during these months. However, it is worth noting that the importance of August SPEI has been increasing as we approach the present, indicating that precipitation (in combination with temperature and evapotranspiration) is gaining importance and will likely play an important role for the growth of beech in the future.
between radial increment and temperatures in the summer months (May, r = −0.28; June, r = −0.36; and August, r = −0.33) and for the period from May to August (r = −0.39).
In the alpine region, temperatures are only important when considered as a combination of July and August temperatures (r = −0.28) or as August temperature alone (r = −0.26).

Temporal Stability of the Climate Signal and Spatial Outreach of the Climate Signal in Beech Tree Rings
Unlike the general response to climate (Figure 2), which illustrates how beech radial increment responds to climate parameters over a longer time span, moving window statistics provide insights into the temporal stability of the response and whether certain climate factors have gained importance as we approach the present time (e.g., the influence of a changing climate).
In the continental region, moving window correlation statistics (Figure 3, bottom) reveal a high, positive, significant, and stable relationship between radial increment and SPEI in June and July throughout the entire studied period. This indicates that aboveaverage precipitation, in combination with average temperature and evapotranspiration, positively impacts beech growth in the continental biogeographical region. Moreover, this influence has become increasingly important over time, as indicated by higher absolute values of the moving correlations as we approach the present. In contrast, the alpine biogeographical region shows a different response than the continental region. The moving window correlation shows a significant response only in July and August, but the response is not stable over time. Specifically, the period between 1960 and 2004 lacks any significant response during these months. However, it is worth noting that the importance of August SPEI has been increasing as we approach the present, indicating that precipitation (in combination with temperature and evapotranspiration) is gaining importance and will likely play an important role for the growth of beech in the future.  The relationship between radial increment and April to August precipitation in the continental region was found to be high, significant, and stable throughout the entire studied period ( Figure A1). Conversely, temperatures between May and August have only become important in recent decades, indicating that the influence of above-average temperatures in the continental region is gaining importance and may pose a serious threat to the growth of beech in the future.
In the alpine region, the radial increment of beech has been affected by precipitation only in recent decades, although this influence is not as pronounced as in the continental region. In addition, the critical months for radial growth (June to August) differ from those in the continental region (April to August, Figure A2). In recent decades, temperatures have also been gaining importance for radial growth in the alpine region.
We also performed a spatial analysis of the climate-growth relationship by correlating grid cells of meteorological data and indexed tree-ring width series for the period 1950-2019 for temperature and precipitation, and from 1950 to 2018 for SPEI ( Figure 4). The spatial extent of the climate signal in tree-rings is smaller for the alpine region ( Figure 4, left panel) than for the continental region ( Figure 4, right panel), both in terms of spatial extent and the magnitude of the correlation values. In particular, the temperature signal in the alpine region is spatially weak and inconclusive. The precipitation signal in the alpine region has a slightly larger spatial extent and corresponds well with the Dinaric Mountain range. The spatial extent of the tree-ring width correlation with SPEI is the largest of all three analyzed parameters in the alpine biogeographical region, although in terms of absolute values it is not significantly higher.   In contrast, the spatial extent of the climatic signal in beech tree rings is extensive in the continental region ( Figure 4, right panel), particularly for precipitation, covering a large area of Pannonian lowland, including the neighboring countries of Hungary, Bosnia and Herzegovina, and Serbia. The signal corresponds well with the precipitation pattern in the Pannonian lowland, indicating that the lowland and the hills surrounding it are under significant threat from summer drought. The spatial extent of correlation between the tree-ring width and SPEI is the highest among the three studied parameters.

Response of the Radial Increment of Beech in Extreme Years
In our study area, all beech trees on all sites show a synchronous response to climate conditions. By means of pointer year analysis (PY), where we identify plot-wise common narrow or wide tree rings, we identified several positive and negative pointer years in both the alpine and continental biogeographical regions ( Figure 5). In general, the alpine region has a greater number of PY (9 positive and 10 negative) than the continental region. The continental region has five positive and eight negative PY, which is unexpected given the weak climate signal in the tree-ring width chronology of beech in the alpine region.
Only a few PY were common to both biogeographical regions. The common positive PY were 1989 and 2014, while the common negative PY included 1957PY included , 1971PY included , 1988PY included , and 2000.
monthly sum of precipitation for the alpine biogeographical region, and the combination of August average monthly temperature and April-August monthly sum of precipitation f continental biogeographical region. In both regions, the analysis includes August SPEI.
In contrast, the spatial extent of the climatic signal in beech tree rings is extens the continental region ( Figure 4, right panel), particularly for precipitation, cover large area of Pannonian lowland, including the neighboring countries of Hungary, B and Herzegovina, and Serbia. The signal corresponds well with the precipitation p in the Pannonian lowland, indicating that the lowland and the hills surrounding under significant threat from summer drought. The spatial extent of correlation bet the tree-ring width and SPEI is the highest among the three studied parameters.

Discussion
Dendroecological methods provide valuable insights into the response of forest trees, as tree-ring widths are strongly controlled by climatic events and as such are an important indicator of changes in the environment and climate [40].
In terms of growth synchronicity, the radial growth of beech in the alpine and continental biogeographical regions is comparable. However, there are differences in terms of absolute radial increment. This discrepancy could be attributed to the fact that the studied beech trees in the alpine region originate from older forest stands where the forest stand structure is closer to old-growth forests [30]. In contrast, in the continental region, beech forest management is more intense, and forest stands are mostly even aged and composed of younger trees, which usually have a more pronounced response to climate than older trees [41]. Both biogeographical regions exhibit an increase in radial growth from the 1960s to the 1980s. However, a decrease in radial growth is observed after 1980, which is consistent with studies of beech radial growth in other parts of Europe [32,37]. Model predictions show that beech growth is expected to decline by as much as 40-50% in the Balkan Peninsula and in the Apennines in the period 2040-2070 [37]. Changing climatic conditions, characterized by more frequent hot spells and drought periods, are among the most important factors contributing to the decline in radial growth [32,42].
When we merge the information gathered with the pointer year analysis and the results from the climate-growth relationship (including temporal analysis using moving window statistics), we can conclude that beech exhibits a strong general climate signal in tree rings and is highly responsive to extreme climatic events, such as extremely dry and hot years.
Beech is a distinct mesophyte, which means that it is not adapted to either particularly dry or particularly wet environments and can therefore benefit from large amounts of precipitation in the summer months [30]. However, the predicted increase in drought events, particularly in the continental biogeographical region, could have a significant negative impact on beech growth and regeneration [32]. Our findings are in accordance with research conducted in Hungary, where the availability of water (precipitation) in the May-July period is critical for the survival of beech [43]. Similar patterns have been observed in Italy, where precipitation plays a major role along the Apennine Mountains [44,45]. In the alpine biogeographical region, temperature appears to be of marginal importance compared to the continental region, where it plays a more important role, especially when combined with a lack of precipitation, as indicated by greater values of the correlations with SPEI-3.
Trees are always good indicators of climatic and environmental conditions. Tree growth in extremely dry and wet years not only shows their immediate response to the unfavorable conditions but also the conditions limiting the survival of the species, since tree vitality can be defined as the ability of a tree to assimilate, survive stress, and react to changing conditions [46]. Therefore, indicators of tree life, growth, and development have to be used [24,47]. Beech growth is above average in years with more precipitation during the growing season and below average in years with low amounts of precipitation and above-average temperatures. Based on the response of beech in PY, it is evident that the change of climate towards warmer and dryer conditions predicted for the future [48] will have dramatic consequences on European beech growth, and it is entirely possible that beech will not be able to adapt to the changed climatic conditions and that we will face increased mortality of beech trees, especially in the continental biogeographical region. Climate change and extreme climatic conditions such as drought are predicted to lead to a decrease in suitable forest sites for European beech [49]. It is predicted that 56-96% of today's beech forests could be out of their current bioclimatic niche by 2050 [43], mostly at the southern border of common beech's range [50]. Extreme climatic events, such as heat waves and dry periods, are expected to occur with increased frequency in Croatia as a whole and even more frequently in the continental biogeographical region of Croatia [51]. Severe and recurrent droughts have been identified as a major factor contributing to accelerated rates of tree decline and forest mortality in Europe [52].
Compared to the continental biogeographical region, the alpine biogeographical region is less exposed to increasing temperatures and a lack of precipitation in the summer months. According to the regression models [53,54], precipitation amounts are lower at lower altitudes, and air temperature is higher in the continental region. This observation is in line with the fact that the climate-growth response of beech in the alpine biogeographical region is less obvious, as correlations between tree-ring widths and different meteorological parameters were not as high as in the continental region. With regards to the effect of droughts, a higher number of dry years is recorded at meteorological stations at lower altitudes in comparison with higher altitudes [55]. This indicates that growing conditions for beech in the alpine region are more favorable than in the continental region, which is clearly visible in the climate-growth response statistics. Similar results were obtained by Kolář et al. [56] for the eastern part of Czech Republic, where beech sites above 800 m are more temperature dependent and sites below 800 m more precipitation dependent [56].
To summarize, the radial growth of beech trees in Croatia's alpine and continental biogeographical regions is affected by climate change, especially in the continental region. Climate influences radial increment to such a degree that further intensification of climate warming may seriously impact the future productivity of European beech in the continental region of Croatia.

Research Plots
We used the infrastructure of UNECE ICP Forests in Croatia for our field work. The 16 × 16 km ICP Forests grid has a pure systematic sampling design, which extends across Europe (referred to as Level I plots). The plots do not have a fixed area: 24 trees are chosen for assessment using a cross-cluster system [57]. Besides defoliation, this database contains information on plot characteristics (location, topography, stand age). For the tree-ring width study, a set of 25 plots, 8 in the alpine and 17 in the continental biogeographical regions of Croatia, were chosen based on the selection criteria that at least 5 out of the 24 trees were European beech (Fagus sylvatica L.) ( Figure 6). Basic information about the sampling locations is presented in Table A2. different meteorological parameters were not as high as in the continental region. With regards to the effect of droughts, a higher number of dry years is recorded at meteorological stations at lower altitudes in comparison with higher altitudes [55]. This indicates that growing conditions for beech in the alpine region are more favorable than in the continental region, which is clearly visible in the climate-growth response statistics. Similar results were obtained by Kolář et al. [56] for the eastern part of Czech Republic, where beech sites above 800 m are more temperature dependent and sites below 800 m more precipitation dependent [56].
To summarize, the radial growth of beech trees in Croatia's alpine and continental biogeographical regions is affected by climate change, especially in the continental region. Climate influences radial increment to such a degree that further intensification of climate warming may seriously impact the future productivity of European beech in the continental region of Croatia.

Research Plots
We used the infrastructure of UNECE ICP Forests in Croatia for our field work. The 16 × 16 km ICP Forests grid has a pure systematic sampling design, which extends across Europe (referred to as Level I plots). The plots do not have a fixed area: 24 trees are chosen for assessment using a cross-cluster system [57]. Besides defoliation, this database contains information on plot characteristics (location, topography, stand age). For the treering width study, a set of 25 plots, 8 in the alpine and 17 in the continental biogeographical regions of Croatia, were chosen based on the selection criteria that at least 5 out of the 24 trees were European beech (Fagus sylvatica L.) ( Figure 6). Basic information about the sampling locations is presented in Table A2.

Sample Collection and Tree-Ring Width Analysis
All European beech trees on the selected Croatian Level I plots ( Figure 6) were cored (two cores per tree) following methods described by Stokes and Smiley [58].
Each core was air-dried, mounted, and sanded with progressively finer sanding paper (up to 600 grit) on an industrial belt sander to achieve a highly polished surface and excellent visibility of tree rings [58]. The cores were then digitized using an ATRICS system [59], and the annual radial growth was measured to the nearest 0.01 mm using CDendro and CooRecorder software v. 9.8.1 (Cybis, Stockholm, Sweden). Each tree-ring series underwent visual and statistical crosschecking using PAST-5 software v. 5.0.610 (SCIEM, Vienna, Austria). In case of any dating or synchronization errors, appropriate corrections were made in CDendro software, and measurements were repeated if necessary.
Individual tree-ring width series were standardized to remove long-term trends [60,61] using a 67% cubic smoothing spline with a 50% frequency cut-off in the dplR library [62] of the R program [63]. Departures of the measured values from the regression curve were calculated as the quotient between the measured tree-ring width and fitted value, resulting in a dimensionless index with a mean of 1. By doing this, we eliminated factors that were not related to climate, such as tree age and the effects of stand dynamics [64]. Index values were pre-whitened using an autoregressive model that was selected based on the minimum Akaike criterion and were combined across all series using a biweight robust estimation of the mean in order to exclude the influence of outliers. The dplR produces two types of chronologies: standardized (STD) and residual (RES). In this research, we used the RES chronology, which is a robust estimate of the arithmetic mean with autocorrelation mathematically removed [65].
Composite chronologies for the alpine and continental biogeographical regions were compiled as follows: (1) all sites from one biogeographical region were merged into a single file and visually and statistically synchronized and cross-dated; (2) they were then standardized using the above-described procedure in the dplR library of the R program; and finally, (3) three different chronologies for each biogeographical region were produced: a raw data chronology, a standardized chronology, and a residual chronology.

Meteorological Data
Climatic data were collected from the KNMI Climate Explorer website [66,67]. For temperature and precipitation, we used monthly gridded data (0.25 × 0.25 • grid) from the E-OBS database [68] covering the period between 1920 and the present. For the drought index, we chose the Standardized Precipitation and Evapotranspiration Index (SPEI) for the period 1901-2018 from the CSIC [69,70], also available at KNMI Climate Explorer.
To identify potential anomalies between gridded and local climate data, we compared the gridded meteorological climate data with shorter local climate data. We used three meteorological stations located in the alpine region (Delnice, Gospić, and Ogulin) and three meteorological stations in the continental region (Zagreb, Bjelovar, and Požega) for comparison with the gridded data. To verify positive and negative pointer years in tree rings, an analysis of monthly and annual climate anomalies (precipitation and temperature percentiles) was conducted according to the modified Conrad-Chapman method [71] based on the deviation from the multi-year average of the period 1961-1990.
The alpine region has a maritime precipitation regime where less than 50% of precipitation falls in the warm part of the year. Precipitation has two maxima, one in early spring and the other in late autumn. It is the area with the highest values of precipitation in Croatia [30]. The continental region in Croatia, on the other hand, has a continental precipitation regime where more than 50% of the precipitation falls in the warm part of the year. Precipitation maxima occur in spring and late summer (Figure 7).

Statistical Data Processing
The expressed population signal (EPS) was used to assess the representation o small sample in relation to the signal of the entire population [38]. EPS values ran between 0 and 1, with values greater or equal to 0.85 considered high enough to indica a common signal across the population [38]. This common signal can be related environmental or climatological factors; however, in many cases, it is a climate sign contained within tree rings (see Section 2.2). In this study, EPS was used as a measure the common signal in the composite chronology.
After de-trending, removing autocorrelation, and calculating EPS, tree-ring ind chronologies were compared to average monthly temperatures, monthly sum precipitation, and 3-and 6-month SPEI using a bootstrapped correlation calculation in t treeclim library [72] of the R program [63]. Additionally, several combinations of month temperature and precipitation data were tested against the standardized tree-ring widt to identify the best possible combination of influential climate variables. A 25-ye window with a 1-year overlap for the calculation of the bootstrapped correlation betwe monthly temperature, precipitation, SPEI, and tree-ring widths was used to assess t temporal stability of the climate-growth relationship. Two SPEI indices, on a 3-and month scale, were tested, and after careful examination of results, it was decided to u only the 3-month SPEI for further analysis.
Spatial analysis based on the correlation between standardized tree-ring widths a the gridded meteorological data (average monthly temperature, monthly sum Climate diagrams for the alpine and continental biogeographical regions in Croatia for the period 1990-2021. Black color within figure-months with precipitation above 100 mm; blue color-wet, humid period. Dark grey color on x-axis-months with mean minimum air temperature <0 • C; light grey color-months with absolute minimum air temperature <0 • C. Numbers on the left y-axis from top to bottom: absolute maximum air temperature, mean maximum air temperature, average fluctuation of air temperature, mean minimum air temperature, and absolute minimum air temperature.

Statistical Data Processing
The expressed population signal (EPS) was used to assess the representation of a small sample in relation to the signal of the entire population [38]. EPS values range between 0 and 1, with values greater or equal to 0.85 considered high enough to indicate a common signal across the population [38]. This common signal can be related to environmental or climatological factors; however, in many cases, it is a climate signal contained within tree rings (see Section 2.2). In this study, EPS was used as a measure of the common signal in the composite chronology.
After de-trending, removing autocorrelation, and calculating EPS, tree-ring index chronologies were compared to average monthly temperatures, monthly sum of precipitation, and 3-and 6-month SPEI using a bootstrapped correlation calculation in the treeclim library [72] of the R program [63]. Additionally, several combinations of monthly temperature and precipitation data were tested against the standardized tree-ring widths to identify the best possible combination of influential climate variables. A 25-year window with a 1-year overlap for the calculation of the bootstrapped correlation between monthly temperature, precipitation, SPEI, and tree-ring widths was used to assess the temporal stability of the climate-growth relationship. Two SPEI indices, on a 3-and 6-month scale, were tested, and after careful examination of results, it was decided to use only the 3-month SPEI for further analysis.
Spatial analysis based on the correlation between standardized tree-ring widths and the gridded meteorological data (average monthly temperature, monthly sum of precipitation, and 3-month SPEI) (source of data: KNMI Climate Explorer) was used to test the spatial strength (outreach) of the climate signal in tree rings. Single months and combi-nations of months were used in the spatial analysis. Each grid cell statistic was tested for significance, and where correlations were statistically significant (p < 0.05), the value was displayed on the map [66].
Pointer year analysis, which examines extreme events in tree-ring width chronologies, was performed for chronologies from both the alpine and continental biogeographical regions. We used a slightly modified criteria for pointer year selection, as described in Schweingruber et al. [73]. A year was recognized as a pointer year when 70% of at least 70 trees per biogeographical region responded with an increase or decrease in tree-ring width in comparison to the prior year. The dplR library in R was used for the calculation of the pointer years.
Statistical analysis was conducted using the R libraries dplR [62] and treeclim [72], and graphs were created using the IgorPRO program. Climate diagrams, following Walther [74], were created in the the KlimaSoft 1.

Data Availability Statement:
The data presented in this study are, due to legal reasons, only partly available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Table A1. Correlations between tree-ring width and average monthly temperature, monthly sum of precipitation, and 3-month SPEI for the alpine (A) and continental (C) biogeographical regions.