No detrimental long-term impacts of coring on tree growth or mortality across European forests

25 1. Tree cores are widely used across a broad range of disciplines in the environmental 26 sciences, most notably as a tool to measure tree growth, estimate tree age, characterise 27 wood anatomy and reconstruct past climate. However, because extracting tree cores is 28 an invasive procedure, concerns about their use are often raised due to perceived risks 29 for tree health. 30 2. Here we comprehensively test the long-term impacts of tree coring on 16 European 31 tree species using a dataset spanning the entire European continent. Over the course 32 of a decade, we tracked the growth and survival of 3334 trees cored in 2012 33 (including trees cored once and twice) and compared them to that of a cohort of 7413 34 neighbouring trees that were never cored. 35 3. We found no evidence that coring had a detrimental impact on either the growth or 36 survival of trees, irrespective of their size, species, climatic environmental or the 37 number of times they were cored. However, we did observe a small positive stem 38 increment response (2.0% for trees cored once and 6.2% for trees cored twice), which 39 we hypothesise is most likely the result of vertical scarring from the coring wound, 40 with potential consequences for the accuracy of repeated diameter measurements 41 collected at the same height. 42 4. Our study supports the use of tree coring as a low-impact method for characterising 43 the growth, age and function of a wide range of tree species. However, to avoid 44 biasing long-term forest census measurements, tree cores should always be collected 45 well above or below the point of measurement of tree stem diameters. 46 Key words: Dendrochronology; forest census data; tree coring; tree growth; tree mortality; 47 coring impact 48 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [3]

49 Tree coring is central to the science of dendrochronology and is used widely in ecology as a 50 tool to reconstruct the age, growth rate and the ecophysiological status of trees. Data 51 extracted from tree cores are used to characterise how tree growth rates vary through time 52 and in response to biotic and abiotic drivers (Jucker et al., 2014; Anderson-Teixeira et al., 53 2022) and quantify forest productivity and carbon storage (Babst et al., 2014; Li, Speer and 54 Thapa, 2024). Tree cores are also useful to understand how trees respond to climate extremes 55 (Anderegg et al., 2015; Sohn, Saha and Bauhus, 2016; Castagneri et al., 2022; Wei et al., 56 2024), measure wood traits related to plant hydraulics and physiology (Fonti et al., 2009; 57 Grossiord et al., 2014) and reconstruct past climate (Esper et al., 2016; Emile-Geay et al., 58 2017; Rodriguez-Caton et al., 2024). Moreover, historical growth reconstruction from tree 59 rings is increasingly being combined with other spatio-temporal datasets, such as sate llite 60 imagery, to track forest responses to climate change at scale (Kannenberg et al., 2019; 61 Levesque et al., 2019; Mathes et al., 2024; Mašek et al., 2024; Morin-Bernard et al., 2024). 62 However, due to its invasive nature, there have long been concerns about the potential 63 adverse effects it could have on tree health (Tsen, Sitzia and Webber, 2016) and because of 64 this, a precautionary approach has meant tree coring is often not permitted in conservation 65 areas and in long-term ecological monitoring sites. 66 Tree cores are extracted using an increment borer, which leaves behind a hole in the trunk 67 that is typically 5–10 mm in diameter (depending on the increment borer used) and extends 68 from the outer bark to the centre of the tree. These bore holes are thought to potentially 69 impact tree health, growth and physiology in several ways. First and foremost, they can serve 70 as a site of infection, particularly for fungal pathogens in warm, wet climates (Florens, 2013, 71 2014; Boura, Péchon and Gigord, 2014). These infections can lead to wood staining and can 72

in wood decay, although from the limited evidence available this appears to be more 73 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [4] common in broadleaf trees (Tsen, Sitzia and Webber, 2016) and less prevalent in conifers 74 (Wunder et al., 2013). The wound caused by extracting the cores can also lead to a localised 75 growth response by trees. For example, Picea abies was shown to have elevated localised 76 growth near coring sites due to vertical scarring and callous tissue overgrowing the wound 77 (Fabiánová and Šilhán, 2021), and large, localised scarring responses have also been reported 78 in Fagus sylvatica (Portier et al., 2023). The effects of tree coring on tree vitality may be 79 particularly pronounced in smaller trees, for which bore holes represent a proportionally 80 much larger wound and which may have fewer resources to repair damaged tissue (Datta et 81 al., 2025). Collectively, this work suggests that further evidence is needed to determine the 82 impacts of coring on tree health, survival and wood quality (Tsen, Sitzia and Webber, 2016; 83 Holl and Road, 2018). 84 Our current understanding of the long-term effects of coring on tree health is pieced together 85 from a relatively small number of studies that have focused on a limited number of tree 86 species and study sites (Mantgem and Stephenson, 2004; Wunder et al., 2011; Helcoski et al., 87 2018; Portier et al., 2023). In particular, only a handful of studies have explored the effects of 88 coring on tree mortality. These studies leveraged large sample sizes (>500 cored trees) and 89 found no adverse effects over a 7–40 year period (Mantgem and Stephenson, 2004; Wunder 90 et al., 2011; Helcoski et al., 2018), but were all conducted at a single location, and only one 91 explored the effects of coring across a significant number (19) of species (Helcoski et al., 92 2018). Similarly, to the best of our knowledge only one study has investigated the long-term 93 impacts of coring on tree growth, reporting no decrease in growth in the 10 years after coring 94 in three common European tree species (Picea abies, Abies alba, and Fagus sylvatica) 95 sampled in Switzerland and the Ukraine (Portier et al., 2023). This limited body of evidence 96 suggests that coring is unlikely to substantially impact tree health. However, we currently 97 lack a comprehensive and systematic assessment of the long-term impacts of coring on tree 98 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [5] growth and mortality that not only captures responses across a diverse range of species but 99 also tests whether effects differ among climatically distinct forest types, tree sizes, and coring 100 practices (e.g., extracting two cores per tree to improve cross dating). 101 Here, we provide the first comprehensive overview of the long-term impacts of tree coring on 102 the growth and survival of 16 common European tree species, comprising 12 broadleaves and 103 four conifers. Over the course of a decade, we tracked the growth and survival of 10,747 trees 104 – 3334 of which were cored in 2012 – distributed across 205 permanent forest plots covering 105 Europe’s major forest types. This includes trees that were cored once (2730) or twice (804) 106 spanning a broad range of sizes (7.5–101.5 cm in stem diameter). Using this unique dataset, 107 we compared the growth and survival of cored and non-cored trees from the same plots and 108 set out to test the following predictions: 109 1. Based on previous studies (Mantgem and Stephenson, 2004; Wunder et al., 2011; 110 Helcoski et al., 2018; Portier et al., 2023), we do not expect tree coring will have a 111 detrimental effect on the long-term growth and survival of most European tree species 112 and forest types. 113 2. If a detrimental effect of coring is observed, we expect this negative impact will be 114 disproportionately pronounced in smaller trees, as these are more vulnerable to 115 internal decay and wounds (Datta et al., 2025). 116 117 3. Whilst often cautioned against (Tsen, Sitzia and Webber, 2016; Portier et al., 2023), 118 we do not expect that collecting a second core will result in any additional negative 119 impact on tree growth or survival compared to taking a single core. As the relative 120 size of the wound to the stem is still small. 121 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [6]

122 FunDivEUROPE forest plot network 123 Our analyses are based on data from the FunDivEUROPE permanent forest plot network 124 (Baeten et al., 2013). The network consists of 209 plots (30×30 m in size) distributed across 125 six sites, spanning three biomes (Mediterranean, Temperate, and Boreal) (Fig. 1), and four 126 distinct forest types across Europe (Mediterranean, Temperate, Boreal, and Hemiboreal). The 127 plots cover a latitudinal gradient of more than 20°, ranging from Mediterranean forests in 128 Spain and Italy, through temperate forests in Germany and Romania, to hemiboreal forests in 129 Poland and boreal forests in Finland. The network was established in 2012 within mature 130 forest stands with no recent management history prior to plot establishment and has since 131 been re-censused in 2017 and 2022. Of the original 209 plots, 205 were surveyed across all 132 three census periods; the remaining plots were abandoned due to harvesting activities that 133 occurred after 2012. 134 Within each plot and at each census interval, all stems with a diameter at breast height (DBH, 135 measured in cm at 1.3 m aboveground) greater than 7.5 cm were tagged, measured, identified 136 to species and their survival was recorded. In addition, the crown illumination index (CII) of 137 each tree was recorded as a measure of light availability to the crown that varies from 1 138 (completely suppressed) to 5 (fully exposed) (Clark and Clark, 1992). In total, we tracked the 139 growth and survival of 10,747 trees belonging to 16 common European tree species, several 140 of which were sampled at more than one site (Table 1). This includes four conifers (Abies 141 alba, Picea abies, Pinus nigra, Pinus sylvestris) and 12 broadleaves (Acer pseudoplatanus, 142 Betula pendula, Carpinus betulus, Castanea sativa, Fraxinus excelsior, Fagus sylvatica, 143 Ostrya carpinifolia, Quercus cerris, Quercus faginea, Quercus ilex, Quercus petraea, 144 Quercus robur). 145 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [7] Tree coring 146 In 2012, when the FunDivEUROPE plots were first established, we acquired tree cores from 147 a subset of trees in each plot to measure their historical growth rates and capture their 148 physiological responses to drought using carbon isotopes (Grossiord et al., 2014; Jucker et 149 al., 2014). In total, 2730 trees were cored for growth reconstruction (approximately 20–25% 150 of trees per plot), while a further 1408 tree cores were collected for isotope analysis (Jucker 151 et al., 2017). Most of the cores collected for isotope analysis overlapped with those sampled 152 for growth, meaning that 804 trees were cored twice. In each case, cores were collected using 153 5.15 mm diameter increment borers (Haglöf AB, Sweden) and were sampled as close as 154 possible to the pith. Trees for growth analysis were selected following a size-stratified 155 random sampling approach (Jucker et al., 2014), while those for isotope analysis were 156 randomly selected from a subset of dominant and co-dominant trees in each plot (Grossiord et 157 al., 2014) (Fig. 1). For a detailed description of how the cores were used to measure rates of 158 tree diameter growth, see Jucker et al. (2014). 159 For the purposes of this analysis, we only retained data from plots with a complete census 160 record between 2012 and 2022. Moreover, we removed all data from Q. ilex trees in Spain, as 161 the majority of these were multi-stemmed and had not been individually tagged in 2012. 162 Consequently, we could not be certain that the same stem was consistently re-surveyed across 163 each census period. In total, this left us with 3337 trees that were cored in 2012 (2532 cored 164 once and 805 cored twice) and 7422 trees that were not cored (Table 1). 165 Impacts of tree coring on growth 166 We assessed the impacts of tree coring on growth using two complementary approaches, the 167 first based on a control–impact design and the second on a before–after design (Christie et 168 al., 2019). First, we compared the growth of trees that were cored in 2012 to those that were 169 not cored in the decade after coring took place using the re-survey data from 2022. To 170 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [8] complement this, for the trees that were cored in 2012 and survived until 2022, we compared 171 their growth rates in the decade before (2001–2011, estimated from tree ring width) and after 172 coring (2012–2022, estimated from measured basal area). In all cases, growth was expressed 173 as a cumulative basal area increment (BAI, in cm2 decade-1). Note that because of 174 measurement errors in DBH that are common with re-census plot data (especially for slow 175 growing trees), a small subset of trees (7.5%) exhibited negative growth values between 2012 176 and 2022 (Fig. S1). These negative growth values were retained in the analysis, as removing 177 them would bias growth estimates as corresponding positive growth errors are much harder to 178 detect and correct (Talbot et al., 2014; Réjou-Méchain et al., 2017). 179 Approach 1: control–impact assessment 180 To compare the growth rates of trees that were cored to those that were not between 2012 and 181 2022, we used the census data to fit a hierarchical Bayesian model in which the BAI of tree i 182 of species s from country c in plot p was modelled following a skewed normal distribution, 183 which best captured the residual error structure of the model (Fig. S2): 184 𝐵𝐴𝐼!,#$,%~𝒮𝒩'𝜇!,#$,%, 𝜎+ (1) 185 where 𝜇!,#$,% is a function of initial tree size (BA2012), the crown illumination index (CII) 186 recorded in 2012 (treated as a continuous variable), a categorical variable describing the 187 coring (C) treatment (factor with three levels: not cored, cored once, cored twice), and the 188 interaction term between coring and initial tree size (C × BA2012): 189 𝜇!,#$,% = 𝛼#$ + 𝛽&#$𝐵𝐴'(&'! + 𝛽'#$𝐶𝐼𝐼! + 𝛽)#$𝐶! + 𝛽*#$(𝐶 × 𝐵𝐴'(&')! + 𝛾% (2) 190 αcs and β1-4cs are country- and species-specific model parameters representing the intrinsic 191 growth rate (αcs), and growth response (i.e., slope) to tree size (β1cs), CII ( β2cs), coring ( β3cs), 192 and how the effect of coring varies depending on tree size (β4cs). Additionally, the intercepts 193 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [9] were allowed to vary among plots (γp) to capture additional variation in BAI among stands. 194 BA2012 and CII were included in the model to account for the well-known effects of tree size 195 and competitive status on growth (Jucker et al., 2014) allowing us to account for differences 196 in the size distribution of trees that were and were not cored (Fig. 1). The interaction term 197 between coring and tree size was included to test whether smaller trees are disproportionally 198 impacted by coring. 199 Country-species-specific model parameters (αsc, β1-4cs) were modelled using a multivariate 200 normal distribution: 201 8 α#$ β&#$ ⋮ β*,- < ∼ MVNormal ⎝ ⎜ ⎛8 α β& ⋮ β* < , S ⎠ ⎟ ⎞ (3) 202 where α represents the community-level intercept, β1-4 the overall effect of covariates on BAI 203 across all species and S is a covariance matrix. Modelling all country-species-level 204 parameters as a multivariate normal distribution allows sharing information across country-205 species combinations thus improving the fit for poorly represented combinations, while 206 preventing overfitting (McElreath, 2020). 207 To account for variance in BAI differing among tree species, climatic regions and depending 208 on tree size, we explicitly modelled the residual variance parameter as the following log-209 linear function (Umlauf and Kneib, 2018): 210 𝑙𝑜𝑔(𝜎!#$) = 𝜀( + 𝜑&𝐵𝐴'(&'!#$ + log(𝛿𝜎#$) (4) 211 where σisc is the residual variance of individual i belonging to species s from country c, e0 is 212 the intercept, j1 captures how variance changes with tree size, and dslogsc represents the 213 baseline variance for species s from country c. 214 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [10] Approach 2: before–after assessment 215 To complement the control–impact analysis, we also tested whether trees grew slower after 216 they were cored. To do this, for each tree that had been cored once, we calculated the 217 differences in BAI (ΔBAI) between the decade after coring (BAI2012-22) and the one before 218 cores were collected (BAI2001-11), where negative values of ΔBAI correspond to slower 219 growth after coring (see Fig. S3 for the same analysis conducted on trees that were cored 220 twice). BAI2001-11 was calculated from the annual growth increments measured directly from 221 the cores, whereas BAI2012-22 was estimated from the repeat census data. We then modelled 222 ΔBAI of tree i belonging to species s from country c in plot p using the same hierarchical 223 Bayesian framework described above but assuming a Student’s t distribution, which best 224 captured the residual error structure of the model (Fig. S4): 225 ∆𝐵𝐴𝐼!,#$,% ~𝑡'𝜇!,#$,%, 𝜎+ (5) 226 where 𝜇!,$#,% is a function of initial tree size (BA2012) and the crown illumination index (CII) 227 recorded in 2012 (treated as a continuous variable): 228 𝜇!,#$,% = 𝛼#$ + 𝛽&#$𝐵𝐴'(&'! + 𝛽'#$𝐶.𝐼! + 𝛾% (6) 229 where αcs and β1-2cs are country- and species-specific model parameters representing the 230 intercept (αcs) and ΔBAI response (i.e., slope) to tree size (β1cs) and crown illumination (β2cs). 231 Initial tree size and crown illumination were included to account for the potential variation in 232 growth regimes. The intercepts were allowed to vary among plots (γp) to capture additional 233 variation in ΔBAI among stands. Country–species–specific parameters (𝛼$#, 𝛽&#$, 𝛽'#$) were 234 modelled using the same multivariate normal hierarchical structure as described in Equation 3 235 for the control-impact assessment model. 236 This analysis makes several assumptions about the comparability of growth trends before and 237 after coring. For example, it overlooks the fact that the climate and the occurrence/severity of 238 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [11] extreme events will have differed between the two time periods, thus resulting in differences 239 in growth that have nothing to do with coring. It also ignores the fact that some trees may 240 have experienced growth declines unrelated to coring due to pest and pathogen outbreaks 241 (e.g., spruce bark beetle, ash dieback and the Oriental chestnut gall wasp). Conversely, some 242 trees may have instead experienced increased growth due to competitive release following the 243 mortality of neighbours. Finally, the analysis also assumes that estimates of BAI before and 244 after coring are directly comparable, even though they were obtained using different 245 methodologies (ring width vs repeat DBH measurements). Despite these sources of 246 uncertainty, this before–after comparison provides a complementary assessment of the 247 potential impacts of coring on growth. 248 Impacts of tree coring on mortality 249 To assess if coring was associated with an increased probability of mortality, we compared 250 the survival rates of trees that were cored in 2012 to those that were not in the decade after 251 coring took place using the re-survey data from 2022. This is analogous to the control–impact 252 assessment on growth described above and made use of the same data and model structure. 253 Specifically, the probability of mortality (P[M=1]) was expressed as a logistic function of initial 254 tree size (BA2012), crown illumination index (CII; continuous variable), coring treatment (C; 255 factor with three levels: not cored, cored once, cored twice), and the interaction term between 256 coring and initial tree size (C × BA2012). 257 P[M=1] was modelled as a Bernoulli distribution with a logit link and allowed to vary among 258 species s, countries c and plots p following a hierarchical model structure: 259 𝑃[01&]~𝐵𝑒𝑟𝑛𝑜𝑢𝑙𝑙𝑖'𝜋!,$#,%+ (6) 260 where 𝜋!,$#,% is a function of initial tree size (BA2012) and the crown illumination index (CII) 261 recorded in 2012 (treated as a continuous variable): 262 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [12] 𝑙𝑜𝑔𝑖𝑡'𝜋!,$#,%+= 𝛼#$ + 𝛽&$# 𝐵𝐴'(&'! + 𝛽'$# 𝐶𝐼𝐼! + 𝛽)$# 𝐶! + 𝛽*$# (𝐶 × 𝐵𝐴'(&')! + 𝛾% (7) 263 Country–species–specific parameters (𝛼$#, 𝛽&3*#$) were modelled using the same 264 multivariate normal hierarchical structure as described in Equation 3 for the control-impact 265 assessment model. 266 Model fitting 267 All models were fit using the brms package in R (Bürkner, 2017). For each model, we used 268 four chains and 3000 iterations per chain, with 1500 discarded as warmup. Chains of all 269 models mixed well with R̂ = 1.00–1.01. The response variable and all model predictors were 270 scaled and centred prior to model fitting. We used uninformative priors, and model parameter 271 posteriors were summarised through their median and 95% highest-density continuous 272 interval (HDCI) using the tidyverse and tidybayes packages (Wickham et al., 2019; Kay, 273 2024). Predictors were considered to have a clear effect on the response variable when the 274 HDCI of their model coefficients did not encompass zero. 275 Model predictions plotted in Fig. 2 and 4 were created using the posterior_linpred() function 276 in the brms R package (Bürkner, 2017). This method describes the uncertainty at the group 277 level and does not include individual level variation (σ). 278

279 Impacts of tree coring on growth 280 Approach 1: control–impact assessment 281 Variation in BAI among trees was strongly and positively associated with both initial tree 282 size (Fig. 2a) and crown illumination index (Fig. S5a), with larger, canopy-dominant trees 283 exhibiting noticeably faster growth rates. On average, a tree with a DBH of 60 cm had a 284 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [13] decadal BAI that was 289.2 cm2 decade-1 greater (355.9 % increase) than that of a tree with a 285 DBH of 20 cm (assuming CII = 3 in both cases). 286 When we compared the BAI of trees that had been cored with those that were not, we found 287 only minimal differences in growth rate and no evidence to suggest that coring detrimentally 288 affected BAI in the decade after coring (Fig. 2a, Fig. S6a). The predicted growth for a tree 289 with DBH of 20 cm that was cored once was 7.5 cm2 decade-1 greater (2.0 % increase) than 290 that of a tree of the same size that was not cored (assuming CII = 3). This positive growth 291 response to coring was slightly more pronounced for trees that were cored twice (BAI 292 increase = 23.8 cm2 decade-1 or 6.2 % relative to non-cored trees, in the above scenario). 293 We also found no evidence to suggest that coring adversely impacted smaller trees more than 294 others. Instead, the positive growth response to coring was most pronounced in smaller trees 295 (a tree with DBH = 10 cm and CII = 3 that was cored once had a BAI that was 6.0 cm2 296 decade-1 greater than that of a tree that was not cored, equivalent to a 4.6 % increase). 297 Moreover, the effects of coring on growth were remarkably consistent across tree species and 298 forest types (Fig. 3a), including when we compared broadleaves and conifers. 299 Approach 2: before–after assessment 300 When we compared the BAI of cored trees in the decade before and after coring took place in 301 2012, we again found no indication of an adverse growth response to coring (Fig. 4). Tree 302 growth rates in the decade before and after coring were strongly positively correlated 303 (Pearson’s correlation coefficient = 0.70; Fig. 4a). Consistent with our comparison of cored 304 and non-cored trees, we observed that trees exhibited slightly elevated BAI in the decade 305 after coring (ΔBAI = 2.3 cm2 decade-1 for a tree with DBH = 20 cm). Once again, these 306

were largely consistent across species and forest types (Fig. 5). We did however 307 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [14] observe that in certain sites, ΔBAI was systematically greater (Poland) or lower (Spain) in the 308 decade following coring. 309 Impacts of tree coring on mortality 310 In the decade between 2012 and 2022, 16.2% of trees across the FunDivEUROPE plots died 311 (Table 1), with the highest mortality rates occurring in Germany and Italy (24.8% and 24.6% 312 decade-1 across species, respectively), with Picea abies in Poland and Germany having the 313 highest species-level mortality rate (68.2% and 66.9% respectively). Similarly to growth, 314 mortality rates were associated with both initial tree size (Fig. 2b) and crown illumination 315 index (Fig. S3b), with larger, canopy-dominant trees exhibiting the lowest mortality rates. On 316 average, a tree with a DBH of 60 cm had a probability of dying of 10.5% decade-1 compared 317 to 4.9% decade-1 for a tree with a DBH of 10 cm (assuming CII = 3 in both cases). 318 We observed no evidence that coring resulted in elevated mortality risk in the decade after 319 coring (Fig. 2b, Fig. S6b). A tree with a DBH of 20 cm had an expected probability of 320 mortality of 8.3% decade-1 if it was cored once, 7.3% decade-1 if it was cored twice, and 7.0% 321 decade-1 if it had not been cored, with confidence intervals around these estimates 322 overlapping broadly (95% range = 4.1–13.6 % decade-1). This lack of a coring effect on tree 323 mortality risk was highly consistent across forest types and species (Fig. 3b), with no 324 evidence to suggest that certain species or climatic region were associated with more 325 pronounced responses to coring (irrespective of whether trees had been cored once or twice). 326

327 Using plot census and tree ring data from the FunDivEUROPE network, we quantified the 328 effect of increment coring on tree growth and mortality across Europe’s dominant tree 329 species and major forest types, spanning Mediterranean, temperate and boreal forests. This 330 study provides the most comprehensive assessment to date of coring’s impact on tree health 331 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [15] across a broad range of forest types and species, finding no negative impact on tree growth 332 and mortality. Our data enabled us to robustly test whether responses to coring vary with tree 333 size and evaluate whether standard dendrochronological practices, such as extracting multiple 334 cores per tree, are associated with long-term impacts on tree health. 335 No evidence of adverse effects of tree coring across forest types 336 Our results indicate that coring has no negative effect on tree growth or mortality 10 years 337 after coring. These findings are consistent with previous studies (Mantgem and Stephenson, 338 2004; Wunder et al., 2011; Helcoski et al., 2018; Fabiánová and Šilhán, 2021; Portier et al., 339 2023) but significantly build upon this evidence by testing across 6 distinct temperate forest 340 types and 16 common European tree species. 341 A subset of the trees within the network was cored twice for isotope analysis, a common 342 dendrochronological practice that is also used to improve cross dating accuracy (Kirdyanov et 343 al., 2018). Although coring twice has previously been discouraged as a precautionary 344 measure (Tsen, Sitzia and Webber, 2016; Portier et al., 2023), our results indicate that, if 345 required, extracting an additional core does not compromise tree health. 346 While we observed no negative impact of coring on tree growth, there was a positive growth 347 response, with a stronger effect in trees cored twice (Fig. S6). For example, a tree with a 348 diameter of 20 cm that was cored once exhibited, on average, 2.0 % more growth than a tree 349 not cored, increasing to 6.2 % when cored twice. These values exceed the reported 350 measurement error associated with diameter tape measurements (Luoma et al., 2017), 351 suggesting that the response cannot be attributed to measurement error. This raises the 352 question of whether the patterns observed in this study reflects a biological response to 353 coring. 354 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [16] Localised positive growth responses have been reported within P . abies as a result of vertical 355 scarring (Fabiánová and Šilhán, 2021), and coring has also been shown to induce pronounced 356 wound responses (Portier et al., 2023). The increased growth observed here may therefore 357

from a wound response to the coring, with a second wound amplifying this effect. 358 Alternatively, the response may be driven by hormesis, whereby a low dose of stress elicits a 359 stimulatory response (Agathokleous, Kitao and Calabrese, 2019). Although hormesis is 360 reported to occur widely in plants induced physical stressors (Agathokleous, Kitao and 361 Calabrese, 2019, 2020; Salinitro et al., 2021), evidence from mature trees is currently 362 lacking, and so the potential contribution of any hermetic process to explain the relationships 363 in our data is uncertain. 364 Irrespective of the underlying mechanism, any growth response to coring is likely to be 365 localised around the coring site. While our results demonstrate coring does not negatively 366 affect tree growth, localised responses to coring risks inflating diameter measurements, 367 potentially leading to overestimates of growth and biomass in permanent monitoring plots. 368 We therefore recommend that increment cores be extracted at least 30 cm above or below the 369 point of diameter measurement as a precautionary measure. 370 With respect to mortality, our model indicates a small reduction in survival among cored trees 371 (Fig. 2b); however, this effect is associated with large uncertainty and is not supported by 372 complementary analyses (Table 1, Fig. S6b). Although larger trees that were cored twice 373 appeared to show reduced survival relative to uncored or singly cored individuals (Fig. 2b), 374 these trees are canopy dominant individuals and exhibit low mortality rates. Consequently, 375 the trend of mortality for trees cored twice should be interpreted with caution. Additionally, 376 the mortality signal directly attributed to coring may be masked in certain locations due to 377 elevated mortality from extreme events over the previous decade. E.g. bark beetle outbreaks 378 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [17] in Picea abies and ash dieback in Fraxinus excelsior within Germany (Lausch, Heurich and 379 Fahse, 2013; Seidl et al., 2014; Fuchs et al., 2024). 380 Smaller trees are not more susceptible to coring 381 We found no evidence that small trees were any more vulnerable to coring than large trees, 382 regardless of whether they were cored once or twice. Although we hypothesised that smaller 383 trees would be more vulnerable to coring, coring wounds are relatively small compared to the 384 minimum size of the stem in this study (7.5 cm diameter), and trees have a high resilience to 385 the loss of xylem (Dietrich et al., 2018). While the immediate hydraulic impacts of coring 386 are therefore likely negligible, we cannot determine the potential longer-term consequences 387 for smaller trees as internal scarring, from coring (Tsen, Sitzia and Webber, 2016), may 388 predispose stems to internal decay over extended periods. Coring can lead to internal scarring 389 (Tsen, Sitzia and Webber, 2016), which may predispose stems to decay over extended 390 periods. This is particularly relevant given that the ecological consequences of stem decay are 391 often more severe for smaller individuals (Datta et al., 2025). Therefore, minimum diameter 392 thresholds for coring should be set along with longer-term monitoring to determine whether 393 size-dependent effects occur after multiple decades. 394 Different tree species and forest types are equally resilient to coring 395 Responses to coring did not differ among tree species or forest types within the 396 FunDivEUROPE network. A common concern is that coring wounds may act as a vector for 397 fungal diseases or other pathogens (Tsen, Sitzia and Webber, 2016), and if this were a 398 significant mechanism, stronger negative effects would be expected in regions with high 399 pathogen load. However, despite elevated mortality associated with bark beetle outbreaks in 400 Picea abies and ash dieback in Fraxinus excelsior within Germany (Lausch, Heurich and 401 Fahse, 2013; Seidl et al., 2014; Fuchs et al., 2024), we found no evidence that single coring 402 affected these species (Fig. 3a/b). This indicates that wounds may not be a significant vector 403 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [18] for disease transmission. However, post coring assessments of wounds would need to be 404 undertaken to confirm this. 405 Challenges of quantifying coring response post-hoc 406 Studies explicitly designed to test the effects of coring on tree health are rare; like most, ours 407 was performed retrospectively. Post-hoc approaches make testing for the impact of coring a 408 non-trivial task as selection bias and unmeasured confounding factors cannot be fully 409 accounted for. These challenges are particularly evident in the before-after growth model. 410 While there was no effect of coring on growth, species-specific responses varied 411 significantly, likely reflecting contrasting environmental conditions between the two decades 412 and ecophysiological differences between species. For instance, P. sylvestris growing in 413 Poland grew faster in the second decade, whereas it displayed consistent growth across both 414 decades in Spain and Finland. 415 There is also evidence of selection bias within this dataset towards coring larger trees with 416 higher crown illumination values, especially among trees cored twice (Fig. 1). While size and 417 crown illumination were accounted for within our models, other factors that correspond with 418 tree vigour may not have been captured. As a result, the positive growth response associated 419 with coring may partially reflect the preferential sampling of healthier individuals rather than 420 a biological effect of coring itself. 421 Recommendations for future work in this area 422 Our results provide strong empirical support for the use of increment coring in European 423 forests, demonstrating the resilience of 16 widespread European tree species across six 424 distinct forest types. Nevertheless, several key research gaps remain. 425 More extensive post-coring assessments of wounds will also be necessary to quantify the 426 extent of scarring and the susceptibly to decay. Although decay following coring has been 427 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [19] investigated in one long-term study (Wunder et al., 2013), this work was limited to a single 428 conifer species. More assessments across both broadleaved and other conifer taxa are needed 429 to test how the response to coring varies among species; do wound responses vary in extent 430 and size and are certain species more resistant to internal scarring? 431 While dendrochronological studies have largely focused on temperate forests, recent 432 methodological advances have been made in tropical dendrochronology; through classical 433 ring boundary analyses (Zuidema et al., 2025) and the advent of radiocarbon dating 434 (Pacheco-Solana et al., 2023). Given the warm and humid conditions in these regions, which 435 may accelerate wood decay, studies in these systems are essential to evaluate whether coring 436 poses greater risks than those in temperate regions. 437 Additionally, 5 mm increment borers are standard to minimise created wounds, but larger (10 438 mm) borers are used for specific analyses (Moore et al., 2009). Therefore, experiments 439 comparing the wound differences between 5 mm and 10 mm borers should be undertaken to 440 provide evidence for best practice guidelines. 441 Finally, to establish causality and minimise selection bias, future studies would benefit from 442 the use of Before-After-Control-Impact designs, often considered the ‘gold standard’ of 443 experimental design as long as sample sizes are sufficiently large (Christie et al., 2019). 444 Furthermore, such studies would allow quantification of how large the affected area is around 445 the coring wound and the potential impact coring has on subsequent diameter measurements. 446 These results would assist with the development of best practice procedures for 447 dendrochronological sampling in permanent forest plots. Given that this study only addresses 448 a relatively limited portion of a tree's lifespan, longer-term, multi-decadal monitoring is now 449 required to assess whether delayed effects of coring – particularly those associated with stem 450 decay – translated into elevated mortality. However, it is worth noting that the longer the time 451 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [20] post coring, the greater the likelihood that tree mortality could be linked to a different driver 452 other than the coring wound. 453

454 Tree coring will always be an integral method in forest ecology to investigate stand dynamics 455 and reconstruct past climates. This study provides support to future dendrochronological 456 studies where concerns to tree health may have been a concern. We find there to be no 457 negative impact on growth or mortality across 16 common European tree species. But 458 highlight a positive growth response in cored trees with potential implications for biasing 459 repeat diameter measurements. As a result, we recommend increment cores not to be 460 collected at the point of diameter measurement. 461

462 The FunDivEUROPE project was funded through the European Union Seventh Framework 463 Programme (grant: 265171). This work was funded by a UKRI Frontier Research award 464 (grant: EP/Y003810/1) awarded to TJ, which also supported RB and DN. 465 Author contribution statement 466 OB coordinated the establishment and re-census of the FunDivEUROPE plot network. TJ 467 generated the tree ring data, with the assistance of OB. RB and TJ designed the study with 468 input from TSO. RB analysed the data, with assistance from DN and FJF. RB wrote the first 469 draft of the manuscript with the assistance of TJ and TSO. All authors contributed 470 substantially to revisions. 471 Data and code availability statement 472 All data and R code underpinning the results of this study will be publicly archived on Zenodo 473 following the review of this paper. 474 475 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [21]

476 Agathokleous, E., Kitao, M. and Calabrese, E.J. (2019) ‘Hormesis: A Compelling Platform for 477 Sophisticated Plant Science’, Trends in Plant Science , 24(4), pp. 318 –327. Available at: 478 https://doi.org/10.1016/j.tplants.2019.01.004. 479 Agathokleous, E., Kitao, M. and Calabrese, E.J. (2020) ‘Hormesis: Highly Generalizable and 480 Beyond Laboratory’, Trends in Plant Science , 25(11), pp. 1076 –1086. Available at: 481 https://doi.org/10.1016/j.tplants.2020.05.006. 482 Anderegg, W.R.L. et al. (2015) ‘Pervasive drought legacies in forest ecosystems and their 483 implications for carbon cycle models’, Science, 349(6247), pp. 528 –532. Available at: 484 https://doi.org/10.1126/science.aab1833. 485 Anderson-Teixeira, K.J. et al. (2022) ‘Joint effects of climate, tree size, and year on annual tree 486 growth derived from tree -ring records of ten globally distributed forests’, Global Change 487 Biology, 28(1), pp. 245–266. Available at: https://doi.org/10.1111/gcb.15934. 488 Babst, F. et al. (2014) ‘Toward consistent measurements of carbon accumulation: A multi-site 489 assessment of biomass and basal area increment across Europe’, Dendrochronologia, 32(2), 490 pp. 153–161. Available at: https://doi.org/10.1016/j.dendro.2014.01.002. 491 Baeten, L. et al. (2013) ‘A novel comparative research platform designed to determine the 492 functional significance of tree species diversity in European forests’, Perspectives in Plant 493 Ecology, Evolution and Systematics , 15(5), pp. 281 –291. Available at: 494 https://doi.org/10.1016/j.ppees.2013.07.002. 495 Boura, A., Péchon, T.L. and Gigord, L.D.B. (2014) ‘Research safeguards protected areas: 496 response to Florens’, Trends in Ecology & Evolution , 29(3), pp. 133 –134. Available at: 497 https://doi.org/10.1016/j.tree.2013.12.007. 498 Bürkner, P.-C. (2017) ‘brms: An R Package for Bayesian Multilevel Models Using Stan’, 499 Journal of Statistical Software , 80, pp. 1 –28. Available at: 500 https://doi.org/10.18637/jss.v080.i01. 501 Castagneri, D. et al. (2022) ‘Meta -analysis Reveals Different Competition Effects on Tree 502 Growth Resistance and Resilience to Drought’, Ecosystems, 25(1), pp. 30 –43. Available at: 503 https://doi.org/10.1007/s10021-021-00638-4. 504 Christie, A.P. et al. (2019) ‘Simple study designs in ecology produce inaccurate estimates of 505 biodiversity responses’, Journal of Applied Ecology , 56(12), pp. 2742 –2754. Available at: 506 https://doi.org/10.1111/1365-2664.13499. 507 Clark, D.A. and Clark, D.B. (1992) ‘Life History Diversity of Canopy and Emergent Trees in 508 a Neotropical Rain Forest’, Ecological Monographs , 62(3), pp. 315 –344. Available at: 509 https://doi.org/10.2307/2937114. 510 Datta, D. et al. (2025) ‘Cavities and the Demographic Performance of Tropical Rainforest 511 Trees’, Ecology Letters, 28(3), p. e70091. Available at: https://doi.org/10.1111/ele.70091. 512 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [22] Dietrich, L. et al. (2018) ‘Losing half the conductive area hardly impacts the water status of 513 mature trees’, Scientific Reports, 8(1), p. 15006. Available at: https://doi.org/10.1038/s41598-514 018-33465-0. 515 Dinerstein, E. et al. (2017) ‘An Ecoregion-Based Approach to Protecting Half the Terrestrial 516 Realm’, BioScience, 67(6), pp. 534–545. Available at: https://doi.org/10.1093/biosci/bix014. 517 Emile-Geay, J. et al. (2017) ‘A global multiproxy database for temperature reconstructions of 518 the Common Era’, Scientific Data , 4(1), p. 170088. Available at: 519 https://doi.org/10.1038/sdata.2017.88. 520 Esper, J. et al. (2016) ‘Ranking of tree -ring based temperature reconstructions of the past 521 millennium’, Quaternary Science Reviews , 145, pp. 134 –151. Available at: 522 https://doi.org/10.1016/j.quascirev.2016.05.009. 523 Fabiánová, A. and Šilhán, K. (2021) ‘The Growth Responses of Picea abies (L.) Karst. to 524 Increment Borer Wounding’, Tree-Ring Research , 77(2). Available at: 525 https://doi.org/10.3959/TRR2020-13. 526 Florens, F.B.V . (2013) ‘Research safeguards protected areas: the important role of 527 governments’, Trends in Ecology & Evolution , 28(9), pp. 504 –505. Available at: 528 https://doi.org/10.1016/j.tree.2013.06.011. 529 Florens, F.B.V . (2014) ‘Research no matter the risks? A reply to Boura et al’, Trends in Ecology 530 & Evolution, 29(3), pp. 134–135. Available at: https://doi.org/10.1016/j.tree.2014.01.006. 531 Fonti, P. et al. (2009) ‘Studying global change through investigation of the plastic responses of 532 xylem anatomy in tree rings’, New Phytologist , 185(1), pp. 42 –53. Available at: 533 https://doi.org/10.1111/j.1469-8137.2009.03030.x. 534 Fuchs, S. et al. (2024) ‘Ash dieback assessments on intensive monitoring plots in Germany: 535 influence of stand, site and time on disease progression’, Journal of Plant Diseases and 536 Protection, 131(5), pp. 1355–1372. Available at: https://doi.org/10.1007/s41348-024-00889-y. 537 Grossiord, C. et al. (2014) ‘Tree diversity does not always improve resistance of forest 538 ecosystems to drought’, Proceedings of the National Academy of Sciences, 111(41), pp. 14812–539 14815. Available at: https://doi.org/10.1073/pnas.1411970111. 540 Helcoski, R. et al. (2018) ‘No Significant increase in tree mortality following coring in a 541 temperate hardwood forest’. 542 Holl, K. and Road, C. (2018) ‘A review of the theory and practice of tree coring on live ancient 543 and veteran trees’, Scottish Natural Heritage Research Report [Preprint], (789). 544 Jucker, T. et al. (2014) ‘Competition for light and water play contrasting roles in driving 545 diversity–productivity relationships in Iberian forests’, Journal of Ecology, 102(5), pp. 1202–546 1213. Available at: https://doi.org/10.1111/1365-2745.12276. 547 Jucker, T. et al. (2017) ‘Detecting the fingerprint of drought across Europe’s forests: do carbon 548 isotope ratios and stem growth rates tell similar stories?’, Forest Ecosystems , 4(1), p. 24. 549 Available at: https://doi.org/10.1186/s40663-017-0111-1. 550 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [23] Kannenberg, S.A. et al. (2019) ‘Linking drought legacy effects across scales: From leaves to 551 tree rings to ecosystems’, Global Change Biology , 25(9), pp. 2978 –2992. Available at: 552 https://doi.org/10.1111/gcb.14710. 553 Kay, M. (2024) ‘tidybayes: Tidy Data and Geoms for (Bayesian) Models’. Available at: 554 http://mjskay.github.io/tidybayes/. 555 Kirdyanov, A.V . et al. (2018) ‘Notes towards an optimal sampling strategy in 556 dendroclimatology’, Dendrochronologia, 52, pp. 162 –166. Available at: 557 https://doi.org/10.1016/j.dendro.2018.10.002. 558 Lausch, A., Heurich, M. and Fahse, L. (2013) ‘Spatio -temporal infestation patterns of Ips 559 typographus (L.) in the Bavarian Forest National Park, Germany’, Ecological Indicators, 31, 560 pp. 73–81. Available at: https://doi.org/10.1016/j.ecolind.2012.07.026. 561 Levesque, M. et al. (2019) ‘Tree -ring isotopes capture interannual vegetation productivity 562 dynamics at the biome scale’, Nature Communications , 10(1), p. 742. Available at: 563 https://doi.org/10.1038/s41467-019-08634-y. 564 Li, H., Speer, J.H. and Thapa, I. (2024) ‘Reconstructing and Mapping Annual Net Primary 565 Productivity (NPP) Since 1940 Using Tree Rings in Southern Indiana, U.S.’, Journal of 566 Geophysical Research: Biogeosciences , 129(8), p. e2023JG007929. Available at: 567 https://doi.org/10.1029/2023JG007929. 568 Luoma, V . et al. (2017) ‘Assessing Precision in Conventional Field Measurements of 569 Individual Tree Attributes’, Forests, 8(2), p. 38. Available at: https://doi.org/10.3390/f8020038. 570 Mantgem, P.J.V . and Stephenson, N.L. (2004) ‘Does coring contribute to tree mortality?’, 571 Canadian Journal of Forest Research , 34(11), pp. 2394 –2398. Available at: 572 https://doi.org/10.1139/x04-120. 573 Mašek, J. et al. (2024) ‘Shifting climatic responses of tree rings and NDVI along environmental 574 gradients’, Science of The Total Environment , 908, p. 168275. Available at: 575 https://doi.org/10.1016/j.scitotenv.2023.168275. 576 Mathes, T. et al. (2024) ‘The effect of forest structure on drought stress in beech forests (Fagus 577 sylvatica L.)’, Forest Ecology and Management , 554, p. 121667. Available at: 578 https://doi.org/10.1016/j.foreco.2023.121667. 579 McElreath, R. (2020) Statistical Rethinking: A Bayesian Course with Examples in R and STAN. 580 2nd edn. New York: Chapman and Hall/CRC. Available at: 581 https://doi.org/10.1201/9780429029608. 582 Moore, J.R. et al. (2009) ‘The effects of site and stand factors on the tree and wood quality of 583 Sitka spruce growing in the United Kingdom’, Silva Fennica , 43(3). Available at: 584 https://www.silvafennica.fi/article/195 (Accessed: 16 January 2026). 585 Morin-Bernard, A. et al. (2024) ‘Integration of tree -ring data, Landsat time series, and ALS -586 derived topographic variables to quantify growth declines in black spruce’, Forest Ecology and 587 Management, 557, p. 121765. Available at: https://doi.org/10.1016/j.foreco.2024.121765. 588 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [24] Pacheco-Solana, A. et al. (2023) ‘Radiocarbon and wood anatomy as complementary tools for 589 generating tree -ring records in Bolivia’, Frontiers in Plant Science , 14. Available at: 590 https://doi.org/10.3389/fpls.2023.1135480. 591 Portier, J. et al. (2023) ‘No evidence that coring affects tree growth or mortality in three 592 common European temperate forest tree species’, European Journal of Forest Research 593 [Preprint]. Available at: https://doi.org/10.1007/s10342-023-01612-6. 594 Réjou-Méchain, M. et al. (2017) : ‘an r package for estimating above -ground biomass and its 595 uncertainty in tropical forests’, Methods in Ecology and Evolution , 8(9), pp. 1163 –1167. 596 Available at: https://doi.org/10.1111/2041-210X.12753. 597 Rodriguez-Caton, M. et al. (2024) ‘A 300 -year tree -ring δ18O-based precipitation 598 reconstruction for the South American Altiplano highlights decadal hydroclimate 599 teleconnections’, Communications Earth & Environment , 5(1), pp. 1 –13. Available at: 600 https://doi.org/10.1038/s43247-024-01385-9. 601 Salinitro, M. et al. (2021) ‘Induction of hormesis in plants by urban trace metal pollution’, 602 Scientific Reports, 11(1), p. 20329. Available at: https://doi.org/10.1038/s41598-021-99657-3. 603 Seidl, R. et al. (2014) ‘Increasing forest disturbances in Europe and their impact on carbon 604 storage’, Nature Climate Change , 4(9), pp. 806 –810. Available at: 605 https://doi.org/10.1038/nclimate2318. 606 Sohn, J.A., Saha, S. and Bauhus, J. (2016) ‘Potential of forest thinning to mitigate drought 607 stress: A meta-analysis’, Forest Ecology and Management , 380, pp. 261 –273. Available at: 608 https://doi.org/10.1016/j.foreco.2016.07.046. 609 Talbot, J. et al. (2014) ‘Methods to estimate aboveground wood productivity from long -term 610 forest inventory plots’, Forest Ecology and Management , 320, pp. 30 –38. Available at: 611 https://doi.org/10.1016/j.foreco.2014.02.021. 612 Tsen, E.W.J., Sitzia, T. and Webber, B.L. (2016) ‘To core, or not to core: the impact of coring 613 on tree health and a best -practice framework for collecting dendrochronological information 614 from living trees’, Biological Reviews , 91(4), pp. 899 –924. Available at: 615 https://doi.org/10.1111/brv.12200. 616 Umlauf, N. and Kneib, T. (2018) ‘A primer on Bayesian distributional regression’, Statistical 617 Modelling, 18(3–4), pp. 219–247. Available at: https://doi.org/10.1177/1471082X18759140. 618 Wei, J. et al. (2024) ‘Drought alters aboveground biomass production efficiency: Insights from 619 two European beech forests’, Science of The Total Environment, 919, p. 170726. Available at: 620 https://doi.org/10.1016/j.scitotenv.2024.170726. 621 Wickham, H. et al. (2019) ‘Welcome to the Tidyverse’, Journal of Open Source Software , 622 4(43), p. 1686. Available at: https://doi.org/10.21105/joss.01686. 623 Wunder, J. et al. (2011) ‘Long-term effects of increment coring on Norway spruce mortality’, 624 Canadian Journal of Forest Research , 41(12), pp. 2326 –2336. Available at: 625 https://doi.org/10.1139/x11-150. 626 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [25] Wunder, J. et al. (2013) ‘Does increment coring enhance tree decay? New insights from 627 tomography assessments’, Canadian Journal of Forest Research , 43(8), pp. 711 –718. 628 Available at: https://doi.org/10.1139/cjfr-2012-0450. 629 Zuidema, P.A. et al. (2025) ‘Pantropical tree rings show small effects of drought on stem 630 growth’, Science, 389(6759), pp. 532 –538. Available at: 631 https://doi.org/10.1126/science.adq6607. 632 633 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [26] Tables 634 Table 1: Overview of the size, growth and mortality rates of trees included in this study. 635 Country Species Cored Not Cored Mean basal area growth (cm2 decade-1) Mean DBH in 2012 (cm) Total number Dead (%) Total number Dead (%) Finland Betula pendula 121 0 175 0 110.1 19.3 Picea abies 203 5.4 654 4.6 102.6 17.8 Pinus sylvestris 176 15.3 473 18.6 101.6 19.3 Poland Betula pendula 77 35.1 112 22.3 194.1 36.0 Carpinus betulus 132 6.1 717 7.1 81.7 18.8 Picea abies 119 79.0 249 63.1 115.4 28.0 Pinus sylvestris 101 9.9 130 6.2 284.4 41.3 Quercus robur 105 5.7 101 5.0 311.9 35.3 Germany Acer psudeoplatanus 83 9.6 124 7.3 92.9 24.1 Fagus sylvatica 238 8.4 329 10.5 120.9 22.9 Fraxinus excelsior 144 25.0 239 36.4 155.5 26.3 Picea abies 67 67.2 208 66.8 169.8 24.1 Quercus petraea 73 9.6 24 8.3 268.4 50.0 Romania Abies alba 95 9.5 206 10.2 211.5 34.0 Acer psuedoplatanus 81 12.3 152 12.5 114.0 32.0 Fagus sylvatica 120 13.3 336 11.6 112.5 28.0 Picea abies 87 12.6 168 16.1 189.3 36.4 Italy Castenia sativa 144 52.8 410 52.0 41.8 18.0 Ostrya carpinifolia 132 16.7 469 23.0 36.2 14.0 Quercus cerris 161 9.3 244 23.0 97.5 22.9 Quercus ilex 166 1.8 262 6.5 53.7 19.3 Quercus petraea 142 11.3 178 23 96.5 22.6 Spain Pinus nigra 165 0.6 336 1.5 48.4 22.5 Pinus sylvestris 104 6.7 235 7.2 49.7 24.4 Quercus faginea 204 4.9 916 4.5 20.7 12.6 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [27] Figures 636 637 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [28] Fig. 1: Variation in the distribution of (a) crown illumination index scores and (b) stem 638 diameters of trees that were either not cored (blue, dashed line), cored once (red, continuous 639 line), and cored twice (black, dotted line). Locations of the FunDivEUROPE sites (c), the 640 colour palette represents the six distinct biomes in Europe (Dinerstein et al., 2017). 641 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [29] 642 Fig. 2: Differences in (a) basal area growth and (b) mortality rates of trees of varying sizes 643 that were either not cored (blue, dashed line), cored once (red, continuous line), and cored 644 twice (black, dotted line). Fitted lines are model predictions with 95% highest posterior 645 density intervals. 646 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [30] 647 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [31] Fig. 3: Variation in the effects of tree coring on (a) growth and (b) mortality rates across 648 species and study sites for trees that were cored once (red, continuous line) or twice (black, 649 dotted line), expressed as the difference relative to trees that were not cored. For growth, 650 points represent posterior means of the standardised model coefficients (±95% posterior 651 density intervals), whole for mortality they show the probability of a cored tree dying 652 compared to non-cored tree (where 0.5 represents no difference between the two). 653 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [32] 654 Fig. 4: Tree growth rates before and after coring. Panel (a) shows the relationship between 655 the basal area growth of trees cored once in 2012 in the decade before and after coring 656 (calculated from the tree ring and census data, respectively). Points are coloured by country, 657 and the line corresponds to a 1:1 relationship. Panel (b) shows the estimated difference 658 (±95% confidence intervals) in basal area increment before and after coring (ΔBAI) as a 659 function of tree size, where values of 0 correspond to no difference in growth over time. The 660 density plot at the top illustrates the size distribution of trees included in the analysis. 661 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint [33] 662 Fig. 5: Variation growth rates before and after coring across species and study sites for trees 663 that were cored once. Points are standardised model coefficients with 95% highest posterior 664 density intervals. A negative value would indicate that cored trees grew slower in the decade 665 after coring. 666 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted February 5, 2026. ; https://doi.org/10.64898/2026.02.03.703454doi: bioRxiv preprint