Higher growth synchrony and climate change‐sensitivity in European beech and silver linden than in temperate oaks

Climatic changes affect the growth dynamics of temperate trees, but these effects might differ between co-occurring ring- and diffuse-porous species as well mesic xeric rear-edge populations. We explore whether recent climate warming has increased sensitivity within-stand synchrony in groups differently. The natural beech–oak ecotone Western Romania at dry margin beech occurrence. Three ring-porous oak (Quercus petraea, Q. frainetto cerris), European (Fagus sylvatica) silver linden (Tilia tomentosa). correlated tree-ring records with monthly seasonal data (period 1940–2017). Regional was assessed through analysis inter-series correlation within populations among using mixed models. In all five including two south-east taxa linden, water availability summer most important climatic determinant radial growth. This factor gained importance since onset rapid after 1980, while impact other factors spring decreased. Within-population a measure overall stress increased, or remained stable, 1980 decreased species, matching declining trends increasing (or stable) oaks. patterns provide valuable information on tree species' drought susceptibility efforts to select suitable for change-adapted forestry. vulnerability is higher than that more drought-resistant due its marked summer-drought Klimatische Veränderungen wirken sich auf die Wachstumsdynamik von Bäumen der gemäßigten Breiten aus, doch könnten diese Auswirkungen zwischen gemeinsam vorkommenden und diffusporigen Arten sowie mesischen xerischen Populationen am Verbreitungsrand unterscheiden. Wir untersuchen, ob rezente Klimaerwärmung Klimaempfindlichkeit Baumarten verändert Synchronisierung des Wachstums innerhalb erhöht hat. Das natürliche Buchen–Eichen Ökoton West-Rumänien an Trockengrenze Buchenvorkommens. Drei ringporige Eichenarten cerris) Zerstreutporer Rotbuche Silberlinde untersuchen anhand Jahrringserien Zeitraumes 1940–2017 räumlichen zeitlichen Muster Klimasensitivität Wachstumssynchronizität fünf Laubbaumarten Breiten. Bei allen einschließlich beiden südosteuropäischen Eichen war das Radialwachstum erster Line Wasserverfügbarkeit im Sommer abhängig. Dieser Faktor hat seit dem Beginn raschen Erwärmung nach Bedeutung gewonnen, während Einfluss anderer Klimafaktoren Frühjahr abgenommen Die als ein Maß für klimatischen Stress bei Buche zugenommen oder blieb stabil, sie den hat, mit abnehmenden rezenten Wachstumstrends Linde zunehmenden (oder stabilen) Trends verbunden ist. liefern wertvolle Informationen über Trockenstressempfindlichkeit Baumarten, um geeignete eine Klimawandel angepasste Forstwirtschaft auszuwählen. Klimaanfälligkeit ist höher trockenheitsresistenteren Eichen, da ihr Wachstum deutlich trockenheitsempfindlicher Widespread occurrence pre-senescent leaf fall, crown dieback mortality have hit Central forests during extreme hot droughts 2003, 2015 2018–2020, when heat, high evaporative demand atmosphere long rainless periods harmed especially fast-growing conifers such Norway spruce, also broadleaf trees like sylvatica L.) (Allen et al., 2015; Schuldt 2020; Senf 2020). It almost certain sensitive timber will not maintain their productivity predicted trend until end century beyond, our capacity predict physiological responses forest limited knowledge tolerance limits still poor. annual temperature increase Europe (Lee 2021) likely shift distribution limit many polewards upwards regions cooler temperatures precipitation (Lenoir 2008; Sykes 1996). For example, envelope models warmer drier lowlands lower montane belts from beech-dominated drought-tolerant, thermophilic communities course 21st (Fischer 2019; Kramer 2010). centre range located central southern Germany (Figure S1.1 Appendix S1), where naturally would cover 2/3 area occurring belt mountains (Leuschner & Ellenberg, 2017). Toward south-eastern Europe, hotter summers aridity are shifting toward elevations moister climates (Czúcz 2011; Fang Lechowicz, 2006). elevation usually give way oak-dominated elevations, which linked mixture (Horvat 1975; Mayer, 1984). Various deciduous replace limit, mainly Quercus cerris Liebl., Ten., pubescens L., various aggregate petraea (Matt.) L. S1). They accompanied by assumed Carpinus betulus C. orientalis Tilia cordata T. tomentosa Moench., Acer tataricum At often only persists small refugia deep valleys north-exposed slopes humid local (Coldea Hohnwald Indreica These suited analysing adaptive response comparison drought-tolerant species. Beech known be drought-sensitive it suffers drought- heat-induced (Brinkmann 2016; Leuschner, Zang 2014). visible vitality losses increases (Dulamsuren 2017; Kasper 2022; Serra-Maluquer 2019), may gradually northern upward direction (Peñuelas 2008). Yet, some studies detected no differences marginal populations, highlight potential adaptation dry-marginal (Hacket-Pain Muffler Tegel Dendrochronological evidence demonstrates submediterranean (Friedrichs 2009; Härdtle 2013; Scharnweber 2011) several hornbeam (Fuchs 2021; Köcher 2014; Zimmermann 2015). However, comparative considerable site conditions, neglect shifts over time, particular possible 1980s. As consequence deteriorating growing show increasingly synchronous environmental (Shestakova 2016, 2018; Tejedor Synchrony here stands common temporal variation trait individuals population, different ecosystem, implications ecosystem functioning stability (Schurman Shestakova 2016). To quantify response, we follow approach al. (2020) use so-called rbar statistic, per definition “the inter-annual variability defined group”, characterize (del Río 2021). Growth can spatially depicting regional global similarities variation, temporarily time. Furthermore, investigated taxonomical levels, is, same (Tejedor change research, across been interpreted indicator susceptibility, whereas indication diversity (Anderegg HilleRisLambers, Camarero Pretzsch A number reported warming-related last century, suggesting causal link this study, investigate within-population co-existing (F. sylvatica), hereafter referred beech, (T. tomentosa; linden), three (Q. frainetto) western Romania. region, conditions similar (Hohnwald establishing replicates landscape scale, studied transects sufficient comparability terms thermal hygric exposition, bedrock type, composition management history (Kasper 2021, Öder analyse correlations time interval 1940–2016 split into before pronounced ca. onwards thereafter. compare further contrast subhumid stands, surviving beyond heat beech. By selecting about 2.5 K mean today, study offers insights under Representative Concentration Pathway (RCP) scenarios 2.6–8.5 2021), 1.0–3.7°C S1.2 following questions guided research: (1) How do growth? (2) Did mid-20th period 1980s? (3) Do how did synchrony? (4) (rear-edge) response? conducted foothills south-western Carpathians 1a). sub-continental warm relatively cold winters considered analogue projected current around 2080s. region thus matches demonstrated plotting future ordination space winter cold, warmth season S1, Kölling Zimmermann, Pronounced took place sites (MSP) fluctuated past decades without clear S1.3 balance (precipitation – evapotranspiration) markedly S1.4 S1; 2022). gradients downhill submontane/montane colline were established (Transects A, B C, Figure 1b, 2022), sequentially >500 m a.s.l., humid-subhumid beech–hornbeam–oak 350–600 finally basal <350 predominantly south-west- south-east-facing dominant soils base-poor acidic eutric Cambisols forests. soil covered places up 100 cm thick loess layer All selected had roughly stem densities comparable competition intensities, mature age (>60 years old), 21–33 height closed canopy (Table S1 While subjected occasional wood-cutting low-intensity coppicing 1960, afterwards transferred state-ownership. Previously coppiced allowed grow (Öder lightly thinned (5%–15% stand volume) pole-wood stage quarters till harvest passed (Ashton Kelty, 2018). Irregular salvage sanitary loggings low intensity (removal <5% volume). No major operations occurred 20 slopes. F. contributed least 85% total numbers forests, dominated cerris, frainetto, >85% remaining stems accompanying tomentosa, betulus, campestre ecotone, beech-oak present, each contributing 30% (except transect B) mostly belonging Five focal selected. (mesic beech), dominants submediterranean, (oak), again beech-oak-hornbeam form 1b Table addition, north-facing sampled extra-zonal (<400 a.s.l.; beech) close proximity belt. wood core sampling 2018 2019 resulted 92 153 cores type (beech vs. xeric). cases upper DBH > 40 sample lacked signs abscission, dieback, damage. Tree extracted breast (1.3 m) increment borer (Haglöf) tension wood, perpendicular any tilting direction. Cores rotten segments excluded analysis. beginning season, incomplete ring omitted analysis; hence, series collectively ended 2017. assessing individual intensity, Hegyi index (CI) calculated (Hegyi, 1974) approximated counting rings. cross-dated (TRW, width mm) relying coefficient agreement (“Gleichläufigkeit” GLK; Eckstein Bauch, 1969), cross-dating (CDI; Dobbertin Grissino-Mayer, 2004) Student's t-value (TVBP; Baillie Pilcher, 1973). chronologies GLK 65%, CDI >2.0, TVBP >3.0 obtain detrended dimensionless (RWI), TRW divided derived 30-year moving-average spline frequency cut-off 50%, first-order autoregression removed subsequently (pre-whitening). Master population value (Tukey's bi-weight robust mean) (RWI) S1.5 quality chronologies, expressed signal (EPS) >0.85, coherence neighbouring (Wigley Sample sizes B: 1950–2017) synchrony. Monthly (winter: previous December–current February; spring: March–May; summer: June–August; autumn: September–November) computed high-resolution gridded (30 arcsec, ~1 km2) downloaded CHELSA (CHELSAcruts) database (Karger effect droughts, Standardized Precipitation-Evapotranspiration Index (SPEI) (Vicente-Serrano 2010) 3-month intervals scales. Climate compiled being averaged means. variables (climate sensitivity) correlating RWI temperature, sums, exposure (3-month SPEI value), spanning meteorological year's June September. done observation (1940–2016) separately (1940–1979; earlier period) (1980–2016; later statistical significance p ≤ 0.05 analyses 1000-fold bootstrapping procedure R-package ‘treeclim’ v2.0.5.1 (Zang Biondi, Synchronous estimated (rbar; 0: 1: totally synchronized dynamics) value, average Pearson given chronology running values (rbar) 20-year moving window R package ‘dplR’ v1.7.1 (Bunn, Significance Mann–Kendall test. When computing associated random error structure, must carefully followed (2016, 2018), model (‘DendroSync’ v0.1.3 package; Alday 2018) spatial networks level (and First, allowing variance–covariance (VCOV) structures grouping (homoscedastic heteroscedastic versions) best VCOV structure based Akaike's Information Criterion (AIC) Bayesian (BIC) (Burnham Anderson, 2004). Species synchronicities then models, 0 (no synchronization populations) 1 (total synchronization) (Alday averages fitting lag 5 2016), 5-year above-mentioned criteria. Means (±SD) presented throughout paper means depicted graphs tables tested Wilcoxon Rank test t-test case pairwise comparisons. procedures performed v4.0.3 (R Core Team, Radial additionally A), found hints low-temperature limitation autumn 2; S1.6 data). Interestingly, negative temperature-growth generally weaker driest C 2). High enhances C), sporadically influential (A B), (A), (A)), Previous seemed less Local (SPEI 3-months aggregation limiting (transect A) (transects C) months influenced and, particular, After showed weakening high-temperature 3). became notable exceptions influence declined 3; S1.7 There consistent decline transects, Differences existed B, reduced so S1.7). lowest moistest B. Overall, highest recorded moderately moist 4a). Compared characterized medium both sites. differed 4b). xeric) (xeric) large declines C; Within-stand significantly term, linearly related long-term 5). dropped 1970s, coincided precipitation. early 1980s S1.3), continuously Both significant decrease observed inversely homoscedastic unstructured (mUN) heterogeneous variances co-variances. obtained results pre-whitening showing compared S1.8 Split (before increase), S1) decreases S1.8c non-pre-whitened S1.9a pre-whitened 1975 broad evaluation (mBE), homogenous 6b; 1975, fit, bifurcations Following remains stable 6b) S1.9b Romania, identified Southern numerous (Bose Bosela Fuchs Hacket-Pain key role atmospheric and/or edaphic underpinned rates productive suffered greater reductions 2000, 2003 2012 strong suggests primarily vapour pressure deficit (for stand-level missing), together controlling instead alone 2011). Physiological mechanisms leading include lowered carbon gain partial stomatal closure turgor loss cambial cells caused reduction tree's hydraulic (Bréda 2006; Müller-Haubold Salomón Other could enhanced investment fine root system compensate drought, formation thi