Multiscale Modelling of European Beech Decline: The Role of Long-Term Climate Deviations and Local Environmental Factors

Multiscale Modelling of European Beech Decline: The Role of Long-Term Climate Deviations and Local Environmental Factors | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Multiscale Modelling of European Beech Decline: The Role of Long-Term Climate Deviations and Local Environmental Factors Jean Lemaire, Michel Vennetier, Bernard Prévosto, Maxime Cailleret This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5417359/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 05 May, 2025 Read the published version in European Journal of Forest Research → Version 1 posted 10 You are reading this latest preprint version Abstract Fagus sylvatica L. is a main forest tree species in Europe but has been subjected to massive decline events over the last decades. This phenomenon has been mainly attributed to the increase in drought frequency and intensity, but it is unclear how the local specificities in stand structure, climatic, soil and topographic conditions interact, and if statistical models are able to capture the high spatial and temporal variability in tree decline. To fulfil this objective, we measured 5380 Fagus sylvatica trees from 308 plots distributed in four regions of France with contrasting environmental conditions, and designed models predicting decline at both regional and national scales. These models aimed at assessing the percentage of stems by plot with at least 50% crown biomass loss based on 229 dendrometric, topographic, soil and climatic variables. The climatic factors explained most of the variability in stand decline, especially the long-term deviations from the 30-years mean in maximal temperature and in hydric deficit. Regional models were the most efficient in predicting beech decline in their calibration areas (Q² varied from 0.26 to 0.42) as they better consider the local environmental factors. They were less effective in the other regions, and the national model was an acceptable compromise on a larger scale. These statistical models provide valuable insights for forest managers and could be improved through a more detailed temporal stand monitoring to control the effects of management and decline dynamics. Fagus sylvatica Statistical modeling Climate Soil factors Topography factors Dendrometry. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. Introduction European beech ( Fagus sylvatica L.) is the most abundant native broadleaf tree species in Europe, spanning from southern Italy to southern Norway and from western Spain to eastern Bulgaria. In addition to its large distribution area, it is also one of the most ecologically and economically important hardwood tree in Europe, playing a pivotal role in numerous forest ecosystems (Houston Durrant et al. 2016 ; Leuschner 2020; Leuschner and Ellenberg 2017 ). Like many tree species that are experiencing rising mortality rates in many regions of the globe (e.g., Allen et al. 2010 ; Choat et al. 2012 ; Hartmann et al. 2018 ; Senf et al. 2018 ; van Mantgem et al. 2009 ), European beech has significantly suffered from droughts and heatwaves associated with climate change in the last decades. Decline episodes have been reported in Western Europe following severe droughts in 1976 and 1989-90 (Lies 1980 ; Nageleisen 1994 ; Peterken and Mountford 1996 ), and its productivity has been declining in the southern part of its distribution range since the 2000s (Peñuelas et al. 2008 ; Piovesan et al. 2008 ; Rozas et al. 2015 ; Serra-Maluquer et al. 2019 ). Massive die-off also occurred following the extreme droughts and heatwaves of 2003 and 2018 (Bréda et al. 2006 ; Braun et al. 2021 ; Rukh et al. 2023 ), often in the core of its distribution range in France (Mirabel and Gaertner 2023 ), Switzerland (Braun et al. 2021 ; Frei et al. 2022 ) and Germany (Langer and Bußkamp 2023 ), but also in Eastern Europe such as in Croatia (Ognjenović et al. 2022 ) or Romania (Chira et al. 2006 ). The drought-induced decline of European beech manifests through various symptoms in the canopy and stem. The first symptom in the case of intensive drought is the premature browning and shedding of leaves (Schuldt et al. 2020 ; Arend et al. 2022 ; Frei et al. 2022 ; Neycken et al. 2024 ). Leaf shedding allows the tree to reduce the intensity of xylem embolism by limiting its water demand but at the cost of carbon assimilation (Bréda et al. 2006 ; Wolfe et al. 2016 ; Schuldt et al. 2020 ). This leaf drop may, however, not be sufficient to prevent embolism in case of extreme drought, particularly on shallow soil (Walthert et al. 2021 ; Arend et al. 2022 ). This may lead to subsequent crown-dieback, with the death of twigs and branches, and ultimately to tree mortality (Chakraborty et al. 2017 ; Frei et al. 2022 ; Leuschner 2020; Schuldt et al. 2020 ). This was observed during summer 2018 and following years, in Central and Western Europe, particularly in France, Switzerland, and Germany. In addition to these direct impacts of extreme drought, lag effects can occur through the decrease in leaf size and shoot growth in the following years (Bréda et al. 2006 ; Nageleisen 2005 ). This process is due to reduced carbon reserves and by the formation of fewer leaf bud primordia which would influence next year’s leaf area (Leuschner, 2020). All together, these impacts increase crown leaf deficit (Roloff 1989 ; Woodcock et al. 1995 ; Eichhorn et al. 2020 ), and are often associated with a sharp reduction in tree radial growth (Braun et al. 2021 ; Arend et al. 2022 ; Rukh et al. 2023 ). For instance, the basal area increment of dominant trees in stands affected by decline, decreased by 10 to 50% of the maximum growth rate during the past 20 to 30 years (Leuschner, 2020). This decline, which manifests in various and often progressive symptoms as explained above, results from a multi-causal process (Franklin et al. 1987 ; Waring 1987 ). For beech, it involves numerous factors related to climate, in particular drought, but also late frosts which can damage leaves, especially when occurring in spring during the critical stage of leaf unfolding (Dittmar et al. 2006 ; Dittmar and Elling 2007 ). Decline can also be aggravated by extreme temperatures and intense radiation which occur during heatwaves and can impact photosynthetic activity and induce leaf photobleaching (Braun et al. 2021 ). Besides the climatic conditions, tree health is influenced by the topographic and soil conditions, the tree and stand characteristics, and often the presence of biotic pathogens (Toïgo et al. 2020 ). For instance, beech trees growing on shallow soils with low water-holding capacity have a greater risk of drought-induced dieback or growth reduction (Chakraborty et al. 2017 ; Rukh et al. 2023 ; Sanders et al. 2012 ; Schmied et al. 2023 ). However, beech may also be at risk on soils with shallow and compact clay layers inducing significant hydromorphy (Cros et al. 1981; Schmull and Thomas 2000 ). The competition status and genetic information also play a role in tree vulnerability. Beech trees subjected to strong competition are more susceptible to decline (Petit-Cailleux et al. 2021 ; Frei et al. 2022 ), and those originating from drier habitats show signs of higher drought tolerance and/or faster recovery from drought stress than moister origins, thanks to local adaptation (Csilléry et al. 2014 ; Leuschner, 2020; Pluess et al. 2016 ). Indeed, such trees experienced a delayed onset of drought-induced reductions in stomatal conductance and photosynthesis, and reduced their fine root biomass to a lesser degree compared to trees from moister origins (Tognetti et al. 1995 ; Rose et al. 2009 ). Stand characteristics can also modulate the impacts of water stress (Bréda et al. 2006 ; Diaconu et al. 2017 ). On the one side, stands with high density and Leaf Area Index (LAI) are more prone to decline due to high water demand (Schmied et al. 2023 ). But on the other side, stands with a reduced LAI, e.g., after heavy thinning, can also show symptoms of decline (Mathes et al. 2024 ) due to a less favourable microclimate at crown level (e.g. De Frenne et al. 2021 ). Therefore, favoring a higher tree diversity through a mixture with other species that are less water-consuming, like oaks or pines, can promote the health of beech trees (Sousa-Silva et al. 2018 ; Mathes et al. 2024 ). Finally, weakened beech trees are more susceptible to attacks by secondary biotic agents (e.g. Diplodia corticola A.J.L. Phillips, A. Alves & J. Luque, Neonectria coccinea (Pers.) Rossman & Samuels, Armillaria mellea ( Vahl) P. Kumm., Taphrorychus bicolor (Herbst, 1793)) which can intensify and accelerate the decline (Mirabel and Gaertner 2023 ; Langer and Bußkamp 2023 ), and lead to widespread mortality in subsequent years (Rouault et al. 2006 ; Nageleisen and Reuter J.-C. 2007; Rasztovits et al. 2014 ). Considering that tree decline risk is projected to increase with future climate change (e.g. Allen et al. 2015 ), it is crucial to better quantify the respective impacts and interactions of the factors involved in this process, and to accurately predict this risk at a spatial scale meaningful to forest managers. A wide variety of models can be used for this purpose such as (i) statistical models that estimate mortality risk based on the correlation between climatic parameters and observed mortality patterns (e.g., Taccoen et al. 2019 ), on past radial growth (e.g., Hülsmann et al. 2016 ), or on the actual presence-absence of the species (species distribution models ; e.g., Cheaib et al. 2012 ); (ii) mechanistic models that explicitly simulate the physiological processes leading to mortality (e.g., Petit-Cailleux et al. 2021 ; Ruffault et al. 2022 ). Models simulating beech leaf deficit are scarcer - despite its strong link with mortality risk (Dobbertin and Brang 2001 ; Dobbertin 2005 ; Petit-Cailleux et al. 2021 ). These models aim at explaining the spatio-temporal changes in leaf deficit using either the climatic, atmospheric, biotic, or soil information, but exclude comprehensive analyses of all four factors and their interrelationships (Seidling et al. 2012 ; Popa et al. 2017 ; Toïgo et al. 2020 ; Rohner et al. 2021 ; Ognjenović et al. 2022 ). In addition, they have been developed using broad-scale information (e.g., with data from the European ICP-Forests monitoring network; Ferretti, 2021 ), and therefore can hardly predict the risk of decline at stand level. This is due to the significant influence factors acting at local scale on forest health, such as micro-climate, -topography, -pedology, and the stand characteristics. In consequence, it is likely that decline models established at small spatial scales are more reliable than those computed at large ones, which blur local peculiarities and compensatory effects (e.g., regional vs. national level; see Chauvier et al. 2021 ; Simon et al. 2023 ). This is particularly the case at the species distribution margins where the mitigating effect of the local soil and topography on the water budget can be very high, but neglected in large-scale models (Mellert et al. 2018 ). Similarly, the transferability of regional models across space (i.e., in other regions) is rarely assessed except in species distribution models (e.g. Suárez-Seoane et al. 2014), while there may be a trade-off between model complexity (~ performance) and transferability (Naas et al. 2024 ). To fill these gaps, our main objective was to develop and compare the predictive accuracy of statistical models of beech leaf deficit (i) in four regions of France with contrasting climatic and site conditions, which were impacted by beech decline for at least the last three decades ; and (ii) on national scale by combining the data from these four regions. We aimed at exploring the similarities and differences between the regional models, assessing their ability to predict beech decline outside their calibration range, and analysing to what extent the national model could accurately predict the decline process at a regional level. 2. Materials and Methods 2.1 Study area and sampling design We used 308 plots distributed across four regions of France (Fig. 1 A; Table 1 ) where beech is abundant and shows significant signs of decline: Hauts-de-France and Normandie (HFN), Jura, the Regional Natural Park of Millevaches (MV), and the Regional Natural Park of Haut-Languedoc (HL). In HFN and HL, previous episodes of decline have been recorded following the severe drought of 1989-90 and the 2003 heatwave (Nageleisen 1994 ; Nageleisen and Reuter, J.-C. 2007; Silva 2010 ). In the Jura region, important decline events occurred in the 1950s (Schaeffer 1955 ) and after the 2018 drought and heatwave (Mirabel and Gaertner 2023 ). These regions cover most of the climatic niche of beech in France: oceanic, semi-continental, mountain and Mediterranean (Joly et al. 2010 ), with mean annual temperature and precipitation spanning from 8.8°C to 13.6°C and 576 mm to 1697 mm, respectively (data from 1981 to 2010; Table 1 ). The plots were situated in the core of the climate-space diagram estimated at the European level (Fig. 1 B), and covered a wide range of soil types, from acidic brown soils at MV and HL (mean pH = 4.8 for both regions) to slightly acidic brown soils or calcareous soils with active lime in HFN and Jura (mean pH = 5.7 and 6.1, respectively). Hydromorphic layers were identified at less than one meter depth in 18 plots, only in the HFN region and were particularly marked in three plots (redoximorphic features exceeding 35%). The mean depth of soil estimated with a pickaxe and hand-dug trial was 65 cm with coarse elements representing from 0–50% of soil volume. The soil texture was predominantly loamy except in the Jura. In this region, 56% of the plots had a soil texture comprising at least 30% of clay, among which 26% had a clay percentage over 50% (heavy clay soils). Table 1 Main characteristics of the study plots by region Mean values are represented in bold, with the minimum and maximum within brackets. Regions : HL = Haut Languedoc, HFN = Hauts-de-France and Normandie, Jura = Jura, MV = Millevaches, NAT = national; XStem50 : percentage of stems by plots with a crown biomass loss ≥ 50%; Declining stands : percentage of plots with XStem50 ≥ 20%, pH : pH measured at 20 cm depth; Act_lime : percentage of plots with active lime in soil profile; Heavy_hydro : percentage of plots with heavy hydromorphy (i.e., complete discoloration of the soil matrix and reduction or rust spots > 35% of soil horizon area); Heavy_clay : percentage of plots with more than 50% of clay in one horizon of the soil profile; TM_year and P_year : mean annual temperature and sum of precipitations over the 1981–2010 period; G : total basal area; G_beech : the proportion of beech in the total basal area; Ddom and Hdom : diameter at breast height and total height of the dominant beech trees of the plot, respectively. Region Number of plots Years of sampling XStem50 (%) Declining stands (%) Soil Climate 1981–2010 Stand characteristics pH Act_lime (%) Heavy_hydro (%) Heavy_clay (%) TM_year (°C) P_year (mm) G (m²/ha) G_beech (%) Ddom (cm) Hdom (m) MV 49 2018, 2019 9.8 (0–45) 22.4 4.8 (4.6–5.2) 0 0 0 9.98 (8.9–12.2) 1313 (996–1578) 29.5 (14–44) 70 (28–100) 40 (33–60) 23 (18–27) HL 61 2015 10 (0–90) 19.7 4.8 (4.5–5.5) 0 0 0 11.18 (8.8–13.6) 1429 (1103–1697) 32.4 (5–55) 87 (20–100) 60 (16–75) 25 (11–35) HFN 120 2015, 2016 10.6 (0–85) 20.8 5.7 (4.7-7) 12.5 2.5 0 10.6 (9.7–11.3) 801 (576–1121) 22.5 (10–42) 73 (20–100) 68 (26–103) 31 (18–41) Jura 78 2020 29.7 (0-100) 52.6 6.1 (4.7–7.7) 19.2 0 25.6 10.34 (9.2–11) 1183 (987–1370) 24 (6.5–42) 70 (20–99) 68 (35–100) 30 (21–37) NAT* 308 2015–2020 15.2 (0-100) 21.2 5.5 (4.5–7.7) 9.7 0.98 6.5 10.6 (8.8–13.6) 1104 (576–1697) 26 (5–55) 75 (25–100) 62 (16–103) 28 (11–41) Within each region, plots selection was carried out using a random stratified sampling approach. As beech is considered as a water-demanding species sensitive to water stress, the stratification was based on the climatic water deficit (P-PET from June to August 1981–2010; in mm). Each region was initially divided into three climatic water deficit zones based on the mean and standard deviation of the climatic water deficit of this region. In each climatic zone, plots were then randomly selected within stands (> 0.5 ha) where beech occupies at least 75% of the canopy cover, as assessed by the National Forest Inventory (© BDFORET v2 IGN). The plots were located at the center of these stands and the minimum distance between two plots was 250 m. Plot area ranged from 800 to 3000 m² (mean 2170 m²). To be selected, plots had to meet the following criteria: (1) absence of signs of logging within the last six years; (2) stand height greater than 16 m or average diameter greater than 10 cm to avoid juvenile stages; (3) plot located in a wooded and accessible area with homogeneous topographic and edaphic conditions; and (4) at least 20 beech trees in the dominant or co-dominant layer. 2.2 Biotic data: health status and stand measurements On each of our 308 plots, we recorded the health status of 20 dominant or co-dominant beech trees during the growing season (except in the Jura where only 10 trees were recorded per plot ; Mirabel and Gaertner, 2023 ) for a total of 5380 trees (Table 1 ). To describe the health status of the selected trees, we used the monitoring protocol of the French Forest Health Department (Saintonge 2023 ), which is based on the European protocol of the ICP-Forests network (Eichhorn et al. 2020 ). For the selected trees, binoculars were used to estimate crown dieback, in two opposite directions, based on the volume proportion of dead and lost branches compared to the potential full crown, and on the crown transparency of the functional crown (i.e., free of competition from neighbouring trees). Crown transparency was assessed as the percentage of total leaf and ramification loss compared to a fully foliated reference beech tree (without considering the dead branches; see Neycken et al. 2024 ). Six classes of crown biomass loss were then determined based on dead branches and crown transparency: A = 0–4.9%, B = 5–24.9%, C = 25–49.9%, D = 50–74.9%, E = 75–99.9%, and F = dead tree (Saintonge 2023 ) (Fig. 2 ). A single operator per region conducted all surveys to limit estimation biases in defoliation assessments, and training programs and intercalibration sessions were conducted to ensure the quality of assessments (e.g. Bussotti et al. 2009 ; Eickenscheidt and Wellbrock, 2014 ; Innes et al. 1993 ). The dominant diameter at breast height (DBH; cm) and the dominant height (Hdom; m) were calculated based on the three largest beeches in diameter, while the mean and coefficient of variation in tree diameters was based on the first 15 beeches closest to the center of the plot (CV_DBH in %; Table 2 ). Stand basal area (m 2 /ha) was measured using a chain relascope for the different species in the stand. It is worth noting that very few or no biotic pathogen were recorded in the stands. 2.3 Soil and topographic data and indices For each plot, a detailed survey of main soil characteristics and local topography was carried out (Table 2 ). Table 2 List of the main climatic, soil, topographic and dendrometric factors used to simulate beech decline STAND DECLINE VARIABLE to predict XStem50 : Percentage of beech trees with at least 50% of crown biomass loss CLIMATE - Frost TYPE 1 - Heat TYPE 2 - Climatic hydric deficit TYPE 3 - Water balance ( including soil) TYPE 4 SOIL TYPE 5 - Available Water Capacity - pH of the upper layer - Proportion of coarse elements - Presence or absence of heavy clay - Presence and depth of hydromorphy - Depth of heavy Hydromorphy - Presence or absence of active lime - Percentage of rock cover on the ground - Presence or absence of active lime Topography TYPE 6 - Slope - Radiation index - Confinement - Topographic position index within 100m and 1500 m radii - Topographic wetness index - General curvature DENDROMETRY TYPE 7 -Tree diameter at breast height (DBH) -Coefficient of variation of beech DBH - Dominant height and DBH of beech - Total basal area - Basal area of beech See Fig. 3 for the climatic variables; see Supplementary Information 1 (SI 1) and 4 (SI 4) for all variables and for the definition of the ‘TYPES’ 2.3.1 Soil data Soil characteristics were estimated directly from a pickaxe and a hand-dug trial pit at the center of each plot. By horizon, soil texture was estimated with the feel method (FAO 2006), and the water content was estimated using the protocol of Baize and Jabiol ( 1995 ) and the pedotransfer functions developed by Jamagne et al. ( 1977 ). Lastly, the available water capacity (AWC; mm) was computed for each plot by multiplying the water content by soil depth considering the proportion of coarse elements. As the root system can extend deeper into the soil than that of the soil pit, we computed the ratio AWC/soil depth i.e. the available water content per centimeter of soil (AWCr; mm/cm) (Algayer et al. 2020 ). The nature of the underlying bedrock was surveyed in the field whenever possible, or else using geological maps. The presence of active limestone was detected in the top 50 cm using 10% hydrochloric acid. The pH was measured at a depth of 20 cm using a ©SoilStick pH meter. Intensity of hydromorphy was noted according to two classes: light (i.e. soil matrix not fully discoloured, with some rust spots or diffuse reduction 35% of soil surface). The depth of the presence of clay in high percentage ( ≥ 45%) was measured. The average percentage of coarse elements was calculated over the entire soil profile. Finally, the percentage of rock cover on the ground was visually estimated. For the Jura region, pH and the presence of active limestone were not measured in the field but were estimated based on NFI data (AgroParisTech 2008 ) and on geological maps. 2.3.2 Topographic indices at plot level and landscape scale The slope, aspect and confinement (i.e., the slope of the line connecting the studied point to the highest part of the landscape in the semicircle defined by the east-south-west direction) were measured using a clinometer and a compass at the center of the plot. Six topographic indices related to plot's water balance were computed based on digital elevation models (DEM) from the National Geographic Institute (“BD ALTI® | Géoservices” n.d.) and using the software © QGIS 3.16 and R 4.2.3 (R Core Team 2023 ): TWI, TPI, and curvature; each index being computed at both 25 m and 75 m resolutions (see details in SI 1). The 25 m resolution refers to plot scale (microtopography), while the 75 m resolution refers to landscape scale (macrotopography). The topographic wetness index (TWI) is an estimate of predicted water accumulation in a defined area. It is calculated as the ratio of the area upslope (i.e., from where water would flow to that point from any given point on the landscape) to the local slope at that point (Galiano et al. 2010 ; Petroselli et al. 2013 ). A high level of TWI corresponds to a higher potential water availability. The topographic position index (TPI) is computed within a 100 m radius for microtopography (TPI100) using the 25 m DEM, and within a 1500 m radius (TPI1500) using the 75 m DEM for macrotopography (Weiss 2001 ). A higher vs lower TPI value indicates a higher vs lower position on the slope. Values close to zero correspond to flat situations or to mid-slopes. Curvature is a measure expressing the extent to which a line deviates from being straight or a surface deviates from being a plane. It is estimated as a second derivative of the surface, and reflects the shape of the slope. The higher the general curvature, the more the plot is situated on convex-shaped topography, favouring water runoff. A low value corresponds to a concave shape, indicating potential water accumulation (Blaga 2012 ; Hengl and Hannes 2021 ). Finally, the radiation index (IKR; inspired from Becker, 1982 ) was computed as the ratio between the mean annual radiation received by the plot (calculated with the DIGITALIS model ; Piedallu and Gégout, 2007 ) and the mean annual radiation that would receive a flat surface corresponding to the horizontal projection of the plot. An IKR > 1.05 corresponds to a warm exposure, while a value below 0.95 indicates a cool situation (Vennetier and Ripert 2005 ). 2.4 Climatic data For each plot, the mean monthly climatic values for the 1981–2010 period were estimated with the AURELHY model for precipitation, mean, minimum and maximum temperatures (Bénichou and Lebreton 1987 ); both models integrating the effects of altitude and topography. For the 30 years preceding the year of measurement, the annual difference of each monthly climatic parameter compared with its 1981–2010 mean was calculated using data from the SAFRAN model (Quintana-Seguí et al. 2008 ). These inter-annual differences in each climatic parameter were then used to correct the 1981–2010 mean from AURELHY and DIGITALIS, in order to downscale them at a finer resolution (method detailed in SI 2). Monthly potential evapotranspiration (PET; mm) was estimated using the Turc method (Turc 1955 ) to determine a monthly climatic water balance (P-PET; mm). These two climatic variables were also calculated for the summer (June - August), two different lengths of the potential growing season (April – October and May to September), and the entire year. We estimated the actual evapotranspiration (AET; mm) of the stand using the Thornthwaite and Mather ( 1957 ) water balance method, incorporating the available soil water capacity estimated from field measurements. The calculated soil water balance allowed us to estimate the number of days during the year when the relative available soil water is lower than 35%, a threshold indicating a water stress for the vegetation (Granier et al. 1999 ). We calculated indices classically used in vegetation studies like the annual hydric deficit PET-AET (Hyd_def; mm) and the ratio of annual hydric deficit AET/PET (Hyd_defr; % ; see Piedallu et al. 2013 ). We subsequently implemented in the models four major types of climatic variables: temperatures related to heatwaves, frost, climatic water deficit, and the water balance incorporating soil and climatic information (Fig. 3). For each climatic parameter, in addition to its (i) absolute value, we also calculated (ii) the absolute difference from the 1981–2010 mean for each period, and (iii) the standardized difference index to the mean 1981–2010. This index was calculated like the standardized precipitation index (SPI; McKee et al. 1993 ) to take into account the deviation from normality of each climatic parameter (SI 3). Finally, we also calculated (iv) the recurrence of years during which a climate threshold was exceeded. In some years (e.g., 2003 and 2018), meteorological conditions can extend far beyond the climatic limits met in average in the natural distribution area of the beech. Although climate variability may have regularly exceeded these limits in the past, it is essential to determine a critical threshold which leads to beech crown dieback. To do so, we computed the above described climatic parameters using the beech distribution area at European scale and the corresponding climatic data (SI 3). After preliminary analyses we used the 0.1 threshold, i.e. the percentile 10% for each climatic parameter calculated over the natural distribution of European beech (SI 3). The absolute values were calculated over five periods: 1, 3, 5, and 15 years preceding the measurement year, and the average 1981–2010; while the absolute difference, standardized difference index, and recurrence of extreme years were calculated over 1, 3, 5, and 15 years preceding the measurement year. 2.5. Statistical approach 2.5.1 Variable of interest Our statistical models aimed at predicting the proportion of trees with at least 50% crown biomass loss (XStem50; %) within each plot; a threshold above which a tree is usually classified as declining (i.e., trees in stages D to F in Fig. 2 ; see Chakraborty et al. 2017 ; Rohner et al. 2021 ; Schmied et al. 2023 ). In a previous study, conducted on Pinus sylvestris in southeastern France, we have shown than XStem50 was a relevant indicator to assess the risk of stand decline (Lemaire et al. 2022 ). 2.5.2. Model development We used partial least squares (PLS) regression (Ter-Braak and Juggins 1993 ) to model beech decline (XStem50) because the number of variables used was large (229 variables, see SI 1). The PLS approach is known to be effective in the case of complex interacting systems (Fernandes 2012 ) with a high number of correlated variables and a limited number of observations (Cramer III et al. 1988 ; Tenenhaus 1998 ). In particular, some climatic and topographic variables based on digital elevation models are highly correlated. We tested a large number of variables with for some of them a high level of correlation, leading to redundancies. To mitigate this effect before computing the PLS regression, we built a dendrogram for each of the 7 types of variables (e.g. 4 types for climate, 1 for topography, 1 for soil and 1 for dendrometry; see Table 2 ). Variables were grouped by clusters within the dendrogram based on their distance ( di ) with di = 1-│ Corr yz │, │ Corr yz │ being the absolute value of the correlation coefficient between variables y and z . Different clusters were determined using a di threshold of 0.15, after testing 10 di thresholds from 0.05 to 0.30 (see SI 4). Within each cluster, the variable with the highest absolute partial standardized coefficient was selected to be tested in the PLS regression. Then, a stepwise method based on the Q² coefficient of Stone-Geisser was used to select the number of components in the PLS regression (SI 4). In the first step of the PLS regression, non-significant variables (p > 0.05) were removed altogether (Tenenhaus 1998 ). In the following steps, among the significant variables, the variable with the lowest standardized coefficient was removed if this led to an increase in Q². The stepwise process continued until the removal of the last variable at stake caused a decrease in Q 2 . The PLS regressions were run with the plsRglm package in R (Bastien et al. 2005 ; Bertrand and Maumy-Bertrand 2018 ). 2.5.3 Predictive accuracy of the regional and national models A model was developed using this stepwise method for each region separately and considering all regions together (thereafter named “NAT” model for ‘national’). To build the NAT model, we randomly resampled the dataset as the number of plots and the mean XStem50 differed among regions (53% of the stands were considered as declining in Jura, while it ranged from 19.7 to 22.4% in the three other regions; Table 1 ). In each region, we selected 10 declining plots and 39 healthy plots to fit the region with the smallest sample size (49 plots in MV; Table 1 ) and the average rate of declining plots in the HFN, MV, and HL regions (21%). We repeated five times this random resampling with replacement, to finally obtain 50 declining plots and 185 healthy ones in each region, for a total of 740 healthy plots and 200 declining plots to build the NAT model. To validate each model and assess their transferability, we compared their ability to predict XStem50 in the region where they have been calibrated, in the three regions outside of their calibration zone, and at the national level. We also tested the national model in each region. The predictive accuracy of the PLS regressions was estimated using the Q 2 coefficient of Stone-Geisser, and the relative importance of each significant variable selected in the models was reflected through the calculation of the weighted loadings (%) (Bastien et al. 2005 ; Bertrand and Maumy-Bertrand 2018 ). To evaluate and compare the performance of the different models in and outside their calibration regions, we also calculated the mean error between the Xstem50 observed in the field and that predicted by the model, and the R² of the linear model fitted on both values. We finally determined if the slope of this regression significantly differed from zero with a t-test. For these calculations, any negative predicted Xstem50 values were adjusted to zero. 3. Results 3.1. Model structure and predictive ability in their calibration region The models of the four regions showed high differences in the accuracy of their predictions (Fig. 4 ), with lower Q² values for Jura (0.26) and Millevaches (MV; 0.27) than for Hauts-de-France/Normandie (HFN; 0.34) and Haut-Languedoc (HL; 0.42). The national model displayed an intermediate value (NAT; 0.31). The cumulative loadings grouped by four major types of variables (i.e. climate, soil, topographic and dendrometric variables; Fig. 4 ) indicated a dominant weight of the climatic parameters in most models: from 47 to 94% in HL, HFN, MV and NAT models. For the Jura region, the three other major types were predominant (cumulative loadings = 63%) but none of these types reached individually a high percentage of loading, neither in any region nor at national scale. More specifically, among the climatic variables, only those related to higher temperatures and water deficit cumulated over 1 to 15 years before measurements were significant, increasing XStem50. No significant variables were identified among the frost parameters and the 30-years climatic average (Table 3 ). In all models, deviations from the 30-years mean and recurrence of critical years exceeding a threshold defined by the percentile 10% in the beech natural area have the highest loadings, except at MV (Fig. 5 ). Table 3 Loadings (%) of the significant variables selected in each regional model and the national model. The sign (+ or -) in parentheses indicates a positive or a negative effect on XStem50, respectively HL HFN Jura MV NAT CLIMATE TEMPERATURE and evapotranspiration Variables Period Years before sampling Absolute value TM_year 3 30.5 (+) 8.3 (+) TX0608 1 19.9 (+) Recurrence TX0608 3 1 (+) Absolute difference compared to climate norms 1981–2010 PET_year 1 14.9 (+) 3.8 (+) TM_year 5 4.9 (+) 15 9.5 (+) PET_year 15 8.2 (+) Standardized Index TM_Year 1 7.6 (+) hydric Deficit Variables Period Years before measures Absolute value Hyd_Defr 1 20 (+) Hyd_Def 14.3 (+) Hyd_Def 3 16.2 (+) Hyd_Def_r 15 4.4 (+) Recurrence Nbdays_HydDef 1 9.6 (+) P-PET0410 17.2 (+) P-PET0509 3 14.2 (+) P-PET0509 15 15 14 (+) P-PET0608 14.3 (+) P-PET_year 7.7 (+) Absolute difference compared to climate norms 1981–2010 P-PET0509 15 17.3 (-) 0.1 (-) Standardized Index Nbdays_HydDef 1 11.9 (+) Hyd_Def 3 1.7 (+) 15 3.1 (+) TOPOGRAPHY Curv_75m 9 (+) 1.7 (+) IKR 5.5 (+) Slope 2.8 (+) 4.6 (+) TWI_25m 1.8 (-) 4.1 (-) TWI_75m 1.1 (-) 11.4 (-) 21.8 (-) SOIL pH 1.5 (+) 6.1 (+) Hydro_depth 4.6 (-) Clay_depth 1 (+) 17.7 (-) 2.5 (+) AWCr 1.5 (-) 3.8 (-) AWC 4.5 (-) 19.8 (-) 3.7 (-) 2.5 (-) Rock_Exp 0.1 (+) DENDROMETRY G 5.3 (-) 26.4 (-) 26.6 (-) 13.2 (-) %G_beech 11.1 (+) 6.6 (+) Hdom 0.7 (-) CV_DBH 3.0 0.1 (+) With TM_year (°C): Mean annual temperature. TX0608 (°C): Mean of the maximum temperature of the 1981–2010 period from June to August. PET_year (mm): Annual potential evapotranspiration. Hyd_Def (mm): Annual hydric deficit. Hyd_Def_r (%): Ratio of annual hydric deficit. Nbdays_HydDef : number of days during which the relative extractable soil water is below 40%. P-PET0410, P-PET0509, P-PET0608 (mm): Climatic water balance from April to October, May to September, and June to August, respectively. P-PET_year (mm): Annual climatic water balance. Curv_75m : General curvature calculated using the 75m DEM. IKR : Radiation index. Slope (%): Slope measured at the center of the plot. TWI_75m and TWI_25m : Topographic wetness indices calculated using the 75m and 25 m DEM resolutions, respectively. Hydro_depth (cm): Depth of light or heavy hydromorphy presence. Clay_depth (cm): Depth of heavy clay presence. AWC (mm): Available water capacity. AWCr (mm/cm): Relative available water capacity per centimeter of soil (AWC/depth of profile). Rock_Exp (%): Percentage of rock exposure on the ground. G (m²/ha): Total basal area. % G_beech (%): Beech proportion in the total basal area. Hdom (m): Dominant height of beech. CV_DBH (%): Coefficient of variation of diameter at breast height XStem50 was lower on soils with high available water capacity in all regions (Table 3 ). However, the effect of heavy clay, with high water-holding capacity, had opposite effects: detrimental in the Jura and at national scale where it increased XStem50, but favourable in HFN. In HL, the soils with higher pH and lower water content were more prone to decline compared to the deeper or more acidic soils. The national model combines all of these soil factors with the same sense of correlation but with different loadings. Plots located in topographic situations unfavourable for water balance showed stronger tree decline. In all regions, the Topographic Wetness Index (TWI) was negatively correlated with XStem50, but the resolution of the concerned DEM varied by region. More specifically, plots situated on convex topography in the Jura and on the national scale, those on south-facing slopes in the Jura, and on steeper slopes in HL and on the national level exhibited a higher XStem50. TWI at75 m was the only significant topographic factor in HFN et MV. Concerning the dendrometric characteristics, stands with a high total basal area were less prone to decline in all regions except in HFN; and this effect explained about 26% of the variance in Xstem50 in the Jura and MV regions. In addition to stand density, the decline was stronger in stands dominated by beech (high proportion of beech in basal area) in HL and on national scale, and with more heterogeneous diameter distributions in HFN and on national scale (but with loadings < 3%). 3.2 Accuracy of the regional and national models outside their calibration range 3.2.1. Regional level Each regional model was by far and logically the best in its calibration region (Fig. 6). In this case, the slope of the linear model (predicted vs observed) is highly significant (p < 0.001) and the mean error is the lowest. Other models can provide inaccurate predictions of plot decline with a mean error ranging from − 28% to + 21% on regional scale. For Jura, only the local and national models were significant (Fig. 6A). The MV model was the least accurate when applied to regions outside its calibration zone, exhibiting a systematic underestimation, while HFN model overestimated Xstem50 outside of its calibration region. In all regions, the models systematically underestimated XStem50 when it was high and overestimated it when it was close to zero (see example in Fig. 7 ). 3.2.2 National level The national model consistently performed better in each region that non-local regional models on the basis of the R² and mean error criteria (Fig. 6). On average across all regions, the national model could explain 32% of the total variance, whereas the regional models accounted for 46%. In HL, the national model explained 60% of the total variance, very close to the regional model (62%). In contrast, the variance explained by the national model was only half of that of the regional model for MV (22% vs 44%). Conversely, the predictive power of each regional model applied on national scale was very low, averaging 11% of the total variance, and represented only 5 to 18% of the variance explained by the national model. 3.2.3. The example of the Haut-Languedoc region. Using the Haut-Languedoc (HL) region as example to compare the predictive accuracy of the 5 models, the HL model was logically the best when applied there, followed by the national model (Figs. 6 and 7 ). Conversely, the MV and Jura models strongly underestimated the observed decline, especially for the highest XStem50 values. Indeed, the Jura model was strongly influenced by soil, topographic and dendrometric parameters and far less by climate. In this region, the environmental and silvicultural conditions strongly differ from Haut Languedoc: the beech forests were (i) submitted to a stronger warming in the 15 years preceding the measurement (+ 0.58°C in the Jura compared to HL; p < 0.001); (ii) more often located on north-facing slopes (mean IKR = 1.0 in Jura and 0.95 in HL; p = 0.008); (iii) located on soils with higher available water capacity (138 mm vs. 90 mm; p < 0.001); (iv) and less dense than in HL (G = 24.1 m²/ha vs. 32.3 m²/ha; p < 0.001). Even though mean water deficit and its inter-annual variability estimated over the previous 3 and 15 years (significant periods in the HFN model ; see Table 3 ) were significantly higher in HL than in HFN (p < 0.05), XStem50 in HL was consistently overestimated by the HFN model (+ 21%). The climatic parameters recorded in these two regions were not significantly different in the year prior to measurement. Furthermore, there were no significant differences in soil and dendrometric parameters between the two regions. 4. Discussion This study, based on multivariate statistical models and conducted on European beech, confirms its high sensitivity to high temperature and drought. Soil characteristics, notably the available water capacity, and topography are important factors mitigating the risk of beech decline. However, there are also other factors that interact to either mitigate or aggravate the decline, such as dendrometric parameters associated with stand structure and composition. Each regional model was by far the most precise in its calibration area, but the larger-scale national model seems to be an acceptable compromise for estimating the risk of decline across all the study regions. 4.1. Performance and structure of the models in their calibration region The different statistical models developed to predict beech decline showed variable performances according to the study region. The best model, i.e. with the highest Q² value (Fig. 4 ), was calibrated in the HL region. In this region submitted to a Mediterranean climate, where beech is at the edge of its distribution range, climate loadings are high indicating that the climatic factors are the main drivers of beech decline, which started in the early 2000s (Silva 2010 ; Cavin and Jump 2017 ). In Jura, forest health was worse than in the other regions (Table 1 ), mainly due to the 2018 drought, which had a significant impact on forest health in Central and Northern Europe (Buras et al. 2020 ; Braun et al. 2021 ; Rukh et al. 2023 ). Sampling in this mountainous area was limited to low-altitude (< 800 m) beech forests (Mirabel and Gaertner 2023 ) where decline was important, and where climatic conditions were systematically unfavourable for beech in 2018. This can probably explain why climate variables, which were quite homogeneous within plots, had low loadings, leading to the lowest predictive accuracy among the five models. But it also clearly highlights the compensating role of the three other groups of factors (topography, soil and dendrometry), which have in Jura their highest loadings. In MV, the decline was less prevalent. At the time of measurement, i.e. the beginning of 2018, beech decline was not intense in this region (Mirabel et al. 2024 ) and consequently difficult to predict. In HFN where the decline of beech started in the early 1990s (Nageleisen 1993 , Pilard-Landeau et al. 1994 ), the regional model had an intermediate predictive accuracy, climatic variables having by far the highest weight. The intensive management that occurred in this region likely decreased the model performance. The different models thus include variables of different nature and with contrasting loadings, which we discuss below. 4.1.1.Influence of climate factors Among the variables that best explained the variability in beech decline within each region and also among regions, the climatic parameters linked to water stress and heat were of primary importance (Table 3 , Figs. 4 and 5 ). This aligns with a vast and consistent literature on beech sensitivity to water deficit and heat, especially its susceptibility to xylem cavitation (Bréda et al. 2006 ; Braun et al. 2021 ; Walthert et al. 2021 ). We did not find any effect of the 30-years climatic average, which was expected as the plots were mainly located in climatic zones favourable to beech (Fig. 1 B). Conversely, the interannual climate variability explained most of the beech decline, particularly exceptionally intense drought and high temperatures, or repeated droughts (see factors of recurrence, difference and standardized indices in Table 3 and Fig. 5 ). This result is consistent with the observation that extensive decline events occurred during periods of elevated temperatures and low precipitation, such as in 2003 and 2018, even in beech core range (Leuschner, 2020). In addition, we observed long-term effects of climate: beech forests that were submitted to a stronger warming, a higher mean hydric deficit, or recurrence of extreme droughts over the last 15 years, showed the highest Xstem50. Such lag-effects have been discussed in numerous studies, and especially highlighted by dendrochronological analyses. For instance, Neycken et al. ( 2022 ) found in Switzerland that beech trees that exhibited severe crown dieback after a year of high hydric stress showed a stronger growth decline than healthy trees over the last 50 years. Drought-induced changes in growth trajectories may be visible before the signs of crown dieback (Neycken et al. 2024 ). 4.1.2 Influence of topography and soil factors The different models demonstrated the importance of soil and topographic factors in either compensating or exacerbating the impact of climatic events on beech health (Table 3 ; Fabiani et al. 2024), mainly because of differences in geological, pedological and topographic conditions among regions. In HDF, beech trees are found in flat areas, while in Jura, they are present both in the plains and at higher elevations in the mountains. At medium and low elevation in Jura, this species can be found on both south- and north-facing slopes. In HL, with a Mediterranean climate, beech trees are more commonly found at high elevations and are primarily located on north-facing slopes (IKR is significantly lower in HL than in Jura). The Topographic Wetness Index (TWI) had a significant effect in all models, indicating that the risk of decline is higher in stands located on convex landform (high curvature) and steep slopes. It is also higher with a southern exposure (Chaplot and Walter 2003 ; Schmidt and Persson 2003 ). These topographic variables are closely related to higher water runoff and lower soil depth (Salvador-Blanes et al. 2006 ; Tesfa et al. 2009 ; Chartin et al. 2011 ), i.e. to a low available water capacity (AWC). AWC is also significantly related to XStem50 in all regions: it is known to play a crucial role in maintaining beech vitality in temperate macroclimatic conditions (Chakraborty et al. 2017 ). In regions with high soil stoniness and compact texture, the relative AWC (AWCr, i.e. the available water content per centimeter of soil) seems to be a better estimator of water availability to roots than the absolute AWC. In these types of soils, it is impossible to assess the depth of soil explored by roots with an auger and pickaxe, and the available water capacity is often underestimated (Arrouays et al. 2014 ; Tetegan et al. 2015 ). The Jura model differed from the other models because soil and topographic variables had high loadings. In Jura, plots located on marls containing more than 50% of clay were highly declining. Such soils are characterized by high compactness, and often a low AWC which are unfavourable to beech (Čermák et al. 1993 ; Obladen et al. 2021 ). Conversely, the presence of clay limited tree declines in HFN. In this region, clay content never exceeds 50% and is often mixed with coarser elements, resulting in less compacted soils. 4.1.3 Influence of stand structure and composition We found that stands with a high total basal area were less prone to decline in HL, Jura, MV; an observation also made at the national level (Table 3 ). This result was not expected and contradicts those of the systematic ICP-Forests network, where stands with high basal area presented a higher level of defoliation (Toïgo et al. 2020 ). Three complementary hypotheses may explain this difference. First, the ICPF network has a long-term monitoring history, while there was only one measurement per plot and no knowledge of past silviculture in our study. The time elapsed between our measurement and the first signs of decline varied across the regions, and could be long (e.g. first reports in HDF in the 1990s while our measurements were conducted in 2015–2016). In the meantime, the probability of sanitary thinning was high, especially because beech is a high-value hardwood (Armand 2002 ) with low wood durability (Pramreiter and Grabner 2023 ) and is often promptly harvested when stands show signs of decline (Brunier et al. 2020 ). It is therefore likely that such thinnings were carried out by selectively harvesting the most declining or dead trees, leading to an underestimation of XStem50 (Lech and Kamińska 2024 ). We tried to avoid this bias by excluding plots that have been thinned in the last six years, but this underestimation is likely for stands with long-term decline history and in regions where harvesting is frequent and rather based on tree health than on tree status in stand structure. The second hypothesis is related to site fertility. At the same level of thinning intensity, stand basal area is generally higher in sites that are more favourable for beech (e.g. soil with higher AWC, located on northern aspect( Rohner et al. 2018 )), which can explain the negative relationship between Xstem50 and stand basal area. The third hypothesis is related to the intensity of thinning, which can lead to contrasted impacts of tree health. If too intense, thinning can increase the water demand at crown level, particularly for large trees, leading to a higher risk of xylem cavitation and mortality (McDowell and Allen 2015 ). In contrast, a moderate thinning can be beneficial by reducing the water uptake at stand level, and thus by boosting growth of the remaining trees (van der Maaten 2013 ; Diaconu et al. 2017 ). Previous studies have shown that beech sensitivity to drought is lower in mixture than in monospecific stands, e.g., when growing with Pinus sylvestris (Metz et al. 2016 ) or other hardwood species (Mölder and Leuschner 2014 ; Vannoppen et al. 2019 ). Indeed, in HL, mixed beech stands were less affected by decline (Table 3 ). The percentage of beech in terms of basal area was higher in HL (p < 0.001, Table 1 ), but with a broader range. In the more mixed stands of this region, this mixture may have a favorable effect on the health of beech, particularly in an area influenced by a Mediterranean climate. 4.2 Assessing the potential extrapolation of the regional and national models In each region, the best model for predicting XStem50 is logically the local model, and the second-best model is the national one (Fig. 6). The prediction of XStem50 by regional models outside their calibration area is less reliable and sometimes completely inaccurate, confirming that increasing the complexity and precision of models can be at the expense of their transferability (Naas et al. 2024 ; but see also Pichler and Hartig 2023 ). This was expected as the selected variables and parameter estimates (~ loadings) of the regional models are closely adjusted to the local climatic, edaphic, topographic, and dendrometric factors. In contrast, the national model integrates and balances the weight of these environmental factors at a larger scale, keeping only the most important and common ones. First, the rate of decline differs among regions, leading to systematic under- or over-estimation in the model predictions when applied in another region. For instance, in MV, declining trees at the time of measurements were quite rare, with a maximum XStem50 of 45% (Table 1 ). Thus, the MV model could hardly predict decline levels as high as those observed in HL, where the maximum XStem50 was 90%. Second, the periods when decline began and dates of measurements differed among regions. Therefore, the periods used to compute the climatic parameters also differ among models. If measurements are delayed many years after the climatic events that trigger the decline process, trees can recover if the climatic conditions become favourable again, and there is a loss of correlation between the climatic signal and XStem50. Third, decline is a complex phenomenon influenced by interactions between climate, soil conditions, stand characteristics, topographic factors, past management as well as presence of biotic agents (Galiano et al. 2010 ; Camarero et al. 2017 ; Chakraborty et al. 2017 ; Lemaire et al. 2022 ). For instance, the Jura model was mainly influenced by soil, topography, and dendrometric factors (cumulative loadings = 90%) and far less by climate, unlike other regions. It strongly underestimated the decline in HL were the soils are sandy-loamy without clay or hydromorphy which both played a key role in increasing XStem50 in Jura. In HL, beech stands were generally located on north-facing slopes with a low IKR, limiting high temperatures and drought in this sub-Mediterranean area, compared to Jura with its colder climate. In the same way, by testing the HFN model on HL data in the south of France, predictions of XStem50 by the HFN model were systematically overestimated by 21% compared to the predictions of the HL model (Fig. 7 ). Conversely, when the HL model was applied to HFN data in northern France, the model predicted a slight decline and underestimated XStem50. The climatic variables related to water deficit and temperature were predominant in these two models (Fig. 4 ). XStem50 measured in HFN was very likely underestimated compared to the actual decline. Indeed, the beech forests in HFN had a higher economic value than those in HL, leading to more intensive timber exploitation in this region (IGN 2012 ). The time period over which climatic variables were considered in both models differed. It was based on the year for HL and on the summer or vegetation period for HFN. The vegetation period of beech trees indeed varied with longitude, latitude, and altitude, and was longer in the warmer plots of HL (Lebourgeois et al. 2002 ). Finally, there were factors not considered in our models that might explain the differences in predicted XStem50, such as phenotypic and genetic adaptations, which could have played an essential role in beech tree resistance to water stress (Leuschner, 2020). It is also possible that beech trees in the four regions exhibited different morphological and physiological adaptations to their own climates, as suggested by Knutzen et al. (2015). For instance, the root system of southern beech forests might be deeper with notable changes in the depth extension of some root diameter classes (Meier et al. 2018), increasing their drought resistance (Bolte et al. 2016 ). All these reasons could explain why it is difficult to apply a statistical model of tree decline outside of its calibration area. 5. Conclusion Our main objective was to model decline intensity – measured by crown biomass loss (XStem50) - of beech forests in four regions of France with contrasted environmental conditions. Using PLS regression techniques, our statistical models predicted XStem50 from a large set of factors with a relatively good prediction (Q² varied from 0.26 to 0.42). Climatic factors were of primary importance. Among them, deviations from the 1981–2010 average related to high temperature and water deficit played the most important role. This illustrates the fact that beech forests are locally adapted to the regional climate in which they grow, but are sensitive to deviations from the “average” regional past climate. With climate change, these deviations are projected to be more intense and more frequent threatening even more the health condition of these forests in the near future. However, we also found that soil and topographic factors with a strong influence on the local water budget could mitigate or aggravate the decline, with specificities among regions according to their local climate and geology, and main stand characteristics. Such a modelling approach can be applied to other European regions but can be improved by including information on past stand dynamics and management, and variation in vulnerability among populations. However, they remain a valuable tool for forest managers to predict which beech forests will be at risk of decline in future climatic conditions and to implement preventive silvicultural actions adapted to local site conditions. Declarations Author Contribution J LEMAIRE : Conceptualization text, figures, Formal Analysis, Writing : draft, review and editing. M VENNETIER et B PREVOSTO : Writing : review and editingM CAILLERET : Formal analysis, writing : review and editing References AGROPARISTECH IFN (2008) Carte du pH de surface des sols forestiers français. https://silvae.agroparistech.fr/public/documents/carte_PH_version1-formatA3.pdf . Accessed 29 Jan 2024 Algayer B, Lagacherie P, Lemaire J (2020) Adapting the available water capacity indicator to forest soils: An example from the Haut-Languedoc (France). Geoderma 357:113962. https://doi.org/10.1016/j.geoderma.2019.113962 Allen CD, Breshears DD, McDowell NG (2015) On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene. Ecosphere 6:art129. https://doi.org/10.1890/ES15-00203.1 Allen CD, Macalady AK, Chenchouni H et al (2010) A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. Ecol Manag 259:660–684. https://doi.org/10.1016/j.foreco.2009.09.001 Arend M, Link RM, Zahnd C et al (2022) Lack of hydraulic recovery as a cause of post-drought foliage reduction and canopy decline in European beech. New Phytol 234:1195–1205. https://doi.org/10.1111/nph.18065 Armand G (2002) Le hêtre autrement. Institut pour le Développement Forestier.262 p Arrouays D, McKenzie N, Hempel J et al (2014) GlobalSoilMap: Basis of the global spatial soil information system. CRC. https://doi.org/10.1201/b16500 Baize D, Jabiol B (1995) Guide pour la description des sols. Editions Quae.430 p Bastien P, Vinzi VE, Tenenhaus M (2005) PLS generalised linear regression. Comput Stat Data Anal 48:17–46. https://doi.org/10.1016/j.csda.2004.02.005 BD ALTI® | Géoservices https://geoservices.ign.fr/bdalti . Accessed 6 Sep 2021 Becker M (1982) Influence relative du climat et du sol sur les potentialités forestières en moyenne montagne. Exemple des sapinières à fétuque (Festuca silvatica Vill.) dans les Vosges alsaciennes. Ann Sci For 39:1–32. https://doi.org/10.1051/forest:19820101 Bénichou P, Lebreton O (1987) Prise en compte de la topographie pour la cartographie des champs pluviométriques statistiques. La météorologie 7è série:23–34 Bertrand F, Maumy-Bertrand M (2018) plsRglm: Partial least squares linear and generalized linear regression for processing incomplete datasets by cross-validation and bootstrap techniques with R. ArXiv181001005 Stat Blaga L (2012) Aspects regarding the signifiance of the curvature types and values in the studies of geomorphometry assisted by GIS. Analele Univ Din Oradea Ser Geogr Year XXII no. 2/2012:327–337 Bolte A, Czajkowski T, Cocozza C et al (2016) Desiccation and Mortality Dynamics in Seedlings of Different European Beech (Fagus sylvatica L.) Populations under Extreme Drought Conditions. Front Plant Sci 7. https://doi.org/10.3389/fpls.2016.00751 Braun S, Hopf S-E, Tresch S et al (2021) 37 Years of Forest Monitoring in Switzerland: Drought Effects on Fagus sylvatica. Glob Chang Oct. https://doi.org/10.3389/ffgc.2021.765782 . 2021 4: Bréda N, Huc R, Granier A, Dreyer E (2006) Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences. Ann Sci 63:625–644. https://doi.org/10.1051/forest:2006042 Brunier L, Delport F, Gauquelin X (2020) Guide de gestion des crises sanitaires en forêt., CNPF-IDF.184 p Buras A, Rammig A, Zang CS (2020) Quantifying impacts of the 2018 drought on European ecosystems in comparison to 2003. Biogeosciences 17:1655–1672. https://doi.org/10.5194/bg-17-1655-2020 Bussotti F, Cozzi A, Cenni E et al (2009) Quality Assurance and measurement errors in monitoring tree crown conditions in Italy. J Environ Monit 11:769–773. https://doi.org/10.1039/B818166G Bussotti F, Feducci M, Iacopetti G et al (2018) Linking forest diversity and tree health: preliminary insights from a large-scale survey in Italy. Ecosyst 5:12. https://doi.org/10.1186/s40663-018-0130-6 Camarero JJ, Linares JC, Sangüesa-Barreda G et al (2017) The Multiple Causes of Forest Decline in Spain: Drought, Historical Logging, Competition and Biotic Stressors. In: Amoroso MM, Daniels LD, Baker PJ, Camarero JJ (eds) Dendroecology: Tree-Ring Analyses Applied to Ecological Studies. Springer International Publishing, Cham, pp 307–323 Caudullo G, Welk E, San-Miguel-Ayanz J (2017) Chorological maps for the main European woody species. Data Brief 12:662–666. https://doi.org/10.1016/j.dib.2017.05.007 Cavin L, Jump AS (2017) Highest drought sensitivity and lowest resistance to growth suppression are found in the range core of the tree Fagus sylvatica L. not the equatorial range edge. Glob Change Biol 23:362–379. https://doi.org/10.1111/gcb.13366 Čermák J, Matyssek R, Kučera J (1993) Rapid response of large, drought-stressed beech trees to irrigation. Tree Physiol 12:281–290. https://doi.org/10.1093/treephys/12.3.281 Chakraborty T, Saha S, Matzarakis A, Reif A (2017) Influence of multiple biotic and abiotic factors on the crown die-back of European beech trees at their drought limit. Flora 229:58–70. https://doi.org/10.1016/j.flora.2017.02.012 Chaplot V, Walter C (2003) Subsurface topography to enhance the prediction of the spatial distribution of soil wetness. Hydrol Process 17:2567–2580. https://doi.org/10.1002/hyp.1273 Chartin C, Bourennane H, Salvador-Blanes S et al (2011) France Geomorphology 135:8–20. https://doi.org/10.1016/j.geomorph.2011.07.020 . Classification and mapping of anthropogenic landforms on cultivated hillslopes using DEMs and soil thickness data — Example from the SW Parisian Basin, Chauvier Y, Thuiller W, Brun P et al (2021) Influence of climate, soil, and land cover on plant species distribution in the European Alps. Ecol Monogr 91:e01433. https://doi.org/10.1002/ecm.1433 Cheaib A, Badeau V, Boe J et al (2012) Climate change impacts on tree ranges: model intercomparison facilitates understanding and quantification of uncertainty. Ecol Lett 15:533–544. https://doi.org/10.1111/j.1461-0248.2012.01764.x Chira D, Danescu F, Rosu C et al (2006) Aspects on European beech in some experimental plots in Oriental Carpanthians. In: Proceedings. International Conference. Beech silviculture in Europe’s largest beech country. IUFRO, pp 111–113 Choat B, Jansen S, Brodribb TJ et al (2012) Global convergence in the vulnerability of forests to drought. Nature 491:752–755. https://doi.org/10.1038/nature11688 Cramer RD III, Bunce JD, Patterson DE, Frank IE (1988) Crossvalidation, Bootstrapping, and Partial Least Squares Compared with Multiple Regression in Conventional QSAR Studies. Quant Struct-Act Relatsh 7:18–25. https://doi.org/10.1002/qsar.19880070105 Csilléry K, Lalagüe H, Vendramin GG et al (2014) Detecting short spatial scale local adaptation and epistatic selection in climate-related candidate genes in European beech (Fagus sylvatica) populations. Mol Ecol 23:4696–4708. https://doi.org/10.1111/mec.12902 De Frenne P, Lenoir J, Luoto M et al (2021) Forest microclimates and climate change: Importance, drivers and future research agenda. Glob Change Biol 27:2279–2297. https://doi.org/10.1111/gcb.15569 Diaconu D, Kahle H-P, Spiecker H (2017) Thinning increases drought tolerance of European beech: a case study on two forested slopes on opposite sides of a valley. Eur J Res 136:319–328. https://doi.org/10.1007/s10342-017-1033-8 Dittmar C, Elling W (2007) Dendroecological investigation of the vitality of Common Beech ( Fagus sylvatica L.) in mixed mountain forests of the Northern Alps (South Bavaria). Dendrochronologia 25:37–56. https://doi.org/10.1016/j.dendro.2007.01.003 Dittmar C, Fricke W, Elling W (2006) Impact of late frost events on radial growth of common beech (Fagus sylvatica L.) in Southern Germany. Eur J Res 125:249–259. https://doi.org/10.1007/s10342-005-0098-y Dobbertin M (2005) Tree growth as indicator of tree vitality and of tree reaction to environmental stress: a review. Eur J Res 124:319–333. https://doi.org/10.1007/s10342-005-0085-3 Dobbertin M, Brang P (2001) Crown defoliation improves tree mortality models. Ecol Manag 141:271–284. https://doi.org/10.1016/S0378-1127(00)00335-2 Eichhorn J, Roskams P, Potocic (2020) ICP Forests Manual: Part IV Visual assessment of crown condition and damaging agents. In: ICP Forests Manual. 55 p. https://www.icp-forests.org/pdf/manual/2020/ICP_Manual_part04_2020_Crown_version_2020-3_update_06-2023.pdf Eickenscheidt N, Wellbrock N (2014) Consistency of defoliation data of the national training courses for the forest condition survey in Germany from 1992 to 2012. Environ Monit Assess 186:257–275. https://doi.org/10.1007/s10661-013-3372-3 FAO (ed) (2006) Guidelines for soil description, 4. ed. Food and Agriculture Organization of the United Nations, Rome.109 p. https://www.fao.org/4/a0541e/a0541e.pdf Fernandes V (2012) En quoi l’approche PLS est-elle une méthode a (re)-découvrir pour les chercheurs en management ? M@n@gement 15:102–123. https://doi.org/10.3917/mana.151.0102 Ferretti M (2021) New appetite for the monitoring of European forests. Ann Sci 78:94. https://doi.org/10.1007/s13595-021-01112-w Franklin JF, Shugart HH, Harmon ME (1987) Tree Death as an Ecological Process. Bioscience 37:550–556. https://doi.org/10.2307/1310665 Frei ER, Gossner MM, Vitasse Y et al (2022) European beech dieback after premature leaf senescence during the 2018 drought in northern Switzerland. Plant Biol 24:1132–1145. https://doi.org/10.1111/plb.13467 Galiano L, Martínez-Vilalta J, Lloret F (2010) Drought-Induced Multifactor Decline of Scots Pine in the Pyrenees and Potential Vegetation Change by the Expansion of Co-occurring Oak Species. Ecosystems 13:978–991. https://doi.org/10.1007/s10021-010-9368-8 Granier A, Bréda N, Biron P, Villette S (1999) A lumped water balance model to evaluate duration and intensity of drought constraints in forest stands. Ecol Model 116:269–283. https://doi.org/10.1016/S0304-3800(98)00205-1 Hartmann H, Moura CF, Anderegg WRL et al (2018) Research frontiers for improving our understanding of drought-induced tree and forest mortality. New Phytol 218:15–28. https://doi.org/10.1111/nph.15048 Hengl T, Hannes I (2021) Geomorphometry: Concepts, Software. Applications - International Society for Geomorphometry, Elsevier Houston Durrant T, De Rigo D, Mauri A et al (2016) European atlas of forest tree species. Office des publications de l’Union européenne, LU Hülsmann L, Bugmann HKM, Commarmot B et al (2016) Does one model fit all? Patterns of beech mortality in natural forests of three European regions. Ecol Appl 26:2465–2479. https://doi.org/10.1002/eap.1388 IGN (2012) IF: Quelles sont les ressources exploitables ? IGN https://inventaire-forestier.ign.fr/IMG/pdf/IF31.pdf Innes JL, Landmann G, Mettendorf B (1993) Consistency of observations of forest tree defoliation in three European countries. Environ Monit Assess 25:29–40. https://doi.org/10.1007/BF00549790 Jamagne M, Bétrémieux R, Bégon JC, Mori A (1977) Quelques données sur la variabilité dans le milieu naturel de la réserve en eau des sols. Bull Tech Inf 324–325:627–641 Joly D, Brossard T, Cardot H et al (2010) Les types de climats en France, une construction spatiale. Cybergeo Eur J Geogr. https://doi.org/10.4000/cybergeo.23155 Langer GJ, Bußkamp J (2023) Vitality loss of beech: a serious threat to Fagus sylvatica in Germany in the context of global warming. J Plant Dis Prot 130:1101–1115. https://doi.org/10.1007/s41348-023-00743-7 Lebourgeois F, Differt J, Granier A et al (2002) Premières observations phénologiques des peuplements du réseau national de suivi à long terme des écosystèmes forestiers (Renecofor). Rev Fr 54:407–418. https://doi.org/10.4267/2042/4930 Lech P, Kamińska A (2024) Mortality, or not mortality, that is the question … How to Treat Removals in Tree Survival Analysis of Central European Managed Forests. Plants 13:248. https://doi.org/10.3390/plants13020248 Lemaire J, Vennetier M, Prévosto B, Cailleret M (2022) Interactive effects of abiotic factors and biotic agents on Scots pine dieback: A multivariate modeling approach in southeast France. Ecol Manag 526:120543. https://doi.org/10.1016/j.foreco.2022.120543 Leuschner C (2020a) Drought response of European beech (Fagus sylvatica L.)—A review. Perspect Plant Ecol Evol Syst 47:125576. https://doi.org/10.1016/j.ppees.2020.125576 Leuschner C (2020b) Drought response of European beech (Fagus sylvatica L.)—A review. Perspect Plant Ecol Evol Syst 47:125576. https://doi.org/10.1016/j.ppees.2020.125576 Leuschner C, Ellenberg H (2017) Vegetation Ecology of Central Europe, Springer, Berlin, Germany. Springer, Berlin https://link.springer.com/book/10.1007/978-3-319-43042-3 Lies E (1980) Le dépérissement du hêtre au Grand-Duché de Luxembourg. Ann Sci For 37:275–277. https://doi.org/10.1051/forest:19800402 Martinez del Castillo E, Zang CS, Buras A et al (2022) Climate-change-driven growth decline of European beech forests. Commun Biol 5:1–9. https://doi.org/10.1038/s42003-022-03107-3 Mathes T, Seidel D, Klemmt H-J et al (2024) The effect of forest structure on drought stress in beech forests (Fagus sylvatica L). Ecol Manag 554:121667. https://doi.org/10.1016/j.foreco.2023.121667 Mauri A, Strona G, San-Miguel-Ayanz J (2017) EU-Forest, a high-resolution tree occurrence dataset for Europe. Sci Data 4:160123. https://doi.org/10.1038/sdata.2016.123 McDowell NG, Allen CD (2015) Darcy’s law predicts widespread forest mortality under climate warming. Nat Clim Change 5:669–672. https://doi.org/10.1038/nclimate2641 McKee TB, Doesken NJ, Kleist J (1993) The relationship of drought frequency and duration to time scales. Eighth Conf Appl Climatol, Anaheim, 17–22 January 1993, 179–184 Mellert KH, Lenoir J, Winter S et al (2018) Soil water storage appears to compensate for climatic aridity at the xeric margin of European tree species distribution. Eur J Res 137:79–92. https://doi.org/10.1007/s10342-017-1092-x Metz J, Annighöfer P, Schall P et al (2016) Site-adapted admixed tree species reduce drought susceptibility of mature European beech. Glob Change Biol 22:903–920. https://doi.org/10.1111/gcb.13113 Mirabel M, Gaertner P-A (2023) Dépérissements de hêtres adultes après 2018 en Franche-Comté. Rev Fr 74:177–186. https://doi.org/10.20870/revforfr.2023.7591 Mirabel M, Saintonge F-X, Caroulle F (2024) Bilan de l’état sanitaire du hêtre en France fin 2023. Dép Santé For Juillet 2024:14 p. https://draaf.bourgogne-franche-comte.agriculture.gouv.fr/IMG/pdf/dsf_bilan_hetre_juin_2024.pdf Mölder I, Leuschner C (2014) European beech grows better and is less drought sensitive in mixed than in pure stands. https://doi.org/10.1007/s00468-014-0991-4 Müller E, Stierlin HR (1990) Sanasilva Tree crown photos., Sanasilva. 98p Naas AE, Keetz LT, Halvorsen R et al (2024) Choice of predictors and complexity for ecosystem distribution models: effects on performance and transferability. Ecography 2024:e07269. https://doi.org/10.1111/ecog.07269 Nageleisen L-M (1994) Le dépérissement actuel de feuillus divers: hêtre, merisier, alisier torminal, érable sycomore, peuplier, châtaignier, charme, aulne glutineux. Rev Fr 554. https://doi.org/10.4267/2042/26583 Nageleisen L-M (2005) Dépérissement du Hêtre: présentation d’une méthode symptomatologique de suivi. Rev Fr 57:255–262. https://doi.org/10.4267/2042/5041 Nageleisen L-M (1993) Les dépérissements d’essences feuillues en France. Rev Fr 45:605–620. https://doi.org/10.4267/2042/26462 Nageleisen L-M, Reuter J-C (2007) Dépérissement de hêtre: retour sur les situations de crise de la hêtraie lors de la dernière décennie. Bilan Santé For. 2006 7 Neycken A, Scheggia M, Bigler C, Lévesque M (2022) Long-term growth decline precedes sudden crown dieback of European beech. Agric Meteorol 324:109103. https://doi.org/10.1016/j.agrformet.2022.109103 Neycken A, Wohlgemuth T, Frei ER et al (2024) Slower growth prior to the 2018 drought and a high growth sensitivity to previous year summer conditions predisposed European beech to crown dieback. Sci Total Environ 912:169068. https://doi.org/10.1016/j.scitotenv.2023.169068 Obladen N, Dechering P, Skiadaresis G et al (2021) Tree mortality of European beech and Norway spruce induced by 2018–2019 hot droughts in central Germany. Agric Meteorol 307:108482. https://doi.org/10.1016/j.agrformet.2021.108482 Ognjenović M, Seletković I, Potočić N et al (2022) Defoliation Change of European Beech (Fagus sylvatica L.) Depends on Previous Year Drought. Plants 11:730. https://doi.org/10.3390/plants11060730 Peñuelas J, Hunt JM, Ogaya R, Jump AS (2008) Twentieth century changes of tree-ring δ13C at the southern range-edge of Fagus sylvatica: increasing water-use efficiency does not avoid the growth decline induced by warming at low altitudes. Glob Change Biol 14:1076–1088. https://doi.org/10.1111/j.1365-2486.2008.01563.x Peterken GF, Mountford EP (1996) Effects of drought on beech in Lady Park Wood, an unmanaged mixed deciduous woodland | Forestry: An International Journal of Forest Research | Oxford Academic. Int J Res 69:125–136. https://doi.org/10.1093/forestry/69.2.125 Petit-Cailleux C, Davi H, Lefèvre F et al (2021) Comparing statistical and mechanistic models to identify the drivers of mortality within a rear-edge beech population. Peer Community J. https://doi.org/10.24072/pcjournal.60 . 1: Petroselli A, Vessella F, Cavagnuolo L et al (2013) Ecological behavior of Quercus suber and Quercus ilex inferred by topographic wetness index (TWI). Trees 27:1201–1215. https://doi.org/10.1007/s00468-013-0869-x Pichler M, Hartig F (2023) Machine learning and deep learning—A review for ecologists. Methods Ecol Evol 14:994–1016. https://doi.org/10.1111/2041-210X.14061 Piedallu C, Gégout JC (2007) Multiscale computation of solar radiation for predictive vegetation modelling. Ann Sci For 64:899–909 Piedallu C, Gégout J-C, Perez V, Lebourgeois F (2013) Soil water balance performs better than climatic water variables in tree species distribution modelling. Glob Ecol Biogeogr 22:470–482. https://doi.org/10.1111/geb.12012 Pilard-Landeau B, Piat J, Joudain M, Fontenelle D (1994) Le dépérissement actuel du hêtre en Picardie. Rev Fr 46 – 5 547–553. https://doi.org/10.4267/2042/26582 Piovesan G, Biondi F, Filippo AD et al (2008) Drought-driven growth reduction in old beech (Fagus sylvatica L.) forests of the central Apennines, Italy. Glob Change Biol 14:1265–1281. https://doi.org/10.1111/j.1365-2486.2008.01570.x Pluess AR, Frank A, Heiri C et al (2016) Genome–environment association study suggests local adaptation to climate at the regional scale in Fagus sylvatica. New Phytol 210:589–601. https://doi.org/10.1111/nph.13809 Popa I, Badea O, Silaghi D (2017) Influence of climate on tree health evaluated by defoliation in the ICP level I network (Romania). IForest - Biogeosciences For 10:554. https://doi.org/10.3832/ifor2202-009 Pramreiter M, Grabner M (2023) The Utilization of European Beech Wood (Fagus sylvatica L.) in Europe. Forests 14:1419. https://doi.org/10.3390/f14071419 Quintana-Seguí P, Moigne PL, Durand Y et al (2008) Analysis of Near-Surface Atmospheric Variables: Validation of the SAFRAN Analysis over France. J Appl Meteorol Climatol 47:92–107. https://doi.org/10.1175/2007JAMC1636.1 R Core Team (2023) R: A Language and environment for statistical computing. [Computer software]. Retrieved from https://cran.r-project.org Rasztovits E, Berki I, Mátyás C et al (2014) The incorporation of extreme drought events improves models for beech persistence at its distribution limit. Ann Sci 71:201–210. https://doi.org/10.1007/s13595-013-0346-0 Rohner B, Kumar S, Liechti K et al (2021) Tree vitality indicators revealed a rapid response of beech forests to the 2018 drought. Ecol Indic 120:106903. https://doi.org/10.1016/j.ecolind.2020.106903 Rohner B, Waldner P, Lischke H et al (2018) Predicting individual-tree growth of central European tree species as a function of site, stand, management, nutrient, and climate effects. Eur J Res 137:29–44. https://doi.org/10.1007/s10342-017-1087-7 Roloff A (1989) Crown development and vitality assessment of selected tree species of temperate latitudes. Crown Dev Vitality Assess Sel Tree Species Temp Latit Rose L, Leuschner C, Köckemann B, Buschmann H (2009) Are marginal beech (Fagus sylvatica L.) provenances a source for drought tolerant ecotypes? Eur J Res 128:335–343. https://doi.org/10.1007/s10342-009-0268-4 Rouault G, Candau J-N, Lieutier F et al (2006) Effects of drought and heat on forest insect populations in relation to the 2003 drought in Western Europe. Ann Sci 63:613–624. https://doi.org/10.1051/forest:2006044 Rozas V, Camarero JJ, Sangüesa-Barreda G et al (2015) Summer drought and ENSO-related cloudiness distinctly drive Fagus sylvatica growth near the species rear-edge in northern Spain. Agric Meteorol 201:153–164. https://doi.org/10.1016/j.agrformet.2014.11.012 Ruffault J, Pimont F, Cochard H et al (2022) SurEau-Ecos v2.0: a trait-based plant hydraulics model for simulations of plant water status and drought-induced mortality at the ecosystem level. Geosci Model Dev 15:5593–5626. https://doi.org/10.5194/gmd-15-5593-2022 Rukh S, Sanders TGM, Krüger I et al (2023) Distinct Responses of European Beech (Fagus sylvatica L.) to Drought Intensity and Length—A Review of the Impacts of the 2003 and 2018–2019 Drought Events in Central Europe. Forests 14:248. https://doi.org/10.3390/f14020248 Saintonge FX (2023) Qualifier plus fréquemment la dégradation des forêts avec DEPERIS. Forêt Innov 5:30–33 Salvador-Blanes S, Cornu S, Couturier A et al (2006) Morphological and geochemical properties of soil accumulated in hedge-induced terraces in the Massif Central, France. Soil Tillage Res 85:62–77. https://doi.org/10.1016/j.still.2004.12.008 Sanders T, Pitman R, Broadmeadow M (2012) Soil type modifies climate-growth response of beech in Southern Britain. Proceedings of the DENDROSYMPOSIUM Scientific Technical Report STR 12/03. p 106–110 10.2312/GFZ.b103-12036 Schaeffer A (1955) Le dépérissement du hêtre dans le département du Doubs.Bulletion Société Forestière Franche-Comté XXVIII:290–291 Schmidt F, Persson A (2003) Comparison of DEM Data Capture and Topographic Wetness Indices. Precis Agric 4:179–192. https://doi.org/10.1023/A:1024509322709 Schmied G, Pretzsch H, Ambs D et al (2023) Rapid beech decline under recurrent drought stress: Individual neighborhood structure and soil properties matter. Ecol Manag 545:121305. https://doi.org/10.1016/j.foreco.2023.121305 Schmull M, Thomas F (2000) Morphological and physiological reactions of young deciduous trees (Quercus robur L., Q. petraea [Matt.] Liebl., Fagus sylvatica L.) to waterlogging. Plant Soil 225:227–242. https://doi.org/10.1023/A:1026516027096 Schneider DP, Deser C, Fasullo J, Trenberth KE (2013) Climate Data Guide Spurs Discovery and Understanding. Eos Trans Am Geophys Union 94:121–122. https://doi.org/10.1002/2013EO130001 Schuldt B, Buras A, Arend M et al (2020) A first assessment of the impact of the extreme 2018 summer drought on Central European forests. Basic Appl Ecol 45:86–103. https://doi.org/10.1016/j.baae.2020.04.003 Seidling W, Ziche D, Beck W (2012) Climate responses and interrelations of stem increment and crown transparency in Norway spruce, Scots pine, and common beech. Ecol Manag 284:196–204. https://doi.org/10.1016/j.foreco.2012.07.015 Senf C, Pflugmacher D, Zhiqiang Y et al (2018) Canopy mortality has doubled in Europe’s temperate forests over the last three decades. Nat Commun 9:4978. https://doi.org/10.1038/s41467-018-07539-6 Serra-Maluquer X, Gazol A, Sangüesa-Barreda G et al (2019) Geographically Structured Growth decline of Rear-Edge Iberian Fagus sylvatica Forests After the 1980s Shift Toward a Warmer Climate. Ecosystems 22:1325–1337. https://doi.org/10.1007/s10021-019-00339-z Silva DE (2010) Ecologie du hêtre (Fagus sylvatica L.) en marge sud-ouest de son aire de distribution. Phdthesis, Université Henri Poincaré - Nancy 1. 198 p. HAL Id: tel-01746152, version 1 Simon A, Katzensteiner K, Wallentin G (2023) The integration of hierarchical levels of scale in tree species distribution models of silver fir ( Abies alba Mill.) and European beech ( Fagus sylvatica L.) in mountain forests. Ecol Model 485:110499. https://doi.org/10.1016/j.ecolmodel.2023.110499 Sousa-Silva R, Verheyen K, Ponette Q et al (2018) Tree diversity mitigates defoliation after a drought-induced tipping point. Glob Change Biol 24:4304–4315. https://doi.org/10.1111/gcb.14326 Taccoen A, Piedallu C, Seynave I et al (2019) Background mortality drivers of European tree species: climate change matters. Proc R Soc B Biol Sci 286:20190386. https://doi.org/10.1098/rspb.2019.0386 Tenenhaus M (1998) La régression PLS: théorie et pratique. Editions TECHNIP Ter-Braak CJT, Juggins (1993) Weighted averaging partial least square regression (WA-PLS): an improved method for reconstructing environnemental variables from species assemblages. Hydrobiologia 269–270:485–502 Tesfa TK, Tarboton DG, Chandler DG, McNamara JP (2009) Modeling soil depth from topographic and land cover attributes. Water Resour Res 45. https://doi.org/10.1029/2008WR007474 Teissier du Cros E, Le Tacon F, Nepveu G et al (1981) Le hêtre. INRA Editions.613 p Tetegan M, Richer de Forges AC, Verbeque B et al (2015) The effect of soil stoniness on the estimation of water retention properties of soils: A case study from central France. CATENA 129:95–102. https://doi.org/10.1016/j.catena.2015.03.008 Thornthwaite CW, Mather JR (1957) Instructions and tables for computing potential evapotranspiration and the water balance. Climatology X:185–311 Tognetti R, Johnson JD, Michelozzi M (1995) The response of European beech (Fagus sylvatica L.) seedlings from two Italian populations to drought and recovery. Trees 9:348–354. https://doi.org/10.1007/BF00202499 Toïgo M, Nicolas M, Jonard M et al (2020) Temporal trends in tree defoliation and response to multiple biotic and abiotic stresses. Ecol Manag 477:118476. https://doi.org/10.1016/j.foreco.2020.118476 Touche J, Calvaruso C, De Donato P, Turpault MP (2022) Five successive years of rainfall exclusion induce nutritional stress in a mature beech stand. Ecol Manag 507:119987. https://doi.org/10.1016/j.foreco.2021.119987 Turc L (1955) Le bilan d’eau des sols. Relations entre les précipitations, l’évaporation et l’écoulement. Ann Agronomiques 6:5 van der Maaten E (2013) Thinning prolongs growth duration of European beech ( Fagus sylvatica L.) across a valley in southwestern Germany. Ecol Manag 306:135–141. https://doi.org/10.1016/j.foreco.2013.06.030 van Mantgem PJ, Stephenson NL, Byrne JC et al (2009) Widespread increase of tree mortality rates in the western United States. Science 323:521–524. https://doi.org/10.1126/science.1165000 Vannoppen A, Kint V, Ponette Q et al (2019) Tree species diversity impacts average radial growth of beech and oak trees in Belgium, not their long-term growth trend. Ecosyst 6:10. https://doi.org/10.1186/s40663-019-0169-z Vennetier M, Ripert C (2005) Evaluer la potentialité forestière d’un site sans observer la flore. Présentation d’une méthode développée en région méditérannéennée calclaire. RDV Tech (ONF) n° 8:15–22 Walthert L, Ganthaler A, Mayr S et al (2021) From the comfort zone to crown dieback: Sequence of physiological stress thresholds in mature European beech trees across progressive drought. Sci Total Environ 753:141792. https://doi.org/10.1016/j.scitotenv.2020.141792 Waring RH (1987) Characteristics of Trees Predisposed to Die. Bioscience 37:569–574. https://doi.org/10.2307/1310667 Weiss AD (2001) Topographic position and landforms analysis. ESRI User Conférence The Nature Conservancy (San Diego, Californie). Poster URL Httpwwwjennessentcomdownloadstpi-Poster-Tnc18x22pdf Wilkens JF, Wagner S (2021) Empirical survival model for European beech (Fagus sylvatica L.) seedlings in response to interactive resource gradients and (a-) biotic conditions within an experimental canopy gap study. Ecol Manag 499:119627. https://doi.org/10.1016/j.foreco.2021.119627 Wolfe BT, Sperry JS, Kursar TA (2016) Does leaf shedding protect stems from cavitation during seasonal droughts? A test of the hydraulic fuse hypothesis. New Phytol 212:1007–1018. https://doi.org/10.1111/nph.14087 Woodcock H, Vollenweider P, Dubs R, Hofer R-M (1995) Crown alterations induced by decline: a study of relationships between growth rate and crown morphology in beech (Fagus sylvatica L). Trees 9:279–288. https://doi.org/10.1007/BF00202018 Additional Declarations No competing interests reported. Supplementary Files SupplementaryInformation.docx Cite Share Download PDF Status: Published Journal Publication published 05 May, 2025 Read the published version in European Journal of Forest Research → Version 1 posted Editorial decision: Revision requested 12 Jan, 2025 Reviews received at journal 08 Jan, 2025 Reviewers agreed at journal 24 Dec, 2024 Reviews received at journal 18 Dec, 2024 Reviewers agreed at journal 13 Dec, 2024 Reviewers agreed at journal 10 Dec, 2024 Reviewers invited by journal 14 Nov, 2024 Editor assigned by journal 09 Nov, 2024 Submission checks completed at journal 09 Nov, 2024 First submitted to journal 08 Nov, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5417359","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":382838666,"identity":"92935c98-4f86-4cff-b34a-a61f02e50adb","order_by":0,"name":"Jean Lemaire","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIiWNgGAWjYBACxgYQWQDEzED8AYjZGxgfEKHFAKKFcQaQ5jnAbECEXVA1zDzEaGFub372gMHAJpq/nffhY5uaO3k8DMxsH/A6rOeYuQGDQVrujMPsxsY5x54VA7Uwz8CrZUaCmQSDweHchsNsbNJAMnE/A/9hvA5jnP/8G1DL/9z5h9nYf1sCtfQAbcGvZQYPyJYDuRuAtjAzEqWlJ6dMIsEgOXfjYTZmyZ5jh4t5mAloMWw/vk3iQ4Vd7rzzxxg//Kg5nMfD3kxASwOQSEASSGDAr4GBQR5dIAGLolEwCkbBKBjhAACy5EIuWnWGUgAAAABJRU5ErkJggg==","orcid":"","institution":"Aix-Marseille University","correspondingAuthor":true,"prefix":"","firstName":"Jean","middleName":"","lastName":"Lemaire","suffix":""},{"id":382838667,"identity":"63428aec-f1d5-4807-abd0-58b991ec7199","order_by":1,"name":"Michel Vennetier","email":"","orcid":"","institution":"Aix-Marseille University","correspondingAuthor":false,"prefix":"","firstName":"Michel","middleName":"","lastName":"Vennetier","suffix":""},{"id":382838668,"identity":"559d281b-594a-45c2-ba65-43b367390638","order_by":2,"name":"Bernard Prévosto","email":"","orcid":"","institution":"Aix-Marseille University","correspondingAuthor":false,"prefix":"","firstName":"Bernard","middleName":"","lastName":"Prévosto","suffix":""},{"id":382838669,"identity":"bacac958-3d4f-4648-9a1c-e8ed512c6593","order_by":3,"name":"Maxime Cailleret","email":"","orcid":"","institution":"Aix-Marseille University","correspondingAuthor":false,"prefix":"","firstName":"Maxime","middleName":"","lastName":"Cailleret","suffix":""}],"badges":[],"createdAt":"2024-11-08 14:53:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5417359/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5417359/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10342-025-01767-4","type":"published","date":"2025-05-05T15:57:07+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":70041866,"identity":"a5c291e7-d260-444f-98df-7f7ef0337459","added_by":"auto","created_at":"2024-11-27 18:15:38","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":239128,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eDistribution of the 308 plots measured in the four regions in France between 2015 and 2020 according to their geographic location (A) and to beech climate-space diagram estimated on European scale (B). \u003c/em\u003eThe presence of beech in France (A) was mapped according to the French national forest inventory (black dots; © IGN). Presence/absence, temperature and precipitation data used in the climate-space diagram (B) were extracted from Mauri et al. (2017), and from the Chelsa model (1981-2010; Schneider et al. 2013). Plots that showed signs of decline (i.e., 20% of the trees with at least 50% of crown biomass loss) are represented with red squares, healthy plots with green circles.\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/bda0c5048af308aa3a106b9b.jpg"},{"id":70041861,"identity":"ec8c0fa6-60a3-4d13-bb92-21d27ddb95d7","added_by":"auto","created_at":"2024-11-27 18:15:37","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":233611,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eReference photos used in the French DEPERIS protocol \u003c/em\u003ewhich combines volume proportion of dead branches and lost branches (DB), and the crown transparency (CT) to estimate by class (A to F) the crown biomass loss (CBL) (Neycken et al. 2024). %CBL was calculated following the formula: %CBL = %DB + (100-%DB)*%CT/100.\u003cem\u003e Photos A ©Müller and Stierlin (1990), B to E Mathieu Mirabel (DSF), F Sylvain Gaudin (CNPF)\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/222e00982dbf44239a2985f9.jpg"},{"id":70041858,"identity":"2a578335-65e0-44e2-a385-9a76536431ee","added_by":"auto","created_at":"2024-11-27 18:15:37","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":53867,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSummary of the climatic \u0026nbsp;\u0026nbsp;variables used to simulate beech decline. Four main categories of variables \u0026nbsp;\u0026nbsp;were selected (left) and combined with five time-periods (center) and four different \u0026nbsp;\u0026nbsp;expressions of climatic parameters (right). TN01\u003c/em\u003e and \u003cem\u003eTN03\u003c/em\u003e : Mean of minimal \u0026nbsp;\u0026nbsp;temperature of January and March (°C); \u003cem\u003eT_year\u003c/em\u003e: \u0026nbsp;\u0026nbsp;Mean annual temperature (°C); \u003cem\u003eTX0608\u003c/em\u003e: \u0026nbsp;\u0026nbsp;Mean of maximal temperature between June and August (°C); \u003cem\u003ePET_year\u003c/em\u003e: Annual potential \u0026nbsp;\u0026nbsp;evapotranspiration (mm); \u003cem\u003eP-ETP\u003c/em\u003e: \u0026nbsp;\u0026nbsp;Climatic Hydric deficit = Precipitation - Annual potential evapotranspiration \u0026nbsp;\u0026nbsp;over the months indicated; AWC: Available water capacity (mm); \u003cem\u003eHyd_def (mm)\u003c/em\u003e: Annual hydric deficit; \u003cem\u003eHyd_def_r (%) \u003c/em\u003e= Ratio of annual hydric \u0026nbsp;\u0026nbsp;deficit; Nbdays_HydDef: annual number of days during which the vegetation is considered as \u0026nbsp;\u0026nbsp;stressed (relative soil water content \u0026lt;= 40%); \u003cem\u003eStandardized Index\u003c/em\u003e: standardized \u0026nbsp;\u0026nbsp;deviation in comparison of average period 1981-2010; \u003cem\u003eRecurrence\u003c/em\u003e: percentage of critical years exceeding a threshold defined by the \u0026nbsp;\u0026nbsp;10th percentile in the beech natural area (%). The absolute difference, the recurrence \u0026nbsp;\u0026nbsp;and the standardized index are calculated for the periods 1, 3, 5, and 15 \u0026nbsp;\u0026nbsp;years before measurements. For more details see SI 1 and SI 3.\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/abaef180c54821af9094fc51.jpg"},{"id":70041860,"identity":"d796103b-5532-4434-8cd3-5e821cbd7b3a","added_by":"auto","created_at":"2024-11-27 18:15:37","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":42089,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePredictive accuracy (Q²) and sums of loadings of the variables (grouped into four major types of variables: CLIMATE, TOPOgraphy, SOIL, DENDROmetry) for the four regional models (HL, HFN, Jura, MV) and the national model (NAT). More details on models’ structure are available in Table 3. Regions HL = Haut Languedoc, HFN = Hauts-de-France and Normandie, Jura=Jura, MV = Millevaches, NAT=national\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/35e68079432bacd5bda8ecfc.jpg"},{"id":70041863,"identity":"76feae1c-4a78-4729-919a-17d3af67ef54","added_by":"auto","created_at":"2024-11-27 18:15:38","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":132338,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePercentage of the total loadings of the climate variables in the regional (HL, HFN, Jura, MV) and national (NAT) models. Variables are expressed as deviation from the 1981-2010 average\u003c/em\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003e\u003cem\u003e(absolute differences and standardized index combined), reccurence or absolute values. The pie charts represent the proportion of variables related to temperature and evapotranspiration (orange) or hydric deficit (blue) in each model. For the definition of the climatic variables, see Fig. 3.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/c9fb49ffa69309a34ff0388e.jpg"},{"id":70041862,"identity":"c3da5329-f0ac-4ea5-a83f-a3a70a84758b","added_by":"auto","created_at":"2024-11-27 18:15:37","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":82652,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePerformance of the different models according to the region where they have been calibrated and applied, in terms of R\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e (A) and mean error (predicted vs observed; in %); (B). The difference between the slope of the linear model (predicted vs observed) and zero is indicated by the asterisks (***p\u0026lt;0.001; **p\u0026lt;0.01; *p\u0026lt;0.05; NS p \u003c/em\u003e\u003cu\u003e\u003cem\u003e\u0026gt;\u003c/em\u003e\u003c/u\u003e\u003cem\u003e 0.05)\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/ab788a3f14761f6454614fb8.jpg"},{"id":70042393,"identity":"686a8d7e-47a6-43e9-9003-c4d44ed29f3d","added_by":"auto","created_at":"2024-11-27 18:23:38","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":107025,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePredictive accuracy of the 4 regional models and of the national model when applied in the Haut Languedoc region (HL). The mean error, R² of the linear model (predicted Xstem50 vs. observed Xstem50), and result of the statistical test if the slope of this linear model differs from zero are detailed in Fig. 6.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Picture7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/05c6ed9659e78f9b82a46177.jpg"},{"id":82537825,"identity":"0c93b5d5-2bc9-4f32-83ae-d42ee6b9bc76","added_by":"auto","created_at":"2025-05-12 16:10:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2616505,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/7a95233e-c831-41fe-a22a-2a046c037fe1.pdf"},{"id":70042392,"identity":"b9c77259-887e-4339-b882-56c952a9e6d2","added_by":"auto","created_at":"2024-11-27 18:23:37","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2243984,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-5417359/v1/33dba07504104050f6d4496c.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Multiscale Modelling of European Beech Decline: The Role of Long-Term Climate Deviations and Local Environmental Factors","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eEuropean beech (\u003cem\u003eFagus sylvatica\u003c/em\u003e L.) is the most abundant native broadleaf tree species in Europe, spanning from southern Italy to southern Norway and from western Spain to eastern Bulgaria. In addition to its large distribution area, it is also one of the most ecologically and economically important hardwood tree in Europe, playing a pivotal role in numerous forest ecosystems (Houston Durrant et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Leuschner 2020; Leuschner and Ellenberg \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Like many tree species that are experiencing rising mortality rates in many regions of the globe (e.g., Allen et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Choat et al. \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Hartmann et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Senf et al. \u003cspan citationid=\"CR119\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; van Mantgem et al. \u003cspan citationid=\"CR136\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), European beech has significantly suffered from droughts and heatwaves associated with climate change in the last decades. Decline episodes have been reported in Western Europe following severe droughts in 1976 and 1989-90 (Lies \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e1980\u003c/span\u003e; Nageleisen \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e1994\u003c/span\u003e; Peterken and Mountford \u003cspan citationid=\"CR87\" class=\"CitationRef\"\u003e1996\u003c/span\u003e), and its productivity has been declining in the southern part of its distribution range since the 2000s (Pe\u0026ntilde;uelas et al. \u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Piovesan et al. \u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Rozas et al. \u003cspan citationid=\"CR106\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Serra-Maluquer et al. \u003cspan citationid=\"CR120\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Massive die-off also occurred following the extreme droughts and heatwaves of 2003 and 2018 (Br\u0026eacute;da et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Braun et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Rukh et al. \u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), often in the core of its distribution range in France (Mirabel and Gaertner \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), Switzerland (Braun et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Frei et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and Germany (Langer and Bu\u0026szlig;kamp \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), but also in Eastern Europe such as in Croatia (Ognjenović et al. \u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) or Romania (Chira et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2006\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe drought-induced decline of European beech manifests through various symptoms in the canopy and stem. The first symptom in the case of intensive drought is the premature browning and shedding of leaves (Schuldt et al. \u003cspan citationid=\"CR117\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Arend et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Frei et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Neycken et al. \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Leaf shedding allows the tree to reduce the intensity of xylem embolism by limiting its water demand but at the cost of carbon assimilation (Br\u0026eacute;da et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Wolfe et al. \u003cspan citationid=\"CR143\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Schuldt et al. \u003cspan citationid=\"CR117\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This leaf drop may, however, not be sufficient to prevent embolism in case of extreme drought, particularly on shallow soil (Walthert et al. \u003cspan citationid=\"CR139\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Arend et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This may lead to subsequent crown-dieback, with the death of twigs and branches, and ultimately to tree mortality (Chakraborty et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Frei et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Leuschner 2020; Schuldt et al. \u003cspan citationid=\"CR117\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This was observed during summer 2018 and following years, in Central and Western Europe, particularly in France, Switzerland, and Germany. In addition to these direct impacts of extreme drought, lag effects can occur through the decrease in leaf size and shoot growth in the following years (Br\u0026eacute;da et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Nageleisen \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). This process is due to reduced carbon reserves and by the formation of fewer leaf bud primordia which would influence next year\u0026rsquo;s leaf area (Leuschner, 2020). All together, these impacts increase crown leaf deficit (Roloff \u003cspan citationid=\"CR103\" class=\"CitationRef\"\u003e1989\u003c/span\u003e; Woodcock et al. \u003cspan citationid=\"CR144\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Eichhorn et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and are often associated with a sharp reduction in tree radial growth (Braun et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Arend et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Rukh et al. \u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). For instance, the basal area increment of dominant trees in stands affected by decline, decreased by 10 to 50% of the maximum growth rate during the past 20 to 30 years (Leuschner, 2020).\u003c/p\u003e \u003cp\u003eThis decline, which manifests in various and often progressive symptoms as explained above, results from a multi-causal process (Franklin et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e1987\u003c/span\u003e; Waring \u003cspan citationid=\"CR140\" class=\"CitationRef\"\u003e1987\u003c/span\u003e). For beech, it involves numerous factors related to climate, in particular drought, but also late frosts which can damage leaves, especially when occurring in spring during the critical stage of leaf unfolding (Dittmar et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Dittmar and Elling \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). Decline can also be aggravated by extreme temperatures and intense radiation which occur during heatwaves and can impact photosynthetic activity and induce leaf photobleaching (Braun et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBesides the climatic conditions, tree health is influenced by the topographic and soil conditions, the tree and stand characteristics, and often the presence of biotic pathogens (To\u0026iuml;go et al. \u003cspan citationid=\"CR132\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). For instance, beech trees growing on shallow soils with low water-holding capacity have a greater risk of drought-induced dieback or growth reduction (Chakraborty et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Rukh et al. \u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Sanders et al. \u003cspan citationid=\"CR111\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Schmied et al. \u003cspan citationid=\"CR114\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). However, beech may also be at risk on soils with shallow and compact clay layers inducing significant hydromorphy (Cros et al. 1981; Schmull and Thomas \u003cspan citationid=\"CR115\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). The competition status and genetic information also play a role in tree vulnerability. Beech trees subjected to strong competition are more susceptible to decline (Petit-Cailleux et al. \u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Frei et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and those originating from drier habitats show signs of higher drought tolerance and/or faster recovery from drought stress than moister origins, thanks to local adaptation (Csill\u0026eacute;ry et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Leuschner, 2020; Pluess et al. \u003cspan citationid=\"CR95\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Indeed, such trees experienced a delayed onset of drought-induced reductions in stomatal conductance and photosynthesis, and reduced their fine root biomass to a lesser degree compared to trees from moister origins (Tognetti et al. \u003cspan citationid=\"CR131\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Rose et al. \u003cspan citationid=\"CR104\" class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eStand characteristics can also modulate the impacts of water stress (Br\u0026eacute;da et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Diaconu et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). On the one side, stands with high density and Leaf Area Index (LAI) are more prone to decline due to high water demand (Schmied et al. \u003cspan citationid=\"CR114\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). But on the other side, stands with a reduced LAI, e.g., after heavy thinning, can also show symptoms of decline (Mathes et al. \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) due to a less favourable microclimate at crown level (e.g. De Frenne et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Therefore, favoring a higher tree diversity through a mixture with other species that are less water-consuming, like oaks or pines, can promote the health of beech trees (Sousa-Silva et al. \u003cspan citationid=\"CR123\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Mathes et al. \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Finally, weakened beech trees are more susceptible to attacks by secondary biotic agents (e.g. \u003cem\u003eDiplodia corticola\u003c/em\u003e A.J.L. Phillips, A. Alves \u0026amp; J. Luque, \u003cem\u003eNeonectria coccinea\u003c/em\u003e (Pers.) Rossman \u0026amp; Samuels, \u003cem\u003eArmillaria mellea (\u003c/em\u003eVahl) P. Kumm., \u003cem\u003eTaphrorychus bicolor\u003c/em\u003e (Herbst, 1793)) which can intensify and accelerate the decline (Mirabel and Gaertner \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Langer and Bu\u0026szlig;kamp \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and lead to widespread mortality in subsequent years (Rouault et al. \u003cspan citationid=\"CR105\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Nageleisen and Reuter J.-C. 2007; Rasztovits et al. \u003cspan citationid=\"CR100\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eConsidering that tree decline risk is projected to increase with future climate change (e.g. Allen et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), it is crucial to better quantify the respective impacts and interactions of the factors involved in this process, and to accurately predict this risk at a spatial scale meaningful to forest managers. A wide variety of models can be used for this purpose such as (i) statistical models that estimate mortality risk based on the correlation between climatic parameters and observed mortality patterns (e.g., Taccoen et al. \u003cspan citationid=\"CR124\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), on past radial growth (e.g., H\u0026uuml;lsmann et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), or on the actual presence-absence of the species (species distribution models ; e.g., Cheaib et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2012\u003c/span\u003e); (ii) mechanistic models that explicitly simulate the physiological processes leading to mortality (e.g., Petit-Cailleux et al. \u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ruffault et al. \u003cspan citationid=\"CR107\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Models simulating beech leaf deficit are scarcer - despite its strong link with mortality risk (Dobbertin and Brang \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Dobbertin \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Petit-Cailleux et al. \u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). These models aim at explaining the spatio-temporal changes in leaf deficit using either the climatic, atmospheric, biotic, or soil information, but exclude comprehensive analyses of all four factors and their interrelationships (Seidling et al. \u003cspan citationid=\"CR118\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Popa et al. \u003cspan citationid=\"CR96\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; To\u0026iuml;go et al. \u003cspan citationid=\"CR132\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Rohner et al. \u003cspan citationid=\"CR101\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ognjenović et al. \u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In addition, they have been developed using broad-scale information (e.g., with data from the European ICP-Forests monitoring network; Ferretti, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and therefore can hardly predict the risk of decline at stand level. This is due to the significant influence factors acting at local scale on forest health, such as micro-climate, -topography, -pedology, and the stand characteristics. In consequence, it is likely that decline models established at small spatial scales are more reliable than those computed at large ones, which blur local peculiarities and compensatory effects (e.g., regional vs. national level; see Chauvier et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Simon et al. \u003cspan citationid=\"CR122\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This is particularly the case at the species distribution margins where the mitigating effect of the local soil and topography on the water budget can be very high, but neglected in large-scale models (Mellert et al. \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Similarly, the transferability of regional models across space (i.e., in other regions) is rarely assessed except in species distribution models (e.g. Su\u0026aacute;rez-Seoane et al. 2014), while there may be a trade-off between model complexity (~\u0026thinsp;performance) and transferability (Naas et al. \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo fill these gaps, our main objective was to develop and compare the predictive accuracy of statistical models of beech leaf deficit (i) in four regions of France with contrasting climatic and site conditions, which were impacted by beech decline for at least the last three decades ; and (ii) on national scale by combining the data from these four regions. We aimed at exploring the similarities and differences between the regional models, assessing their ability to predict beech decline outside their calibration range, and analysing to what extent the national model could accurately predict the decline process at a regional level.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study area and sampling design\u003c/h2\u003e \u003cp\u003eWe used 308 plots distributed across four regions of France (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) where beech is abundant and shows significant signs of decline: Hauts-de-France and Normandie (HFN), Jura, the Regional Natural Park of Millevaches (MV), and the Regional Natural Park of Haut-Languedoc (HL). In HFN and HL, previous episodes of decline have been recorded following the severe drought of 1989-90 and the 2003 heatwave (Nageleisen \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e1994\u003c/span\u003e; Nageleisen and Reuter, J.-C. 2007; Silva \u003cspan citationid=\"CR121\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). In the Jura region, important decline events occurred in the 1950s (Schaeffer \u003cspan citationid=\"CR112\" class=\"CitationRef\"\u003e1955\u003c/span\u003e) and after the 2018 drought and heatwave (Mirabel and Gaertner \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThese regions cover most of the climatic niche of beech in France: oceanic, semi-continental, mountain and Mediterranean (Joly et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), with mean annual temperature and precipitation spanning from 8.8\u0026deg;C to 13.6\u0026deg;C and 576 mm to 1697 mm, respectively (data from 1981 to 2010; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The plots were situated in the core of the climate-space diagram estimated at the European level (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB), and covered a wide range of soil types, from acidic brown soils at MV and HL (mean pH\u0026thinsp;=\u0026thinsp;4.8 for both regions) to slightly acidic brown soils or calcareous soils with active lime in HFN and Jura (mean pH\u0026thinsp;=\u0026thinsp;5.7 and 6.1, respectively). Hydromorphic layers were identified at less than one meter depth in 18 plots, only in the HFN region and were particularly marked in three plots (redoximorphic features exceeding 35%). The mean depth of soil estimated with a pickaxe and hand-dug trial was 65 cm with coarse elements representing from 0\u0026ndash;50% of soil volume. The soil texture was predominantly loamy except in the Jura. In this region, 56% of the plots had a soil texture comprising at least 30% of clay, among which 26% had a clay percentage over 50% (heavy clay soils).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eMain characteristics of the study plots by region\u003c/em\u003e Mean values are represented in bold, with the minimum and maximum within brackets. \u003cem\u003eRegions\u003c/em\u003e: HL\u0026thinsp;=\u0026thinsp;Haut Languedoc, HFN\u0026thinsp;=\u0026thinsp;Hauts-de-France and Normandie, Jura\u0026thinsp;=\u0026thinsp;Jura, MV\u0026thinsp;=\u0026thinsp;Millevaches, NAT\u0026thinsp;=\u0026thinsp;national; \u003cem\u003eXStem50\u003c/em\u003e: percentage of stems by plots with a crown biomass loss\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026ge;\u003c/span\u003e\u0026thinsp;50%; \u003cem\u003eDeclining stands\u003c/em\u003e: percentage of plots with XStem50\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026ge;\u003c/span\u003e\u0026thinsp;20%, \u003cem\u003epH\u003c/em\u003e: pH measured at 20 cm depth; \u003cem\u003eAct_lime\u003c/em\u003e : percentage of plots with active lime in soil profile; \u003cem\u003eHeavy_hydro\u003c/em\u003e: percentage of plots with heavy hydromorphy (i.e., complete discoloration of the soil matrix and reduction or rust spots\u0026thinsp;\u0026gt;\u0026thinsp;35% of soil horizon area); \u003cem\u003eHeavy_clay\u003c/em\u003e: percentage of plots with more than 50% of clay in one horizon of the soil profile; \u003cem\u003eTM_year\u003c/em\u003e and \u003cem\u003eP_year\u003c/em\u003e: mean annual temperature and sum of precipitations over the 1981\u0026ndash;2010 period; \u003cem\u003eG\u003c/em\u003e: total basal area; \u003cem\u003eG_beech\u003c/em\u003e: the proportion of beech in the total basal area; \u003cem\u003eDdom\u003c/em\u003e and \u003cem\u003eHdom\u003c/em\u003e: diameter at breast height and total height of the dominant beech trees of the plot, respectively.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"15\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c15\" colnum=\"15\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRegion\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNumber of plots\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eYears of sampling\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eXStem50 (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDeclining stands (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003eSoil\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c11\" namest=\"c10\"\u003e \u003cp\u003eClimate 1981\u0026ndash;2010\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c15\" namest=\"c12\"\u003e \u003cp\u003eStand characteristics\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003epH\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eAct_lime (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eHeavy_hydro\u003c/p\u003e \u003cp\u003e(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eHeavy_clay (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eTM_year\u003c/p\u003e \u003cp\u003e(\u0026deg;C)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eP_year\u003c/p\u003e \u003cp\u003e(mm)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eG\u003c/p\u003e \u003cp\u003e(m\u0026sup2;/ha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eG_beech\u003c/p\u003e \u003cp\u003e(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003eDdom\u003c/p\u003e \u003cp\u003e(cm)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c15\"\u003e \u003cp\u003eHdom\u003c/p\u003e \u003cp\u003e(m)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMV\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e49\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2018, 2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e9.8\u003c/b\u003e (0\u0026ndash;45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e22.4\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e4.8\u003c/b\u003e (4.6\u0026ndash;5.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003e9.98\u003c/b\u003e (8.9\u0026ndash;12.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003e1313\u003c/b\u003e (996\u0026ndash;1578)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003e29.5\u003c/b\u003e (14\u0026ndash;44)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003e70\u003c/b\u003e (28\u0026ndash;100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u003cb\u003e40\u003c/b\u003e (33\u0026ndash;60)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e\u003cb\u003e23\u003c/b\u003e (18\u0026ndash;27)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e61\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e10\u003c/b\u003e (0\u0026ndash;90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e19.7\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e4.8\u003c/b\u003e (4.5\u0026ndash;5.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003e11.18\u003c/b\u003e (8.8\u0026ndash;13.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003e1429\u003c/b\u003e (1103\u0026ndash;1697)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003e32.4\u003c/b\u003e (5\u0026ndash;55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003e87\u003c/b\u003e (20\u0026ndash;100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u003cb\u003e60\u003c/b\u003e (16\u0026ndash;75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e\u003cb\u003e25\u003c/b\u003e (11\u0026ndash;35)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHFN\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e120\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2015, 2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e10.6\u003c/b\u003e (0\u0026ndash;85)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e20.8\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e5.7\u003c/b\u003e (4.7-7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e12.5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e2.5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003e10.6\u003c/b\u003e (9.7\u0026ndash;11.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003e801\u003c/b\u003e (576\u0026ndash;1121)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003e22.5\u003c/b\u003e (10\u0026ndash;42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003e73\u003c/b\u003e (20\u0026ndash;100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u003cb\u003e68\u003c/b\u003e (26\u0026ndash;103)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e\u003cb\u003e31\u003c/b\u003e (18\u0026ndash;41)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eJura\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e78\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e29.7\u003c/b\u003e (0-100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e52.6\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e6.1\u003c/b\u003e (4.7\u0026ndash;7.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e19.2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e25.6\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003e10.34\u003c/b\u003e (9.2\u0026ndash;11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003e1183\u003c/b\u003e (987\u0026ndash;1370)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003e24\u003c/b\u003e (6.5\u0026ndash;42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003e70\u003c/b\u003e (20\u0026ndash;99)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u003cb\u003e68\u003c/b\u003e (35\u0026ndash;100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e\u003cb\u003e30\u003c/b\u003e (21\u0026ndash;37)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNAT*\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e308\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2015\u0026ndash;2020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e15.2\u003c/b\u003e (0-100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e21.2\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e5.5\u003c/b\u003e (4.5\u0026ndash;7.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e9.7\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003e0.98\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003e6.5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003e10.6\u003c/b\u003e (8.8\u0026ndash;13.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003e1104\u003c/b\u003e (576\u0026ndash;1697)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003e26\u003c/b\u003e (5\u0026ndash;55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003e75\u003c/b\u003e (25\u0026ndash;100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e\u003cb\u003e62\u003c/b\u003e (16\u0026ndash;103)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e\u003cb\u003e28\u003c/b\u003e (11\u0026ndash;41)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWithin each region, plots selection was carried out using a random stratified sampling approach. As beech is considered as a water-demanding species sensitive to water stress, the stratification was based on the climatic water deficit (P-PET from June to August 1981\u0026ndash;2010; in mm). Each region was initially divided into three climatic water deficit zones based on the mean and standard deviation of the climatic water deficit of this region. In each climatic zone, plots were then randomly selected within stands (\u0026gt;\u0026thinsp;0.5 ha) where beech occupies at least 75% of the canopy cover, as assessed by the National Forest Inventory (\u0026copy; BDFORET v2 IGN). The plots were located at the center of these stands and the minimum distance between two plots was 250 m. Plot area ranged from 800 to 3000 m\u0026sup2; (mean 2170 m\u0026sup2;).\u003c/p\u003e \u003cp\u003eTo be selected, plots had to meet the following criteria: (1) absence of signs of logging within the last six years; (2) stand height greater than 16 m or average diameter greater than 10 cm to avoid juvenile stages; (3) plot located in a wooded and accessible area with homogeneous topographic and edaphic conditions; and (4) at least 20 beech trees in the dominant or co-dominant layer.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Biotic data: health status and stand measurements\u003c/h2\u003e \u003cp\u003eOn each of our 308 plots, we recorded the health status of 20 dominant or co-dominant beech trees during the growing season (except in the Jura where only 10 trees were recorded per plot ; Mirabel and Gaertner, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) for a total of 5380 trees (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). To describe the health status of the selected trees, we used the monitoring protocol of the French Forest Health Department (Saintonge \u003cspan citationid=\"CR109\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), which is based on the European protocol of the ICP-Forests network (Eichhorn et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor the selected trees, binoculars were used to estimate crown dieback, in two opposite directions, based on the volume proportion of dead and lost branches compared to the potential full crown, and on the crown transparency of the functional crown (i.e., free of competition from neighbouring trees). Crown transparency was assessed as the percentage of total leaf and ramification loss compared to a fully foliated reference beech tree (without considering the dead branches; see Neycken et al. \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Six classes of crown biomass loss were then determined based on dead branches and crown transparency: A\u0026thinsp;=\u0026thinsp;0\u0026ndash;4.9%, B\u0026thinsp;=\u0026thinsp;5\u0026ndash;24.9%, C\u0026thinsp;=\u0026thinsp;25\u0026ndash;49.9%, D\u0026thinsp;=\u0026thinsp;50\u0026ndash;74.9%, E\u0026thinsp;=\u0026thinsp;75\u0026ndash;99.9%, and F\u0026thinsp;=\u0026thinsp;dead tree (Saintonge \u003cspan citationid=\"CR109\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). A single operator per region conducted all surveys to limit estimation biases in defoliation assessments, and training programs and intercalibration sessions were conducted to ensure the quality of assessments (e.g. Bussotti et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Eickenscheidt and Wellbrock, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Innes et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e1993\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe dominant diameter at breast height (DBH; cm) and the dominant height (Hdom; m) were calculated based on the three largest beeches in diameter, while the mean and coefficient of variation in tree diameters was based on the first 15 beeches closest to the center of the plot (CV_DBH in %; Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Stand basal area (m\u003csup\u003e2\u003c/sup\u003e/ha) was measured using a chain relascope for the different species in the stand. It is worth noting that very few or no biotic pathogen were recorded in the stands.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Soil and topographic data and indices\u003c/h2\u003e \u003cp\u003eFor each plot, a detailed survey of main soil characteristics and local topography was carried out (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eList of the main climatic, soil, topographic and dendrometric factors used to simulate beech decline\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSTAND DECLINE VARIABLE\u003c/p\u003e \u003cp\u003eto predict\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eXStem50 : Percentage of beech trees with at least 50% of crown biomass loss\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCLIMATE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e- Frost \u003csub\u003e\u003cb\u003eTYPE 1\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003cp\u003e- Heat \u003csub\u003e\u003cb\u003eTYPE 2\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e- Climatic hydric deficit \u003csub\u003e\u003cb\u003eTYPE 3\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003cp\u003e- Water balance (\u003cem\u003eincluding soil)\u003c/em\u003e \u003csub\u003e\u003cb\u003eTYPE 4\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSOIL\u003c/b\u003e \u003csub\u003e\u003cb\u003eTYPE 5\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e- Available Water Capacity\u003c/p\u003e \u003cp\u003e- pH of the upper layer\u003c/p\u003e \u003cp\u003e- Proportion of coarse elements\u003c/p\u003e \u003cp\u003e- Presence or absence of heavy clay\u003c/p\u003e \u003cp\u003e- Presence and depth of hydromorphy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e- Depth of heavy Hydromorphy\u003c/p\u003e \u003cp\u003e- Presence or absence of active lime\u003c/p\u003e \u003cp\u003e- Percentage of rock cover on the ground\u003c/p\u003e\u003cp\u003e- Presence or absence of active lime\u003c/p\u003e\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTopography\u003c/b\u003e \u003csub\u003e\u003cb\u003eTYPE 6\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e- Slope\u003c/p\u003e \u003cp\u003e- Radiation index\u003c/p\u003e \u003cp\u003e- Confinement\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e- Topographic position index within 100m and 1500 m radii\u003c/p\u003e \u003cp\u003e- Topographic wetness index\u003c/p\u003e \u003cp\u003e- General curvature\u003c/p\u003e\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDENDROMETRY\u003c/b\u003e \u003csub\u003e\u003cb\u003eTYPE 7\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-Tree diameter at breast height (DBH)\u003c/p\u003e \u003cp\u003e-Coefficient of variation of beech DBH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e- Dominant height and DBH of beech\u003c/p\u003e \u003cp\u003e- Total basal area\u003c/p\u003e \u003cp\u003e- Basal area of beech\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e\u003cem\u003eSee Fig.\u0026nbsp;3 for the climatic variables; see Supplementary Information 1 (SI 1) and 4 (SI 4) for all variables and for the definition of the \u0026lsquo;TYPES\u0026rsquo;\u003c/em\u003e\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1 Soil data\u003c/h2\u003e \u003cp\u003eSoil characteristics were estimated directly from a pickaxe and a hand-dug trial pit at the center of each plot. By horizon, soil texture was estimated with the feel method (FAO 2006), and the water content was estimated using the protocol of Baize and Jabiol (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e1995\u003c/span\u003e) and the pedotransfer functions developed by Jamagne et al. (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e1977\u003c/span\u003e). Lastly, the available water capacity (AWC; mm) was computed for each plot by multiplying the water content by soil depth considering the proportion of coarse elements. As the root system can extend deeper into the soil than that of the soil pit, we computed the ratio AWC/soil depth i.e. the available water content per centimeter of soil (AWCr; mm/cm) (Algayer et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe nature of the underlying bedrock was surveyed in the field whenever possible, or else using geological maps. The presence of active limestone was detected in the top 50 cm using 10% hydrochloric acid. The pH was measured at a depth of 20 cm using a \u0026copy;SoilStick pH meter. Intensity of hydromorphy was noted according to two classes: light (i.e. soil matrix not fully discoloured, with some rust spots or diffuse reduction\u0026thinsp;\u0026lt;\u0026thinsp;35% of soil surface) or heavy (i.e. totally discoloured matrix with rust spots or diffuse reduction\u0026thinsp;\u0026gt;\u0026thinsp;35% of soil surface). The depth of the presence of clay in high percentage (\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026ge;\u003c/span\u003e\u0026thinsp;45%) was measured. The average percentage of coarse elements was calculated over the entire soil profile. Finally, the percentage of rock cover on the ground was visually estimated. For the Jura region, pH and the presence of active limestone were not measured in the field but were estimated based on NFI data (AgroParisTech \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2008\u003c/span\u003e) and on geological maps.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.3.2 Topographic indices at plot level and landscape scale\u003c/h2\u003e \u003cp\u003eThe slope, aspect and confinement (i.e., the slope of the line connecting the studied point to the highest part of the landscape in the semicircle defined by the east-south-west direction) were measured using a clinometer and a compass at the center of the plot. Six topographic indices related to plot's water balance were computed based on digital elevation models (DEM) from the National Geographic Institute (\u0026ldquo;BD ALTI\u0026reg; | G\u0026eacute;oservices\u0026rdquo; n.d.) and using the software \u0026copy; QGIS 3.16 and R 4.2.3 (R Core Team \u003cspan citationid=\"CR99\" class=\"CitationRef\"\u003e2023\u003c/span\u003e): TWI, TPI, and curvature; each index being computed at both 25 m and 75 m resolutions (see details in SI 1). The 25 m resolution refers to plot scale (microtopography), while the 75 m resolution refers to landscape scale (macrotopography). The topographic wetness index (TWI) is an estimate of predicted water accumulation in a defined area. It is calculated as the ratio of the area upslope (i.e., from where water would flow to that point from any given point on the landscape) to the local slope at that point (Galiano et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Petroselli et al. \u003cspan citationid=\"CR89\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). A high level of TWI corresponds to a higher potential water availability. The topographic position index (TPI) is computed within a 100 m radius for microtopography (TPI100) using the 25 m DEM, and within a 1500 m radius (TPI1500) using the 75 m DEM for macrotopography (Weiss \u003cspan citationid=\"CR141\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). A higher vs lower TPI value indicates a higher vs lower position on the slope. Values close to zero correspond to flat situations or to mid-slopes. Curvature is a measure expressing the extent to which a line deviates from being straight or a surface deviates from being a plane. It is estimated as a second derivative of the surface, and reflects the shape of the slope. The higher the general curvature, the more the plot is situated on convex-shaped topography, favouring water runoff. A low value corresponds to a concave shape, indicating potential water accumulation (Blaga \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Hengl and Hannes \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFinally, the radiation index (IKR; inspired from Becker, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e1982\u003c/span\u003e) was computed as the ratio between the mean annual radiation received by the plot (calculated with the DIGITALIS model ; Piedallu and G\u0026eacute;gout, \u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) and the mean annual radiation that would receive a flat surface corresponding to the horizontal projection of the plot. An IKR\u0026thinsp;\u0026gt;\u0026thinsp;1.05 corresponds to a warm exposure, while a value below 0.95 indicates a cool situation (Vennetier and Ripert \u003cspan citationid=\"CR138\" class=\"CitationRef\"\u003e2005\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Climatic data\u003c/h2\u003e \u003cp\u003eFor each plot, the mean monthly climatic values for the 1981\u0026ndash;2010 period were estimated with the AURELHY model for precipitation, mean, minimum and maximum temperatures (B\u0026eacute;nichou and Lebreton \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e1987\u003c/span\u003e); both models integrating the effects of altitude and topography. For the 30 years preceding the year of measurement, the annual difference of each monthly climatic parameter compared with its 1981\u0026ndash;2010 mean was calculated using data from the SAFRAN model (Quintana-Segu\u0026iacute; et al. \u003cspan citationid=\"CR98\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). These inter-annual differences in each climatic parameter were then used to correct the 1981\u0026ndash;2010 mean from AURELHY and DIGITALIS, in order to downscale them at a finer resolution (method detailed in SI 2).\u003c/p\u003e \u003cp\u003eMonthly potential evapotranspiration (PET; mm) was estimated using the Turc method (Turc \u003cspan citationid=\"CR134\" class=\"CitationRef\"\u003e1955\u003c/span\u003e) to determine a monthly climatic water balance (P-PET; mm). These two climatic variables were also calculated for the summer (June - August), two different lengths of the potential growing season (April \u0026ndash; October and May to September), and the entire year. We estimated the actual evapotranspiration (AET; mm) of the stand using the Thornthwaite and Mather (\u003cspan citationid=\"CR130\" class=\"CitationRef\"\u003e1957\u003c/span\u003e) water balance method, incorporating the available soil water capacity estimated from field measurements. The calculated soil water balance allowed us to estimate the number of days during the year when the relative available soil water is lower than 35%, a threshold indicating a water stress for the vegetation (Granier et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). We calculated indices classically used in vegetation studies like the annual hydric deficit PET-AET (Hyd_def; mm) and the ratio of annual hydric deficit AET/PET (Hyd_defr; % ; see Piedallu et al. \u003cspan citationid=\"CR92\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe subsequently implemented in the models four major types of climatic variables: temperatures related to heatwaves, frost, climatic water deficit, and the water balance incorporating soil and climatic information (Fig.\u0026nbsp;3). For each climatic parameter, in addition to its (i) absolute value, we also calculated (ii) the absolute difference from the 1981\u0026ndash;2010 mean for each period, and (iii) the standardized difference index to the mean 1981\u0026ndash;2010. This index was calculated like the standardized precipitation index (SPI; McKee et al. \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e1993\u003c/span\u003e) to take into account the deviation from normality of each climatic parameter (SI 3). Finally, we also calculated (iv) the recurrence of years during which a climate threshold was exceeded. In some years (e.g., 2003 and 2018), meteorological conditions can extend far beyond the climatic limits met in average in the natural distribution area of the beech. Although climate variability may have regularly exceeded these limits in the past, it is essential to determine a critical threshold which leads to beech crown dieback. To do so, we computed the above described climatic parameters using the beech distribution area at European scale and the corresponding climatic data (SI 3). After preliminary analyses we used the 0.1 threshold, i.e. the percentile 10% for each climatic parameter calculated over the natural distribution of European beech (SI 3).\u003c/p\u003e \u003cp\u003eThe absolute values were calculated over five periods: 1, 3, 5, and 15 years preceding the measurement year, and the average 1981\u0026ndash;2010; while the absolute difference, standardized difference index, and recurrence of extreme years were calculated over 1, 3, 5, and 15 years preceding the measurement year.\u003c/p\u003e\u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.5. Statistical approach\u003c/h2\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.5.1 Variable of interest\u003c/h2\u003e \u003cp\u003eOur statistical models aimed at predicting the proportion of trees with at least 50% crown biomass loss (XStem50; %) within each plot; a threshold above which a tree is usually classified as declining (i.e., trees in stages D to F in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e; see Chakraborty et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Rohner et al. \u003cspan citationid=\"CR101\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Schmied et al. \u003cspan citationid=\"CR114\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In a previous study, conducted on \u003cem\u003ePinus sylvestris\u003c/em\u003e in southeastern France, we have shown than XStem50 was a relevant indicator to assess the risk of stand decline (Lemaire et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e2.5.2. Model development\u003c/h2\u003e \u003cp\u003eWe used partial least squares (PLS) regression (Ter-Braak and Juggins \u003cspan citationid=\"CR126\" class=\"CitationRef\"\u003e1993\u003c/span\u003e) to model beech decline (XStem50) because the number of variables used was large (229 variables, see SI 1). The PLS approach is known to be effective in the case of complex interacting systems (Fernandes \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) with a high number of correlated variables and a limited number of observations (Cramer III et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; Tenenhaus \u003cspan citationid=\"CR125\" class=\"CitationRef\"\u003e1998\u003c/span\u003e). In particular, some climatic and topographic variables based on digital elevation models are highly correlated.\u003c/p\u003e \u003cp\u003eWe tested a large number of variables with for some of them a high level of correlation, leading to redundancies. To mitigate this effect before computing the PLS regression, we built a dendrogram for each of the 7 types of variables (e.g. 4 types for climate, 1 for topography, 1 for soil and 1 for dendrometry; see Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Variables were grouped by clusters within the dendrogram based on their distance (\u003cem\u003edi\u003c/em\u003e) with \u003cem\u003edi\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1-│\u003cem\u003eCorr yz\u003c/em\u003e│, │\u003cem\u003eCorr yz\u003c/em\u003e│ being the absolute value of the correlation coefficient between variables \u003cem\u003ey\u003c/em\u003e and \u003cem\u003ez\u003c/em\u003e. Different clusters were determined using a \u003cem\u003edi\u003c/em\u003e threshold of 0.15, after testing 10 \u003cem\u003edi\u003c/em\u003e thresholds from 0.05 to 0.30 (see SI 4). Within each cluster, the variable with the highest absolute partial standardized coefficient was selected to be tested in the PLS regression. Then, a stepwise method based on the Q\u0026sup2; coefficient of Stone-Geisser was used to select the number of components in the PLS regression (SI 4). In the first step of the PLS regression, non-significant variables (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05) were removed altogether (Tenenhaus \u003cspan citationid=\"CR125\" class=\"CitationRef\"\u003e1998\u003c/span\u003e). In the following steps, among the significant variables, the variable with the lowest standardized coefficient was removed if this led to an increase in Q\u0026sup2;. The stepwise process continued until the removal of the last variable at stake caused a decrease in Q\u003csup\u003e2\u003c/sup\u003e. The PLS regressions were run with the plsRglm package in R (Bastien et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Bertrand and Maumy-Bertrand \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e2.5.3 Predictive accuracy of the regional and national models\u003c/h2\u003e \u003cp\u003eA model was developed using this stepwise method for each region separately and considering all regions together (thereafter named \u0026ldquo;NAT\u0026rdquo; model for \u0026lsquo;national\u0026rsquo;). To build the NAT model, we randomly resampled the dataset as the number of plots and the mean XStem50 differed among regions (53% of the stands were considered as declining in Jura, while it ranged from 19.7 to 22.4% in the three other regions; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). In each region, we selected 10 declining plots and 39 healthy plots to fit the region with the smallest sample size (49 plots in MV; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and the average rate of declining plots in the HFN, MV, and HL regions (21%). We repeated five times this random resampling with replacement, to finally obtain 50 declining plots and 185 healthy ones in each region, for a total of 740 healthy plots and 200 declining plots to build the NAT model.\u003c/p\u003e \u003cp\u003eTo validate each model and assess their transferability, we compared their ability to predict XStem50 in the region where they have been calibrated, in the three regions outside of their calibration zone, and at the national level. We also tested the national model in each region. The predictive accuracy of the PLS regressions was estimated using the Q\u003csup\u003e2\u003c/sup\u003e coefficient of Stone-Geisser, and the relative importance of each significant variable selected in the models was reflected through the calculation of the weighted loadings (%) (Bastien et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Bertrand and Maumy-Bertrand \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). To evaluate and compare the performance of the different models in and outside their calibration regions, we also calculated the mean error between the Xstem50 observed in the field and that predicted by the model, and the R\u0026sup2; of the linear model fitted on both values. We finally determined if the slope of this regression significantly differed from zero with a t-test. For these calculations, any negative predicted Xstem50 values were adjusted to zero.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Model structure and predictive ability in their calibration region\u003c/h2\u003e \u003cp\u003eThe models of the four regions showed high differences in the accuracy of their predictions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e), with lower Q\u0026sup2; values for Jura (0.26) and Millevaches (MV; 0.27) than for Hauts-de-France/Normandie (HFN; 0.34) and Haut-Languedoc (HL; 0.42). The national model displayed an intermediate value (NAT; 0.31).\u003c/p\u003e \u003cp\u003eThe cumulative loadings grouped by four major types of variables (i.e. climate, soil, topographic and dendrometric variables; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e) indicated a dominant weight of the climatic parameters in most models: from 47 to 94% in HL, HFN, MV and NAT models. For the Jura region, the three other major types were predominant (cumulative loadings\u0026thinsp;=\u0026thinsp;63%) but none of these types reached individually a high percentage of loading, neither in any region nor at national scale.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMore specifically, among the climatic variables, only those related to higher temperatures and water deficit cumulated over 1 to 15 years before measurements were significant, increasing XStem50. No significant variables were identified among the frost parameters and the 30-years climatic average (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). In all models, deviations from the 30-years mean and recurrence of critical years exceeding a threshold defined by the percentile 10% in the beech natural area have the highest loadings, except at MV (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cem\u003eLoadings (%) of the significant variables selected in each regional model and the national model. The sign (+\u0026thinsp;or -) in parentheses indicates a positive or a negative effect on\u003c/em\u003e XStem50, \u003cem\u003erespectively\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHL\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHFN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eJura\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNAT\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"8\" nameend=\"c8\" namest=\"c1\"\u003e \u003cp\u003eCLIMATE\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"8\" nameend=\"c8\" namest=\"c1\"\u003e \u003cp\u003eTEMPERATURE and evapotranspiration\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eVariables\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003ePeriod\u003c/b\u003e\u003c/p\u003e \u003cp\u003eYears before sampling\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eAbsolute value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eTM_year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e30.5 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e8.3 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eTX0608\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e19.9 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRecurrence\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eTX0608\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e\u003cb\u003eAbsolute difference compared to climate norms 1981\u0026ndash;2010\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003ePET_year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14.9 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.8 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eTM_year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e5\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4.9 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e9.5 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003ePET_year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.2 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eStandardized Index\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eTM_Year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.6 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c8\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003ehydric Deficit\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eVariables\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003ePeriod\u003c/b\u003e\u003c/p\u003e \u003cp\u003eYears before measures\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e\u003cb\u003eAbsolute value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eHyd_Defr\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eHyd_Def\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14.3 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eHyd_Def\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.2 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eHyd_Def_r\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4.4 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e \u003cp\u003e\u003cb\u003eRecurrence\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eNbdays_HydDef\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9.6 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eP-PET0410\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e17.2 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eP-PET0509\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14.2 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eP-PET0509\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eP-PET0608\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14.3 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eP-PET_year\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e7.7 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAbsolute difference compared to climate norms 1981\u0026ndash;2010\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eP-PET0509\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e17.3 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.1 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e\u003cb\u003eStandardized Index\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eNbdays_HydDef\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11.9 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eHyd_Def\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.7 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e15\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.1 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c8\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTOPOGRAPHY\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCurv_75m\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e9 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.7 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eIKR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5.5 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSlope\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.8 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4.6 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTWI_25m\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.8 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4.1 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTWI_75m\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.1 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e11.4 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e21.8 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c8\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSOIL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003epH\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.5 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e6.1 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHydro_depth\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4.6 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClay_depth\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e17.7 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.5 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAWCr\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.5 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.8 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAWC\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.5 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e19.8 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.7 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.5 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRock_Exp\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.1 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"8\" nameend=\"c8\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDENDROMETRY\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eG\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.3 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e26.4 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e26.6 (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e13.2 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003e%G_beech\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11.1 (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e6.6 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHdom\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.7 (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCV_DBH\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.1 (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWith \u003cem\u003eTM_year\u003c/em\u003e (\u0026deg;C): Mean annual temperature. \u003cem\u003eTX0608\u003c/em\u003e (\u0026deg;C): Mean of the maximum temperature of the 1981\u0026ndash;2010 period from June to August. \u003cem\u003ePET_year\u003c/em\u003e (mm): Annual potential evapotranspiration. \u003cem\u003eHyd_Def\u003c/em\u003e (mm): Annual hydric deficit. \u003cem\u003eHyd_Def_r\u003c/em\u003e (%): Ratio of annual hydric deficit. \u003cem\u003eNbdays_HydDef\u003c/em\u003e: number of days during which the relative extractable soil water is below 40%. \u003cem\u003eP-PET0410, P-PET0509, P-PET0608\u003c/em\u003e (mm): Climatic water balance from April to October, May to September, and June to August, respectively. \u003cem\u003eP-PET_year\u003c/em\u003e (mm): Annual climatic water balance. \u003cem\u003eCurv_75m\u003c/em\u003e: General curvature calculated using the 75m DEM. \u003cem\u003eIKR\u003c/em\u003e: Radiation index. \u003cem\u003eSlope\u003c/em\u003e (%): Slope measured at the center of the plot. \u003cem\u003eTWI_75m\u003c/em\u003e and \u003cem\u003eTWI_25m\u003c/em\u003e: Topographic wetness indices calculated using the 75m and 25 m DEM resolutions, respectively. \u003cem\u003eHydro_depth\u003c/em\u003e (cm): Depth of light or heavy hydromorphy presence. \u003cem\u003eClay_depth\u003c/em\u003e (cm): Depth of heavy clay presence. \u003cem\u003eAWC\u003c/em\u003e (mm): Available water capacity. \u003cem\u003eAWCr\u003c/em\u003e (mm/cm): Relative available water capacity per centimeter of soil (AWC/depth of profile). \u003cem\u003eRock_Exp\u003c/em\u003e (%): Percentage of rock exposure on the ground. \u003cem\u003eG\u003c/em\u003e (m\u0026sup2;/ha): Total basal area. %\u003cem\u003eG_beech\u003c/em\u003e (%): Beech proportion in the total basal area. \u003cem\u003eHdom\u003c/em\u003e (m): Dominant height of beech. \u003cem\u003eCV_DBH\u003c/em\u003e (%): Coefficient of variation of diameter at breast height\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eXStem50 was lower on soils with high available water capacity in all regions (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). However, the effect of heavy clay, with high water-holding capacity, had opposite effects: detrimental in the Jura and at national scale where it increased XStem50, but favourable in HFN. In HL, the soils with higher pH and lower water content were more prone to decline compared to the deeper or more acidic soils. The national model combines all of these soil factors with the same sense of correlation but with different loadings.\u003c/p\u003e \u003cp\u003ePlots located in topographic situations unfavourable for water balance showed stronger tree decline. In all regions, the Topographic Wetness Index (TWI) was negatively correlated with XStem50, but the resolution of the concerned DEM varied by region. More specifically, plots situated on convex topography in the Jura and on the national scale, those on south-facing slopes in the Jura, and on steeper slopes in HL and on the national level exhibited a higher XStem50. TWI at75 m was the only significant topographic factor in HFN et MV.\u003c/p\u003e \u003cp\u003eConcerning the dendrometric characteristics, stands with a high total basal area were less prone to decline in all regions except in HFN; and this effect explained about 26% of the variance in Xstem50 in the Jura and MV regions. In addition to stand density, the decline was stronger in stands dominated by beech (high proportion of beech in basal area) in HL and on national scale, and with more heterogeneous diameter distributions in HFN and on national scale (but with loadings\u0026thinsp;\u0026lt;\u0026thinsp;3%).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Accuracy of the regional and national models outside their calibration range\u003c/h2\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec16\" class=\"Section3\"\u003e \u003ch2\u003e3.2.1. Regional level\u003c/h2\u003e \u003cp\u003eEach regional model was by far and logically the best in its calibration region (Fig.\u0026nbsp;6). In this case, the slope of the linear model (predicted \u003cem\u003evs\u003c/em\u003e observed) is highly significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and the mean error is the lowest. Other models can provide inaccurate predictions of plot decline with a mean error ranging from \u0026minus;\u0026thinsp;28% to +\u0026thinsp;21% on regional scale. For Jura, only the local and national models were significant (Fig.\u0026nbsp;6A). The MV model was the least accurate when applied to regions outside its calibration zone, exhibiting a systematic underestimation, while HFN model overestimated Xstem50 outside of its calibration region.\u003c/p\u003e \u003cp\u003eIn all regions, the models systematically underestimated XStem50 when it was high and overestimated it when it was close to zero (see example in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section3\"\u003e \u003ch2\u003e3.2.2 National level\u003c/h2\u003e \u003cp\u003eThe national model consistently performed better in each region that non-local regional models on the basis of the R\u0026sup2; and mean error criteria (Fig.\u0026nbsp;6). On average across all regions, the national model could explain 32% of the total variance, whereas the regional models accounted for 46%. In HL, the national model explained 60% of the total variance, very close to the regional model (62%). In contrast, the variance explained by the national model was only half of that of the regional model for MV (22% vs 44%).\u003c/p\u003e \u003cp\u003eConversely, the predictive power of each regional model applied on national scale was very low, averaging 11% of the total variance, and represented only 5 to 18% of the variance explained by the national model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section3\"\u003e \u003ch2\u003e3.2.3. The example of the Haut-Languedoc region.\u003c/h2\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eUsing the Haut-Languedoc (HL) region as example to compare the predictive accuracy of the 5 models, the HL model was logically the best when applied there, followed by the national model (Figs.\u0026nbsp;6 and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e7\u003c/span\u003e). Conversely, the MV and Jura models strongly underestimated the observed decline, especially for the highest XStem50 values. Indeed, the Jura model was strongly influenced by soil, topographic and dendrometric parameters and far less by climate. In this region, the environmental and silvicultural conditions strongly differ from Haut Languedoc: the beech forests were (i) submitted to a stronger warming in the 15 years preceding the measurement (+\u0026thinsp;0.58\u0026deg;C in the Jura compared to HL; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001); (ii) more often located on north-facing slopes (mean IKR\u0026thinsp;=\u0026thinsp;1.0 in Jura and 0.95 in HL; p\u0026thinsp;=\u0026thinsp;0.008); (iii) located on soils with higher available water capacity (138 mm vs. 90 mm; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001); (iv) and less dense than in HL (G\u0026thinsp;=\u0026thinsp;24.1 m\u0026sup2;/ha vs. 32.3 m\u0026sup2;/ha; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eEven though mean water deficit and its inter-annual variability estimated over the previous 3 and 15 years (significant periods in the HFN model ; see Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) were significantly higher in HL than in HFN (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), XStem50 in HL was consistently overestimated by the HFN model (+\u0026thinsp;21%). The climatic parameters recorded in these two regions were not significantly different in the year prior to measurement. Furthermore, there were no significant differences in soil and dendrometric parameters between the two regions.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study, based on multivariate statistical models and conducted on European beech, confirms its high sensitivity to high temperature and drought. Soil characteristics, notably the available water capacity, and topography are important factors mitigating the risk of beech decline. However, there are also other factors that interact to either mitigate or aggravate the decline, such as dendrometric parameters associated with stand structure and composition. Each regional model was by far the most precise in its calibration area, but the larger-scale national model seems to be an acceptable compromise for estimating the risk of decline across all the study regions.\u003c/p\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e4.1. Performance and structure of the models in their calibration region\u003c/h2\u003e \u003cp\u003eThe different statistical models developed to predict beech decline showed variable performances according to the study region. The best model, i.e. with the highest Q\u0026sup2; value (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e), was calibrated in the HL region. In this region submitted to a Mediterranean climate, where beech is at the edge of its distribution range, climate loadings are high indicating that the climatic factors are the main drivers of beech decline, which started in the early 2000s (Silva \u003cspan citationid=\"CR121\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Cavin and Jump \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In Jura, forest health was worse than in the other regions (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), mainly due to the 2018 drought, which had a significant impact on forest health in Central and Northern Europe (Buras et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Braun et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Rukh et al. \u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Sampling in this mountainous area was limited to low-altitude (\u0026lt;\u0026thinsp;800 m) beech forests (Mirabel and Gaertner \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) where decline was important, and where climatic conditions were systematically unfavourable for beech in 2018. This can probably explain why climate variables, which were quite homogeneous within plots, had low loadings, leading to the lowest predictive accuracy among the five models. But it also clearly highlights the compensating role of the three other groups of factors (topography, soil and dendrometry), which have in Jura their highest loadings. In MV, the decline was less prevalent. At the time of measurement, i.e. the beginning of 2018, beech decline was not intense in this region (Mirabel et al. \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and consequently difficult to predict. In HFN where the decline of beech started in the early 1990s (Nageleisen \u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e1993\u003c/span\u003e, Pilard-Landeau et al. \u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e1994\u003c/span\u003e), the regional model had an intermediate predictive accuracy, climatic variables having by far the highest weight. The intensive management that occurred in this region likely decreased the model performance. The different models thus include variables of different nature and with contrasting loadings, which we discuss below.\u003c/p\u003e \u003cdiv id=\"Sec21\" class=\"Section3\"\u003e \u003ch2\u003e4.1.1.Influence of climate factors\u003c/h2\u003e \u003cp\u003eAmong the variables that best explained the variability in beech decline within each region and also among regions, the climatic parameters linked to water stress and heat were of primary importance (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e). This aligns with a vast and consistent literature on beech sensitivity to water deficit and heat, especially its susceptibility to xylem cavitation (Br\u0026eacute;da et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Braun et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Walthert et al. \u003cspan citationid=\"CR139\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe did not find any effect of the 30-years climatic average, which was expected as the plots were mainly located in climatic zones favourable to beech (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Conversely, the interannual climate variability explained most of the beech decline, particularly exceptionally intense drought and high temperatures, or repeated droughts (see factors of recurrence, difference and standardized indices in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e). This result is consistent with the observation that extensive decline events occurred during periods of elevated temperatures and low precipitation, such as in 2003 and 2018, even in beech core range (Leuschner, 2020). In addition, we observed long-term effects of climate: beech forests that were submitted to a stronger warming, a higher mean hydric deficit, or recurrence of extreme droughts over the last 15 years, showed the highest Xstem50. Such lag-effects have been discussed in numerous studies, and especially highlighted by dendrochronological analyses. For instance, Neycken et al. (\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) found in Switzerland that beech trees that exhibited severe crown dieback after a year of high hydric stress showed a stronger growth decline than healthy trees over the last 50 years. Drought-induced changes in growth trajectories may be visible before the signs of crown dieback (Neycken et al. \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section3\"\u003e \u003ch2\u003e4.1.2 Influence of topography and soil factors\u003c/h2\u003e \u003cp\u003eThe different models demonstrated the importance of soil and topographic factors in either compensating or exacerbating the impact of climatic events on beech health (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e; Fabiani et al. 2024), mainly because of differences in geological, pedological and topographic conditions among regions. In HDF, beech trees are found in flat areas, while in Jura, they are present both in the plains and at higher elevations in the mountains. At medium and low elevation in Jura, this species can be found on both south- and north-facing slopes. In HL, with a Mediterranean climate, beech trees are more commonly found at high elevations and are primarily located on north-facing slopes (IKR is significantly lower in HL than in Jura).\u003c/p\u003e \u003cp\u003eThe Topographic Wetness Index (TWI) had a significant effect in all models, indicating that the risk of decline is higher in stands located on convex landform (high curvature) and steep slopes. It is also higher with a southern exposure (Chaplot and Walter \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Schmidt and Persson \u003cspan citationid=\"CR113\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). These topographic variables are closely related to higher water runoff and lower soil depth (Salvador-Blanes et al. \u003cspan citationid=\"CR110\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Tesfa et al. \u003cspan citationid=\"CR127\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Chartin et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), i.e. to a low available water capacity (AWC). AWC is also significantly related to XStem50 in all regions: it is known to play a crucial role in maintaining beech vitality in temperate macroclimatic conditions (Chakraborty et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In regions with high soil stoniness and compact texture, the relative AWC (AWCr, i.e. the available water content per centimeter of soil) seems to be a better estimator of water availability to roots than the absolute AWC. In these types of soils, it is impossible to assess the depth of soil explored by roots with an auger and pickaxe, and the available water capacity is often underestimated (Arrouays et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Tetegan et al. \u003cspan citationid=\"CR129\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe Jura model differed from the other models because soil and topographic variables had high loadings. In Jura, plots located on marls containing more than 50% of clay were highly declining. Such soils are characterized by high compactness, and often a low AWC which are unfavourable to beech (Čerm\u0026aacute;k et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e1993\u003c/span\u003e; Obladen et al. \u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Conversely, the presence of clay limited tree declines in HFN. In this region, clay content never exceeds 50% and is often mixed with coarser elements, resulting in less compacted soils.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003e4.1.3 Influence of stand structure and composition\u003c/h2\u003e \u003cp\u003eWe found that stands with a high total basal area were less prone to decline in HL, Jura, MV; an observation also made at the national level (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). This result was not expected and contradicts those of the systematic ICP-Forests network, where stands with high basal area presented a higher level of defoliation (To\u0026iuml;go et al. \u003cspan citationid=\"CR132\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Three complementary hypotheses may explain this difference. First, the ICPF network has a long-term monitoring history, while there was only one measurement per plot and no knowledge of past silviculture in our study. The time elapsed between our measurement and the first signs of decline varied across the regions, and could be long (e.g. first reports in HDF in the 1990s while our measurements were conducted in 2015\u0026ndash;2016). In the meantime, the probability of sanitary thinning was high, especially because beech is a high-value hardwood (Armand \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) with low wood durability (Pramreiter and Grabner \u003cspan citationid=\"CR97\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and is often promptly harvested when stands show signs of decline (Brunier et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). It is therefore likely that such thinnings were carried out by selectively harvesting the most declining or dead trees, leading to an underestimation of XStem50 (Lech and Kamińska \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). We tried to avoid this bias by excluding plots that have been thinned in the last six years, but this underestimation is likely for stands with long-term decline history and in regions where harvesting is frequent and rather based on tree health than on tree status in stand structure.\u003c/p\u003e \u003cp\u003eThe second hypothesis is related to site fertility. At the same level of thinning intensity, stand basal area is generally higher in sites that are more favourable for beech (e.g. soil with higher AWC, located on northern aspect( Rohner et al. \u003cspan citationid=\"CR102\" class=\"CitationRef\"\u003e2018\u003c/span\u003e)), which can explain the negative relationship between Xstem50 and stand basal area.\u003c/p\u003e \u003cp\u003eThe third hypothesis is related to the intensity of thinning, which can lead to contrasted impacts of tree health. If too intense, thinning can increase the water demand at crown level, particularly for large trees, leading to a higher risk of xylem cavitation and mortality (McDowell and Allen \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). In contrast, a moderate thinning can be beneficial by reducing the water uptake at stand level, and thus by boosting growth of the remaining trees (van der Maaten \u003cspan citationid=\"CR135\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Diaconu et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrevious studies have shown that beech sensitivity to drought is lower in mixture than in monospecific stands, e.g., when growing with \u003cem\u003ePinus sylvestris\u003c/em\u003e (Metz et al. \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) or other hardwood species (M\u0026ouml;lder and Leuschner \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Vannoppen et al. \u003cspan citationid=\"CR137\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Indeed, in HL, mixed beech stands were less affected by decline (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). The percentage of beech in terms of basal area was higher in HL (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), but with a broader range. In the more mixed stands of this region, this mixture may have a favorable effect on the health of beech, particularly in an area influenced by a Mediterranean climate.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Assessing the potential extrapolation of the regional and national models\u003c/h2\u003e \u003cp\u003eIn each region, the best model for predicting XStem50 is logically the local model, and the second-best model is the national one (Fig.\u0026nbsp;6). The prediction of XStem50 by regional models outside their calibration area is less reliable and sometimes completely inaccurate, confirming that increasing the complexity and precision of models can be at the expense of their transferability (Naas et al. \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; but see also Pichler and Hartig \u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This was expected as the selected variables and parameter estimates (~\u0026thinsp;loadings) of the regional models are closely adjusted to the local climatic, edaphic, topographic, and dendrometric factors. In contrast, the national model integrates and balances the weight of these environmental factors at a larger scale, keeping only the most important and common ones.\u003c/p\u003e \u003cp\u003eFirst, the rate of decline differs among regions, leading to systematic under- or over-estimation in the model predictions when applied in another region. For instance, in MV, declining trees at the time of measurements were quite rare, with a maximum XStem50 of 45% (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Thus, the MV model could hardly predict decline levels as high as those observed in HL, where the maximum XStem50 was 90%. Second, the periods when decline began and dates of measurements differed among regions. Therefore, the periods used to compute the climatic parameters also differ among models. If measurements are delayed many years after the climatic events that trigger the decline process, trees can recover if the climatic conditions become favourable again, and there is a loss of correlation between the climatic signal and XStem50. Third, decline is a complex phenomenon influenced by interactions between climate, soil conditions, stand characteristics, topographic factors, past management as well as presence of biotic agents (Galiano et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Camarero et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Chakraborty et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Lemaire et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). For instance, the Jura model was mainly influenced by soil, topography, and dendrometric factors (cumulative loadings\u0026thinsp;=\u0026thinsp;90%) and far less by climate, unlike other regions. It strongly underestimated the decline in HL were the soils are sandy-loamy without clay or hydromorphy which both played a key role in increasing XStem50 in Jura. In HL, beech stands were generally located on north-facing slopes with a low IKR, limiting high temperatures and drought in this sub-Mediterranean area, compared to Jura with its colder climate. In the same way, by testing the HFN model on HL data in the south of France, predictions of XStem50 by the HFN model were systematically overestimated by 21% compared to the predictions of the HL model (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e7\u003c/span\u003e). Conversely, when the HL model was applied to HFN data in northern France, the model predicted a slight decline and underestimated XStem50. The climatic variables related to water deficit and temperature were predominant in these two models (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e). XStem50 measured in HFN was very likely underestimated compared to the actual decline. Indeed, the beech forests in HFN had a higher economic value than those in HL, leading to more intensive timber exploitation in this region (IGN \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The time period over which climatic variables were considered in both models differed. It was based on the year for HL and on the summer or vegetation period for HFN. The vegetation period of beech trees indeed varied with longitude, latitude, and altitude, and was longer in the warmer plots of HL (Lebourgeois et al. \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2002\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFinally, there were factors not considered in our models that might explain the differences in predicted XStem50, such as phenotypic and genetic adaptations, which could have played an essential role in beech tree resistance to water stress (Leuschner, 2020). It is also possible that beech trees in the four regions exhibited different morphological and physiological adaptations to their own climates, as suggested by Knutzen et al. (2015). For instance, the root system of southern beech forests might be deeper with notable changes in the depth extension of some root diameter classes (Meier et al. 2018), increasing their drought resistance (Bolte et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). All these reasons could explain why it is difficult to apply a statistical model of tree decline outside of its calibration area.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eOur main objective was to model decline intensity \u0026ndash; measured by crown biomass loss (XStem50) - of beech forests in four regions of France with contrasted environmental conditions. Using PLS regression techniques, our statistical models predicted XStem50 from a large set of factors with a relatively good prediction (Q\u0026sup2; varied from 0.26 to 0.42). Climatic factors were of primary importance. Among them, deviations from the 1981\u0026ndash;2010 average related to high temperature and water deficit played the most important role. This illustrates the fact that beech forests are locally adapted to the regional climate in which they grow, but are sensitive to deviations from the \u0026ldquo;average\u0026rdquo; regional past climate. With climate change, these deviations are projected to be more intense and more frequent threatening even more the health condition of these forests in the near future. However, we also found that soil and topographic factors with a strong influence on the local water budget could mitigate or aggravate the decline, with specificities among regions according to their local climate and geology, and main stand characteristics. Such a modelling approach can be applied to other European regions but can be improved by including information on past stand dynamics and management, and variation in vulnerability among populations. However, they remain a valuable tool for forest managers to predict which beech forests will be at risk of decline in future climatic conditions and to implement preventive silvicultural actions adapted to local site conditions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eJ LEMAIRE : Conceptualization text, figures, Formal Analysis, Writing : draft, review and editing. M VENNETIER et B PREVOSTO : Writing : review and editingM CAILLERET : Formal analysis, writing : review and editing\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAGROPARISTECH IFN (2008) Carte du pH de surface des sols forestiers fran\u0026ccedil;ais. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://silvae.agroparistech.fr/public/documents/carte_PH_version1-formatA3.pdf\u003c/span\u003e\u003cspan address=\"https://silvae.agroparistech.fr/public/documents/carte_PH_version1-formatA3.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 29 Jan 2024\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlgayer B, Lagacherie P, Lemaire J (2020) Adapting the available water capacity indicator to forest soils: An example from the Haut-Languedoc (France). Geoderma 357:113962. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.geoderma.2019.113962\u003c/span\u003e\u003cspan address=\"10.1016/j.geoderma.2019.113962\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllen CD, Breshears DD, McDowell NG (2015) On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene. Ecosphere 6:art129. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1890/ES15-00203.1\u003c/span\u003e\u003cspan address=\"10.1890/ES15-00203.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllen CD, Macalady AK, Chenchouni H et al (2010) A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. Ecol Manag 259:660\u0026ndash;684. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2009.09.001\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2009.09.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArend M, Link RM, Zahnd C et al (2022) Lack of hydraulic recovery as a cause of post-drought foliage reduction and canopy decline in European beech. New Phytol 234:1195\u0026ndash;1205. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.18065\u003c/span\u003e\u003cspan address=\"10.1111/nph.18065\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArmand G (2002) Le h\u0026ecirc;tre autrement. Institut pour le D\u0026eacute;veloppement Forestier.262 p\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArrouays D, McKenzie N, Hempel J et al (2014) GlobalSoilMap: Basis of the global spatial soil information system. CRC. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1201/b16500\u003c/span\u003e\u003cspan address=\"10.1201/b16500\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaize D, Jabiol B (1995) Guide pour la description des sols. Editions Quae.430 p\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBastien P, Vinzi VE, Tenenhaus M (2005) PLS generalised linear regression. Comput Stat Data Anal 48:17\u0026ndash;46. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.csda.2004.02.005\u003c/span\u003e\u003cspan address=\"10.1016/j.csda.2004.02.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBD ALTI\u0026reg; | G\u0026eacute;oservices \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://geoservices.ign.fr/bdalti\u003c/span\u003e\u003cspan address=\"https://geoservices.ign.fr/bdalti\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed 6 Sep 2021\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecker M (1982) Influence relative du climat et du sol sur les potentialit\u0026eacute;s foresti\u0026egrave;res en moyenne montagne. Exemple des sapini\u0026egrave;res \u0026agrave; f\u0026eacute;tuque (Festuca silvatica Vill.) dans les Vosges alsaciennes. Ann Sci For 39:1\u0026ndash;32. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1051/forest:19820101\u003c/span\u003e\u003cspan address=\"10.1051/forest:19820101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB\u0026eacute;nichou P, Lebreton O (1987) Prise en compte de la topographie pour la cartographie des champs pluviom\u0026eacute;triques statistiques. La m\u0026eacute;t\u0026eacute;orologie 7\u0026egrave; s\u0026eacute;rie:23\u0026ndash;34\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBertrand F, Maumy-Bertrand M (2018) plsRglm: Partial least squares linear and generalized linear regression for processing incomplete datasets by cross-validation and bootstrap techniques with R. ArXiv181001005 Stat\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlaga L (2012) Aspects regarding the signifiance of the curvature types and values in the studies of geomorphometry assisted by GIS. Analele Univ Din Oradea Ser Geogr Year XXII no. 2/2012:327\u0026ndash;337\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBolte A, Czajkowski T, Cocozza C et al (2016) Desiccation and Mortality Dynamics in Seedlings of Different European Beech (Fagus sylvatica L.) Populations under Extreme Drought Conditions. Front Plant Sci 7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpls.2016.00751\u003c/span\u003e\u003cspan address=\"10.3389/fpls.2016.00751\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun S, Hopf S-E, Tresch S et al (2021) 37 Years of Forest Monitoring in Switzerland: Drought Effects on Fagus sylvatica. Glob Chang Oct. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/ffgc.2021.765782\u003c/span\u003e\u003cspan address=\"10.3389/ffgc.2021.765782\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 2021 4:\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBr\u0026eacute;da N, Huc R, Granier A, Dreyer E (2006) Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences. Ann Sci 63:625\u0026ndash;644. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1051/forest:2006042\u003c/span\u003e\u003cspan address=\"10.1051/forest:2006042\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrunier L, Delport F, Gauquelin X (2020) Guide de gestion des crises sanitaires en for\u0026ecirc;t., CNPF-IDF.184 p\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBuras A, Rammig A, Zang CS (2020) Quantifying impacts of the 2018 drought on European ecosystems in comparison to 2003. Biogeosciences 17:1655\u0026ndash;1672. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/bg-17-1655-2020\u003c/span\u003e\u003cspan address=\"10.5194/bg-17-1655-2020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBussotti F, Cozzi A, Cenni E et al (2009) Quality Assurance and measurement errors in monitoring tree crown conditions in Italy. J Environ Monit 11:769\u0026ndash;773. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1039/B818166G\u003c/span\u003e\u003cspan address=\"10.1039/B818166G\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBussotti F, Feducci M, Iacopetti G et al (2018) Linking forest diversity and tree health: preliminary insights from a large-scale survey in Italy. Ecosyst 5:12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40663-018-0130-6\u003c/span\u003e\u003cspan address=\"10.1186/s40663-018-0130-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCamarero JJ, Linares JC, Sang\u0026uuml;esa-Barreda G et al (2017) The Multiple Causes of Forest Decline in Spain: Drought, Historical Logging, Competition and Biotic Stressors. In: Amoroso MM, Daniels LD, Baker PJ, Camarero JJ (eds) Dendroecology: Tree-Ring Analyses Applied to Ecological Studies. Springer International Publishing, Cham, pp 307\u0026ndash;323\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCaudullo G, Welk E, San-Miguel-Ayanz J (2017) Chorological maps for the main European woody species. Data Brief 12:662\u0026ndash;666. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.dib.2017.05.007\u003c/span\u003e\u003cspan address=\"10.1016/j.dib.2017.05.007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCavin L, Jump AS (2017) Highest drought sensitivity and lowest resistance to growth suppression are found in the range core of the tree Fagus sylvatica L. not the equatorial range edge. Glob Change Biol 23:362\u0026ndash;379. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/gcb.13366\u003c/span\u003e\u003cspan address=\"10.1111/gcb.13366\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eČerm\u0026aacute;k J, Matyssek R, Kučera J (1993) Rapid response of large, drought-stressed beech trees to irrigation. Tree Physiol 12:281\u0026ndash;290. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/treephys/12.3.281\u003c/span\u003e\u003cspan address=\"10.1093/treephys/12.3.281\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChakraborty T, Saha S, Matzarakis A, Reif A (2017) Influence of multiple biotic and abiotic factors on the crown die-back of European beech trees at their drought limit. Flora 229:58\u0026ndash;70. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.flora.2017.02.012\u003c/span\u003e\u003cspan address=\"10.1016/j.flora.2017.02.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChaplot V, Walter C (2003) Subsurface topography to enhance the prediction of the spatial distribution of soil wetness. Hydrol Process 17:2567\u0026ndash;2580. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/hyp.1273\u003c/span\u003e\u003cspan address=\"10.1002/hyp.1273\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChartin C, Bourennane H, Salvador-Blanes S et al (2011) France Geomorphology 135:8\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.geomorph.2011.07.020\u003c/span\u003e\u003cspan address=\"10.1016/j.geomorph.2011.07.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Classification and mapping of anthropogenic landforms on cultivated hillslopes using DEMs and soil thickness data \u0026mdash; Example from the SW Parisian Basin,\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChauvier Y, Thuiller W, Brun P et al (2021) Influence of climate, soil, and land cover on plant species distribution in the European Alps. Ecol Monogr 91:e01433. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ecm.1433\u003c/span\u003e\u003cspan address=\"10.1002/ecm.1433\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheaib A, Badeau V, Boe J et al (2012) Climate change impacts on tree ranges: model intercomparison facilitates understanding and quantification of uncertainty. Ecol Lett 15:533\u0026ndash;544. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1461-0248.2012.01764.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1461-0248.2012.01764.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChira D, Danescu F, Rosu C et al (2006) Aspects on European beech in some experimental plots in Oriental Carpanthians. In: Proceedings. International Conference. Beech silviculture in Europe\u0026rsquo;s largest beech country. IUFRO, pp 111\u0026ndash;113\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoat B, Jansen S, Brodribb TJ et al (2012) Global convergence in the vulnerability of forests to drought. Nature 491:752\u0026ndash;755. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nature11688\u003c/span\u003e\u003cspan address=\"10.1038/nature11688\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCramer RD III, Bunce JD, Patterson DE, Frank IE (1988) Crossvalidation, Bootstrapping, and Partial Least Squares Compared with Multiple Regression in Conventional QSAR Studies. Quant Struct-Act Relatsh 7:18\u0026ndash;25. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/qsar.19880070105\u003c/span\u003e\u003cspan address=\"10.1002/qsar.19880070105\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCsill\u0026eacute;ry K, Lalag\u0026uuml;e H, Vendramin GG et al (2014) Detecting short spatial scale local adaptation and epistatic selection in climate-related candidate genes in European beech (Fagus sylvatica) populations. Mol Ecol 23:4696\u0026ndash;4708. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/mec.12902\u003c/span\u003e\u003cspan address=\"10.1111/mec.12902\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Frenne P, Lenoir J, Luoto M et al (2021) Forest microclimates and climate change: Importance, drivers and future research agenda. Glob Change Biol 27:2279\u0026ndash;2297. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/gcb.15569\u003c/span\u003e\u003cspan address=\"10.1111/gcb.15569\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDiaconu D, Kahle H-P, Spiecker H (2017) Thinning increases drought tolerance of European beech: a case study on two forested slopes on opposite sides of a valley. Eur J Res 136:319\u0026ndash;328. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10342-017-1033-8\u003c/span\u003e\u003cspan address=\"10.1007/s10342-017-1033-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDittmar C, Elling W (2007) Dendroecological investigation of the vitality of Common Beech (\u003cem\u003eFagus sylvatica\u003c/em\u003e L.) in mixed mountain forests of the Northern Alps (South Bavaria). Dendrochronologia 25:37\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.dendro.2007.01.003\u003c/span\u003e\u003cspan address=\"10.1016/j.dendro.2007.01.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDittmar C, Fricke W, Elling W (2006) Impact of late frost events on radial growth of common beech (Fagus sylvatica L.) in Southern Germany. Eur J Res 125:249\u0026ndash;259. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10342-005-0098-y\u003c/span\u003e\u003cspan address=\"10.1007/s10342-005-0098-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDobbertin M (2005) Tree growth as indicator of tree vitality and of tree reaction to environmental stress: a review. Eur J Res 124:319\u0026ndash;333. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10342-005-0085-3\u003c/span\u003e\u003cspan address=\"10.1007/s10342-005-0085-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDobbertin M, Brang P (2001) Crown defoliation improves tree mortality models. Ecol Manag 141:271\u0026ndash;284. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0378-1127(00)00335-2\u003c/span\u003e\u003cspan address=\"10.1016/S0378-1127(00)00335-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEichhorn J, Roskams P, Potocic (2020) ICP Forests Manual: Part IV Visual assessment of crown condition and damaging agents. In: ICP Forests Manual. 55 p. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.icp-forests.org/pdf/manual/2020/ICP_Manual_part04_2020_Crown_version_2020-3_update_06-2023.pdf\u003c/span\u003e\u003cspan address=\"https://www.icp-forests.org/pdf/manual/2020/ICP_Manual_part04_2020_Crown_version_2020-3_update_06-2023.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEickenscheidt N, Wellbrock N (2014) Consistency of defoliation data of the national training courses for the forest condition survey in Germany from 1992 to 2012. Environ Monit Assess 186:257\u0026ndash;275. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10661-013-3372-3\u003c/span\u003e\u003cspan address=\"10.1007/s10661-013-3372-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFAO (ed) (2006) Guidelines for soil description, 4. ed. Food and Agriculture Organization of the United Nations, Rome.109 p. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.fao.org/4/a0541e/a0541e.pdf\u003c/span\u003e\u003cspan address=\"https://www.fao.org/4/a0541e/a0541e.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFernandes V (2012) En quoi l\u0026rsquo;approche PLS est-elle une m\u0026eacute;thode a (re)-d\u0026eacute;couvrir pour les chercheurs en management ? M@n@gement 15:102\u0026ndash;123. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3917/mana.151.0102\u003c/span\u003e\u003cspan address=\"10.3917/mana.151.0102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerretti M (2021) New appetite for the monitoring of European forests. Ann Sci 78:94. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s13595-021-01112-w\u003c/span\u003e\u003cspan address=\"10.1007/s13595-021-01112-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFranklin JF, Shugart HH, Harmon ME (1987) Tree Death as an Ecological Process. Bioscience 37:550\u0026ndash;556. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/1310665\u003c/span\u003e\u003cspan address=\"10.2307/1310665\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrei ER, Gossner MM, Vitasse Y et al (2022) European beech dieback after premature leaf senescence during the 2018 drought in northern Switzerland. Plant Biol 24:1132\u0026ndash;1145. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/plb.13467\u003c/span\u003e\u003cspan address=\"10.1111/plb.13467\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGaliano L, Mart\u0026iacute;nez-Vilalta J, Lloret F (2010) Drought-Induced Multifactor Decline of Scots Pine in the Pyrenees and Potential Vegetation Change by the Expansion of Co-occurring Oak Species. Ecosystems 13:978\u0026ndash;991. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10021-010-9368-8\u003c/span\u003e\u003cspan address=\"10.1007/s10021-010-9368-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGranier A, Br\u0026eacute;da N, Biron P, Villette S (1999) A lumped water balance model to evaluate duration and intensity of drought constraints in forest stands. Ecol Model 116:269\u0026ndash;283. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0304-3800(98)00205-1\u003c/span\u003e\u003cspan address=\"10.1016/S0304-3800(98)00205-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHartmann H, Moura CF, Anderegg WRL et al (2018) Research frontiers for improving our understanding of drought-induced tree and forest mortality. New Phytol 218:15\u0026ndash;28. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.15048\u003c/span\u003e\u003cspan address=\"10.1111/nph.15048\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHengl T, Hannes I (2021) Geomorphometry: Concepts, Software. Applications - International Society for Geomorphometry, Elsevier\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHouston Durrant T, De Rigo D, Mauri A et al (2016) European atlas of forest tree species. Office des publications de l\u0026rsquo;Union europ\u0026eacute;enne, LU\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026uuml;lsmann L, Bugmann HKM, Commarmot B et al (2016) Does one model fit all? Patterns of beech mortality in natural forests of three European regions. Ecol Appl 26:2465\u0026ndash;2479. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/eap.1388\u003c/span\u003e\u003cspan address=\"10.1002/eap.1388\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIGN (2012) IF: Quelles sont les ressources exploitables ? IGN \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://inventaire-forestier.ign.fr/IMG/pdf/IF31.pdf\u003c/span\u003e\u003cspan address=\"https://inventaire-forestier.ign.fr/IMG/pdf/IF31.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInnes JL, Landmann G, Mettendorf B (1993) Consistency of observations of forest tree defoliation in three European countries. Environ Monit Assess 25:29\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/BF00549790\u003c/span\u003e\u003cspan address=\"10.1007/BF00549790\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJamagne M, B\u0026eacute;tr\u0026eacute;mieux R, B\u0026eacute;gon JC, Mori A (1977) Quelques donn\u0026eacute;es sur la variabilit\u0026eacute; dans le milieu naturel de la r\u0026eacute;serve en eau des sols. Bull Tech Inf 324\u0026ndash;325:627\u0026ndash;641\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoly D, Brossard T, Cardot H et al (2010) Les types de climats en France, une construction spatiale. Cybergeo Eur J Geogr. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4000/cybergeo.23155\u003c/span\u003e\u003cspan address=\"10.4000/cybergeo.23155\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLanger GJ, Bu\u0026szlig;kamp J (2023) Vitality loss of beech: a serious threat to Fagus sylvatica in Germany in the context of global warming. J Plant Dis Prot 130:1101\u0026ndash;1115. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s41348-023-00743-7\u003c/span\u003e\u003cspan address=\"10.1007/s41348-023-00743-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLebourgeois F, Differt J, Granier A et al (2002) Premi\u0026egrave;res observations ph\u0026eacute;nologiques des peuplements du r\u0026eacute;seau national de suivi \u0026agrave; long terme des \u0026eacute;cosyst\u0026egrave;mes forestiers (Renecofor). Rev Fr 54:407\u0026ndash;418. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4267/2042/4930\u003c/span\u003e\u003cspan address=\"10.4267/2042/4930\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLech P, Kamińska A (2024) Mortality, or not mortality, that is the question \u0026hellip; How to Treat Removals in Tree Survival Analysis of Central European Managed Forests. Plants 13:248. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/plants13020248\u003c/span\u003e\u003cspan address=\"10.3390/plants13020248\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLemaire J, Vennetier M, Pr\u0026eacute;vosto B, Cailleret M (2022) Interactive effects of abiotic factors and biotic agents on Scots pine dieback: A multivariate modeling approach in southeast France. Ecol Manag 526:120543. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2022.120543\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2022.120543\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeuschner C (2020a) Drought response of European beech (Fagus sylvatica L.)\u0026mdash;A review. Perspect Plant Ecol Evol Syst 47:125576. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ppees.2020.125576\u003c/span\u003e\u003cspan address=\"10.1016/j.ppees.2020.125576\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeuschner C (2020b) Drought response of European beech (Fagus sylvatica L.)\u0026mdash;A review. Perspect Plant Ecol Evol Syst 47:125576. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ppees.2020.125576\u003c/span\u003e\u003cspan address=\"10.1016/j.ppees.2020.125576\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeuschner C, Ellenberg H (2017) Vegetation Ecology of Central Europe, Springer, Berlin, Germany. Springer, Berlin \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://link.springer.com/book/10.1007/978-3-319-43042-3\u003c/span\u003e\u003cspan address=\"https://link.springer.com/book/10.1007/978-3-319-43042-3\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLies E (1980) Le d\u0026eacute;p\u0026eacute;rissement du h\u0026ecirc;tre au Grand-Duch\u0026eacute; de Luxembourg. Ann Sci For 37:275\u0026ndash;277. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1051/forest:19800402\u003c/span\u003e\u003cspan address=\"10.1051/forest:19800402\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartinez del Castillo E, Zang CS, Buras A et al (2022) Climate-change-driven growth decline of European beech forests. Commun Biol 5:1\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s42003-022-03107-3\u003c/span\u003e\u003cspan address=\"10.1038/s42003-022-03107-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMathes T, Seidel D, Klemmt H-J et al (2024) The effect of forest structure on drought stress in beech forests (Fagus sylvatica L). Ecol Manag 554:121667. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2023.121667\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2023.121667\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMauri A, Strona G, San-Miguel-Ayanz J (2017) EU-Forest, a high-resolution tree occurrence dataset for Europe. Sci Data 4:160123. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/sdata.2016.123\u003c/span\u003e\u003cspan address=\"10.1038/sdata.2016.123\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcDowell NG, Allen CD (2015) Darcy\u0026rsquo;s law predicts widespread forest mortality under climate warming. Nat Clim Change 5:669\u0026ndash;672. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nclimate2641\u003c/span\u003e\u003cspan address=\"10.1038/nclimate2641\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcKee TB, Doesken NJ, Kleist J (1993) The relationship of drought frequency and duration to time scales. Eighth Conf Appl Climatol, Anaheim, 17\u0026ndash;22 January 1993, 179\u0026ndash;184\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMellert KH, Lenoir J, Winter S et al (2018) Soil water storage appears to compensate for climatic aridity at the xeric margin of European tree species distribution. Eur J Res 137:79\u0026ndash;92. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10342-017-1092-x\u003c/span\u003e\u003cspan address=\"10.1007/s10342-017-1092-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMetz J, Annigh\u0026ouml;fer P, Schall P et al (2016) Site-adapted admixed tree species reduce drought susceptibility of mature European beech. Glob Change Biol 22:903\u0026ndash;920. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/gcb.13113\u003c/span\u003e\u003cspan address=\"10.1111/gcb.13113\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMirabel M, Gaertner P-A (2023) D\u0026eacute;p\u0026eacute;rissements de h\u0026ecirc;tres adultes apr\u0026egrave;s 2018 en Franche-Comt\u0026eacute;. Rev Fr 74:177\u0026ndash;186. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.20870/revforfr.2023.7591\u003c/span\u003e\u003cspan address=\"10.20870/revforfr.2023.7591\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMirabel M, Saintonge F-X, Caroulle F (2024) Bilan de l\u0026rsquo;\u0026eacute;tat sanitaire du h\u0026ecirc;tre en France fin 2023. D\u0026eacute;p Sant\u0026eacute; For Juillet 2024:14 p. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://draaf.bourgogne-franche-comte.agriculture.gouv.fr/IMG/pdf/dsf_bilan_hetre_juin_2024.pdf\u003c/span\u003e\u003cspan address=\"https://draaf.bourgogne-franche-comte.agriculture.gouv.fr/IMG/pdf/dsf_bilan_hetre_juin_2024.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026ouml;lder I, Leuschner C (2014) European beech grows better and is less drought sensitive in mixed than in pure stands. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00468-014-0991-4\u003c/span\u003e\u003cspan address=\"10.1007/s00468-014-0991-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026uuml;ller E, Stierlin HR (1990) Sanasilva Tree crown photos., Sanasilva. 98p\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaas AE, Keetz LT, Halvorsen R et al (2024) Choice of predictors and complexity for ecosystem distribution models: effects on performance and transferability. Ecography 2024:e07269. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ecog.07269\u003c/span\u003e\u003cspan address=\"10.1111/ecog.07269\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNageleisen L-M (1994) Le d\u0026eacute;p\u0026eacute;rissement actuel de feuillus divers: h\u0026ecirc;tre, merisier, alisier torminal, \u0026eacute;rable sycomore, peuplier, ch\u0026acirc;taignier, charme, aulne glutineux. Rev Fr 554. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4267/2042/26583\u003c/span\u003e\u003cspan address=\"10.4267/2042/26583\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNageleisen L-M (2005) D\u0026eacute;p\u0026eacute;rissement du H\u0026ecirc;tre: pr\u0026eacute;sentation d\u0026rsquo;une m\u0026eacute;thode symptomatologique de suivi. Rev Fr 57:255\u0026ndash;262. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4267/2042/5041\u003c/span\u003e\u003cspan address=\"10.4267/2042/5041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNageleisen L-M (1993) Les d\u0026eacute;p\u0026eacute;rissements d\u0026rsquo;essences feuillues en France. Rev Fr 45:605\u0026ndash;620. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4267/2042/26462\u003c/span\u003e\u003cspan address=\"10.4267/2042/26462\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNageleisen L-M, Reuter J-C (2007) D\u0026eacute;p\u0026eacute;rissement de h\u0026ecirc;tre: retour sur les situations de crise de la h\u0026ecirc;traie lors de la derni\u0026egrave;re d\u0026eacute;cennie. Bilan Sant\u0026eacute; For. 2006 7\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNeycken A, Scheggia M, Bigler C, L\u0026eacute;vesque M (2022) Long-term growth decline precedes sudden crown dieback of European beech. Agric Meteorol 324:109103. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.agrformet.2022.109103\u003c/span\u003e\u003cspan address=\"10.1016/j.agrformet.2022.109103\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNeycken A, Wohlgemuth T, Frei ER et al (2024) Slower growth prior to the 2018 drought and a high growth sensitivity to previous year summer conditions predisposed European beech to crown dieback. Sci Total Environ 912:169068. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.scitotenv.2023.169068\u003c/span\u003e\u003cspan address=\"10.1016/j.scitotenv.2023.169068\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eObladen N, Dechering P, Skiadaresis G et al (2021) Tree mortality of European beech and Norway spruce induced by 2018\u0026ndash;2019 hot droughts in central Germany. Agric Meteorol 307:108482. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.agrformet.2021.108482\u003c/span\u003e\u003cspan address=\"10.1016/j.agrformet.2021.108482\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOgnjenović M, Seletković I, Potočić N et al (2022) Defoliation Change of European Beech (Fagus sylvatica L.) Depends on Previous Year Drought. Plants 11:730. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/plants11060730\u003c/span\u003e\u003cspan address=\"10.3390/plants11060730\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePe\u0026ntilde;uelas J, Hunt JM, Ogaya R, Jump AS (2008) Twentieth century changes of tree-ring δ13C at the southern range-edge of Fagus sylvatica: increasing water-use efficiency does not avoid the growth decline induced by warming at low altitudes. Glob Change Biol 14:1076\u0026ndash;1088. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1365-2486.2008.01563.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1365-2486.2008.01563.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeterken GF, Mountford EP (1996) Effects of drought on beech in Lady Park Wood, an unmanaged mixed deciduous woodland | Forestry: An International Journal of Forest Research | Oxford Academic. Int J Res 69:125\u0026ndash;136. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/forestry/69.2.125\u003c/span\u003e\u003cspan address=\"10.1093/forestry/69.2.125\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetit-Cailleux C, Davi H, Lef\u0026egrave;vre F et al (2021) Comparing statistical and mechanistic models to identify the drivers of mortality within a rear-edge beech population. Peer Community J. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.24072/pcjournal.60\u003c/span\u003e\u003cspan address=\"10.24072/pcjournal.60\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 1:\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetroselli A, Vessella F, Cavagnuolo L et al (2013) Ecological behavior of Quercus suber and Quercus ilex inferred by topographic wetness index (TWI). Trees 27:1201\u0026ndash;1215. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00468-013-0869-x\u003c/span\u003e\u003cspan address=\"10.1007/s00468-013-0869-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePichler M, Hartig F (2023) Machine learning and deep learning\u0026mdash;A review for ecologists. Methods Ecol Evol 14:994\u0026ndash;1016. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/2041-210X.14061\u003c/span\u003e\u003cspan address=\"10.1111/2041-210X.14061\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiedallu C, G\u0026eacute;gout JC (2007) Multiscale computation of solar radiation for predictive vegetation modelling. Ann Sci For 64:899\u0026ndash;909\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiedallu C, G\u0026eacute;gout J-C, Perez V, Lebourgeois F (2013) Soil water balance performs better than climatic water variables in tree species distribution modelling. Glob Ecol Biogeogr 22:470\u0026ndash;482. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/geb.12012\u003c/span\u003e\u003cspan address=\"10.1111/geb.12012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePilard-Landeau B, Piat J, Joudain M, Fontenelle D (1994) Le d\u0026eacute;p\u0026eacute;rissement actuel du h\u0026ecirc;tre en Picardie. Rev Fr 46 \u0026ndash; 5 547\u0026ndash;553. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4267/2042/26582\u003c/span\u003e\u003cspan address=\"10.4267/2042/26582\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiovesan G, Biondi F, Filippo AD et al (2008) Drought-driven growth reduction in old beech (Fagus sylvatica L.) forests of the central Apennines, Italy. Glob Change Biol 14:1265\u0026ndash;1281. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1365-2486.2008.01570.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1365-2486.2008.01570.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePluess AR, Frank A, Heiri C et al (2016) Genome\u0026ndash;environment association study suggests local adaptation to climate at the regional scale in Fagus sylvatica. New Phytol 210:589\u0026ndash;601. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.13809\u003c/span\u003e\u003cspan address=\"10.1111/nph.13809\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePopa I, Badea O, Silaghi D (2017) Influence of climate on tree health evaluated by defoliation in the ICP level I network (Romania). IForest - Biogeosciences For 10:554. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3832/ifor2202-009\u003c/span\u003e\u003cspan address=\"10.3832/ifor2202-009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePramreiter M, Grabner M (2023) The Utilization of European Beech Wood (Fagus sylvatica L.) in Europe. Forests 14:1419. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/f14071419\u003c/span\u003e\u003cspan address=\"10.3390/f14071419\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuintana-Segu\u0026iacute; P, Moigne PL, Durand Y et al (2008) Analysis of Near-Surface Atmospheric Variables: Validation of the SAFRAN Analysis over France. J Appl Meteorol Climatol 47:92\u0026ndash;107. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1175/2007JAMC1636.1\u003c/span\u003e\u003cspan address=\"10.1175/2007JAMC1636.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team (2023) R: A Language and environment for statistical computing. [Computer software]. Retrieved from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cran.r-project.org\u003c/span\u003e\u003cspan address=\"https://cran.r-project.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRasztovits E, Berki I, M\u0026aacute;ty\u0026aacute;s C et al (2014) The incorporation of extreme drought events improves models for beech persistence at its distribution limit. Ann Sci 71:201\u0026ndash;210. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s13595-013-0346-0\u003c/span\u003e\u003cspan address=\"10.1007/s13595-013-0346-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRohner B, Kumar S, Liechti K et al (2021) Tree vitality indicators revealed a rapid response of beech forests to the 2018 drought. Ecol Indic 120:106903. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ecolind.2020.106903\u003c/span\u003e\u003cspan address=\"10.1016/j.ecolind.2020.106903\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRohner B, Waldner P, Lischke H et al (2018) Predicting individual-tree growth of central European tree species as a function of site, stand, management, nutrient, and climate effects. Eur J Res 137:29\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10342-017-1087-7\u003c/span\u003e\u003cspan address=\"10.1007/s10342-017-1087-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoloff A (1989) Crown development and vitality assessment of selected tree species of temperate latitudes. Crown Dev Vitality Assess Sel Tree Species Temp Latit\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRose L, Leuschner C, K\u0026ouml;ckemann B, Buschmann H (2009) Are marginal beech (Fagus sylvatica L.) provenances a source for drought tolerant ecotypes? Eur J Res 128:335\u0026ndash;343. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10342-009-0268-4\u003c/span\u003e\u003cspan address=\"10.1007/s10342-009-0268-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRouault G, Candau J-N, Lieutier F et al (2006) Effects of drought and heat on forest insect populations in relation to the 2003 drought in Western Europe. Ann Sci 63:613\u0026ndash;624. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1051/forest:2006044\u003c/span\u003e\u003cspan address=\"10.1051/forest:2006044\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRozas V, Camarero JJ, Sang\u0026uuml;esa-Barreda G et al (2015) Summer drought and ENSO-related cloudiness distinctly drive Fagus sylvatica growth near the species rear-edge in northern Spain. Agric Meteorol 201:153\u0026ndash;164. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.agrformet.2014.11.012\u003c/span\u003e\u003cspan address=\"10.1016/j.agrformet.2014.11.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuffault J, Pimont F, Cochard H et al (2022) SurEau-Ecos v2.0: a trait-based plant hydraulics model for simulations of plant water status and drought-induced mortality at the ecosystem level. Geosci Model Dev 15:5593\u0026ndash;5626. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5194/gmd-15-5593-2022\u003c/span\u003e\u003cspan address=\"10.5194/gmd-15-5593-2022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRukh S, Sanders TGM, Kr\u0026uuml;ger I et al (2023) Distinct Responses of European Beech (Fagus sylvatica L.) to Drought Intensity and Length\u0026mdash;A Review of the Impacts of the 2003 and 2018\u0026ndash;2019 Drought Events in Central Europe. Forests 14:248. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/f14020248\u003c/span\u003e\u003cspan address=\"10.3390/f14020248\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaintonge FX (2023) Qualifier plus fr\u0026eacute;quemment la d\u0026eacute;gradation des for\u0026ecirc;ts avec DEPERIS. For\u0026ecirc;t Innov 5:30\u0026ndash;33\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalvador-Blanes S, Cornu S, Couturier A et al (2006) Morphological and geochemical properties of soil accumulated in hedge-induced terraces in the Massif Central, France. Soil Tillage Res 85:62\u0026ndash;77. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.still.2004.12.008\u003c/span\u003e\u003cspan address=\"10.1016/j.still.2004.12.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanders T, Pitman R, Broadmeadow M (2012) Soil type modifies climate-growth response of beech in Southern Britain. Proceedings of the DENDROSYMPOSIUM Scientific Technical Report STR 12/03. p 106\u0026ndash;110 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2312/GFZ.b103-12036\u003c/span\u003e\u003cspan address=\"10.2312/GFZ.b103-12036\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchaeffer A (1955) Le d\u0026eacute;p\u0026eacute;rissement du h\u0026ecirc;tre dans le d\u0026eacute;partement du Doubs.Bulletion Soci\u0026eacute;t\u0026eacute; Foresti\u0026egrave;re Franche-Comt\u0026eacute; XXVIII:290\u0026ndash;291\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmidt F, Persson A (2003) Comparison of DEM Data Capture and Topographic Wetness Indices. Precis Agric 4:179\u0026ndash;192. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1023/A:1024509322709\u003c/span\u003e\u003cspan address=\"10.1023/A:1024509322709\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmied G, Pretzsch H, Ambs D et al (2023) Rapid beech decline under recurrent drought stress: Individual neighborhood structure and soil properties matter. Ecol Manag 545:121305. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2023.121305\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2023.121305\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmull M, Thomas F (2000) Morphological and physiological reactions of young deciduous trees (Quercus robur L., Q. petraea [Matt.] Liebl., Fagus sylvatica L.) to waterlogging. Plant Soil 225:227\u0026ndash;242. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1023/A:1026516027096\u003c/span\u003e\u003cspan address=\"10.1023/A:1026516027096\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchneider DP, Deser C, Fasullo J, Trenberth KE (2013) Climate Data Guide Spurs Discovery and Understanding. Eos Trans Am Geophys Union 94:121\u0026ndash;122. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/2013EO130001\u003c/span\u003e\u003cspan address=\"10.1002/2013EO130001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchuldt B, Buras A, Arend M et al (2020) A first assessment of the impact of the extreme 2018 summer drought on Central European forests. Basic Appl Ecol 45:86\u0026ndash;103. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.baae.2020.04.003\u003c/span\u003e\u003cspan address=\"10.1016/j.baae.2020.04.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeidling W, Ziche D, Beck W (2012) Climate responses and interrelations of stem increment and crown transparency in Norway spruce, Scots pine, and common beech. Ecol Manag 284:196\u0026ndash;204. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2012.07.015\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2012.07.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSenf C, Pflugmacher D, Zhiqiang Y et al (2018) Canopy mortality has doubled in Europe\u0026rsquo;s temperate forests over the last three decades. Nat Commun 9:4978. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-018-07539-6\u003c/span\u003e\u003cspan address=\"10.1038/s41467-018-07539-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSerra-Maluquer X, Gazol A, Sang\u0026uuml;esa-Barreda G et al (2019) Geographically Structured Growth decline of Rear-Edge Iberian Fagus sylvatica Forests After the 1980s Shift Toward a Warmer Climate. Ecosystems 22:1325\u0026ndash;1337. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10021-019-00339-z\u003c/span\u003e\u003cspan address=\"10.1007/s10021-019-00339-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSilva DE (2010) Ecologie du h\u0026ecirc;tre (Fagus sylvatica L.) en marge sud-ouest de son aire de distribution. Phdthesis, Universit\u0026eacute; Henri Poincar\u0026eacute; - Nancy 1. 198 p. HAL Id: tel-01746152, version 1\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimon A, Katzensteiner K, Wallentin G (2023) The integration of hierarchical levels of scale in tree species distribution models of silver fir (\u003cem\u003eAbies alba\u003c/em\u003e Mill.) and European beech (\u003cem\u003eFagus sylvatica\u003c/em\u003e L.) in mountain forests. Ecol Model 485:110499. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ecolmodel.2023.110499\u003c/span\u003e\u003cspan address=\"10.1016/j.ecolmodel.2023.110499\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSousa-Silva R, Verheyen K, Ponette Q et al (2018) Tree diversity mitigates defoliation after a drought-induced tipping point. Glob Change Biol 24:4304\u0026ndash;4315. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/gcb.14326\u003c/span\u003e\u003cspan address=\"10.1111/gcb.14326\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTaccoen A, Piedallu C, Seynave I et al (2019) Background mortality drivers of European tree species: climate change matters. Proc R Soc B Biol Sci 286:20190386. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1098/rspb.2019.0386\u003c/span\u003e\u003cspan address=\"10.1098/rspb.2019.0386\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTenenhaus M (1998) La r\u0026eacute;gression PLS: th\u0026eacute;orie et pratique. Editions TECHNIP\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTer-Braak CJT, Juggins (1993) Weighted averaging partial least square regression (WA-PLS): an improved method for reconstructing environnemental variables from species assemblages. Hydrobiologia 269\u0026ndash;270:485\u0026ndash;502\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTesfa TK, Tarboton DG, Chandler DG, McNamara JP (2009) Modeling soil depth from topographic and land cover attributes. Water Resour Res 45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2008WR007474\u003c/span\u003e\u003cspan address=\"10.1029/2008WR007474\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeissier du Cros E, Le Tacon F, Nepveu G et al (1981) Le h\u0026ecirc;tre. INRA Editions.613 p\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTetegan M, Richer de Forges AC, Verbeque B et al (2015) The effect of soil stoniness on the estimation of water retention properties of soils: A case study from central France. CATENA 129:95\u0026ndash;102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.catena.2015.03.008\u003c/span\u003e\u003cspan address=\"10.1016/j.catena.2015.03.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThornthwaite CW, Mather JR (1957) Instructions and tables for computing potential evapotranspiration and the water balance. Climatology X:185\u0026ndash;311\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTognetti R, Johnson JD, Michelozzi M (1995) The response of European beech (Fagus sylvatica L.) seedlings from two Italian populations to drought and recovery. Trees 9:348\u0026ndash;354. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/BF00202499\u003c/span\u003e\u003cspan address=\"10.1007/BF00202499\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTo\u0026iuml;go M, Nicolas M, Jonard M et al (2020) Temporal trends in tree defoliation and response to multiple biotic and abiotic stresses. Ecol Manag 477:118476. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2020.118476\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2020.118476\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTouche J, Calvaruso C, De Donato P, Turpault MP (2022) Five successive years of rainfall exclusion induce nutritional stress in a mature beech stand. Ecol Manag 507:119987. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2021.119987\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2021.119987\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTurc L (1955) Le bilan d\u0026rsquo;eau des sols. Relations entre les pr\u0026eacute;cipitations, l\u0026rsquo;\u0026eacute;vaporation et l\u0026rsquo;\u0026eacute;coulement. Ann Agronomiques 6:5\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Maaten E (2013) Thinning prolongs growth duration of European beech (\u003cem\u003eFagus sylvatica\u003c/em\u003e L.) across a valley in southwestern Germany. Ecol Manag 306:135\u0026ndash;141. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2013.06.030\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2013.06.030\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Mantgem PJ, Stephenson NL, Byrne JC et al (2009) Widespread increase of tree mortality rates in the western United States. Science 323:521\u0026ndash;524. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/science.1165000\u003c/span\u003e\u003cspan address=\"10.1126/science.1165000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVannoppen A, Kint V, Ponette Q et al (2019) Tree species diversity impacts average radial growth of beech and oak trees in Belgium, not their long-term growth trend. Ecosyst 6:10. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40663-019-0169-z\u003c/span\u003e\u003cspan address=\"10.1186/s40663-019-0169-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVennetier M, Ripert C (2005) Evaluer la potentialit\u0026eacute; foresti\u0026egrave;re d\u0026rsquo;un site sans observer la flore. Pr\u0026eacute;sentation d\u0026rsquo;une m\u0026eacute;thode d\u0026eacute;velopp\u0026eacute;e en r\u0026eacute;gion m\u0026eacute;dit\u0026eacute;rann\u0026eacute;enn\u0026eacute;e calclaire. RDV Tech (ONF) n\u0026deg; 8:15\u0026ndash;22\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWalthert L, Ganthaler A, Mayr S et al (2021) From the comfort zone to crown dieback: Sequence of physiological stress thresholds in mature European beech trees across progressive drought. Sci Total Environ 753:141792. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.scitotenv.2020.141792\u003c/span\u003e\u003cspan address=\"10.1016/j.scitotenv.2020.141792\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaring RH (1987) Characteristics of Trees Predisposed to Die. Bioscience 37:569\u0026ndash;574. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/1310667\u003c/span\u003e\u003cspan address=\"10.2307/1310667\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeiss AD (2001) Topographic position and landforms analysis. ESRI User Conf\u0026eacute;rence The Nature Conservancy (San Diego, Californie). Poster URL Httpwwwjennessentcomdownloadstpi-Poster-Tnc18x22pdf\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilkens JF, Wagner S (2021) Empirical survival model for European beech (Fagus sylvatica L.) seedlings in response to interactive resource gradients and (a-) biotic conditions within an experimental canopy gap study. Ecol Manag 499:119627. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2021.119627\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2021.119627\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWolfe BT, Sperry JS, Kursar TA (2016) Does leaf shedding protect stems from cavitation during seasonal droughts? A test of the hydraulic fuse hypothesis. New Phytol 212:1007\u0026ndash;1018. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.14087\u003c/span\u003e\u003cspan address=\"10.1111/nph.14087\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWoodcock H, Vollenweider P, Dubs R, Hofer R-M (1995) Crown alterations induced by decline: a study of relationships between growth rate and crown morphology in beech (Fagus sylvatica L). Trees 9:279\u0026ndash;288. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/BF00202018\u003c/span\u003e\u003cspan address=\"10.1007/BF00202018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"european-journal-of-forest-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"efor","sideBox":"Learn more about [European Journal of Forest Research](http://link.springer.com/journal/10342)","snPcode":"10342","submissionUrl":"https://submission.nature.com/new-submission/10342/3","title":"European Journal of Forest Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Fagus sylvatica, Statistical modeling, Climate, Soil factors, Topography factors, Dendrometry.","lastPublishedDoi":"10.21203/rs.3.rs-5417359/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5417359/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e \u003cem\u003eFagus sylvatica\u003c/em\u003e L. is a main forest tree species in Europe but has been subjected to massive decline events over the last decades. This phenomenon has been mainly attributed to the increase in drought frequency and intensity, but it is unclear how the local specificities in stand structure, climatic, soil and topographic conditions interact, and if statistical models are able to capture the high spatial and temporal variability in tree decline. To fulfil this objective, we measured 5380 \u003cem\u003eFagus sylvatica\u003c/em\u003e trees from 308 plots distributed in four regions of France with contrasting environmental conditions, and designed models predicting decline at both regional and national scales. These models aimed at assessing the percentage of stems by plot with at least 50% crown biomass loss based on 229 dendrometric, topographic, soil and climatic variables.\u003c/p\u003e \u003cp\u003eThe climatic factors explained most of the variability in stand decline, especially the long-term deviations from the 30-years mean in maximal temperature and in hydric deficit. Regional models were the most efficient in predicting beech decline in their calibration areas (Q\u0026sup2; varied from 0.26 to 0.42) as they better consider the local environmental factors. They were less effective in the other regions, and the national model was an acceptable compromise on a larger scale. These statistical models provide valuable insights for forest managers and could be improved through a more detailed temporal stand monitoring to control the effects of management and decline dynamics.\u003c/p\u003e","manuscriptTitle":"Multiscale Modelling of European Beech Decline: The Role of Long-Term Climate Deviations and Local Environmental Factors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-27 18:15:32","doi":"10.21203/rs.3.rs-5417359/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-01-12T10:06:45+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-01-08T14:39:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"116440568484882796572658465704773518914","date":"2024-12-24T13:21:45+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-12-18T11:25:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"222739325391243960042164980363597704653","date":"2024-12-13T09:23:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"33765504417387038175837936224044821004","date":"2024-12-10T07:59:21+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-11-14T10:12:43+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-11-09T06:11:45+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-11-09T06:10:56+00:00","index":"","fulltext":""},{"type":"submitted","content":"European Journal of Forest Research","date":"2024-11-08T14:39:38+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"european-journal-of-forest-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"efor","sideBox":"Learn more about [European Journal of Forest Research](http://link.springer.com/journal/10342)","snPcode":"10342","submissionUrl":"https://submission.nature.com/new-submission/10342/3","title":"European Journal of Forest Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"d76010a1-5e6d-4c11-9c21-bfb88c6096b7","owner":[],"postedDate":"November 27th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-05-12T16:08:28+00:00","versionOfRecord":{"articleIdentity":"rs-5417359","link":"https://doi.org/10.1007/s10342-025-01767-4","journal":{"identity":"european-journal-of-forest-research","isVorOnly":false,"title":"European Journal of Forest Research"},"publishedOn":"2025-05-05 15:57:07","publishedOnDateReadable":"May 5th, 2025"},"versionCreatedAt":"2024-11-27 18:15:32","video":"","vorDoi":"10.1007/s10342-025-01767-4","vorDoiUrl":"https://doi.org/10.1007/s10342-025-01767-4","workflowStages":[]},"version":"v1","identity":"rs-5417359","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5417359","identity":"rs-5417359","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}