Sampling sites
The study region is the Ore Mts in NW Czechia, Central of Europe (see Fig. 1). This mountain range stretches along the border of Czechia and Germany in the SW-NE direction. It is about 150 km long, approximately 40 km wide, and its highest mountain is Klínovec (1,244 m a.s.l.). The relief is formed by a long-tilted block with a steep scarp towards Czechia, with an elevational change of up to 600 m. On top of the mountain range there are plateaus and uplands at elevations not exceeding 900 m a.s.l. (Melichar and Krása 2009). Gneiss, granite, phyllite and mica schist form the bedrock of the Ore Mts. The regional climate is moderately cold, with mean July temperatures of ca 12–15°C and a mean annual temperature of 5.4°C. Mean July temperatures in the upper parts of the mountains sometimes drop to below 10°C. The average annual precipitation is about 750 mm, with around 450 mm during the vegetation period lasting 112 days (Plíva and Žlábek 1986).
Four sites were chosen for this study (see Fig. 1). All of them lie in a mesophytic region in the central to eastern part of the Ore Mts. The choice of the sites was based on the presence of agricultural landforms and the tree species of focus, the sycamore maple (Acer pseudoplatanus L.). The sycamore maple was identified as the only tree species inhabiting all the sites in dendrochronologically sound amounts. The sites vary in elevation and slope, and there are two types of soil and bedrock (see Online Resource 1). The study sites are located at places known as Blatno, Kninice, Dlouha Louka and Nova Ves, further referred to as the B-, K-, D- and N-site. All the sites except for the K-site fall within a cold climate region; the Kninice site belongs to a moderately warm one (Web Map Service: climatic regions map of Czechia – basic characteristic of estimated pedologic-ecological units © VÚMOP, 2008). The character of the vegetation cover at the B-site and the K-site differs from that at the D-site and N-site. The first two sites are characterized by a dense canopy cover on the landforms whereas the latter two have a much sparser woody vegetation cover, as is visible in the orthophoto maps, see Fig. 1. The co-occurring tree vegetation at these sites differs, too. At the B-site, the tree vegetation consists of dominating sycamore maple and common hazel (Corylus avellana), bird cherry (Prunus avium), common beech (Fagus sylvatica), European crab apple (Malus sylvestris) and European white birch (Betula pendula). By contrast, at the K-site, the sycamore maple does not dominate, being part of a variable mixture of common hazel, common beech, Norway maple (Acer platanoides), bird cherry and European white birch. At the N-site, sycamore maple dominates, accompanied by mountain ash (Sorbus aucuparia) and common hazel, complemented by bird cherry. At the D-site, sycamore maple dominates, mostly as polycormons, complemented by mountain ash and bird cherry trees.
Sycamore maple ecology
The sycamore maple (Acer pseudoplatanus L.) was selected because it grows spontaneously on the landforms under study. This native deciduous tree species can live for more than 350 years, reaching heights of ca 30 m and trunk diameters of 60–80 cm (Praciak 2013). The species grows at higher elevations compared to most broadleaved trees, and the presence of suitable soil is a stronger determining factor than the climate (Jones 1945). Sycamores grow best on well-drained and fertile soils (Hein et al. 2009), tolerating intermediately shaded conditions under the canopy, particularly when juvenile (Jones 1945), which explains its successful growth in established woodlands (Petritan et al. 2007). Microclimates at the bottom of ravines and slopes are convenient for sycamores, based on their shade tolerance and preference for nutrient-rich soil. The species can also grow on calcareous substrates associated with coarse screes, steep rocky slopes, ravines and cliffs, which corresponds to the stone content of ASWs. Sycamore stands can be found on base-rich rocks as well as on acidic soils (Jones 1945). Their roots increase the stability of slopes and mitigate erosion, acting highly effectively against rockfalls (Norris et al. 2008). The sycamore grows fast on suitable sites (Praciak 2013). The species often dominates in mixed deciduous woodlands or forests of the maple-lime type, designated as protected priority habitat 9180 Tilio-Acerion forests of slopes, screes and ravines (European Commission 2013) It tolerates low summer temperatures and pollution (Praciak 2013). Compared to field and Norway maples, the sycamore maple is not such a drought-tolerant species (Röhrig and Ulrich 1991). That corresponds with the results of Morecroft et al. (2008), who observed slower growth of sycamores and lower photosynthetic rate in the canopy during dry periods, when the soil is dry. The effect of global warming on the sycamore maple is still a subject of debates (Hemery et al. 2010; Carón et al. 2015a). Initial studies in experimental plant ecology about seed germination suggest that Acer species react negatively to extreme warming conditions and are weakly resilient to excess or irregular water availability (Carón et al. 2015b).
Woody vegetation bi-temporal spatial analysis
The landscape of the Ore Mts has undergone a period of land-use disintensification due to land abandonment after the Second World War. The main change has been a reduction in land-use for agriculture and pastoral purposes. This has resulted in the lack or drop in ASW management. We used archive aerial imagery from 1953 (Web Map Service: Historical ortophotographic map (1950s) © CENIA 2022, Czech Environmental Information Agency) and an ortophotographic map from 2019/2021 (Web Map Service - Ortophotographic map of Czechia © ČUZK 2022, Czech Office for Surveying, Mapping and Cadastre) as sources of information on the extent of the vegetation cover (both maps are accessible through web map services). The extent of the tree canopy, reflecting management intensity, was digitized to compare the rate of woody vegetation-dynamics on ASWs among the study sites over the last 70 years.
Sample collection and processing
At all the sites, we collected tree cores from sycamore maple individuals growing on stone walls. We only sampled trees with single trunks (without branching at the roots) to avoid the effects of branching, which induces multiple pith formation and adds noise to the data. We collected one core per tree at breast height using a standard 0.5 mm increment borer. We followed the procedures as prescribed in the standard textbook ‘Methods of Dendrochronology’ (Cook and Kairiukstis 1990). At all sites we collected metadata such as the location of the tree (coordinates obtained using a GARMIN GPS device), elevation and the surrounding dominant and co-dominant vegetation. After extracting the cores from the trees, we secured the cores in polythene straws and labelled them. We then transported the cores to the laboratory, where we mounted them on wooden holders and sanded them using a belt sander. In the next step, we scanned the polished cores using a high-resolution Epson flatbed scanner.
Development of ring width series, detrending, site-chronologies
We measured ring widths from scanned images of tree cores using the CooRecorder software and assembled ring width chronologies. Using CDendro software, which is an extension to the COO Recorder (CYBIS 2006), we cross-dated individual ring width chronologies from every site and stored the ring width measurements in files in RWL format. We cross-dated individual cores using a combination of block-wise correlation and mean inter-series correlation analyses for all individuals within each site and proceeded in a similar manner for all the sites. We recorded all the missing rings or false rings during the cross dating. Their summary is presented in Table 1. We imported the RWL files into R (Core Team 2021) and detrended the chronologies using a cubic smoothing spline to remove the age-related growth trend and to retain the high-frequency trend. These detrended chronologies were thus prepared for further analyses. We plotted the detrended ring with indices as boxplots for every year and plotted the master chronology as a line. This was necessary to visualize the data over the aspects of growth variability for every year along with the mean growth for the entire time window from 1949 to 2020.
Table 1
Age structure of trees studied at all study sites in the Ore Mts
Site | B-site | D-site | K-site | N-site |
Mean tree age | 64 | 65 | 65 | 62 |
Number of samples | 20 | 19 | 19 | 19 |
Standard deviation | 20 | 31 | 15 | 33 |
Oldest individual sampled | 116 | 115 | 100 | 112 |
Number of missing rings | 2 | 4 | 2 | 3 |
Intra annual density fluctuations (IADFs) | 6 | 8 | 3 | 6 |
Climate correlation analyses
We correlated the detrended cross-dated ring width indices with monthly temperature and precipitation values. Monthly temperature and precipitation data were acquired from the CRUTEM dataset available from the World Meteorological Organization online database (WMO 2022). The climate data is gridded over 0.5°-×-0.5° grid generated by the data provided by regional climate stations, which are then interpolated across space. We carried out a climate correlation analysis in the R statistical programming language (Core Team 2021), using the dplR and TREECLIM packages (Zang and Franco 2022). We used the master chronology generated by averaging all the individual chronologies from every site and correlated it with temperature and precipitation data. We used a moving window correlation analysis with a window size of 20 years. The correlations were bootstrapped to ensure stability of values. The correlation values were then plotted as colour matrix plots (Fig. 4 and Fig. 5).
Comparison of radial growth within and among sites
To gain a better understanding of how individual chronologies have contributed to the master chronology (individual growth variability) at every site, we examined various parameters of the chronologies, for example mean inter-series correlation (r), express population signal (EPS) and glk (Gleichläufigkeit). We further noted the within-site growth variability as well as among-site growth variability.
Pointer-year analysis
Negative growth extreme years were distinguished from detrended tree ring width chronologies for all the sites by a threshold of plus/minus one standard deviation. Only extremes occurring at three or all sites were considered and only negative growth extremes were studied, as they reflect unfavourable environmental conditions. The negative pointer year analysis that we carried out was done only for the time period starting from 1949 to 2020, because this time period had consistent sample depth. We compared the ring width data to mean March to May (i.e. spring) and June to September (i.e. summer) temperatures and precipitation values. We examined local historical records to see if there were extreme climatic events to confirm the pointer years (Treml 2011; Brazdil et al. 2015).