Please summarrise the following paper, be very comprehensive:
Norway spruce forest management in the Czech Republic is linked to the solar cycle under conditions of climate change – from tree rings to salvage harvesting
Václav Šimůnek1*, Zdeněk Vacek1, Stanislav Vacek1, Michal Švanda2,3, Vojtěch Hájek1 and Giuseppe D’Andrea1
1 Czech University of Life Sciences Prague, Faculty of Forestry and Wood Sciences, Kamýcká 129, CZ – 165 21 Prague 6 – Suchdol, Czech Republic
2 Astronomical Institute of the Czech Academy of Sciences, Fričova 298, 25165 Ondřejov, Czech Republic
3 Astronomical Institute, Charles University, Faculty of Mathematics and Physics, V Holešovičkách 2, 18000 Praha, Czech Republic
Received: 20 April 2023 Accepted: 13 September 2024
Abstract
Norway spruce (Picea abies [L.] Karst.) is the most important economic tree species in Central Europe. It has been affected by extensive windstorms and bark beetle disturbances, which have intensified in recent years. Spruce stands are subject to regular reoccurring disturbances with different intensities. This study examines timber harvests in the Czech Republic and the tree-ring series of Norway spruce from six permanent research plots across the country. The timber harvest shows a cyclical connection with the sunspot number, and of these, the percentage of spruce logging shows a substantial link with the sunspot number while the percentage of salvage logging indicates a relation with Total Solar Irradiance (TSI) and seasonal temperature. The seasonal North Atlantic Oscillation (NAO) indicated a link with logging types as well. However, the connection between TSI and SunSpot Number (SSN) with logging types was statistically significant, while the tree rings were more related to seasonal NAO than timber harvests. The tree Ring Width Index (RWI) series shows reduced growth, which precedes the peak in the timber harvests by 1 year. Timber harvests peak one year after minimum RWI, and they coincide with solar minimum years. In summary, average tree ring series reach predominantly their lowest values one year before the solar minimum, coinciding with increased spruce and salvage logging. Conversely, during the solar maximum, harvests are lower, and tree ring series remain relatively stable. In the spectral analysis, tree-ring series and all types of timber harvests show a quasi-11-year cycle. Spruce trees indicate higher tree-ring growth and lower timber harvests during solar maximum compared to solar minimum and this may be associated with specific meteorological conditions that may be affected by solar variability. European forestry is unfamiliar with the effects of the solar cycle; nevertheless, this phenomenon is present in both spruce tree rings and the aggregate of timber harvests. Further research on this issue will be necessary, but it is already apparent that regularly recurring calamities in spruce stands are likely to be reflected in all of Central Europe.
Key words: Forest management / Sunspot number / Total solar irradiance / Picea abies (L.) Karst. / North Atlantic Oscillation
© V. Šimůnek et al., Published by EDP Sciences 2024
Licence Creative CommonsThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
1 Introduction
The forests of Central Europe – Norway spruce (Picea abies [L.] Karst.), in particular – are crucially affected by climate change. Forest management of this most widespread tree species, on which Czech forestry economically depends, is expected to face increasing difficulties (Thiele et al., 2017). Spruce stands have suffered large-scale damage during the recent years of climate change (MAF, 2018; Toth et al., 2020). Europe is challenged by more frequent wildfires or intense windthrow with uneven distribution of precipitation and increasing temperatures caused by climate change (Kundzewicz et al., 2006; Vacek et al., 2023). Climate change has also influenced coniferous tree silviculture in the Czech Republic and neighboring countries, which have recently experienced damage from periodic droughts (Ashraf et al., 2015; Alvarez et al., 2016; Noce et al., 2016; Tumajer et al., 2017). Spruce stands account for the largest tree species representing, 49.5% in the Czech Republic. The entire forestry sector is based on coniferous tree production, while 71% are coniferous stands, the rest (29%) are deciduous stands (MAF, 2020). Norway spruce is one of the coniferous species less adaptable to drought during the summer months due to its poor ability to reach moisture deeper in the soil profile (Hartl-Meier et al., 2018).
The poor condition of spruce stands is also accompanied by outbreaks of bark beetle infestation, reoccurring predominately in dry years and thus, enhancing the disintegration of spruce stands (Kopáček et al., 2015; Nováková & Edwards-Jonášová, 2015). More frequent logging operations lead to infestations with other secondary pests, such as fungal diseases Heterobasidion spp. and Armillaria spp. (Aosaar et al., 2020; Blomquist et al., 2020). One of the common fungal diseases, Heterobasidion annosum (Fr.) Bref., weakens the base of spruce trunks, making them more susceptible to breakage and overturning (Piri, 1996; Vacek et al., 2020). The health status of spruce is also complicated by the disease of the root system, Armillaria spp., which weakens spruce stands during drought, increasing their susceptibility to bark beetle attacks (Holuša et al., 2018). As a result of climate change, dieback has led to the decline of spruce stands in recent years (Čermák et al., 2019). Game browsing and bark stripping further undermine the health of forest stands, making them more sensitive to climatic factors, especially drought, which reduces the increment and stability of spruce stands (Cukor et al., 2019; Krisans et al., 2020).
Stimulated by the climate, populations of Ips typographus (L.) and other bark beetles increase and attack healthy trees as well (Marini et al., 2017; Netherer et al., 2019). The general decline of spruce in the Czech Republic is compounded by the overall aging of forest stands: the amount of stands over 81 years of age is progressively increasing which can be seen in the report from 2021, along with the amount of salvage logging, even in the older forests (MAF, 2021). Old spruce stands, in particular, are more sensitive to temperature fluctuations (Kukumägi et al., 2017; Vacek et al., 2019). Older spruce forests are sensitive to higher temperatures due to their reduced ability to adapt to climate changes, resulting in excessive drought and subsequently increasing tree vulnerability to pests such as bark beetles (Hlásny et al., 2017). In addition, the older the forest stand, the higher the risk of disturbance (Vacchiano et al., 2013) and of Ips typographus that primarily attacks weakened trees older than 60 years (Schroeder & Lindelöw, 2002; Hlásny & Turčáni, 2013).
Climate change has proven to have a negative impact on forestry economic indicators affecting spruce management, which ostensibly fails to eliminate negative climate variability (Hlásny et al., 2017). Higher short-term salvage logging, induced by ecological calamities, lowers timber prices, which directly impacts forest management (Remeš et al., 2020; Toth et al., 2020). All impacts combined reduce the economic value of forests in the long term (Hanewinkel et al., 2013) and potentially challenge atmospheric carbon fixation as well (Dobor et al., 2020). Climatic fluctuations force forest management to adapt silviculture to nature’s unexpected changes (Spiecker, 2003).
The periodic influence of the 11-year solar cycle on climate may offer a partial answer to the cyclical disintegration of forests. The 11-year cycle is observed in the forest environment and tree rings (Matveev et al., 2017; Komitov, 2021). Sunspot cycles have been observed to influence many factors such as the ozone layer, stratospheric temperature, El Niño-Southern Oscillation, zonal winds, etc. (Rigozo et al., 2007; Maycock et al., 2016, 2018; Ball et al., 2019; Tartaglione et al., 2020). Solar activity can influence the temperature and precipitation extremes of the Earth’s atmosphere, e.g., during colder winter periods or during the summer European weather patterns (Gray et al., 2010; Rimbu et al., 2021).
The solar activity cycle also manifests in total solar irradiance, which causes the TSI fluctuations of 0.1% (Kopp et al., 2016). A direct connection is assumed between the 11-year solar cycle and temperatures in specific geographic regions during certain multi-decadal periods, specifically in February, March, June, and September (Lüdecke et al., 2020) and indirectly, also with the North Atlantic Oscillation (NAO) and the Atlantic Multidecadal Oscillation (AMO) – (Bice et al., 2012). Jet streams exhibit non-uniformity in terms of wind speed and altitude along their length, often featuring fragmented sections with the greatest baroclinicity. The variability and mean flow tend to maximize over oceanic regions (Blackmon, 1976). Recent studies have presented conflicting results regarding the solar cycle influences on winter climate over the North Atlantic/European region. Polar front jet stream variability is responsible for instances of extreme weather and is crucial for regional climate change, particularly affecting western Europe and eastern North America (Brugnara et al., 2013; Hall et al., 2015; Ma et al., 2018). The modulation effect of solar variability on the North Atlantic Oscillation (NAO) is currently a topic of controversy. In recent discussions, arguments both in favor (Drews et al., 2022) of and against (Chiodo et al., 2019; Spiegl et al., 2023) the solar impact on NAO have been analyzed.
The Quasi-Biennial Oscillation (QBO) is a repeating wind pattern in the Earth’s stratosphere that changes direction approximately every 28–29 months. During the QBO’s east phase, there is an increased likelihood of sudden stratospheric warming, leading to a warm polar stratosphere and potential cooling over Northern Europe. Observations show a positive correlation between 30-hPa North Pole troposphere temperatures and the 11-year solar cycle during the QBO’s west phase, while no correlation is observed during its east phase. Observations indicate an enhanced likelihood of sudden stratospheric warming during the QBO east phase, leading to a warming of the polar stratosphere and potentially causing cooling over Northern Europe in the troposphere by triggering cold air outbreaks (Labitzke, 2005). This short-term 28-month cycle significantly influences stratospheric and tropospheric temperatures (Baldwin et al., 2001; Coy et al., 2017). The circulation over Europe is also linked with El Niño-Southern Oscillation (ENSO) which affects weather patterns worldwide. It involves the periodic warming (El Niño) and cooling (La Niña) of sea surface temperatures in the central and eastern tropical Pacific Ocean. In Europe, it can influence winter weather by altering the position and strength of the jet stream, leading to shifts in atmospheric pressure patterns. El Niño is associated with milder and wetter conditions, while La Niña tends to bring colder and drier weather. These ENSO-related changes can impact European regions, affecting precipitation, temperature, and even extreme events (Brönnimann, 2007; Cai et al., 2010). Variations in the cloud cover show similarities with the variability modes of the ENSO, NAO, and QBO. Several publications highlight a solar influence on the summer monsoon, primarily through variations in TSI. For instance, studies have shown that changes in TSI can impact the Asian summer monsoon, affecting weather patterns (Wang et al., 2005; Shi et al., 2014).
The European climate is significantly impacted by the NAO and shifting wind patterns, which respond differently to climate change along latitude gradients. The NAO has a modest early-season influence on summer temperatures (Kjellström et al., 2013) and affects both precipitation and airflow patterns during dry seasons (Tsanis and Tapoglou, 2019). The NAO has positive and negative phases. A negative phase can bring cooler, drier conditions to Central Europe, while the positive phase does the almost opposite (Vicente-Serrano et al., 2011; Yao & Luo, 2014; Steirou et al., 2017; Tatli & Menteş, 2019).
The NAO and solar cycle dynamics also manifest in annual tree growth. While NAO-tree ring responses vary across Europe, they are often tied to local climate and seasonal growth effects. Central Europe has witnessed changing NAO-related tree-ring responses over the past century, with distinct trends in the 20th century compared to the 19th century (St. George, 2014). Notably, during February, NAO can indirectly impact tree-ring growth by influencing precipitation and groundwater levels (Akhmetzyanov et al., 2023).
Various dendrochronological research studies also demonstrated the influence of solar cycles on the radial growth of different tree species, in studies from Northwest Russia (Shumilov et al., 2011; Kasatkina et al., 2019), Chile (Rigozo et al., 2002), the Tibetan Plateau (Wang & Zhang, 2011), or Central and Southern Europe (Surový et al., 2008; Šimůnek et al., 2021b). The effect of the solar cycle on the growth of trees has been observed in the annual rings of trees in the past. The fossil finds revealed the presence of an 11–15-year cycle in tree-rings data in central Italy (Cecchini et al., 1996). There are also studies describing the coincidence of solar cycles with forest fire risk (Uğur & Feriha, 2017). Studying climate patterns like the North NAO and solar cycle can enhance our understanding of mechanisms. This knowledge, coupled with insights into the cyclic processes, has the potential to significantly aid forestry and timber-processing management in adapting to these fluctuations through enhanced prediction. Forest management planning and ecological stability of the stands will benefit from the knowledge of forest calamity cycles, which will also result in reduced economic losses.
This study aims to compare the radial growth of spruce with harvesting and the course of the solar cycle. It describes the tree-ring growth of Norway spruce at different plots in the Czech Republic, types of forest timber harvests for the entire Czech Republic, and data for total, salvage, and spruce harvests. Two time periods were evaluated: 1900–2019 and 1961–2019. Tree-ring growth of Norway spruce and different types of timber harvest in the Czech Republic were compared with the sunspot number, total solar irradiance, air temperature, seasonal temperature, precipitation, seasonal precipitation, annual NAO, and seasonal NAO. The datasets used in this study describe the relationship in the oscillations between timber harvests and tree-ring series with respect to the course of precipitation, temperatures, NAO, total solar irradiance, and the sunspot cycle. An important objective of this study is not only to demonstrate the climatic influences affecting timber harvesting and tree-ring series but to also understand the sequence and order of factors preceding fluctuations in salvage logging and spruce timber harvests. The influence of the 11-year cycle on the development of timber harvesting and ring-width index has not been jointly compared so far. Therefore, with this study, we hope to generate a new perspective on the relative roles of the various influencing factors and the involved process in forest calamities in central Europe and their possible connection with tree rings and climate data fluctuations.
2 Methodology
2.1 Study area
The study area for the dendrochronological analysis is located in the Czech Republic. The research plots are situated according to altitude and occupy lower-lying, middle, and mountain positions. A total of six research plots were evaluated, of which the first two – the Karlštejn (1_KAR) and Kostelec (2_KOS) plots – belong to the lower positions, while the Jeseník (3_JES) and Podkrkonoší (4_POD) plots are mid-altitude positions, and Orlické hory (5_ORL) and Bažinky (6_BAZ) plots classify as mountain sites. Dendrochronological samples from these plots were obtained from forest stands with dominant Norway spruce. The research plots of 6_BAZ, 3_JES, and 5_ORL are located in the no-logging areas, where no logging operations have been conducted since 1985, and prior to that, a shelterwood system based on selection logging had been applied. In contrast, the research plots of 4_POD, 1_KAR, and 2_KOS are commercial forest stands managed in close-to-nature ways based on a shelterwood system, target trees, and the tending of the growing stock. Trophic and mesotrophic cryptopodzols to nutrient-poor, oligotrophic soils of gleys and podzols dominate on the surveyed sites. Soil moisture conditions in the study plots range from dry to gleyed soils. Table 1 provides more information on the individual research plots.
Table 1Basic plot and stand characteristics of Norway spruce research plots in 2019.
The precipitation and air temperature conditions vary across the Czech Republic, therefore, meteorological stations nearest to the research plots were used for the dendrochronological analyses (Fig. 1a). The basic habitat and stand characteristics of the plots are given in Table 1. In the Czech Republic, lower altitudes are primarily influenced by the continental climate: according to Koppen’s classification (Köppen, 1936), the lower altitude sites are in the climatic region Cfb – temperate oceanic climate with warm summer without dry seasons. The middle elevations fall within region Dfb – a warm-summer humid continental climate characterized by hot summers and cold winters. The higher elevations are region Dfc – subarctic cold (continental) climate with cold summers without a dry season (Köppen, 1936). The annual air temperature in the Czech Republic ranges from +9.5 to −0.4 °C, with an average of 7.9 °C, and the annual precipitation ranges from 410 to 1705 mm with an average of 650 mm for the entire country in relation to altitude, from 115 to 1603 m a.s.l. The average altitude in the Czech Republic is 430 m a.s.l. The lowest average air temperature in January is −2.0 °C, and the highest average air temperature in June is 17.8 °C (ČHMÚ, 2022).
thumbnail Figure 1
Location of research plots 1–6 (grey dots) and the meteorological stations used for dendrochronological analyses (black flags); the grey pentagon indicates the capital city Prague (upper map). Research plots number: 1–1_KAR; 2–2_KOS; 3–3_JES; 4–4_POD; 5–5_ORL; 6–6_BAZ (a). The change of the fraction of forest area and total volume of forest stands (b) as well as forest ownership (c) in the Czech Republic in ten-year intervals from 1950 to 2020.
2.1.1 Demographic evolution of the forests in the Czech Republic
Forest ecosystems in the Czech Republic have undergone significant transformations from 1900 to the present. Following the aftermath of the First World War, the forest cover declined to approximately 30%. However, subsequent efforts in afforestation on previously unused lands resulted in an increase in forest cover to around 34% by 2020 (Fig. 1b). The most extensive afforestation took place after the Second World War when privately owned lands were nationalized and converted into state forests. Another significant event affecting forestry was the air pollution catastrophe from high concentrations of SO2 in the mountainous regions of the Czech Republic during the seventies and eighties, particularly in the mountain ranges bordering Germany and Poland (Materna, 1989; Vacek et al., 2015a).
The political regime changes in 1992 resulted in the transformation of state forests into private ownership (as depicted in Fig. 1c). Since the 1990s, the existing forest stands have gradually aged, accompanied by an increased adoption of mechanization in forest management practices. This mechanization trend has led to a reduction in the number of forestry employees from 70,000 in 1985 to 14,000 in 2020 (CZSO, 2021). The reduction in the number of employees led to insufficiently fast processing of bark beetle-infested wood, allowing the calamities to grow to larger proportions, subsequently exerting pressure on the price of bark beetle-infested wood (Hlásny et al., 2021a). The situation after the restitution (return) of forest stands to private ownership in the 1990s also resulted in the fragmentation of smaller land properties, which in many cases led to an increase in bark beetles and delays in the processing of infested wood, a situation that continues to persist until today (Seidl et al., 2016).
Since the year 2000, there has been a notable increase in forest stand mortality, primarily affecting Norway spruce and other coniferous species. Climate change is considered the main driver behind this calamity, with higher average temperatures and inadequate precipitation during critical climatic periods. These unfavorable conditions have contributed to the bark beetle graduation calamities which are the most significant secondary problem after the drought of the current Czech forestry (MAF, 2021).
2.2 Data collection
For the dendrochronological analysis of the samples, increment cores were taken from Norway spruce using a Pressler auger (Haglöf, Långsele, Västernorrland, Sweden) perpendicular to the trunk axis at a height of 1.3 m above the ground. Structurally homogeneous spruce stands with a stand density of 0.8 (80%) to 1 (100%) were selected for sampling. Samples were collected from square 50 × 50 m study plots from healthy co-dominant and dominant trees compared to sub-dominant and suppressed trees (Remeš et al., 2015) with an average diameter at breast height (DBH) > 25 cm. Height was measured with a Laser Vertex hypsometer (Haglöf, Långsele, Västernorrland, Sweden) to an accuracy of 0.1 m for all trees with increment cores obtained. For all trees, the DBH was also measured with a Mantax Blue metal caliper (Haglöf, Långsele, Västernorrland, Sweden) with a precision of 1 mm. A total of 318 samples were collected for dendrochronological analysis, with sample core numbers per plot ranging from 40 to 58 per plot (Table 3). The cores were measured using a LINTAB measuring stage with an Olympus microscope. The measuring table gauges to an accuracy of 0.01 mm, and TSAP-Win software (Rinntech, 2010) was used to record the increment cores. Subsequent cross-dating of the core samples was performed in the CDendro program (Cybis Elektronik & Data AB, Sweden) so that the cross-correlation index was CC > 25 for each sample (Larsson, 2013).
Monthly temperature and precipitation data for the whole Czech Republic were provided and statistically prepared by the Czech Hydrometeorological Institute, Prague1. This study also used data from the longest measuring climatic station in the Czech Republic, Klementinum (altitude 191 m a.s.l.; 50°5′12″N, 14°24′56″E). The vegetation season extends from May to September. A detailed overview of the meteorological stations used and their relationship to the research plots is shown in Table 2.
Table 2Location and meteorological characteristics of Norway spruce research plots in 2019.
The annual mean sunspot number data were provided by the Royal Observatory of Belgium, Brussels2. Data on the annual TSI were taken from the composite of the TSI measurements and reconstruction by Greg Kopp running up to the year 2019 (Dudok de Wit et al., 2017). Additionally, for the period prior to 1978, the TSI reconstruction by Wu et al. (2018) using the SATIRE model was incorporated. The TSI reconstruction was made available by G. Kopp on his webpage3. The TSI model is composed of a total irradiance model that captures 92% variance of sunspot number and absolute scale creation is reported and explained by Kopp & Lean (2011).
Timber harvest data were obtained from the Forest Management Institute in Brandýs nad Labem, Czech Republic4 and the Czech Statistical Office in Prague, Czech Republic5. The total harvested timber is the volume of large timber (apart from logging residues), including self-production (self-production refers to wood collection by individuals, either partially or entirely, in exchange for a predetermined fee or without any charge). Timber from logging and silvicultural operations includes salvage logging. The total coniferous timber logging in the Czech Republic also includes salvage logging in coniferous stands. The total spruce timber logging in the Czech Republic includes salvage logging too. Salvage logging data cover all types of salvage logging and calamities caused by abiotic and biotic factors. It also includes dead-standing trees, isolated breaks, uprootings, all of the trees felled for trapping bark beetles, and individual trees in which harmful insects (bark beetles, etc.) spend the winter (Nownes, 2012).
To eliminate political and economic influences on timber harvesting, salvage logging, and spruce logging have been converted into percentage shares of the main harvests. This step eliminates quantity fluctuations in timber harvesting dependent on economic planning and better captures overall fluctuations in these types of logging. Since salvage logging is a practical forestry operation carried out as quickly as possible to assess the damage, it is possible to say that climatic fluctuations best mirror the development of salvage logging to weather fluctuations, bark beetle damage, and other pests. Given that conifers are known to be more sensitive to climatic fluctuations and sudden events, spruce, which is also the most represented conifer species in the Czech Republic, is most harvested during salvage logging.
The data on the NAO are obtained from the webpage of the Climatic Research Unit, University of East Anglia6. The data of the annual and monthly NAO index are derived from Gibraltar and South-west Iceland. During the winter season, the contrast between the standardized sea level pressure in Gibraltar and Southwest Iceland serves as a valuable indicator of the NAO intensity (Jones et al., 1997; CRU-UEA, 2023).
Figure 2 shows the main data used for different types of logging, sunspot number, TSI, precipitation, temperature, and NAO in this study (climatic data from the meteorological stations are not shown due to the extensive amount of data).
thumbnail Figure 2
Open access or institutional data used in this study: (a) Dynamics of total timber harvest, total spruce timber logging, and salvage logging in the Czech Republic, the amount of timber in mill. m3 – million cubic meters of harvested timber; (b) Total solar irradiance (TSI); (c) Precipitation from the longest monitored meteorological station, Klementinum (1900–2019), and from the Czech Republic (in 1961–2019); (d) Air temperature from the Czech Republic (in 1961–2019) and from the longest measuring meteorological station, Klementinum (1900–2019); (e) North Atlantic Oscillation and sunspot number.
2.3 Data analysis
Dendrochronological data for Norway spruce (the data are described in result Sects. 3.1 and 3.2) were processed with the R software (Team R Core, 2022) using the “dplR” package (Bunn, 2008; Zang et al., 2018). Detrending of each tree was performed by the “dplR” package (Bunn, 2008, 2010). A two-step negative exponential detrending with spline (of n = 67% of series length) was used on each tree. Detrending removes the age trend while preserving low-frequency climate signals (Cook et al., 1990; Shumilov et al., 2011). The detrended tree-ring growth data are denoted as tree ring-width index (RWI) further. The expressed population signal (EPS) was calculated for the detrended data. The EPS serves as a metric for assessing the reliability of a chronology and is determined by a combination of inter-series correlations R-bar and sample size (Fritts, 1966; Anthony et al., 2003). The EPS represents the reliability of a chronology as a fraction of the joint variance of the theoretical infinite tree population. The limit for using the data for comparison against climate data was a significant EPS threshold, EPS > 0.85 (Bunn et al., 2018). Also, the signal-to-noise ratio (SNR) was calculated, which represents the signal strength of chronology, and inter-series correlations (R-bar). The SNR is a statistical measure that compares the strength of the desired signal in a data set of series to the level of background noise. A higher value of SNR represents a stronger climatic signal relative to noise. The R-bar describes the similarity of tree-ring patterns across multiple samples. R-bar is the average pairwise correlation coefficient between individual trees within a chronology. A higher R-bar value indicates stronger coherence among the tree-ring patterns (Fritts, 1976). The first-order autocorrelation (ar1) was also calculated. The ar1 refers to the degree of correlation between a data point and the preceding one in a time series of tree ring chronology. The EPS, SNR, R-bar, and ar1 indexes were calculated according to the instructions in the “dplR” (Bunn et al., 2018) and are based on general dendrochronological theories (Fritts, 1976; Speer, 2010).
Table 3 shows basic descriptive data obtained for the research plots from which tree-ring analyses were conducted to describe tree-ring spruce series measured (LINTAB). These indicators in the table are calculated for the period from 1900 to 2019 that is used in the study. Data for this table was acquired using the dendrochronological method described in Section 2.2. The table clearly presents the course of mean ring width (RW) and mean min–max (mm), which was derived from undetrended tree-ring curves. The indicators ar1, R-bar, EPS, and SNR of the detrended tree-ring series are used in our results.
Table 3Characteristics of tree-ring chronologies for Norway spruce in research plots for time period 1900–2019.
To investigate the influence of sunspot number and TSI on tree-ring growth and timber harvesting, all data were analysed. Precipitation, temperature, and NAO data were also evaluated in relation to tree ring growth, as they play a crucial role in the forest environment. In addition, cyclical or causal order relationships were described using spectral analysis, and superposed epoch analysis (SEA) on the main tree ring data and types of timber harvesting. Furthermore, the cyclical relationships between timber harvesting and tree ring series were investigated, which could potentially explain their mutual relationship in comparison to solar activity.
Spectral analyses of the data were performed with the Statistica 13 software (StatSoft, 2013). The calculation was performed with the “Single Fourier (Spectral) Analysis” function, using the output of the “Periodogram” plot by “Period.” The periodogram values are interpreted in terms of variance (sums of squares) of the data at the respective period. The periodogram ascribes values to periods. Additionally, multiple linear regression analyses were calculated using this software.
The statistical analysis, conducted with the Statistica software (StatSoft, Tulsa, CA, USA), a multiple linear regression approach was employed to explore the relationships between various independent variables and the dependent variable. The Multiple R field shows the coefficient of multiple correlation, representing the positive square root of R-square (the coefficient of multiple determination). This statistic is handy in multivariate regression, particularly when dealing with multiple independent variables, to describe the relationship between these variables. An overall F-test (F) was performed to assess the association between the dependent variable and the set of independent variables, with the F-value and its corresponding p-value serving as crucial indicators (F = Regression Mean Square/Residual Mean Square). The overall F-test is founded on the null hypothesis that all coefficients are zero. Therefore, it serves to identify models that are clearly mis-specified, although it is not designed to quantify how good a model is. In multiple linear regression, it is typically assumed that the residuals (observed minus predicted values) follow a normal distribution. While F is generally robust even if this assumption is violated, it is prudent to review the distributions of the main variable of interest before reaching final conclusions. F is used to show relationships between the dependent variable and the set of independent variables. Additionally, an Adjusted R 2 value was computed, akin to the R 2 coefficient, but adjusted to account for the degrees of freedom by considering the error sum of squares and total sums of squares. The df is degrees of freedom (different in the numerator and denominator) while “Residual S2” is the error sums of squares and the “Total S2” is the total sums of squares. The R 2 Adjusted is described by equation:
[
(
Residual
S
2
/
df
)
/
(
Total
S
2
/
df
)
]
.
The results are presented as standardized (b*) and non-standardized (b) regression coefficients along with their standard errors and statistical significance. The b* coefficients offer insights into the relative contributions of each independent variable when all variables are standardized to a mean of 0 and a standard deviation of 1, aiding in the prediction of the dependent variable (StatSoft, 2013). The first multiple linear regression used the dependent percentage of salvage logging and independent sunspot number, TSI, seasonal temperature CZ, seasonal precipitation CZ, NAO, and seasonal NAO. The second multiple linear regression used dependent percentage of spruce logging and independent sunspot number, TSI, seasonal temperature CZ, seasonal precipitation CZ, NAO, seasonal NAO.
The data used for analyses were calculated into yearly sampled (annual) data series. Specifically, for tree rings (dendrochronological data), timber harvesting, sunspot number, and total solar irradiance (TSI) the data represent the annual mean. Precipitation, (air) temperature, and NAO data were obtained from monthly data sets. Monthly mean temperatures, monthly mean NAO, and precipitation were averaged to annual and seasonal temperatures/precipitation, respectively, to create annual data.
Calculations for “Seasonal temperature”, “Seasonal NAO” and “Seasonal precipitation” were performed according to the length of the vegetation season at each plot (described in Sect. 2.2). Seasonal temperature and seasonal NAO are the arithmetic mean over the months of the vegetation season. The seasonal precipitation is the sum of precipitation during the vegetation season.
The salvage and spruce timber logging databases were expanded to include the percentage of total timber harvest. The “Percentage of spruce timber logging” and “Percentage of salvage logging” data series represent the percentage of the logging out of the total timber harvest in a single year. Percentage shares of all types of timber harvest were calculated for each year from 1961 to 2020.
Analyses for the 1961–2020 period were conducted by comparing the spruce research plots to data provided by the nearest meteorological station. Different types of timber harvests were compared with average meteorological data for the entire Czech Republic. Sunspot number and TSI were also used for the 1961–2020 period. Ring-width indices in 1900–2020 were compared with different meteorological data, unlike the case for 1961–2020. This is due to the absence of meteorological data before 1961 at local climate stations. The 1900–2020 period compares only Norway spruce plots with meteorological data from the Klementinum climate station, with sunspot numbers and TSI.
Principal component analysis (PCA) was performed in the CANOCO 5 program (Microcomputer Power) to evaluate the relations between RWI, TSI, sunspot numbers, NAO, timber harvest types, temperature, and precipitation. Before the analysis, the data were standardized and centralized. The results of PCA were illustrated by unconstrained ordination diagrams of species and environmental variables projected.
The Superposed Epoch Analysis (SEA; Chree, 1913) was calculated to reduce the importance of anomalies caused by other effects than the solar cycle. The SEA is a statistical technique used to reveal either periodicity within a time sequence or to find a correlation between two time series. Repeated events are defined as key times in the reference time series. The SEA key times were defined as the solar maximum years (maximal year mean in 1968, 1979, 1989, 2000, 2014) and determined the solar minimum years (minimum mean year in 1964, 1976, 1986, 1996, 2008). Subsets are then extracted from the second time series within a specified range around each key time and superposed to form a composite. Reference years for solar minima and maxima were determined automatically from the sunspot-number series by searching for the local minima or maxima of the sunspot-number curve. The averaging allows for the reduction of the stochastic background variability and enhancement of the possible signal synchronized with the key times. For each time in the composite, the uncertainty of the average may be computed as a root-mean-squared error of the mean.
Unfortunately, in our case only a few key epochs are identified and the composite series around the key times averages only a handful of realisations. In such a case the root-mean-squared of the mean may not properly assess the true variance of the mean.
Around the key times from the overall variability of the series, the existing autocorrelation of series was taken into account, and a Monte-Carlo-like approach was applied (Laken & Čalogović, 2013). For each time series in question, we constructed additional 1000 composites of an equal length with randomly chosen epochs. For each time lag in the composite, we then obtained 1000 realisations of the mean, which allowed us to estimate the confidence intervals by using 5th and 95th percentile of the empirical histogram (for the chosen statistical threshold, which is p = 0.05). Should the mean of the reference composite at the given time lag be within the confidence interval, we cannot reject the hypothesis that the composite mean, even though it appears extreme within the composite, is due to chance. On the other hand, the mean of the composite outside the chosen confidence interval indicates a statistically significant deviation.
3 Results
3.1 Tree rings during the solar cycles
The tree-ring growth of Norway spruce undergoes numerous fluctuations. The tree-ring growth during 1900–2019 in all six research plots of this study is depicted in Figure 3, along with the sunspot number and TSI. Figure 3 reveals that RWI in all research plots roughly follows the sunspot number and TSI. The agreement appears to be somewhat better between TSI and RWI of the research plots (Fig. 3b) than between sunspot number and RWI (Fig. 3a). The RWI often shows low values during the solar minimum when the sunspot number or TSI is also lower. During the solar maximum, when the sunspot number or TSI shows higher values, the RWI also exhibits higher values. The graphs also show that there can be a decrease in RWI values before the solar maximum. This can be observed, for example, in the data from 1946 or 1990. The solar cycle is evident in the RWI of almost all the research plots, yet this phenomenon also has variations in the form of time shifts, or it can be dispersed by other factors, for instance, the air pollution calamities of the 1970s and 1980s.
thumbnail Figure 3
The time series of the ring-width index (RWI) of Norway spruce in six research plots in the Czech Republic in 1900–2019 compared with (a) The sunspot number and RWI; (b) RWI in six research plots with total solar irradiance (TSI).
The complex relationships of the radial width index of individual research plots, temperature, precipitation, TSI, NAO, and sunspot number may be visualized by employing the principal component analysis. The PCA ordinations diagram is given in Figure 4 for the period 1928–2019. The first ordination axis of PCA explains 48% of data variability, the first two axes together explain 74%, and the first four axes 95%. The total variation is 27.7, and supplementary variables account for 20.6% (adjusted explained variation 14.9%). The horizontal axis illustrates the sunspot number, seasonal NAO, and TSI, and the vertical axis depicts the temperature and precipitation data set. Precipitation and temperature in the vegetation season had greater explanatory variability than annual data. In low-altitude areas (1_KAR, 2_KOS), tree-ring growth observed a positive correlation with annual precipitation (r = 0.20 to 0.44, p = 0.01 to 0.05) and especially with precipitation in vegetation season (r = 0.21 to 0.48, p = 0.01 to 0.04). On the other hand, RWI was positively correlated with annual temperature (r = 0.20, p = 0.05) and temperature in vegetation season (r = 0.22, p = 0.04) in mountain areas (6_BAZ). However, in contrast to precipitation and temperature, the sunspot number and TSI influenced the tree-ring growth of spruce in all plots, with TSI being a highly observed indicator (r = 0.29 to 0.33, p < 0.01 on 4_POD, 5_ORL, 6_BAZ). Among the observed indicators, the NAO and seasonal NAO appear to be the least important factors influencing the RWI research plots.
thumbnail Figure 4
Unconstrained ordination diagram of species and environmental variables projected showing results of the principal component analysis of relationships between the radial width index (RWI) of individual research plots, temperature (TemAnn – annual temperature, TemSea – seasonal temperature), precipitation (PrecAnn – annual precipitation, PreSea – seasonal precipitation), total solar irradiance (TSI), North Atlantic oscillation (NAO), NAOSea – seasonal NAO and sunspot number (SunNum).
3.2 Norway spruce and timber harvest
The tree-ring growth of Norway spruce since 1961 (Fig. 5a) shows that this tree species has undergone significant fluctuations. This figure describes the evolution of the percentage of spruce logging and salvage logging against the tree-ring growth in higher detail for comparison. The datasets show a common fluctuation with the solar cycle as characterized by the sunspot number (Fig. 5a). However, the common fluctuation trend of the RWI with the solar cycle is not straightforward, as shown by the 6_BAZ research plot, which was negatively affected by air pollution in 1980–1990 (inordinately high concentrations of SO2 from nearby power plants). Another asymmetry between tree-ring growth and the solar cycle is evident in 1989 for the research plots of 1_KAR and 2_KOS. All plots show the RWI concurrent with the solar cycle from 1961 to 2019. However, there are exceptions that disperse the concurrence of the RWI with the solar cycle at individual plots.
thumbnail Figure 5
The change of used data from 1961 to 2019 compared to sunspot number: (a) Ring-width index (RWI) chronologies of Norway spruce on research plots in the Czech Republic; (b) Percentage of spruce logging and percentage of salvage logging.
The percentage of spruce timber logging also indicated common fluctuation trend with the solar cycle, which is clearly visible in Figure 5b. This figure shows that in recent years, there has been a higher amount of total logging in spruce stands. The percentage of spruce timber logging does not precisely follow the solar cycles, but there is a hint of a common fluctuation trend. The percentage of salvage logging in Figure 5b has been distinctly following solar cycles since 1994, with previous years showing more frequent peaks of salvage logging.
The complex interactions between the timber harvest, radial width index of individual research plots, temperature, precipitation, TSI, NAO, and sunspot number are shown in Figure 6 for the period 1961–2019. The first ordination axis of PCA explains 42% of data variability, the first two axes together explain 67%, and the first four axes are 85%. Total variation is 696.0, supplementary variables account for 34.3% (adjusted explained variation 23.6%). The sunspot number was negatively correlated (p < 0.05) with all timber harvest types (r = −0.35 to −0.44). Annual and seasonal temperatures were positively correlated (r = 0.46 to 0.51; p > 0.05) with all timber harvest types except the percentage of salvage logging. Precipitation has less effect on harvest indicators (r = −0.25 to 0.03). Radial width indices were negatively correlated (r = −0.28 to −0.48; p > 0.05) with total salvage logging, percentage of spruce logging, and percentage of salvage logging. Generally, temperature, sunspot number, and TSI were more explanatory variables in the context of radial growth and timber harvest types compared to low-effect NAO and precipitation. Even though seasonal NAO exhibited a more noticeable negative relationship with logging compared to RWI, it did not indicate high r values for our assessed timber harvests (r = −0.07 to −0.12).
thumbnail Figure 6
Unconstrained ordination diagram of species and environmental variables projected showing results of the principal component analysis of relationships between total timber harvest (TotTimHar), total spruce logging (TotSprLog), percentage of spruce logging (SprLog%), total salvage logging (TotSalLog), percentage of salvage logging (SalLog%), the radial width index (RWI) of individual research plots, temperature (TemAnn – annual temperature, TemSea – seasonal temperature), precipitation (PrecAnn – annual precipitation, PreSea – seasonal precipitation), total solar irradiance (TSI), North Atlantic oscillation (NAO) and sunspot number (SunNum).
The application of multiple linear regression (MLR) analysis, detailed in Table 4, complements the findings from Principal Component Analysis (PCA). For the percentage of salvage logging, the MLR model yields an overall multiple R of 0.47, with a statistically significant intercept (p = 0.04). Among the variables considered, TSI exhibits the most substantial influence, indicated by a coefficient (b*) of −0.59. Additionally, seasonal temperature has a significant impact, with a b* of 0.38 at p = 0.03. In summary, in predicting the percentage of salvage logging, TSI plays a prominent role in the MLR analysis, followed by seasonal temperature.
Table 4Multiple linear regression analysis in 1961–2019 between dependent percentage of salvage logging or percentage of spruce logging and independent studied factors sunspot number, TSI, season temperature CZ, season precipitation CZ, NAO, seasonal NAO; significant results at p < 0.05 are in bold; seasonal temperature, seasonal precipitation, and seasonal NAO are detailed in the methodology section with the vegetation season.
For the percentage of spruce logging, the MLR model shows a slightly higher multiple R of 0.50 compared to the salvage logging case. However, the intercept is not statistically significant. Among the independent variables, the sunspot number holds the highest coefficient (b*) of −0.54, with the sole significant p-value of 0.04.
3.3 Feedback of solar cycles to tree rings and timber harvest
The application of the SEA (superposed epoch analysis) method to datasets of RWI and percentage of salvage logging and percentage of spruce logging (Fig. 7) reveals patterns during the solar maximum and minimum in the period 1961–2019. The SEA analysis describes the average course of the examined data during selected epochs in solar maximum or minimum.
thumbnail Figure 7
Plots of epoch-superposed subsets in 1961–2019 from solar maximum (max) and solar minimum (min) in relative −6 years before and 14 years after the key epoch; RWI research plots in max a) and in min b); percentage of spruce logging and percentage of salvage logging (logg.) in max c) and in min d); the error bars show the one-standard-deviation half-widths.
The RWI behaves differently during solar maxima and minima. During solar maxima at Lag 0, there are varying increases in RWI for each research plot (mean RWI = 1.0–1.3). It can be observed that during solar maxima, there are variable increases in spruce RWI, but the overall growth reaches in each plot peaks mostly from Lag −2 to Lag 2 (mean RWI = 0.95–1.08) with the largest RWI observed in Lag 1 (mean RWI = 1.05–1.24). During solar minima, all research plots experience the lowest RWI at mostly Lag −1 (mean RWI = 0.66–1.02). A synchronized gradual decline in RWI from Lag −6 (mean RWI = 0.92–1.08) across all research plots is observed as well. From Lag −1 (mean RWI = 0.66–1.02), there is an increase in RWI for all research areas, continuing until Lag 2 (mean RWI = 1.06–1.18).
The percentage of spruce logging during solar maxima SEA has its lowest percentage at Lag 0, which is in the range of 68–71%, and the highest logging occurs at Lag 6, reaching at range 73–81%. However, during solar minima, the percentage of spruce logging exhibits its highest values directly at Lag 0, with range of 72–82%. As we move away from Lag 0 in SEA min, the percentage of spruce logging decreases, with logging values irregularly decreasing until Lag 3 to range 68–70%. On the other hand, a gradual and regular decline in the percentage of spruce logging in SEA minimum is observed as we move into the negative Lag −5 with a range 69–71%.
The percentage of salvage logging exhibits greater variations in logging during the cycle than the percentage of spruce logging. During solar maxima, the percentage of salvage logging shows its lowest values mostly at Lag −2 with a range of 24–30%. Overall, the percentage of salvage logging during solar minima records its highest value at Lag 0, reaching the range of 39–59%. From Lag 0 in SEA minimum, salvage logging irregularly decreases on both the positive and negative sides of Lag, similar to the percentage of spruce logging. In the SEA minimum for logging from Lag 10% of spruce logging is in the high range of 75–81% and the percentage of salvage logging is in the high 18 range of 57–73%.
3.4 Spectral analysis of timber harvests and tree rings in 1961–2019
The single spectral analyses in Figure 8 show an example of RWI plot 4_POD and 6_BAZ cyclical relationships in Norway spruce indexed tree rings for the period 1900–2019. The rest of the research RWI plots are very similar to the examples. The exceptions are the research plots 4_POD (age 112 years) and 1_KAR (age 92 years), which have a shorter age period. The periodogram values show the variance of the data at the respective period. Ring-width indices (RWI) from each Norway-spruce research plot show 9-year to 14-year cycles in their tree-ring growth. Furthermore, four of the six research plots show 35-year cycles, which generally have lower periodogram values than the 9-year to 14-year cycles. The exception is the 6_BAZ research plot, which shows higher periodogram values for 60-year cycles. However, even the 6_BAZ research plot has an 11-year cycle in its tree-ring growth.
thumbnail Figure 8
Single spectral analysis of used datasets, ring-width indixes (RWI) in 1900–2019 for RWI, and types of timber logging in 1961–2019.
Timber harvest types in Figure 8 have also experienced 10-year to 14-year cycles, similar to the RWI. However, timber harvest types from 1961 to the present day have a shorter period. Regarding cyclicity, the total, coniferous, and spruce timber harvests have the most similarities, showing the same 10-year and 25-year cycles. Salvage logging follows 11-year and 35-year cycles, evidenced by spectral analysis. The percentage of salvage logging and spruce timber logging indicate 10-year to 14-year cycles and 35-year cycles, respectively. Observed RWI and harvest data show a peak at around the 11-year cycle in spectral analysis.
4 Discussion
4.1 Anthropogenic and meteorological influences in logging
Norway spruce is sensitive to weather extremes (Tumajer et al., 2017; Netherer et al., 2019). The Czech Republic also suffered from excessive sulphur dioxide deposition in the 1970s and 1980s, created by the flue gas from coal-fired power plants (Mikulenka et al., 2020; Šimůnek et al., 2021a). This negatively affected tree growth, primarily in the highest parts of the mountainous areas (Matějka et al., 2010; Vacek et al., 2015a; Šimůnek et al., 2019). The 6_BAZ research plot (Fig. 3a) responded with a sharp drop in RWI to the period of the air pollution calamity (1980–1990). This RWI drop is most evident at the highest altitude research plot of 6_BAZ, which is illustrated in Figure 5a. This reduction in increment, induced by air pollution, led to forest dieback in the area and an increase in deadwood (Vacek et al., 2015b). Air pollution also had an impact on the salvage, total, spruce, and coniferous timber harvests in other mountainous areas, evidenced in Figure 5, as an average of 6.2 million m3 of timber was harvested in salvage loggings. As a result of the air pollution disaster, about 47,300 m3 of timber was logged in the mountainous areas of the Czech Republic, i.e., a total of 2% of the Czech forest area, primarily in the Krušné hory, Lužické hory, Jizerské hory, Krkonoše, and Orlické hory Mountains (Vacek et al., 2003).
Drought, coupled with rising temperatures, particularly in recent years, has been identified as a contributing factor to increased salvage logging in spruce stands (Netherer et al., 2019). The multiple linear regression analysis in Table 4 underscores seasonal temperature as a significant factor (p = 0.03) influencing salvage logging. In contrast, there is no apparent link between seasonal precipitation and salvage logging as indicated by the multiple linear regression analysis. One possible explanation for this could be the upsurge in salvage logging following windstorms and bark beetle overpopulation in deciduous forests. The process of salvage logging, in turn, modifies the micro-climate of the plot (Peterson & Leach, 2008). Elevated temperatures contribute to more frequent bark beetle swarming, leading to additional generations within a single season (Marini et al., 2013). Consequently, this escalation in salvage logging is reflected in the overall logging of spruce timber (Marini et al., 2017).
Precipitation is linked with tree-ring growth, especially at lower altitudes in the 1_KAR and 2_KOS research plots, that faced more frequent drought periods in this region and are (Fig. 4) accompanied by low RWI (Fig. 5) in recent years. At lower elevations, Norway spruce is more negatively affected by irregular precipitation during the vegetation season than at higher altitudes (Rybníček et al., 2012), which significantly increases salvage logging in the Czech Republic (MAF, 2021). During drought periods, Ips typographus reproduce excessively and infest healthy trees too (Schroeder & Lindelöw, 2002; Turčáni & Hlásny, 2007). It is a necessary time to remove infested epicenters from the forest before the bark beetle increases its numbers and attack other mostly nearby trees (Økland & Berryman, 2004).
The challenges surrounding timber extraction in the Czech Republic are exacerbated by the aging of forest stands. Thanks to the Forestry Act in the 1990s, the rotation period of forests in the Czech Republic was extended. This resulted in forests aging from 102 years in 1960 to 114 years in 2018. It is precisely this increase in the age of forest stands that made them more susceptible to climate fluctuations (Šimůnek et al., 2020b). As forest stands age is higher, the occurrence of windthrow and bark beetle infestations becomes more frequent. Research has shown that the risk of disturbances in these stands increases with the higher age of forest stands (Nascimbene et al., 2013). This is further supported by the observed higher salvage logging activities in the late 20th and early 21st century (Fig. 2a). Therefore, salvage logging is continuing higher also in percentage (Fig. 5b).
Furthermore, the complexity of 20th-century calamity dynamics is compounded by the change in ownership of smaller forest properties, which were restituted during the 1990s (Fig. 1c). Many of these smaller properties are owned by individuals who cannot effectively harvest and manage broken and infested wood. As a result, the populations of bark beetle pests proliferate, gaining strength and causing further damage (Zahradník & Zahradníková, 2019; MAF, 2021). The inability to process harvested timber from salvage logging time is evident in our data regarding the percentage of salvage logging. Before 1995, there were higher frequencies of wood processing, but, for example, since 2000, the calamities have been longer and much more intense. The inability to process wood is also related to the reduced number of employees in forestry, which only complicates the processing of random timber harvesting (Seidl et al., 2016).
4.2 Spruce tree rings and timber harvest comparison to the solar cycle
The sunspot cycles represented by the sunspot number and TSI observed link with all logging types from 1961 to 2019. Multiple linear regression analysis of our results (Table 4) showed a notable link between the percentage of spruce logging with a sunspot number. The PCA in Figure 6 revealed a strong positive link between RWI and the studied types of timber harvests, as well as TSI and sunspot number. A recent study described this phenomenon in the Czech Republic, where salvage logging observed correlation (p < 0.05) with the sunspot area (Šimůnek et al., 2020b). Solar cycles have also been confirmed in other European locations on the annual rings of other important forest tree species, such as Scots pine (Pinus sylvestris L.) – (Shumilov et al., 2011; Matveev et al., 2017; Kasatkina et al., 2019) or European beech (Fagus sylvatica L.) (Šimůnek et al., 2020a; Komitov, 2021). We hypothesize that increased solar activity, characterized by a higher sunspot number and larger total solar irradiance (TSI), enhances tree growth (as reflected in increased RWI), which may subsequently influence bark beetle infestations and trigger changes in logging practices.
While the impact of TSI is significant for tree rings and timber harvests, particularly influencing RWI in research plots (see Fig. 4), it is intricately connected with the sunspot number. The connection between sunspot number and TSI is complex when compared to various solar parameters indirectly linked to tree rings. TSI itself undergoes a 0.1% variability (approximately 0.08 ± 0.2 K) over an 11-year cycle, reaching its peak during solar maxima (Kopp, 2021). As the sunspot number increases, so does the TSI, resulting in a higher level of UV radiation. However, the effects of the solar cycle are multifaceted, involving the broader solar spectrum. Notably, the entire light spectrum varies with the solar activity cycle (Tsiropoula, 2003; Yeo et al., 2017). To comprehend this intricate interplay between solar dynamics and our climate indicators, it is crucial to explore the combined influence of TSI, spectral variations, and the many other solar components.
Our results of multiple linear regression (Table 4) suggest an important role (b* = −0.59 at p < 0.04) of TSI with the percentage of salvage. Thus, TSI has slightly a greater effect on RWI and types of loggings than sunspot number according to the Figures 4 and 6 (PCA). The TSI with RWI can be linked to the influence of UV radiation of the Sun, which also changes throughout the solar cycle (Floyd et al., 2002, 2003). The UV (more precisely UV-B) radiation can influence tree ring growth by inducing oxidative stress in trees, which can lead to narrower tree rings (Madronich, 1993). Additionally, increased UV-B radiation can stimulate the production of secondary metabolites like lignin and flavonoids in response to stress, potentially affecting ring patterns (Ballaré et al., 2001).
Our PCA results for RWI also indicate the same variability with respect to the sunspot number and TSI. The effect of TSI on climate has been indirectly associated with the South Asian summer monsoon cycle, which is imprinted in tree rings (Shi et al., 2014). Our results also revealed that TSI and temperatures are associated with RWI, where according to the PCA (Fig. 4), the influence of temperature on RWI increases with increasing altitude. The impacts of irradiance and radiation variations on spruce tree rings are significantly higher at mountain altitudes compared to spruce trees in lowlands (Černý et al., 2020). This is also supported by results from regression analysis where even the percentage of salvage logging observed an important link with TSI and temperature while the main portion of the salvage loggings are from middle and high altitudes. There is also the assumption that temperature is less influenced by the solar cycle influences than precipitation in North-Eastern Europe (Chapanov & Gorshkov, 2019). Therefore, there is an inclination to the fact that the temperature initiates higher timber harvests, which is a long-term trend longer than the solar cycle. We suggest that the presence or absence of precipitation – whether in the form of drought or wet periods – has an immediate impact on salvage logging, resulting in either a decrease or increase in the volume of harvested timber. Consequently, we hypothesize that the patterns of these dry and wet periods affecting the health status of the trees are likely linked to some mechanism associated with the solar cycle.
According to the SEA shown in Figure 7, the spruce research plot’s RWI has mostly the higher variable growth around solar maximum, indicating that the tree-ring series growth was associated with the solar cycle through indirect climate effects. However, during solar minima, the lowest RWI at Lag −1 is observed mostly across all research plots, and a synchronized gradual decline in RWI is mostly observed from Lag −5 onwards across all research areas. This may indicate that the average RWI is influenced by a climatic feedback mechanism associated with the solar cycle. Similar results between observed lags of sunspot number and RWI were found in European beech in the Krkonoše Mountains and Broumov region in the northeast Czech Republic (Šimůnek et al., 2021b). These observed lags are likely to be associated with the NAO, which influences precipitation and temperatures in Europe (Wibig & Piotrowski, 2018; Kotsias et al., 2020). Our results show that, among the considered factors, NAO contributes to RWI to the least extent according to PCA (Fig. 4). However, in the PCA (Fig. 6) for timber harvesting, it is evident that seasonal NAO may play a role in climate-related factors influencing timber harvesting while the multiple linear regression analysis did not confirm any strong evidence.
Before the peak of percentage spruce harvesting and percentage salvage logging, all RWI plots, according to SEA analysis, exhibited their lowest values mostly at Lag −1, which is one year prior to the solar minimum. However, during the solar maximum, significant long-term calamities were not recorded, and even in the recent 30 years, the interval between calamities has extended further according to the 11-year cycle. This difference between RWI and timber harvests is due to the chronic weakening of spruce stands, which is indicated by higher salvage logging due to increased vulnerability for forest stands (Kolář et al., 2017) This low tree-ring growth can be an early warning indicator of extensive spruce stand disintegration along with other indicators such as the abundance of Ips typographus populations (Marini et al., 2017).
The solar cycle is firstly observed in the tree-ring growth of spruce, and then, with a 1-year delay, will manifest itself in the percentage of salvage logging while RWI in the solar maximum is not significantly stressed. According to the multiple linear regression analysis, a significant relationship was even found between sunspot numbers and the percentage of spruce logging, indicating that logging activities can be linked to solar cycle related changes in temperature and precipitation in the Czech Republic. Some studies describe bark beetle infestations in relation to solar radiation and its effect on the habitat, especially during the peak and decline phases of the beetle outbreak (Mezei et al., 2019). This link between TSI and calamities may relate to a higher infestation of the forests, which slightly prolongs the vegetation season effect on higher tree-ring growth. A mostly one-year preceding Lag indicated low tree rings to sunspot minimum and with a maximal salvage and spruce logging is something new. This Lag effect may be due to sudden stratospheric warming events (SSWs), altering stratospheric circulation and affecting the jet stream, storm paths, rainfall, and temperatures. During solar minimum/easterly QBO years, midwinter SSW frequency slightly increases (0.67 per year at 90% confidence). However, study observations alone cannot confirm the robustness of the solar-QBO-SSW relationship (Baldwin et al., 2021). The apparent observation of a solar cycle in RWI and timber harvests may be a false attribution to the solar cycle due to the sparse database of a few solar cycles. Even, artificial (false) solar fluctuations correlations and coherences may be even higher than, those generated by genuine solar variability in for example we found that this was observed in QBO phases (Salby & Shea, 1991).
Solar activity can induce changes in temperature and pressure, influencing the dynamics of the atmosphere and correlating with fluctuations in the NAO. These similarities can also be identified over decadal periods, with a Lag of a few years relative to the solar cycle. Some studies emphasize that solar variability can impact atmospheric circulation, interacting with pressure systems and manifesting in decadal NAO signals (Scaife et al., 2013; Gray et al., 2017; Ye et al., 2023). Overall, it can be concluded that there is a connection between solar variability and decadal changes in the NAO over Europe, and from our data used in Figure 2, the coincidence of fluctuations between the NAO and the sunspot number can be visually observed, mainly since the 1980s until the present.
The NAO phases are probably an important reason why we can observe such cycles on tree rings and types of timber harvesting (Fig. 6). The NAO varies between positive and negative phases, which can be identified as potential drivers affecting forest growth and timber harvesting in Central Europe. The NAO, with its distinct phases, can introduce varying weather patterns that influence temperature and precipitation. These variations can be observed in tree-ring data, influencing the radial growth of trees. The NAO phase is linked to the formation of distinct climate patterns and atmospheric pressure systems, which can influence tree-ring series in the region. The complex interaction between NAO phases, the solar cycle, and their effects on tree-ring series and timber harvesting has been the subject of extensive research (Jones et al., 1997; Piovesan & Schirone, 2000; Trouet et al., 2009; Leonelli et al., 2017). The connection between the NAO, the solar cycle, tree-ring series, and timber harvesting in Central Europe is a multifaceted relationship, and understanding these interactions contributes to a more comprehensive insight into the region’s forest ecosystems.
4.3 Cyclical presence of the solar cycle on spruce from rings to timber harvest
The results of the spectral analyses (Fig. 8) recorded high similarity to the 11-year cycle for almost all the time series from RWI to different types of timber harvest. There is an assumption that solar-induced changes in the atmosphere might be the reason of observed results, but the solar cycle does not always last the same length of time (Chiodo et al., 2012, 2019; Scafetta, 2012; Hathaway, 2015). The solar cycle is present in the RWI from all research plots, as shown by the results of PCA from 1900 to 2019 (Fig. 4). All harvest types also show an 11-year cycle in the data, which is most evident in the salvage logging (Fig. 8). And it is the salvage logging that take place concurrently to the solar cycle (Šimůnek et al., 2020b), which is confirmed by the results with percentage of salvage logging. Norway spruce harvest indicates common fluctuation trend to the solar cycle, but the concurrence is not as clearly visible as in the case of salvage logging. In our results, the solar cycle manifests in various forms, ranging from tree rings to timber harvesting. However, it is essential to note that these signals may not be solely attributed to a pure solar source.
The presence of the solar cycle in our data is influenced by climatic factors that are likely remotely linked to the NAO. Although this phenomenon was not statistically supported by regression analysis, the tree-ring series, as analyzed through PCA, shows a relationship with the seasonal NAO. The alternation of NAO phases in Central Europe is significantly impacted by the solar cycle. During the positive phase of the NAO, there is a higher occurrence of cloudy days, reducing solar radiation (Foukal et al., 2006). Conversely, during the negative phase of the NAO, there are typically more sunny days, promoting increased solar radiation and a positive impact on the solar cycle (Lockwood et al., 2014). Therefore, we argue that the influence of TSI can be observed in our datasets.
Additionally, the results related to the percentage of spruce timber logging are significant to sunspot number (Table 4). Spruce timber logging also follows a longer, 58-year cycle, partially visible in Figure 8, which is consistent with the natural growth and commercial rotation of Norway spruce, which undergoes these cyclic disturbances (Mikol et al., 2020; Vacek et al., 2021). Before the 1990s, the cyclical rotation of spruce forest harvesting in the Czech Republic was around 80 years, but then it was increased to almost 115 years of the age of the forest stands, which is partly reflected in the trend of total and spruce timber harvesting (Remeš et al., 2020). Medium-term 25-year cycles are observed for total, spruce timber, and coniferous timber logging (Fig. 8). In terms of numerical similarity, this cycle could be related to the 22-year Hale cycle, which recurs through meteorological indicators such as precipitation (Laurenz et al., 2019) and temperature (Lüdecke et al., 2020). Similar cyclical results to our study were also recorded in the tree-ring growth of Scots pine in the western Russian steppe, where 10–12-year, 22-year, and 32–36-year cycles were recorded. This 35-year cycle can be defined according to the Brückner-Egeson-Lockyer cycle, which is associated with climatic fluctuations that are documented in climatic indicators (Halberg et al., 2010). There were even 70–90-year fluctuations in the moisture regime (Matveev et al., 2017).
In the conditions of the Czech Republic, the solar cycle manifests itself first on spruce that has been weakened by drought and cyclical worsening of growing conditions, which is well shown in Figure 5, where the RWI of spruce at the research plots reacts cyclically with the solar cycle. During periods of poor growth conditions, spruce trees are physiologically affected by higher temperatures and lack of precipitation (Rybníček et al., 2010; Tumajer et al., 2017). It is related to the solar cycle through variability in the irradiance at the Earth’s surface (Dasi-Espuig et al., 2016; Kopp et al., 2016), lack of precipitation during the growing season (Laurenz et al., 2019), and fluctuations in temperatures (Lüdecke et al., 2020). All of this is related to air circulation, manifesting in windy calamities. At the same time, there is a gradual overpopulation of bark beetles which increases salvage logging (Šimůnek et al., 2020b).
Our results show a repeatable pattern similar in the RWI and timber harvests to the 11-year solar cycle on the seasonal weather regime that is probably linked with drought period in which trees are sensitive. River Ammer (southern Germany) flood frequency variability and summer extreme events were observed in seasonal data in central Europe, which can be explained by a solar modulation of eastern European-western Russia summer blocking and associated upstream upper-level wave-breaking activity (Rimbu et al., 2021).
4.4 Opportunities and challenges in the declining of spruce forests: understanding limitations and potential outcomes
Wind, bark beetle, and air pollution disturbances primarily increase salvage, total, spruce timber, and coniferous timber logging in the Czech Republic (Hlásny & Turčáni, 2013; Mitchell, 2013; Vacek et al., 2013; Hlásny et al., 2014). The substantial increase in percentage (Fig. 7) of Norway spruce logging grows from a mean of 69% (range 68–71%) to a mean of 77% (range 73–81%.) and the percentage of salvage logging is from mean 35% (range 22–42%) to 60% (range 47–69%) from the solar maximum to solar minimum. In Figure 3, our data describe that spruce is cyclically stressed, as apparent in the concurrent low RWI. Increased spruce logging follows, and simultaneously, the salvage logging increases too (e.g., around 2007 or 2019). The rise in spruce timber logging exerts significant pressure on the forestry sector, which is forced to sell high volumes of timber within a short period of time. This destabilizes market prices of timber (Jandl, 2020; Hlásny et al., 2021b).
Coniferous forests are consequently susceptible to a range of secondary diseases and pests and are particularly vulnerable to warmer and drier climates (Seidl et al., 2008). The overall situation is exacerbated by the lower quality of timber, damaged by windthrows and bark beetles (Kärhä et al., 2018; Netherer et al., 2019). However, the decline of spruce forests provides more space and support for European beech, which is more adaptable to climatic fluctuations than Norway spruce (Kolář et al., 2017; Vacek et al., 2019, 2021). Considering that salvage logging and timber harvesting occur in cycles similar to the eleven-year solar cycle, we suggest these cycles might affect the wood industry. Given the impact of these cycles on forestry, we could use this knowledge for long-term planning in the timber and forestry sectors.
However, the tree-ring growth does not always coincide with the solar cycle. This is apparent from 1900 to 1919 when the trees were not stressed due to a lack of precipitation (Fig. 2c). During this period, there was a more stable distribution of precipitation and temperature across the Czech Republic, particularly during the vegetation season. During this time, the RWI does not exhibit a clear solar cycle pattern in nearly all plots. The RWI has notably decreased since 1980 at the 6_BAZ research plot during the human-induced air pollution period.
Internal decadal variability of the climate system, incorporating TSI, sunspot number, and the NAO, can exert significant influence. For example, TSI is associated with sunspots, affecting temperature patterns (Usoskin et al., 2016). Simultaneously, NAO has the potential to influence regional climate by impacting precipitation and temperature distributions, particularly in Northern Europe. Higher temperatures and altered precipitation regimes can affect tree growth, showing a positive response in certain circumstances (Carrer & Urbinati, 2004). Nevertheless, our data display a lag in RWI response, like a reduction a year before the maximum timber harvests, possibly attributed to the influence of NAO (Helama et al., 2009). The NAO, by affecting moisture availability, can either dry or moisten the climate in Central Europe, creating conditions unsuitable for tree growth (Cook et al., 2010).
5 Conclusion
Norway spruce harvest is one of the primary forestry activities in the Czech Republic. It becomes apparent that climatic fluctuations greatly affect spruce timber harvesting, as do cyclical calamities. The ring-width indices of Norway spruce from six research plots indicate a relationship between timber harvesting and its RWI. The types of loggings are greatly influenced by temperatures, seasonal NAO, and TSI. Regarding the percentage of spruce logging, the sunspot number was the highest indicator on the multiple linear regression analysis. Percentage of salvage logging is significantly linked to TSI and seasonal temperature. The average RWI series reveals predominantly low tree ring growth with a lag time of one year prior to the occurrence of the maximal timber harvests. Timber harvests are higher during Lag 0 with the solar minimum when spruce and salvage logging typically peaks and then decreases to solar maximum. Spruce logging and tree-ring series show a close yet not fully and precisely described relationship with the 11-year solar cycle. The results of this study can help to understand the cyclical relationships between spruce tree rings and the management of this tree species. A comprehensive view of cyclical fluctuations in forestry can assist forest management to better prevent unexpected calamities in the future and thus reduce the impacts of climate change. The findings of this study can be applied to improve action against major forest calamities in Central Europe. Nevertheless, a more in-depth investigation of this topic is essential for its future application.