Faculty of Science and Forestry
DROUGHT SENSITIVITY OF SILVER FIR AND SCOTS PINE OVER THE 20
thCENTURY IN NORTHERN CZECH REPUBLIC
Thaís da Silva Reis Prado
MASTER’S THESIS FORESTRY (CBU) ________________________________________________________________
JOENSUU 2019
Reis Prado, Thaís. 2019. Drought sensitivity of Silver fir and Scots pine over the 20th century in Northern Czech Republic. University of Eastern Finland, Faculty of Science and Forestry, School of Forest Sciences. Master of Science in Agriculture and Forestry with specialisation in Forest Ecology and Silviculture, 36 pp.
ABSTRACT
The influence of climate change on natural disturbances of forests, such as drought, has been increased over the past two centuries. Therefore, it is important to understand how the tree growth responds to the increase in natural disturbances such as drought events and their severity. The main aim of this thesis was to analyse in a natural forest landscape in Suchý vrch Nature Reserve, in Jeseník, Czech Republic, the drought sensitivity of Silver fir (Abies alba Mill.) and Scots pine (Pinus sylvestris L.) over the 20th century. This was done by correlating the dendrochronological data from both tree species with climate data, using the Palmer Drought Severity Index (PDSI), the Standardized Precipitation-Evapotranspiration Index (SPEI), temperature and precipitation, respectively. The dendrochronological data and climate data used in this study covered the period of 1900 to 2010. The data analysis was conducted using R software. First, the dendrochronological data was detrended using three different methods: Spline, Modified Negative Exponential curve and Autoregressive model residuals (Ar), in order to provide the mean chronologies for tree growth. Secondly, the detrended tree growth data (the mean chronologies) were correlated to the climate data. The results showed that the Spline detrending method was the best approach to discern disturbances noise and climate signal. Furthermore, the first half of the 20th century showed a different growth trend than the second half for both tree species. Based on this study, Silver fir responded, in general, more to the climate variability than Scots pine did. However, also the atmospheric pollution (e.g. nitrogen deposition) likely affected the observed variation in growth over the years in addition to drought and temperature variability.
Keywords: climate change, dendrochronological data, drought, Pinus sylvestris, Abies alba, tree growth
Acknowledgements
It is time to remember all the people that in some way have been involved in this adventure of doing a master´s at the CBU programme, at the School of Forest Sciences, University of Eastern Finland.
First, I would like to express my gratitude to my supervisors, Dr. Miloš Rydval from the Czech University of Life Sciences, Prague, and professor Heli Peltola from the School of Forest Sciences, University of Eastern Finland, Joensuu campus. Your guidance has helped me to improve and finish my thesis. I am also very grateful for Kristyna Svobodova for her assistance, as she has helped and encouraged me on every step since the beginning of the master´s thesis idea.
The decision of leaving Brazil to study abroad would not be possible without the support and care of my family and friends (the old and the new ones). Thank you so much for being by my side at any circumstance. I am blessed to have all of you in my life! Special thanks to Edu, who has always been backing me up and encouraging me to do my best.
And finally, thank you to whom I believe my capacity and motivation come from every day, thank you God for always being there for me.
May 2019, Joensuu, Finland Thaís da Silva Reis Prado
Table of Contents
1. INTRODUCTION 5
1.1 Influence of climate change on drought, growth and mortality of trees 5 1.2 Measurements, crossdating and detrending of dendrochronological data 6
1.3 Calculation of drought indices 9
1.4 Research objectives 11
2. MATERIALS AND METHODS 12
2.1 Study area and location of sampled trees 12
2.2 Preparation of dendrochronological data 13
2.3 Detrending and building of chronologies 15
2.4 Climate data 16
2.5 Calculation of tree growth–drought relationship 17
3. RESULTS 18
3.1 Chronology development 18
3.2 Detrending methods 19
3.3 Correlations between growth and climate data 22
4. DISCUSSION 25
4.1 Evaluation of the detrending methods used 25
4.2 Evaluation of growth-trend based on the chronologies 26
4.3 Evaluation of climate tree-growth relationships 27
5. CONCLUSION 29
References 30
Appendix 1 34
Appendix 2 35
Appendix 3 36
1. INTRODUCTION
1.1 Influence of climate change on drought, growth and mortality of trees
The influence of climate change on natural disturbances of forests has been recognized at least for the past two centuries (Schelhaas et al. 2003). The ongoing climate change has already resulted in temperature increase and changes in precipitation regimes (IPCC 2013). Along with climate change, the frequency and severity of different natural disturbances are expected to increase in the future (Dale et al. 2001; Seidl et al. 2017). Climate-driven tree mortality is considered as one of the major climatic risks under a warming climate. This is due to the combined impacts of drought and increasing temperatures on tree mortality (Allen et al. 2010, 2015). Droughts are also expected to become more frequent and severe in the future, which may become lethal for some plants under warming conditions (McDowell et al. 2008; Allen et al. 2015). The stress and the risk of tree mortality during a drought event may increase both directly via physiological impacts (Adams et al. 2009) or indirectly via pathogens and pests’
effects (Weed et al. 2013).
Additionally, atmospheric pollution may influence the dynamics and growth of forest ecosystems in many regions. The levels of atmospheric pollution, for example, in parts of central Europe, have been during the second half of the twentieth century among the highest on the continent (Rydval & Wilson 2012). For example, the Sudetes, a mountain range in Central Europe, was under strong stress caused by air pollution in the second half of the 20th century (Alcamo 1987; Molski 1987). The fast development of the power industry took place in the area situated west of the Sudetes, in the early 1960s, where large mining and electric power centres were created, such as Turów in Poland, large parts of northwest Czech Republic, and Hagenwerder in Germany. Brown coal with high concentrations of sulphur was used in the electric power stations and as the energy production increased, so did the emission of sulphur dioxide (SO2), nitrogen oxides (NOx), and dust. The consequences of air pollution could be seen mostly in the Sudetes, due to the predominance of westerly winds (Koźmiński et al. 2001).
Some previous studies have shown that many tree species experienced a strong decline in growth, which often induced also tree death. At the higher elevations from the sea level, the
stands suffered most, because they were there more directly exposed to air pollution (Miś and Rączka 2002; Feliksik and Wilczyński 2003).
Mountain forests, including temperate forests, are nowadays experiencing changes in tree species distributions as well as upslope elevational shifts, including greater mortality of trees at lower-elevation range limits (Allen et al. 2015). A diverse suite of environmental processes and ecosystem services are fundamentally affected by broad-scale tree mortality, even in places where tree mortality does not cause species range changes or biome shifts (Aber et al. 1995;
Breshears et al. 2011).
1.2 Measurements, crossdating and detrending of dendrochronological data
Dendroclimatological studies may help to understand the influence of tree age, altitude and site conditions on the climate sensitivity of tree growth in different tree species (Babst et al. 2013).
In addition, the establishment of regional-scale networks may enable to link spatial and species- specific growth response patterns with thermal and moisture variations within a particular climatic zone (Frank & Esper 2005). However, it is needed to distinguish the climate signal and noise from disturbances when studying climate sensitivity of growth (e.g. using tree ring widths) in forests (Rydval et al. 2018).
The growth of trees is a result of several causal factors acting within the tree and within the tree´s environment and depends on the characteristics of the tree species, as well as on the impact of many external factors. Weather conditions usually have a short-term effect on the growth of trees (Fritts 1976). A relatively consistent growth rhythm, mainly subject to climate conditions, can be determined by healthy and dominant trees in stands, although their natural rhythm of growth can be negatively affected by a range of environmental disturbances (Ermich et al. 1976; Spiecker 2002). The tree ring-widths can vary with fluctuations in environmental conditions, but also due to tree age (from pith to bark). Therefore, the age-related influence should be removed from tree ring measurements before further analyses of climate sensitivity of tree growth (Fritts 1976). For the dendrochronological data, the tree ring samples are first pre-prepared in the laboratory, and thereafter the data is properly measured and cross-dated before detrending methods are applied to the tree ring width measurements. The climate data
are needed for the location and period of interest, and besides temperature and precipitation, drought indices are often used to quantify the influence of drought.
The standard practice regarding dendrochronological data preparation in laboratory conditions consists of preparing the wood by drying the tree cores, mounting the dried cores on wooden frames, and shaving or sanding the cores until the tree-rings become clearly visible. Afterwards, tree-ring widths are measured using a stereomicroscope and measuring stage (Speer 2010). The method of Duncan (1989) can be used to estimate pith offset for cores that do not contain the pith, in order to obtain tree age estimation. The next step involves crossdating the samples, which is a standard dendrochronological approach that can, for instance, be conducted with CDendro software (Larsson 2015). Crossdating is vital in assuring that each tree ring can be absolutely dated and associated with a specific calendar date (Fritts 1976). An error can be easily produced by a simple ring count without crossdating, due to missing rings and false rings, causing the chronology to be misdated by one or more years, for instance (Speer 2010).
Detrending involves the estimation and removal of the tree’s natural biological growth trend.
The standardization includes the division or subtraction of each tree ring measurement series by the growth trend to produce detrended series expressed by dimensionless ring-width index (RWI) units and then averaged in order to produce what is known as a mean chronology (Cook and Kairiukstis 1990; Fritts 2001). One factor alone can severely affect the growth of trees, although most likely, there are multiple factors affecting their growth. The Aggregate Tree Growth Model (Rt, see Cook 1985, 1992) is used in order to conceptualize the tree response and to understand the influence of different variables on tree growth, i.e. Rt = f(Gt, Ct, D1t, D2t, Et), where Rt is ring width at year t, Gt is the age (or size)-related growth trend, Ct is climate in year t, D1t is the endogenous disturbance within the stand, D2t is the exogenous disturbance from outside the stand, and Et is the error term incorporating all the signals that are not controlled by the above variables.
The Aggregate Tree Growth Model demonstrates that for each year the ring width growth is influenced by a complex range of variables. Also, trees have an inherent age-related growth trend, they react to current climate conditions, their growth in a particular year can be influenced by previous year´s climate, and they can also be influenced by disturbances (originating from within or outside of the stand). The error term incorporates any unexplained variability (Speer 2010). The standardization is considered to be a powerful technique used to minimize the noise
in a chronology and to increase the signal of interest, meaning that it can be used to remove the influence of disturbances and competition, and biological age-trends (Fritts 1976; Speer 2010).
Different methods can be used to obtain the ring-width indices. The common methods used for dendrochronological data, for instance, are the smoothing spline (Spline), modified negative exponential curve (ModNegExp), and the residuals of an autoregressive model (Ar) (Cook and Kairiukstis 1990; Fritts 2001).
A cubic smoothing spline is an example of an empirical model, it uses a flexible curve that can adjust at a regular interval (Cook 1985). The Spline method uses a spline where the frequency response is 0.50 at a wavelength of 0.67 multiplied by the “series length in years”, or it can be differently specified. It attempts to remove the low-frequency variability due to biological or stand effects (Cook and Kairiukstis 1990; Fritts 2001). Splines are a more organic fit to the data than a straight line or exponential fit, but they do remove different amounts of variance at different temporal scales. This method eliminates the potential multi-decadal and longer climatic and non-climatic trends, but it preserves the year-to-year and decadal variability that provides the most robust analytical base for correlation analyses (Speer 2010).
The modified negative exponential curve (ModNegExp) method aims to fit a classic nonlinear model of biological growth (f(t) = a exp (b t) + k), where the argument of the function is time.
If the function is non-decreasing or some values are not positive, a nonlinear model cannot be fit, in which case a linear model is used; many researchers are particularly sceptical of the use of a classic nonlinear model of biological growth for detrending, the ModNegExp method can be useful when a more rigid detrending approach is needed and for example when a consistent form of the detrending curve is desired (Cook and Kairiukstis 1990; Fritts 2001). The negative exponential curve is deterministic, meaning that it follows a model of tree growth (Cook 1985).
The Ar method or pre-whitening approach is the residual of the Ar model fitted to the original tree ring series, which are then averaged. It removes all but the high-frequency variation in the series and therefore only represents interannual variability (Cook and Kairiukstis 1990; Fritts 2001).
Descriptive statistics, including mean series intercorrelation (RBAR), and expressed population signal (EPS), are also calculated for the mean chronologies (Briffa and Jones 1990). The mean
chronology is the average of the standardized indices of individual trees for a sampled site (Fritts 1976). The RBAR represents the mean pairwise correlation coefficient among the tree- ring series (Wigley et al. 1984). The running RBAR examines the signal strength throughout the chronology, it is a good measure of the common signal strength through time (Cook et al.
2000).
The EPS assesses the degree to which the chronology, based on a finite number of trees, represents a hypothetically complete chronology (Wigley et al. 1984). If the EPS value descends below a predetermined level, it means that the chronology is starting to be influenced by individual tree-level signal rather than a coherent stand-level signal. The value of 0.85 has been traditionally used as an appropriate cut-off point in assessing chronology robustness (e.g.
Wigley 1984; Speer 2010).
1.3 Calculation of drought indices
The hydrologic cycle intensification is expected in a warmer world (Dore 2005). The literature shows a variety of drought indices developed to quantify the different aspects of drought.
However, no single index can describe all the spatiotemporal characteristics of drought, be applicable in drought research in different regions, and evaluate the influence of drought on the environment and society. In addition, distinct indices might yield different results for an identical drought event. For instance, the trend of global drought remains controversial because of the use of different drought indices (Sheffield et al. 2012; Dai 2012). In general, different drought indices focus on different aspects or physics of drought and their calculation requires different variables (Yang et al. 2017). The two drought indices, Palmer drought severity index (PDSI) and the Standardized Precipitation Evapotranspiration Index (SPEI), are commonly used in research involving drought and are considered as a good drought index for use in studying climate change influences (Vicente-Serrano et al., 2012, Wang and Chen 2012, Yang et al 2017).
The Palmer drought severity index (PDSI), defined by Palmer (1965), has become the most widely used meteorological drought index for the purpose of drought monitoring and research in the United States (Alley 1984). According to Fuchs (2012), the PDSI is very effective in
measuring the impacts sensitive to soil moisture conditions, such as in agriculture production.
Moreover, it is also used as a drought monitoring tool to start actions on drought contingency plans. Typically, it is monthly computed, but in modified versions, it can be computed on a weekly timescale.
Drought is defined by the PDSI as a period of months to years that is characterized by precipitation lower than climatically appropriate precipitation values for existing conditions, also called as CAFEC values. The PDSI incorporates previous and current moisture supply and demand into a hydrological model, including precipitation, soil moisture, streamflow and potential evapotranspiration (PET). The differences between the actual precipitation and CAFEC values, based on a model of physical water-balance, are computed and scaled with empirical parameters to obtain PDSI values (Yang et al. 2017). Some limitations include that PDSI does not perform well in regions with variability extremes of rainfall or runoff and all precipitation is treated as rain, which means that snowfall, snow cover, and frozen ground are not included in the index; as a consequence the timing of PDSI values may be inaccurate in snowy regions in the winter and spring months. The Self-Calibrated Palmer Drought Severity Index (scPDSI) calibrates automatically the behaviour of the index at any location, as a result, the conditions of any climate should be realistically represented by the index within the definition of the PDSI (Fuchs 2012).
The Standardized Precipitation Evapotranspiration Index (SPEI) was proposed by Vicente- Serrano et al. (2010, 2012), to overcome the issue of another drought index, the Standardized Precipitation Index (SPI), which is unsuitable for use in the current global warming circumstance and future climate scenarios (Yang et al 2017). The SPEI uses the difference between monthly precipitation and evapotranspiration instead of monthly precipitation anomalies; this drought index combines the PDSI sensitivity to changes in evapotranspiration demand and the SPI multiscalar nature. Consequently, drought has been detected in many areas by using this drought index. The time scale can vary from 1 to 48 months on SPEI values (Vicente-Serrano et al., 2012; Wang and Chen, 2012). The SPEI also has some limitations, it is sensitive to the method used for calculating the potential evapotranspiration and it requires the use of long base period samples, at least 30 to 50 years of natural variability data (Vicente- Serrano & National Center for Atmospheric Research Staff, 2015).
1.4 Research objectives
The characterization of current and past drought-growth relationships can contribute to a better understanding of possible climate change consequences on future forest growth. With this in the mind, the aim of this thesis was to analyse the drought sensitivity of Silver fir (Abies alba Mill.) and Scots pine and (Pinus sylvestris L.) over the 20th century by correlating the dendrochronological data from both tree species with climate data, using the Palmer Drought Severity Index (PDSI), Standardized Precipitation-Evapotranspiration Index (SPEI), temperature and precipitation, respectively. The specific steps of the research were:
i. Dendrochronological data preparation,
ii. Use different detrending methods (smoothing Spline, modified negative exponential curve and the residual of an Ar model) to develop the chronologies for each tree species and thereafter use the most relevant method for the further correlation analysis,
iii. Compilation of monthly climate data, including mean temperatures, the sum of
precipitation, SPEI12 values, and PDSI values over the 20th century for the study area location,
iv. Correlation between tree growth (tree ring width data) and climate data, using PDSI, SPEI, temperature and precipitation.
2. MATERIALS AND METHODS
2.1 Study area and location of sampled trees
The dendrochronological data was obtained from Suchý vrch Nature Reserve, a Protected Landscape area located in Jeseník District, in the Olomouc Region, the northern area of the Czech Republic (Figure 1). The mean coordinates of the area were 17,347E and 50,156N and the altitude of the study area was in the range of 690 to 940 m above sea level. The climate was cold and temperate, with significant rainfall (classified as humid continental climate by the Köppen-Geiger, 1954). The Nature Reserve belongs to the Jeseník geomorphological area and the whole Hrubý Jeseník (High Ash Mountains) a mountain range of the Eastern Sudetes;
characterized by sea stones, relics of exposed habitats, and stone forests on slopes, with relatively homogeneous stand domination of spruce.
Figure 1. Location of the Jeseník District in the Czech Republic area.
The sampled trees were selected, focusing on old scattered pines and firs; the pines were located on exposed sites near the peak and stone debris, especially on the western slope, while the firs were located on less exposed stands of stone fir and beech trees (Figure 2). The mean DBH was for Silver fir and Scots pine 401 and 363 mm, respectively.
Figure 2. Suchý vrch Nature Reserve and location of the sampled trees.
2.2 Preparation of dendrochronological data
The preparation of dendrochronological data was performed according to standard practice, which means that tree cores were air-dried, mounted on wooden frames, and shaved with a razor blade until clear visibility of the tree-rings was achieved. After the preparation stage, the tree-ring widths were measured using a stereomicroscope and a LintabTM transversing stage measuring device coupled with TSAPWinTM software (Rinntech, Heidelberg, Germany). The method of Duncan (1989) was used to estimate missing rings for cores that missed the pith, in order to obtain tree age estimation. Standard dendrochronological approaches were used for crossdating the samples, conducted with CDendro (Larsson 2015).
The data include 40 tree-ring series of fir and 48 tree-ring series of pine (Figures 3 and 4). The last year for tree-ring series was 2016. A few series started from the beginning of the 17th century.Full chronology spam ranged from 1588 to 2016 for fir and from 1605 to 2016 for
pine.Corresponding ranges for chronology spam with EPS greater than 0.85 were 1831 to 2016 for both tree species.
Figure 3. All 40 series of Abies alba (Silver fir).
Figure 4. All 48 series of Pinus sylvestris (Scots pine).
2.3 Detrending and building of chronologies
The dendrochronology program library (dplR) package (Bunn 2008) in R software (R Core Team 2017) was used for chronology building and detrending. Descriptive statistics, including mean series intercorrelation (RBAR) and expressed population signal (EPS), were calculated
for Silver fir and Scots pine mean chronologies (Briffa and Jones 1990). The chronologies were built by aggregating tree-ring series using Turkey´s robust bi-weight mean (Cook and Kairiukstis 2013). This averaging approach appears to be better as a result of generality, high breakdown point, and their efficiency (Huber 1981). The detrending methods used in this study were Spline, ModNegExp, and Ar, which were introduced in detail in the introduction chapter.
The standardization process was done by series, and each of them was divided by the growth trend estimated by the detrending methods, resulting in dimensionless indices of the ring-width index. For the mean detrended chronologies by the Spline method, a cubic smoothing spline with 50% frequency response cut-off at 50 years was used for detrending (Cook and Peters 1981). Mean detrended chronologies were also developed using the ModNegExp method, or a linear model was used if a nonlinear model could not be fit (Cook and Kairiukstis 1990; Fritts 2001). Detrending used to develop mean Ar (prewhitened) chronologies was performed by calculating and averaging the residuals of an AR model fitted to the tree ring series.
2.4 Climate data
The climate data, including monthly mean temperatures, the sum of precipitation, SPEI12 values, and PDSI values, were obtained via KNMI Climate Explorer available at http://climexp.knmi.nl (Trouet and Oldenborgh 2013). The mean location coordinates were used to extract the information, more details in Table 1 below.
Table 1. Description of the temperature, precipitation and drought index datasets extracted from KNMI Climate Explorer ((http://climexp.knmi.nl).
Variables Datasets
Temperature 1901-2016: CRU TS 4.01 (land) 0.5° dataset, lat 51.2 / lon 17.3 Precipitation 1901-2016: CRU TS 4.01 (land) 0.5°
Drought indices 1902-2013: CSIC SPEI drought index 12 months CRU self-calibrating PDSI 1901-2016 0.5° Global 3.25
The mean annual temperature from 1901 to 2010 was 6.0°C (± 0.8°C), and mean annual precipitation was 280 mm (± 50 mm). The temperature and precipitation fluctuation over the study period can be seen in Figure 5. Regarding SPEI12 and PDSI, drought indices, positive values indicate wet conditions and negative values dry conditions.
Figure 5. Average temperature and the sum of precipitation fluctuation at Suchý vrch Nature Reserve from 1901 to 2016.
2.5 Calculation of tree growth–drought relationship
The treeclim R package (Zang and Biondi 2015) was used to calculate the moving correlations in a 30-year moving window with a one-year offset, using Pearson's correlation coefficients and a significance threshold of 0.05. A bootstrap moving correlation function was applied between tree-ring chronologies and the monthly climatic data, including the selected drought indices, mean temperature and the sum of precipitation. This was done to evaluate the growth response of Scots pine and Silver fir to drought. Bootstrap is a statistical technique used to estimate population quantities by averaging estimates from multiple small data samples.
3. RESULTS
3.1 Chronology development
The chronologies were developed for the period from 1903 to 2010, although there are series from both species starting around the year of 1600 (see Appendix 1). The sample depth or the total number of chronological series in the period of interest starts from 30 up to 40 series for Silver fir and from 36 up to 48 series for Scots pine (Figures 6). Regarding the descriptive statistics, the EPS values were 0.903 for Silver fir and 0.922 for Scots pine, for both species the values were above the 0.85 threshold. The RBAR, inter-series correlation measures, were 0.216 for Silver fir and 0.222 for Scots pine, all the descriptive statistics in detail are available in Appendix 2. Table 2 summarizes the statistics results for both species.
Table 2. The total number of series of each species forming the mean chronology from 1903 to 2010. EPS values, RBAR measures, and mean DBH (mm) values for both species.
Total number of series EPS RBAR Mean DBH (mm)
Silver fir 40 0.903 0.216 400.79
Scots pine 48 0.922 0.222 362.70
Figure 6. Sample depth and Spline ring-width index (grey line) during the period of interest.
6A: Abies alba data. 6B: Pinus sylvestris data.
3.2 Detrending methods
The chronologies from both species in Figure 7, show quite similar growth trends until about the middle of the 20th century for the mean detrended chronologies obtained with the Spline, ModNegExp and Ar methods. There is a severe decrease in the growth trend detected by the Spline and ModNegExp methods starting around 1945 and lasting until approximately 1960.
The two species start to present a different growth pattern in the second half of the 20th century.
However, the growth patterns of the two species appear to start behaving similarly again around the year 2000, which is the case until the end of the period of analysis. Furthermore, the Spline detrended chronologies strike a balance by retaining some decadal scale variability (information), compared to the chronologies obtained with the Ar method. Moreover, the age- related trend is effectively removed along with any non-climatic trends such as those related to disturbance, which would not be achieved with the ModNegExp detrending method. Due to the advantages of using the Spline method compared with the ModNegExp and Ar methods, the Spline detrended chronologies were selected for the next step, the moving correlation analysis between growth and climate data.
Figure 7. Mean detrended chronologies for both species using different detrending methods.
The chronologies range from 1903 to 2010, the ring-width chronology for Abies alba is shown as the red line and Pinus sylvestris is marked by the blue line. 7A: Mean detrended Spline 50 years. 7B: Mean detrended ModNegExp. 7C: Mean detrended Ar.
3.3 Correlations between growth and climate data
The moving correlation coefficients between growth and moisture conditions were calculated with SPEI12_September (Figure 8A) and PDSI (Figure 8B), using a 30-year moving window with a one-year offset. Both drought indices express the same trend for both species. When we look at the correlations of SPEI12 and PDSI, Silver fir generally still retains its climatic sensitivity to these drought indices, whereas the Scots pine does not. The first half of the 20th century shows that the growth trend of Silver fir is significantly positively correlated with the moisture condition, which means more moisture results in more growth and less growth is caused by dry conditions. The correlations appear to break down during the second half of the 20th century for both drought indices.
Figure 9 only show the months with significant correlations between the tree ring chronologies and temperature as well as precipitation data during the period of interest. The figures containing climate-growth relationships for all months can be found in Appendix 3.
The correlation between growth and temperature (Figure 9) in the first half of the 20th century shows significant negative values for Silver fir especially in May, and for the months of May and June for Scots pine. These significant negative values mean that higher temperatures decrease the growth and lower temperatures increase the growth. The second half of the 20th century shows significant positive values for Silver fir in March, April, June and September.
There are significant positive values for Scots pine mainly in March. The higher temperatures of these months support the growth as lower temperatures decrease the growth.
The correlation between growth and precipitation shows mainly significant positive values for both species over the 20th century (Figure 9). Especially in March, June, July and August for Silver fir, and in June and July for Scots pine. The higher precipitation of these months supports the growth as lower precipitation decreases the growth.
Figure 8. Moving correlation between growth and moisture conditions over the period from 1903 to 2010 for Abies alba and Pinus sylvestris. The x-axis indicates the midyear values of 30-year moving windows with a one-year offset, and the y-axis represents Pearson’s correlation coefficient. Windows with significant correlations (p < 0.05) are displayed in large symbols.
An increase in the trend values indicates an increase in growth and a decrease in the trend values indicates a decrease in growth. 8A: Moving correlation between growth and SPEI12 values.
8B: Moving correlation between growth and PDSI values.
Figure 9. Moving correlations between growth and temperature (above) and precipitation (below) over the period 1903 to 2010 for Abies alba and Pinus sylvestris. The x-axis indicates the midyear values of 30-year moving windows with a one-year offset, and the y-axis represents the Pearson’s correlation coefficient. Windows with significant correlations (p < 0.05) are displayed in large symbols. An increase in the trend values indicates an increase in growth and a decrease in the trend values indicates a decrease in growth.
4. DISCUSSION
4.1 Evaluation of the detrending methods used
The drought sensitivity of Silver fir and Scots pine over the 20th century was analyzed using dendrochronological data. More specifically, the ring width series data were prepared and detrended using three different methods, namely smoothing Spline, modified negative exponential curve and the residual of an Ar model. The mean chronologies from the Spline method were further on used to perform a correlation analysis. The moving correlation analysis was done between the tree ring width chronologies and monthly climate data (mean temperatures, precipitation sum, SPEI12 values, and PDSI values) over the 20th century.
Each detrending method has its own merits and can show us different aspects of growth and its relationship to climate. For instance, the Ar residual chronology can be used to give an interannual representation of growth and its relation to climate, without showing any longer- term relationship other than year-to-year. The ModNegExp method is much less flexible and retains most of the growth trends, which can also be useful for studying the chronologies from a different perspective. The Spline method can be more or less flexible than the 50-years Spline used in this study, which can also be used to retain or remove growth trends at different timescales. However, in principle, all these chronologies may provide useful information about how tree radial growth is responding to climatic inputs over different timescales and periods.
The selected method for further data analyses was the Spline because the age-related trend and non-climatic trends (disturbance) could be removed without completely removing the decadal scale variability. The ModNegExp method showed a quite similar effect, although the disturbance influence was not removed as effectively. The Ar method removed the most longer- term information of all the methods, retaining only the interannual variability.
4.2 Evaluation of growth-trend based on the chronologies
After analysing the mean chronologies, it is possible to notice a different behaviour in the growth trend in the first and second half of the 20th century from both species. The first half of the 20th century shows a quite similar growth trend for Silver fir and Scots pine. In the second half of the 20th century Silver fir and Scots pine show a different pattern in their growth compared with the first half of the century. There is also a greater and longer lasting growth decline in Silver fir around mid-century. The period of the 1980s to 2000 shows also higher growth rates for Scots pine compared to Silver fir. This could possibly be related to some disturbances, which seemed to benefit more Scots pine than Silver fir.
Around the years 1945 until 1960, the chronologies from the Spline and ModNegExp methods show a severe decrease in the radial growth of both species. There was likely a climatic trigger for the mid-1940s decline when considering summer drought. The year 1947 was one of the driest summers in the 20th century (Levanič, 2013). Normally the summer drought factor would only cause a quite short-term effect, such as reduced growth for one or two years. The second half of the 20th century has also been known as a period of increased atmospheric environmental pollution from power stations, affecting the Sudetes region (Alcamo 1987; Molski 1987). Based on previous studies, severe air pollution can severely stress the trees leading to a productivity decrease, or even tree death if the stress is severe enough and lasts for a longer time (Miś and Rączka 2002; Feliksik and Wilczyński 2003).
Generally, when an increase of SO2 concentrations in the atmosphere occurs there is a suppression of growth rates of trees. When this concentration decreases, there is observed growth recovery, respectively. It has been suggested a clear connection to between the suppression severity and recovery strength (Filipiak and Ufnalski 2004). Studies related to the investigation of SO2 effects on coniferous tree species suggest also a similar reaction to pollution, despite tree species, geographical location and site-specific conditions (Rydval, 2012). Moreover, it has been suggested that Silver fir suits well (even best) for studies regarding the sensitivity of trees to environmental pollution (Elling et al. 2009). The decrease in the growth trend may also be linked to the consequences of environmental pollution in the area.
However, it is necessary to have sufficient information regarding the pollution loading during that period in order to infer about the prolonged growth decline. Probably different factors have
also interactive effects on growth and its decline, e.g. climate-induced drought together with increased stress conditions for trees due to environmental pollution.
In the case of the pollution-related stress being so severe that it is causing tree death (neighbouring trees die), the remaining trees will have more resources than before due to lower competition for space, light, nutrients, and water. A growth release or productivity increase will then be expected, and it might even last a longer time, e.g. a decade or two. The mean chronologies (Spline and ModNegExp) of this study for the end of the 20th century show this behaviour, i.e. after the year 2000 Silver fir and Scots pine both appear to have a similar growth pattern again.
4.3 Evaluation of climate tree-growth relationships
The same growth trend can be seen for Silver fir and Scots pine for the moving correlation with both drought indices. Silver fir appears to have more climatic sensitivity to SPEI12 and PDSI values. On the other hand, Scots pine does not respond to the drought indices in general, suggesting that some other variables may be influencing its growth. According to Martínez- Vilalta et al. (2012), local climate can have a relatively minor effect on the radial growth response of Scots pine. Its growth response can be primarily determined by tree-level characteristics (age and previous growth rate), stand basal area and other tree species competition. The significant positive correlation between Silver fir growth and the drought indices suggests that during the first half of the 20th century this tree species was significantly influenced by the precipitation sum and consequently soil moisture availability. This was because of its increased growth when moisture availability was greater, and growth decline during drier conditions.
In the second half of the 20th century, both drought indices show a breakdown in the correlations, suggesting that something other than moisture conditions is influencing the tree growth response. Even though Silver fir was, in general, more responsive to climate data than Scots pine, in the period around the 1950s there was no significant response of Silver fir to the drought indices. The studies of Miś and Rączka (2002) and Feliksik and Wilczyński (2003) suggest that this growth decline was due to the atmospheric environmental pollution that
affected the Sudetes area in the second half of the 20th century. This may have contributed to the loss of climatic sensitivity in Silver fir.
The observed negative significant values for the moving correlation comparison between tree growth and temperature for Silver fir and Scots pine (in the first half of the 20th century) are suggesting that high temperatures during some months are causing a decrease in the tree growth, in opposite to lower ones (in both tree species). The positive significant values (in the second half of the 20th century) are suggesting that higher temperatures increase tree growth in opposite to lower temperatures in Silver fir (on March, April, June and September) and Scots pine (on March), respectively. The significant positive values for both species for the moving correlation comparison between tree growth and precipitation over the 20th century show that higher precipitation sum increases tree growth in opposite to lower precipitation sum in Silver fir (March, June, July and August) and Scots pine (June and July), respectively.
5. CONCLUSION
Regarding the methodologies used in this study, it is possible to conclude that the Spline detrending method was the best approach to discern disturbances noise and climate signal in studies of climate sensitivity in natural forests such as the one in Suchý vrch Nature Reserve.
However, each detrending method has its own merits and can show us different aspects of the relationship between growth and climate. The Palmer Drought Severity Index and the Standardized Precipitation Evapotranspiration Index showed also a similar trend for both tree species which indicates that both drought indices can be used quite equally for this kind of studies.
Although there are also some clear differences in the growth patterns of different tree species in certain time periods, the main difference between them was their growth response to climate fluctuations. The relationship with climate fluctuations was generally stronger for Silver fir than for Scots pine. Towards the end of the 20th century, the drought (or precipitation sum) response in Scots pine was more disrupted and lasted a longer time than the Silver fir response. This was perhaps also related to some physiological differences between these species. Also, the response to temperature is seemingly becoming more important during this period compared to the drought (or precipitation sum) response.
The growth response of both tree species to climate showed that moisture stress and low temperature do not affect the growth of these tree species alone, but also the atmospheric pollution likely play an important role in their growth responses. The breakdown in the relationship between drought index or precipitation sum with tree growth, later in the century is almost certainly due to the increased degree of air pollution loading at that time. This possibly has also affected at least partially the growth responses of trees to climatic conditions. In order to better understand the interactive effects of different climatic factors and air pollution on tree growth responses, additional research is still needed in the future.
References
Aber, J.D., Ollinger, S.V., Federer, C.A., Reich, P.B., Goulden, M.L., Kicklighter, D.W., Melillo, J.M., Lathrop Jr., R.G. 1995. Predicting the effects of climate change on water yield and forest production in the northeastern United States. Climate Research 5:207-222.
Adams, H.D., Guardiola-Claramonte, M., Barron-Gafford, G.A., Villegas, J.C., Breshears, D.D., Zou, C.B., Troch, P.A., and Huxman, T.E. 2009. Temperature sensitivity of drought- induced tree mortality portends increased regional die-off under global-change- type drought. Proceedings of the National Academy of Sciences USA 106:7063–7066.
Alcamo. 1987. Future sulfur deposition scenarios in forested areas of Europe. In: Kairiukstis, L., Nilsson, S., Straszak, A. (eds.) Forest decline and reproduction: regional and global consequences, Proceedings of the workshop held in Krakow, Poland, 23–27 March 1987, IIASA, Laxenburg Austria, pp 33–43.
Allen, C.D., Breshears, D.D., and McDowell, N.G. 2015. On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene.
Ecosphere 6(8):129.
Allen, C.D., et al. 2010. A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. Forest Ecology and Management 259:660–684.
Alley, W.M. 1984. The Palmer Drought Severity Index: Limitations and Assumptions.
Journal of Climate and Applied Meteorology., 23, 1100–1109.
Babst, F., Poulter, B., Trouet, V., Tan, K., Neuwirth, B., Wilson, R., Carrer, M., Grabner, M., Tegel, W., Levanic, T., Panayotov, M., Urbinati, C., Bouriaud, O., Ciais, P., Frank, D.
2013. Site- and species-specific responses of forest growth to climate across the European continent, Global Ecology and Biogeography, 22 (6), 706-717.
Breshears, D.D., López-Hoffman, L., Graumlich, L.J. 2011. When ecosystem services crash:
preparing for big, fast, patchy climate change. Ambio 40:256-263.
Briffa, K.R., Jones, P.D. 1990. Basic chronology statistics and assessment. In: Methods of Dendrochronology: Applications in the Environmental Sciences. Kluwer Academic Publishers, pp. 137-152
Cook, E.R., Buckley, B.M., D'Arrigo, R.D., and Peterson, M.J. 2000. Warm-season temperatures since 1600 B.C. reconstructed from Tasmanian tree rings and their relationship to large scale sea surface temperature anomalies. Climate Dynamics 16(2/3): 79–91.
Cook, E.R. 1985. A time series analysis approach to tree ring standardization. PhD Thesis.
University of Arizona, Tucson.
Cook, E.R. and Kairiukstis, L.A., (eds). 1990. Methods of Dendrochronology: Applications in the Environmental Sciences. Springer. ISBN-13: 978-0-7923-0586-6.
Cook, E.R., Kairiukstis LA eds. (2013) Methods of dendrochronology: applications in the environmental sciences. Springer Science & Business Media
Cook, E.R. 1992. A conceptual linear aggregate model for tree rings. In: Cook E.R. and Kairiukstis, L.A. (eds.) Methods of dendrochronology: Applications in the environmental sciences. Kluwer Academic Publishers. London. 98–104 pp.
Cook, E.R., Peters, K. 1981. The smoothing spline: a new approach to standardizing forest interior tree ring width series for dendroclimatic studies. Tree-ring bullet
Dai, A. 2012. Increasing drought under global warming in observations and models. Nat Clim Change, 3: 52–58
Dale, V.H., Joyce, L.A., McNulty, S., Neilson, R.P., Ayres, M.P., Flannigan, M.D., Hanson, P.J., Irland, L.C., Lugo, A.E., Peterson, C.J., Simberloff, D. 2001. Climate change and forest disturbances: climate change can affect forests by altering the frequency, intensity, duration, and timing of fire, drought, introduced species, insect and pathogen outbreaks, hurricanes, windstorms, ice storms, or landslides. BioScience 51:723-734.
Dore, M.H. 2005. Climate change and changes in global precipitation patterns: what do we know? Environment International 31:1167-1181.
Duncan, R.P. 1989. An evaluation of errors in tree age estimates based on increment cores in kahikatea (Dacrycarpus dacrydioides). New Zealand natural sciences 16:1-37.
Elling, W., Dittmar, C., Pfaffelmoser, K., & Rötzer, T. 2009. Dendroecological assessment of the complex causes of decline and recovery of the growth of silver fir (Abies alba Mill.) in Southern Germany. Forest Ecosystem Management, 257, 1175–1187.
Ermich, K., Rutkowski, B., Bednarz, Z., Feliksik, E. 1976. The degree of similarity of dendrochronological curves as an indicator of site conditions. Tree-Ring Bull 36:1–8.
Feliksik, E., Wilczyński, S. 2003. Tree-rings as indicators of environmental change. EJPAU 6, 2, ser Forestry.
Filipiak, M., & Ufnalski, K. 2004. Growth reaction of European Silver fir [Abies alba Mill.]
associated with air quality improvement in the Sudeten Mountains. Polish Journal of Environmental Studies, 13(3), 267–273.
Frank, D.C. & Esper, J. 2005. Temperature reconstructions and comparisons with instrumental data from a tree-ring network for the European Alps. International Journal of Climatology, 25, 1437–1454.
Fritts, H.C. 2001. Tree Rings and Climate. Blackburn Press: Caldwell 2001. 567pp.
Fritts, H.C. 1976. Tree rings and climate. London, New York and San Francisco: Academic Press. xii + 567 pp.
Fuchs, B. 2012. Palmer Drought Severity Index (PSDI and scPDSI). National Drought Mitigation Center. University of Nebraska-Lincoln. Caribbean Drought Workshop May 22- 24, 2012.
Huber, P.J. 1981. Robust statistics, New York: John Wiley & Sons, Inc., ISBN 0-471-41805- 6, MR 0606374
IPCC. 2013. Climate change 2013: the physical science basis. Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change.
Cambridge University Press, Cambridge, UK.
Köppen, W., Geiger, R. 1954: Klima der Erde (Climate of the earth). Wall Map 1:16 Mill.
Klett-Perthes, Gotha.
Koźmiński, C., Michalska, B., Nidzgorska-Lencewicz, J. 2001. Wind directions and velocities. In: Koźmiński C, Michalska B (eds) Atlas of climatic risk to crop cultivation in Poland, University of Agriculture in Szczecin, p 25.
Larsson, L. 2015. CooRecorder and Cdendro programs of the CooRecorder/Cdendro package version 549 665 7.8. http://www.cybis.se/forfun/dendro/
Levanič, T., Popa, I., Poljanšek, S. and Nechita, C.2013. A 323-year long reconstruction of drought for SW Romania based on black pine (Pinus nigra) tree-ring widths. International journal of biometeorology, 57(5), pp.703-714.
Martínez-Vilalta, J., López, B.C., Loepfe, L., Lloret, F. 2012. Stand- and tree-level determinants of the drought response of Scots pine radial growth. Oecologia 168: 877–888.
McDowell, N.G., Pockman, W.T., Allen, C.D., Breshears, D.D., Cobb, N., Kolb, T., Plaut, J., Sperry, J., West, A., Williams, D.G., and Yepez, E.A. 2008. Mechanisms of plant survival and mortality during drought: Why do some plants survive while others succumb to drought? New Phytologist 178:719–739.
Miś, R., Rączka, G. 2002. Effect of long-term industrial pollution on the increase in height of Scots pine (Pinus sylvestris L.). In: Siwecki R (ed) Biological reactions of trees of industrial pollution, Inst Dendrologii PAN, Kórnik, pp 431–440.
Molski, B. 1987. Fir and pine phloem and wood growth retardation due to air pollution. In:
Kairiukstis, L., Nilsson, S., Straszak, A. (eds.) Forest decline and reproduction: regional and global consequences, Proceedings of the workshop held in Krakow, Poland, 23–27 March 1987, IIASA, Laxenburg, Austria, pp 115–122.
Palmer, W.C. 1965. Meteorological Drought Research Paper 45. Washington DC: US Weather Bureau
Rydval, M. & Wilson, R. 2012. The impact of industrial SO2 pollution on north Bohemia conifers. Water, Air, and Soil Pollution, 233: 5727-5744.
Rydval, M., Druckenbrod, D.L., Svoboda, M., Trotsiuk, V., Janda, P., Mikoláš, M., Čada, V., Bače, R., Teodosiu, M., Wilson, R. 2018. Influence of sampling and disturbance history on climatic sensitivity of temperature-limited conifers. The Holocene: 0959683618782605.
Schelhaas, M.J., Nabuurs, G.J., Schuck, A. 2003. Natural disturbances in the European forests in the 19thth and 20th centuries. Global Change Biology 9:1620-1633.
Seidl, R., Thom, D., Kautz, M., Martin-Benito, D., Peltoniemi, M., Vacchiano, G., Wild, J., Ascoli, D., Petr, M., Honkaniemi, J., Lexer, M.J. 2017. Forest disturbances under climate change. Nature climate change 7:p.395.
Sheffield J, Wood E F, Roderick M L. 2012. Little change in global drought over the past 60 years. Nature, 491: 435–438
Speer, J.H. 2010. Fundamentals of tree-ring research. Tucson: University of Arizona Press.
Spiecker, H. 2002. Tree rings and forest management in Europe. Dendrochronologia 20/2–
1:191–202.
Trouet, V., Van Oldenborgh, G.J. 2013. KNMI Climate Explorer: a web-based research tool for high resolution paleoclimatology. Tree-Ring Research 69:3-13.
Vicente-Serrano, S.M., Beguería, S., López-Moreno, J.I., Angulo, M., El Kenawy, A. 2010.
A new global 0.5° gridded dataset (1901–2006) of a multiscalar drought index: Comparison with current drought index datasets based on the Palmer drought severity index. J Hydrometeorol, 11:1033–1043.
Vicente-Serrano, S.M., Beguería, S., Lorenzo-Lacruz, J., Camarero, J.J., López-Moreno, J.I., Azorin-Molina, C., Revuelto, J., Morán-Tejeda, E., Sanchez-Lorenzo, A. 2012.
Performance of drought indexes for ecological, agricultural, and hydrological applications.
Earth Interact, 16: 1–27.
Vicente-Serrano, S.M. & National Center for Atmospheric Research Staff (Eds). Last modified 18 Jul 2015. "The Climate Data Guide: Standardized Precipitation Evapotranspiration Index (SPEI)." Retrieved from https://climatedataguide.ucar.edu/
climate-data/ standardized-precipitation-evapotranspiration-index-spei.
Wang L, Chen W. 2012. Characteristics of multi-timescale variabilities of the drought over last 100 years in Southwest China (in Chinese). Adv Meteorol Sci Technol, 4: 21–26 Weed, A.S., Ayres, M.P., and Hicke, J.A. 2013. Consequences of climate change for biotic disturbances in North American forests. Ecological Monographs 83:441–470.
Wigley, T.M.L., Briffa, K.R., & Jones, P.D. 1984. On the average value of correlated time series, with applications in dendroclimatology and hydrometeorology. American Meteorological Society, 23, 201–213.
Yang, Q., Li, M.X., Zheng, Z.Y., Ma, Z.G. 2017. Regional applicability of seven meteorological drought indexes in China. Science China Earth Sciences, 60:745–760, doi:
10.1007/s11430-016-5133-5.
Zang, C., Biondi, F. 2015. Treeclim: an R package for the numerical calibration of proxy- climate relationships. Ecography 38:431-436.
Appendix 1
Raw series of Abies alba (Silver fir) and Pinus sylvestris (Scots pine).
Appendix 2
Descriptive statistics for Silver fir and for Scots pine.
Silver fir
Scots pine
Appendix 3
Moving correlation between growth and temperature and precipitation sum, respectively, over the period of 1903 to 2010 for all months. The x-axis indicates the starting year of 30-year moving window with a one-year offset, and the y-axis represents Pearson’s correlation coefficient. Windows with significant correlations (p < 0.05) are displayed in large symbols.
An increase in the trend values indicates an increase in growth, a decrease in the trend values indicates a decrease in growth.
