HEAT AS A FACTOR DETERMINING THE ACTIVITY OF CAMBIUM IN BLACK ALDER (ALNUS GLUTINOSA (L.) GAERTN.) ON MOIST SITES

Sławomir Wilczyński¹, Rafał Podlaski^2
1 Department of Forest Protection and Forest Climatology, Agricultural University of Cracow, Poland
² Department of General Biology and Nature Protection, Institute of Biology, Pedagogical University of Kielce, Poland

ABSTRACT

The effects of air temperature, precipitation, and duration of the direct solar radiation (sunshine) on radial increment of black alder (Alnus glutinosa (L.) Gaertn.) are discussed in this paper. The investigated 21 trees were growing on a fertile site with a high water table. Each tree was cored, and on cores the tree-ring width was measured. On the basis of these measurements the mean tree-ring chronology was developed which represented the growth rhythm of trees in the investigated site. The black alder trees were most strongly responding to temperature. The reaction to sunshine duration was a little weaker, while reaction to precipitation was insignificant. The effect of previous and current year temperatures on radial increment was significant. A positive effect on tree growth had a low temperature in May, high in July and October of the previous year, and high in May, July and August of the current year. Frosts in January and prolonged winter with low temperatures were harmful to trees. The duration of the direct solar radiation in the current growing season had a significant influence on tree growth. The increased amount of sunshine in May, July and August had a positive effect on radial increment. A lack of significant correlation between precipitation and increment was probably the result of moisture conditions of the investigated site. However, there was some influence of precipitation of the cool part of the year on tree growth in the coming growing season.

Key words: Alnus glutinosa, dendroclimatology, dendroecology, tree-ring width, Swietokrzyski National Park, Poland.

INTRODUCTION

The dendroclimatic analyses concerning relationships between weather conditions and tree increment permit to obtain a quite exact estimation of tree sensitivity to meteorological factors [11]. This type of research should include various sites located in various climatic regions, because site characteristics frequently decide on the nature of climate – increment relationships [28]. The ecological requirements of black alder are quite well known, but its susceptibility to the weather factor was in the first place studied when trees were young (seedlings and young trees). Natural black alder stands mainly occur in central and northern Poland, but they are also present in mountains where they may be found up to 500 m in elevation. Black alder mainly occurs on terraces of river and lake valleys [23]. It prefers fertile soils of moderate moisture and of neutral or little acid reaction, and with a stable high water table [5, 7, 18]. A long drought has a harmful effect on trees, especially in their young age [20]. Black alder prefers a temperate climate with the mean annual temperature of 4 to 7.5°C. Its moisture requirements are satisfied by annual precipitation above 400 mm [16].

The site with trees under investigations was characterised by a high water table permanently supplied with rivulet waters. In respect of climate there were optimum conditions for growth of black alder. However, the seasonal extreme temperatures and precipitation could have a negative effect on trees. This may be deduced from climatic diagrams made for this area and presented in Fig. 1. In spite of the fact that black alder belongs to frost resistant species the long lasting frosts in winter have a negative effect on the condition of trees, and this results in increment drop [16]. However, this species is quite resistant to short-term temperature drops in winter. Beginning with September the frost resistance of shoots and buds increases reaching its maximum in February. In March this resistance gradually decreases. In April the frost resistance of previous year shoots and buds suddenly decreases, while in May young developing shoots and flowers are most susceptible to late frosts [7].

In north-western Poland black alder as a rule begins to bloom in early April, and in some years already in late March [27]. In the Swiętokrzyskie Mountains this usually happens in mid-April. The shedding of leaves most often begins in September and lasts till November, depending on temperature [26, 27]. The roots begin an active growth already in late winter, depending on temperature and snow cover [19]. Black alder is considered to be a light demanding species [2], but its light requirements are much greater when trees are older [21]. Nutrients stored by black alder are in great extent transferred to the soil with falling leaves. Thus, it does not store the biogenes for winter and the next growing season [34]. Black alder produces fruits every year, but their abundance varies, and this sometimes may affect the growth potential of trees [3].

Well known ecology of this species demarcates the area for searching the relationships between individual meteorological elements and the radial growth of trees. It should be pointed out, however, that most of characteristics of black alder presented above concern its young age. The aim of our study was to determine the climate-radial increment relationships in fully mature trees, above 50 years of age. The following three biologically important meteorological elements were taken into consideration: air temperature, duration of the direct solar radiation (sunshine) (being the carriers of heat energy), and precipitation.

MATERIAL AND METHODS

Trees used in this study were about a hundred years old. They were growing on a fertile very-fine sandy alluvial soil in the ash-alder (Fraxino-Alnetum) reparian forest in compartment 8A of the Swiętokrzyski National Park. According to data of the near meteorological station in Kielce this area is characterised by a relatively short growing season (185 days) (Fig. 1). As a rule the mean temperature of winter months is below 0 °C, and in the extreme case it was –11 °C (Fig. 1). The mean temperature of the warmest months over a long-term period of time was below 18°C. The maximum precipitation occurs in summer, and this is also true for the direct solar radiation (sunshine). Temperature was most variable in winter, while the variability of precipitation and sunshine was the highest during the growing season. Annual values of temperature and precipitation are almost right in the middle of black alder climatic optimum, while the extreme monthly or seasonal values may be out of the tolerance range of this species [23], and therefore they may limit the growth of trees.

Trunks of 21 trees were cored, taking one core from each tree, 1.3 m above the ground level. On cores the tree-ring width was measured. Tree-ring dating was verified using the computer program COFECHA [14]. To eliminate the trend and long-term fluctuations, the tree-ring series were standardized using the computer program ARSTAN [4]. On the basis of indexed series the tree-ring chronology was computed for the period 1960–2004. The chronology, thanks to the process of autoregressive modelling (AR) of tree-ring series, was cleared of autocorrelation characteristic for tree-ring chronologies. Such a chronology is called a residual chronology. Indexing of tree-ring series was carried out according to the algorithm:

(1)

with I_i – increment index value for year i ; R_i – actual value of tree-ring width in year i ; Y_i – smoothed value of tree-ring width in year i, read from the fit-curve.

The principal components analysis was used to estimate the homogeneity of tree growth responses. This method permits to reduce the number of cases describing variables and permits to classify them. Tree-ring series (variables) are strongly correlated with one another. Besides, they are described by a very large number of cases (tree-ring widths). This creates problems in interpretation and makes detection of their structure and relationships difficult. The principal components analysis (PCA) facilitates the accomplishment of these two tasks through transformation of original variables into their smaller numbers which are correlated with one another.

In searching for factors having the greatest influence on variation of radial increments the method of correlation and similarity based on the coefficient of agreement (GL%) was used [6]. This coefficient was computed according to the formula:

[%]

(2)

with m – number of sections of compared curves of a similar progress direction (increase or decrease); n – number of compared years.

The relationships between the air temperature, sunshine, and precipitation and the tree-ring width for the period 1960 – 2004 were estimated using correlation and response functions (a modified version of multiple regression) [11, 12, 13]. Each time 17 predictors were taken into account, i.e. monthly values of temperature, sunshine, and precipitation from May of the previous year to September of the year in which the tree-ring was formed. The increment indexes of residual chronology from the period 1960–2004 were the predictands. Coefficients of determination of the multiple regression permitted to estimate the contribution of temperature, precipitation, and sunshine to variation of radial increment. The computer program RESPO was used for these computations [15].

To estimate changes in climate-increment relationships taking place in time the computer program DENDROCLIM2002 was used [1]. This program utilizes the bootstrap and response function. The correlation and regression functions were computed for multiple time intervals. The bootstrapped response and correlation functions were computed for single and multiple time intervals. Three variants were used: a variant of moving intervals in which a constant length is progressively slid by one year; a variant of backward evolutionary intervals in which interval length is incremented by one starting from the most recent year; and a variant of forward evolutionary intervals in which interval length is incremented by one starting from the least recent year [1].

The climatic conditions prevailing during positive and negative signature years were also analysed [17, 24, 25]. It was assumed that during positive signature years 90% of trees increased the tree-ring width in comparison with width of the ring produced in the previous year. While in negative signature years the tree-ring width decreased in 90% of trees.

In each of the above cases the period from 1960 to 2004, i.e. when age of trees was greater than 50 years, was analysed.

RESULTS

The indexed tree-ring series of 21 trees were relatively similar, although this similarity was diversified in time (Fig. 2). There were years when all trees responded in a similar manner, but there were also numerous cases when this homogeneity of increment responses was much smaller. The mean coefficient of agreement of tree-ring series of 21 trees was 75.5% (p<0.001 for n=44). It was therefore very high. Coefficients of agreement (GL) between individual series were as a rule highly significant, although there were cases when there was no agreement between some series. Similar relations concerned the correlation of tree-ring series. The mean coefficient of correlation of 21 series was 0.467 (p<0.001, n=45) (Table 1).

Years of an especially high homogeneity of increment responses were marked as signature years. In total there were 19 signature years found among 45 years analysed. There were 12 positive and 7 negative years (Fig. 2). This was a relatively high number of signature years in comparison with other tree species [9, 29, 30].

The spread of indexed tree-ring series in relation to the first four principal components indicated that the first principal component integrates the tree-ring series (Fig. 3), while the next three differentiate them into groups. The first principal component (PC1) described 48.6% of total variation of tree-ring series, the second (PC2) 10.2%, the third (PC3) 8.51%, and the fourth (PC4) 6.1%. In total the first four components described 73.4% of total variation of tree-ring widths.

The identification of the nature of factors determining the variation of tree-ring widths permitted to conclude that only the first factor was of a climatic nature. Only values of the first principal component (PC1) showed a significant correlation and agreement with climatic parameters. These were the mean temperature of May of the previous year (negative correlation and agreement), the mean temperature of May of the current year (positive correlation and agreement), the mean temperature of the period July – August, and the total sunshine (positive correlation and agreement) (Fig. 4). The remaining factors described by the remaining three principal components must have had a non-climatic nature. They were probably accidental factors connected with growth history and individual features of trees.

In order to confirm results described above the analysis of correlation and response function of tree-ring widths with various climatic factors was carried out for the period 1960–2004 (Fig. 5). Its results in majority confirmed the results of PCA analysis. Tree-ring width was negatively correlated only with mean temperature of May of the previous year, while there was a positive correlation with temperature of July and October of the previous year as well as with temperature of January, May, July, and August of the current year. Besides, a high amount of sunshine in February had a negative influence on diameter increment, while an increased amount of sunshine in May, July, and August of the current year had a positive influence (Fig. 5). There was no significant correlation between diameter increment and precipitation.

It also turned out that temperature of air had the greatest effect on variation of tree-ring width (R²=41.1%). Much smaller was the effect of sunshine (R² = 28.0%), while the effect of precipitation was minimal (R² = 6.4%). The analysis of the bootstrap correlation and response function for variable intervals gave similar results (Tables 2, 3, and 4). A high temperature in May of the previous year had a significant negative effect on width of black alder tree-rings. This relationship was stable in time and it was very strong. Also a positive effect on radial increment of temperature of July and October of the previous year as well as temperature of March, April, May, July, and August of the current year was revealed. However, these relations were not as stable in time as it was in the case of temperature of May of the previous year (Table 2). Furthermore, it was found that overabundance of sunshine in June of the previous year negatively affected the radial increment of the current year. Also there was a distinct positive effect on tree-ring width of the increased amount of sunshine in May, July, and August of the year in which the ring is formed. The effect of sunshine in August was the most stable in time (Table 3). The relationship precipitation – diameter increment was very weak (Table 4). However, there was a positive effect of high precipitation in June of the previous year on growth of young black alder trees. This was revealed by results of the bootstrap correlation method in the variant of backward evolutionary intervals (Table 4). The findings presented above showed a high similarity of results obtained by various statistical methods.