One of indirect indicators of climate change under effect natural and anthropogenic factors is width of a radial increment of year tree rings. Dendrochronology (dendroindication) of natural and anthropogenic changes of the nature promotes analysis of a conditions and dynamics of geocomplexes at the retrospective and can be used for the forecasts. The estimation of relationships between dendrochronologic and climatic indexes creates the basis for renovation of elements of a climate for the period determined by age of tree-ring-type chronologies. Rather short duration of the hydrometeorological time series and the features of spatial distribution of the weather stations allow making analysis only for processes and phenomena, restricted by relatively short time-space scales. The region for investigations occupy of the Eastern Fennoscandia including taiga sub-zones of the Kola Peninsula, Karelia, Vologda and Arkhangelsk regiones. Dendrochronologic investigations in these territories were investigated before by Bitvinskas, Kairatis (1978), Vessart (1978), Lovelius (1992), Jaskelainen, Kozlov, (1993), Sinkevich (1996), Zetterberg (1986), Lindholm (1996).

The dating of natural processes and phenomena, anthropogenic effects on the nature, historical events implements with using the dendrochronological scales. The main advantage of tree-ring method is in obtaining relatively long series of observations. Material for our dendrochronologic researches are: a pine Pinus sylvestris L., a fir-tree of European (Picea abies (L.) Karst.) and Siberian (Picea abovata Ledeb.) species, a Siberian larch (Larix sibirica Ledeb.), and a juniper Juniperus communis L.

The dendrochronologic samples were taken during the field expeditions at the period from 1995 till 2001 at Karelia, Arkhangelsk, Vologda and Murmansk districts from the living trees and from historical and mineral (fossil) arbors, and the sites, where the samples were collected, presents on Figs. 5 and 6.

The samples were taken by Pressler's age bore with a working section 45 cm. In exeptional cases the disks were carved. For review and scoring of tree-rings a binocular microskope was used. The annual increments of tree rings was calculated with accuracy in 0,05 mm. The measurements of arboreal discs were made on two radiuses - from center punches on perpendicular lines of tree-rings.

The dendrochronologic samples were taken during the field expeditions at the period from 1995 till 1999 at Karelia, Arkhangelsk, Vologda and Murmansk districts from the living trees and from historical and mineral (fossil) arbors, and the sites, where the samples were collected, presents on Fig. 7.

For interpretation obtained dendrochronologic time series the standardizing of the natural data was applied. It is necessary for equalization of the physiologically conditioned differences of the natural data, i. e. for reduction them to indexes.

An optimal method for computation of tree-ring indexes is the calculation by technique of a "corridor", offered by S. Shijatov (1986). In accordance to this method on the diagram of tree-ring series two curves are marking: on the most maximum and minimum values (Fig. 7). At calculation of indexes of increment the attitude of points within the limit of a "corridor" take into account. The transformed in accordance of method of a "corridor" diagram of annual tree-ring increments is shown on Fig. 8, and tree-ring diagram transformed in accordance to the method of a "corridor" showed on Fig. 9.

For simplification of calculations, width of a corridor was taken equal 100 or 200% (in last case the values of indexes changes from 0 to 200%). For calculation of indices, the following equation was used:

I = 100[(L_f - L_min)/(L_max - L_min)],

or, respectively

I = 200[(L_f - L_min)/(L_max - L_min)],

where I is index of increment, per cent; L_f is width of a year ring, L_min is difference between an abscissa axis and minimum curve, L_max is difference between an abscissa and maximum curve (Shijatov, 1986).

Fig. 5. Locations of dendrochronologic sample sites in Karelia.

Fig. 6. Locations of dendrochronologic sample sites in the Kola Peninsula

Fig. 7. The 451-year Pine scots from Tolvojarvi, Karelia

Fig.8. An example of annual tree-ring time series.

Fig. 9. Tree-ring diagram transformed in accordance to the method of a "corridor".

Advantages of this method are:

• visualization and capability of validation of sampling of models;
• possibility to reveal not only short-term, but also long-term climatically conditioned oscillations in tree-ring series;
• an absence of a non-homogeneous in tree-ring series;
• the duration of initial time series is not reduced.

The indexing of absolute numbers of sampled dendrochronologic time series on a method of a "corridor" was make with application of a software "Indexa. V. 08/99", designed by V. Mazepa (Institute of plant ecology and animals, Ural Division of the Russian Academy of Sciences, Ekaterinburg). Obtained series were tested with using statistical methods (trend, spectral, regression, auto- and cross-correlation analysis, etc.) with using of a computer statistical program. The method of cross-dating was applied for more accurate definition of the characteristics of dendrochronological series, received for the groups of samples, taken from one locations. First of all for the cross analysis served the samples, taken from the oldest trees, and from historical wooden facilities and mineral timbers. This samples were collected from architectural constructions, the data building which one are approximately known under the archive data. Some samples of mineral timber were obtained at historical corduroy road "Osudareva doroga", which was built at 1702.

In a process of our study the basic principles of dendrochronological researches (principle of monotony, principle of limiting factor, principle of ecological valence etc.) were applied (Bitvinskas, 1974).

For the creation of temporary scales of a increment of tree-year rings of different localities of territory of the studied area were used the representative dendrochronologic time series (models) obtained from unimpaired alive arbors of different age. The indexes of each model were mated in generalized series, which one were finally built by more continuous tree-ring series. Cross-dating of the models with mineral and historical timber did allowed in some cases to prolong the scales on 100-200 years and more. The duration of the scales is changed from 187 to 488 years (see Fig. 10).

Fig. 10. Historical timber from the ancient road from Onega Lake to the White Sea

Accommodation of the obtained arboreal chronologies in a southern part of Karelia, in a middle taiga sub-zone of softwood forests, gave possibility to generalize indexes of the local scales into unified time series of increment of tree-rings. For the territories of the north of Karelia and for the Kola Peninsula the local unified three-ring type chronologies were created too. Obtained local dendrochronological scales were analyzed on the separate years and periods in the comparison with the hydrometeorological data of instrument observations on the climate stations, located in Karelia (Petrozavodsk, Olonets, Sortavala, Loukhi and Kestenga), and with the instrumental data on the water level regime of Lake Onega, as will be show at next chat of the report.

By result of the analysis of the tree-ring data in Finland, it was obtained that in a pine (Pinus sylvestris L.) the variations of the contents of an isotope of carbon ¹⁴C most intimately correlate with summer air temperature in a vegetation period (Filatov, 1997). The correlation coefficients between mean month temperature and concentration of CO₂ in tree-rings for the Northern Finland are the highest for July and August, during the period of the maximum growth of the trees.

It is established that the activity of growth depends on a course of annual air temperature. The matching of indexes of dendrochronological scales obtain from a pine (Pinus sylvestris L.) on both territories Southern Karelia and Southern Finland (Zetterberg, 1990a) demonstrates some resemblance of processes of a increment of annual tree-rings of a pine in adjacent areas (Fig. 11).

Fig.11. The generalized dendrochronological scales (5-year moving averages) for the territory of Southern Karelia (1) and Southern Finland (2), obtained by Zetterberg (1990a).

The basic information of the created scales is given in Table 1.

On these diagrams the period of large amplitudes of timber increment marked in XVI-XVIII centuries is well excreted. Extremely unfavorable for growth of a pine in a middle taiga zone of Karelia and Finland were 1590 and 1640 decades. From a beginning XIX century the amplitude of accretions is decrease, that can be a consequent of the total tendency on warming during final stage of a "small drift epoch". In accordance to instrumental observations, the maximum annual air temperature in Karelia and Finland was marked in 1990 decades of present century. Positive trend in increasing of annual precipitation also was marked. The rather cold period was observed from middle 1950 up to middle 1970 decades, that is in a good correspondence with the dendrochronological data, obtained in our researches.

Table 1.
Duration of the local dendrochronological scales of the Eastern Fennoscandia

No	Scale	Period, years	Duration, Years
1	Pulozero	1506-1995	490
2	Andomskaya height	1737-1998	262
3	Eastern coast of Lake Onega	1736-1997	262
4	Andomskaya height (larches)	1798-1998	200
5	Bolshoy Klimenetsky isl., Lake Onega	1722-1998	277
6	Eastern coast of Lake Ladoga	1755-1997	243
7	Western coast of Lake Onega	1549-1997	449
8	Paanajarvi	1647-1997	351
9	Lojmola	1743-1996	254
10	Tolvojarvi	1548-1998	451
11	Vendjury	1734-1997	264
12	Pryazha	1606-1998	393
13	Shotozero	1770-1993	224
14	Sjamozero	1720-1998	279
15	Kinejarvi	1809-1997	189
16	Khibiny	1559-1998	440
17	Khibiny (fir-tree)	1776-1998	223
18	White Sea	1762-1998	237

Comparison of dendrochronological scales of a middle taiga zones both of Middle Ural (Shijatov, 1986) and Southern Karelia have shown the high synchronism and synphasic of descending changes in tree-rings increment of pines, located on the big distance (thousand kilometers and more) from each other. The coefficient of linear correlation between this two time series is not high (r = 0,3), however qualitative resemblance of this curves indicates a manifested coincidence of processes, which one exerted influence on change of tree-ring increment, apparently, having climatic components.

Thus, it is possible to approve that during the 450-year period the climatic processes in the territory of northern-west part of Russia, down to Ural, wore synchronic nature. Both on Middle Ural, and in Eastern Fennoscandia operating one limiting factor does not show, a composit complex of climatic factors are influence on increment of tree-rings, and first of all are mean air temperature and precipitation during the vegetation period of year.

Characteristic of a radial increment of woody plants growing in different parts of taiga forests is the availability of enough regular and well fixed perennial oscillations of different duration. In the majority dendrochronologic time series obtained in our researches, there are some cycles of different duration, which one being imposed against each other, extremely complicate their selection and interpretation. For construction of dendrochronologic time series in the present investigation the arbors age per 150-350 years were used. Therefore only cycles with duration not more than few decades were reliable estimated. The obtained indexes were subjected to a statistical analysis in comparison to time series of air temperature (May-August), instrumentally obtained on the 5 climate stations (Petrozavodsk, Olonets, Sortavala, Loukhi and Kestenga).

The correlation analysis shows that there is a gentle climate signal in the series of year increment of timber. The correlation coefficients are in the interval from r = 0.20 up to r = 0.35. The different models of time smoothing were applied with the smoothing periods at 3, 5, 7, 11, 21, and 31 years. Thus the correlation coefficients was increased up to r = 0.48 – 0.53. Nevertheless, the visual analyses shows a matched course of the 5-year smoothed diagrams both dendrochronologic time series and series of the air temperature for the warm period of year, coefficient of synchronism thus reaches 0.75 (Fig. 12).

Fig.12. Time series (5-year moving averages) and linear trends of tree-ring indexes for the sub-zone of middle taiga (1) and mean air temperature for May-August in Petrozavodsk (2).

The stationarity of generalized dendrochronologic time series for the territory of Eastern Fennoscandia and air temperature time series for the climate station Petrozavodsk was estimated with using the auto-correlation function. The fluctuations of these function show clearly, the auto-correlation function long does not dump, that confirms the quasiperiodic nature of the oscillations. Thus, the process of increment of timber for the studied area is steady enough, each value of annual increment correlates with previous, at least, with a correlation coefficient r = 0.75, and it is possible to made conclusion about enough high degree of predictability of process.

With the purpose of detection of relationship between both air temperature and tree-ring increment time series, and for the determination of possible common cycles their mutual correlation analysis was make also. The spectral analysis and the reciprocal relationship in frequency area with using of the function of coherency and phase difference estimation of the time series were made. The resemblance in long-term changes of tree-ring-type chronologies and air temperature for the warm period of year in the studied territory shows in a spectral density analysis, which one demonstrates significant cyclic changes in a increment of timber with duration about 40 and 13 years and significant of mean temperature for May-August on the same frequencies.

The contribution of these oscillation makes about 35-40 per cent in a dispersion of a increment of timber and air temperature accordingly. The smaller conterminous are marked also in the field of frequencies conforming to time scales about 19-22 years and about 5-7 years. In a spectrum of tree-ring time series the peaks of frequencies in the field corresponds to 9-11 years are visible also. The greatest values the coefficient of coherence of two parsed spectra reached in area of 19-22 years oscillations, the 13-year cycles are less significant.

The well-marked peaks in spectra corresponds to 40-year quasiperiodic oscillations. In spectra of other chronologies, obtained for Southern Karelia, the periods about 11, 13, 18, 22, and 29-30 years are present. Close to century and century oscillations were estimated only for tree-ring chronologies, which one have age more than 300 years. However century cycles for the majority of old-age pine timbers are not always fixed. Only in separate dendrochronologic series of a middle taiga zone of the Eastern Fennoscandia the cycles closed to 110-year were found. This cycles can be united from two 50-year cycles, as it was marked by Shijatov (1986). Some cycles with the super-century duration are marked in the studied area also.

Fig.13. The spectral density function for the generalized tree-ring chronology of the Eastern Fennoscandia

The cycle with duration about 4750 years found in the select dendrochronologies of the taiga sub-zone of Eastern Fennoscandia also described by Shijatov (1986) for the Ural sub-region of Polar, sub-Polar and Southern Ural. This cycle, closed by 3035 years, usually concerns to the so-call Brickner cycle (Shnitnikov, 1968; Shijatov, 1986). The cycle with duration about 2122 years most often meets in oscillations of tree-ring increment (Bitvinskas, 1974; Lovelius, 1973, 1979; Stupneva, Bitvinskas, 1978; Shijatov, 1986), and detecting of this cycle in geophusical and biological phenomena the majority of investigators connect to reacting to the conforming oscillations of solar activity (Shnitnikov, 1968; Maximov, 1995). The cycles, detected in our researches, coincide the data obtained by Stupneva (1981) for the northern and southern sub-zones of taiga of the Eastern Fennoscandia. As result of application of a method of auto-correlation for analysis of tree-ring dendrochronologies of studied area the next conclusions were made:

• the stationarity in oscillations of dendrochronologic scales depends on a degree of extremeness of conditions of a habitation and effect of the exogenic factors.
• at decomposing of the obtained frequencies of the dendroscales on time of their development it is possible to mark, that the cycles within periods of 11-22 years dominate at the end of XVIII and at the first half of XIX c. At the end of XIX and beginning of XX centuries there is a rearrangement of oscillations of tree-ring increment to the side of more high 5-7 years cycles.
• at XX century the oscillations close to century and century duration strengthen.

Thus, because the described cycles meet in the majority of obtained dendrochronologies, it is possible with a large degree of probability to suspect, that all called cycles have exogenic nature. But it is very difficult to observe all links of solar-earth relationships, and it is possible only to state their influence on the timber increment. The qualitative comparison of a solar activity (presented by Wolf's numbers) and tree-ring increment time series for the period from 1700 up to 1984 years was made also. Visually it is possible to mark noticeable resemblance of a radial increment of year rings of obtained time series and oscillations of a solar activity.

For a quantitative estimation of the relationship the correlation coefficients were calculated, which one equals from 0.32 up to 0.50 for different cases, both without smoothing and for smoothed time series. Thus, the relationships between a solar activity and tree-ring increment shows quantitative not clearly.

Data analysis of dendrochronologic time series showed that the tendencies of climate variability may be defined at the level of internal to century cycles. The present researches allow to made a conclusion about existing quasiperiodic variations of climate. Probably, there is a dependence of tree-ring increment from joint influence of climatic factors (air temperature for the warm period of year, temperature and humidity of soil, precipitation etc.). It is not obviously possible yet to select of one universal factor clearly influencing on the radial increment of trees in investigated territory.

For detection of dynamics of the natural and anthropogenic processes at inshore geocomplexes of the largest water bodies of the studied territory the fieldworks for collecting and testing tree-ring samples of a pine (Pinus sylvestris L.) and a fir-tree (Picea abies L.) growing near to Lake Onega were make at 1997-1999 years. From all this samples, selected in inshore complexes, including insular, 108 samples were selected for analysis.

The dendroindication of the samples in comparison with the water level dynamic of Lake Onega did showed that years with the maxima of increment of tree-rings of arbors, growing nearby the lake, are in accordance to the years with the high average water level (1888, 1889, 1894, 1910, 1924, 1936, 1946, 1955, 1958, 1962, 1977, 1979, 1984, 1988). The years of minimum accretions coincide with years of minimum horizons of water level (1883, 1891, 1908, 1914, 1921, 1928, 1934, 1940, 1945, 1947, 1948, 1951, 1956, 1960, 1965, 1972, 1980, 1985, 1986, 1990). It is necessary to note, that coincide not only years of extremes in both tree-ring and water level oscillations, but also amplitudes of these changes.

The considerable increase of a water level conducts to the greater increment of tree-ring, as well as decrease of water horizons produced a reverse reaction of a model arbor. For example, in 1940 the lowest value of water level of Lake Onega was marked, and per the same year the most thin ring (0,15 mm) for a model arbor was marked, the mean increment for the long-term period for this time series was 1,37 mm.

Having compared the diagrams of water level and increment of tree-rings, we have marked the fact of coincidence of oscillations during period of a natural water regime of Lake Onega, as well as after 1952, when the Svir hydroelectric power station in the outflow of Svir river was built, and natural regime of the lake was changed. Since1956 the oscillations of water level became less, and since the same year the notably reduction of amplitude increment of tree-ring is marked. The resemblance of a model diagrams is watched for arbors growing in inshore geocomplexes, remote from each other on 50-100 km.

The matching of oscillations of year tree-rings increment of arbors growing in immediate locality from water, with level variation give possibility to restore the level regime for the period from which the instrumental water level observations are absent. In judgement of Galazij (1967), the woody plants settle on pebbly barrier beaches in 10-15 years after their formation. A closed-grained substratum, sand and loam soils populate by arbors usually within 2-5 years.

Therefore, after analysis samples taken at the inshore zone of Lake Onega, we have received following conclusions : the maiden barrier beach of the lake was derivated in 1825-1828, the average water level in this period was, in interquartile, on 0,4-0,5 m higher, than modern; the second barrier beach, adjusted for an initial increment of timber, was formed in the beginning 1520 years. Therefore, the average water level of Lake Onega at this period was even higher on 1,5-2,0 m. The variations of water level in the different parts of the lake are concordanced and both time of formation of barrier beaches and altitudes of a water per elapsed centuries are corresponds to them.

The obtained reconstruction of an average water level of Lake Onega in the beginning of XVI century is agreed with the conclusions make by Shnitnikov (1968). He has obtained, that during XIV-XVI centuries the long-time increasing of the common moistening, at least, in the Northern hemisphere, was marked. In this period in a water areas of the Baltic sea and other large water bodies the often strong gales blasting piers reconfiguring shore lines were characteristic. The forming of powerful barrier beaches around Lake Onega, which has been dated by us as formed at 20-30 decades of XVI century, is connected, is most interquartile, with strong gales of the given epoch. These barrier beaches are specially great in south-east part of a lake, where their relative altitudes makes about 2-2,5 m. The outcomes about the long-term changes of a precede regime of Lake Onega, obtained with using of dendrochronologic method, are agreed with the outcomes relatively to lakes of Finland by Alestalo (1971).

The joint analysis of dendro- and climatical series has shown that in increment of timber in the Eastern Fennoscandia there is a relation with climatic and other geophysical factors. However, universal exogenic effect producing an enough strong mating signal in variability of width tree rings for taiga sub-zones of studied territory in not revealed. It is necessary to note that temperature of the warm period of year (June-August) exerts the most significant influence on oscillation of growth of pine timber stands in different sites of Karelia and southern part of the Kola Peninsula. Besides, the relationship between an increment of annual rings and solar activity variations is detected, which one, probably, influences on growth of a wooden plants indirectly.

In dendrochronologic series of a sub-zone of northern taiga the noticable displacement of a cyclicity in low frequency band is observed, and the cycles with duration at 30-35 and 110 years are detected. Secular cycles in regions of northern limits of a lignosa propagation are marked in tree-ring chronologies of the Taimyr (Lovelius, 1970, 1979), Polar Ural (Shijatov, 1986), Ural-Siberian sub-Arctic (Vaganov et al., 1996), Middle Siberia (Panjushkina, 1997) and others.

In the most chronologies of middle taiga, 40-, 19-22-, 13-, 9-11-, and 5-7-year cycles are found, the nature of which one is conditioned, in all probability, by oscillations of mean air temperature for May-August, because their periods are characteristic for cycles of both temperature meteorological and solar activity time series. However, for stands of trees of a taiga zone of the Eastern Fennoscandia the applicability of dendrochronologic forecasts of meteorological elements is possible while only at a level of an estimation of the tendencies and qualitative changes.

The joint analysis of indirect indicators of climate changes gives the basis to suppose, that the existence of "small ice drift epoch" or "stage of Fernay" in the Eastern Fennoscandia at XIV-XIX centuries also took place. In this period there was a cooling-down and humidifying of a climate, that was result to increasing of total moistening and rising of water levels of lakes, in particular Lake Onega.

The coupled analysis of dendrochronological and climatic series has shown that there is a relationship between climatic and other geophysical factors in increment of timber in the Eastern Fennoscandia. However, the universal exogenic effect producing the strong mating signal in the width of tree rings for taiga sub-zones of studied territory is not revealed for the territory of Karelia.

References can be found here.