Effect of Climatic Conditions on Tree-Ring Widths in Black Locust ( Robinia pseudoacacia L . ) in the city of Wroc ł aw

The response of trees to weather conditions, expressed in the tree-ring widths, depends on the extent of urbanization in the area. Specifi c climatic conditions in an urban heat island can be expected to result in growth differences to trees growing in non-urban areas. Accordingly, the aim of this study was to characterize the effect of climatic conditions on tree-ring widths of black locust in the city of Wrocław. Materials consisted of wooden discs taken from felled trees at four sampling sites (4 streets of Wrocław) at a height of 1.3 m from the ground. Meteorological data were obtained for the period 1971-2013 from the Institute of Meteorology and Water Management (IMGW) in Wrocław. Analysis of the multiannual period showed a signifi cant negative trend in annual ring widths of black locust in the city of Wrocław. A code WROB was assigned to the chronology. The annual ring widths averaged 3.4 mm, ranging from 1.6 to 5.6 mm. In the urban conditions of Wrocław, the air temperature and precipitation signifi cantly infl uenced the annual ring widths of black locust. The results indicate the need for further research on a larger number of samples.


INTRODUCTION
1. UVOD Black locust (Robinia pseudoacacia L.) was one of the fi rst North American forest species introduced in Europe ( Boring and Swank, 1984).Its natural range includes the eastern and central parts of the United States: the Appalachian Mountains, the Ozark Plateau and Ouachita Mountains (Feldhake, 2001;Call, 2002;Zajączkowski and Wojda, 2012).In the area of modern Poland, black locust was fi rst planted in the 19 th century.Since then, it has spread in Poland, not only through artifi cial plantings but also via self-seeding (Ślusarczyk, 2012).Currently, it can be found almost everywhere in the country, with the highest concentrations in the western part, in both deciduous and coniferous forests (Wojda et al., 2014).Outside forest ecosystems, black locust can also be found in buffer strips, roadsides, shelter belts and city alleys (Łukaszewicz, 2010).It is especially appreciated in the reconstruction of historical parks and gardens (Majdecki, 1993).In the light of recent research, black locust can also be recommended as an energy crop in harsh habitat conditions, where it can be competitive with other energy crops (Rédei et al., 2010;Böhmi et al., 2011; Kraszkiewicz, 2013; Stolarski et al., 2013).
The response of trees to weather, especially the air temperature and precipitation, expressed in treering widths and the course of phenological phases, changes under the infl uence of human impact in the area (Walkovszky, 1998;Gillner et al., 2014).Trees from urban heat islands are expected to show differences in this respect compared to trees growing in nonurban areas.However, the dendroclimatological analysis of trees from urban areas has been poorly researched so far (Cedro and Nowak, 2006;Meier and Scherer, 2012).Namely, most commonly studied are trees growing within their natural ranges in the open landscape, and far less often trees that can be found outside the natural range (Tsakov and Alexandrov, 2005;Feldhake, 2001;Zhang et al., 2013;Čufar et al., 2014).
Therefore, the aim of this study was to characterize the time course of the tree ring widths of black locust in the urban environment of the city of Wrocław and assess its relations with the air temperature and precipitation.

MATERIJAL I METODE
Samples were made from 16 black locust trees located in four streets in Wrocław: Sucha (WS), Długa (WD), Witelona (WW) and Krucza (WK) (Figure 1).Individual sequences of ring widths are given in parentheses.Each tree was assigned a symbol and a number (1 to 16); for example WK15 and WK16 denoted trees felled in Witelota Street (Figure 2).Habitat conditions were similar in all sampling sites, with a predominance of surface covered by concrete and tarmac, etc., and low share of soils with open access to rainwater and sunlight.Discs were taken from trees at a height of 1.3 m above the ground.After drying and polishing the discs, the widths of tree-rings were measured in millimetres.The planned felling of the trees in 2014 provided the opportunity to obtain the disc samples, based on administrative decisions issued by the authorized body of the City of Wrocław.Direct sampling was performed by the arborysta.comcompany.
Meteorological data for the period 1971-2013, i.e. the average air temperature (Ta -analyzed year, PTa -previous year, °C) and total precipitation (Rfanalyzed year, PRf -previous year, mm) were obtained The individual sequences of tree-ring widths in black locust were used to assess similarities in the chronology of the analyzed trees (Figure 2).The most outlying sequences, WS1 and WS2, were excluded from further analysis due to a weak relationship with the pattern observed for the entire group of 16 trees.Local chronology was then established based on a correlation analysis and Student's t test performed on the remaining group of 14 trees.The sequences of ring widths were used to determine the local chronology.The individual sequences of 14 trees, as well as a local chronology of black locust WROB, are presented using the following statistical indicators: mean values (x̄), standard deviation (Sd) and minimum (Min) and maximum (Max) values.Additionally, for annual ring widths, their prevalence was determined in one millimetre intervals, while for the local chronology the following parameters were determined: skewness, kurtosis, fi rst-order autocorrelation (r a ), correlation coeffi cient (r) for linear trend, deviations from the long-term average in the period 1971-2013 (ΔWROB), and differences between successive years of WROB tree ring widths in the local chronology (ΔyWROB).
To evaluate the effect of the air temperature and precipitation on the annual ring width of black locust, master chronology was used, i.e. the arithmetic average of the 14 ring width sequences of the analyzed tree species.First, the chronology was transformed into an indexed chronology by removing the linear trend and long-term fl uctuations, according to the formula proposed by Fritts (1976), and also after removing the fi rst order autocorrelation in the residual chronology (Koprowski, 2006; Speer, 2010; Bijak et al., 2012).
The evaluation of the accuracy of the relationship between annual ring widths and air temperature and precipitation in the period 1971-2013 was determined using a Pearson's correlation coeffi cient (Dobosz, 2001).Assessment of the combined infl uence of meteorological conditions was performed using the multiple stepwise regression equation.Multiple regression parameters were determined by the method of least squares.Fit of the regression to the empirical data was measured by the coeffi cient of determination R 2 (%) with the error of the regression equation Sy (mm).A variable partial correlation analysis was used to determine the contribution of each of the selected factors to the prediction of the explanatory variable.The calculated partial correlation coeffi cients were squared and expressed in % (as r 2 ).In the correlation and regression analyses, the dependent variable was annual tree ring widths in the form of residual chronology, while the independent variable was meteorological elements from the months of the current year (from January to September) and the previous year (June to December).Similar periods were taken into account in dendroclimatic studies by Fritts (1976), Cedro and Nowak (2006) and Koprowski (2012).Statistical analyses were performed using STATISTICA 10.The scientifi c name of the analyzed tree species was taken from "Dendrology" by Seneta and Dolatowski (2009).

REZULTATI
The local chronology of black locust with the signature WROB for four sampling sites (WS, WD, WW, WK) located in the main streets in Wrocław was 134 years in length and represented the period 1880-2013 (Figure 2, Figure 3).The longest dendrogram (WD7) equalled the local chronology, while the shortest ones (WS5, WW13, WW14 and WK16) were 43 years in length .In the period 1971-2013, the tree-ring widths were widest in the tree WS3 (6.0 mm) followed by the trees WS5 and WK16 (4.4 and 4.2 mm, respectively) (Figure 4).The variability of annual ring widths, expressed as standard deviation, ranged from 0.7 mm (WS7) to 2.6 mm (WS3).Both the maximum and minimum annual ring widths in individual trees included in the local chronology varied signifi cantly.Maximum annual ring widths ranged from 3.7 mm (WS6) to 12.4 mm (WS3), whereas minimum ring widths ranged from 0.5 mm (WD8) to 2.6 mm (WS3).
Diagrams showing the incidence of tree-rings in the adopted ranges revealed that over 77 % of all growth rings obtained in the years 1971 to 2013 ranged from 1 to 5 mm (Figure 5).Lower (<1 mm) and higher ranges (from 5 to 6 mm) showed a signifi cantly lower incidence (about 5 % and 8 %, respectively).Growth rings wider than 6 mm occurred in 9 % of cases, while those >10 mm occurred in only 1 % of cases.The lower quartile for each dendrochronological curve ranged from 0.9 mm (WD8, WW9) to 3.9 mm (WS3), and the upper quartile from 2.5 mm (WD7) to 8.4 mm (WS3).The interquartile range varied from 1.0 (WS6) to 4.5 mm (WS3) and showed a moderate variability of annual growth rings among the individual sequences.
The average tree-ring width of black locust WROB for the period 1971-2013 was 3.4 mm, with a standard deviation of 1.2 mm (Table 1).The narrowest absolute  In the fi rst 18 analyzed years (i.e.period 1971-1988), the deviation from the average ring width in local chronology WROB was positive and ranged from 0.1 mm in 1987 to 2.2 mm in 1974 (Figure 6).A nega- tive deviation was observed in the period 1989-2013 (with the exception of 1990, in which the deviation was 0 mm); the highest deviation was observed in 2013 (-1.8 mm).Differences were also found between adjacent years ranging from -0.7 mm between 1979 and 1978 to 0.8 mm between 1974 and 1973 (Figure 7).The differences, both positive and negative, were distinctly higher in the fi rst half of the analyzed period.In 14 years, growth ring widths were wider than in the previous year, and in only two years they were identical to the previous years.Annual ring widths of black locust in Wrocław WROB depended on the air temperature and precipitation (Figure 8).The annual ring widths in urban conditions correlated negatively with the air temperature in May-September of the analyzed year, and in December and August of the previous year, while they correlated positively with the air temperature in January-April of the analyzed year and in the period June-July and September-November of the previous year.Among the analyzed 16 months, only three months of the analyzed year demonstrated a significant effect of thermal conditions on tree-ring widths: positive in February (Ta2, r = 0.29, P ≥ 0.05) and March (Ta3, r = 0.28, P ≥ 0.05), negative in July (Ta7, r = -0.34,P ≥ 0.01).
In the years 1971-2013, precipitation positively infl uenced ring widths in February-March and May-August of the analyzed year and in the period June and September-November of the previous year, while negatively in January, April and September of the analyzed year and in the period July-August and in December of the previous year.A signifi cant relationship was found in September (Rf9, r = -0.35,P ≥ 0.01) of the analyzed year  Multiple regression analysis, showing the combined effect of weather conditions (precipitation and air temperature) on ring widths, indicated that the biggest impact was exerted by precipitation in three months: PRf11, PRf12 and Rf9, while thermal conditions proved to be insignifi cant (Table 2).Precipitation, signifi cant at p≥0.05, determined the annual ring widths in black locust by about 30 %; the greatest impact was shown by Rf9 (r 2 = 14.9 %, P ≥ 0.05).
Precipitation in months that signifi cantly contributed to a reduction in annual ring widths, averaged approximately 35 mm in PRf12 and about 50 mm in Rf9 (Table 3).The highest precipitation in PRf12 and Rf9 amounted to 100 mm in 1975 and 110 mm in 2013, with the lowest (4 mm) in 1973 and 7 mm in 1975.The precipitation in PRf11 was characterized by a significantly lower variability (by about 10-20 %) than PRf12 and Rf9, and a signifi cant negative trend in the years 1971-2013 (r = 0.35, P ≥ 0.05).The extreme levels of PRf11 were the lowest among the three analyzed months (Max = 98 mm, Min = 2 mm).

RASPRAVA
There are few papers on the variability of annual ring widths of trees in the technosphere and its dependence on meteorological conditions.This is mainly due to the diffi culty in obtaining permission from the local authorities for sampling with the use of a Pressler drill, and also due to the diffi culty in collecting the appropriate sample size from a given type of sampling site (with similar conditions of growth and development for a given tree species).Trees in urban areas, including black locust, are exposed to many adverse environmental factors infl uencing the course of their vegetation, as refl ected, for example, in tree-ring width.The most important include the deteriorated quality and quantity of soil, pollution of the ground with debris, calcium, chlorine and sodium ions, as well as altered meteorological parameters (Szczepanowska, 2001;Łukasiewicz, 2010;Rahmana et al., 2014).
Most publications in the fi eld of biology and dendroclimatology deal with trees in forests outside urbanized areas.Although research on black locust has been   1984).Wrocław averages from -9.0 °C to 4.2 °C and from 15.9 °C to 21.5 °C, respectively (Dubicki et al., 2002).The U.S. precipitation ranges between 1020-1830 mm and is on average about 250 % higher than in Poland.

ZAKLJUČAK
In the years 1971-2013, a signifi cant negative trend of annual ring width of black locust was observed in Wroclaw WROB.The annual growth averaged 3.4 mm and ranged from 1.6 to 5.6 mm, most frequently ranging from 2 to 3 mm.Differences between successive years ranged from -0.7 mm between 1979 and 1978 to 0.8 mm between 1974 and 1973.
In this study, a signifi cant infl uence of air temperature and precipitation on the annual ring width of black locust was demonstrated.Multiple regression analysis showed that the annual ring widths of the analyzed tree species were mainly affected by precipitation in November and December of the previous year and September of the analyzed year, with different temporal distributions in individual years.
The results indicate that further research should be made on a larger number of samples.

Table 3
(Boring andl., 2013)7))tics designated for preFeliksik et al., 2007;Kalbarczyk, 2010) infl uencing (in the entire system of meteorological elements) the annual ring widths in the local chronology of black locust WROB.Years 1971-2013 Tablica 3. Statistička obilježja određena za oborine (PRF, RF, mm) koje znatno utječu (u cijelom sustavu meteoroloških elemenata) na širinu goda u lokalnoj kronologiji bagrema WROB; razdoblje 1971.-2013.outforexamplebyBoringandSwank(1984),FarrarandEvert(1997),Feldhake(2001),Zhangetal.(2013),andWangetal. (2013), a completely different climate and locations in those studies do not allow for comparison of their results with our paper.In the period 1971-2013, the average annual ring width of black locust in the sampling sites in Wrocław was 3.4 mm, ranging from 1.6 to 5.6 mm.Similar values, obtained in a forest in SE Poland (Forest District Krosno), were recorded byFeliksik et al. (2007), where the average annual ring width was 3.3 mm and ranged from 1.2 to 5.9 mm.It should be emphasized that comparisons concerning average tree-ring widths are especially prone to error.The average ring width at individual sampling sites does not take into account the differences arising from the age of the trees and the habitat conditions(Cailleret and Davi, 2011).The dendroclimatological analysis of the relationship between weather conditions and annual ring width in 1971-2013 indicates a signifi cant infl uence of the air temperature and precipitation on the ring width of black locust.The air temperature, lower than the average in January-April (especially in February and March) and higher than the average in May-September (especially in July and September), adversely affected the tree-ring widths.A similar negative effect was induced by a wet September and December and a dry November.Similarly,Feliksik et al. (2007)found that cold winters and low air temperature in early spring, as well as low amount of precipitation during the summer (June-July), were the most important meteorological elements inhibiting the radial growth of black locust.The air temperature from late winter to early spring affects the physiological processes involved in the initiation of cambial cell production and in differentiation of xylem cells in the tree(Begum et al., 2013).Importantly, premature initiation of cambial activity increases the risk of damage to biomass growth because of a sudden drop in air temperature below the physiological tolerances of the tree.The climate in the natural range of the black locust is distinctly different in terms of air temperature and (especially) precipitation in comparison with the climate of south-western Poland(Boring and Swank  1984, Dubicki et al., 2002;Feliksik et al., 2007;Kalbarczyk, 2010).For example, the average air temperature in January in the eastern and central parts of the United States ranges from about -4.0 °C to +7.0 °C, and in August from 18.0 to 27.0 °C(Boring and Swank, V -coeffi cient of variability / koefi cijent varijabilnosti (%), ** signifi cant at P ≥ 0.05 / signifi kantno pri P ≥ 0,05; for further explanations see Table1/ druga objašnjenja pogledati u tablici 1. carried