Tree growth is affected by various environmental changes such as increased atmospheric CO₂ concentration, temperature, atmospheric pollution, and nitrogen deposition (Chmura et al. 2011, Ito et al. 2011). The effects of these environmental factors on tree growth have been studied worldwide. Forest decline in Asia has been linked to acid precipitation (Hirano et al. 2007), air pollution and global warming (Woo 2009).

Impacts of atmospheric NO₂ concentration on forest decline of Pinus densiflora in Japan have been reported by Kume et al. (2000) and decreases in annual ring width of Pinus thunbergii stands exposed to SO₂ concentration have also been reported by Kim and Fukazawa (1997). A threat to C. japonica and Chamaecyparis obtusa forests in Japan from soil acidification caused by increased nitrogen deposition have also been noted (Ito et al. 2011). As the effects of atmospheric pollutants and global warming are a consequence of long-term chronic exposure, gradual changes in forest growth, structure, and composition are observed (Kume et al. 2000, Chmura et al. 2011).

In Korea, tree ring formation and growth are considered useful sources of information for understanding the effects of environmental changes. Historical changes in tree growth can be evaluated from annual ring area as better tree growth produces a larger ring area. A few studies have already investigated relationships between annual ring growth and environmental changes in temperature, precipitation pH, and atmospheric pollution for Pinus densiflora stands (Park and Yadav 1998, Choi et al. 2005, Kwak et al. 2009, Kwak et al. 2011). Information on growth responses of forest trees to historical environmental changes plays an important role in understanding how these changes have affected tree growth.

In this study, environmental factors that constrain or stimulate annual tree-ring growth of C. japonica (Japanese cedar) are concerned such as temperature, precipitation, atmospheric pollutants, and precipitation pH in southern Korea. Cryptomeria japonica is a fast growing species. It is widely planted in the warm-temperate forest zone, and plays an important role in the forests of Korea. This species appears to have promising potential for chronology-building and climatic reconstruction as it shows a sufficiently strong correlation between tree growth and climatic factors (Kojo 1987). Investigation into how environmental changes have historically affected annual ring area in different regions is one option for evaluating tree growth.

The objective of this study was to compare annual tree ring growth at two study sites in South Korea. The effects of climate factors (temperature and precipitation) and air pollution (acid precipitation, NO₂, SO₂, O₃ and CO₂ concentrations) on annual ring growth were examined.

[Table 1.] Characteristics of location and tree ring growth in the study sites

Characteristics of location and tree ring growth in the study sites

[Fig. 1.] Geographical locations of Cryptomeria japonica stands.

The study was performed at two sites: Haenam (34˚21′28″N, 126˚33′38″E) and Jangseong (35˚27′19″N, 126˚47′01″E) in southern Korea (Table 1, Fig. 1). The soil at these sites is classified as shallow gravelly silt loam with thin brown to dark brown color (Rural Development Administration of Korea 2000) with an average pH of 4.80 in Haenam (range: 4.66 to 4.98) and 4.35 in Jangseong (range: 4.30 to 4.44); total C concentration of 26.24 g C kg^？1 in Haenam (range: 17.55 to 47.65) and 32.72 g C kg^？1 in Jangseong (range: 8.40 to 78.84); total N of 2.32 g N kg^？1 in Haenam (range: 1.34 to 3.82) and 3.09 g N kg^？1 in Jangseong (range: 1.32 to 6.18) (Table 2).

The dominant understory vegetation at the study sites were Parthenocissus tricuspidata, Quercus serrata, Pteridium aquilinum, Rhus tricocarpa, Corylus heterophylla,

[Fig. 2.] Changes in mean annual temperature and precipitation in Haenam (a) and Jangseong (b).

Styrax japonicus, Zanthoxylum schinifolium, Stephanandra incisa, Lindera glauca, Paederia scandens, etc.

The environmental data set used for the correlation analysis included climate variables (temperature and precipitation), atmospheric pollutants (NO₂, SO₂, O₃, and CO₂), and precipitation pH. At each study site, all the meteorological and environmental monitoring data were obtained at two monitoring stations. For Haenam, climate variables were obtained from the Wando monitoring station located 14 km from the study site; atmospheric pollutant concentrations were obtained from the Mokpo monitoring station. In Jangseong, atmospheric pollutant concentrations were obtained from the Gwangju monitoring station, while climate variables were obtained from the Jeongup monitoring station, which is 14 km from the study site (Ministry of Environment of Korea 2010). During the 37 year period between 1973 and 2009 for which climate data are available, the mean annual temperature and precipitation were 14.0℃ and 1496.7 mm, 13.0℃ and

[Table 2.] Soil chemical properties in the study sites

Soil chemical properties in the study sites

[Fig. 3.] Changes in mean concentration of atmospheric pollutants in Haenam (a) and Jangseong (b).

1304.1 mm in Haenam (Wando), Jangseong (Jeongup), respectively (Fig. 2).

At Haenam (Mokpo) station between 1995 and 2009, mean annual SO₂ concentration decreased from 12.0 to 4.0 nL L^？1, while mean annual concentrations of NO₂ and O₃ fluctuated between 10.0 and 20.0 nL L^？1 and between 24.0 and 35.0 nL L^？1, respectively (Fig. 3a). At Jangseong (Gwangju) station, atmospheric pollutant concentration data were available from 1989 to 2009. During this period, mean annual SO₂ concentration decreased from 21.0 to 4.0 nL L^？1, while mean annual concentrations of O₃ and NO₂ increased from 7.0 to 26.0 nL L^？1 and from 11.0 to 21.0 nL L^？1, respectively (Fig. 3b).

At Gwangju station, atmospheric CO₂ concentration increased from 370.7 ppm in 1999 to 392.5 ppm in 2009 (Fig. 4a). The mean precipitation pH at two sites showed a decreasing trend from 5.8 to 4.5 in Haenam (Mokpo) and from 5.5 to 4.9 in Jangseong between 1992 and 2009 (Fig. 4b).

A 20 × 20 m plot was established at each study site and randomly selected trees were cut down to collect tree-ring disks (Table 1). Three tree-ring disks were collected at breast height at each site. The ages of tree disks at the study sites were determined and ranged from 35- to 37-years old in Haenam, and from 32- to 34-years old in Jangseong.

Tree ring samples were sanded and polished to measure

[Fig. 4.] In mean CO2 concentration (a) and precipitation pH (b) in study sites.

the ring width. The ring width measurement was based on four radii for each disk sample. The radii were crossdated to identify to exact year in which each tree ring was formed (Baillie and Pilcher 1973). The ring widths of the disk were measured along each radius with an accuracy of 0.01mm, using the program CDendro and CooRecorder 7.4 (Cybis Elektronik and Data AB, Salsjobden, Sweden). The annual basal area increments were calculated assuming that tree rings are concentric circles to minimize agerelated growth trends of ring width and were used as an indicator of tree growth (Choi et al. 2007). The annual ring growth was calculated as an average for each region.

Response and correlation function analysis (Blasing et al. 1984) was used to examine relationships between annual ring growth and climate variables (temperature and precipitation), atmospheric pollutants (SO₂, NO₂, O₃, and CO₂), and precipitation pH (Kwak et al. 2011). Response function analysis is a form multiple regression. Because variation in annual ring width was likely to be affected by multiple environmental factors, these relationships were explored by Pearson correlation analysis.

All the response and correlation functions were determined for the period 1973？2009, which was common to both tree-ring data and the regional climate records. For relationships between air pollutants and annual ring area, the response and correlation functions were determined

[Fig. 5.] Annual ring width growth (a), and annual ring area (b) of C. Japonica in the study sites.

for SO₂, NO₂, and O₃ from 1995 in Haenam, and 1989 in Jangseong, for precipitation pH from 1992 in Haenam and Jangseong, for CO₂ concentration from1999 at both sites. Therefore, the determination of these relationships was conducted using annual data.

All statistical analyses were performed using SPSS 11.5 statistical software package (SPSS, Chicago, Illinois, USA). The level of significance for all statistical tests was an α value of 0.05. The significance of annual trends of tree ring growth was assessed by the analysis of time series using year as an independent variable.

Annual ring growth rates among the two sites were not significantly different (p > 0.05), but the growth of annual rings within each site was significantly different (p < 0.01). Annual growth ring widths at all sites showed a decreasing pattern over time (Fig. 5a). Variation in annual ring widths in Jangseong had the high growth rate of 0.09 mm compared to that in Haenam. Mean tree ring widths were 3.57 mm in Haenam and 3.66 mm in Jangseong (Table 1).

Annual ring area in the all sites increased over time (p > 0.05). However, annual ring area in Jangseong increased up until 1998 and thereafter it decreased gradually, whereas in Haenam it increased gradually (Fig. 5b).

Annual ring growth in Jangseong was positively correlated (p < 0.05) with temperature; this correlation was not significant at Haenam (Table 3). Annual ring growth at both sites was not affected by precipitation. For regression analysis, the significant correlation of temperature with annual ring growth in Jangseong was not useful for estimating past temperature because the correlation coefficient was too small (data not shown).

The effects of atmospheric pollutants such as acid precipitation, NO₂, SO₂, O₃ and CO₂ on the tree-ring growth differed between sites. Annual ring growth in Haenam was negatively correlated with SO₂ concentration (p < 0.01) but not related with NO₂, O₃, and CO₂ concentrations. Annual ring growth in Jangseong was positively correlated with NO₂ concentration (p < 0.05) and negatively correlated (p < 0.05) with SO₂ and CO₂ concentrations (Table 3). Precipitation pH was negatively correlated (p < 0.01) with annual ring growths in Haenam; there was no effect at Jangseong (Table 3).

Inter-correlations of annual ring area (y) with air pollution and acid precipitation were assessed using linear correlation for multiple regressions at the two sites. In Jangseong, y was positively correlated with CO₂ and NO₂ (y = - 0.47 CO₂ ？ 0.61NO₂ + 212.77, r² = 0.90, p < 0.001). In Haenam, this equation was correlated with precipitation pH (y = -4.91pH + 44.0, r² = 0.40, p < 0.01).

[Table 3.] Pearson correlation between annual ring area and environmental variables in the study area

Pearson correlation between annual ring area and environmental variables in the study area

Annual ring growth varied markedly among the trees (n = 3) at each site, but the increment in annual ring growth was not different between sites (n = 2) (Fig. 5, Table 1). This suggests that variation in annual ring increment of individual trees depends on the environmental factors at each site such as climate, air pollution, and soil fertility. A positive correlation of temperature with annual ring area increment of trees in Jangseong was suggested that annual ring growth increased with increasing temperature. Lebourgeois et al. (2005) suggested that the positive effect of increased temperature on tree-ring growth might be related with soil water capacity. In Jangseong, mean annual precipitation was 1,304 mm (Fig. 2b), but this was not correlated with annual ring area. Thus, this site might have adequate moisture for tree growth. However, the lack of correlation between temperature and precipitation at this site may be due to the periods of water limitation (Savard 2010). Temperature changes between 1973 and 2009 were about from 0.1℃ to 1.5℃ which were about 1℃ lower comparing to temperature in Haenam (Fig. 2a). Increases in temperature might be directly affecting tree growth and indirectly through interactions with other stressors and disturbances (Chmura et al. 2011), but at Haenam, tree growth was not affected by temperature. The correlation between temperature and annual ring area in Jangseong has analyzed the regression model. Temperature has been claimed to be relatively unimportant for tree growth. Increase in tree growth in Jangseong may be extending the combination of increases in nitrogen deposition, CO₂ concentration, and temperatures (Hyvonen et al. 2007, Bytnerowicz et al. 2007).

There were significant correlations between atmospheric pollutants and annual tree-ring growth at Haenam and Jangseong. The negative correlation between SO₂ concentration and mean annual ring increment in Haenam and Jangseong (Table 3) suggested that annual ring growth increased with decreasing SO₂ concentration. In a previous study, decreases in atmospheric SO₂ concentration due to national policy were not considered high enough to affect tree growth (Kume et al. 2000). Therefore, the significant effect of SO₂ concentration on increased growth of annual ring area in Haenam and Jangseong could be due to the acid deposition (Kwak et al. 2011). A significant relationship between NO₂ concentration and annual ring area of trees in Jangseong (Table 3) could be attributed to annual ring area increment increasing with NO₂. In an earlier study, the rate of forest decline of P. densiflora in Japan was negatively correlated with atmospheric NO₂ concentration (Kume et al. 2000). The positive effect of NO₂ concentration on annual tree ring area in Jangseong may be attributed to nitrogen deposition that can have various effects on forests (Kwak et al. 2011, Bytnerowicz et al. 2007). Increasing CO₂ concentration between 1999 and 2009 (Fig. 4a) affected annual tree-ring growth of C. japonica at Jangseong. The influence of increased CO₂ on annual ring growth could be direct effects on physiological processes of photosynthesis, respiration, and transpiration (McPherson et al. 1994, Chapin III et al. 2009, Chmura et al. 2011). Increases in CO₂ concentration may reflect both growth and water use efficiency of plants (Bytnerowicz et al. 2007). In this study, annual ring growth decreased with increasing CO₂ concentration because of stomatal closure or assimilated by leaves that activates photosynthetic enzyme (Kwak et al. 2009, Choi and Lee 2012). This result was consistent with the study of Clark et al. (2003) who found reductions in tree growth with increased atmosphere CO₂ concentration in tropical rain forest. The significant correlation between CO₂ concentration and annual ring area at Jangeong suggests that elevated atmospheric CO₂ might be attributable to changes in tree growth rings (Yazaki et al. 2004).

For regression analysis at Jangseong, the combination of both CO₂ and NO₂ concentrations was positively correlated with annual ring area; this correlation suggested that annual ring growth increased with increased CO₂ and elevated nitrogen deposition (Bytnerowicz et al. 2007). N deposition originated from NO₂ (Kwak et al. 2011, Bytnerowicz et al. 2007) is claimed to be the most important factor of the increased tree growth. In effect, total N concentration in soil in Jangseong was 3.09 g kg^-1 and it was higher than that in Haenam (Table 2); this may be the reason why annual ring widths in this site were larger than that in Haenam. Increased CO₂ could lead to stomata closure to protect the plant from air pollution (Bytnerowicz et al. 2007). Therefore, integrated effects of elevated CO₂, N deposition and climate changes on tree growth may be linked to acid deposition (Bytnerowicz et al. 2007) because of soil acidification with an average pH of 4.35 (Ito et al 2011) in Jangseong.

Acid deposition is formed from NO₂, SO₂, and O₃ due to the burning of fossil fuels. At Haenam, the negative correlation between precipitation pH and annual ring area (Table 3) may reflect decreases in annual ring growth with increasing precipitation pH between 1992 and 2009. This correlation may also be linked to nitrogen deposition (Kwak et al. 2009b, Kwak et al. 2011) due to the H⁺ input from acid rain (Shan 1998).The nitrogen content from acid precipitation can be sufficient to stimulate tree growth (Shan 1998). The decreasing precipitation pH (Fig. 4b) may seem to reflected increased N deposition originated from NO_x emission that is known to be depleted in ¹⁵N relative to the soil mineral N due to soil acidification (Kwak et al. 2011). The soil was acidic with pH of below 5.5 (Pritchett & Fisher 1987). In our study, mean pH value of soil in Haenam was 4.8 and total N concentration in soil was 2.32 g kg^-1 (Table 2), and leaching of nutrients at this site with the acid deposition may result in changes in growth (Ito et al. 2011). Long-term acid deposition may also have altered forest structure and function (McNulty and Boggs 2010). The effect of increased precipitation acidity on decreased tree ring growth may be attributed to changes in forest nutrient cycles such changed soil nitrogen dynamics (Choi et al. 2007). The significant correlation between precipitation pH and annual ring area in Haenam also suggests that tree ring data might be useful as an indicator of precipitation pH. Kwak et al. (2009b, 2011) also suggested that tree ring chemistry of P. densiflora may be used to estimate historical precipitation acidity in Korea. Therefore, the combination between SO₂ concentration and precipitation pH may form the acid deposition and affect tree growth of C. japonica in Haenam.

1. At two study sites in southern Korea, annual ring growth of C. japonica was correlated with environmental factors and annual increment of ring area was not different between regions. However, the growth rate of annual ring width in Jangseong was higher than that in Haenam because of the higher mineral N concentrations in the soil.

2. At all sites, the annual tree-ring growth was affected by integration of climate change and atmospheric pollutants in the forest ecosystem of Jangseong. The positive correlations of temperature and NO₂ concentration with annual ring growth of trees could suggest that the ring growth increased with increasing temperature and nitrogen deposition in Jangseong. In addition, the negative effects of SO₂ and CO₂ concentrations on tree-ring growth might be reflected in the increase in annual ring area with decreased atmospheric SO₂ and CO₂ concentrations due to national policy and the respiration of plants by inhibiting photosynthesis in Jangseong. Therefore, the majority of increased annual ring growth of trees in this site could interact among increased temperature, nitrogen deposition, and CO₂. However, effects of atmospheric SO₂, NO₂, CO₂ on annual growth rings were affected by air pollution in forest ecosystem. The combination of effects of atmospheric pollutants and temperature changes on tree growth might be the problem of acidification in the forest ecosystem of Jangseong.

3. In Haenam, a negative correlation of tree-ring growth with SO₂ and precipitation pH might be reflected in the nitrogen deposition, and these correlations may be changes in forest nutrient cycles caused by the acid rain. Precipitation pH affected the annual growth of trees in Haenam and annual ring growth might be useful as an indicator to estimate precipitation pH.

4. These results suggested that annual ring growth of C. japonica in Jangseong was more correlated with environmental variables than that in Haenam. However, the radial growth at two sites might be affected by changes in nitrogen deposition and climate. The further growth of C. japonica forest is at risk from the long-term effects of acid deposition from fossil fuel combustion. To sustain forest productivity in each site, changes in silviculture practice such as forest soil management practice may be employed in future studies.