Although disturbance by logging has been substantially decreased, the forest fire, still as a threatening factor to forest, has been growing in its scale according to increase of forest tree accumulation in these days. Forest of Kangwon-do is accounted for 83% of the Kangwon-do area and has excellent forest landscape. On the other hand, deforestation area by human interference was more than other province (Lee 1995).

Particularly, large forest fires have intensively occurred in the east coastal region of Korea. The biggest damage occurred in 2000 due to forest fire, simultaneously along east coastal regions (Goseong, Donghae, Kangneung, Samcheok, and Uljin-gun). This accounted for about 0.36% out of the total forest area of Korea and 80 times larger than that of Yeouido, 2.9 km² gross area (Choung et al. 2004). The most common vegetation types in the east costal region of Korea are composed of pine tree species containing the high combustible material, so it brings about the widespread forest fire with strong winds of this region. Forest fire is one of the most common disturbance factors in terrestrial ecosystems. It is well known that the frequency and intensity of forest fire influences on large changes in physicochemical properties of forest soils, the species composition of forest vegetation, and various space formation in the forest (Lee et al. 1988). Furthermore, above-ground parts of organisms and organic material is removed by abiotic environmental factors inducing changes (Mun and Choung 1996). Live fuels usually act as a heat sink during combustion until the moisture has been driven out of them: since the moisture content of dead wood is low, combustion of dead fuels drives the moisture out of living fuels (Bond and van Wilgen 1996). The extent and duration of these effects depend firstly upon fire severity, which, in turn, is controlled by several environmental factors that affect the combustion process, such as amount, nature, and moisture of live and dead fuel, air temperature and humidity, wind speed, and topography of the site. Numerous findings on the effects of fire on soil properties are available in the literature (Certini 2005). The damage of forest fire causes changes in forest structure, such as reduction of biomass, and has a fatal effect on the function of ecosystem, such as cycle of material. In addition, the detrimental effects caused by forest fires have been substantially varied physical and chemical properties of forest ecosystems, according to strength and sustainment time of forest fire, moisture content of soil, times when forest fire occurs, and strength of rain after forest fire (Chandler et al. 1983). The forest fire dramatically affects the nutrient cycling and the physical, chemical, and biological properties of the underlying soil (DeBano 1991). However, changes in chemical properties of the soils due to the forest fire are known to be restored to the previous state within two to three years after the oxidation (Woo et al.1985, Lee et al.1988, Woo and Lee 1989). In general, fire-induced changes in most of the soil nutrient cycles are slight and ephemeral except for the nitrogen (N) and phosphorus (P). The availability of these nutrients generally is increased by the combustion of soil organic matter and the increase is strictly dependent upon type of nutrient, burnt tree species, soil properties, and pathway of leaching processes (Kutiel and Shaviv 1992). Also, the fire creates various scales of spaces within the forest according to its scale, and causes change of non-biological environment factors by removing biological entities and organic materials on the ground (Mun and Choung 1996). As it affects on soil respiration, with organic materials on the ground remove due to this, importance of understandings about the relationship between forest fire and soil respiration has been expanded (Jeong 2007). Soil respiration consists of heterotrophic respiration (mainly by soil microorganism and animals) (Anderson 1982, Landsberg and Gower 1997) and autotrophic respiration (mainly by living plant roots) (Gough and Seiler 2004, Jassal and Black 2006). The process of soil respiration, a procedure of emitting CO₂ in the air, which is generated in the oxidization procedure of carbon composites, root respiration of plant, and respiration of soil animal and microorganism, acts as one of major phenomena that emit CO₂ from forest ecosystem (Chae et al. 2003). About 10% of CO₂ contained in the air is generated from soil, which accounts for about 10 times larger than emission of CO₂ according to consumption of fossil fuel (Raich and Schlesinger 1992). The study of soil respiration has been actively conducted in Korea. Kim (2006) has clarified soil carbon cycle and CO₂ emission rate from soil of the Pinus densiflora forest in Sambong Mt. of Hamyang-gun, Kyeongnam, and Lee (2011) has conducted monitoring of the amount of soil respiration in the temperate deciduous forest of Kwangneung. Lee and Mun (2001) have evaluated amount of CO₂ annually emitted in the air, soil temperature and soil moisture for soil respiration in the study of soil respiration of the acorn forest, and Moon (2004) has analyzed respiration amount according to the type of trees with comparison in the study of soil respirations of P. densiflora, Quercus variabilis, and Platycarya strobilacea in the region of Jinju, Kyeongnam, and assessed effects of soil temperature and soil moisture for the amount of soil respiration. In addition, Yi (2003) has compared differences among stands by measuring the generation rate of CO₂ in soil of Q. variabilis forest in Chuncheon, Kangwon-do, and assessed the annual generation rate of CO₂ for each season with figuring out a correlation between soil temperature and soil moisture. However, there has been a lack of studies for a correlation between emission rate of CO₂ from soil and impact factors and the annual generation rate of CO₂ from soil.

Therefore, this study is intended to conduct a quantitative evaluation of the amount of CO₂ generated from soil and to clarify the correlation between impact factors and the annual generation rate of CO₂ from soil after going through natural restoration process for 12 years in the region in Samcheok, Kangwon-do, where had been disturbed by forest fire in 2000.

The subject area for this study was located in Samcheok- si, Gangwon-do, with P. densiflora as its major vegetation, where had been widely damaged, equivalent to 70% out of the total area damaged by forest fire in 2000, in the east coastal region (Fig. 1). According to the Korea Meteorological Administration (2011) from the Kangwon Regional Meteorological Station, the annual average temperature and the amount of precipitation in this region from March 2011 through February 2012 were 12.4 °C and 1160 mm, respectively (Fig. 2). Characteristics of vegetation and soil in unburned and burned study sites were shown in Table 1. The permanent quadrat (20 m × 20 m) in both study sites was installed in 2001. Afterwards, each set of quadrats of 5 m × 5 m and 10 m × 10 m was installed in 2007 in the most representative places with homogeneity in the burned site and the unburned study site (Fig. 1). The burned site is a long-term ecological research site designated by Ministry of Environment, in the region of Mt. Geombong where is located in Imwon-ri, Wondeokeup, and where its vegetation was completely burnt by forest fire in 2000 up to crown layer. 12 years had passed since the site was damaged by forest fire, and Quercus mongolica was dominating throughout the shrub layer in there, with some distribution of Quercus serrata and Lespedeza bicolor. The unburned site is located in Dongmak-ri, Keundeok-myeon (N 37°19′42.0″, E 129°12′10.6″). Pinus densiflora was dominating in the tree layer of the unburned site, and diameter at breast height and the height of trees were 26.06 ± 7.18 cm and 15.46 ± 0.47 m (mean ± standard deviation), respectively, while Q. serrata and Q. mongolica were dominating with some distribution of Q. variabilis, Castanea crenata, and Rhododendron mucronulatum var. ciliatum.

[Fig. 1.] Map of the study area in Samcheok-si, Gangwon-do (blue dot indicates unburned study site; red dot, burned study site).

[Fig. 2.] Seasonal patterns of precipitation and air temperature in the study region from 2011 to 2012.

[Table 1.] Characteristics of the permanent research plots

Characteristics of the permanent research plots

Small quadrats (120 cm × 80 cm) were installed in the unburned and burned study sites, and the emission of CO₂ per unit soil surface area per unit time was measured in the soil by installing each 6 collars with 7 cm and 5 cm of height (taking into consideration the height of litter layer at each site) and 10 cm of diameter (being considered Infrared Gas Analyzer width) inside the quadrats (Fig. 3). The amount of CO₂ emitted from the collars was measured once a month at 10 min interval with connecting infrared gas analyzer LI-6400 (Li-Cor Inc., Lincoln, NE, USA) and the closed chamber method (Chae et al. 2003, Joo et al. 2011) was used to measure soil respiration.

[Fig. 3.] The installation drawings of collars.

Soil temperature was measured at the soil of 10 cm depth taking into consideration the forest litter layer (about 3-5 cm in depths). The air temperature at 1.5 m height was automatically measured at 10 minute interval from April 2011 to March 2012 by installing Watch Dog 1000 Series Data Loggers (Spectrum Technologies Inc., Aurora, IL, USA) at each small quadrat in the burned unburned study sites (Fig. 4). The data for Kangwon Regional Meteorological Station (Korea Meteorological Administration 2011) is quoted for the amount of precipitation.

[Fig. 4.] The installation photographs of data logger used in this study.

Q₁₀ values, the sensitivity of soil respiration to the change of temperature (Fang and Moncrieff 2001), indicate the change of soil respiration and can be assessed by temperature and soil respiration. A large Q₁₀ value means that the soil respiration is profoundly changed and sensitive according to the change of temperature. Q₁₀ values were calculated with following formula [2], after value of b was obtained with applying soil respiration rates (R_soil) and temperature (T) to exponential function [1]:

where R_soil is soil respiration rate (mg CO₂ m^-2 h^-1), T is temperature (°C), and a and b are constant.

For the analysis of correlation between the soil respiration rates and the environmental factors, such as soil temperature, air temperature, content of moisture in soil, and amount of precipitation, regression analysis and ANOVA analysis were conducted with statistical analysis program, SPSS ver. 12.0 (SPSS Inc., Chicago, IL, USA).

During the entire study period from March 2011 to February 2012, seasonal variations in monthly measured mean soil and air temperature in the unburned and burned sites were shown in Fig. 5. The annual average soil temperature in the unburned site and the burned site for the during the study period were 11.2°C and 11.7°C, respectively. The annual average air temperature in the unburned site and the burned site during the study period were 10.6°C and 10.8°C, respectively.

[Fig. 5.] Seasonal variations of measured monthly mean soil temperature (a) and air temperature (b) in the unburned and burned sites during the study period from March 2011 to February 2012.

The emission of CO₂ from soil in the burned site ranged from 41.37 ± 7.78 to 417.82 ± 54.98 mg CO₂ m^-2 h^-1, while that in the unburned site ranged from 42.66 ± 6.41 to 1175.14 ± 280.03 mg CO₂ m^-2 h^-1 (Fig. 6). In both sites, soil respiration decreased after peaking in August (Fig. 6). The annual mean emission of CO₂ from soil in the burned site was 194.51 ± 123.35 mg CO₂ m^-2 h^-1, while it was 353.00 ± 322.45 mg CO₂ m^-2 h^-1 in the unburned site, indicating about 2 times higher than that in the crown site. With ANOVA analysis of monthly CO₂ emission from soil in the unburned site and the burned site, the difference of CO₂ emission rate between both sites was statistically significant within 0.1% in months from April through November, while it was statistically significant around 5% in December; there was no significance in July and October (Fig. 6). Moon (2004) reported the annual mean soil respiration rate of 430 mg CO₂ m^-2 h^-1 at the P. densiflora forest in the region of Jinju, Kyeongnam. In this study, the similar result was obtained in the unburned site consisting of dominant P. densiflora forest stand. The amount of soil respiration showed a tendency similar to the result of Singh and Gupta (1977) indicating that it varied within the range from 10 to 1000 mg CO₂ m^-2 h^-1, with being synthesized after measurement in various forest ecosystem of temperate regions.

[Fig. 6.] Seasonal patterns of soil respiration rate for both sites during the study period from March 2011 to February 2012 (*, P < 0.05; ***, P < 0.001).

Correlation between CO₂ emission and environmental factors, such as soil temperature, air temperature, and amount of precipitation, was shown in Fig. 7 and Table 2. The relationship between CO₂ emission from soil and soil temperature showed a high significance (P < 0.001) both in the burned site (R² = 0.932) and the unburned site (R² = 0.942), as well as a high positive correlation; however, the relationship between CO₂ emission from soil and air temperature had a correlation lower than soil temperature both in the burned site (P < 0.01, R² = 0.668) and the unburned site (P < 0.001, R² = 0.729) (Table 2). This result is identical to the report from a study of soil respiration (Witkamp1966, Son and Kim 1996, McHale et al. 1998, Knapp et al. 1998, Lee and Mun 2001, Yi 2003) demonstrating that soil respiration was led by soil temperature and that there was a high correlation between them. The regression equation induced by real data of soil respiration and soil temperature was used to measure the amount of soil respiration continuously. In this study, the highest soil temperature and the lowest soil temperature were measured once a month, and a high positive correlation between soil respiration and soil temperature were exhibited. Q₁₀ values, the sensitivity of soil respiration to the change of soil temperature, are 4.572 and 2.408 in the unburned site and the burned site, respectively, and were more sensitive in the unburned site than in the burned site. These results were higher than the Q₁₀ value of 2.38 obtained by Moon (2004) at the P. densiflora forest in the region of Jinju, Kyeongnam. However, in our unburned site, the Q₁₀ value was comparable to those of 4.66 and 4.57 obtained by Kim (2008) and Noh et al. (2010), respectively, in the P. densiflora forests. In the burned site, the Q₁₀ value was higher than 1.77 after 1 year since forest fire had occurred (Jeong 2007). This value reflected the difference caused by the number of years of forest ecosystem restoration after forest fire. The Q₁₀ value was incorporated within the range of 1.3−3.3 for various temperate forest soils in studies of the global CO₂ flux reviewed by Raich and Schelesinger (1992), which noticed the relations among the vegetation, climate and soil respiration.

[Fig. 7.] Relationship between soil respiration and soil temperature (a) and air temperature (b). The white and the black circles represent data from the unburned site and the burned site, respectively.

[Table 2.] Relationships between the soil respiration and temperatures (soil and air) during the study period from March 2011 to February 2012

Relationships between the soil respiration and temperatures (soil and air) during the study period from March 2011 to February 2012

We assessed CO₂ emission from soil during the study period in the unburned and burned sites, using a relational expression with soil temperature in the highest correlation with CO₂ emission from soil (Table 3). The total CO₂ emission from soil (t CO₂ ha^-1 yr^-1, 1t = 10³ kg) was 35.59 and 14.69 in the unburned site and the burned site, respectively, during the study period. In the unburned site, the total CO₂ emission (t CO₂ ha^-1 yr^-1) was higher than 24.0 in the study of soil respiration in P. densiflora forest in the region of Jinju, Kyeongnam (Moon 2004); 27.32 in the study of soil respiration in P. densiflora forest in Mt. Sambong Hamyang, Kyeongnam (Kim 2006); 21.20 in the study of soil respiration in P. densiflora forest after forest fire (Jeong 2007); and 26.72 in P. densiflora forest in Mt. Mudeung (Kim 2008). However, the obtained annual total soil respiration in various P. densiflora forests of Korea was incorporated within the range between 10.0 and 46.0 from the study of Raich and Nadelhoffer (1989). In the burned site, the value was higher than 10.0 t CO₂ ha^-1 yr^-1 after 1 year passed since forest fire in this study, and this indicated the difference between changes in the forest restoration period after being damaged due to forest fire.

[Table 3.] Seasonal and annual soil respiration at the unburned site and burned site during the study period from March 2011 to February 2012

Seasonal and annual soil respiration at the unburned site and burned site during the study period from March 2011 to February 2012

This study was conducted in the P. densiflora forests in the east coastal region of Samcheok-si, Gangwon-do, where went through restoration process for 12 years after burned among the sites with being damaged due to forest fire, to make a quantitative assessment of the correlation between the CO₂ emission rates from the soil and environmental factors, and the annual total soil CO₂ emission in the damaged sites.

1. The emission of CO₂ from soil ranged from 41.37 ± 7.78 to 417.82 ± 54.98 mg CO₂ m^-2 h^-1 in the burned site, and from 42.66 ± 6.41 to 1175.14 ± 280.03 mg CO₂ m^-2 h^-1 in the unburned site. It decreased after peaking in August in both sites.

2. The relationship between CO₂ emission from soil and soil temperature showed a high significance (P < 0.001), as well as a high positive correlation, in both of the burned site (R² = 0.932) and the unburned site (R²= 0.942). However, the relationship between CO₂ emission from soil and air temperature was less correlative than that from soil temperature in both of the burned site (P < 0.01, R² = 0.668) and the unburned site (P < 0.01, R² = 0.729).

3. The total CO₂ emission from soil (t CO₂ ha^-1 yr^-1) was 35.59 and 14.69 in the unburned site and the burned site, respectively, during the study period.

4. In 2012, our burned site was not completely recovered from loss of soil organic matter and nutrient elements, and the vegetation structure undergoing secondary succession had been changed from the pine forest to the oak forest. Even after 12 years passed since forest fire, the vegetation regeneration was so slow that there was the difference in soil respiration between the burned and unburned sites.

5. The annual CO₂ emission from soil in this study was assessed by the correlation between the environmental factors during the 1 year. In further studies, the long-term and continuous measurement of CO₂ emission from soil should be necessary to assess it in accuracy. In addition, the studies of restoration method for CO₂ reduction are also be needed, by assessing the amount of CO₂ storage and emission from soil in the process of restoring vegetation after being damaged by forest fire, and by measuring CO₂ emission from soil in the initial stage after forest fire occurred.