Climate change, desertification, melting ice caps, rising sea level, and destruction of ecosystem have been under progress as global warming continued to accelerate since mid-1980s. The increase of CO₂ concentration was noted as the main cause of mentioned phenomena (Jones et al. 1986, Cannell et al. 1992, Kim 2012). The global CO₂ concentration increased by 40% from 278 ppm (1975) to 390.5 ppm (2011) (Ciais et al. 2013). The main cause of this global increase was the usage of fossil fuels which significantly contributed to changes in land-use (IPCC 2007). Therefore, much attention is paid to the forest ecosystem which is a carbon sink (Hu and Wang 2008) and to the change and role of carbon cycle. The carbon cycle of terrestrial ecosystem was significantly disturbed after the industrialization. The carbon emission during 1980-1990 was 140 billion tons C ha^-1 and carbon absorption was only 101 billion tons C ha^-1 resulting with surplus of 39 billion tons C ha^-1 (Sabine et al. 2004). The carbon storage of terrestrial ecosystem can be mainly divided into forest and soil (Lee 2010). The carbon cycle starts by fixing carbon from the atmosphere through photosynthesis of plants and they are absorbed into the soil in form of organic matter such as leaves, woods, and fruits. The accumulated carbon in the soil is then released into the atmosphere through plant, microbial decomposition and respiration. Thus, further research on forest matter production is required in order to analyze their carbon absorption, emission, and storage.

The matter production is a result of growth of communities within the forest that are formed by the interaction between environment and vegetation. The research has a significant meaning in a way that it revaluates forest resources by investigating and estimating the productivity, ecological properties, and carbon fixation in the forest (Park et al. 2005). The forests have different carbon storage capacity depending on physical, chemical, and biological environment such as dominant tree species, forest age, locational condition, and climate. However, it has been reported that forest ecosystem is responsible for 90% of aboveground carbon storage and 40% of belowground carbon storage (Waring and Schlesinger 1985). The standing biomass of plant communities becomes an index for matter production within the ecosystem and biomass allocation is an important factor in analyzing the feature and type of matter accumulation in the ecosystem (Whittaker and Marks 1975). The estimate of standing biomass can be calculated by using diameter at breast height (DBH) of trees and allometric equation between each organs (Kittredge 1944, Shidei 1960). It has been reported by Kim and Kim (1988) in the domestic forestry research that belowground biomass increases with the increase in aboveground biomass and that belowground biomass accounts for 20-30% of aboveground biomass.

The organic carbon in soil accounts for approximately more than 50% of total Earth’s carbon (Vitousek 1991) and it is crucial to investigate the carbon accumulation in soil and carbon emission from soil to atmosphere. Therefore, the ecosystem and climatic condition must be investigated because carbon cycle of soil is controlled by such factors (Lee et al. 2010), and recently temperature and moisture correlation data, which is the environmental factors of soil respiration, have been required (Shutou and Nakane 2004). Alkali absorption method which applies CO₂ absorbing chemicals such as lime and potassium hydroxide under closed chamber conditions were used in the beginning to measure soil respiration but Infrared Gas Analyzer (IRGA) is used today (Lee et al. 2010). The litter production is a fundamental process in maintaining the functions of forest ecosystem and it is determined by factors such as temperature, rainfall, insects, tree density, and climate (Bray and Gorham 1964). The accumulated organic matter on the forest floor of forest ecosystem mostly consists of tree leaves that form the forest (Ovington and Heitkamp 1960), and CO₂ is released again during the absorption process in the soil when plants are decomposed (Berg and Agren 1984). Hence, the litter production is an important factor that controls the primary production of the forest (Cole and Rapp 1981, Meentemeyer et al. 1982).

In this way, the forest carbon distribution, absorption, and emission can be determined through estimated carbon budget by quantifying the amount of carbon in each factor. It was purported that temperate forest ecosystem plays an important role in decreasing the atmospheric carbon (Tans et al. 1990), and Quercus spp., which are deciduous trees, dominate south central regions of Korea (Kim et al. 1981). Worak National Park situated in central Korea was the 17^th National Park selected and it is situated at 128°02′-128°21′ East longitude, and 36°47′-36°58′ North latitude having total area of 287,977 km². There has not been a research on carbon budget of Q. serrata forest even though there were researches on Pinus koraiensis S. et Z. plantation by Pyo et al. (2003), Q. variabilis BL. Forest by Choi (2007) and Namgung (2010) at Worak National Park, Q. mongolica FISCH. Forest by Shin (2012), and P. densiflora S. Et Z. forest by Lee (2014a) with regards to carbon budget of domestic forest ecosystem. Q. serrata Thunb. ex Murrayis recognized as climax species of forest vegetation succession as well as the potential natural vegetation (Lee and You 2012), and it is widely distributed across central lowlands and southern regions of Korea (Chung and Lee 1965).

This research investigated the main plant community, Q. serrata forest amongst Quercus forests which represents temperate deciduous forests of Korea, in Worak National Park from 2010 to 2013. The following areas were studied in order to explain the matter production of forest ecosystem, budget and distribution of carbon: annual litter production, litter organic carbon, standing biomass, net primary production, organic carbon in plants and soil, organic carbon in litter on forest floor, and soil respiration.

Worak National Park consists of following main vegetation: 32.4% deciduous forests, 16.7% Pinus-Quercus forests, and 12.4% Q. mongolica forest (Oh et al. 2005). It has continental climate as it is situated in southern central regions of Korean Peninsula and it also reflects some of the features of mountain climate because it is surrounded by rugged mountain ranges. It has relatively heavy rainfall affected by the orographic rainfall of Taebaek, Sobaek, Charyeong, and Noryung mountain ranges.

In November 2009, 20 m × 20 m permanent quadrat (N 36°50′08.5″, E 128°17′27.4″) was installed in the research site for Q. serrate community. It is located in Dongsa-myeon, Moonan-gol of Mt. Worak 398 m above sea level with the slope of 22°. The average forest age was 18.4 years, average height was 12.7, average DBH was 14.2 cm, and tree density was 1200 tree/ha. Staphylea bumalda DC., Securinega suffruticosa (Pall.) Rehder, and Weigela subsessilis (Nakai) L. H. Balley were distributed in the shrub layer whereas Carex siderosticta Hance, Lespedeza maximowiczii C.K. Schneid, and Atractylodes ovata (Thunb.) DC. were found in the herb layer. It was found that annual average temperature was 12.4℃ and rainfall was 1258.9 mm with reference to 30 years of climatic data stored at Jecheon Meteorological Observatory 20 km away from the research site. The annual average temperature was 9.63-10.11℃ and annual average rainfall was 1381.30-2230.50 mm (Korea Meteorological Administration 2010, 2011, 2012, 2013) during the study period (2010-2013) but exceptionally high rainfall was observed in 2011.

The harvesting method is the best method to measure the standing biomass of tree layer after felling the forest (Kang and Kwak 1998) but allometric equation which can be applied to each plant was used instead to measure the standing biomass as felling is prohibited in National Parks. The diameter at breast height (D) and height (H) of the trees with diameter greater than 3.0 cm within the permanent quadrat of the research site were measured every April from 2010 to 2013. Allometric Equation of Q. serrata Thunb. ex Murray (Koh 2002), Q. mongolica (Kim and Yoon 1972), Q. variabilis (Kang and Kwak 1998), and P. sargentii (Kwak et al. 2004) were used to estimate the standing biomass. The annual litter production collected from the litter trap was used to calculate the standing biomass of leaves and reproductive organs.

The standing biomass of shrub layer was estimated by installing three quadrats (2 m × 2 m) around permanent quadrat in the region of same community and leaves and twigs produced in that year were collected. The same procedure was conducted for herb layer by installing three quadrats (1 m × 1 m). The collected samples were weighed after being dehydrated at 60℃ in the dryer until it reached constant weight and then the standing biomass per area unit was calculated. It was reported that the proportion of belowground standing biomass in woody plants in general accounts for 15-35% of entire standing biomass depending on forest age and environment (Rodin and Bazilevich 1967).This research estimated and applied 25% as the ratio of belowground standing biomass (roots) to aboveground standing biomass (shoots) (Johnson and Risser 1974).

The annual net primary production (△W = W2 - W1) was calculated by subtracting current year’s standing biomass (W1) from next year’s standing biomass (W2). The annual net primary production of leaves and reproductive organs were calculated using the annual litter production collected from the litter trap because they wither in same year they are produced. The net primary production of shrub and herb layer was used the standing biomass.

The plant organic carbon was estimated to be 45% of dehydrated plant mass in tree, shrub, and herb layer (Houghton et al. 1983). The absorption of organic carbon by plants was calculated based on organic carbon found in plants which was measured. The annual uptake of organic carbon (△C = C2 - C1) was calculated by subtracting current year’s organic carbon (C1) from next year’s organic carbon (C2).

The litter was collected in intervals of two months (excl. winter season) from March 2010 to December 2013 by installing three 1 m × 1 m litter trap 1 m above ground in November 2009 to measure the litter production. The collected litter was classified into leaf, stem (bark and branch), reproductive organs (flower and seed), and miscellaneous (other trees etc.) which were then weighed after being dehydrated at 60℃ condition in the dryer for more than 48 hours. Subsequently, the litter production per area unit was calculated using the litter collected from the litter trap. The organic carbon of litter production was estimated at 45% of dehydrated mass (Houghton et al. 1983).

Four 25 cm × 25 cm small quadrats were installed outside the permanent quadrat and the litter on forest floor was collected after classifying the layer into Litter and Fermentation layer. It was then weighed after being dehydrated in the dryer at 65℃ condition for 48 hours in the laboratory. The organic carbon of forest floor litter was estimated at 45% of dehydrated mass (Houghton et al. 1983).

The soil was collected in 10 cm intervals until it reached 50 cm in depth at three random points outside the permanent quadrat and it was taken to the laboratory sealed in plastic bags. The collected soil was used to analyze organic matter content, organic carbon content, accumulated organic carbon content, and rock volume. The soil was collected in 10 cm intervals up to 50 cm by using a soil sampler which is a stainless cylinder 5 cm in diameter and 10 cm in height. It was taken to the laboratory sealed in plastic bag to be weighed after being dehydrated at 105℃ condition in the dryer until it reached constant weight. The measured weight value was then divided by the volume in order to calculate soil bulk density. The rock volume was calculated by dividing the weight of particles greater than 2 mm by the total weight after filtering the dehydrated soil particles with 2 mm filter sieve. The 5.000 g air dried fine soil in porcelain crucible was weighed after being dehydrated in the dryer at 105℃ condition for 48 hours and it was heated in electric furnace at 600℃ condition for six hours. The ash content was subtracted from the dry mass and it was then divided by 1.724 in order to convert the value into soil organic carbon content (Black 1965). The following equation was used to calculate accumulated organic carbon content according to the method proposed by Wang et al. (2002) considering the rock volume ratio:

A portable measuring device, EGM-4 (PP Systems, Amesbury, UK), was used to measure the soil respiration. The measuring method of this device is one of the SRC-1 (PP Systems) based on closed method and it was easy to measure with infrared gas analyzer at the measuring point using portable battery. The soil respiration was measured by CO₂ released from the soil into the closed chamber (g CO₂ m^-2 hr^-1) according to elapsed time after removing the litter layer and fixing a cylinder chamber on the soil surface. The root respiration was calculated by 46% of estimated value after estimating the soil respiration using the correlation between soil respiration and soil temperature excluding the maximum and minimum value measured quarterly at 10 random points in the permanent quadrat of this research site (Koo et al. 2005).

The quantification of carbon budget using biometric method is a method of completing the entire frame of carbon budget by quantifying the carbon mass of each factor within the ecosystem and the carbon transfer process (Lee et al. 2010). The net ecosystem production (NEP), a carbon budget at ecosystem level, is calculated by subtracting heterotrophic respiration (HR), which releases carbon into the atmosphere through animal and microbial carbon decomposition process, from net primary production (NPP).

The standing biomass of the research site was 122.626, 133.625, 144.233, 155.528 ton/ha for 2010, 2011, 2012, 2013, respectively, and total average over the study period was 139.003 ton/ha. The plant organic carbon was 55.182, 60.131, 64.905, 69.988 ton C ha^-1 , respectively, and total average over the study period was 62.551 ton C ha^-1. In addition, the standing biomass and organic carbon increased annually (Table 1).The research results related to standing biomass of Q. serrata forest were reported as follows: Koh (2002) reported aboveground biomass of 52.631 ton/ha at Mt. Songgwang, and Q. serrata forest of 94.20 ton/ha at Mt. Songni; Park and Lee (2002) reported 56.07 ton/ha of 34 year old Q. serrata forest at Gwangyang, 44.00 ton/ha of 25 year old Q. serrata forest at Muju, and 86.32 ton/ha of 37 year old Q. serrata forest in Pohang region; and lastly, Park and Moon (1994) reported 115.3 ton/ha of 26-29 year old Q. serrata forest at Mt. Mohu in Jeollanam-do. The standing biomass of above-mentioned sites were lower than the standing biomass of this research but the standing biomass (221 ton/ha) of Q. serrata forest in Bukhansan National Park estimated by Lee (2011) and standing biomass (261.2 ton/ha, 1,473 trees/ha) and organic carbon (136 ton C ha^-1) of deciduous forest (Q. serrata dominant) in Gwangneung estimated by Lim et al. (2003) were higher than this research result.

[Table 1.] Standing biomass (ton/ha), plant organic carbon (ton C ha-1), net primary production (ton ha-1 yr-1) and uptake of organic carbon (ton C ha-1 yr-1) of Quercus serrata forest from 2010 to 2013

Standing biomass (ton/ha), plant organic carbon (ton C ha-1), net primary production (ton ha-1 yr-1) and uptake of organic carbon (ton C ha-1 yr-1) of Quercus serrata forest from 2010 to 2013

The annual net production was 10.999, 10.608 and 11.295 ton ha^-1yr^-1 for 2010, 2011, 2012, respectively, and total average was 10.967 ton ha^-1yr^-1 (Table 1). This result was similar to the result (11.36 ton ha^-1yr^-1) reported by Chang and Kim (1983) for Q. serrata forest in Mt. Jiri. However, it was higher than 4.870 ton ha^-1yr^-1 of Q. serrata forest in Suncheon reported by Lim (1985), and estimates reported by Park and Lee (2002) in Gwangyang (6.18 ton ha^-1yr^-1), Muju (6.53 ton ha^-1yr^-1), and Pohang region (7.66 ton ha^-1yr^-1). Nonetheless, the result was slightly lower than 14.5 ton ha^-1yr^-1 of Park and Moon (1994). The standing biomass and plant organic carbon show similar trends in their correlation. The reason why standing biomass, annual net production, and plant organic carbon show different values to other research is because forest age and tree density of different research sites have influenced the study. Kimmins (1987) stated that the standing biomass and net production increases with aging of the forest in case of young forests given that community structure of the forest is similar and Ohtsuka et al. (2010) also reported that the increase of biomass and net production of younger forest is greater than the older forest. In addition, the NPP of forest ecosystem typically changes with the forest age, and it starts to decrease once dense pioneer community reaches its peak (Marks 1974, Smith and Long 2001, Binkley et al. 2002). Kang and Kwak (1998) reported that there could be a difference in matter production as it is influenced by tree density, forest age, and locational condition even though Quercus spp. consist of same species. Furthermore, Kwak and Kim (1992) reported that the change in annual net production of tree layer has a close connection to changes in tree density, death rate, and production cycle rate. Thus, the research result was higher than the other research results because the forest was younger and tree density was higher.

The uptake of organic carbon was 4.949, 4.774, 5.083 ton C ha^-1yr^-1 for 2010, 2011, 2012, respectively, and the average was 4.935 ton C ha^-1yr^-1 (Table 1). The research result showed similar result to 4.3 ton C ha^-1yr^-1 of deciduous forest in Gwangneung reported by Lim et al. (2003) and it was higher than 1.834 ton C ha^-1yr^-1 uptake of organic carbon at Mt. Songgwang in Suncheon reported by Lim (1985). The previously mentioned differences in forest age, tree density, and high NPP of young forest seem to affect the uptake of organic carbon as well. In particular, there was a tree in this community damaged by wild animals in 2012 but it was recovering very fast. The uptake of organic carbon in this community seems to be higher than other communities as Ohtsuka et al. (2010) states that the damage recovery is faster in younger forests which affects the production.

The amount of annual organic carbon accumulating on the forest floor from the litter production was 1.834, 1.769, 1.440, 1.641 ton C ha^-1yr^-1 during study period (2010-2013) and total average was 1.671 ton C ha^-1yr^-1 (Fig. 1). The organic carbon of litter showed significant annual change. The organic carbon decreased sharply in 2012 but it increased again. The sharp decrease seems to be from the damage on leaf and stem of trees caused by the typhoon (Bolaven). The high ratio of stands that have low DBH and fast recovery of young Q. serrata forest seem to have affected the increase in the production and uptake of organic carbon. The leaves had the highest distribution of organic carbon on forest floor each year and the rest were in following order: stem, reproductive organ, then miscellaneous. The litter production and organic carbon was always high during fall and it was the lowest during winter each year except in 2012 (Fig. 1). A similar trend was observed in the period and change in litter production of P. densiflora, Q. mongolica, and R. pseudoacacia forests in Namsan (Jeong et al. 2013). Q. serrata forest, representative Quercus forest in Mt. Worak, showed the lowest organic carbon though P. densiflora forest of Lee (2014a) in the same Mt. Worak showed a similar value of 1.68 ton C ha^-1yr^-1, whereas Q. mongolica forest of Shin (2012) was 2.59 ton C ha^-1yr^-1, and Q. variabilis forest of Namgung and Mun (2009) was 2.42 ton C ha^-1yr^-1. In the other regions, Gwangneung deciduous forest (Q. serrata dominant) was 2.10 ton C ha^-1yr^-1 (Lee 2009) and Geumsan Q. serrata forest was 2.323 ton C ha^-1yr^-1 (Kim et al. 2013) showing higher value than this research. It is known that there can be changes in litter production because it has significant relation to temperature, rainfall, and pests unless there is a disturbance such as typhoon (Lousier and Parkinson 1975, Li et al. 2005, Kim et al. 2013).

[Fig. 1.] Seasonal fluctuations of litter production (ton C ha-1) in Quercus serrata forest from 2009 to 2013. Win, winter; Spr, spring; Sum, summer.

The organic carbon of forest floor in the Q. serrata forest during the study period was on average 1.959 ton C ha^-1 for L layer, 2.115 ton C ha^-1 for F layer, and total average was 4.075 ton C ha^-1 (Fig. 2).The organic carbon of F layer was always greater than L layer except during spring and it decreased from spring to summer but it increased from summer to winter (Fig. 2). Decomposition of forest floor litter is affected by animals as well as microbes in soil and they are affected by moisture, temperature and chemical properties of leaves (Rochow 1974, Singh and Gupta 1977, Meentemeyer 1978, Leirós et al. 1999, Mun 2004). It seemed that there are seasonal differences inorganic carbon of forest floor because the decomposition process accelerates during summer due to high temperature and rainfall.

[Fig. 2.] Seasonal fluctuations of organic carbon (ton C ha-1) on forest floor of Quercus serrata forest during the study period. Spr, spring; Sum, summer; Win, winter.

The organic carbon of forest floor in Gwangneung deciduous forest was 5.4 ton C ha^-1 (Lee et al. 2010) and Q. serrata forest of Piagol in Mt. Jiri was 4.29 ton C ha^-1 (Chang and Kim 1983) which showed higher values than this research. However, similar or higher values of 3.97, 3.58, and 4.96 ton C ha^-1 (National Institute of Environmental Research 2013) were observed during the same period (2010-2013) in the long-term ecological research of P. densiflora, A. holophylla, and Q. mongolica in Mt. Jeombong. The difference in organic carbon of forest floor seems to be from the differences in environmental factors of each community. Namgung and Mun (2009) stated that the difference is affected by the slope, forest age, tree density, and size of the leaf, etc. In addition, the differences in litter on forest floor seem to be caused by wild animals and artificial disturbances as well apart from natural phenomena. The research site is in the region where landslide often occurs and litter on forest floor was often lost in each investigation due to landslide and wild animals (boars etc.). There were significant differences in forest floor litter in comparison to other researched forests because the soil was bear and likewise for organic carbon. The naturally damaged deadwoods (stem etc.) and fallen leaves have high nutrient content and they can be leached fast which influences the increase of distribution and production of carbon as it accumulates on the forest floor in a long term. Moreover, Triska and Cromack (1980) stated that deadwoods enhance species diversity as they provide microhabitats. Therefore, more research on forest floor litter is required in order to understand the status of organic carbon and carbon cycle of forest ecosystem (Mun 2004).

The organic carbon in soil which was measured up to 50 cm in depth during the study period in Q. serrata forest was in total 64.119 ton C ha^-1yr^-1and the average was 12.824 ± 2.097 ton C ha^-1yr^-1 (Fig. 3). The organic carbon in soil decreased as it went deeper from top soil (17.937 ton C ha^-1yr^-1) to 50 cm (8.981 ton C ha^-1yr^-1) in depth except at 40 cm (11.564 ton C ha^-1yr^-1) (Fig. 3). The seasonal accumulated organic carbon content in soil decreased as soil depth increased except in summer. Spring season was 8.390-16.759 ton C ha^-1yr^-1, summer season was 12.279-22.275 ton C ha^-1yr^-1, fall season was 9.490-21.286 ton C ha^-1yr^-1, and winter season was 5.765-11.426 ton C ha^-1yr^-1. Summer season had the highest value and the value decreased in order of fall, spring, and winter season. Also, summer season was highest and then fall, spring, and winter with regards to the depth of the soil showing similar trend (Fig. 3). This is a similar trend to Q. serrata forest in Bukhansan National Park where amount of carbon decreases according to depth of the soil. The total organic carbon was 49 ton C ha^-1yr^-1 (Lee 2010) which was lower than this research but Q. serrate dominant forest in Gwangneung had 92 ton C ha^-1yr^-1 which was higher than this research (Lim et al. 2003).The organic carbon in soil is affected by amount of plants falling on the soil and it is different depending on seasonal variation without doubt (Leirós et al. 1999). The decomposition occurred very well in the forest floor layer due to high temperature and rainfall of summer season. It seemed that carbon storage of soil was the highest during summer season because decomposed carbon and organic matters accumulated in soil which showed a change of seasonal difference. In addition, storage capacity of organic carbon in soil is different by region and age classes (Jeong et al. 1998). There are significant amount of nutrient, and carbon amongst aboveground matter production, that moves belowground (Chen et al. 2000). Therefore, the decomposition of fine roots and roots increase organic carbon in soil and it seems that differences in organic carbon in soil is caused by biological factors such as production and decomposition of litter and environmental factors such as temperature, soil depth, and locational conditions. The organic carbon of Q. serrata forest in Mt. Bukhan had the lowest amount of organic carbon in soil due to shallow soil depth in comparison to other research sites within the National Park (Lee 2011). It seems that such result is different from this research because it has lower tree density (800 tree/ha) than Q. serrata forest in Mt. Worak.

[Fig. 3.] Seasonal fluctuations of organic carbon (ton C ha-1) along the soil depth of Quercus serrata forest during the study period.

The organic carbon from soil respiration of Q. serrata forest was 7.338, 7.544, 7.183 ton C ha^-1 yr^-1 for 2010, 2011, 2012, respectively (Fig. 4), root respiration was 3.375, 3.470, 3.304 ton C ha^-1 yr^-1, annual microbial respiration was 3.962, 4.074, 3.879 ton C ha^-1 yr^-1, and total average was 3.972 ton C ha^-1 yr^-1. The annual soil respiration was highest in 2011 and lowest in 2012. The soil respiration in each year was highest during summer season in August, lowest in January in 2010 and 2011, but it was lowest in February in 2013 (Fig. 4). With regards to research on soil respiration of Q. serrata forest, Lee (2009) reported 3.65 ton C ha^-1yr^-1 for Q. serrata and C. laxiflora communities (80-200 years) using Open Flow (OF) method, and Chae et al. (2003) reported 4.13 ton C ha^-1yr^-1 for mixed forest of Q. serrata, P. koraiensis, and Q. mongolica (70-80 years) using Closed Dynamic Chamber (CDC) method. Katagiri (1988) reported 5.03 (upper slope), 4.16 (middle), 4.30 (lower slope) ton C ha^-1yr^-1 according to altitude for Q. serrata and C. crenata communities in Shimane, Japan using Akali Absorption (AA) method and Katagiri et al. (1979) reported 4.09 ton C ha^-1 yr^-1 which showed similar value to this research result. Nakane et al. (1996) reported 9.40 ton C ha^-1 yr^-1 for Q. serrata forest (102 years) in Hiroshima, Japan, Yim et al. (2002) reported 9.52 ton C ha^-1 yr^-1for Q. serrata forest (70 years) in the same region, and 5.68 ton C ha^-1 yr^-1 for Q. serrata dominant mixed forest in Tokyo all using AA method. Takahashi et al. (2004) reported 6.42 ton C ha^-1 yr^-1 for Q. variabilis and Q. serrata forest using CO₂ gradient method, whereas Yim et al. (2002) reported 8.88 ton C ha^-1 yr^-1 using CDC method showing higher values than this research result. The scale of soil respiration is similar to NPP and it is influenced by soil temperature, moisture, rainfall, typhoon, and root respiration but especially the rainfall increases soil respiration (Lee et al. 2006). The seasonal change in soil respiration of Q. serrata forest had the same trend in many other researches (Buchmann 2000) and close correlation between soil respiration and soil temperature is already well-known to scholars (Witkamp 1969, Dulohery et al. 1996, Lee 2014b). It was reported that soil moisture greatly influences soil respiration as well (Liu et al. 2006). The sharp increase in soil respiration during early summer is related to the expansion time of leaf in crown and high metabolism during the leaf production will also affect the increase in soil respiration (Kawamura et al. 2001, Lee et al. 2003). The soil respiration is the sum of heterotrophs respiration and respiration of plant roots and therefore it is important to measure them separately. The mineralization of organic carbon in soil decomposed by microbial activity is influenced by moisture and temperature of the soil (Leirós et al. 1999). In addition, there will be difference in the biomass of microbes in soil depending on vegetation type, organic carbon in soil, and environmental factors because it is affected by the level of organic carbon in soil. Thus, it is determined that decomposition of forest floor litter, high level of organic carbon in soil, plant production, decomposition and production of roots accelerated by microbial activity in summer under high atmospheric temperature, soil temperature, and rainfall increased the soil respiration in summer season. It has been reported that microbial respiration is more reactive and therefore higher in thin roots than roots with thick diameter in Q. serrata forest (Matsumoto et al. 2009). It seems that respiration was higher in older forests as they have more root diversity than the forest studied in this research.

[Fig. 4.] Seasonal variation of organic carbon of soil respiration in the Quercus serrata forest from 2010 to 2013.

The budget and distribution of organic carbon per carbon storage during the study period of Q. serrata forest in Mt. Worak is shown in Fig. 5. The total average organic carbon was 130.745 ton C ha^-1, aboveground part of plant was 50.041 ton C ha^-1, and belowground part was 12.510 ton C ha^-1. The organic carbon of litter on forest floor was 4.075 ton C ha^-1 and the soil was 64.119 ton C ha^-1 having the highest distribution. The total average of carbon fixation in plants from photosynthesis was 4.935 ton C ha^-1yr^-1 and amount of organic carbon released from soil respiration to microbial respiration was 3.972 ton C ha^-1yr^-1 during the study period.

[Fig. 5.] Compartment model showing the distribution and flow organic carbon of Quercus serrata forest for four years in the study area. The box indicates standing carbon (ton C ha-1) and arrow is the flux of carbon (ton C ha-1 yr-1). The bracket gives information about uptake of carbon (ton C ha-1 yr-1). NEP, net ecosystem production; NPP, net primary production.

The annual NEP estimated from annual organic carbon production of the Q. serrata forest and soil respiration was 0.963 ton C ha^-1yr^-1. As result of this research, the Q. serrata forest showed to be a positive factor in improving the atmospheric environment as it absorbs more carbon then it releases through respiration. There are no researches on NEP of Q. serrata forest in Korea at present except for the research result of 0.36 ton C ha^-1yr^-1 by Lee (2014b). However, NEP (0.23 ton C ha^-1yr^-1) estimated by using NPP of 4.3 ton C ha^-1yr^-1 (Lim et al. 2003) and HR of 4.07 ton C ha^-1yr^-1 (Min 2006) from the research results of Gwangneung deciduous forest (Q. serrata dominant) was lower than the result obtained in this research. The NEP of deciduous forest in cool-temperate regions of Japan, where secondary succession of young trees (incl. Q. serrata) is occurring, was 0.9 ton C ha^-1yr^-1 showing similar values to this research. Ohtsuka et al. (2010) stated that the NEP of young forests, together with their biomass, will rapidly increase as their NPP was greater than the forest (50 year) located 260 m higher within the same region. The young forests and mature forests have greater capability to act as carbon sinks than old forests (Bond-Lamberty et al. 2004), and the rainfall increases respiration of the ecosystem and affects annual carbon budget (Lee et al. 2006). In addition, the Q. serrata forest of this research with properties of tree density (1200 trees/ha), number of individual 0-20 cm trees per DBH (43 ea), forest age (18 year), and annual average temperature (12.4℃) had greater and active matter production than the forest researched by Ohtsuka et al. (2010) with tree density (1085 trees/ha), forest age (18 year), average temperature (7.3℃) and elements (850 trees/ha, 23 ea, 37 year) of Q. serrata forest researched by Lee (2014b) in same period and location. In addition, the Q. serrata forest also had greater carbon assimilation being a young forest rendering high NEP.

Therefore, the NEP was related to previously analyzed results (standing biomass, organic carbon in litter, forest floor, soil, and soil respiration etc.). The tree diameter was important as uptake of organic carbon was highest in the plant stem and the soil played a crucial role as the carbon storage. Lastly, the understanding of carbon budget will be important in the future because the amount of carbon in soil decreases together with the climate change (Kirschbaum 1995).