Introduction

Litterfall inputs and litter decomposition represent important components of the carbon (C) and nitrogen (N) cycles in forest ecosystems because the turnover of litter is a major pathway by which C and N enter forest soil (Bray and Gorham 1964; Gower and Son 1992; Berg and Laskowski 2006; Kamei et al. 2009). The inputs of C and N in litterfall depend on several ecological factors and forest management activities such as forest type, climate, site quality, stand increment, stand age, stand density, fertilization, and thinning (Berg and Laskowski 2006; Kim 2008; Kim et al. 2009). Forest type can have a significant influence on litterfall C and N because of the difference in input of leaf litter or non-leaf litter produced by different species (Bray and Gorham 1964; Kamei et al. 2009). Thus, a balance between litterfall and decomposition processes controls development of the soil organic layer, and hence the C and N content in the soil (Finzi et al. 1998; Berg and Laskowski 2006; Hansen et al. 2009). However, it is not easy to compare litterfall and litter decomposition of different forest types due to the potentially confounding effects of site and management factors (Gower and Son 1992; Chen and Li 2003; Kamei et al. 2009).

Soil C and N status can change upon establishment of plantations because different tree species have different nutrient requirements (Son and Gower 1992; Binkley 1995), litterfall inputs and litter decomposition mechanisms (Finzi et al. 1998; Hansen et al. 2009; Kamei et al. 2009). However, conflicting results have been reported on the influence of forest type on soil C and N status (Son and Gower 1992; Kamei et al. 2009) with no discernable effect (Finzi et al. 1998) of different tree species. Although numerous studies to evaluate the effect of forest types on nutrient distribution are reported in the literature (Binkley 1995; Finzi et al. 1998; Hansen et al. 2009), major uncertainties remain about the importance and behavior of C and N status in coniferous plantations in Korea.

Red pine (Pinus densiflora S. et Z.), rigitaeda pine (P. rigida × P. taeda), and larch (Larix leptolepis Gordon) have been the three most important coniferous tree species planted throughout Korea during the last 40 years. Red pine occupies more than 23.5% (1.5 million ha) of Korean forest lands. Larch forests were planted on about 600,000 ha between 1957 and 1990 (Korea Forest Service 2006). Rigitaeda pine, a hybrid of pitch (P. rigida Mill.) and loblolly (P. taeda L.) pines, shows good growth characteristics with enhanced cold resistance compared with pitch or loblolly pine. There is a need for information evaluating the direction and rates of change in C and N status in the soil due to litter input, and litter decomposition by various plantation types because tree species play a primary role in C and N cycling in forest ecosystems (Binkley 1995). The objectives of this study were to quantify the impacts of plantation establishment on C and N status of needle litter, litter decomposition and soil in even-aged larch and pine plantations under similar site conditions. This entailed measuring (1) C and N inputs via litterfall, (2) C and N dynamics in decomposing needle litter, and (3) the distribution of soil C and N.

Materials and methods

The study was conducted in the Sambong Exhibition Forests located in Hamyanggun, Gyeongsangnamdo, and administered by Seobu National Forest Office, Korea Forest Service. The annual mean precipitation in this area is 1,322 mm and the annual mean temperature is 12.8°C. Experimental plots consisting of one deciduous (larch) and two evergreen coniferous plantations (red pine, rigitaeda pine) were located adjacent to each other on moderately productive sites (Table 1). All three plantations were established in 1963 on northeast facing slopes (5–15°) with small pits and mounds. The study sites were of identical macroclimatic condition, quality, and stand age. The understory species in the larch plantation site included Viburnum dilatatum Thunb., Lindera erythrocarpa Makino, Rubus parvifolius L., Quercus serrata Thunb., Q. acutissima Carruth., Q. variabilis Bl., Castanea crenata S. et Z., Schisandra chinensis Baill., Zanthoxylum schinifolium S. et Z., and Elaeagnus umbellate Thunb. The understory species in the red pine plantation site were Rhododendron mucronulatum Turcz., Q. serrata Thunb., Q. aliena Bl., Lindera glauca Bl., L. obtusioloba Bl., Smilax china L. and Juglans mandshurica Maxim., and those in the rigitaeda pine plantation site were Styrax japonica S. et Z., Stephanandra incise Zabel, Z. schinifolium S. et Z., Cornus controversa Hemsl., Q. aliena Bl., Q. serrata Thunb., Symplocos chinensis var. leucocarpa for. pilosa Ohwi, J. mandshurica Maxim. and Rhus sylvetris S. et Z. Data were collected from three 20 m × 10 m plots within each plantation site. The general characteristics of the tree species with regard to their mean stand densities, diameter at breast height (DBH), and basal area of the three coniferous plantations are shown in Table 1. The differences in DBH and stand basal area among the three plantations were due to common forest management practices such as thinning.

Table 1 General characteristics of the three coniferous plantations at the study sites. Values in parenthesis represent ranges
Full size table

To measure litterfall, we installed three circular litter traps made of 1.5 mm nylon net and with 0.25 m2 surface area at randomly chosen locations in each plot at 60 cm above the forest floor for each plantation. The total number of litter traps used in the whole study was 27. Litter was collected at monthly intervals between July 2006 and June 2008. The collected litter from each trap was transported to a laboratory and then oven-dried at 65°C for 48 h. The dried samples were then separated into needles and other components, and the weight of each portion was recorded. The collected litterfall was ground in a Wiley mill to pass through a 40-mesh stainless steel sieve. The C and N contents in needle litter were determined using an Elemental Analyser (EA1110, ThermoQuest Italia, Milan, Italy). The results of N inputs obtained from this study may have been underestimated as we measured the N concentration in needle litter collected only in the heavy litterfall season, although most of the needle litterfall from these three coniferous plantations takes place in November of each year (Jeong et al. 2009).

The C and N outputs due to decomposing needle litter were estimated using the litterbag technique. Fresh needle litter from each treatment was collected from the forest floor during late November 2006. After collection, the litter was air-dried at room temperature for 14 days. A sample of 10 g was then weighed out and placed in a 30 cm × 30 cm nylon net bag with a mesh size of 0.1 mm. Sub-samples from the litter were also taken to determine the oven-dried mass after heating at 65°C for 48 h. Fifteen litterbags for each plot of the plantations were placed randomly on the forest floor in December 2006. The litterbags were held fast by 10 cm long metal pins, and there were five subsequent collections of litterbags (in March 2007, July 2007, December 2007, March 2008, and June 2008) from each plot during the study period. Each litterbag sample was then oven-dried at 65°C for 48 h and the mass loss rates determined. The C and N concentrations of decomposed needle litter samples were determined by Elemental Analyser (EA1110, ThurmoQuest).

To compare soil environmental factors among the three plantations, soil temperature was measured monthly at a depth of 20 cm using a soil temperature probe (Model STP-1) attached to an EGM-4 environmental gas monitoring system (PP systems, Hitchin, UK). Soil samples to measure soil water content were collected monthly at 20 cm depth using an Oakfield soil sampler. The core samples were placed in plastic bags, transported to a laboratory and dried in an oven for 48 h at 105°C to quantify the soil gravimetric water content.

Soil samples were collected in October 2002 from three randomly selected points in each plot. At each point, a soil pit of 50 cm × 50 cm was dug to collect soil samples at three different depths (i.e., 0–10, 10–20, and 20–30 cm) because the top soil at 0–30 cm depth accounted for a large proportion of the soil C content in forest ecosystems (IPCC 2003). Soil samples for bulk density determination were collected at soil depths of 0–10, and 10–20 cm by using 100 cm3 stainless steel cans. Five bulk soil samples were also collected from each depth and passed through a 2 mm sieve prior to soil C and N analysis. The organic C concentration and total N of the mineral soils were determined by the loss-on-ignition (LOI) method at 375°C for 16 h (Soon and Abboud 1991) and by the Kjeldahl method (Bremmer and Mulvaney 1982), respectively. The data were subjected to analysis of variance at P < 0.05 and Tukey’s test for mean separation analysis (SAS Institute 1989).

Results and discussion

C and N inputs via litterfall

The mean C concentration of needle litter for the two study years was significantly higher (P < 0.05) in the red pine (510 g kg−1) plantation than in the larch (489 g kg−1) or the rigitaeda (489 g kg−1) pine plantations (Table 2). The values in this study are comparable to the C concentrations reported by Moore et al. (2006) of 497 g kg−1 for Jack pine (P. banksiana Lambert) and 488 g kg−1 for Tamarack (L. laricina K. Koch) needle litter in Canada. Other studies reported that the variation in C concentration in different tree species is determined by genetic and environmental factors such as site condition, stand characteristics and forest management practices (Bert and Danjon 2006; Zhang et al. 2009). The mean C input by needle litter was significantly higher (P < 0.05) in the pine plantations (1,976 kg ha−1 year−1 for red pine; 1,891 kg ha−1 year−1 for rigitaeda pine) than in the larch plantations (1,217 kg ha−1year−1). The C input by needle litter in this study is similar to that determined for a 40-year-old red pine stand (1,652 kg ha−1 year−1) and a 36-year-old larch plantation (1,181 kg ha−1 year−1) in the same region (Kim 2008; Kim et al. 2009). The C amount in needle litter of different tree species could be related to physiological responses, such as a species-specific allocation of resources between woody and photosynthetic components (Berg and Laskowski 2006). The C input from needle litter in this study was determined by the differences in needle litter mass between the larch and the pine plantations rather than the difference in C concentration (Table 2). There was no significant difference in C input from needle litter produced from the red pine and rigitaeda pine plantations (P > 0.05) despite a considerable difference in stand basal area and stand density between the two pine plantations (Table 1). The similarity in values of C input by needle litter produced by the two pine plantations could be associated to the canopy closure in these mature plantations, as the annual needle litterfall remains relatively constant after canopy closure (Bray and Gorham 1964; Berg and Laskowski 2006; Hansen et al. 2009).

Table 2 Carbon (C) and nitrogen (N) concentration and inputs from needle litter at three coniferous plantations. Values in parenthesis represent standard errors
Full size table

The mean N concentration of needle litter for the two study years was 11.1, 5.7, and 5.6 g kg−1 in the larch, rigitaeda pine, and red pine plantations, respectively (Table 2). The higher N concentration of larch needle litter compared to the two pine needle litters could be due to the inherent characteristic of higher N uptake by the deciduous larch compared to the evergreen pine tree species (Son and Lee 1997; Finzi et al. 1998; Kim 1999). Other studies have attributed the difference in N concentration of needle litter to abiotic factors such as climatic variables, nutrient supply to the site, canopy leaching characteristics; and biotic factors such as the trees’ genetic characteristics and retranslocation rates before senescence (Berg and Laskowski 2006; Hansen et al. 2009; Kamei et al. 2009). The N concentration values observed in this study are comparable to the N concentration values of 4.8 and 5.9 g kg−1 reported for Scot’s pine and White pine, respectively (Berg and Laskowski 2006). However, the values are slightly higher than the values of 3.7 g kg−1 for P. rigida and 8.2 g kg−1 for L. leptolepis plantations found in Gyeonggi Province, Korea (Kim et al. 2005). The mean input of N returned to the forest floor via needle litter was 27.61, 21.95, and 21.36  kg ha−1 year−1 by the larch, rigitaeda pine, and red pine plantations, respectively. The distinct difference in N input was associated with the lower N concentration in pine needle litter compared to larch needle litter. The value of N inputs in this study was higher than the 13.2 kg ha−1 year−1 reported for P. rigida or 23.1 kg ha−1 year−1 reported for L. leptolepis plantations in Gwangneung, Korea (Kim et al. 2005).

C and N dynamics in decomposing needle litter

A significant difference in needle litter decomposition (P < 0.05) was observed during the initial stage (March, July 2007) of the decomposition processes of the three coniferous plantations (Fig. 1). The mass loss rates in decomposing needle litter were more rapid for larch (9.7%) than for rigitaeda pine (5.4%) or red pine (5.1%) needle litter during the 222-day incubation of initial needle litter. Both red pine and rigitaeda pine showed almost similar mass loss at all five collections (Fig. 1).

Fig. 1
figure 1

Mass remaining (a), carbon concentration (b), nitrogen concentration (c), carbon/nitrogen ratio (d), carbon remaining (e), and nitrogen remaining (f) in decomposing needle litter of three coniferous plantations (filled diamond, Larix leptolepis; open square, Pinus densiflora; filled triangle, Pinus rigida × Pinus taeda). Vertical bars Standard error. Data represent mean values (n = 9); different letters indicate a significant difference at P = 0.05

Full size image

Needle litter decomposition can be influenced by abiotic factors, such as soil temperature, soil moisture and nitrogen availability, and by biotic factors such as decay organisms present in the soil, as well as the N and lignin concentration (Berg and Laskowski 2006). In this study, abiotic factors such as soil temperature and water content, which are regarded as the main factors influencing litter decomposition, were significantly different among the three coniferous plantations (Fig. 2). However, the rapid mass loss in larch may be attributed to the difference in substrate quality between larch and pine needle litter (Finzi et al. 1998) and/or to the high leaching loss of soluble labile C components during the early stages of the decomposition process (Li et al. 2009) because soil temperature and water content were significantly lower in the larch than in the red pine plantations (Fig. 2). For example, the initial C/N ratio, which is the most widely used index of litter quality, was significantly lower in larch (44) than in rigitaeda pine (89) or red pine (97) needle litter (Fig. 1). Many studies have reported that litter with high N concentrations and a low C/N ratio is decomposed more rapidly than those with low N concentrations and a high C/N ratio (Finzi et al. 1998; Berg and Laskowski 2006; Kamei et al. 2009). However, mass loss in decomposing larch litter at later stages of decomposition (after 356 days of incubation) was not affected by higher N concentration or lower C/N ratio compared with pine needle litter. This was because the decay rates at later stages of decomposition are determined by accumulation of recalcitrant constituents rather than by water soluble compounds (Berg and Laskowski 2006).

Fig. 2
figure 2

Means of monthly measured temperature (a) and water content (b) in soil at a depth of 20-cm of three coniferous plantations (filled square, L. leptolepis; shaded square, P. densiflora; open square, P. rigida × P. taeda). Vertical bars Standard error. Mean (n = 9); different letters indicate a significant difference at P = 0.05

Full size image

The C concentration and remaining C in decomposing needle litter decreased rapidly in larch litter compared to both the pine needle litters during the study period (Fig. 1). The lower C concentration in the larch needle litter could be attributed to rapid C mineralization throughout the decomposition process (Polyakova and Billor 2007) and/or absorption of dust particles from the air and subsequent precipitation without filtering effects (Binkley 1995) under open canopy during the winter season. In addition, the ash content from decomposing litter after the 461-day incubation (March 2008) of initial needle litter was higher in larch litter (77 g kg−1) than in either of the pine needle litters (46 g kg−1 for red pine; 36 g kg−1 for regitaeda pine). The C release of each needle litter type was approximately proportional to the mass loss of needle litter, which decreased with increasing incubation times. The C remaining over the study period did not exceed the initial value in undecomposed needle litter.

The N concentration and remaining N in decomposing needle litter were significantly lower in both pine plantations than in the larch plantation (Fig. 1). The N remaining followed a pattern of N concentration in decomposing needle litter. In contrast to C remaining, the N remaining in decomposing larch needle litter showed higher N gains than in either of the pine needle litters because the N in larch needle litter is readily immobilized by microorganisms due to its high concentration compared with pine needle litter. Nitrogen from the three needle litters is initially leached from the litter, and then immobilized (Fig. 1). Nitrogen immobilization in decomposing needle litter occurred generally during the growing season when microbial activities can be stimulated by increased soil temperature (Baker et al. 2001). The microbial immobilization of N occurred via uptake of N by fungal hyphae from the surrounding litter or mineral soil (Berg and Laskowski 2006).

C and N distribution in soil

Soil bulk density was significantly different among the three plantations (P < 0.05). The bulk density in the surface depth (0–10 cm) was lower in the red pine than in other two plantations (Table 3). Coarse fragment content (>2 mm) of each depth except for the surface depth was not significantly different (P > 0.05) among the three plantations.

Table 3 Bulk density, coarse fragment (>2 mm), soil carbon and nitrogen content at each depth of three coniferous plantations. Values in parenthesis represent standard errors
Full size table

The soil organic C concentration among the three plantations was not significantly different at a depth of 0–10 cm, but that at other depths showed a significant difference. The organic C concentration in the subsurface depth (10–30 cm) was significantly higher in the red pine than in the other two plantations (Table 3). It is not easy to compare the property of soil C in the three coniferous plantation sites because of the non-availability of data on soil conditions prior to plantation establishment, and the high variation in nutrient concentration, bulk density and coarse fragment of mineral soil layers (Binkley 1995).

The C and N distribution of the surface soil layer might reflect differences in the quality and quantity of litterfall inputs, litter decomposition dynamics, and the production and turnover of fine roots, which are the principal pathways for the return of C and N to the soil (Finzi et al. 1998; Wang et al. 2009). Despite having different mechanisms affecting the C and N cycle from litterfall inputs and decomposition processes, the absence of any significant difference in the soil C and N status of the larch and pine plantations might be due to other factors such as fine roots, which are also a main source affecting soil C and N contents. For example, Hwang et al. (2007) found that the production of fine roots was significantly greater in 44-year-old Japanese larch (1,762 kg ha−1 year−1) than in 40-year-old pitch pine (1,108 kg ha−1 year−1) plantations. This may explain why larch with low C input from needle litter showed a similar soil organic C content as both the pine plantations at surface depth. In addition, the slow decomposition rates of pine needle litter may result in more C accumulation in the forest floor and less C incorporation into mineral soils compared with larch needle litter. In contrast to the surface depth, the organic C concentration at subsurface depths (10–30 cm) was significantly higher in the red pine than in the larch or rigitaeda plantations. High understory vegetation cover (above 90% cover rates) by R. mucronulatum in the red pine plantation compared with the other two plantations could increase sources of organic matter such as fine roots and litterfall under dense understory vegetation. Chen and Li (2003) reported that shrubs can store more C at the deep soil horizon because shrubs may allocate more biomass to roots, especially fine roots.

The soil N concentration and content were not significantly different (P > 0.05) among the three coniferous plantations, with the exception of 10–20 cm depth in the rigitaeda plantation. The soil N content was generally higher in the larch than in both the pine plantations except at 20–30 cm depth (Table 3). The high soil N content in the larch plantation could be attributed to a combination of high N input by needle litterfall and N remaining in decomposing needle litter (Fig. 1, Table 2). In addition, N input by fine roots was higher in 44-year-old Japanese larch (23.1 kg ha−1 year−1) than in 40-year-old pitch pine (10.9 kg ha−1 year−1) plantations (Hwang et al. 2007).

Conclusions

Carbon and N status arising from needle litter input and decomposing needle litter differed considerably between a deciduous larch and two evergreen pine plantations established on sites with similar conditions. However, the C and N content at 0–10 cm soil depth were relatively little influenced by needle litter and decomposition dynamics over about 45 years of plantation establishment. Further studies will be needed to evaluate C and N contributions by the forest floor and fine roots to ascertain the impact of different tree species.