Litter Production and Leaf-Litter Decomposition in Natural and Monoculture Plantation Forests of Castanopsis kawakamii in Subtropical China

YANG Yu-Sheng 1,2** , LIN Peng 1 , GUO Jian-Fen 2 , LIN Rui-Yu 2 , CHEN Guang-Shui 2 , HE Zong-Ming 2


The amount and pattern of litterfall, and leaf-litter decomposition associated with its quality were studied in a natural forest of Castanopsis kawakamii (NF) and adjacent monoculture plantations of C. kawakamii (CK) and Chinese fir (Cunninghamia lanceolata, CF) in Sanming, Fujian, China. Mean annual total litterfall over 3 years of observations (from 1999 to 2001) was 11.01 t·ha -1 in the NF, 9.54 t·ha -1 in the CK and 5.47 t·ha -1 in the CF respectively. Of the total annual litterfall in the three forests, leaf litter constituted 59.70%, 71.98% and 58.29%, respectively. Litterfall in the NF and CK showed a unimodal distribution pattern. For the CF, the litterfall pattern was trimodal. The annual percent leaf litter mass loss was the highest for C. kawakamii in the NF (98.16%) and the lowest for Chinese fir (60.78%). Ratios of C/N and lignin/N had significantly negative influences on decay rate coefficients, while initial N and water soluble compounds exerted significantly positive influences. The results of this study demonstrate that the natural forest has a greater capability for maintaining site productivity than monoculture plantations due to higher amount and quality of litter coupled with faster litter decomposition. Therefore, conservation of the natural forest is recommended as a practical measure in forest management to realize sustainable development of forestry in mountainous areas of southern China.


Due to rapid increase of human population and subsequent demands for timber, fuel material, and other forest products, many natural forests in the world have been converted into plantations to meet these demands. However, problems of reducing community diversity, stability and sustainability of woodlands in monoculture plantations have aroused people's worries. In South China where high rainfall, steep slopes and fragile soil are characteristic, native broadleaved forests have been cleared for the last several decades, and successive monoculture plantations of economical conifers are established following forest land clearcutting, slash burning and soil preparation. Whereas yield decline and land deterioration in such disturbed ecosystem have become serious (Yang 1993; Yu 1996).

Natural forest of Castanopsis kawakamii located in National Nature Reserve of Xiaohu in Sanming, Fujian represents the precious evergreen broadleaved C. kawakamii forest in mid-subtropical China and unique in the world with its high purity (85% of relative prominence for C. kawakamii), old age (~150 year) and large area (~700 ha) (Lin et al. 1986). In 1966, part of natural C. kawakamii forest was clear-cut and pure Chinese fir (Cunninghamia lanceolata) and C. kawakamii plantations were established in 1967. These plantations and adjacent natural forest had homogeneous substrate (similar mineralogy, depths, and horizonation). Several studies have reported community structure and species diversity in natural C. kawakamii forest (Lin et al. 1986; Yang 1993). However, there is few information on litter comparison between natural and monocultural forests of the same tree species. Therefore, the primary aims of this study, covering a 3 year period, were to (i) examine the patterns of litterfall in natural C. kawakamii forest (NF) and two monoculture plantations of C. kawakamii (CK) and Chinese fir (CF), (ii) determine the relationship between decomposition rate and litter quality.

Materials and Methods

Site descriptions

The study was carried out from January 1999 to December 2001 in Xiaohu work area of Xinkou Experimental Forestry Centre of Fujian Agricultural and Forestry University, Sanming, Fujian, China ( 26°11′30″N, 117°26′00″E). The region has a middle sub-tropical monsoonal climate, with a mean annual temperature of 19.1 ºC, and a relative humidity of 81%. The mean annual precipitation is 1749 mm, mainly occurs from March to August (Fig.1). Mean annual evapotranspiration is 1585 mm. The growing season is relatively long with an annual frost-free period of around 300 days. The soil is classified as red soil derived from sandy shale and thickness exceeds 1.0 m. In 1999, five 20m×20m plots were randomly established at the midslope position in study sites of NF, CK and CF respectively. Various forest characteristics and some surface soil (0-20cm) properties under the three sites are described in Table 1.

Litter collection

Fifteen 0.5 m×1.0 m litter traps made of nylon mesh (1 mm mesh size) were arranged cater-cornered in each stand and were raised 25cm above the ground, and litterfall was collected at 10-day intervals from January 1999 to December 2001. The collected litter at each time was oven-dried at 80℃ to constant weight. At the end of each month, the oven-dried litter was combined and sorted into leaves, small branches (<2 cm in diameter), flowers, fruits, and miscellaneous material (insect fecal, unidentified plant parts, etc.). Further, leaf and small branch litter collected in the NF were separated into two classes, viz. C. kawakamii and other tree species in tree layer. Thereafter monthly mass of each fraction was determined.

Leaf-litter decomposition

The litterbag technique was used to quantify litter decomposition rate. In April 1999, freshly fallen/senescent foliage of C. kawakamii and other tree species in the NF and two plantation tree species were collected on nylon mesh screens which were placed in the experimental stands. Sub-samples from leaf litter of each species were retained for the determination of initial chemical composition. Except for leaf litter of single tree species of C. kawakamii in the NF and CK and Chinese fir, leaves of other species of trees in the NF and mixed-leaf of equal amount of the individual C. kawakamii and other tree species in the NF were employed for decomposition experiment. A known amount of air-dried leaf litter (20g) of each species or species combination was put into a 20cm×20cm, 1.0 mm mesh size nylon bag. For each type, 80 bags were prepared and randomly placed on the forest floor in the respective stands at the end of April 1999. After 30, 60, 90, 150, 210, 270, 330, 390, 510, 630 and 750 days of sample placement, six litterbags of each type were recovered at random from each forest site, and transported to the laboratory. The adhering soil, plant detritus and the "ingrowth" roots were excluded, and the bags were then oven-dried at 80℃ to constant weight for the determination of remaining weight.

Chemical analyses

Litter sub-samples for determination of initial chemical composition were oven-dried, ground and passed through a 1-mm mesh screen. For the determination of C, the plant samples were digested in K2Cr2O7-H2SO4 solution by oil-bath heating and then C concentration was determined by titration. For determination of N, P, K, Ca and Mg, the samples were digested in the solution of H2SO4-HClO4, and then N concentration was determined on the KDN-C azotometer, P concentration was analyzed colorimetrically with blue phospho-molybdate, K by flame photometry, and Ca and Mg concentrations were determined by the atomic absorption method (Department of National Forestry 2000). The initial organic constituents of fresh leaf litter samples including lignin, cellulose, hemicellulose, coarse protein, alcohol and water soluble compounds were determined by proximate chemical analysis (Wen et al. 1984).

Statistical analyses and calculations

The data on litterfall amount, percentage of leaf litter mass remaining during the first year and initial chemistry of leaf litter were analysed for differences between forests using one-way ANOVA. The multiple comparison was determined with SSR test at a significance level of 0.05 (SAS Institute 1998).

The model for constant potential weight loss is represented by the following equation (Olson, 1963): x / x0 = exp (-kt) , where x is the weight remaining at time t, x0 is the initial weight, the constant k is the decomposing coefficient, and t is the time. Correlation coefficients (r) between k and the chemical properties of leaf litter were also worked out.



There were significant differences (P<0.05) in the litter production among study forests (Table 2). Average annual litterfall (1999-2001) ranged from 5.47 t·ha -1 of the CF to 11.01 t·ha -1 of the NF. Of the total annual litterfall in the three forests, leaf litter constituted 59.70%, 71.98% and 58.29% respectively.

Total litterfall followed a unimodal distribution pattern for the NF and CK, with a distinct peak in April of each year. The CF showed trimodal pattern and the litterfall peaks occurred in April (or May), August and November respectively (Fig.2).

Chemical composition of leaf litter

Concentrations of Ca, Mg and C in Chinese fir needle were higher than those in other leaves (Fig. 3). No significant differences in N and P concentrations among various leaves were observed except for K. For concentrations of organic components, there were only significant differences for alcohol and water soluble compounds (P<0.05).

Leaf-litter decomposition

The significantly fastest decomposition was found for leaf litter of C. kawakamii in the NF in which only 1.84% of the initial mass remained at the end of the first year of study (P<0.05). Mass loss in Chinese fir needle litter was the least with 39.22% mass remaining undecomposed after one year (Fig. 4). Decomposition of various leaf litter over the 750-day period was characterized by an initial faster rate of disappearance, followed by a subsequent slower rate. For instance, C. kawakamii leaf litter in the NF lost 90.61% of initial weight in the first 150 day period, compared with 9.36% of that in the later 600 day period.



Litterfall production in forest ecosystem is determined by climatic condition, species composition and successional stage in its development (Haase 1999). In this study, the observed litterfall of the NF (11.01t·ha -1 ·yr -1 ) was in the upper part of the range recorded for subtropical evergreen broadleaved forests (Liang 1994; Zheng et al. 1995) and even equivalent to or higher than that in some tropical rain forests elsewhere in the world (Haase 1999; Songwe et al. 1988). The litter production in the CK and CF was lower than that in the NF but similar to that recorded in climatically comparable plantations (Tian et al. 1989; Zheng et al. 1995). The high species diversity of trees, soil fertility level and the standing crop (563.465 m 3 · ha -1 ) in the NF compared with monoculture plantations may explain the higher litterfall in the NF. Comparison between litterfall in the CK and CF showed significant differences (Table 2), which could be attributed to tree behaviour (Xu and Hirata 2002).

For the NF and CK, a major peak of litterfall was observed in April during 3-year period, corresponding to most of old leaves replaced by new ones in the spring. This rhythm was coincided with that of broadleaved forest in Dinghu mountain (Weng et al. 1993) and subtropical rain forest of Hexi (Zheng et al. 1995). Available studies concerning Chinese fir plantations mostly showed two maximum litterfall (Liang 1994; Tian et al. 1989), whereas our study showed the third smaller peak in August which was probably ascribed to high actual evapotranspiration (AET) and slow-growth characteristic of Chinese fir in the period (Yu 1996).

Leaf-litter decomposition related to its quality

At a regional scale with similar climatic conditions, litter decomposition rates are primarily controlled by litter quality (Aerts 1997). The higher amounts of alcohol/water soluble compounds seemed to stimulate the litterfall decomposition process, which was shown by 9.39% of dry weight remaining for leaves of C. kawakamii in the NF after 150 days of the onset of decomposition (Fig. 4). However, only strong relationship between initial concentration of water soluble compounds and decay rates of various leaf litter, can be recognized in our study (r= 0.98).

High initial concentrations of N in litter or P in litter at sites with low P availability have generally been considered to increase decomposition rates (Aerts 1997). Correlation studies confirmed that initial N and P concentrations (r=0.474 and 0.258 respectively) exerted positive influences on litter decay rate coefficients in this study. Unlike N and P, a high initial lignin concentration is expected to retard the decomposition process. Chinese fir needle had higher initial lignin concentration than other leaves (Fig. 3), which was in agreement with its lower decay rate (r= -0.235). Ratios of C/N (r= -0.821) and lignin/N (r= -0.563) showed significant negative relationships with decay rate coefficient (k). Previous workers also have elucidated such relationships (Cornelissen 1996; Singh et al. 1999). In this paper, annual decomposition rate of mixed-leaf (92.24%) was faster than that of other leaves of tree species in the NF (80.24%), showing somewhat stimulative decomposition of mixed foliar litters. A further study should be conducted to explore the possible interaction of mixed litter with different quality.


Aerts, R.,1997. Climate, leaf chemistry and leaf litter decomposition in terrestrial ecosystems: a triangular relationship. Oikos 79: 439~449

Cornelissen, J. H. C., 1996. An experimental comparisons of leaf decomposition rates in a wide range of temperate plant species and types. J. Ecol. 84: 573~582

Department of National Forestry, 2000. Forest soil analysis methods (in Chinese). Beijing: Chinese Criteria Press.

Haase, R., 1999. Litterfall and nutrient return in seasonally flooded and non-flooded forest of the Pantanal, Mato Grosso, Brazil. For. Ecol. Manage. 117(1-3): 129~147

Liang, H. W., 1994. Studies on the litterfall of two forest types in mid-altitude of Laoshan mountain in Tianlin Country. Chinese Journal of Ecology (in Chinese) 13 (1): 21~26

Lin, P., X. Z. Qiu, 1986. Study on the Castanopsis kawakamii forest in the Wakeng area of Sanming city, Fujian province. Acta Phytoecol. Geobot. Sinica (in Chinese) 10(4): 241~252

Olson, J. S., 1963. Energy storage and the balance of producers and decomposers in ecological systems. Ecology 44: 323~331

SAS Institute, 1998. The SAS System for Windows, version 7 ed. SAS Institute, Inc., Cary, NC.

Singh, K. P., P. K. Singh, S. K. Tripathi, 1999. Litterfall, litter decomposition and nutrient release patterns in four native tree species raised on coal mine spoil at Singrauli, India. Biol. Fertil. Soils 29: 371~378

Songwe, N.C., F. W. Fasehun, D. U. U. Okali, 1988. Littefall and productivity in a tropical rainforest, Southern Bakundu Forest Reserve, Cameroon. J. Trop. Ecol. 4: 25~37

Tian, D. L., K. Zhao, 1989. Studies on the litter in a Chinese fir plantation ecosystem I. Amount, composition and dynamics of litter. Journal of Central-South Forestry College (in Chinese) 9: 38~44

Wen, Q. X., L. J. Du, X. H. Zhang, 1984. Analysis for soil organic matter (in Chinese). Beijing: China Agriculture Press, 256~271

Weng, H., Z. A. Li, M. Z. Tu, et al. 1993. The production and nutrient contents of litter in forest of Dinghushan. Acta Phytoecol. Geobot. Sinica (in Chinese) 17 (4): 299~304

Xu, X. N. and E. Hirata, 2002. Forest floor mass and litterfall in Pinus luchuensis plantations with and without broad-leaved trees. For. Ecol. Manage. 157: 165~173.

Yang, Y. S., Z. W. Li, A. Q. Liu, 1993. Studies on soil fertility for natural forest of Castanopsis kawakamii replaced by broadleaf plantation. Journal of Northeast Forestry University (in Chinese) 21(5): 14~21

Yu, X. T., 1996. Silviculture of Chines fir (in Chinese). Fuzhou: Fujian Science and Technology Press.

Zheng, W. J., C. Shao, L. M. Wang, et al. 1995. Dynamic of nutrient elements in litterfall of subtropical rain forest of Hexi in Fujian. Journal of Tropical and Subtropical Botany (in Chinese) 3 (4): 38~43

Table 1. Forest characteristics and soil properties of the study sites

Forest type



NF (1)



Mean tree height (m)




Mean tree diameter at breast height (cm)




Stand density (stem·ha -1 )




Stand volume (m 3 · ha -1 )




Standing crop of forest floor (t·ha -1 )




Soil (top 0-20cm depth)


Bulk density (g·cm -3 )




Organic matter (%)




Total N (%)




Total P (%)




Available P (mg·kg -1 )




Notes: (1) Castanopsis kawakamii is only involved.

Table 2. Quantity (kg·ha -1 ·yr -1 ) and composition (%, in parentheses) of litterfall in three forests

Forest type


Leaf of other tree species *

Subtotal of leaf

Small branch

Branch of other tree species *

Subtotal of branch












± 125.99a



± 529.36a









± 9.36b


± 42.19b

± 532.39b













Notes: Values are means ±s.d. of five plots at each forest over 3 years. Means followed by different letters on the same column indicate significant differences at P<0.05. *Other tree species in the NF indicate those species in the tree layer except for C. kawakamii.

Fig.1. Temperature and rainfall patterns for the study area

● Monthly rainfall     ○ Monthly mean temperature

Fig. 2. Monthly variation in total litterfall in the three forests (n=15 per forest, error bars indicate ± s.d.)

●the NF     ○ the CK     △ the CF

Fig. 3. Initial chemical composition of leaf litter from the three forests (n=5, error bars indicate ± s.d.). Concentrations followed by different letters denote a significant difference at P<0.05. ASC, Alcohol soluble compounds; WSC, Water soluble compounds; CP, coarse protein

Fig. 4. Percentage of dry mass remaining in various leaf litter type over a 750 day period


1 College of Life Science, Xiamen University, Xiamen 361005, China
2 College of Forestry, Fujian Agriculture and Forestry University, Nanping 353001, China
** Author for correspondence.
Tel: (086) 0599-8504990;
E-mail: < ffcyys@public.npptt.fj.cn >

* Foundation item: The Post-doctoral Research Foundation of China, the Supporting Program for University Key Teacher by the Ministry of Education of China and the Key Basic Research Project of Fujian Province (2000F004).