Changes in soil organic carbon, nitrogen, pH and bulk

0 downloads 0 Views 622KB Size Report
*Key Laboratory of Forest Plant Ecology, Northeast Forestry University, ... duration is needed to rehabilitate the underground function of soil via the ... The answers to these questions are crucial for RFTF .... (ASA1, Institute of Geophysical and Geochemical Exploration, ...... former farmland has sequestrated 1.5 billion cubic.
Global Change Biology (2011), doi: 10.1111/j.1365-2486.2011.02447.x

Changes in soil organic carbon, nitrogen, pH and bulk density with the development of larch (Larix gmelinii) plantations in China W A N G W E N - J I E *, Q I U L I N G *, Z U Y U A N - G A N G *, S U D O N G - X U E *, A N J I N G *, W A N G H O N G - Y A N *, Z H E N G G U A N - Y U *, S U N W E I * and C H E N X I - Q U A N w *Key Laboratory of Forest Plant Ecology, Northeast Forestry University, Harbin 150040, China, wForestry College, Northeast Forestry University, Harbin 150040, China

Abstract Under the government of China’s environmental program known as Returning Farmland To Forests (RFTF), about 28 million hectares of farmland have been converted to tree plantation. This has led to a large accumulation of biomass carbon, but less is known about underground carbon-related processes. One permanent plot (25 years of observation) and four chronosequence plot series comprising 159 plots of larch (Larix gmelinii) plantations in northeastern China were studied. Both methods found significant soil organic carbon (SOC) accumulation (96.4 g C m2 yr1) and bulk density decrease (5.7 mg cm3 yr1) in the surface soil layer (0–20 cm), but no consistent changes in deeper layers, indicating that larch planting under the RFTF program can increase SOC storage and improve the physical properties of surface soil. Nitrogen depletion (4.1–4.3 g m2 yr1), soil acidification (0.007–0.022 pH units yr1) and carbon/ nitrogen (C/N) ratio increase (0.16–0.46 per year) were observed in lessive soil, whereas no significant changes were found in typical dark-brown forest soil. This SOC accumulation rate (96.4 g m2 yr1) can take 39% of the total carbon sink capacity [net ecosystem exchange (NEE)] of larch forests in this region and the total soil carbon sequestration could be 87 Tg carbon within 20 years of plantation by approximating all larch plantations in northeastern China (4.5 Mha), showing the importance of soil carbon accumulation in the ecosystem carbon balance. By comparison with the rates of these processes in agricultural use, the RFTF program of reversing land use for agriculture will rehabilitate SOC, soil fertility and bulk density slowly (o69% of the depletion rate in agricultural use), so that a much longer duration is needed to rehabilitate the underground function of soil via the RFTF program. Global forest plantations on abandoned farmland or function to protecting farmland are of steady growth and our findings may be important for understanding their underground carbon processes. Keywords: chronosequences, constant mass correction, Larix gmelinii plantations, long-term permanent plot, power analysis, returning farmland to forests, SOC accumulation, soil bulk density, soil nitrogen, soil pH

Received 17 October 2010; revised version received 5 February 2011 and accepted 19 March 2011

Introduction Soil organic carbon (SOC; approximately 1.5  1018 g) in active exchange with the atmosphere constitutes approximately two-thirds of the carbon in terrestrial ecosystems (Post et al., 1982), and small changes in the SOC may radically alter the balance of forest ecosystems. Agricultural practices, involving the clearing and use of original forests or grassland, can cause the loss of a large fraction of soil organic matter (SOM) (Tiessen et al., 1982; Mann, 1986). These practices, together with an increase in the amount of land used for agriculture in the last 200 years, has led to a decrease in carbon stored in soils and a net release of carbon into the atmosphere Correspondence: Zu Yuan-Gang, tel. 1 86 451 821 90092, fax 1 86 451 821 90092, e-mail: [email protected]; Wang Wen Jie, e-mail: [email protected]

r 2011 Blackwell Publishing Ltd

(Houghton et al., 1983), which has strongly influenced the global carbon balance (Sellers et al., 1997). In the black soil region in Northeastern China, over 50% of original SOC in the soil has been lost due to decades of agricultural practice after the establishment of the People’s Republic of China in 1949 (Xin et al., 2002; Song et al., 2005). Nitrogen loss has decreased soil fertility (Tiessen et al., 1994; Pimentel et al., 1995) and in northeastern China near half has been lost (Ding & Liu, 1980; Wang et al., 2002). Furthermore, past excesses of agricultural practice in China have led to natural disasters such as the widespread serious river floods of 1998 (HOMWR, 2002), which led to a new environmental protection policy implemented by China’s central government, known as Returning Farmland to Forests (RFTF) (Li, 2004). RFTF is an environmental initiative that is based on natural principles (Li, 2004). Upto the end of 2009, the central government of China had invested 233.2 billion 1

2 W . W E N - J I E et al. RMB in the program and a further 200 billion RMB will be invested in it from 2010 to 2021. As a result, millions of hectares of farmland were afforested from 1999 to 2009 (see http://politics.people.com.cn/GB/1026/ 12477229.html as at September 13, 2010). The RFTF policy originally aimed to recover the hydrological protecting function of forests or grasslands so as to reduce flood disasters as well as extreme weather, but carbon sequestration is also important and the present study seeks to clarify belowground soil carbon-related processes. Several questions arise. First, can the RFTF practice rapidly increase SOC storage? Second, can the soil fertility, chemistry and soil physical properties, in particular the soil nitrogen, pH value and bulk density, be returned to its original level as in virgin forest soil? Can the soil fertility support the long-term growth of plantation (biomass carbon accumulation) and soil carbon accumulation? The answers to these questions are crucial for RFTF practices regarding the local carbon balance and the implications for global warming in both the short- and long-term. Moreover, global plantation forests are upto 264 million hectares in 2010 and their establishment on land that until then was not classified as forest, such as farmlands functioning for agricultural protection or abandoned farmland for ecological function just like RFTF practice in this study (FAO, 2010). The answers may be meaningful in a global scale at this viewpoint. Today’s data shortage should be remedied by systematic studies (Li, 2004; Zu et al., 2009). Study of a typical plantation ecosystem and of soil carbon, soil nitrogen and soil pH changes will provide the data necessary to make informed decisions in the future (e.g. Berthrong et al., 2009). Larch trees comprise one of the main species used in RFTF practices in northeastern China (Hou et al., 2004); more than half of the total afforested area in northeastern China has been planted by this species (Wang et al., 2005a; Sun et al., 2007). The black soil [mainly dark brown forest soil and lessive (Baijiang) soil in forest region] in northeastern China contains abundant organic matter and is of high fertility compared with other soils (Forest Soil group of CAS, 1980; HLJTR, 1992), but has been dramatically affected by global warming (Piao et al., 2009) and by historical agricultural reclamation (Song et al., 2005). The high biomass productivity of larch forests is well known (Jiang et al., 2002; Wang et al., 2005a; Sun et al., 2007) and is an important carbon sink at global level (Gower & Richards, 1990; Schulze et al., 1999), but the soil carbon beneath these forests have scarcely been studied systematically (Zu et al., 2009). We therefore studied the dynamics of soil carbon in larch forests. Soil carbon dynamics studies often involve spatial– temporal substitution methods, such as the chronosequence method (Covington, 1981; Wang et al., 2006),

which is a way of making long-term studies of vegetation succession and soil dynamics; with more replicated samples, random errors should be smaller, although systematic errors cannot be excluded if different sites are taken to have some properties in common (Johnson & Miyanishi, 2008; Walker et al., 2010) and statistical analysis, such as power analysis may detect the data probability (power) for finding significant changes within heterogeneous soils (Kravchenko & Robertson, 2011). In contrast, the most accurate method for measuring change in soil carbon is to repeatedly sample a single site over time and ensure that soils are collected, processed and analyzed in a consistent manner (Paul et al., 2002; Zhou et al., 2006). Very few field sites have been consistently studied for a timescale of years or decades (Zhou et al., 2006; Wirth et al., 2009), however, so it is which made it difficult to make a full views of large scale soil carbon balance. A combination of the two methods in a single study of the soil carbon balance may give more precise results, and this is what is carried out in the present study. We chose in this work to study one permanent plot and four chronosequence plot series consisting of 159 sites in larch plantations. We aim to determine the annual rate of change of SOC, soil nitrogen, soil pH and soil bulk density after returning farmland to forest. This will reveal the importance of soil carbon processes in the ecosystem carbon budget and provide information on possible soil fertility problems during land rehabilitation. Comparison with published work of ecosystem carbon sinks together with knowledge of changes during past agricultural reclamation in history should be a useful guide for RFTF practice, as well as soil carbon dynamics for global forest plantations.

Material and methods

Study sites All study sites are located in the central region of Larix gmelinii plantations, in which the RFTF program was implemented (Li, 2004). At the Laoshan site, a permanent plot and a chronosequence plot series with 57 plots were used, whereas three chronosequence plot series were surveyed at the Maoershan, Dongshan and Daqingchuan sites (Fig. 1). This region has a continental monsoon climate, with average annual temperature 0.3 to 2.6 1C and precipitation of 676–724 mm. Accompanying saplings, shrubs and herbs include Betula platyphylla, Fraxinus mandushurica, Quercus mongolica, Acanthopanax senticosus, Euonymus pauciflorus, Agrimonia obtusifolia, Chelidonium majus and Cacalia hastate. The soil in this region, as shown in picture 1 in Appendix S1, is generally normal dark-brown forest soil (i.e., Mollic Bori-Udic Cambosols, at Maoershan, Dongshan and Daqingchuan) and some regions are characterized as lessive (Baijiang) soil (Mollic Albi-Boric Argosols

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S

3

Fig. 1 Larix gmelinii distribution in China and the study sites in this paper.

Table 1

Situation of the four chronosequence plot series and the permanent plot Laoshan

Maoershan

Dongshan

Daqingchuan

Coordination Soil type

45120 0 N, 127134 0 E Lessive soil

Permanent plot Plot number for Chronosequence series Age ranges (years)

1 57

45124 0 N, 127133 0 E Dark brown forest soil – 37

46157 0 N, 129110 0 E Dark brown forest soil – 30

47100 0 N, 129107 0 E Dark brown forest soil – 35

0 (farmland)–40

0 (farmland)–47

0 (farmland)–52

334(SD 5 30)

243(SD 5 11)

247(SD 5 39)

Altitude a.s.l. (m)

0 (farmland)–50 Permanent plot, 52 295(SD 5 35)

–, no data available.

according to China Soil Taxonomy), at Laoshan (Chen, 1986; Gong et al., 2007). Both soils are generally originated from forest lands (former is generally from Korean pines-broadleaved mixed forests, while the latter is generally from hygrophytes such as woody plants of ash and birch) (HLJTR, 1992) and the same origination from forests may avert the interfering influences both from variable soil origin and development of plantation forests (Ce´spedes-Payret et al., 2009).

Plot selection, collection of soil samples and tree age determination For the chronosequence plot series, the stand selection criteria restrict the differences between the sites so that the length of time since the farmland was returned to forest practice is the

main variable influencing uneven changes in forest soil among the stands selected. As well as the stand age, topographic position, slope and elevation might influence soil carbon and soil fertility (Garten et al., 1994; Enoki et al., 1996). The plots were selected to have a similar elevation of about 280 m (SD 5 43 m) and slope of o20%, so as to minimize the differences from nonchronosequence sources. The history of the plot before larch plantation was obtained from the forestry farm’s inventory data and from asking local formers. All of the plots in this study were historically farmland, producing soybean or corn in the short term (as short as several years) or long term (decades). As shown in Table 1, a total of 159 plots were surveyed in the four chronosequence plot series: 57 plots in the Laoshan site, 37 plots in the Maoershan site, 30 plots in the Dongshan site and 35 plots in the Daqingchuan site.

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

4 W . W E N - J I E et al. At each of the 159 plots, a 20 m  20 m quadrat was surveyed. Four soil profiles in each plot (approximately 2 m long, 1 m width and 0.8 m deep) were dug out. After the exclusion of recognizable litter at the soil surface (A0 layer), 100 cm3 of soil was sampled at each layer of depth, 0–20, 20–40, 40–60 and 60– 80 cm. Finally, four soil samples from the same soil layers of four soil profiles were mixed as a composite sample of the soil of the plot at each soil depth. Using this method, four soil samples (0–20, 20–40, 40–60, 60–80 cm) were collected in each plot of the chronosequence plot series. At the permanent plot, soil samples were collected in 1983 and 2008 using similar procedures. The larches in the permanent plot (200 m  200 m) (Laoshan) were planted in 1958. One investigation of soil carbon, soil nitrogen and soil pH was carried out in 1983. This comprised 10 soil profiles with depth over 80 cm; two profiles made up a composite sample so that there were a total of five samples for each layer (0–20, 20–40, 40–60 and 60–80 cm). Using the same sampling method, 24 soil profiles (depth 80 cm) were dug out in 2008 and four profiles made up a composite sample, so that six composite samples for each layer (using the same mixing sequence) resulted for laboratory analysis. The 1983 data have been published (NEFU, 1984; Chen, 1986). The age of larch trees were determined using an increment borer (Zhonglinweiye, Beijing). At least five wood cores in each plot were drilled, and tree rings were counted to give the mean age of the plantation. Although 30–57 plots were surveyed in each chronosequence plot series to obtain a nearly uniform distribution of plot numbers in each age class, no larch trees were planted in some years in certain regions (e.g., no larch trees were planted at the Laoshan site from 1992 to 2000, so that no 8–16-year-old plantations were observed in this site); this can be seen in Figs 2–6.

SOC, total nitrogen, pH and bulk density measurement in the laboratory Soil cups of size 100 cm3 were used to collect soil from each layer of the soil profile, and the samples were placed in a cloth bag and aired in a dry ventilated room until they reached constant weight, in order to determine the soil bulk density. After roots in the soil sample were carefully picked out, the samples were ground with a wooden rolling pin and passed through a 2 mm soil sieve. The gravels were sieved out of the soil sample and the gravel weight and volume were measured. The o2 mm component of the soil sample was smashed in a motor pulverizer for approximately 3 min and passed through a 0.25 mm soil sieve. The o0.25 mm soil samples were stored in our laboratory before measurement of SOC, soil nitrogen and soil pH. SOC was measured by the heated dichromate/titration method (Bao, 2000). In this procedure, potassium dichromate (K2Cr2O7) and concentrated H2SO4 are added to between 0.1 and 0.5 g of soil. The solution was fully mixed and halted overnight, then the tubes were put into auto-controlled 170– 180 1C oil bath for 5 min. The FeSO4 solution with an exact concentration was used to titrate the digested solution with two to three drops of phenanthroline as color indicator (brick-

red is the endpoint of the titration). Total soil nitrogen concentration was determined by the Semimicro-Kjeldahl method (Bao, 2000). The soil samples and the standard soil sample (ASA1, Institute of Geophysical and Geochemical Exploration, Langfang, China) were measured together. In each digestion a standard reference sample tube was added, and the final SOC content and soil nitrogen of the reference supposes that digestion is complete. The pH of the soil solution (one part soil to five parts water) was measured with an acidity meter (Sartorius PT-21, Shanghai, China).

Data analysis The SOC and soil nitrogen storage was calculated using the following formula: SOC or soil nitrogen ðg m2 Þ ¼ a  rb  0:2  ð1  c  xÞ ð1Þ  ð1  Vgravel Þ; where a is the concentration of SOC (g kg1) or soil nitrogen (g kg1) in soil, rb is the soil bulk density (g cm3 5 ton m3), Vgravel is the volume percentage of gravel (%) and 0.2 represents the constant depth of each soil layer. Because land-use changes may significantly affect the bulk density, the fixed depth calculation may bias-estimate the total carbon or nitrogen storage owing to more or less inclusion of original soils in the calculation (Davidson & Ackerman, 1993). For getting a constant mass at different ages in this study, a correcting factor 1c  x for each larch age was used for offsetting soil bulk density decrease with larch age. 1c  x is the age-dependent correcting factor, where c is the slope value between soil bulk density and larch age (as shown in Fig. 5 of this paper) and x is the age of larch trees. When there was no significant relations, c is 0 and it means no need to make any constant soil mass correction. The SOC or soil nitrogen changes during development of the larch plantation were calculated as the gradient of the linear relation between forest age and SOC or soil nitrogen storage. Similarly, the changes of pH and soil bulk density were also calculated as the gradient of the relation between tree age and pH or bulk density. Regression analysis for the linear correlation was performed by SPSS 13.0 (SPSS, Chicago, USA). The parameters of statistical significance of linear correlation, such as slope value and its standard error, 95% lower and upper confidence limits as well as 95% lower and upper predicting bands and coefficient of determination (r2) and P-level, were computed. Large within-treatment errors can induce the data having a low statistical power for detecting significant differences when they really exists and a power analysis can tell what are the changes they are able to detect with these data (Kravchenko & Robertson, 2011). Thus, a power analysis was also performed by a JFM 9.0 software (SAS, Cary, USA) to find the statistical power of our data for detecting significance of the linear correlations for all the relations. Two parameters, i.e. LSN (the least significant number: the number of observations that would produce a specified P-value o0.05, if the data has the same structure and estimates as the current sample) and Power (the probability of getting a significant linear slope at

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S

5

Fig. 2 Age–soil organic carbon (SOC) relations for different chronosequence plot series at Laoshan, Maoershan, Dongshan and Daqingchuan. In the significant linear correlations, 95% confidence bands were within the two dot lines and 95% prediction bands were within the two dashed lines. Power is the probability of getting a significant linear slope at Po0.05 for the correlation analysis. The least significant number (LSN) is the data number that would produce a specified P-value o0.05, if the data has the same structure and estimates as the current sample.

or below a given P-value o0.05 of the data), were calculated via the power analysis. For the permanent plot at Laoshan site, the SOC and soil nitrogen stocks in 1983 and 2008 were calculated according to Eqn (1). The differences in SOC, soil nitrogen,

soil pH and bulk density between these two sampling years are taken to be due to the 25-year development of the larch plantation, and mean annual changes of these parameters were calculated by dividing the total changes by 25.

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

6 W . W E N - J I E et al.

Fig. 3 Age–nitrogen (N) relations for different chronosequence plot series at Laoshan, Maoershan, Dongshan and Daqingchuan. Notes are the same as in Fig. 2.

Results

Changes of SOC, soil nitrogen, pH and bulk density in the four chronosequence plot series and their power analysis Laoshan chronosequence plot series (57 plots). As shown in left column of the Fig. 2 and Table 2, significant linear

correlation between SOC stocks and larch age was observed (gradient 5 0.0658, r2 5 0.292, Po0.05) in the surface soil layer (0–20 cm), whereas in the deeper soil layers a tendency to decrease was found, and the deep layer at 60–80 cm showed a significant decrease (slope 50.0236, r2 5 0.142, Po0.05). The significant

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S

7

Fig. 4 Age–soil pH relations for different chronosequence plot series at Laoshan, Maoershan, Dongshan and Daqingchuan. Notes are the same as in Fig. 2.

linear relationship at 0–20 cm corresponds to a SOC accumulation rate of 65.8 g C m2 yr1 (95% lower and upper confidence limits were, respectively, 38.1 and 93.5 g C m2 yr1), while the depletion rate at 60–80 cm was about 23.6 g C m2 yr1(95% lower and upper confidence limits were, respectively, 39.3 and 7.9 g C m2 yr1). The soil carbon stock from 0 to 80 cm (derived by summing the various layers) does not change significantly with the development of larch forest plantation at the Laoshan site. No significant changes in soil nitrogen storage during larch plantation development were found in the top 40 cm (r2o0.0451, P40.05), while significant decreases were found in the 40–60 cm layer (slope 50.0024, r2 5 0.1518, Po0.05) and the 60–80 cm layer

(slope 50.0019, r2 5 0.1233, Po0.05) (Fig. 3 left column and Table 2). The rates of nitrogen depletion rate were 2.4 and 1.9 g m2 yr1 for the 40–60 and 60–80 cm soil layers, respectively. Their 95% lower and upper confidence limits were 3.9, 0.9 g m2 yr1 for the 40–60 cm, and 3.2, 0.5 g m2 yr1 for the 60–80 cm, respectively (Table 2). The overall soil nitrogen depletion rate for 0–80 cm depth was 5.6 g m2 yr1 (r2 5 0.0765, Po0.05) and 95% lower and upper confidence limits were, respectively, 10.8 and 0.3 g m2 yr1 (see Table 2). No significant changes in soil pH during larch plantation development took place in the top 20 cm (P40.05), but in the deeper layers a significant decrease in soil pH was found (Fig. 4, left column); r2 for the 20– 40, 40–60 and 60–80 cm layers was, respectively, 0.14,

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

8 W . W E N - J I E et al.

Fig. 5 Age–bulk density relations for different chronosequence plot series at Laoshan, Maoershan, Dongshan and Daqingchuan. Notes are the same as in Fig. 2.

0.32 and 0.28 and the corresponding linear gradients were 0.0078, 0.0111 and 0.0096 (Fig. 4, left column and Table 2). Based on these slope values, the pH of the different soil layers decreases by 0.20, 0.28 and 0.24, respectively, during the 25 years of development of the larch plantation (1983–2008). Their 95% upper confidence limits were 0.32, 0.39 and 0.34, while their 95% lower confidence limits were 0.07, 0.17 and 0.14 (see Table 2) (slopes  25 years). The soil bulk density at the surface layer (0–20 cm) decreased significantly with larch age (slope 50.0067, r2 5 0.2115, Po0.05), but no significant relations were found for the other soil layers (r2o0.06, P40.05) (Fig. 5, left column and Table 2). The rate of decrease of the bulk density was 6.7 mg cm3 yr1 for the top 0–20 cm soil and its 95% upper and lower confidence limits were, respectively, 10.2 and 3.2 mg cm3 yr1 (Table 2). As shown in left column of Fig. 6, the carbon/nitrogen

(C/N) ratio increased with larch plantation development in 0–20 cm soil layer (slope 5 0.2825, r2 5 0.1068, Po0.05) and the overall 0–80 cm profile (slope 5 0.1774, r2 5 0.1465, Po0.05), but no such tendency was found in any of the soil layers (r2o0.04, P40.05) (Fig. 6, Table 2). From the slope value of C/N ratio with time during larch plantation development, the rate of increase for the 0–20 cm soil was 0.2825 per year with 95% upper and lower confidence limits of 0.0596 and 0.5055 per year, and for the 0–80 cm layer was 0.1774 per year and its 95% upper and lower confidence limits were, respectively, 0.0606and 0.2943 per year (Table 2).

Maoershan chronosequence plot series (37 plots) A tendency for SOC to increase was observed in the top 60 cm of soil, but significant linear correlations (Po0.05) was found only in the surface soil layer, 0–20 cm

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S

9

Fig. 6 Age–C/N ratio relations for different chronosequence plot series at Laoshan, Maoershan, Dongshan and Daqingchuan. Notes are the same to Fig. 2. C, carbon; N, nitrogen.

(slope 5 0.1394, r2 5 0.4605, Po0.05) and 20–40 cm (slope 5 0.0696, r2 5 0.1428, Po0.05) (Fig. 2, middle left column and Table 2). In the deep layer at 60–80 cm, a slight decrease but no significant correlation was found (r2 5 0.0114, P40.05). The significant linear relationship revealed that SOC accumulates in the 0–20 and 20–40 cm soil layers during larch plantation development at rates of 139.4 and 69.6 g C m2 yr1 (95% lower confidence limits, 87.6 and 11.1 g C m2 yr1; 95% upper confidence

limits, 191.2 and 128.2 g C m2 yr1) (Table 2). Analysis of pooled data indicates that SOC in 0–80 cm increased significantly with forest age (slope 5 0.2273, r2 5 0.2643, Po0.05), with an overall SOC accumulation rate of upto 0.2273 kg C m2 yr1(227.3 g C m2 yr1) and its 95% lower and upper confidence limits were, respectively, 97.2 and 357.4 g C m2 yr1 (Table 2). Soil nitrogen storage tended to increase with larch plantation development, but only a significant linear

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

10 W . W E N - J I E et al. Table 2 Slope and intercept values of the significant linear correlations between larch ages and different soil parameters as well as the statistics for slopes Slopes Site and layer Age-SOC Laoshan 0–20 cm Laoshan 60–80 cm Maoershan 0–20 cm Maoershan 20–40 cm Maoershan 0–80 cm Dongshan 0–20 cm Dongshan 0–80 cm Daqingchuan 0–20 cm Age-N Laoshan 40–60 cm Laoshan 60–80 cm Laoshan 0–80 cm Maoershan 0–20 cm Maoershan 0–80 cm Dongshan 0–20 cm Age-pH Laoshan 20–40 cm Laoshan 40–60 cm Laoshan 60–80 cm Age-bulk density Laoshan 0–20 cm Maoershan 0–20 cm Maoershan 20–40 cm Dongshan 0–20 cm Daqingchuan 0–20 cm Age-C/N Laoshan 0–20 cm Dongshan 20–40 cm Laoshan 0–80 cm Maoershan 0–80 cm Daqingchuan 40–60 cm

Values

SE

t-Value

Prob > jtj

95%LCL

95%UCL

Intercepts

0.0658 0.0236 0.1394 0.0696 0.2273 0.1242 0.1548 0.0579

0.014 0.008 0.026 0.029 0.064 0.024 0.070 0.022

4.8 3.0 5.5 2.4 3.5 5.2 2.2 2.6

0.000 0.004 0.000 0.021 0.001 0.000 0.035 0.013

0.0381 0.0393 0.0876 0.0111 0.0972 0.0749 0.0115 0.0129

0.0935 0.0079 0.1912 0.1282 0.3574 0.1736 0.2981 0.1029

3.40 1.91 5.09 4.02 14.11 8.12 18.88 6.82

0.00243 0.00187 0.0056 0.01074 0.0190 0.00600

0.001 0.001 0.003 0.004 0.009 0.002

3.2 2.8 2.1 3.0 2.1 2.6

0.002 0.007 0.037 0.004 0.041 0.015

0.0039 0.0032 0.0108 0.0036 0.0008 0.0013

0.0009 0.0005 0.0003 0.0179 0.0371 0.0107

0.17 0.14 0.81 0.43 1.18 0.56

0.0078 0.0111 0.0096

0.003 0.002 0.002

3.0 5.1 4.7

0.004 0.000 0.000

0.0129 0.0155 0.0136

0.0027 0.0067 0.0055

5.57 5.61 5.58

0.0067 0.0069 0.0062 0.0037 0.0027

0.002 0.003 0.002 0.002 0.001

3.8 2.5 2.8 2.2 2.1

0.000 0.018 0.008 0.037 0.039

0.0102 0.0125 0.0107 0.0071 0.0052

0.0032 0.0013 0.0017 0.0002 0.0001

1.39 1.21 1.33 1.07 1.11

0.2825 0.1733 0.1774 0.1162 0.1666

0.111 0.070 0.058 0.052 0.082

2.54 2.46 3.04 2.23 2.04

0.014 0.020 0.004 0.032 0.049

0.0596 0.0292 0.0606 0.2220 0.0007

0.5055 0.3174 0.2943 0.0104 0.3326

11.40 12.27 12.10 14.26 9.32

The slope values correspond to the significant correlations in Figs 2–6. LCL, lower confidence limit; UCL, upper confidence limit.

relations were found in 0–20 cm layer (slope 5 0.0107, r2 5 0.2093, Po0.05) (Fig. 3, middle left column and Table 2). The pooled soil nitrogen storage from 0 to 80 cm was also significantly correlated with larch age (slope 5 0.019, r2 5 0.114, Po0.05). The soil nitrogen accumulation rate for 0–20 cm soil layer was 10.7 g m2 yr1 (95% lower and upper confidence limits were, respectively, 3.6 and 17.9 g N m2 yr1) (Table 2). In the case of 0–80 cm soil profile, the rate was upto 19.0 g m2 yr1 (Table 2). No significant changes in soil pH at different layers were observed (r2o0.02, P40.05) (Fig. 4, middle left column), showing that larch plantation growth did not cause any significant changes in soil acidity in this chronosequence plot series.

The soil bulk density decreased significantly with larch age in the 0–20 cm soil layer (slope 50.0069, r2 5 0.1501, Po0.05) and the 20–40 cm soil layer (slope 50.0062, r2 5 0.1832, Po0.05), whereas the bulk density in the other soil layers did not change significantly with larch age (r2o0.009, P40.05) (Fig. 5, middle left column and Table 2). The rate of reduction of the soil bulk density was 6.9 mg cm3 yr1 at 0–20 cm depth and 6.2 mg cm3 yr1 at 20–40 cm depth. No consistent pattern was found in the relations between the soil C/N ratio and larch age in different soil layers (r2o0.09, P40.05), while a significant decrease was observed in the pooled 0–80 cm soil profile (slope 50.1162, r2 5 0.1243, Po0.05) (Fig. 6, middle left column and Table 2). Its 95% lower and upper

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S confidence limits were, respectively, 0.2220 and 0.0104 (Table 2).

Dongshan chronosequence plot series (30 plots) The relation between SOC and larch age was significant in the 0–20 cm soil layer (slope 5 0.1242, r2 5 0.4871, Po0.05) in the Dongshan plot series, but not in the 20–40, 40–60 and 60–80 cm layers (r2o0.06, P40.05) (Fig. 2, middle right column and Table 2). According to the significant relation for 0–20 cm depth, the accumulation rate of SOC was 124.2 g C m2 yr1, and its 95% lower and upper confidence limits were, respectively, 74.9 and 173.6 g C m2 yr1 (Table 2). The overall 0–80 cm soil stock of organic carbon also increased significantly with the development of the larch plantation (slope 5 0.1548, r2 5 0.1448, Po0.05) (Table 2). As shown in middle right column of Fig. 3, no significant correlations between soil nitrogen storage and larch age were observed in the differing soil layers (r2o0.02, P40.05) except the 0–20 cm layer, a significant positive linear relation (slope 5 0.006, r2 5 0.1931, Po0.05) (Table 2). In this chronosequence series, accumulation of soil nitrogen was observed in surface 20 cm layer at a rate of 6.0 g N m2 yr1 (95% lower and upper confidence limits were, respectively, 1.3 and 10.7 g N m2 yr1) (Table 2). Differently, no consistent pattern in the soil pH value during larch plantation development was found in different soil layers (r2o0.005, P40.05) (see Fig. 4, middle right column). The soil bulk density decreased significantly with larch age in the 0–20 cm soil layer (slope 50.0037, r2 5 0.1462, Po0.05), whereas the bulk density in other soil layers did not change significantly with larch age (r2o0.02, P40.05) (Fig. 5, middle right column and Table 2). The rate of decrease of the surface soil bulk density was 3.7 mg cm3 yr1 (95% lower and upper confidence limits, 7.1 and 0.2 mg cm3 yr1) (Table 2). The C/N ratio generally increased, but only in the 20–40 cm layer was the relation statistically significant (slope 5 0.1733, r2 5 0.1781, Po0.05)( see Fig. 6, middle right column and Table 2). The rate of increase of the C/N ratio for 20–40 cm soil was 0.1733 per year with 95% lower and upper confidence limits of 0.0292 and 0.3174 per year.

Daqingchuan chronosequence plot series (35 plots) The relationship between SOC and larch age was significant in the 0–20 cm soil layer (slope 5 0.0579, r2 5 0.1722, Po0.05), but was not significant in the 20–40, 40–60 and 60–80 cm layers, nor in the pooled data for depth 0–80 cm (r2o0.07, P40.05) (Fig. 2, right column and Table 2). The rate of accumulation from 0 to 20 cm was 57.9 g C m2 yr1 with 95% lower and upper confidence limits of 12.9 and 102.9 g C m2 yr1, and

11

there is SOC accumulation at the surface but not in the deep soil layers. No significant correlation between soil nitrogen storage and larch age was observed in the various soil layers or in the pooled data for the profile from 0 to 80 cm (r2o0.07, P40.05). It follows that larch plantation growth does not cause any change in soil nitrogen storage (Fig. 3, right column). No consistent pattern in soil pH value was found during larch plantation growth in the various soil layers (r2o0.03, P40.05) (see Fig. 4, right column). We found, however, that the bulk density of soil in the 0–20 cm layer decreased significantly with larch age (slope 5 0.0027, r2 5 0.1226, Po0.05), but not in other soil layers (r2o0.02, P40.05)(Fig. 5, right column and Table 2). The rate of decrease of the surface soil bulk density was 2.7 mg cm3 yr1 with 95% lower and upper confidence limits of 5.2 and 0.1 mg cm3 yr1 (Table 2). The C/N ratio generally increases, but only in the 40–60 cm layer was the relation statistically significant (slope 5 0.1666, r2 5 0.1123, Po0.05); with a rate of increase of 0.1666 per year (95% lower and upper confidence limits: 0.0007 and 0.3326 per year) (see Fig. 6, right column and Table 2).

Power analysis of SOC–age, soil nitrogen–age, pH–age and bulk density changes–age relations at different chronosequence plot series For the significant slopes for the age–SOC relations at 0–20 cm layers (Laoshan, Maoershan and Dongshan), power analysis indicated our data could yield an 100% chance (power 5 1.00) of detecting the larch growth induced SOC accumulation, while in some others (Daqingchuan 0–20 cm, Maoershan 20–40 cm and 0–80 cm of Maoershan and Dongshan), our data had 57–92% chance (power 5 0.57–0.93) (Fig. 2). All the nonsignificant correlations showed o45% power to detect the possible significant changes if it really exists, and generally, much more data are needed for observing significant correlations (LSN 5 57–5045) with a median of 273 given that data structure is the same to current samples (Fig. 2). There was a quite high (71–89%) chance to detect the SON changes at Maoershan 0–20 cm soil, Dongshan 0–20 cm soil and Laoshan 40–80 cm soils, while general low chance (54–55%) in the pooled 0–80 cm soils at Laoshan and Maoershan was observed although significant correlations were observed (Fig. 3). Comparing with the SOC–age power analysis, the LSN for SON– age analysis is more similar to the data number of our sampling at each chronosequence series (Fig. 3) and a larger dataset may improve the statistical power. In the case of nonsignificant correlations, our data had very low power (probability) from 5% to 35% to detect the

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

12 W . W E N - J I E et al. Table 3

pH, C, N, C/N and bulk density changes from 1983 to 2008 in the permanent plot at Laoshan site Changes per year

SOC storage (kg C m2)

N storage (kg N m2)

pH

Bulk density (g cm3)

C/N ratio

Soil layer (cm)

Year 1983

Year 2008

Difference

Permanent plot

Chronosequence plots

0–20 20–40 40–60 60–80 Total (0–80) 0–20 20–40 40–60 60–80 Total (0–80) 0–20 20–40 40–60 60–80 Average 0–20 20–40 40–60 60–80 Average 0–20 20–40 40–60 60–80 Average

4.67 2.35 1.77 1.07 9.86 0.32 0.16 0.11 0.11 0.69 5.84 5.81 5.88 5.93 5.86 1.28 1.41 1.44 1.48 1.4 14.65 14.88 16.58 9.76 13.97

7.04 2.50 1.32 1.04 11.89 0.33 0.11 0.06 0.10 0.596 5.58 5.31 5.21 5.18 5.32 1.08 1.39 1.52 1.54 1.38 25.24 22.92 36.84 16.50 25.38

2.37 0.15 0.45 0.03 2.03 0.012 0.048 0.053 0.014 0.104 0.26 0.51 0.68 0.75 0.55 0.2 0.03 0.08 0.06 0.02 10.59 8.04 20.26 6.74 11.41

94.8 5.8 18.1 1.3 81.2 0.47 1.94 2.13 0.56 4.14 0.011 0.02 0.027 0.03 0.022 8 0.8 3.2 2.4 0.8 0.42 0.32 0.81 0.27 0.46

65.8 0 0 23.6 42.2 0 0 2.4 1.9 4.3 0 0.0078 0.0111 0.0096 0.0071 6.7 0 0 0 1.7 0.2825 0 0 0 0.1774

(2.02) (1.42) (0.96) (0.32) (0.23) (0.02) (0.01) (0.01) (0.18) (0.18) (0.03) (0.04) (0.10) (0.06) (0.04) (0.01) (6.4) (9.0) (8.9) (2.9)

(0.27) (0.98) (0.24) (0.13) (0.09) (0.02) (0.04) (0.07) (0.10) (0.26) (0.20) (0.14) (0.02) (0.04) (0.07) (0.04) (16.5) (9.4) (25.1) (10.3)

Units for change per year: pH, unit yr1; carbon and nitrogen storage, g m2 yr1; C/N ratio, yr1; bulk density, mg cm3 yr1. Chronosequence plot data show a linear slope value as in Figs 2–6; a 0 indicates that the linear correlation is not statistically significant (P40.05).

significant changes of SON if it really exists. The LSN was generally very large (hundreds and even thousands) with a median 444 (Fig. 3), which is much larger than our sampling data number. The pH data also had a high probability (85–100%) for finding the significant decrease in soil pH in 20–80 cm soil of Laoshan site, while there was o15% chance to detecting the significant pH changes for all others if it really exists (power 5 0.05–0.15) (Fig. 4). LSN for the nonsignificant correlations were from 159 to 13 483 with a median of 1494, which is much larger than any possible sampling protocol for a soil study (Fig. 4). Soil bulk density data had 55%–97% chances for detecting the significant density decrease in surface 0–20 cm layer of various chronosequence series. However, in the deeper layers, the possibility sharply decreased to 5%– 45% (power 5 0.05–0.45) and no significant relations were found (Fig. 5). LSN number for these nonsignificant correlations ranged from 66 to several million with a median of 469, showing that it is almost impossible to detect any significant changes in practices (data number is too large) (Fig. 5).

In the case of C/N ratio, our data had 5–66% chances for detecting significant C/N changes with larch growth, while five significant correlations were observed when the power was higher 51% (Fig. 6). For all the nonsignificant correlations, the LSN was median at 172 with the largest number of 656 095; showing enlargement of sample number may improve statistical power for detecting significant difference (Fig. 6). However, such large dataset is almost impracticable and conclusion of no significant change is reasonable.

Comparison of SOC, soil nitrogen, pH, soil bulk density and C/N ratio changes in the permanent plot and chronosequence plot series The data for the permanent plot in Table 3 show that SOC of 0–80 cm soil increased by 2.03 kg C m2 during the 25 years at an average rate of 81.2 g C m2 yr1. The change is mainly in the surface layers (94.8 g C m2 yr1), whereas in the deep layers below 40 cm there were decreases in SOC of 19.4 g C m2 yr1. In the chronosequence plot series, the top 20 cm accumulated SOM at

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S 65.8 g C m2 yr1 and the overall 0–80 cm layer at 42.2 g C m2 yr1. A 25-year soil nitrogen depletion of about 104 g N m2 was observed (annual depletion rate 4.14 g N m2 yr1) in the permanent plot. Slight nitrogen accumulation rate was observed in 0–20 cm layer at 0.47 g N m2 yr1, while significant nitrogen depletion was observed in deeper layer, i.e., 1.94, 2.13, and 0.56 g N m2 yr1 for 20–40, 40–60, and 60–80 cm layers, respectively. In the chronosequence analysis, no significant changes in 0–40 cm soil layers were observed, while annual nitrogen depletion rate for 40–60 and 60–80 cm soil layers were 2.43 and 1.87 g N m2 yr1, respectively (Table 3). The overall 0–80 cm nitrogen depletion rates ranged from 4.14 to 4.3 g N m2 yr1 for both methods (Table 3). As shown in the permanent plot, decreases in pH, varying from 0.26 to 0.75, were observed in differing soil layers. The annual rate of decrease of pH (obtained by dividing by 25) ranged from 0.011 to 0.030 per year, with an average of 0.022 per year. The chronosequence data also showed a tendency for the pH to decrease, at rates varying from 0 to 0.0111 per year, with an average of 0.0071 per year (Table 3). The bulk density of the surface soil (0–40 cm) in the permanent plot decreased at a rate of 0.80– 8.0 mg cm3 yr1, whereas at depth 40–80 cm it increased at a rate of 2.40–3.20 mg cm3 yr1. In the chronosequence, the bulk density of the surface soil (0–20 cm) tended to decrease at 6.7 mg cm3 yr1 and no significant changes were found in deeper layers (Table 3). The permanent plot and the chronosequence both showed the same tendency for the C/N ratio to increase with the growth of larch trees (Table 3). In the chronosequence significant increases in the top 20 cm layers and in the total 0–80 cm soil profile were found at rate of 0.2825 and 0.1774 mg cm3 yr1. In the permanent plot, the C/N ratio increased at a rate varying from 0.27 to 0.81 mg cm3 yr1, with an average of 0.46 mg cm3 yr1 (Table 3).

Discussion

Crosschecking of chronosequence method and permanent plot method The chronosequence method used in long-term studies is well established, for example in studies of nutrient cycling (Vitousek et al., 1995), restoration (Aide et al., 2000), carbon sequestration processes (Grunzweig et al., 2004) and productivity and carbon flux (Litvak et al., 2003). There have nevertheless been criticisms of this methodology (Johnson & Miyanishi, 2008), and crosschecking of it against the permanent plot methodology is worthwhile. In this study, a long-term perma-

13

nent plot (1983–2008) and a chronosequence plot series (consisting of 57 plots) were chosen in the same region, to facilitate crosschecking and soil type is lessive soil originated from forests (Table 3, supporting information, Appendix S1). The chronosequence and permanent plot show similar trends in SOM and soil nitrogen, soil pH and soil bulk density changes with larch plantation development. Both methods, for instance, indicate the accumulation of SOC, depletion of soil nitrogen, decrease in soil pH and soil bulk density and increase C/N ratio (Table 3). Moreover, the magnitudes of these changes are in fairly similar ranges. For example, the rates of accumulation of SOM in the top 20 cm were 94.8 and 65.8 g m2 yr1; also similar were the nitrogen depletion rate in the overall 0–80 cm soil profile (4.14 vs. 4.3 g m2 yr1), and the rates of increase of the C/N ratio in the overall 0–80 cm soil profile were, respectively, 0.46 and 0.18 per year. The vertical distribution of the changes in the soil profile is somewhat different in the two methodologies, however. For the permanent plot the C/N ratio and bulk density changes were observed in the different layers of the 0–80 cm soil, whereas the chronosequence method finds significant changes only in the overall 0–80 cm soil profile or 0–20 cm surface soil. In the permanent plot it was possible to discern the tendency of the pH to decrease in the overall soil profile, but the chronosequence plot series failed to find significant changes in the top 20 cm soil layer (Table 3); similar result also can be seen in C/ N ratio (Table 3). This crosscheck shows that proper use of the chronosequence plot series method is capable of reasonably accurate estimation of soil carbon, nitrogen and pH-related dynamics, especially given the large dataset of 57 plots used in this study.

Larch plantation development led to obvious SOC accumulation The importance of soil carbon storage has attracted the attention of many scientists (Hansen, 1993; Wang et al., 2006; Luyssaert et al., 2008; Springsteen et al., 2010). From the gradient of the linear relation between SOC and larch age, larch plantations in northeastern China can accumulate SOC at a rate ranging from 57.9 to 139.4 g m2 yr1 (on average 96.4 g m2 yr1) in the top 20 cm of soil and power analysis indicates that our data are powerful (72–100% chance) to detecting such significant SOC accumulation, whereas deep layers are generally affected only slightly by the growth of larch trees, as shown by the nonsignificant changes in SOM with larch age as well as low power for detecting such difference (Fig. 2, Tables 2 and 3). This is because the larch has a shallow root system, often not more than 50 cm (Wang et al., 2005b). Long-term micrometeorolo-

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

14 W . W E N - J I E et al. gical studies at the Laoshan site have found net ecosystem exchange (NEE) and net primary productivity (NPP) in the larch plantation of about 264 g m2 yr1 (Qiu et al., 2011) and 486 (SD 5 103) g m2 yr1 (Wang et al., 2005a). The belowground soil carbon accumulation consequently corresponds to 36% of the ecosystem carbon sink capacity, and 20% of NPP. Many scientists have calculated the carbon sequestration rate of soil during forest development or vegetation succession, and comparison of results will improve our understanding of soil carbon process (Table 4). Three distinct conclusions exist. One group of scientists found that soil could deposit a large amount of organic carbon, with a fairly high SOM accumulation rate during forest development (Covington, 1981; Hansen, 1993; Garten, 2002; Zhou et al., 2006; Luyssaert et al., 2008; Springsteen et al., 2010 and others in Table 4). A total of 17 data from references in this class show that this SOM accumulation rate ranges from 11 to 238 g C m2 yr1, averaging approximately 66.2 (SD 5 46.5) g C m2 yr1. A second group of scientists found that, although the SOC pool is quite large, the SOM accumulation rate is fairly small (Richter et al., 1999; Wirth et al., 2002; Schedlbauer & Kavanagh, 2008) (see Table 4). By pooling 15 previous studies in Table 4, the SOM accumulation rate during forest development ranges from 0 to 7.6 g C m2 yr1 and averages 4.6 (SD 5 2.3) g C m2 yr1. This is just 7% of the mean value of the high SOM accumulation rate given in previous references (66.2 g C m2 yr1; SD 5 46.5) (see Table 4). Some scientists also assert that, although forest litter can return some biomass carbon, forest soil might act as a source of atmospheric CO2, since heterotrophic respiration in soil could efflux large amounts of stored SOM to the atmosphere (Berthrong et al., 2009). The mean SOC depletion rate is 3.5 (SD 5 5.6) g C m2 yr1 (Table 4). Our finding supports the first group of scientists and shows that larch plantations sequester carbon into soil (0– 20 cm) at a rate ranging from 57.9 to 139.4 g C m2 yr1 (Fig. 2 and Tables 2 and 3). As a result, underground soil carbon processes in the black soil region of northeastern China should be re-examined in studying the forest ecosystem carbon balance. Correction to constant mass in needed when changes in soil bulk density occur after land-use changes (Davidson & Ackerman, 1993; Henderson et al., 2004; Pin˜eiro et al., 2009) and the general decreasing tendency in soil bulk density after returning farmland to forests (Fig. 5 and Table 4) were included in the calculation of this paper. This correction to constant mass enhances the SOC increase during larch forest development. When without correction, the SOC accumulation rate at surface soil (0–20 cm) ranges from 32.2 to 81.8 g C m2 yr1 (on average 57.2 g m2 yr1) (data not shown here), while such rate is from 57.9 to

139.4 g C m2 yr1 with an average 96.4 g m2 yr1 (Table 4). In the calculation process for constant mass correction, more soil (e.g. 1–2 cm in depth) at aged larch plantations has to be included for offsetting influences from bulk density decreases, which made the aged plantation have more SOC or SON stocks comparing with those of noncorrection of constant mass. It is just like a reverse process of soil reclamation or erosion (Davidson & Ackerman 1993; Henderson et al., 2004; Pin˜eiro et al., 2009). In their cases, much shallower soil depth should be considered because soil density is increased, thus remove the decreasing tendency in SOC. In our study, larger soil depth should be considered, thus strongly enhancement in the increase rate of SOC accumulation has been observed. For the functioning of a soil ecosystem, the turnover of SOC is probably more significant than the sizes of SOC stocks (Six & Jastrow, 2006). An understanding of SOC turnover is crucial for determining the quantitative and temporal responses of local, regional, or global carbon and nutrient budgets to perturbations caused by human activities or climate changes (George et al., 2010). In this study, the two parameters, slope and intercept of the significant linear relations (Table 2), are indicators for annual changes of SOC as well as steady state SOC storage before larch afforestation in theory. Mean residence time of SOC could be approximated as the ratio between steady state SOC storage (intercept) and annual SOC changes (slope) (Six & Jastrow, 2006; Yu et al., 2008). Calculated from Table 2, mean residence time of SOC at surface 20 cm layers ranged from 36.5 years at Maoershan sites to 117.8 years at Dongshan sites (averaged of four sites 67.9 years). Deeper soils generally had a longer residence time, e.g., the SOC residence time for Laoshan 60–80 cm (80.9 years) and Maoershan 20–40 cm (57.8 years) were over 1.5 times higher than those in 0–20 cm soil. Different fractions of SOC (labile carbon, mineral-occluded carbon, etc.) have different turnover rates (Six & Jastrow, 2006; George et al., 2010), however, our study cannot give a detail on their turnover characteristics. More attention to the different fraction of SOC will be highlighted in the future studies.

Can the soil physico-chemical properties support longterm larch growth? Following afforestation by fast-growing species such as larch, soil fertility may be reduced as a result of the large absorption of various nutrients into the biomass; less returns to the soil (Berthrong et al., 2009), and continuous cropping of larch plantations may degrade the soil physical properties necessary for good water transportation and nutrient storage (Lu et al., 1999), as

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S Table 4

Soil organic carbon (SOC) accumulation rate in larch plantation compared with other studies

No.

Forest type

Present study 1 Larch plantation at Laoshan 2 Larch plantation at Laoshan 3 Larch plantation at Maoershan 4 5

6

Larch plantation at Dongshan Larch plantation at Daqingchuan Average

SOC deposition rate (g C m2 yr1)

94.8 for 0–20 cm 81.2 for 0–80 cm 65.8 for 0–20 cm 23.6 for 20–40 cm 139.4for 0–20 cm 69.6 for 20–40 cm 227.3 for 0–80 cm 124.2 for 0–20 cm 154.8 for 0–80 cm 57.9 for 0–20 cm

Oak forest

4

Hybrid poplar forest

90.24 163  16

30% boreal and 70% 130  80 temperate forests (519 plots) 6 Conifer forest 26 7 Deciduous forest 35 8 Mixed oak forest 52 9 Conifer forest 38.5 10 Old growth forest 61 11 Mixed tropical 28.3 12 Mixed tropical 11.8 13 Mixed tropical 60.6 14 Larix olgensis 38.5 plantation 15 Mixed conifer 57.2 16 Scots pine plantation 78–238 17 Grassland to 18 woody shrubs 18 Average 66.2(SD 5 44.5) Low or near zero SOM accumulation rate in references 1 Pinus taeda Pine-forest 4.1 2 Tropical secondary forest 0 5

3 4 5 6 7 8

Pinus sylvestris, Vacc. type Pseudotsuga menziesii Mixed conifer Picea abies Fagus sylvatica Nothofagus pumilo

Location

References

45116 0 N, 127134 0 E

Long term permanent plot in this study Laoshan chronosequence plot series in this study Maoershan chronosequence plot series in this study

45116 0 N, 127134 0 E 45124 0 N, 127133 0 E

46157 0 N, 129110 0 E 47100 0 N, 129107 0 E

Dongshan chronosequence plot series in this study Daqingchuan chronosequence plot series in this study

35116 0 N, 88129 0 W

Garten (2002)

White mountain, New Hampshire San Dimas Experimental Forest, Los Angeles North Central Forest Experiment Station, USDA forest service World mean value

Covington (1981)

96.4(SD 5 35.5) for 0–20 cm

High SOM accumulation rate in references 1 The pine and 40–170 hardwood forest 2 Hardwoods 56 3

15

0.6 7.6 4.4 7.6 4.5 1.8

Ulery et al. (1995) Hansen (1993)

Luyssaert et al. (2008)

42100 0 N, 85158 0 W 42100 0 N, 85158 0 W Rhode Island 29134 0 N, 94133 0 E 23109 0 N, 112133 0 E Kompiai, New Zealand Porce, Colombia Puerto Rico, USA 127128 0 –129117 0 E, 42142 0 – 44130 0 W California, USA 27147 0 E, 59119 0 N 46146 0 N, 100155 0 W

Morris et al. (2007) Morris et al. (2007) Hooker & Compton (2003) Tian et al. (2009) Zhou et al. (2006) Street (1982) Sierra et al. (2007) Brown & Lugo (1990) Wang et al. (2006)

34.51N, 821W 10142 0 N, 84110 0 W

Richter et al. (1999) Schedlbauer & Kavanagh (2008) Wirth et al. (2002) Entry & Emmingham (1998) Black & Harden (1995) Gleixner et al. (2009) Gleixner et al. (2009) Gleixner et al. (2009)

Central Siberia Oregon, USA California, USA Fichtelgebirge, Germany Leinefelde, Germany Patagonia, Chile

Sollins et al. (1983) Reintam et al. (2002) Springsteen et al. (2010)

Continued

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

16 W . W E N - J I E et al. Table 4. (Contd.)

No. 9 10 11 12 13 14

Forest type

SOC deposition rate (g C m2 yr1)

Mixed deciduous Nothofagus solandri Mixed tropical Mixed tropical Scots pine plantation Deciduous forest

6.2 8.2 2.4 2.7 3.3–5.1 3.6

15 Average 4.6 (SD 5 2.3) SOM depletion (negative SOM accumulation) in references 1 Pinus sylvestris, lichen type 0.8 2 Pseudotsuga menziesii 14.2 3 Tsuga heterophylla 0.1 4 Pinus ponderosa 8.5 5 Fagus sylvatica 0.6 6 Fagus sylvatica 0.4 7 Mixed tropical 0.1 8 Average 3.5 (SD 5 5.6)

Location

References

Pennsylvania, USA New Zealand Para, Brazil Sarapiqui, Costa Rica 59115 0 –59117 0 N, 27140 0 –27146 0 E (45148 0 N, 90105 0 W) (46115 0 N, 89121 0 W).

Hoover et al. (2002) Davis et al. (2003) Smith et al. (1998) Guariguata et al. (1997) Karu et al. (2009) Tang et al. (2009)

Central Siberia Washington, USA Oregon, USA Oregon, USA Leinefelde, Germany Hainich, Germany Paragominas, Brazil

Wirth et al. (2002) Klopatek (2002) Boone et al. (1988) Law et al. (2003) Gleixner et al. (2009) Gleixner et al. (2009) De Camargo et al. (1999)

SOM, soil organic matter.

well as soil acidity which regulates nutrient availability (Chen & Xiao, 2006; Berthrong et al., 2009). These studies provide valuable information on changes in soil fertility and physical properties during forest development, but seldom consider the status of soil before the forest is established. Some larch plantations are afforested from farmland, in which the soil is usually heavily degraded in its physical and chemical soil properties (Ding & Liu, 1980; Xin et al., 2002; Song et al., 2009). If the SOC increases with larch development (as in the present study) then this might indicate high soil fertility (Tiessen et al., 1994), although the low temperatures in this region may hinder the return of nitrogen and other elements to the status necessary for direct absorption by plants (HLJTR, 1992). Direct analysis of soil physicochemical properties is needed. In the lessive soil, the permanent plot underwent obvious decreases during larch forest development in soil nitrogen and pH, but significant increases in the C/N ratio. These findings are verified by the chronosequence plot series at the same site (Tables 2 and 3 and Figs 3–6). Moreover, the significant increasing C/N ratio at the Laoshan site may be due to the increasing SOC but decreasing nitrogen content (Table 3), indicating that the accumulated carbon in soil is less readily broken down when there is a shortage of nitrogen (Gong et al., 2007). In the typical dark-brown forest soil, however, three chronosequence plot series (Maoershan, Dongshan and Daqingchuan) yielded no significant changes in soil nitrogen or soil pH value; most of the C/N ratio–age relations (12 out of 15) show no signifi-

cant changes and the statistical power of the data for quantifying such significant correlations is quite low even for the significant relations (power o0.66) (Fig. 6). As shown in the soil profile picture in the Appendix S1, the soil at the Laoshan site is a kind of lessive (Baijiang) soil originated from lowland forests (Chen, 1986), whereas the other three soils are typical darkbrown forest soil from local climax forests (NEFU, 1984; Gong et al., 2007). Lessive soil type has only been studied at Laoshan sites (permanent plot and one chronosequence plot series), thus there is potential confounding effects of site by soil type and microclimate. The soil texture and clay mineralogy of lessive soil are significantly different from those of dark-brown forest soil (HLJTR, 1992). These differences could exert a strong influence on the soil fertility and crop production, and be responsible for the dynamics of SOC, nitrogen and soil pH by absorbing nutrients into biomass and returning litter and roots back to the soil system (Richter et al., 1999; Gong et al., 2007). Our findings indicate that, for larch plantations in lessive soil, the soil should be careful tended to prevent reduction in its fertility, while there is no cause for concern in the dark-brown forest soil in this region, at least for 50 years. Proper return of harvested residues and introduction of nitrogen-fixing species may be necessary in lessive soil plantations to maintain soil fertility for aboveground biomass production and belowground carbon accumulation (Sun et al., 2009). Soil bulk density is an important physical parameter for soil nutrient storage, water transportation and gas

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S

17

Fig. 7 Comparison of agricultural- and Returning Farmland To Forests (RFTF)-induced changes: soil organic carbon (SOC) changes (a), soil nitrogen changes (b) and soil bulk density changes (c). Data for agricultural-induced soil organic matter (SOM) depletion are from the Forest soil group of CAS (1980), Xin et al. (2002), Qiu et al. (2005) and Song et al. (2005). Data for agricultural-induced soil nitrogen depletion are from the Forest Soil group of CAS (1980), Ding & Liu (1980), Wang et al. (2002) and Xin et al. (2002). Data for agricultural induced soil bulk density changes are from Ding & Liu (1980), Lu et al. (2005) and Song et al. (2009). The data for the RFTF program is from the present study.

penetration (Wang et al., 2010). The bulk density of the surface soil decreased with larch plantation development, and all four chronosequence plot series showed similar patterns (Fig. 5 and Table 2), indicating that the growth of larch trees caused the soil to be lighter and more porous. The soil bulk density is strongly correlated with SOC (Zhou et al., 2006; Xie et al., 2007); with the growth of larch forest, most returning litter was deposited at the surface of the soil. SOC therefore accumulates at the surface (Fig. 2) and it is reasonable to suppose that the bulk density will decrease, as shown in Table 2 and Fig. 5. This reduction in bulk density may be favorable for the growth of larch plantations in both lessive soil and typical dark-brown forest soil in this region.

Implications for the RFTF program implemented in China and a global view The timescale needed for the RFTF program is important information in governmental policy decisions

(Li, 2004; Zu et al., 2009). Comparison between the processes of agriculture and RFTF may help. Agricultural practice usually causes large decreases in SOC and soil nitrogen storage, and in Northeastern China it generally increases the soil bulk density and degrades its physical quality (Ding & Liu, 1980; Lu et al., 1999; Xin et al., 2002; Song et al., 2009). After the land has been returned to forest, can the original soil physico-chemical properties be recovered? For ease of comparison of the degradation rates with the rates of recovery in our study, the depletion rate of SOC storage during this forestland to farmland process are calculated to range from 170 to 473 g m2 yr1, with an average of 321 g m2 yr1 (SD 5 98.9) in the top soil layer; the value for soil nitrogen ranged from 1.2 to 70 mg m2 yr1 (averaging 20.3 mg m2 yr1; SD 5 25.0) (see Fig. 7) (recalculation is based on a soil bulk density of 1.2 g cm3 from Xie et al., 2007). The rate of SOC accumulation during the returning of farmland to forest was only 30% of the depletion rate during agricultural

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

18 W . W E N - J I E et al. use, showing that RFTF can recover SOC level, but the time period needed is much longer than the agricultural duration. Nitrogen depletion has obviously decreased from 20.3 mg m2 yr1 during agricultural use, while slight accumulation rate at 3.4 mg m2 yr1 (SD 5 6.4) in RFTF practice was observed (Fig. 7), revealing that the RFTF practice has greatly slowed the rate of nitrogen loss. The rate of decrease of bulk density during RFTF is upto 5.7 mg cm3 yr1, which is 69% of the rate of increase during agricultural use (8.3 mg cm3 yr1) (Fig. 7). Large-scale transformation of original forests, grassland and wetland to farmland has been going on for 60 years, since the 1950s (Zheng, 2007). The rehabilitation time for soil will be much longer than the duration of agricultural use. This time period (ca. 30 years from governmental plan) (Li, 2004) is too short; more time is needed for return to the original carbon sink capacity as well as the regaining of soil fertility. RFTF practices, as one of China’s environmental initiatives (Li, 2004), is the reverse of the process of claiming farmland from clear-cutting of original forests. Until the end of 2009, a total of 28 million hectares of former farmland has sequestrated 1.5 billion cubic meters of biomass, which is about 4.05  1014 g carbon, or more than half of China’s total carbon evolution by industry in 2004, which was 8.08  1014 g (Bai et al., 2006). Biomass production is vital, but little is known about the dynamics of soil carbon (Li, 2004). At present, over 4.5 million hectare land have been afforested by larch plantation, which is almost 70% of total plantation in this region (Sun et al., 2009). Based on our SOC accumulation rate in this study and referenced biomass data, an approximate calculation (area  biomass carbon is total carbon accumulated by biomass, while area  SOC accumulation rate  time period is total carbon accumulated in soil) was done in this study. The soil could sequestrate 52–125 Tg (average 87 Tg) during 20 years growth of these plantations, while biomass carbon sequestration in the same region could be upto 293–383 Tg (average 338 Tg) in the same period of time (high, low and mean biomass data, respectively, 130, 170 and 150.2 ton ha1 were from Wang et al., 2005a; Sun et al., 2007). Carbon sequestration into soil (87 Tg) could be upto 26% of biomass carbon sequestration (338 Tg) during the same growth period. Through various modeling methods, Piao et al. (2009) found 42% of whole national sink capacity (0.177 Pg C yr1) lies in SOM and our estimation of 26% should be a data supporting for their modeling. In the case of China, ignoring the SOC changes during the RFTF process would therefore lead to a great underestimate of the carbon sink function of larch plantations in this region. At the global scale, the planted forest area has steadily increased in all regions of world since 1990. The

world planted forest area increased by more than 3.6 million hectares per year from 1990 to 2000, by 5.6 million hectares per year from 2000 to 2005, and by 4.2 million hectares per year from 2005 to 2010. Given this trend, a further rise in the planted forest area upto 300 million hectares by 2020 can be anticipated (FAO, 2010). East Asia, Europe and North America reported the greatest area of planted forests, together accounting for about 75% of global planted forest area and in East Asia planted forests make up 35% of the total forest area; most of these are found in China (FAO, 2010). Within this huge amount of forest plantation, many of them are planted in degraded or abandoned farmland as well as are used as agricultural protection forests (FAO, 2010). Their soil carbon process may be similar to the process of RFTF (just like this study). Aboveground biomass carbon accumulation is obvious and evident (Wang et al., 2005a; Sun et al., 2007), and our study may give some important hints for the underground soil carbon dynamics, which is very scarce but with large data variation in today carbon balance studies (as shown in Table 4).

Conclusion By analyzing data from one permanent plot and four chronosequence plot series consisting of 159 plots in northeastern China, we observed significant SOC accumulation (96.4 g C m2 yr1) and soil bulk density decrease (5.7 mg cm3 yr1) in the surface layer (0–20 cm) of soil, but no clear changes in soil pH or soil nitrogen storage during the development of larch plantations. These results indicate that the RFTF practice of setting up larch plantations can increase SOC storage; long-term, lessive soil may need intervention to improve its fertility, but normal dark-brown forest soil should not. The 4.5 million hectares of larch planted under RFTF can sequestrate approximately 87 Tg carbon in the soil, which is 26% of that stored in biomass; these data show the importance of soil carbon in the total ecosystem carbon budget. Comparison of agriculture-induced changes in SOC, soil nitrogen and soil bulk density with the RFTF process shows that the latter is much slower, so that longer times than the duration of agricultural use are necessary to rehabilitate the underground carbon storage facility and soil fertility.

Acknowledgements This study was supported financially by China’s Ministry of Science and Technology (2011CB403205), China’s National Foundation of Natural Sciences (40873063), China’s postdoctoral foundation (201003406 & 20080430126) and basic research fund for national universities from Ministry of Education of China (DL09CA17). Thanks are due to subject editor, Pete Smith and three anonymous reviewers, and their comments have greatly

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

S O I L O R G A N I C C A R B O N I N L A R C H P L A N TAT I O N S improved the scientific significance of this paper. W. W. J. appreciates the kind help from Dr Wei Xiaorong (Institute of Soil and Water Conservation, Chinese Academy of Sciences, China), Dr. Hou Jihua (Beijing Forestry University, China) and Prof Peter B. Reich (Department of Forest Resources, University of Minnesota) during the major revision of the original manuscript.

19

Gower ST, Richards JH (1990) Larches: deciduous conifers in an evergreen world – in their harsh environments, these unique conifers support a net carbon gain similar to evergreens. Bioscience, 40, 818–826. Grunzweig JM, Sparrow SD, Yakir D, Chapin FS (2004) Impact of agricultural land-use change on carbon storage in boreal Alaska. Global Change Biology, 10, 452–472. Guariguata MR, Chazdon RL, Denslow JS, Dupuy JM, Anderson L (1997) Structure and floristic of secondary and old-growth forest stands in low land Costa Rica. Plant Ecology, 132, 107–120. Hansen EA (1993) Soil carbon sequestration beneath hybrid poplar plantations in the

References Aide TM, Zimerman JK, Pascarella JB, Rivera LW, Humfredo MV (2000) Forest regeneration in a chronosequence of tropical abandoned pastures: implications for restoration ecology. Restoration Ecology, 8, 328–338. Bai B, Li XC, Liu TF et al. (2006) Preliminary study on CO2 industrial point sources and their distribution in China. Chinese Journal of Rock Mechanics and Engineering, 25, 2918–2923. Bao SD (2000) The Method of the Soil and Agriculture Chemical Analysis. China Agriculture Press, Beijing. Berthrong ST, Jobba´gy E, Jackson RB (2009) A global meta-analysis of soil exchangeable cations, pH, carbon, and nitrogen with afforestation. Ecological Applications, 19, 2228–2241. Black TA, Harden JW (1995) Effect of timber harvest on soil carbon storage at Blodgettexperimental-forest, California. Canadian Journal of Forest Research – Rev Can Rech For, 25, 1385–1396. Boone RD, Sollins P, Cromack K (1988) Stand and soil changes along a mountain hemlock death and regrowth sequence. Ecology, 69, 714–722. Brown S, Lugo AE (1990) Effects of forest clearing and succession on the carbon and nitrogen-content of soils in Puerto-Rico and US Virgin Islands. Plant and Soil, 124, 53–64. Chen LX, Xiao Y (2006) Evolution and evaluation of soil fertility in forest land in Larix gmelinii plantations at different development stages in Daxinganling forest region. Science of Soil and Water Conservation, 4, 50–55. Chen XQ (1986) The effect of community structure of artificial larch wood on physicochemical properties of Baijiang soil. Journal of Northeast Forestry University, 14, 113–116. Ce´spedes-Payret C, Pin˜eiro G, Achkar M, Gutie´rrez O, Panario D (2009) The irruption of new agro-industrial technologies in Uruguay and their environmental impacts on soil, water supply and biodiversity: a review. International Journal of Environment and Health, 3, 175–197. Covington W (1981) Changes in forest floor organic matter and nutrient content following clear cutting in northern hardwoods. Ecology, 62, 41–48. Davidson EA, Ackerman IL (1993) Changes in soil carbon inventories following cultivation of previously untilled soils. Biogeochemistry, 20, 161–193. Davis MR, Allen RB, Clinton PW (2003) Carbon storage along a stand development sequence in a New Zealand Nothofagus forest. Forest Ecology and Management, 177, 313–321. De Camargo PB, Trumbore SE, Martinelli LA, Davidson EA, Nepstad DC, Victoria RL (1999) Soil carbon dynamics in re-growing forest of eastern Amazonia. Global Change Biology, 15, 693–702. Ding RX, Liu ST (1980) A study on the fertility of black soil after reclamation. Acta Pedologica Sinica, 17, 20–32. Enoki T, Kawaguchi H, Iwatsubo G (1996) Topographic variations of soil properties and stand structure in a Pinus thunbergii plantation. Ecological Research, 11, 299–309. Entry JA, Emmingham WH (1998) Influence of forest age on forms of carbon in Douglas-fir soils in the Oregon Coast Range. Canadian Journal of Forest Research–Rev Can Rech For, 28, 390–395. FAO (2010) Global Forest Resources Assessment 2010 Main Report. FAO, Rome, pp. 90–95. Forest Soil Group of CAS (1980) Forest Soil in Northeast China. Science Press, Beijing. Garten CT Jr (2002) Soil carbon storage beneath recently established tree plantations in Tennessee and South Carolina, USA. Biomass and Bioenergy, 23, 93–102. Garten CT Jr, Huston MA, Thoms CA (1994) Topographic variation of soil nitrogen dynamics at Walker Branch watershed, Tennessee. Forest Science, 40, 497–512. George SJ, Kelly RN, Greenwood PF, Tibbett M (2010) Soil carbon and litter development along a reconstructed biodiverse forest chronosequence of South-Western Australia. Biogeochemistry, 101, 197–209. Gleixner GD, Tefs C, Jordan A et al. (2009) Soil carbon accumulation in old-growth forests. In: Old-Growth Forests: Function, Fate and Value (eds Wirth C, Gleixner G, Heimann M), pp. 231–266. Springer-Verlag, Berlin. Gong ZT, Zhang GL, Chen ZC et al. (2007) Pedogenesis and Soil Taxonomy. Science Press, Beijing.

North Central United States. Biomass and Bioenergy, 5, 431–436. Henderson D, Ellert B, Naeth MA (2004) Grazing and soil carbon along a gradient of Alberta rangeland. Journal of Range Management, 57, 402–410. HLJTR (1992) Soil of Heilongjiang Province, P.R. China. Science and Technology press of Heilongjiang Province , Harbin, P.R. China. HOMWR (Hydrographic office of Ministry of Water Resource, P.R. China) (2002) Storm Flood of Songhua River in 1998. China Waterpower Press, Beijing. Hooker TD, Compton JE (2003) Forest ecosystem carbon and nitrogen accumulation during the first century after agricultural abandonment. Ecological Applications, 13, 299–313. Hoover CM, Magrini KA, Evans RJ (2002) Soil carbon content and character in an oldgrowth forest in northwestern Pennsylvania: a case study introducing pyrolysis molecular beam mass spectrometry (py-MBMS). Environmental Pollution, 116, 269–275. Hou YK, Duan SG, Zhao S (2004) Main Tree Species for China Returning Farmland to Forests (Northern China Part). China Forestry Publishing House, Beijing. Houghton RA, Hobbie JE, Melillo JM, Moore B, Peterson BJ, Shaver GR, Woodwell GM (1983) Changes in the carbon content of the terrestrial biota and soils between 1860 and 1980: a net release of CO2 to the atmosphere. Ecological Monographs, 53, 235–262. Jiang H, Michael AJ, Peng CH, Zhang YL, Liu JX (2002) Modeling the influence of harvesting on Chinese boreal forest carbon dynamics. Forest Ecology and Management, 169, 65–82. Johnson EA, Miyanishi K (2008) Testing the assumptions of chronosequences in succession. Ecology Letters, 11, 419–431. Karu H, Szava-Kovats R, Pensa M, Kull O (2009) Carbon sequestration in a chronosequence of Scots pine stands in a reclaimed opencast oil shale mine. Canadian Journal of Forest Research, 39, 1507–1517. Klopatek JM (2002) Below ground carbon pools and processes in different age stands of Douglas-fir. Tree Physiology, 122, 197–204. Kravchenko AN, Robertson GP (2011) Whole-profile soil carbon stocks: the danger of assuming too much from analyses of too little. Soil Science Society of America Journal, 75, 235–240. Law BE, Sun O, Campbell J, Van Tuyl S, Thornton P 2003 Changes in carbon storage and fluxes in a chronosequence of ponderosa pine. Global Change Biology, 9, 510–524. Li SD (2004) Research on Conversion of Farmland to Forests in China. Science Press, Beijing. Litvak M, Miller S, Wofsy SC, Goulden M (2003) Effect of stand age on whole ecosystem CO2 exchange in the Canadian boreal forest. Journal of Geophysical Research, 108, 8225–8235. Lu CY, Chen X, Shi Y, Zheng J, Zhou QL (2005) Study on the change characteristics of black soil quality in northeast China. System Sciences and Comprehensive Studies in Agriculture, 21, 182–189. Lu XJ, Gao H, Xu SW, Zheng KX (1999) Studies on the influences of continuous planting of larch on soil physical properties and tree growth. Journal of Liaoning Forestry Science and Technology, 5, 10–12. Luyssaert SE, Schulze ED, Bo¨rner A et al. (2008) Old-growth forests as global carbon sinks. Nature, 455, 213–215. Mann LK (1986) Changes in soil carbon storage after cultivation. Soil Science, 142, 279–288. Morris SJ, Bohm S, Haile-Mariam S, Paul EA (2007) Evaluation of carbon accrual in afforested agricultural soils. Global Change Biology, 13, 1145–1156. NEFU (1984) Basic Data for Maoershan Experimental Forest farm of northeast Forestry University. NEFU Press, Harbin, P.R. China. Paul KI, Polglase PJ, Nyakuengama JG, Khanna PK (2002) Change in soil carbon following afforestation. Forest Ecology and Management, 168, 241–257. Piao SL, Fang JY, Ciais P, Peylin P, Huang Y, Sitch S, Wang T (2009) The Carbon balance of terrestrial ecosystems in China. Nature, 458, 1009–1013. Pimentel D, Harvey C, Resosudarmo P et al. (1995) Environmental and economic costs of soil erosion and conservation benefits. Science, 267, 1117–1123. Pin˜eiro G, Paruelo JM, Jobbagy EG, Jackson RB, Oesterheld M (2009) Grazing effects on belowground C and N stocks along a network of cattle exclosures in temperate and subtropical grasslands of South America. Global Biogeochemical Cycles, 23, GB2003, doi: 2010.1029/2007GB003168.

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x

20 W . W E N - J I E et al. Post WM, Emanuel WR, Zinke PJ et al. (1982) Soil carbon pools and world life zones. Nature, 298, 156–159.

Vitousek PM, Turner DR, Kitayama K (1995) Foliar nutrients during long-term soil development in Hawaiian montane rain forest. Ecology, 76, 712–720.

Qiu JJ, Wang LG, Tang HJ, Li H, Li CS (2005) Studies on the situation of soil organic carbon storage in croplands in Northeast of China. Agricultural Sciences in China, 4, 101–105. Qiu L, Zu YG, Wang WJ, Sun W, Su DX, Zheng GY (2011) CO2 flux characteristics and their influences on the carbon budget of a larch plantation in Maoershan region of Northeast China. China Journal of Applied Ecology, 22, 1–8. Reintam L, Kaar E, Rooma I (2002) Development of soil organic matter under pine on

Walker LR, Wardle DA, Bardgett RD, Clarkson BD (2010) The use of chronosequences in studies of ecological succession and soil development. Journal of Ecology, 98, 725–736. Wang CM, Ouyang H, Shao B et al. (2006) Soil carbon changes following afforestation with Olga Bay Larch (Larix olgensis Henry) in Northeastern China. Journal of Integrative Plant Biology, 48, 503–512. Wang JK, Wang TY, Zhang XD, Guan LZ, Wang QB, Hu HX, Zhao YC (2002) An approach to the changes of black soil quality (I) – changes of the indices of black soil

quarry detritus of open-cast oil shale mining. Forest Ecology and Management, 17, 191–198. Richter DD, Markewitz D, Trumbore SE, Wells CG (1999) Rapid accumulation and turnover of soil carbon in a re-establishing forest. Nature, 400, 56–58. Schedlbauer JL, Kavanagh KL (2008) Soil carbon dynamics in a chronosequence of secondary forests in northeastern Costa Rica. Forest Ecology and Management, 255,

with the year(s) of reclamation. Journal of Shenyang Agricultural University, 332, 43–47. Wang WJ, He HS, Zu YG et al. (2010) Addition of HPMA affects seed germination, plant growth and properties of heavy saline-alkali soil in northeastern China: comparison with other agents and determination of the mechanism. Plant and Soil, 339, 177–191. Wang WJ, Zu YG, Wang HM et al. (2005a) Plant biomass and productivity of Larix

1326–1335. Schulze ED, Lloyd J, Kelliher FM et al. (1999) Productivity of forests in the Eurosiberia boreal region and their potential to act as a carbon sink – a synthesis. Global Change Biology, 5, 703–722. Sellers PJ, Dickinson RE, Randall DA et al. (1997) Modeling the exchanges of energy, water, and carbon between continents and the atmosphere. Science, 275, 502–509.

gmelinii forest ecosystems in Northeast China: intra- and inter- species comparison. Eurasian Journal of Forest Research, 8, 21–41. Wang WJ, Zu YG, Wang HM, Hirano T, Sasa K, Koike T (2005b) Effect of collar insertion on soil respiration in a larch forest measured with a LI-6400 soil CO2 flux system. Journal of Forest Research, 10, 57–60. Wirth C, Czimczik CJ, Schulze ED (2002) Beyond annual budgets: carbon flux at

Sierra CA, del Valle JI, Orrego SA et al. (2007) Total carbon stocks in a tropical forest landscape of the Porce region, Colombia. Forest Ecology and Management, 243, 299–309. Six J, Jastrow JD (2006) Organic matter turnover. In: Encyclopedia of Soil Science, (Vol. 2 (ed. Lal R), pp. 1210–1215. Taylor & Francis, London. Smith K, Gholz HL, DeAssis Oliveira F (1998) Litter fall and nitrogen-use efficiency of plantations and primary forest in the eastern Brazilian Amazon. Forest Ecology and Management, 10, 209–220.

different temporal scales in fire-prone Siberian Scots pine forests. Tellus Series B: Chemical and Physics Meteorology, 54, 611–630. Wirth C, Gleixner G, Heimann M (2009) Old-Growth Forests: Function, Fate and Value. Springer-Verlag, Berlin. Xie ZB, Zhu JG, Liu G et al. (2007) Soil organic carbon stocks in China and changes from 1980s to 2000s. Global Change Biology, 13, 1989–2007. Xin G, Yan L, Wang JH, Guan LZ (2002) Changes of organic carbon in black soils with

Sollins P, Spycher G, Topik C (1983) Processes of soil organic-matter accretion at a mudflow chronosequence, Mt Shasta, California. Ecology, 64, 1273–1282. Song GH, Li LQ, Pan GX, Zhang Q (2005) Top soil organic carbon storage of China and its loss by cultivation. Biogeochemistry, 74, 47–62. Song R, Liu L, Wu CS, Guo JX (2009) Reclamation on organic matter content and structural properties in steppe soil of northeast Songnen plain. Chinese Journal of

the different reclamation years. Chinese Journal of Soil Science, 33, 332–335. Yu WT, Jiang ZS, Zhou H, Ma Q (2008) Effect of different land use patterns on soil microbial biomass carbon and its turnover rate in an aquic soil. Chinese Journal of Ecology, 27, 1302–1306. Zheng JZ (2007) The Great Northern Wilderness in Northeast China: 60 years. People Publisher of Heilongjiang Province, Harbin, P.R. China.

Grassland, 31, 91–95. Springsteen A, Loya W, Liebig M, Hendrickson J (2010) Soil carbon and nitrogen across a chronosequence of woody plant expansion in North Dakota. Plant and Soil, 328, 369–379. Street JM (1982) Changes of carbon inventories in live biomass and detritus as a result of the practice of shifting agriculture and the conversion of forest to pasture: case studies in Peru, New Guinea and Hawaii. In: Geography in the Third World (ed. Jahi

Zhou GY, Liu SG, Li Z et al. (2006) Old-growth forests can accumulate carbon in soils. Science, 314, 1417. Zu YG, Wang WJ, Cui S, Liu W, Koike T (2009) Soil CO2 efflux, carbon dynamics, and change in thermal conditions from contrasting clear-cut sites during natural restoration and uncut larch forests in northeastern China. Climatic changes, 96, 137–159.

AI), pp. 249–258. Penerbit University, Kebangsaan, Malaysia. Sun YJ, Zhang J, Han AH, Wang XJ (2007) Biomass and carbon pool of Larix gmelinii young and middle age forest in Xing’an Mountains Inner Mongolia. Acta Ecologica Sinca, 27, 1756–1762. Sun ZH, Jin GZ, Mu CC (2009) Study on the Keeping Long-Term Productivity of Larix olgensis Plantation. Science Press, Beijing. Tang JW, Bolstad PV, Martin JG (2009) Soil carbon fluxes and stocks in a Great Lakes forest chronosequence. Global Change Biology, 15, 145–155. Tian YQ, Xu XL, Song MH, Zhou CP, Gao Q, Ouyang H (2009) Carbon sequestration in two alpine soils on the Tibetan Plateau. Journal of Integrative Plant Biology, 51, 900–905. Tiessen H, Cuevas E, Chacon P (1994) The role of soil organic matter in sustaining soil fertility. Nature, 371, 783–785. Tiessen H, Stewart JWB, Bettany JR (1982) Cultivation effects on the amounts and concentration of carbon, nitrogen, and phosphorus in grassland soils. Agronomy Journal, 74, 831–835. Ulery AL, Graham RC, Chadwick OA, Wood HB (1995) Decade-scale changes of soil carbon, nitrogen and exchangeable cations under chaparral and pine. Geoderma, 65, 121–134.

Supporting Information Additional Supporting Information may be found in the online version of this article: Appendix S1. Typical soil profile from different chronosequence plot series. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

r 2011 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2011.02447.x