Effects of land use change from natural forest to plantation on C, N and natural abundance of 13C and 15N along a climate gradient in eastern China

Soil C and N turnover rates and contents are strongly influenced by climates (e.g., mean annual temperature MAT, and mean annual precipitation MAP) as well as human activities. However, the effects of converting natural forests to intensively human-managed plantations on soil carbon (C), nitrogen (N) dynamics across various climatic zones are not well known. In this study, we evaluated C, N pool and natural abundances of δ13C and δ15N in forest floor layer and 1-meter depth mineral soils under natural forests (NF) and plantation forest (PF) at six sites in eastern China. Our results showed that forest floor had higher C contents and lower N contents in PF compared to NF, resulting in high forest floor C/N ratios and a decrease in the quality of organic materials in forest floor under plantations. In general, soil C, N contents and their isotope changed significantly in the forest floor and mineral soil after land use change (LUC). Soil δ13C was significantly enriched in forest floor after LUC while both δ13C and δ15N values were enriched in mineral soils. Linear and non-linear regressions were observed for MAP and MAT in soil C/N ratios and soil δ13C, in their changes with NF conversion to PF while soil δ15N values were positively correlated with MAT. Our findings implied that LUC alters soil C turnover and contents and MAP drive soil δ13C dynamic.

contents and C/N ratio were significantly different among the study sites (P < 0.001, for all except for N content P < 0.01) and their values were considerably affected following conversion of natural forest (NF) to plantation forest (PF) ( Table 1). Furthermore, forest floor C mean content was lower in NF with 399.8 ± 86.1 g kg −1 across six sites as compared to PF with 438.6 ± 70.9 g kg −1 while an opposite trend was observed in forest floor N content with 13.6 ± 1.8 g kg −1 in NF and 10.3 ± 2.6 g kg −1 in PF. We observed an increase of forest floor C varying from 2-17% among sites following the order: DH > JF > MH > QY > HT > XY. Moreover, a general increase of C/N ratio and decrease of soil N in forest floor were observed except at XY site ( Table 2). Soil C ranged from 2.98 g kg −1 to 49.55 g kg −1 whereas soil N was in the range of 0.23 g kg −1 to 4.23 g kg −1 in mineral soil across the sampling sites (Table 2). There was a significant difference (P < 0.01) in soil C, N content and C/N ratio among sites ( Table 3). The statistical analysis showed that soil C and C/N ratios were significantly (P < 0.05) altered following the conversion from NF to PF but soil N did not change significantly (P = 0.214). In addition, we observed a gain of the mean values of soil C content over 1 m by +6% (HT), +39% (XY) and a loss varied between −38% to −12% at DH > MH > JF > QY (Fig. 1a). Soil C/N ratio increased at XY, HT and JF sites ( Table 2). On the other hand, although an increase of soil N contents after NF conversion to PF was observed at XY, there were no obvious differences among sites (Fig. 1b). Our results also indicated a decrease in soil C, N values and C/N ratio with soil depth. The statistical analyses affirmed this trend and revealed that soil depth significantly (P < 0.01) alters C, N values and C/N ratio. However, the conversion from NF to PF strongly influenced C and C/N ratio (Table 3). In addition, a significant and positive correlation was found between C and N (r = 0.879, P < 0.01) ( Table 4).
The natural abundance of 13 C and 15 N in forest floor and mineral soils. Forest floor δ 13 C ranged between −31.58‰ and −27.73‰ in all sites sampled while forest floor δ 15 N ranged between −3.65‰ to 0.20‰ in NF, and between −4.66‰ to 3.66‰ in plantations. In general, forest floor δ 13 C values were significantly lower in NF than PF in XY and HT (Fig. 2) while the change in forest floor δ 15 N was not obvious except for DH and JF ( Fig. 3). There were significant (p < 0.001) differences in forest floor δ 13 C and δ 15 N values among sites. In addition, LUC significantly increased forest floor δ 13 C values (Table 1) but not δ 15 N (P = 0.155).
Soil δ 13 C and δ 15 N values significantly (P < 0.001) varied among sites in mineral soil after the conversion from NF to PF. In addition, soil depth, LU and site × soil depth interaction were significantly altered after the conversion from NF to PF (Table 3). Soil δ 13 C and δ 15 N were enriched with depth after the conversion from NF to PF (Figs 2, 3), especially at the northern (MH and QY) and southern sites (DH and JF) indicating the high decomposition rate of soil organic matter (SOM). However, there were no differences in soil δ 13 C between NF and PF at QY and HT. Furthermore, soil δ 13 C was significantly different between NF and PF among 0-10 cm, 10-20 cm, 20-40 cm, 40-60 cm and 80-100 cm depths at DH and JF. Soil δ 15 N was enriched along soil profiles with the mean values of 2.98‰ in 0-10 cm, 4.29‰ in 10-20 cm and 5.35‰ in 20-40 cm but it decreased between 40-100 cm depth layers.
Relationships of climatic factors to C and N contents, C/N ratios and δ 13 C and δ 15 N. The mean annual temperature (MAT) ranged from −4.14 °C at MH to 21.08 °C at DH, and mean annual precipitation (MAP) increased from 436 mm to 2499 mm from northern to southern China (Table 5). We observed a linear relationship between soil C/N ratio and MAT (Fig. 4a), and a quadratic relationship between soil C/N ratio and MAP in NF (Fig. 4b). Soil δ 13 C were related to MAT and MAP in both NF and PF (Fig. 4c,d), with a linear relationship of soil δ 13 C and MAT in PF (r = 0.22, P = 0.004; Fig. 4c), and a quadratic relationship of soil δ 13 C and MAT in NF (r = 0.29, P = 0.004; Fig. 4c), and there were quadratic relationships of soil δ 13 C and MAP in both NF and PF (Fig. 4d).

Discussion
Effects of forest conversion on carbon and nitrogen contents, and C/N ratios in forest floor and mineral soils. The present study provides a quantitative overview of C and N contents, and natural abundances of δ 13 C and δ 15 N in forest floor layer and mineral soil layers over 1-meter depth under NF and converted plantations across eastern China. Firstly, we found high forest floor N, but low C and C/N ratios in forest floor layer under NF as compared to PF. This point to the better quality of leaf litter materials in NF stands and large release potential of N during decomposition of the litter. This result was consistent with that found by Chen et al. 35 who suggested the lower C/N ratios in forest floor litter in NF compared to hoop pine (Araucaria cunninghamii) plantations in Australia. The increase in forest floor C and C/N values and decrease in forest floor N may indicate that forest management practices provide favorable conditions for decomposing microorganisms thus alter their values as previously reported. In China, forest management practices have been directed towards timber production and sustained yield of wood supply, a practice which has significantly altered the balance between heterotrophic litter decomposition and litter inputs, thus impacting on the C content in the forest floor. This trend is consistent with the results of Vesterdal et al. 36 who reported that the accumulation of nutrients in the forest floor is altered with increasing thinning intensity. There were significant differences in C, N and C/N within the study sites because of the effects of the forest conversion, including tree species that may affect the decomposition rate and the turnover of these elements in the soil 37,38 . Tree species differ in their carbon sequestration potential, LUC by changing species composition, tree density and forest structure altering their sequestration potential. Similar results have been reported by Vesterdal et al. 39 who observed that forest floor C and N contents and C/N ratio were strongly affected by trees species. Besides, Fonseca et al. 37 reported that forest floor C and N under the coniferous species had a large quantity of organic materials poorly decomposed, while a high rate of transformation of forest floor and incorporation in mineral soil have been observed in broad-leaved species. In our study, the forest floor litter was mainly composed of deciduous or evergreen broad-leaf litter in NF, while forest floor in PF is almost purely needle litter. Broad leaves are generally thought to produce mull forest floors that are richer in  Table 1. The two-way ANOVA results for all soil variables in forest floor. n = 6 (Sites), n = 18 (Land use); *, ** and *** indicate a significant level at P < 0.05, P < 0.01 and P < 0.001, respectively.
nutrients and promote rapid decomposition 40 . In addition, the higher forest floor N observed in NF at MH and QY is probably a result of the low temperature in North part of China which negatively affects the decomposition process and consequently leads a larger accumulation of organic matter. Thus, this study revealed that forest floor C and N content is strongly affected by human disturbances, consequently by LUC.
In mineral soil, C values were significantly altered after the conversion from NF to PF among site probably because of the difference of soil type, vegetation and trees species. The vegetation cover influences the storage of its elements in that it reduces the arrival of solar radiation directly to the soil. Yet, litter decomposition rates are controlled by the temperature and moisture which directly affect soil microbial activity. These findings are consistent with those of Jobbágy & Jackson 41 who reported that the variation of soil C with depth in the profile varies strongly with vegetation type. Ramesh et al. 42 also noted that quality and quantity of different soil organic carbon   www.nature.com/scientificreports www.nature.com/scientificreports/ pools change with time depending on the rate of photosynthetic C addition and their losses through decomposition. Our study therefore suggested that LUC alters the carbon-holding capacity of soil in short carbon retention capacity of soil. Furthermore, the changes of soil C and N contents after NF converted to PF were not obvious. We observed a gain in soil C values at XY and HT sites and loss at DH, MH, JF, QY sites after the conversion from NF to PF suggesting that LUC from NF to PF influences C inputs. Similarly, Lewis et al. 43 found that the effects of change from NF converted to introduced Pinus sp. plantations were highly site-specific and might have a positive, negative, or no influence on the variation of soil C values. Smith et al. 34 also demonstrated that conversion from natural Amazonian forest to plantations altered soil organic C with an increase in surface under Euxylophora  Table 3. The two-way ANOVA results for all soil variables over 1 m depth layer. n = 36 (Site, depths), n = 108 (Land use); *, ** and *** indicate a significant level at p < 0.05, p < 0.01 and p < 0.001, respectively. www.nature.com/scientificreports www.nature.com/scientificreports/ paraensis Hub. plantation and decreases under Pinus caribaea var. honduensis Barrett and Golfari. The result could be related to different quantity and quality of C input through root exudation, litter inputs and different management practices between NF and PF 44 . Although biological N fixation is the primary source of nitrogen input 45 , soil N values did not change significantly within the study sites after the conversion from NF to PF in the present study indicating that LUC did not alter significantly the balance between N input and loss.
In general, soil C and N where stored in the 0-20 cm depth segment of the overall profile, i.e. 67 and 57%, respectively. This is consistent with the results found by Batjes et al. 46 who reported 50% the amount of OC  Table 4. Pearson's coefficients correlation between soil variables, land use and depth. * Correlation is significant at the 0.05 level (2-tailed). **Correlation is significant at the 0.01 level (2-tailed). n.s (not significant), n = 216, n = 108 (age), LU (land use), MAP (mean annual precipitation), MAT (mean annual temperature). www.nature.com/scientificreports www.nature.com/scientificreports/ located in the upper 30 cm of the soil organic carbon in the layer 0-100 cm. This result suggests a greater soil C sequestration within 0 to 20 cm soil depth. Furthermore, we found that soil C and N content decreased with soil depth may be due to the vertical distribution of roots. Indeed, Jobbágy et al. 41 , found that root distributions affect the vertical placement of C in the soil, and above-and below-ground allocation affects the relative amount of C that eventually falls to the soil surface from shoots. In addition, it well known that N status is a crucial factor driving forest soil C dynamics and high N availability can promote a greater soil C sequestration 13,47 . This trend is in line with our study which reported a significant and positive correlation between C and N probably through the effects on organic matter decomposition 45 . Effects of forest conversion on natural abundance of 13 C and 15 N in forest floor and mineral soils. Stable isotopic abundances of δ 13 C and δ 15 N have been used as powerful index to evaluate the long-term alterations of C and N cycles 16,17,19,20,22 . In the present study forest floor δ 13 C were enriched after LUC in subtropical and tropical regions while δ 15 N depleted. The depleted natural abundance of 13 C and 15 N in litterfall could account for low values of forest floor δ 13 C and δ 15 N in NF because of the generally higher above-ground litterfall input in NF compared to PF 7,48 . Moreover, we found that soil δ 13 C were strongly enriched at XY and HT sites after LUC. These findings might also be related to the less above-ground litterfall and below-ground roots inputs in these sites. Hertel et al. 49 confirmed this trend by reporting that the conversion of tropical forest into plantations decrease C flux with fine root mortality to soil organic C pool. Moreover, soil organic C and N accumulation and stability have been suggested to be strongly influenced by litter quality, resulting in more stable organic C accumulated in NF soils with high quality of litter substrates 50 . On the other hand, the intensive management practices in PF (e.g., clear-cutting and slash burning, site preparation and pruning) accompanied by higher temperature www.nature.com/scientificreports www.nature.com/scientificreports/ and precipitation conditions in tropical and subtropical (at the DH and JF sites) regions could enrich soil 13 C and 15 N in plantation because of more soil C and N loss 27,44 . However, soil organic C accumulation and stability were also influenced by soil matrix structure 51 , and were related to the saturation of SOC, climatic zones and ecosystem types 52 . It was difficult to explain clearly the effect of LUC on soil δ 15 N variation within the study sites because of the complex interactions between abiotic and biotic factors.
A number of studies observed an increasing tendency of δ 13 C and δ 15 N with soil depth 20,21,23,53 . In our study, we observed an enrichment of soil δ 13 C following the increase of depth layer by 3.09‰ in surface mineral soil (0-10 cm) after LUC probably because deeper soil layer has a greater humification of organic matter    www.nature.com/scientificreports www.nature.com/scientificreports/ decomposition or largely to the increasing residence time of the organic C in the soil 54 . This result was in line with Brunn et al. 20 who found the mean 13 C enrichment between δ 13 C values of the Oi horizon and δ 13 C values of soil organic matter (SOM) in 10 cm soil depth was 3.4 + 0.2‰ under temperate beech forests. Several processes might cause enrichment of δ 13 C with increasing soil depth e.g., the depletion of 13 C in the atmosphere due to combustion of fossil fuel, and the considerable presence of litter and roots with depleted 13 C in the upper soil. Accoe et al. 55 observed that the average rate of change in soil δ 13 C is directly related to organic matter decomposition rates in different parts of the soil profile. Thus, the extent of change in 13 C-abundance with increasing soil depth may indicate the quality or stability of SOM under continuous C3 vegetation 55 . Furthermore, isotopic fractionation during microbial metabolism of SOM and the downward cycling of hydrophilic 13 C enriched decomposing products with dissolved organic carbon fluxes possibly contributing to the establishment of the vertical δ 13 C depth trends 54,56 . The pattern of soil δ 13 C observed in PF of DH showed an increase from 0 to 40 cm and then a decrease until 100 cm depth layer reflecting the more complex processes of microbial degradation and mixing of soil C of different ages 57 . On the other hand, the low soil δ 15 N values observed at the surface soil across the study sites particularly at XY could be due to the continuous addition of plant residues with extremely low δ 15 N values 58 .

Effects of climatic factors on carbon, nitrogen and their isotopic abundances. Climatic variables
have strong impacts on soil organic C and N contents 59,60 , and natural abundance of soil 13 C 20,22 and 15 N 61,62 . There were linear and non-linear regressions of soil C/N ratios and δ 13 C with MAT and MAP in both NF and plantations, which indicated that SOC turnover rates were largely determined by MAT and MAP. Our result was consistent with Wang et al. 22 who examined large-scale controls of climate over patterns of SOC turnover across terrestrial biomes worldwide using a meta-analysis of soil δ 13 C in previous literatures and demonstrated that SOC turnover was substantially accelerated with increasing MAT. In addition, the distribution of soil δ 13 C was slightly correlated with MAP after LUC probably due to the replacement of vegetation types which determines changes in the relative vertical distribution of soil C along rainfall gradients 41 after LUC. This result is in line with those found by Burke et al. 63 who reported that precipitation clearly has a direct role regionally and globally in the amount of soil C stored. Thus, our findings revealed that MAP is a key factor controlling soil C accumulation and decomposition after the conversion from NF to PF. According to Jia et al. 64 , vegetation type and soil type rather than MAT explained the variability in soil δ 13 C along the 400 mm isohyet in China.
Previous studies suggested that climate is the primary factor controlling soil δ 15 N 61 . Contrary to these results, we did not observe clear relationship between climate factors and δ 15 N in the present study. This could be due to the climate change which may change soil δ 15 N values through altered precipitation patterns, elevated temperatures, and more frequent and extreme weather events. In addition, climate change can affect temperature-sensitive biogeochemical processes, including N mineralization, nitrification and denitrification, soil respiration, litter decomposition, as well as root dynamics and plant productivity 65 , consequently, altering the rate of δ 15 N accumulation.

Materials and Methods
Study sites and soil sampling. In August 2017, soil samples were collected from one natural forest (NF) and one plantation forest (PF) stand at six sites including Mohe (MH), Qingyuan (QY), Xinyang (XY), Huitong (HT), Dinghushan (DH), Jianfengling (JF) across eastern China forest (Fig. 5, Table 5). To minimize the impact of local terrain on vegetation or trees, the topography of all selected plots was uniform and flat. In each one of these, the forest floor litter was collected using three 10 cm × 10 cm wooden frames in each stand. Sampling was done carefully in order to avoid contamination with the mineral material. Forest floors were very thin in most stands and were dried at 65 °C for 72 h and roots were sorted out and weighed to determine dry mass. Mineral soils were sampled in the same points where forest floors had been removed. For soil sampling, three pits were dug at 1-meter depth, and mineral soil samples were collected at 0-10, 10-20, 20-40, 40-60, 60-80 and 80-100 cm along soil profiles. A total of 252 samples were collected from 6 sites × 2 land uses × 3 pits × 7 layers. Mean annual temperature (MAT) and mean annual precipitation (MAP) data at six sites were recorded from the adjacent climate observing stations during 1960-2014.
Soil C and N contents, and δ 13 C and δ 15 N analysis. Forest floor litter samples were oven-dried at 65 °C for 72 h, and mineral soil samples were air-dried at room temperature for 2 weeks before passing through a 2 mm sieve to remove coarse stones, plant residues, and roots. Then, the forest floor and soil samples were ground to fine powders using a ball mill (JXFSTPRP-64, Jingxin Co., Ltd, China) for measurements of soil C and N contents, and δ 13 C and δ 15 N.
Soil C and N pools and stable isotope 13 C and 15 N composition were measured using an isotope ratio mass spectrometer (IRMS) (IsoPrime 100, Isoprime Ltd., UK), connected to a CN elemental analyzer (Vario MICRO cube, Elementar, Germany). Carbon and nitrogen stable isotope abundances were calculated as δ 13 C and δ 15 N (‰) using the following formula: where R sample is the 13 C: 12 C or 15 N: 14 N ratio in the samples and R standard is the 13 C: 12 C or 15 N: 14 N ratio in the standard. The Vienna Pee Dee Belemnite (VPDB) and atmospheric N 2 (δ 15 N = 0.0‰) were used as the standard, respectively. The precision of isotopic composition was checked using internal standards i.e. acetanilide, L-histidine, D-glutamic and glycine 28 . In general, the analytical precision for δ 13 C and δ 15 N was better than 0.2‰.

Statistical analysis.
The results are the average of the replicates determined from three subsamples of the same site. Statistically significant differences were determined by P < 0.05 unless otherwise stated. The sample