The effects of nitrogen fertilization on N2O emissions from a rubber plantation

To gain the effects of N fertilizer applications on N2O emissions and local climate change in fertilized rubber (Hevea brasiliensis) plantations in the tropics, we measured N2O fluxes from fertilized (75 kg N ha−1 yr−1) and unfertilized rubber plantations at Xishuangbanna in southwest China over a 2-year period. The N2O emissions from the fertilized and unfertilized plots were 4.0 and 2.5 kg N ha−1 yr−1, respectively, and the N2O emission factor was 1.96%. Soil moisture, soil temperature, and the area weighted mean ammoniacal nitrogen (NH4+-N) content controlled the variations in N2O flux from the fertilized and unfertilized rubber plantations. NH4+-N did not influence temporal changes in N2O emissions from the trench, slope, or terrace plots, but controlled spatial variations in N2O emissions among the treatments. On a unit area basis, the 100-year carbon dioxide equivalence of the fertilized rubber plantation N2O offsets 5.8% and 31.5% of carbon sink of the rubber plantation and local tropical rainforest, respectively. When entire land area in Xishuangbanna is considered, N2O emissions from fertilized rubber plantations offset 17.1% of the tropical rainforest’s carbon sink. The results show that if tropical rainforests are converted to fertilized rubber plantations, regional N2O emissions may enhance local climate warming.


NH 4
+ -N fertilizer was responsible for higher soil N 2 O emissions than nitrate nitrogen (NO 3 − -N) fertilizer, as in their study the nitrification rates were higher from NH 4 + -N-treated plots than from NO 3 − -N-treated plots in subtropical forests. Few studies have examined changes in soil N 2 O emissions in response to the ongoing deposition of different forms of N in tropical forests in China. Hence, to support both estimations of the contributions of tropical forest soils to local and global N 2 O emission budgets and predictions of future trends in N 2 O emissions, it is important to quantify the N 2 O emissions from these forest soils in response to various N inputs. Therefore, the results of such a study will improve our understanding of the effects of chronic N inputs to tropical forest soils.
High global demand for natural rubber has resulted in sharp price increases. The unique climatic and growing conditions required by rubber trees mean that plantations are mainly concentrated in areas to the south of the Amazon River, western Peru, southern Bolivia and Brazil, and in some developing countries in Southeast Asia 30 . Rubber plantations in China have traditionally been planted in the south and southeast of the country. However, the ongoing expansion of rubber plantations has resulted in further loss of virgin tropical rainforest in other areas [31][32][33][34] . For example, in Xishuangbanna, southwest China, rubber plantations cover an area of ~4.7 × 10 5 ha (24.6% of the total land area), which represents about half the tropical rainforest in the region 34 .
Intensive soil-management practices, along with chronic additions of synthetic N fertilizers, mean that rubber plantations are potential N 2 O sources in tropical regions. Furthermore, while high rainfall, soil moisture content, soil temperature, and soil nutrient availability provide favorable environmental conditions for microbial activities, they are also important environmental factors that regulate N 2 O emissions from tropical forest soils. The main objective of the present study was therefore to improve our understanding of how soil N 2

Results
Site conditions at the rubber plantation. The results show that temporal variations in soil temperature (0-5 cm soil depth) followed a unimodal pattern, with the highest temperature recorded in the middle of the rainy season. The total rainfall from 1 April 2012 to 31 March 2013 was 1292.6 mm, and from 1 April 2013 to 31 March 2014 was 1187.3 mm; 80.0% of the rainfall and 83.6% of the rainfall events occurred during the rainy season. Only 13 rain events occurred during the dry season (Fig. 1a). The soil water content increased sharply because of high rainfall that occurred at the end of April after a relatively long dry period (Fig. 1b,c). There were strong positive correlations between soil temperature and soil water content in all the treatments and at the flux tower (Table S1). The temperature and soil water content measurements made at the flux tower, along with other parameters, were included in the correlation analyses.
The temporal variations in soil temperature (F = 0.316, p = 0.929) and soil water content (F = 0.368, p = 0.692) for the different treatments were not significant. However, we found significant differences in dissolved organic carbon (DOC) (F = 2.20, p = 0.037), microbial biomass carbon (MBC) (F = 32.34, p < 0.0001), NH 4 + -N (F = 6.38, p < 0.0001), NO 3 − -N (F = 4.12, p = 0.0002), and mineral N (F = 6.55, p < 0.0001) over the monitoring period ( Fig. 2). There were no clear seasonal trends in C and N concentrations (Fig. S1). The DOC, MBC, dissolved nitrogen (DN), NH 4 + -N, and dissolved organic nitrogen (DON) were higher in the unfertilized plots than fertilized plots, whereas NO 3 − -N was higher in the fertilized plots than unferitilized plots (Table 1). This clearly indicates that the fertilizer had a considerable effect on soil chemical properties in the rubber plantations after 2 years of N fertilizer applications.
Fertilizer effects on N 2 O emissions in the rubber plantation. The N 2 O flux rates showed similar seasonal trends in the unfertilized platform plots (NN), unfertilized slope plots (NN+ ), unfertilized trench plots (NNt), fertilized platform plots (UN), and fertilized slope plots (UN+ ) (Fig. 3a-e), with high flux rates during the rainy season when the soil temperature, soil water contents, and rainfall were high, and low flux rates during the dry season, when the soil water contents and soil temperature were low (Fig. 1a,b). The trends in N 2 O fluxes at the dry season fertilizer trench plots (UNts) and rainy season fertilizer trench plots (UNta) were different from those from the unfertilized platform, unfertilized slope, unfertilized trench, fertilized platform, and fertilized slope treatments. A month after fertilizer application during the dry season, the N 2 O fluxes increased sharply to a maximum and remained at a relatively high level about 4 months until the end of August (Fig. 3f). Similarly, during the rainy season the N 2 O fluxes increased rapidly at ~15 days after fertilizer application and remained high for about 3 months (Fig. 3g). After fertilizer application, it took much less time to reach the peak N 2 O flux in the rainy season than in the dry season. Likewise, the peak N 2 O flux that occurred after fertilizer application was maintained for a longer duration (about 4 months) in the rainy season than in the dry season (about 3 months). Because the fertilizer was applied during April and September, the area-weighted N 2 O flux from the fertilized rubber plantation showed a bimodal pattern (Fig. 4).
The mean N 2 O flux rates varied significantly (F = 23.14, p < 0.0001) among the different treatments during the monitoring period ( Fig. 3; Table S3). The flux rate was highest at the fertilized trench during the rainy season (0.58 ± 0.11 mg N m −2 h −1 ), closely followed by the dry season fertilized trench (0.55 ± 0.092 mg N m −2 h −1 ), and was lowest at the unfertilized trench (0.014 ± 0.0016 mg N m −2 h −1 ). The area-weighted N 2 O flux rate was significantly higher (t = 8.1, p < 0.0001, df = 118) at the N fertilized plots (0.046 ± 0.032 mg N m −2 h −1 ) than the unfertilized plots (0.029 ± 0.026 mg N m −2 h −1 ). The N 2 O fluxes from both fertilized and unfertilized plots were higher during the rainy season than during the dry season. The annual cumulative N 2 O fluxes from the fertilized and unfertilized treatments were 4.0 and 2.5 kg N ha −1 yr −1 , respectively. The N 2 O emission factor was 1.96% at the rubber plantation.

Factors influencing N 2 O emissions at the rubber plantation.
With the exception of the fertilizer trench in the rainy season, regression analysis shows significant correlations among soil temperature, soil water content, and N 2 O flux rates for the different treatments (Table S2). In contrast, there were no clear correlations between C fractions and the N 2 O flux rates or between the N fractions and the N 2 O flux rates (data not shown). The N 2 O flux rate was higher from the fertilized plots than from the unfertilized plots (Fig. 4), which reflects the fact that soil water content and temperature had a greater influence on the flux rates in unfertilized than did fertilized treatment, and that fertilizer applications decreased the influence of soil water content and soil temperarue on flux rates (Fig. 5). Regression analysis between soil temperature and N 2 O flux (Fig. 5) shows that Q10 dependence on the temperature of the unfertilized treatment (5.0) was higher than that of the fertilized treatment (3.9). Furthermore, Pearson correlations show that of the different C and N fractions, NH 4 + -N was significantly and positively correlated with the N 2 O flux rate from the fertilized (r = 0.45, p = 0.0063, n = 36) and unfertilized treatments (r = 0.37, p = 0.030, n = 36).
We used Pearson correlation analysis to assess the relationships among N 2 O flux, C and N fractions, and SWC in the different treatments. The results indicate that the N fractions, such as DN, DON, mineral N, NH 4 + -N, and NO 3 − -N, were significantly (p < 0.01) and positively correlated with the average N 2 O fluxes (Table 3). Stepwise linear regression analysis also indicated that the NH 4 + -N content was the key control on variations in the N 2 O fluxes from the different treatments. The regression equation was as follows: where y is the annual N 2 O flux (kg N ha −1 yr −1 ) and x is the average NH 4

Discussion
Nitrogen fertilization has a significant influence on N 2 O dynamics and contributes to high N 2 O fluxes 5,35,36 . The results from this study show that the N 2 O flux was higher from the fertilized rubber plantation (4.0 kg N ha −1 yr −1 ) than from the unfertilized rubber plantation (2.5 kg N ha −1 yr −1 ). The increased fertilizer ratio (160%) of the rubber plantation was less than the ratios reported for subtropical forests in southeast China (403-762%), where the fertilizer application rates ranged from 40 to 120 kg N ha −1 yr −1 16 . Although the amount of N fertilizer (75 kg ha −1 yr −1 ) applied to the rubber plantation was higher than that reported in some previous studies 10, 16,37,38 , which showed the contribution of inorganic N fertilizer applications to N 2 O emissions in the rubber plantation at Xishuangbanna were much lower. This result can be explained as follows. Firstly, fertilization in the rubber plantation followed the furrow fertilization model (Fig. 6). This meant that the fertilized trench plot was a hot spot of N 2 O emissions from which the emissions were significantly higher than from the slope and terrace plots ( Fig. 3; Table S3). The fertilized trench plots, however, accounted for only a small areal proportion ( Fig. 6) of the rubber plantation, therefore the overall contribution of the fertilized trench plots on N 2 O emissions were low. Secondly, there were no significant (p > 0.05) differences among the N 2 O fluxes from the fertilized and unfertilized slope and terrace plots ( Fig. 3; Table S3). Hence, the area-weighted mean N 2 O flux from the fertilized plot was only 1.6 times greater than that from the unfertilized plot. As increased amounts of fertilizer are applied, N 2 O emissions increase [39][40][41] . The N 2 O emissions from the fertilized rubber plantation (N application rate was 75 kg ha −1 yr −1 ) of the present study were lower than those from plantations where the N fertilizer application rates were higher (130-360 kg N ha −1 yr −1 ), such as papaya orchards (3.8-5.4 kg N ha −1 yr −1 ), coffee plantations (4.3 kg N ha −1 yr −1 ), coffee + tree legumes in Costa Rica (5.8 kg N ha −1 yr −1 ), and lychee orchards (4.5 ± 1.1 kg N ha −1 yr −1 ) [39][40][41] . The fertilization emission factor (1.96%) reported here was higher than those reported for the lychee (1.56%) 39 , papaya (0.90%) 41 , coffee+tree legume orchards (~1.0%) 40 in tropical regions, and the default factor (1.0%) reported by the IPCC 5 . This shows that the   (Table 3). Furthermore, stepwise linear regression showed that the NH 4 + -N content explained 94.5% of the variance in N 2 O among the treatments, and that the NH 4 + -N fraction dominated the spatial variations among the treatments. After 2 years without fertilizer application, the microbiomass community, which did not show the same degree of variation in the MBC and MBN (Fig. 2b,d) 21    The NH 4 + -N content, along with other factors, contributed to spatial variations in the N 2 O emissions from the different treatments. It was also one of the main controls on the area weighted mean values of N 2 O in the fertilized and unfertilized rubber plantation. This suggests that nitrification rather than denitrification controlled N 2 O emissions in the rubber plantation. It also contrasts with results from other studies in rubber plantation sites that are controlled by nitrification 10,42,43 , in which N 2 O emissions have generally been attributed to the combined effects of environmental, soil, vegetation, and human factors. While the soil pH in the rubber plantation of the present study (Table S4) was similar to that reported for other studies in tropical regions 12 Table 2. Pearson's correlations between the average N 2 O flux, soil water content, and fractions of C and N after the different treatments in the rubber plantation. * * Correlation is significant at the 0.01 level (twotailed). * Correlation is significant at the 0.05 level (two-tailed). n = 7.
Scientific RepoRts | 6:28230 | DOI: 10.1038/srep28230 of the filled pore space was less than 80% (Fig. S2), which triggered changes in the threshold of the soil microbiomass community for denitrication [21][22][23] . Thus, NH4 + -N and N 2 O emissions were significantly and positively correlated in the present study predicted nitrification may the main N 2 O emission way in this rubber plantation.
To understand the influence of the previously mentioned factors on N 2 O emissions, further studies are needed to (1) determine whether nitrification is the main control on N 2 O emissions in each treatment, (2) calculate the N 2 O:nitrification ratio in the rubber plantation, and (3) examine the N 2 O emission processes in isotope incubation experiments, in the field and the laboratory. Regardless of whether nitrification or denitrification dominates, the soil water content and temperature are the most important controls on N 2 O seasonal dynamics, and drive higher N 2 O emissions in the rainy season than in the dry season in most forests. The area-weighted mean N 2 O fluxes from each treatment indicate that the N 2 O fluxes were higher in the rainy season than in the dry season (Figs 2 and 4). With the exception of the fertilizer trench plot during the rainy season, the seasonality in the N 2 O flux rate can be explained by the significant correlations between N 2 O flux and soil temperature, and between N 2 O flux and soil water content (Figs 3 and 5; Table S1). Seasonal variations in N 2 O fluxes from the rainy season fertilizer trench plot were not well correlated with soil temperature, because the flux rate increased gradually after the rainy season fertilizer application as the soil temperature decreased (Figs 1a and 3g). Thus, the timing of fertilizer applications may also have a significant influence on the relationship between soil temperature and N 2 O fluxes 35,36,39 , and may cause variations in N 2 O flux rates within a given plantation. The N 2 O flux reached a peak more quickly after fertilization in the rainy season fertilizer trench plot than in the dry season fertilizer trench plot (Fig. 3f,g). Because of the relatively high soil water content and temperature in the later part of the rainy season (Fig. 1b), there was more microbiomass to produce N 2 O emissions 24,26 , so the increased N mineralization rate and N 2 O flux were sustained for a shorter time in the rainy season than after fertilization in the dry season 39 . In addition, fertilizer application reduced the sensitivity of N 2 O emissions to temperature and moisture fluctuations because of variations in the N input to the substrate ( Fig. 5; Fig. S1). The fact that N 2 O was more responsive to variations in soil moisture than soil temperature indicates that N cycling processes are relatively sensitive to variations in soil moisture because of the relatively small variations in temperature in the rubber plantation 24,25,40,42 .
The soil C and N fractions are the major energy sources for microbial activity and the substrate, respectively, which produce N 2 O emissions in tropical forests soil [42][43][44][45][46] . However, in this study there were no clear correlations between N 2 O fluxes and any of the different soil N or C fractions for any of the treatments. In addition, transfers in the microbiomass community via soil water may not be clear from the MBC and MBN dynamics [24][25][26] and, based on our results, the MBC and MBN for each treatment are not correlated to N 2 O fluxes. Thus, variations in the N and C fractions had little influence on temporal variations in N 2 O fluxes in all the treatments. Furthermore, the lack of clear trends in the contents of the C and N fractions in all the treatments (Fig. S1) may mean that a combination of other factors influenced N 2 O emissions. Temperature, soil moisture, and microbial bioactivity stimulate the N cycle 14,42,44 . Variations in N uptake from the surface and interflow runoff discharges of N influence the supply of N 2 O emissions to the substrate 13,[45][46][47] . Rubber tree phenology and changes in fine root dynamics in the rhizosphere microenvironment influence N and C availability [47][48][49] . Furthermore, variations in litter inputs and soil microbial biomass influence the soil C and N fractions 39,[44][45][46][47] , and, depending on the substrate and the energy supply, fertilization will have a considerable influence on N 2 O emissions 33,43,44 . Overall, N 2 O from the fertilized rubber plantation contributes about 1227.6 kg C ha −1 yr −1 to the 100-year GWP, which accounts for 6.0% of the rubber plantation net ecosystem exchange (NEE) (9.04 t C ha −1 yr −1 ) 49 . This figure is far less than that reported for an acacia plantation in Sumatra, Indonesia (10.0%) 37 , and shows that the N 2 O fluxes from the rubber plantation made a relatively small, yet significant, contribution to the GWP 17 . However, the NEE is similar to that reported for a Chinese chestnut plantation, where CO 2 emissions accounted for more than 95.0% of the total GWP, regardless of the understory management treatment 50 . Comparison with the results from the unfertilized rubber plantation showed that mineral fertilizer only contributed 1.5 kg N ha −1 yr −1 or a GWP of 708.8 kg CO 2 ha −1 yr −1 , which offset the NEE of the rubber plantation by 2.1%. A modeling study 7 has shown that the relatively low contribution of N 2 O to the GWP in the rubber plantation can be attributed to increases in the C sink potential and reductions in the GWP arising from the use of N fertilizer 17,50,51 .
In this study, the ratio of the rubber plantation N 2 O to the primary tropical forest NEE was estimated at 1.67 t C ha −1 yr −1 52 (Table 3). The ratio of the N 2 O GWP of the fertilized rubber plantation to the primary tropical rainforest NEE was relatively high (31.5%), and was 8.6 times greater than the N 2 O GWP:NEE ratio of the tropical rainforest. Most of the conversion from tropical rainforest to rubber plantation in Xishuangbanna has occurred since the 1970s [31][32][33][34] . In Xishuangbanna during 2014, rubber plantation and tropical rainforest covered areas of 47.12 × 10 4 hm 2 and 86.83 × 10 4 hm 2 , respectively 34 . Using the land use ratio (i.e., the ratio of the fertilized rubber plantation N 2 O GWP to the primary tropical forest NEE), we calculated that for the GWP the rubber plantation will offset 17.1% of the NEE of the primary tropical rainforest in Xishuangbanna. This indicates that as the area of the rubber plantation increases, the local tropical rainforest carbon sink will decrease because of the contribution of fertilizer in the rubber plantation, and there will be positive greenhouse effect feedbacks to the local climate with particular influences on precipitation and temperature 53 . Greenhouse gas emissions change with variations in land use and vegetation life cycles 1,2,5,6,8 . In the present study, we investigated only one stage of the life cycle of a mature rubber plantation. Hence, to obtain more accurate information about N 2 O feedback from fertilized rubber plantations to the local GWP, and the total accumulated greenhouse gas feedback to local climate change, future studies should investigate the whole life cycle of the rubber plantation.

Materials and Methods
Site description. The study site was a managed rubber plantation (21°55′ 30″ N, 101°15′ 59″ E; elevation 580 m) that was planted in 1993 in Xishuangbanna, Yunnan Province, China. The annual average temperature in this area is 21.7 °C and the mean annual precipitation is 1557 mm 53 . There is a rainy season from May to Scientific RepoRts | 6:28230 | DOI: 10.1038/srep28230 October and a dry season from November to March; the dry season is further divided into a cool, dry period from November to February, and a hot, dry period during March and April 54 . The tropical rainforest was felled in 1990, and rubber trees were planted in May 1993. Rubber trees were planted 2 m apart in rows. The spacing between adjacent rows was either narrow (3 m) or wide (19 m), and the trees were planted at a density of 495 rubber trees per hectare. The mean canopy height of the rubber plantation area ranged from 20 to 30 m. The main rubber-tapping period is from May to November. The plantation was not irrigated. Fertilizer was applied twice each year, during the dry season at the end of March (or the beginning of April) and during the rainy season at the end of August or the beginning of September. The study site is on a slope (15%) and is covered with a thick layer of oxisol, the physical and chemical characteristics of which are listed in Table S4.
Experimental design. In March 2012 we established a 1 ha study plot in the rubber plantation. We set up a randomized block experiment to compare the effects of mineral fertilizers on N 2 O emissions from the terraces, slope, and fertilizer trench. Taking into account the slope gradient, slope direction, and buffer zone, three fertilized and unfertilized blocks were installed from the bottom to the top of the slope around an eddy meteorological observation tower (on top of the hill) (Fig. 6). There were three terraces (NN), three fertilizer trenches (NNt), two narrow slopes, and two wide slopes (NN+ ) in the unfertilized plots, and four terraces (UN), three dry season fertilizer trenches (UNts), three rainy season fertilizer trenches (UNta), two narrow slopes, and two wide slopes (UN+ ) in the fertilized plots. In line with local farming practices, 1 kg of mineral fertilizer (Hubei Sanning Chemical, China) containing 15% N as (NH 2 ) 2 CO, 15% P as NH 4 H 2 PO 4 , and 15% K as KCl was applied to each rubber tree every year; this amounted to an application rate of 75 kg N ha −1 yr −1 , split into two applications (1 April and 6 September in 2012 and 2013). The fertilizer was applied by hand into a trench that was 150 cm long, 20 cm wide, and 10 cm deep between two rubber trees at a rate of 0.5 kg per tree per application.

Measurement of N 2 O emissions.
We measured the N 2 O fluxes simultaneously in all the fertilized and unfertilized treatments for 2 years from 1 April 2012 to 30 March 2014 using a static opaque chamber technique 20,41,52,53 . A chamber base collar made of PVC casing (covering an area of 0.12 m 2 ) was inserted to a depth of 0.10 m in the center of each field plot to measure the gas flux. The chamber base collars were kept in place throughout the entire measurement period. Chambers, 0.12 m 2 in area and 0.2 m high, were manually mounted onto the base collars to measure gas fluxes. To measure the N 2 O flux, five gas samples were taken from each chamber during the 1 h incubation at 15 min intervals. The gas samples were collected and stored in 100 ml plastic syringes and then transported to the laboratory within 3 hours of sampling. All the gas samples were analyzed immediately in the laboratory using a gas chromatograph (GC, Agilent 7890A, Agilent Technologies, CA, USA) equipped with an electron capture detector at 330 °C following the DN-Ascarite method with N 2 as the carrier gas and an ascarite filter to remove CO 2 from the samples 20,51,55,56 . The N 2 O samples were isothermally separated in two stainless steel columns (both with an inner diameter of 2 mm, and either 1 or 2 m long) packed with Porapak Q, 80-100 mesh, at 55 °C. To check the accuracy of the sample analysis, four standard N 2 O samples (National Center for Standard Matters, Beijing, China) with concentrations of 350 ppbv were analyzed by the GC system every 10 samples. Results of the GC analyses were accepted when the four gas standard calibrations produced variation coefficients that were lower than 1%.
The air pressure and headspace air temperature were recorded when the samples were collected from the chamber enclosure. The gas flux was calculated from the change in concentration within the enclosure, the recorded air temperature, and pressure. We used either a linear or nonlinear fitting approach to determine the initial rate of change in the concentration and the N 2 O flux [55][56][57] . Sampling and measurements were carried out weekly throughout the observation period. All samples were collected between 09:00 and 11:00 AM (local time) on the sampling day so that the flux rate was representative of the mean daily flux 51 . Emissions for days lacking measurements were estimated as the arithmetic mean flux of the two closest days for which measurements were available. The daily estimates were summed to obtain the total weekly, monthly, and annual emissions.
Auxiliary measurements. Soil volumetric water content and soil temperature. During gas sampling, the soil volumetric water content (0-12 cm soil depth) was determined with time-domain reflectometry (TDR100, Campbell Scientific, USA) in the soil close to the gas sampling plot. At the same time, the soil temperature (0-10 cm) and the air temperature were recorded with a needle thermometer. Missing soil water content and temperature data were calculated by calibrating the collected data with the soil water content (0-5 cm) and temperature, which was recorded half-hourly with a data logger fixed to the eddy flux tower (Campbell Scientific). Rainfall was recorded half-hourly by an automatic recorder attached to the top of the eddy flux tower.
Soil sampling and C/N fraction analysis. Throughout the soil N 2 O flux observation period, soil samples (~200 g) were collected at a depth of 0-20 cm every month from a point close to the static chambers using a stainless steel auger (3 cm in diameter). Once collected, soils were immediately passed through a 2 mm sieve to remove roots, gravel, and stones. All the soil analyses were completed within 48 hours. The soil samples were extracted in a 2.0 mol L −1 KCL solution (soil:water ratio of 1:10) and shaken for 1 h (160 rpm). The soil suspension was then filtered before NH 4 + -N and NO 3 − -N concentrations were determined with a continuous flow autoanalyzer (AutoAnalyzer 3; Germany). At the same time, a portion of each of the treated soil samples was pooled to make a composite sample for analysis of microbial biomass nitrogen (MBN), microbial biomass carbon (MBC), dissolved organic carbon (DOC), and total dissolved nitrogen (TDN). Soil MBC and MBN were determined by the chloroform fumigation-extraction method 58 . Four replicate samples of 7.0 g of each treated soil sample were fumigated with ethanol-free chloroform for 24 h at 25 °C in a sealed incubator in the dark. Three 7.0 g samples of each treated soil were not fumigated. After fumigation, the chloroform was completely removed from the soils, and both the fumigated and unfumigated samples were extracted with 35 ml of freshly prepared 0.05 M Scientific RepoRts | 6:28230 | DOI: 10.1038/srep28230 K 2 SO 4 (soil:water ratio of 1:5), capped, and shaken at 300 rpm for 1 h. The suspensions were then centrifuged for 10 min at 5000 × g and the supernatants were filtered through 0.45 μ m nitrocellulose membrane filters (Pall Life Science Company, Beijing, China). These filtered samples were analyzed for DOC and DN concentrations using Pt-catalyzed high-temperature combustion (680 °C) and a total organic carbon/total nitrogen analyzer (LiquiTOC II, Elementar Analyzer System, Germany). DOC and TDN were determined on the unfumigated filters, and the differences in DOC and TDN between the unfumigated and fumigated filters were the MBC and MBN, respectively. Mineral N was the sum of NH 4 + -N and NO 3 − -N, and DON was the difference between TDN and mineral N. Statistical analyses. The N 2 O flux from the fertilized and unfertilized areas was calculated as the area-weighted mean of three or more replicates from each treatment, as follows: , and all of the data complied with the assumptions of the analysis methods that we used. The N 2 O emission factor (EF) for rubber plantation soils after fertilizer had been applied was calculated as follows: 0 where E and E 0 are the N 2 O emissions from the fertilized and unfertilized fields during the same period, respectively, and N is the amount of nitrogen fertilizer applied 9 .