Effects of Biochar Addition on CO2 and N2O Emissions following Fertilizer Application to a Cultivated Grassland Soil

Carbon (C) sequestration potential of biochar should be considered together with emission of greenhouse gases when applied to soils. In this study, we investigated CO2 and N2O emissions following the application of rice husk biochars to cultivated grassland soils and related gas emissions tos oil C and nitrogen (N) dynamics. Treatments included biochar addition (CHAR, NO CHAR) and amendment (COMPOST, UREA, NO FERT). The biochar application rate was 0.3% by weight. The temporal pattern of CO2 emissions differed according to biochar addition and amendments. CO2 emissions from the COMPOST soils were significantly higher than those from the UREA and NO FERT soils and less CO2 emission was observed when biochar and compost were applied together during the summer. Overall N2O emission was significantly influenced by the interaction between biochar and amendments. In UREA soil, biochar addition increased N2O emission by 49% compared to the control, while in the COMPOST and NO FERT soils, biochar did not have an effect on N2O emission. Two possible mechanisms were proposed to explain the higher N2O emissions upon biochar addition to UREA soil than other soils. Labile C in the biochar may have stimulated microbial N mineralization in the C-limited soil used in our study, resulting in an increase in N2O emission. Biochar may also have provided the soil with the ability to retain mineral N, leading to increased N2O emission. The overall results imply that biochar addition can increase C sequestration when applied together with compost, and might stimulate N2O emission when applied to soil amended with urea.


Introduction
Biochar application to agricultural soils is a promising management practice that has the potential to mitigate climate change and increase soil quality [1][2][3][4][5]. However, it is not yet widely used in agricultural fields as a common practice, because the effects of biochar appear to be dependent on the characteristics of the soil and the biochar [6][7][8][9].
If biochar is a completely inert material that does not interact with soil components, only the C sequestration potential of biochar needs to be considered, and changes in the physical, been reported to have opposite effects on N 2 O emission. Higher N 2 O emission was observed from rice paddy soil amended with biochar made from swine manure by [32][33][34]; the changes in N 2 O emission were attributed to the size of the available inorganic N pool [32][33][34].
In this study, we monitored changes in CO 2 and N 2 O emissions from cultivated grassland soil in response to biochar amendment. Although the area of cultivated grassland in South Korea is not large, the government plans to extend this area because of increased demand for domestic forage. Grassland in Korea is also important because soil C storage in this ecosystem ranges from 3847gC m -2 to 9568 gC m -2 , which is approximately twice as large as that in Korean rice paddy systems [32]. Conventional management of grassland soil includes application of compost and/or N fertilizer. Compost application generally increases soil CO 2 evolution, while the application of N fertilizer is closely related to an increase in N 2 O emission [35][36][37]. We investigated changes in CO 2 and N 2 O evolution from soil treated both with biochar and other amendments. Two different types of amendments (compost and urea) were chosen because these are commonly used in current grassland management in Korea. To the best of our knowledge, this is the first field site experiment in South Korea to investigate the effects of biochar on the soil ecosystem. We related changes in CO 2 and N 2 O emissions to soil C and N dynamics and evaluated whether application of biochar is a sustainable management practice to simultaneously control greenhouse gas emissions and soil N availability.

Experimental site and basic physicochemical properties of biochar and soil
The field experiment was established on cultivated grassland located in Chunan-Si, Chungcheongnam-Do, Korea (East longitude 127°, North latitude 36°), where the average annual temperature and precipitation are 12.5°Cand 1226.5 mm, respectively (Korea Meteorological Administration).The field site was set up on Oct. 25, 2011 and the permission for the location was issued and managed by National Institute for Animal Science (NIAS), Korea. The mixseeding rates were tall fescue (Festuca arundinacea) 8 kg ha -1 , orchard grass (Dactylis glomerata) 15 kg ha -1 , perennial ryegrass (Lolium perenne) 5 kg ha -1 , and white clover (Trifolium repens) 2 kg ha -1 .
The biochar used was a commercial product sold by the Farmers' Association in Gangjingun, Korea. Biochar was produced by pyrolyzing rice husks at 500-600°Con a small-scale pyrolysis reactor (DCH-400, 1.4 m×5.2 m×5 m (L×W×H) from Daewon GSI Co., Korea.The amount of rice husk biochar processed was 400 kg h -1 . The residence time was 110-120 min and the overall yield was 43%. The particle size of rice husks was 5mm or less before pyrolysis.
Biochar and soil pH were determined at a 1:5 ratio of air-dried biochar or soil to deionized water(w/v). Biochar surface area was measured by the BET method with N 2 gas using millground biochar. Soil texture was determined by a hydrometer and the cation exchange capacity (CEC) of soil was measured using an unbuffered salt extraction method [38].Total carbon (TC) content in the soil and biochar were analyzed by combustion analysis using a Carlo Erba NS 1500 C/N analyzer (Carlo Erba, Milan, Italy). NH 4 + -N and NO 3 --N concentrations in the biochar and soil were determined through 2M potassium chloride (KCl) extraction and colorimetric methods [39].Metal content was analyzed by a spectroscopic method usingICP MS (Perkin-Elmer ICP-OES OPTIMA5300DV) to check whether the biochar contained metals that might be toxic to the soil ecosystem. The physicochemical properties of the biochar and soil are shown in Table 1.

Experimental design and treatments
Treatments included biochar addition (NO CHAR, CHAR) and amendment (COMPOST, UREA, NO FERT). We used factorial design for six treatment combinations: NO CHAR/COMPOST, NO CHAR/UREA, NO CHAR/NO FERT, CHAR/COMPOST, CHAR/UREA, and CHAR/NO FERT. As we performed three replicates of each type, eighteen plots were completely randomized to the 22 m×37 m experimental field. The size of each plot was 5m×7m and individual plots were separated by protection rows that were 1m in width.
Biochar was applied once a year in fall after harvest (Oct. 25, 2011 andSep. 21, 2012). The form of biochar applied was the particle of < 3 mm. The application rate of biochar was 0.3% by weight, which is equivalent to3.3 ton/ha when calculated based on a 10-cm depth field application. Biochar was incorporated into the soil profile down to a depth of 10 cm using shovels. The non-biochar amended plots were also mixed with a shovel. Soil amendment was conducted twice per year in fall and spring (Oct 25, 2011, March 21, 2012, Sep21, 2012, March 21, 2013. Amendment rates were140kg ha -1 yr -1 of UREA and 1500 kg ha -1 yr -1 of COMPOST, which are equivalent to 51.8 kg Nha -1 yr -1 and 11.3 kg Nha -1 yr -1 , respectively.

Gas sampling and analysis
Gas samples were taken every month from 10/31/2011 to 05/01/2013 except over winter using a chamber method [29].Two chambers (20-cm diameter, 25-cm height) were inserted 5 cm deep into the soil for each plot. On sampling dates, the chambers were closed with airtight lids for 40 minutes and gas samples were withdrawn from the headspace of the closed chamber using a 10-ml three-way syringe (BD Luer-LokTip).
Gas samples were analyzed using a gas chromatograph (Agilent 7890A, USA) equipped with two detectors. CO 2 was detected using a thermal conductivity detector (TCD) and N 2 O was detected using an electron capture detector (ECD).    Gas fluxes were calculated from the changes in headspace concentration over the measured period using the following equation [34]: where, dGas/dt is the difference in gas concentrations between the initial and end time points, V is the volume of the chamber, A is the surface area which the chamber covers, P is the atmospheric pressure, MW is the molecular weight of the gas, R is a gas constant, 8314 J mol -1 K -1 , and T is the absolute temperature.

Soil sampling and analysis
On the same dates as gas sampling, soil samples were collected from each plot from a depth of 0-15 cm using a soil core sampler (4.9-cm i.d., Forest supplier, USA). Samples were sealed in marked plastic bags and taken to the laboratory after sampling. For soil temperature measurements, a mini thermometer probe (Testo, 905-T1) was inserted to a depth of 10 cm. To measure soil gravimetric water content, approximately 15 g of soil was taken from each plastic bag and dried in the oven at 105°C for 24 h. Soil bulk density was determined by taking a soil core (4.9-cm diameter, 15-cm depth) and drying the soil at 105°C for 24 h to determine the soil dry weight contained in the known volume. Soil bulk density and gravimetric water content were used to determine the water-filled pore space (WFPS) of each soil core [39].
Temporal soil samples were passed through a 2-mm sieve and air-dried for 2 weeks before analysis. Labile organic C content of biochar was measured by the amount of hot water extractable C (HWC) following the method described in [40].To investigate N mineralization in soil, we measured NH 4 + -N and NO 3 --N. Soil NH 4 + -N and NO 3 --N concentrations were determined through 2M potassium chloride (KCl) extraction and colorimetric methods [41].Soil microbial activity was evaluated with the fluorescein diacetate (FDA) hydrolysis method [42].We analyzed the changes in soil pH, CEC, microbial biomass C and water holding capacity (WHC) for the last soil samples which was taken on 05/22/2013. Microbial analysis was measured by the CHCl 3 fumigation extraction method [43,44] and the WHC in the soil was determined using modification of the method described in [45].
To further examine the effect of biochar on soil N dynamics in the UREA soil, we set up a postulated balance of added urea N. The balance started with the amount of urea N applied which was 5180 mgN m -2 . We assumed that 40% of applied urea N would be recovered in the plant biomass and soil organic N pool and this amount was not affected by biochar amendment [46]. We also assumed that 20% of applied urea were lost via NH 3 volatilization process [47] and the amount of volatilization loss was not affected by biochar amendment. The remaining of the urea N was divided N 2 O emission, soil mineral N content (NH 4 + + NO 3 -), and N loss from leaching. We calculated the amount of N leaching by subtracting measured N 2 O emission and soil mineral N content from the remaining amount of the urea N.

Statistical analyses
Analysis of variance (ANOVA) was performed using the MIXED procedure of SAS 9.2 [48] on CO 2 emissions, N 2 O emissions, TC contents, bulk density, water filled pore space, microbial activity (FDA activity), HWC content, soil NH 4 + and NO 3 contents. Biochar treatment, urea and compost amendment, and date were fixed effects. The ANOVA was performed separately on the soil pH, CEC, microbial biomass C, and WHC for the final soil sample and for this ANOVA, date was not considered as a fixed effect because the measurement was conducted only once. Least square means of the parameters were used to compare date, fertilization, and biochar effects. Pearson's correlation coefficients among the CO 2 emissions, N 2 O emissions, soil temperature, soil water content, and HWC content were calculated using the CORR procedure of SAS [48]. Statistical significance was set to the 5%probability level for all analyses.

Results and Discussion
Carbon dioxide emission and soil C dynamics The overall CO 2 emission pattern was positively correlated with soil temperature ( Table 2, r = 0.610 ÃÃÃ ), which has been reported by many researchers [22,49,50]. In contrast, the CO 2 pattern was not correlated with soil water content, most likely because the gravimetric soil water content on our sampling dates was very low (range, 3.5-25.5%) (Fig 1A). An average throughout the 19 months of the field experiment revealed that overall CO 2 emission from the COMPOST soil was significantly greater than that from the UREA and NO FERT soils (Table 3, Fig 1C). Greater CO 2 emission from compost-treated soil than untreated soil has been reported previously by several researchers [51,52]. Soluble C measured by hot water extractable C (HWC) from compost was hypothesized to stimulate the microbial community, resulting in higher CO 2 evolution, and extractable organic C showed a high correlation with CO 2 evolution (Table 3, r = 0.586 ÃÃÃ ).However, unlike the consistent trend in HWC with CO 2 emission, the patterns in FDA activity and MBC were not consistent with that of CO 2 emission (Fig 2) and the correlations of them with CO 2 emission rate were not significant. We attributed this inconsistency to the lower sensitivity of FDA activity and MBC to the treatment. It was widely reported that FDA activity and MBC are highly correlated [53] and they both represent "potential" microbial activity because FDA activity is measured after sufficient substrate is added and MBC includes dead biomass of microbes. Hence, we would say that "actual" microbial activity in the COMPOST soil might have been stimulated but it was not detected by these parameters.
The temporal pattern of CO 2 emissions differed according to soil amendment type and biochar addition (Table 3II). Addition of biochar did not significantly affect CO 2 emissions from any of the amended or control soils at the start of the experiment until 06/01/2012 (Table 3, Fig  1C). During Jul 2012, when the soil temperature and CO 2 emission level were both very high, biochar addition had a negative effect onCO 2 flux in the COMPOST soil ( Fig 1B).From Aug 2012 to May 2013, we again did not observe any effects of biochar addition on CO 2 emission from any of the soils. The reason why the biochar treatment only had an effect on CO 2 emission when the temperature was high may in part be due to temporal changes in HWC ( Table 4). The HWC content on 07/22/2012 was the highest measured, and on the same day, it was higher in the CHAR soils than in the NO CHAR soils. This result implies that labile C contained in the COMPOST soil was respired more freely in the NO CHAR soil than the CHAR soil. The suppression of CO 2 emission by biochar was also reported in [19]; these authors attributed this to the high adsorptive affinity of biochar for existing organic C. We also argue that the higher content of labile C in the COMPOST soil protected from mineralization by biochar, resulting in a low CO 2 evolution rate in this treatment. The pattern of soil CO 2 emission was consistent with that of total C content (Fig 3). On average, soils with char addition had a 139.28% soil C content in the 19 months after biochar was first applied. In soils with biochar addition, total C was slightly higher in Nov 2012 than May 2013. The reason is that soil samples collected in Nov 2012 contained a lot of newly  Table 3. Results of analysis of variance by date, biochar and amendments. incorporated biochar particles, because we applied biochar in Sep 2012 for the second time, therefore there was not enough time for the biochar to become completely mixed with the soil. In May 2013, the soil C content was the highest in the COMPOST soil with biochar, and this result is consistent with low CO 2 emission from the COMPOST soil amended with biochar during the hot summer days. Overall, the data imply that when biochar is applied together with compost, higher C sequestration can be expected, probably because decomposition of compost C is lowered by adsorption of labile C by biochar.

Nitrous oxide emissions and soil available N dynamics
The overall patternofN 2 O emission showed a correlation with soil temperature ( Table 2, r = 0.216 Ã ), but no correlation with soil water content. Soil water status has been reported to have a strong effectonN 2 O emission by several research groups [23,39,54,55]. However, because the soil water content on our sampling dates was very low (less than 25% gravimetric water content) (Fig 1A), we did not observe a significant correlation between soil water content andN 2 O emission.  Table 4. Seasonal change in the soil hot water extractable C (HWC) concentrations by biochar and amendments. Comparison was made within the column by biochar and amendments within one date. N 2 O emission was significantly influenced by the interaction between amendment and biochar treatment (Fig 4A and 4B). In the COMPOST and NO FERT soils, biochar amendment did not change N 2 O emission at any point during the experiment. In contrast, biochar addition to UREA soil significantly stimulated N 2 O emission by 49% on average compared to soil without biochar. This result is consistent with the results reported in [56], namely thatN 2 O emission was significantly increased by biochar addition, especially when added together with mineral fertilizer N. Other researchers have also reported that biochar treatment of fertilized soil results in higher N 2 O emission than from the fertilized soil alone [32,39,57]. However, contrary to our results, it was reported that biochar treatment significantly reduced N 2 O emission from rice paddy soils with no N fertilization [19,29]. In the mechanism-based research of [29], 11 of 15 agricultural soils were found to emit a lower level of N 2 O upon biochar amendment, and these authors discussed the interaction between black C and N dynamics. However, Effect of Biochar on Greenhouse Gas Emission from a Grassland Soil no change in N 2 O emission in response to biochar addition was also noted when urea N was added [20]. These inconsistent results indicate that the effect of biochar on soil N 2 O emission is dependent on water status, pH, C status, and N fertilizer management of soils [4,33,58].
If we assume that heterotrophic nitrification is the main process resulting in N 2 O emission from the UREA soil, the higher N 2 O emission resulting from CHAR treatment of this soil can be explained as follows. In the CHAR treatment of UREA soil, consistently higher amounts of HWC and mineral N were observed (Table 4, Fig 5). Hence, the combination of a higher amount of labile C and mineral N can explain the higher rate of N 2 O emission from UREA soil with CHAR treatment compared to UREA soil with NO CHAR treatment due to stimulation of heterotrophic nitrification. This interaction was especially important at our field site because our soil was severely C limited due to a very low organic C content (0.4% weight basis). We did not consider biochar-derived mineral N as making a direct contribution to the higher mineral N content in the UREA soil because the mineral N content in the biochar has been reported to be very low [60] and was not increased by biochar addition to the COMPOST and NO FERT soils. This argument is further supported by the microbial FDA activity and microbial biomass C results (Fig 2). Higher FDA activity and microbial biomass C in the UREA soil with biochar addition indirectly indicated higher microbial autotrophic and heterotrophic nitrification in this soil.
Another possible mechanism to explain the higher N 2 O emission from the soil amended with both urea and biochar is as follows. Theoretical distribution of urea-derived mineral N in the UREA soil revealed that in the biochar-amended soil, N loss through N 2 O emission was higher by 48.89% and the soil mineral N content was higher by 28.13% than in soil without biochar. In contrast, N loss via leaching was significantly lower in CHAR soil than in the NO CHAR soil (Fig 6). This result is consistent with [61], which observed that soil with biochar addition showed less NO 3 leaching than soil without biochar. They proposed that the soil with biochar had better water-holding capacity due to the enhanced volume provided by soil mesoaggregates. In this study, we also observed higher WHC in the UREA and NO FERT soils with biochar addition (Fig 7), further supporting that N loss through leaching is reduced in biochar-treated soil.
In summary, C-poor soil treated with urea fertilizer and biochar was able to retain a higher amount of mineral N than untreated soil. This could supply sufficient available N for plant growth and explain the reported increases in plant yield in response to biochar addition ). The amount of NH 3 volatilization, plant biomass N and soil organic N were assumed to be the same between NO CHAR and CHAR treatments and not shown in the graph. Different letters beside the bars indicate significant differences between the NO CHAR and CHAR treatments at a 5% probability level.
doi:10.1371/journal.pone.0126841.g006 [62,63]. However, the higher retainment of mineral N within the soil might cause greater N loss through N 2 O emissions, which is a strong greenhouse gas. Therefore, when applying biochar and urea to C-poor soils, the growth pattern of plants should be considered to determine the best timing to maximize the plant's utilization and minimize N 2 O loss.

Conclusions
As grassland ecosystems have the potential to sequester significant amounts of C, development of good management practices for these systems is needed to enhance their potential to mitigate climate change by maximizing C sequestration and minimizing greenhouse gas emissions. When both compost and biochar were applied to soil, less CO 2 evolved from the soil during the summer, probably due to adsorption of labile C by the biochar. Soil C sequestration was highest in the COMPOST soil with biochar addition, consistent with the pattern of CO 2 evolution.
Soil N 2 O emissions were significantly influenced by interactions between the biochar and amendment materials. Higher N 2 O emission was observed in UREA soil with biochar than without it, while in the COMPOST and NO FERT soils, there was no difference in N 2 O emissions between soils with and without biochar. We proposed two possible mechanisms to explain the higher N 2 O emission from UREA soil amended with biochar. The first mechanism was enhanced autotrophic and heterotrophic nitrification due to interaction between labile C from the biochar and mineral N. When soil is C limited, as was the soil in our study, the addition of biochar could provide the soil with labile C, resulting in enhanced N mineralization (ammonification) and heterotrophic nitrification, finally leading to an increase in N 2 O emission. The second mechanism was enhanced retention of mineral N content by the biochar. Mineral N could easily be leached out if there was no holding mechanism in the soil. Biochar addition reduced N loss due to leaching in the UREA soil, leading to greater retention of available N, which resulted in higher N 2 O emissions from the soil. The overall N 2 O emission results indicate that care should be taken when applying biochar together with urea fertilizer. To develop a sustainable management strategy for fertilizer and biochar application, the timing of N fertilizer and biochar application should take into consideration of N loss through both N 2 O emission and leaching. In future studies, we intend to trace fertilizer N using an isotope to gain a mechanistic understanding of the effects of biochar on soil N dynamics.