1 Effects of steel slag and biochar amendments on CO2, CH4, and N2O flux, and 2 rice productivity in a subtropical Chinese paddy field 3 Chun Wanga,b, Weiqi Wanga,b,*, Jordi Sardansc,d, Ankit Singlae,*, Congsheng Zenga,b, 4 Derrick Yu Fo Laif, Josep Peñuelasc,d a 5 Key Laboratory of Humid Subtropical Eco-geographical Process, Ministry of 6 Education, Fujian Normal University, Fuzhou 350007, China 7 b Institute of Geography, Fujian Normal University, Fuzhou 350007, China c 8 CSIC, Global Ecology Unit CREAF- CSIC-UAB. 08913 Bellaterra, Catalonia, 9 Spain d 10 CREAF. 08913 Cerdanyola del Vallès, Catalonia, Spain e 11 Regional Centre of Organic Farming, Ministry of Agriculture and Farmers Welfare, 12 Government of India, Bhubaneswar, 751021, India f 13 Department of Geography and Resource Management, The Chinese University of 14 Hong Kong, Hong Kong SAR, China 15 16 17 * 18 Author to whom correspondence should be addressed 19 E-Mail: Weiqi Wang: wangweiqi15@163.com; Tel.:+86-591-83465214; Fax: 20 +86-591-83465397. 21 Ankit Singla: ankitsingla2607@yahoo.co.in; Tel:+91-674-2721281 22 23 24 25 This is a post-peer-review, pre-copyedit version of an article published in Environmental geochemistry and health (Ed. Springer). The final authenticated version is available online at: DOI 10.1007/s10653-018-0224-7 1 26 Abstract Steel slag, a by-product of the steel industry, contains high amounts of 27 active iron oxide and silica which can act as an oxidizing agent in agricultural soils. 28 Biochar is a rich source of carbon, and the combined application of biochar and steel 29 slag is assumed to have positive impacts on soil properties as well as plant growth, 30 which are yet to be validated scientifically. We conducted a field experiment for two 31 rice paddies (early and late paddy) to determine the individual and combined effects 32 of steel slag and biochar amendments on CO2, CH4, and N2O emission, and rice 33 productivity in a subtropical paddy field of China. The amendments did not 34 significantly affect rice yield. The seasonal CO2 flux in each treatment was 35 correlated positively with soil temperature, and negatively with the water content, 36 salinity, and soil pH during most of the study period. The seasonal CH4 flux was 37 positively correlated with the soil salinity and water content in all treatments except 38 the biochar treatment in the early-paddy, and also with the soil temperature during 39 most of the study period. It was observed that CO2 was the main greenhouse gas 40 emitted from all treatments of both paddies. Steel slag decreased the cumulative CO2 41 flux in the late paddy. Biochar as well as steel slag+biochar treatment decreased the 42 cumulative CO2 flux in the late paddy and complete year (early and late paddy); 43 while, steel slag+biochar treatment also decreased the cumulative CH4 flux in the 44 early paddy. The biochar, and steel slag+biochar amendments decreased the 45 global-warming potential (GWP). Interestingly, the cumulative annual GWP was 46 lower for the biochar (55,, 422 kg CO2-eq ha-1), and steel slag+biochar (53, 965 kg 47 CO2-eq ha-1) treatments than the control (68,, 962 kg CO2-eq ha-1). Total GWP per 48 unit yield was lower for the combined application of steel slag+biochar (8951 kg 49 CO2-eq Mg-1 yield) compared to the control (12,, 805 kg CO2-eq Mg-1 yield). This 50 study suggested that the combined application of steel slag and biochar could be an 51 effective long-term strategy to reduce greenhouse gases emission from paddies 52 without any detrimental effect on the yield. 53 54 Keywords Paddy · greenhouse gases · steel slag · biochar · rice productivity 2 55 Introduction 56 Rice is a main cereal crop that currently feeds more than 50% of the global population 57 (Haque et al. 2015). Rice production will be required to be increased by 40% by the 58 end of 2030 to meet the demand for food of the growing population worldwide (FAO 59 2009). China has the second largest area of rice cultivation in the world, and the 60 emission of greenhouse gases (GHGs) from the rice cultivation account for 40% of 61 the total agricultural source of GHGs, especially the emissions of methane (CH4), and 62 nitrous oxide (N2O) (Myhre et al. 2013; Singla and Inubushi 2014; Singla et al. 2015). 63 Ninety percent of the paddies in China are in the subtropics, mainly in Fujian, Jiangxi 64 and Hunan Provinces. Developing strategies to increase the cost-effectiveness of rice 65 agriculture, enhancing rice yield, and mitigating GHG emission from paddies in 66 subtropical China are thus of national and global importance. 67 The application of exotic materials such as biochar (Zhang et al. 2010) or steel 68 slag (Wang et al. 2012a) is an important way to improve rice yields and mitigate GHG 69 emissions. The application of biochar has reduced N2O emissions from paddies 70 (Zhang et al. 2010; Wang et al. 2012b); however, the impact of biochar on CH4 71 emission from paddy field is still not clear, and it largely depends on the types of 72 biochar used (Zhang et al. 2012; Singla et al. 2014a). On the other hand, the 73 application of steel slag can reduce CH4 emissions (Furukawa and Inubushi 2002; Ali 74 et al. 2008; Singla and Inubushi 2015; Liu et al. 2016), although the reduction can 75 depend on the soil type, and paddies management (Xie et al. 2013). However, the 76 effect of steel slag on N2O emission is more complex (Zhu et al. 2013). Steel slag is 77 rich in iron (Fe), and the application of Fe-rich material can increase the amount of 78 iron plaques on rice roots; thereby, limiting the emission of GHGs to the atmosphere 79 (Huang et al. 2012). Biochar application has been reported to increase soil carbon (C) 80 (Cui et al. 2017); while decreasing inorganic nitrogen (N) (Nguyen et al. 2017). The 81 impact of biochar or steel slag on GHGs emission and rice yield in subtropical 82 paddies are lesser reported compared to temperate paddies (Furukawa and Inubushi 83 2002; Wang et al. 2014a). In our previous study, we demonstrated steel slag as an 3 84 effective amendment to reduce CH4 flux and increase rice yield over a short growing 85 season in a subtropical paddy in Fujian Province (Wang et al. 2014a). However, the 86 effect on N2O emission was uncertain during the growing period (Wang et al. 2015a). 87 The rational use of steel slag due to its high availability in China could become a 88 useful and cost-effective tool in the management of rice paddies (Xie and Xie 2003). 89 The effect of steel slag or biochar on CO2 emission from paddies has also been 90 less studied compared to CH4, and N2O emissions, and our understanding of the 91 impacts of the combined application of biochar and steel slag remains poor. We 92 assume that the combination of steel slag and biochar can superimpose the effects of 93 slag or biochar alone; thereby, reducing GHGs emission. Keeping these points in view, 94 the present study had the following objectives: a) to determine the effects of 95 individual or combined applications of steel slag and biochar on CO2, CH4, and N2O 96 fluxes in paddy field, and b) to assess the impacts of these amendments on rice 97 productivity in early and late paddies. We also aimed to provide a scientific base for 98 the selection and adaptation of countermeasures for mitigating GHG emissions from 99 rice cultivation. 100 Materials and methods 101 Study area and experimental fields 102 A field experiment during the early paddy season (16 April-16 July) and the late 103 paddy season (25 July-6 November) in year 2015 was conducted at the Fujian 104 Academy of Agricultural Sciences, Fujian, southeastern China (26.1°N, 119.3°E) 105 (Fig. S1). Air temperature and humidity during the study period are shown in Fig. 106 S2. The proportions of sand, silt and clay in the top 15 cm of the soil were: 28, 60, 107 and 12%, respectively. Other physicochemical properties of the soil were: bulk 108 density, 1.1 g cm-3; pH (1:5 with H2O), 6.5; organic C content, 18.1 g kg-1; total N 109 content, 1.2 g kg-1, and total phosphorus (P) content, 1.1 g kg-1 (Wang et al. 2014a, 110 2018a, b). The experimental plots were laid out in a randomized block design, with 111 triplicate plots (10 m2) for each of the four treatments (including a control). PVC 112 boards (0.5 cm thick, 30 cm high) were installed along the margins of each plot to 113 prevent the exchange of water and nutrients across different treatment plots. In each 4 114 plot, rice seedlings (early rice, Hesheng 10 cultivar; late rice, Qinxiangyou 212) were 115 transplanted to a depth of 5 cm with a spacing of 14 × 28 cm using a rice transplanter. 116 The field was plowed to a depth of 15 cm with a moldboard plow and was leveled 117 two days before rice transplantation and immediately after plowing. 118 The experiment with the following treatments was conducted in a completely 119 randomized block design: 1) control; 2) steel slag (8 Mg ha-1); 3) biochar (8 Mg ha-1); 120 and 4) steel slag (8 Mg ha-1)+biochar (8 Mg ha-1). Steel slag (Table S1) was obtained 121 from the Jinxing Iron & Steel Co., Ltd., Fujian. Biochar (produced by the pyrolysis 122 of rice straw at 600 °C for 90 min) was obtained from the Qinfeng Straw 123 Technology Co., Ltd., Jiangsu (Table S1). In our previous study, the application of 8 124 Mg ha-1 of steel slag (Wang et al. 2015 a) increased crop yield, and reduced GHG 125 emissions without increasing heavy metals in the soil or the rice grains over multiple 126 growing seasons. 127 Each experimental plot received the equal amount of water and mineral fertilizer. 128 All experimental plots for both seasons were flooded from 0 to 37 days after 129 transplantation (DAT), and the water level was maintained at 5-7 cm above the soil 130 surface by an automatic water-level controller. Each plot was drained between 37 to 131 44 DAT in both paddies. Afterward, the soil of each treatment was kept moist 132 between 44 to 77 DAT for the early rice, and 44 to 91 DAT for the late rice. The 133 paddy was drained two weeks before the harvest (77 DAT for the early rice, 91 DAT 134 for the late rice). The early rice was harvested at 92 DAT, and the late rice was 135 harvested at 106 DAT. The mineral fertilizer (N-P2O5-K2O: 16-16-16%; Jingzhou, 136 Keda Fertilizer Co., Ltd.), and urea (46% N) were applied in three doses: one day 137 before transplantation (N, P2O5, K2O @ 42, 40, 40 kg ha-1, respectively), during the 138 tiller-initiation stage (7 DAT; N, P2O5, K2O @ 35, 20, 20 kg ha-1, respectively) and 139 during the panicle-initiation stage (56 DAT; N, P2O5, K2O @ 18, 10 ,10 kg ha-1, 140 respectively). 141 Measurement of CO2, CH4, and N2O fluxes 142 Static closed chambers were used to measure CO2, CH4, and N2O emissions during 5 143 the study period. The chambers were made of PVC and consisted of two parts, an 144 upper opaque compartment (100 cm height, 30 cm width, 30 cm length) placed on a 145 permanently installed bottom collar (10 cm height, 30 cm width, 30 cm length). Each 146 chamber had two battery-operated fans to mix the air inside the chamber headspace, 147 an internal thermometer to monitor temperature changes during gas sampling and a 148 gas-sampling port with a neoprene rubber septum at the top of the chamber for 149 collecting gas samples from the headspace. To minimize the soil disturbance during 150 gas sampling, a wooden boardwalk was built for accessing the treatment plots. 151 Gas flux from each chamber was measured weekly. Gas samples were collected 152 from the chamber headspace using a 100-ml plastic syringe with a three-way 153 stopcock. The syringe was used to collect gas samples from the chamber headspace 154 0, 15, and 30 min after chamber deployment (Wang et al. 2015a). The samples were 155 immediately transferred to 100-ml air-evacuated aluminum foil bags (Delin Gas 156 Packaging Co., Ltd., Dalian, China) sealed with butyl rubber septa, and transported 157 immediately to the laboratory for the analysis of CO2, CH4, and N2O emission. 158 CO2, CH4, and N2O concentrations in the headspace air samples were 159 determined by a gas chromatography using a stainless steel Porapak Q column (2 m 160 length, 4 mm OD, 80/100 mesh) (CO2 and CH4 using a Shimadzu GC-2010, and 161 N2O using a Shimadzu GC-2014, Kyoto, Japan). A methane conversion furnace, 162 flame ionization detector (FID) and electron capture detector (ECD) were used for 163 the determination of the CO2, CH4, and N2O concentrations, respectively. The 164 operating temperatures of the column, injector and detector for the determination of 165 CO2, CH4, and N2O were adjusted to 45, 100 and 280 °C; to 70, 200 and 200 °C, and 166 to 70, 200 and 320 °C, respectively. Helium (99.999% purity) was used as a carrier 167 gas (30 ml min-1), and a make-up gas (95% argon and 5% CH4) was used for the 168 ECD. The gas chromatograph was calibrated before and after each set of 169 measurements using 503, 1030, and 2980 μl CO2 l-1 in He; 1.01, 7.99, and 50.5 μl 170 CH4 l-1 in He, and 0.2, 0.6, and 1.0 μl N2O l-1 in He (CRM/RM Information Center 171 of China) as standards. CO2, CH4, and N2O fluxes were then calculated as the rate of 6 172 change in the mass of CO2, CH4, and N2O per unit of surface area and per unit of 173 time. Three injections were used for each analysis. One sample was injected into the 174 GC for each analysis. The detection limits of the instrument were 1 ppm for CO2 and 175 CH4, and 0.05 ppm for N2O. We used linear calculations for CO2, CH4, and N2O 176 fluxes. 177 Global warming potential (GWP) 178 CO2 is typically used as the reference gas for estimating GWP, and a change in the 179 emission of CH4 or N2O was converted into “CO2-equivalents”. The GWP for CH4 is 180 34 (based on a 100-year time horizon and a GWP for CO2 of 1), and the GWP for 181 N2O is 298 (Myhre et al. 2013). The GWP of the combined emissions of CO2, CH4, 182 and N2O was calculated by the equation: 183 GWP = cumulative CO2 emission × 1 + cumulative CH4 emission × 34 + 184 cumulative N2O emission × 298 185 Measurement of soil properties 186 Soil samples in three replicates were collected from each treatment. The samples 187 were transported to the laboratory and stored at 4 °C until the analysis. Soil 188 temperature, pH, salinity, and water content in the top 15 cm of each plot were 189 measured in situ on each sampling day. Temperature and pH were measured with a 190 pH/temperature meter (IQ Scientific Instruments, Carlsbad, USA), salinity was 191 measured using a 2265FS EC meter (Spectrum Technologies Inc., Paxinos, USA) and 192 water content was measured using a TDR 300 meter (Spectrum Field Scout Inc., 193 Aurora, USA). 194 Statistical analysis 195 Differences in CO2, CH4, and N2O emissions among the treatments were tested for 196 statistical significance by general mixed models, using plots as random factors, and 197 using plots and time as nested factors within plots as random independent factors 198 when time was included in the analysis. We used the “nlme” (Pinheiro et al. 2016) R 199 package with the “lme” function. We chose the best model for each dependent 200 variable using Akaike information criteria. We used the MuMIn (Barton, 2012) R 7 201 package in the mixed models to estimate the percentage of the variance explained by 202 the model. We conducted Tukey’s post hoc tests to detect significant differences in 203 the analyses for more than two communities using the “multcomp” (Hothorn et al. 204 2013) R package with the “glht” function. 205 We also applied the data Normal test. The difference of treatments, sampling time, 206 and the interaction effects on GHG, and the soil properties were determined by 207 Repeated-measures analysis of variance (RM-ANOVAs). The relationships between 208 GHG fluxes and soil properties were determined by Pearson correlation analysis. 209 The significance of treatments was tested by Bonferroni’s post hoc tests (at P< 0.05). 210 These statistical analyses were performed using SPSS Statistics 18.0 (SPSS Inc., 211 Chicago, USA). We also performed multivariate statistical analyses using a general 212 discriminant analysis (GDA) to determine the overall differences in gas fluxes and 213 the soil traits in the samples from the amended treatments. We also took into account 214 the component of the variance due to the different DATs as an independent 215 categorical variable. The GDAs were performed using Statistica 6.0 (StatSoft, Inc., 216 Tulsa, USA). 217 Results 218 CO2 flux 219 The fluxes of CO2 from the early paddy varied significantly across sampling dates 220 (P< 0.01, Table S2) but not for the interactions between treatment and sampling date 221 or between the treatments (P> 0.05), except for the steel slag+biochar treatment (P= 222 0.05). Fluxes were low (< 381.85 mg m-2 h-1) during the initial growth period of the 223 early paddy (< 22 DAT) (Fig. 1a). However, the fluxes increased with the rice 224 growth and biomass, and peaked at 64 DAT as: 3448.49, 3605.36, 3530.20, and 225 3259.63 mg m-2 h-1 in the control, steel slag, biochar, and steel slag+biochar 226 treatments, respectively. 227 CO2 fluxes for the late paddy differed significantly across treatments and sampling 228 dates, and for the interactions between treatment and sampling date (P< 0.05, Table 229 S2). The CO2 fluxes were significantly lower in steel slag+biochar treatments 8 230 compared to the control (P< 0.05, Fig. 1b). The CO2 flux in each treatment increased 231 with rice growth and it was the highest at 36 DAT (2216.37 mg m-2 h-1), 78 DAT 232 (2226.11 mg m-2 h-1), 71 DAT (1842.26 mg m-2 h-1), and 50 DAT (1979.23 mg m-2 233 h-1) in the control, steel slag, biochar, and steel slag+biochar treatments, respectively. 234 CH4 flux 235 The fluxes of CH4 for the early paddy varied significantly across sampling dates (P< 236 0.01, Table S2) but not for the interactions between treatment and sampling date or 237 between the treatments (P> 0.05), except for the steel slag+biochar treatment (P< 238 0.05). Fluxes were low (< 5.02 mg m-2 h-1) during the initial growth period of the 239 early paddy (before 29 DAT), and increased to peaks at 36 DAT as: 2.94, 12.49, 240 15.00, and 7.40 mg m-2 h-1 (Fig. 2a) in the control, steel slag, biochar, and steel 241 slag+biochar treatments, respectively. Afterwards, it decreased steadily until the 242 harvesting of rice. 243 CH4 flux from the late paddy differed significantly across sampling dates and for 244 the interactions between treatment and sampling date (P< 0.05, Table S2), except for 245 the biochar treatment. However, it did not differ significantly among the treatments 246 (P> 0.05). Unlike for the early paddy, CH4 flux for the late rice was significantly 247 higher after transplantation (< 22 DAT) (Fig. 2b), and the fluxes from each treatment 248 were significantly lower after 22 DAT. The CH4 flux was the highest at 15 DAT for 249 the control (35.83 mg m-2 h-1), and at 8 DAT for the steel slag, biochar, and steel 250 slag+biochar treatments (33.81, 29.68 and 28.74 mg m-2 h-1, respectively). The CH4 251 flux from each treatment remained low (< 1.73 mg m-2 h-1) during the later period of 252 growth until the harvesting of rice in November, even though the paddy was 253 re-flooded at 50 DAT. 254 N2O flux 255 N2O flux for the early paddy varied significantly across sampling dates (P< 0.01, 256 Table S2) but not for the interactions between treatment and sampling date or among 257 the treatments (P> 0.05). The temporal pattern of N2O flux in each treatment was 258 almost similar during most of the observations (Fig. 3a). The fluxes were the highest 9 259 from the control, steel slag, and biochar treatments for the early paddy at 64 DAT 260 (347.39, 242.32, and 266.12 μg m-2 h-1, respectively) (Fig. 3a). Interestingly, the flux 261 for N2O from steel slag+biochar treatment was the highest at 78 DAT (128.93 μg m-2 262 h-1). 263 N2O flux for the late paddy differed significantly among sampling dates and 264 between the biochar and steel slag+biochar treatments (P< 0.05; Table S2, Fig.3b) 265 but not for the interactions between the treatment and sampling date (P> 0.05, Table 266 S2). Fluxes in the control treatment for the late paddy peaked at 85 DAT (137.06 μg 267 m-2 h-1); whereas, for steel slag, biochar, and steel slag+biochar treatments, it was the 268 highest at 106 DAT (120, 180, and 114 μg m-2 h-1, respectively). 269 Rice yield, cumulative flux and GWP of CO2, CH4, and N2O 270 The average rice yield was higher in the steel slag, biochar and steel slag+biochar 271 treatments than the control for both the early and late paddies, however, the 272 differences were not statistically significant (Table 1). The general mixed models 273 indicated that the treatments significantly affected late and total annual (early and 274 late) CO2 emissions, early-paddy CH4 emissions, GWP, and GWP by yield 275 production for late and cumulative (early and late) gas emissions (Table 1, S3). The 276 total cumulative CO2 flux for the late paddy and total year (early and late) was the 277 lowest in the biochar treatment followed by the steel slag+biochar, and steel slag 278 treatments. Cumulative CH4 and N2O (Table 1) fluxes did not differ significantly 279 among the four treatments; however, both were significantly lower for the steel 280 slag+biochar treatment than the control in the early paddy (P< 0.05). The average 281 fluxes of both gases for both paddies were lower in most of the amended treatments 282 than the control treatment which indicates a tendency of lower fluxes of CH4, and 283 N2O in the biochar and steel slag treatments compared to the control. 284 Differences in the soil properties among treatments 285 Soil pH (Fig. 4a, b), soil salinity (Fig. 4c, d) and water content (Fig. 4g, h) for the 286 early and late paddies differed significantly among sampling dates, treatments and 287 interactions between treatment and sampling date (P< 0.01, Table S4). Soil 10 288 temperature (Fig. 4e, f) for the early and late paddies differed significantly among 289 sampling dates (P< 0.01) but not for the interactions between treatment and 290 sampling date or between the treatments (P> 0.05). Soil pH (Fig. 4a, b), soil salinity 291 (Fig. 4c, d) and water content (Fig. 4g, h) for the early and late paddies were 292 significantly higher in all three amended treatments than the control (P< 0.05). Soil 293 temperature (Fig. 4e, f), however, did not differ significantly among the treatments 294 (P> 0.05). 295 Relationships between gaseous flux and the soil properties 296 For the early paddy, seasonal CO2 flux correlated positively with the soil temperature, 297 and negatively with the soil-water content (P< 0.01, Table S5). Seasonal CO2 flux 298 was also negatively correlated with soil salinity and pH during most of the study 299 period. Seasonal CH4 flux was positively correlated with soil salinity (P< 0.05) and 300 water content in all treatments (P< 0.05), except for the biochar treatment for the 301 early paddy. Seasonal CH4 flux was also positively correlated with soil salinity, 302 water content and temperature in all treatments (P< 0.05) for the late paddy. 303 Seasonal N2O flux was not clearly correlated with any of the parameters for either 304 cropping. The GDA identified all these trends, with lower gas fluxes and soil 305 temperatures and higher soil pHs, water contents and salinities in the amended 306 treatments than the control (Fig.5 a, b). The variance in the multi-dimensional space 307 generated by the various gas fluxes and the soil variables differed significantly 308 between the amended treatments and the control for both the early and late paddies 309 (Tables S6 and S7). CH4 fluxes and soil salinity, water content and pH contributed 310 significantly to the models for both the early and late paddies (Tables S8 and S9). 311 Discussion 312 Effects of amended treatments on CO2 flux 313 In this study, CO2 flux varied seasonally, increasing with the rice growth and 314 temperature. It has been observed that the temperature is highly correlated with CO2 315 production and emission (Liu et al. 2011; Treat et al. 2014) as rise in temperature 316 increases soil microbial activities (Vogel et al. 2014), and alters plant respiration 11 317 (Slot et al. 2013). In our study, CO2 flux decreased significantly in the steel slag, 318 biochar, and steel slag+biochar treatments for the late paddy as well as for the sum 319 of both paddies (Table 1). Biochar and steel slag are known to be alkaline; hence, 320 their application can increase soil pH (Liu et al. 2011; Ma et al. 2013). The increase 321 in the soil pH will increase absorption of CO2 to the paddy water; resulting into the 322 decreased CO2 flux (Table 1). Moreover, the steel slag and biochar also contain 323 calcium (Ca) (Table S1) which may combine with CO2 to form CaCO3. The CaCO3 324 can be deposited in the soil and decreases CO2 emission (Phillips et al. 2013). Some 325 studies have also suggested that biochar amendments to the soil can potentially 326 induce a positive priming effect, with an increase in the decomposition of resident 327 soil organic matter (SOM) (Luo et al. 2011). The effect of biochar on the 328 mineralization of SOM depends on the production temperature: biochar produced at 329 low temperatures (250-400 °C) stimulates C mineralization, whereas biochar 330 produced at high temperatures (525-650 °C) like in our study, suppresses C 331 mineralization (Saarnio et al. 2013); ultimately, decreasing CO2 emission. 332 The application of steel slag may increase soil Fe3+ concentration, thereby 333 enhancing the formation of iron plaques around rice roots, and thus limiting the 334 transport of nutrients, water, and dissolved organic carbon (DOC) to the roots 335 (Huang et al. 2012). Transport by rice plants is the most important pathway of GHG 336 emission to the atmosphere (Wassmann and Aulak 2000). Iron plaques decrease root 337 ventilation which results into lesser transportation of CO2 through the internal 338 system of interconnected gas lacunae of the plants; thereby, lowering the soil CO2 339 emission (Tavares et al. 2015). The combined application of steel slag and biochar 340 for the sum of both paddies showed a tendency of the lowest CO2 emission; however, 341 it was not significantly lower than biochar treatment (Table 1). 342 Effects of amended treatments on CH4 flux 343 In agreement to the report of Minamikawa et al. (2014), CH4 emission from each 344 treatment in our study were lower soon after rice transplantation, and during 345 drainage periods and the final ripening stage (Fig. 2a, b). It was reported that 12 346 lowering the water level in the rice field decreases CH4 production by decreasing the 347 abundance of the methanogenic archaeal population (Minamikawa et al. 2014). It has 348 been observed that Fe3+ is an alternative electron acceptor that will use C substrates 349 before methanogens (Jiang et al. 2013); thus, decreasing the amount of CH4 350 production following the applications of steel slag (Wang et al. 2014a; Singla and 351 Inubushi 2015). 352 Biochar amendment increases the soil ventilation (Revell et al. 2012) which 353 results in decreased CH4 production (Lehmann 2007). Our statistical analysis found 354 no significant difference for CH4 flux among the treatments but a lower average CH4 355 flux in all three amended treatments (Table 1) suggest a tendency for reducing CH4 356 flux under long-term applications. The effect of biochar or any other organic 357 amendments including steel slag on CH4 emission will depend on the physical and 358 chemical properties of the organic amendment, the type of soil, the microbiological 359 circumstances and the management of water and fertilizer. Liu et al. (2011) observed 360 a decrease in CH4 production under waterlogged incubated soil after the 361 application of bamboo and rice-straw biochar pyrolyzed at 600 °C. The raw 362 material and pyrolytic conditions for biochar production can collectively affect the 363 availability of C from biochar, thus affecting CH4 emission (Liu et al. 2011; Singla 364 and Inubushi 2014). The biochar prepared from straw (as in our study) or corn are 365 generally more porous which may or may not decrease CH4 emission (Table 1; Liu 366 et al. 2011). 367 Effects of amended treatments on N2O flux 368 In the present study, N2O flux was generally low throughout the growing seasons 369 (Fig. 3a, b). The paddies in our study region are highly N limited (Wang et al. 370 2014b), so, together with the low levels of soil O2, most of the N2O produced would 371 have probably been reduced to N2, which caused apparently very low emissions or 372 even a net uptake of N2O (Fig. 3a, b; Zhang et al. 2010). The application of biochar 373 and steel slag in all the three treatments tended to lower the average N2O emission 374 compared to the control; however, the difference was not significant (Table 1; S2). 13 375 Steel slag acts as an oxidizing agent which may stimulate N2O emission by 376 enhancing the rate of nitrification. The application of steel slag in our study did not 377 increase N2O emissions, a result also reported by Singla and Inubushi (2015). 378 Biochar, an alkaline material, can stimulate N2O reductase activity; thereby, inducing 379 the reduction of N2O to N2 (Yanai et al. 2007). The porous structure of biochar may 380 also absorb NH4+-N and NO3-N (Singla et al. 2014b); hence, decreasing N2O 381 emission (Cayuela et al. 2010). In contrary to this, the application of biochar may 382 also increase N2O emission (Yanai et al. 2007; Saarnio et al. 2013; Xie et al. 2013) 383 or may not cause any significant change in N2O emission (Kammann et al. 2012; Xie 384 et al. 2013). These studies suggest that the amount of N2O emission will also depend 385 on the physical and chemical properties of the biochar, the raw material used for 386 biochar preparation, the type of soil, microbiological activity and composition and 387 the management of water and fertilizer. Bruun et al. (2012) used wheat straw as the 388 raw material for biochar preparation, which reduced N2O emission. On the other 389 hand, corn (Xie et al. 2013) or biogas-digested slurry (Singla et al. 2014b) as the raw 390 material of biochar preparation increased N2O emissions. The observations by Bruun 391 et al. (2012) were closer to those of our study due to the similarity of the raw 392 material used for biochar preparation. 393 In addition, the absence of a consistent effect of the steel slag and biochar on N2O 394 flux from the paddy could be attributed to several possible mechanisms: an 395 inhibition of the enzymatic reduction of N2O by higher levels of Fe3+ (Huang et al. 396 2009), an increase in the production of hydroxylamine by the biological oxidation of 397 NH4+ favored by higher Fe3+ concentrations (Noubactep 2011). These possible 398 mechanisms may have been responsible for not only the lack of significant 399 differences between the treatments but also for the average lower N2O emissions in 400 the treatments containing steel slag and biochar (Table 1). 401 Effect of amended treatments on GWP and rice yield 402 The application of steel slag and biochar resulted in overall lower GWP (kg CO2-eq 403 Mg-1 yield) compared to the control (Table 1). CO2 contributed the most towards the 14 404 cumulative GWP (kg CO2-eq ha-1) and was also responsible for the significant 405 difference in the cumulative GWP (kg CO2-eq ha-1; kg CO2-eq Mg-1 yield) for the 406 sum of both paddies. The rice yield did not differ significantly between treatments 407 but tended to increase with the application of steel slag and biochar (Table 1). The 408 application of silicate fertilizers or biochar should have a positive impact on the rice 409 growth and yield parameters (Ali et al. 2008; Singla et al. 2014a; Singla and 410 Inubushi 2015). The increase in the soil pH (Fig. 5a, b) by the application of 411 slag-type fertilizers or biochar may also increase the content of available phosphate 412 in paddy soil, thereby increasing the rice yield. The inputs of organic C can stimulate 413 soil microbial activity and nutrient recycling (Antil et al. 2009). Additionally, the 414 activities of some enzymes, e.g. alkaline phosphatase, aminopeptidase, 415 N-acetylglucosaminidase and urease can increase with the additions of organic 416 matter to the soil (Bailey et al. 2010). Organic matter may also influence crop yield, 417 depending on soil type, crop type and many other environmental factors (Kolb et al. 418 2009; Bruun et al. 2012), and the application of biochar may even decrease crop 419 yield (Xie et al. 2013; Singla et al. 2014b). The most interestingly, GWP by yield (kg 420 CO2-eq Mg-1 yield) was the lowest in the steel slag+biochar treatment (Table 1). 421 Conclusions 422 The cumulative CO2 emission and GWP for the sum of both paddies (early and late 423 paddy) were the lowest for biochar, and steel slag+biochar treatments. The 424 application of biochar and steel slag alone or in the combination showed a tendency 425 of reducing CH4 and N2O emission, and increasing the rice yield. The most 426 importantly, steel slag+biochar treatment resulted in the lowest cumulative GWP by 427 yield, indicating that the combined application of both could be an effective 428 long-term strategy for reducing GHGs emission and increasing the rice yield. 429 Acknowledgments 430 The authors would like to thank Dengzhou Gao, Miaoying Wang and Qiuli Zhu for 431 their assistance during field sampling. Funding was provided by the National 432 Science Foundation of China (41571287, 31000209), Natural Science Foundation 15 433 Key Programs of Fujian Province (2018R1034-1), Outstanding Young Research 434 Talents in Higher Education of Fujian Province (2017), the European Research 435 Council Synergy grant ERC-SyG-2013-610028 IMBALANCE-P, the Spanish 436 Government grant CGL2013-48074-P and the Catalan Government grant SGR 437 2014-274. 438 Conflicts of Interest 439 The authors declare no conflicts of interest. 440 16 441 References 442 Ali, M. A., Oh, J. H., Kim, & P. J. (2008). Evaluation of silicate iron slag amendment 443 on reducing methane emission from flood water rice farming. Agriculture, 444 Ecosystems & Environment, 128, 21–26. 445 Antil, R. S., Janssen, B. H. & Lantinga, E. A. (2009). Laboratory and greenhouse 446 assessment of plant availability of organic N in animal manure. Nutrient Cycling 447 in Agroecosystems, 85, 95-106. 448 Bailey, V. L., Fansler, S. J., Smith, J. L. & Bolton, Jr. H. (2010). Reconciling 449 apparent variability in effects of biochar amendment on soil enzyme activities by 450 assay optimization. Soil Biology and Biochemistry, 43, 296-301. 451 Barton, K., (2012). MuM. In: Multi-model inference. R package version 1.7.2. 452 http://CRAN. R-283 project.org/package=MuMIn. at http://cran.r-project.org/ 453 package=MuMIn>. 454 Bruun, E. W., Ambus, P., Egsgaard, H. & Hauggaard-Nielsen, H. (2012). Effects of 455 slow and fast pyrolysis biochar on soil C and N turnover dynamics. Soil Biology 456 and Biochemistry, 46, 73-79. 457 Cayuela, M. L., Oenema, O., Kuikman, P. J., Bakker, P.R. & Van Groenigen, J. W. 458 (2010). Bioenergy by-products as soil amendments? Implications for carbon 459 sequestration and greenhouse gas emissions. GCB Bioenergy, 2, 201–213. 460 Cui, Y. F., Meng, J., Wang, Q. X., Zhang, W. M., Cheng, X. Y., Chen, W. F., (2017). 461 Effects of straw and biochar addition on soil nitrogen, carbon, and super rice 462 yield in cold waterlogged paddy soils of North China. Journal of Integrative 463 Agriculture, 16, 1064-1074. 464 FAO. [Food and Agricultural Organization of the United Nations]. (2009). 465 OECD-FAO Agricultural Outlook 2011–2030. 17 466 Furukawa, Y. & Inubushi, K. (2002). Feasible suppression technique of methane 467 emission from paddy soil by iron amendment. Nutrient Cycling in 468 Agroecosystems, 64, 193–201. 469 Haque, M. M., Kim, S. Y., Ali, M. A. & Kim, P. J. (2015). Contribution of 470 greenhouse gas emissions during cropping and fallow seasons on total global 471 warming potential in mono-rice paddy soils. Plant and Soil, 387, 251–264. 472 Hothorn, T., Bretz, F., Wesrfall, P., (2013). Package “mulcomp” (WWW document). 473 U.R.L. http://cran.stat.sfu.ca/web/packages/mulcomp/mulcomp.pdf. (accessed 474 16.03. 17). 475 Hou, H., Peng, S., Xu, J., Yang, S. & Mao, Z. (2012). Seasonal variations of CH4 476 and N2O emissions in response to water management of paddy fields located in 477 Southeast China. Chemosphere, 89, 884–892. 478 Huang, B., Yu, K. & Gambrell, R. P. (2009). Effects of ferric iron reduction and 479 regeneration on nitrous oxide and methane emissions in a rice soil. Chemosphere, 480 74, 481–486. 481 Huang, Y., Chen, Z. & Liu, W. (2012). Influence of iron plaque and cultivars on 482 antimony uptake by and translocation in rice (Oryza sativa L.) seedlings exposed 483 to Sb (III) or Sb (V). Plant and Soil, 352, 41–49. 484 Jiang, G., Sharma, K. R. & Yuan, Z. (2013). Effects of nitrate dosing on 485 methanogenic activity in a sulfide-producing sewer biofilm reactor. Water 486 Research, 47, 1783–1792. 487 Kammann, C., Ratering, S., Eckhard, C. & Müller, C. (2012). Biochar and hydrochar 488 effects on greenhouse gas (carbon dioxide, nitrous oxide, and methane) fluxes 489 from soils. Journal of Environmental Quality, 41, 1052-1066. 18 490 Kolb, S. E., Fermanich, K. J. & Dornbush, M. E. (2009). Effect of charcoal quantity 491 on microbial biomass and activity in temperate soils. Soil Science Society of 492 America Journal, 73, 1173-1181. 493 Lehmann, J. (2007). Bio-energy in the black. Frontiers in Ecology and the 494 Environment, 5, 381–387. 495 Liu, Q., Liu, B., Ambus, P., Zhang, Y., Hansen, V., Lin, Z., Shen, D., Liu, G., Bei, Q., 496 Zhu, J., Wang, X., Ma, J., Lin, X., Yu, Y., Zhu, C. & Xie, Z. (2016). Carbon 497 footprint of rice production under biochar amendment – a case study in a 498 Chinese rice cropping system. Global Change Biology Bioenergy, 8, 148-159. 499 Liu, Y., Yang, M., Wu, Y., Wang, H., Chen, Y. & Wu, W. (2011). Reducing CH4 and 500 CO2 emissions from waterlogged paddy soil with biochar. Journal of Soils and 501 Sediments, 11, 930–939. 502 Luo, Y., Durenkamp, M., Nobili, M. D., Lin, Q., & Brookes, P. C. (2011). Short term 503 soil priming effects and the mineralisation of biochar following its incorporation to 504 soils of different pH. Soil Biology & Biochemistry, 43(11), 2304-2314. 505 Ma, J., Wang, Z. Y., Stevenson, B. A., Zheng, X. J. & Li, Y. (2013). An inorganic 506 CO2 diffusion and dissolution process explains negative CO2 fluxes in 507 saline/alkaline soils. Scientific Reports, 3, 400. 508 Minamikawa, K., Fumoto, T., Itoh, M., Hayano, M., Sudo, S. & Yagi, K. (2014). 509 Potential of prolonged midseason drainage for reducing methane emission 510 from rice paddies in Japan: a long-term simulation using the DNDC-Rice 511 model. Biology and Fertility of Soils, 50, 879–889. 512 Myhre, G., Shindell, D., Bréon, F.M., Collins, W., Fuglestvedt, J., Huang, J., Koch, 513 D., Lamarque, J.F., Lee, D., Mendoza, B., Nakajima, T., Robock, A., Stephens, 514 G., Takemura, T. & Zhang, H., (2013). Anthropogenic and Natural Radiative 515 Forcing. In: Climate Change 2013: The Physical Science Basis. Contribution 19 516 of Working Group I to the Fifth Assessment Report of the Intergovernmental 517 Panel on Climate Change [Stocker, T.F., Qin, D., Plattner, G.K., Tignor, M., 518 Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (eds.)]. 519 Cambridge University Press, Cambridge, United Kingdom and New York, NY, 520 USA. pp 714. 521 Noubactep, C. (2011). On the mechanism of microbe inactivation by metallic iron. 522 Journal of Hazardous Materials, 198, 383–386. 523 Nguyen, T. T. N., Xu, C. Y., Tahmasbian, I., Che, R., Xu, Z., Zhou, X., Wallace, 524 H.M., & Bai, S.H. (2017). Effects of biochar on soil available inorganic 525 nitrogen: a review and meta-analysis. Geoderma, 288, 79–96. 526 Phillips, A. J., Lauchnor, E., Eldring, J., Esposito, R., Mitchell, A. C., Gerlach, R., 527 Cunningham, A. B. & Spangler, L. H. (2013). Potential CO2 leakage reduction 528 through biofilm-induced calcium carbonate precipitation. Environmental Science & 529 Technology, 47, 142–149. 530 Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., Core, T. R., (2016). nlme: Linear and 531 Nonlinear Mixed Effects Models. R package version 3.1-126, http://CRAN. 532 Revell, K. T., Maguire, R. O. & Agblevor, F. A. (2012). Influence of poultry litter 533 biochar on soil properties and plant growth. Soil Science, 177, 402–408. 534 Saarnio, S., Heimonen, K. & Kettunen, R. (2013). Biochar addition indirectly affects 535 N2O emissions via soil moisture and plant N uptake. Soil Biology and 536 Biochemistry, 58, 99-106. 537 Singla, A. & Inubushi, K. (2014). Effect of biochar on CH4 and N2O emission from 538 soils vegetated with paddy. Paddy and Water Environment, 12, 239-243. 539 Singla, A. & Inubushi, K. (2015). Effect of slag-type fertilizers on N2O flux from 540 komatsuna vegetated soil and CH4 flux from paddy vegetated soil. Paddy and 20 541 Water Environment, 13, 43-50. 542 Singla, A., Dubey, S. K., Ali, M. A., Inubushi, K. (2015). Methane flux from paddy 543 vegetated soil: a comparison between biogas digested liquid and chemical 544 fertilizer. Wetlands Ecology and Management, 23, 139-148. 545 Singla, A., Dubey, S.K., Singh, A. & Inubushi, K. (2014a). Effect of biogas digested 546 slurry-based biochar on methane flux and methanogenic archaeal diversity in 547 paddy soil. Agriculture, Ecosystems & Environment, 197, 278-287. 548 Singla, A., Iwasa, H. & Inubushi, K. (2014b). Effect of biogas digested slurry 549 based-biochar and digested liquid on N2O, CO2 flux and crop yield for three 550 continuous cropping cycles of komatsuna (Brassica rapa var. perviridis). Biology 551 and Fertility of Soils, 50, 1201-1209. 552 Slot, M., Wright, S. J., Kitajima, K. (2013). Foliar respiration and its temperature 553 sensitivity in trees and lianas: in situ measurements in the upper canopy of a 554 tropical forest. Tree Physiology, 33, 505–515. 555 Tavares, R. L. M., Farhate, C. V. V., de Souza, Z. M., Junior, N. L. S., Torres, J. L. 556 R. & Campos, M. C. C. (2015). Emission of CO 2 and soil microbial activity in 557 sugarcane management systems. African Journal of Agricultural Research, 10, 558 975–982. 559 Treat, C. C., Wollheim, W. M., Varner, R. K., Grandy, A. S., Talbot, J. & Frolking, S. 560 (2014). Temperature and peat type control CO2 and CH4 production in Alaskan 561 permafrost peats. Global Change Biology, 20, 2674–2686. 562 Vogel, J. G., Bronson, D., Gower, S. T. & Schuur, E. A. G. (2014). The response of 563 root and microbial respiration to the experimental warming of a boreal black 564 spruce forest. Canadian Journal of Forest Research, 44, 986–993. 565 Wang, W., Li, P., Zeng, C. & Tong, C. (2012a). Evaluation of silicate iron slag as a 21 566 potential methane mitigating method. Advanced Materials Research, 468, 567 1626–1630. 568 Wang, J., Pan, X., Liu, Y., Zhang, X. & Xiong, Z. (2012b). Effects of biochar 569 amendment in two soils on greenhouse gas emissions and crop production. Plant 570 and Soil, 360, 287–298. 571 Wang, W., Lai, D. Y. F., Li, S., Kim, P. J., Zeng, C., Li, P. & Liang, Y. (2014a). Steel 572 slag amendment reduces methane emission and increases rice productivity in 573 subtropical paddy fields in China. Wetlands Ecology and Management, 22, 574 683–691. 575 Wang, W., Sardans, J., Zeng, C., Zhong, C., Li, Y. & Peñuelas, J. (2014b). Response 576 of soil nutrient concentrations and stoichiometry to increased human disturbance 577 in a subtropical tidal wetland. Geoderma, 232, 459–470. 578 Wang, W., Sardans, J., Lai, D. Y. F., Wang, C., Zeng, C., Tong, C., Liang, Y. & 579 Peñuelas, J. (2015a). Effects of steel slag application on greenhouse gas 580 emissions and crop yield over multiple growing seasons in a subtropical paddy 581 field in China. Field Crops Research, 171, 146–156. 582 Wang, W., Lai, D. Y. F., Wang, C., Pan, T., & Zeng, C. (2015b). Effects of rice straw 583 incorporation on active soil organic carbon pools in a subtropical paddy field. 584 Soil & Tillage Research, 152, 8–16. 585 Wang, W., Lai, D.Y.F., Abid, A.A., Neogi, S., Xu, X., ＆Wang, C. (2018a). Effects 586 of steel slag and biochar incorporation on active soil organic carbon pools in a 587 subtropical paddy field. Agronomy, 8(8):135. 588 Wang, W., Zeng, C., Sardans, J., Zeng, D., Wang, C., Bartrons, M., ＆Peñuelas, J. 22 589 (2018b). Industrial and agricultural wastes decreased greenhouse-gas emissions 590 and increased rice grain yield in a subtropical paddy field. Experimental 591 Agriculture, 54 (4), 623–640. 592 Wassmann, R. & Aulakh, M. S. (2000). The role of rice plants in regulating 593 mechanisms of methane emissions. Biology and Fertility of Soils, 31, 20–29. 594 Xie, W. & Xie, X. (2003). Cleansing production technology of iron and steel industry 595 in China. Energ. Metal. Ind., 22, 49-53. 596 Xie, Z., Xu, Y., Liu, G., Liu, Q., Zhu, J., Tu, C., Amonette, J.E., Cadisch, G., Yong, J. 597 W. H. & Hu, S. (2013). Impact of biochar application on nitrogen nutrition of 598 rice, greenhouse-gas emissions and soil organic carbon dynamics in two paddy 599 soils of China. Plant and Soil, 370, 527-540. 600 Yanai, Y., Toyota, K. & Okazaki, M. (2007). Effects of charcoal addition on N2O 601 emissions from soil resulting from rewetting air-dried soil in short-term 602 laboratory experiments. Soil Science and Plant Nutrition, 53, 181–188. 603 Zhang, A., Bian, R., Pan, G., Cui, L., Hussain, Q., Li, L., Zheng, J., Zheng, J., Zhang, 604 X., Han, X. & Yu, X. (2012). Effects of biochar amendment on soil quality, 605 crop yield and greenhouse gas emission in a Chinese rice paddy: A field study 606 of 2 consecutive rice growing cycles. Field Crops Research, 127, 153–160. 607 Zhang, A., Cui, L., Pan, G., Li, L., Hussain, Q., Zhang, X., Zheng, J. & Crowley, D. 608 (2010). Effect of biochar amendment on yield and methane and nitrous oxide 609 emissions from a rice paddy from Tai Lake plain, China. Agriculture, 610 Ecosystems & Environment, 139, 469–475. 611 Zhu, X., Silva, L. C. R., Doane, T. A. & Horwath, W. R. (2013). Iron: The Forgotten 612 Driver of Nitrous Oxide Production in Agricultural Soil. PloS ONE, 8, e60146. 613 23 614 615 Legend to Figures 616 Fig. 1 Changes in CO2 emissions for the early (a) and late (b) paddies in the 617 treatments. Error bars indicate one standard error of the mean of triplicate 618 measurements. Different letters indicate significant different among treatments 619 (P<0.05). 620 Fig. 2 Changes in CH4 emissions for the early (a) and late (b) paddies in the 621 treatments. Error bars indicate one standard error of the mean of triplicate 622 measurements. 623 Different letters indicate significant different among treatments (P<0.05). 624 Fig. 3 Changes in N2O emissions for the early (a) and late (b) paddies in the 625 treatments. Error bars indicate one standard error of the mean of triplicate 626 measurements. 627 Different letters indicate significant different among treatments (P<0.05). 628 Fig. 4 Changes in soil pH (a), temperature (c), salinity (e) and water content (g) for 629 the early paddy and soil pH (b), temperature (d), salinity (f) and water content (h) for 630 the late paddy in the treatments. Error bars indicate one standard error of the mean of 631 triplicate measurements. 632 Different letters indicate significant different among treatments (P<0.05). 633 Fig. 5 Discriminant general analysis of the samples with treatment as a categorical 634 dependent variable; CO2, CH4 and N2O emissions and soil temperature, pH, salinity 635 and water content as independent continuous variables and sampling date as a 636 controlling variable for the late (a) and early (b) paddies. 24