Introduction

Rice cultivation has been identified as a major source of atmospheric greenhouse gas emissions (GHGs), mainly in the form of methane (CH4), globally releasing approximately 31–83 Tg year−1 (Reay et al. 2010). With a projected increase in global rice consumption of around 45 % by 2050 (Rejesus et al. 2012), efforts to intensify and boost rice production to meet the increasingly global demand will lead to a further rise in CH4 emissions into the atmosphere. Increasing rice production without increasing the release of GHG emissions therefore presents a major scientific and societal challenge.

In rice cultivation, rice straw (RS) is either removed from the fields for use as animal fodder, left in the fields or burned in situ. The burning of RS is generally not recommended as it causes nutrient loss, air pollution and damage to human health (Dobermann and Fairhurst 2002; Knoblauch et al. 2011). The incorporation of crop residues into the fields has therefore been widely promoted to preserve organic matter and nutrients in the soil. However, it has been reported that the incorporation of crop residues into the rice field increases CH4 emissions significantly (Naser et al. 2007; Wang et al. 2012b; Wassmann et al. 1996) as CH4 production in rice fields largely depends on the availability of readily decomposable organic materials that serve as a source of both carbon (C) and energy for microorganisms (Wassmann and Aulakh 2000). The effect of organic matter application depends on the amounts of the labile fraction in the applied and inherent soil organic matter at flooding (Yagi and Minami 1990). Therefore, the application of fresh organic materials will lead to considerable CH4 emissions. Various mitigation options have been proposed to reduce CH4 emissions from rice cultivation. These include modified water management, i.e. alternating wetting and drying (Itoh et al. 2011; Lu et al. 2000; Ly et al. 2013; Towprayoon et al. 2005), off-season application of organic materials and application of organic materials with more stabilised carbon, i.e. well-decomposed organic materials or biochar (incomplete combustion of organic materials) (Knoblauch et al. 2011; Shin et al. 1996).

Over the past few years, there has been increasing interest in biochar (BC) owing to its potential for carbon sequestration in soils, nutrient retention and GHG mitigations while simultaneously increasing crop yields (Feng et al. 2012; Ippolito et al. 2012; Sarkhot et al. 2012; Zhang et al. 2010). Woolf et al. (2010) estimated that annual net emissions of CO2, CH4 and N2O could be reduced by about 12 % of current anthropogenic CO2-eq emissions due to global sustainable implementation of BC, without compromising food security, habitat or soil conservation. Nevertheless, studies showed inconsistent results about the responses of CH4 emissions following BC amendment. Some studies (Knoblauch et al. 2011; Singla and Inubushi 2014; Wang et al. 2012a; Zhang et al. 2010; Zhang et al. 2012) reported a remarkable increase in CH4 emissions following BC amendment in paddy soils. The increase in CH4 emissions in BC-amended soils could be due to the labile C from BC as a predominant source of methanogenic substrates (Knoblauch et al. 2011). Other studies meanwhile found that CH4 emissions decreased with the addition of BC (Feng et al. 2012; Liu et al. 2011), and this can be explained by the inhibition of methanogenic activity, stimulation of methanotrophic activity and enhancement of CH4 adsorption in soil (Liu et al. 2011). Indeed, the underlying mechanisms of the effects of BC addition on CH4 emissions may vary with soil type, water regime, type of biomass feedstock and charring condition (temperature and heating time) (Spokas et al. 2009; Yu et al. 2013). Nevertheless, there have been few investigations into the rate of C mineralisation from BC compared to fresh organic material following application in two contrasting water management regimes, namely saturated conditions and flooded conditions. The objectives of this present study were to examine the effects of water regimes (continuously flooded and alternating saturated-flooded) and RS, BC and mineral fertilisers on GHGs, primarily CH4 and N2O. It was hypothesised that (1) alternating saturated-flooded conditions will increase the speed of RS decomposition and thus result in significantly lower CH4 emissions at the later stage compared to continuously flooded treatments; (2) due to C stability in BC, the effects of water regimes on CH4 emissions from BC will be less pronounced compared to RS and (3) N2O emissions rise as the rate of mineral fertiliser application increases.

Materials and methods

The trial was carried out in a controlled-environment growth chamber at the Faculty of Science, University of Copenhagen, Denmark. The temperature was maintained at 28 °C during the first month of the experiment, but was reduced to 25 °C during the day and 22 °C during the night. The light regime was 12 h a day (from 5 to 17 h), and the light intensity was 400 μE.

Soil column and gas sampling chamber design

Cultivation of rice and measurement of GHG emissions from the soil and rice system were conducted in cylindrical soil columns on which cylindrical headspace chambers could be mounted for gas sampling. The soil columns and gas sampling chambers were made of transparent Plexiglas (diameter 14 cm and height 30 cm for soil column; 60 cm high for gas sampling chamber). A ring frame with a water seal was placed on top of the soil column to ensure the headspace was air-tight. Inside the gas sampling chamber, two mini electric fans (4 cm, 12 V DC) were installed at the top and bottom, facing opposite directions, to ensure homogenous air mixture. Gas samples were collected through a rubber septum placed on the top of the chamber. The temperature inside the chamber was measured via a thermometer probe inserted through another rubber septum. The air-tightness of the chamber was tested using infrared gas analysis (IRGA) (WMA-2 CO2 analyser, PP Systems, UK), and CO2 leakage was found to be minimal (<1 %) over a deployment time of 2 h.

Experimental set-up, plant establishment and management

The experiment was a 2 × 3 factorial, completely randomised design, and the treatments included combinations of two irrigation regimes: continuously flooded (CF) and alternate wetting and drying (AWD), and three organic amendments (no amendment-control, RS and BC).

The soil used in this experiment is a rain-fed lowland rice soil from Cambodia, collected from the plough-layer (25 cm) on a farmer’s plot (11°3′38.69″N, 104°34′52.38″E) in Ta Suon village, Trapeang Thom Khang Cheung commune, Tramkak district, Takeo province. It has a sandy texture (70 % sand), is very acidic in reaction (pH 3.82) and has a low organic carbon content (5.2 g kg−1), low cation exchange capacity (CEC) and low nutrient availability (Table 1). The soil was classified as belonging to the Prateah Lang group (local classification) or Plinthustalfs (Bell and Seng 2004; Hin et al. 2005). This soil type accounts for about 28 % of soil used for rice production in Cambodia (Bell and Seng 2004).

Table 1 Soil, rice straw and biochar properties
Full size table

The soil was homogenously mixed and packed into 18 columns (3 kg each, dry weight basis). Rice straw and BC were applied as a basal amendment at a rate of 18.75 g dry weight per column (0.16 g total N) for RS and 6.25 g per column (0.09 g total N) for BC (Table 2). The application rate of RS is equivalent to 15 t ha−1 (soil bulk density 1.2 g cm−3 at 20 cm depth) (Table 2). The application rate of BC is 5 t ha−1, equivalent to 15 t of RS (based on a conversion factor of 3 kg straw to 1 kg biochar) (Table 2). Rice straw and BC were homogenously mixed with soil before being packed into the column. The RS used in this experiment was collected after the spring rice harvest in 2012 in Vietnam. Biochar was produced from RS by a pyrolysis process of dry RS (at 450 °C for 45 min in a closed iron container, externally heated by RS burning), with a char yield of 30 %. Then RS and BC materials were collected from Vietnam and transported to the laboratory at Copenhagen University for the experiment. Their properties are given in Table 1.

Table 2 Treatment details
Full size table

After packing the soil, tap-water (0.014 mg ammonium, 2.4 mg nitrate, 76 mg sulphate, 113 mg Ca, 4 mg K and 0.007 mg P L−1, pH 7.5, electrical conductivity 66 mS m−1) was added to each column, at saturated level in the AWD treatments and flooding (2 cm depth) in the CF treatments. During the first 2 weeks, the CF treatments received additional watering twice to keep them flooded, however no water was added to the AWD treatments during this period. Details on water regimes in both treatments, CF and AWD, are indicated in Fig. 1. Two weeks after watering, a single 14-day-old seedling of a Cambodia’s premium fragrant rice cultivar Phka Rumduol (Oryza sativa L.) was transplanted in each column of all treatments. This rice cultivar is a medium maturity variety, and it was awarded the “World’s Best Rice” for two consecutive years, during the World Rice Conference in Bali in 2012 and in Hong Kong in 2013. On the transplanting day, NPK mineral fertilisers (11–5–18) were applied (deep application) to all treatments. Mineral fertiliser was applied at a rate equivalent to 360 kg N ha−1 (0.45 g N per column) split into three doses: 50 % on the first application (during transplanting), 25 % on the second application (20 days after transplanting) and 25 % on the third application (37 days after transplanting). In the RS and BC-amended treatments, the amount of mineral nitrogen fertiliser was adjusted according to the level of N content in the RS, and BC applied to these treatments to ensure that the nitrogen amount of all treatments was similar. The total amount of mineral fertiliser was 232 kg N ha−1 (0.29 g N per column) for the RS treatment and 284 kg N ha−1 (0.36 g N per column) for the BC (Table 2). In terms of water regimes, after transplanting and fertilisation, the AWD treatments were flooded to the same level as the CF treatments. The AWD treatments were dried once (32 days after transplanting), which lasted for 5 days. Drying was via evapotranspiration, eliminating nutrient losses by drainage.

Fig. 1
figure 1

Variations in CH4 fluxes for the six treatments (top panel) and changes in redox potentials (bottom panel). The water regime is indicated in the middle, with saturated soil (hatched rectangle), flooding (grey rectangle) and drying (white rectangle), and (downwards arrow) indicates the mineral fertiliser application day. Error bars were removed for better visibility. CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar, MF mineral fertiliser

Full size image

Gas sampling, analysis and calculation

Gas samples were collected 28 times over the growth period, mostly once every 2–3 days during the first 2 months. The chamber deployment time was 40 min, during which three gas samples were collected (at time 0 and after 20 and 40 min). Gas samples were taken using a 60-ml syringe. Immediately after sampling, a used 3-ml vial (ExetainerLabco, UK) was flushed with 45–50 ml of the gas sample through two injection needles inserted in the septum. Samples were collected first from replicate 1 of each treatment (9 am), followed by replicates 2 (10 am) and 3 (11 am).

The concentrations of CH4, N2O and CO2 were simultaneously measured using a gas chromatograph (Bruker 450-GC 2011) equipped with detectors for CH4, N2O and CO2. Methane was determined using a flame ionisation detector (FID) at 300 °C, N2O was determined by electron capture detector (ECD) at 350 °C and CO2 was determined by thermal conductivity detector (TCD) at a temperature of 200 °C. The oven temperature was set at 50 °C. The carrier gas for CO2 and CH4 is He (99.999 %), and for N2O it is argon with 5 % CH4 at a flow rate of 60 ml min−1.

The gas fluxes were calculated using the following equation given by (Smith and Conen 2004):

$$F = \left( {\frac{\varDelta C}{\varDelta t}} \right)*\left( \frac{v}{A} \right)*\left( \frac{M}{V} \right)*\left( {\frac{P}{{P_{0} }}} \right)*\left( \frac{273}{T} \right),$$
(1)

where ∆C is the change in concentration of the gas of interest in time interval ∆t, v and A are the chamber volume and soil surface area, respectively, M is the molecular weight of the gas of interest, V is the volume occupied by 1 mol of the gas at standard temperature and pressure (22.4 L), P is the barometric pressure (mbar), P 0 is the standard pressure (1,013 mbar), and T is the average temperature inside the chamber during the deployment time (K).

The cumulative emission of CH4, N2O or CO2 over the rice-growing cycle was calculated from the integration of the area under the curve of each measurement point. The area between two adjacent intervals of the measurement days was calculated using the trapezoid formula as follows:

$$A_{\text{t(ab)}} { = }\frac{{\left( {t_{\text{b}} { - }t_{\text{a}} } \right) .\left( {F_{\text{ta}} { - }F_{\text{tb}} } \right)}}{ 2},$$
(2)

where A t(ab) is the area of the two adjacent intervals of the measurement days (between t a and t b), t a and t b are the dates of the two measurements, respectively, and F ta and F tb are the fluxes of the gas of interest at the two measurement dates, respectively.

Therefore, the cumulative emission of CH4, N2O or CO2 over the rice-growing cycle can be calculated using the following formula:

$${\text{Cumulative}}\;{\text{emission}}\;{\text{of}}\;{\text{CH}}_{ 4} , {\text{N}}_{ 2} {\text{O}}\;{\text{or}}\;{\text{CO}}_{ 2} { = }\sum {A_{\text{t(ab)}} }.$$
(3)

Global warming potential (GWP) over a hundred-year period was calculated by multiplying seasonal CH4 emissions by a factor of 25 and seasonal N2O emissions by a factor of 298 to convert them into CO2-equivalents. Yield-scaled GWP (CO2-equivalents per unit of paddy grain produced) was calculated by dividing CO2-equivalents by paddy grain produced.

Additional measurements

Four platinum electrodes were permanently installed (CF-RS, AWD-RS, AWD-BC and AWD-Control) to monitor the redox potentials under both water treatments. The redox measurements were taken on each of the sampling days. At the end of the experiment (124 days after transplanting), paddy grain at harvest maturity was harvested; the aboveground plant biomass was also collected, oven-dried and weighed.

Statistical analysis

Statistical tests were performed using SPSS 16 for Windows. The interaction effects of water and nutrient management were tested using two-way ANOVA. The simple main effect analysis was conducted to test the effects of water within each treatment (pair-wise comparison of treatments with the same nutrient amendment under different water managements, i.e. CF-RS vs. AWD-RS, CF-BC vs. AWD-BC and CF-Control vs. AWD-Control). One-way ANOVA was also performed to examine the effect of nutrient amendment under each water regime. Differences between treatment means under each regime were analysed using the least significant difference (LSD) method.

Results

Methane emissions and redox potentials

Daily fluxes of CH4 and the changes in redox potentials are depicted in Fig. 1. Methane emissions from all treatments were observed just 2 days after watering, when the redox potentials ranged from −18 to −133 mV. The first peak of CH4 emissions occurred at 5–9 days after watering soil for RS-amended treatments in both water regimes. Redox potentials during the peak emission were in an oxidised state (109 mV) for the AWD-RS treatment, but in a reduced state (−127 mV) in CF-RS. A significantly high CH4 emission from CF-RS at 9 days after watering was due to an extremely high CH4 emission from a replicate (replicate 1), which was up to 129 mg column−1 day−1. At 14 days after watering (1 day after rice transplanting), CH4 emissions from all treatments were similar, ranging from 0.07 mg column−1 day−1 to 0.50 mg column−1 day−1. After rice transplanting, CH4 emissions from the CF-RS treatment gradually increased, while CH4 emissions from the AWD-RS treatment lagged behind. However, CH4 emissions from the CF-RS and AWD-RS treatments reached the second peak at a similar period, which was at about 35 days after watering (22 days after rice transplanting). A significant drop in CH4 emissions from AWD-RS was identified during the drying period, with lagged emissions again after re-flooding. The gradual increase in CH4 emissions from CF-BC was observed only after about 65 days after watering (52 days after transplanting).

Within the first 14 days, carbon turnover was faster in treatments under the AWD water regime than those under the CF water regime (Fig. 2). Carbon mineralisation was highest in the RS treatments. Under both water regimes, C loss was mainly in the form of CO2-C, accounting for more than 90 % of the total C loss except in the CF-RS treatment where C loss via CH4-C accounted for about 51 % of the total C mineralisation (Fig. 2). As a ratio to total C (SOC + carbon from applied organic materials), C loss ranged from 0.11 % in the CF-BC treatment to 1.9 % in the AWD-RS treatment (Fig. 2).

Fig. 2
figure 2

C loss within 14 days. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05) and capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image

Total CH4 emissions during the experimental period are shown in Fig. 3. Total CH4 emissions were highest in the straw-amended treatments followed by BC, while the treatments without organic amendments had the lowest CH4 emissions. By comparing treatments with the same nutrient amendments in both water regimes, CH4 emissions in organic amendment treatments were lower in AWD than in the CF water regime, namely 44 % lower in the RS treatments (p < 0.01) and 29 % lower in BC treatments (p > 0.05). However, total CH4 emissions were higher under the AWD water regime for the treatments without organic amendment (Control).

Fig. 3
figure 3

Total CH4 fluxes. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05) and capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image

Nitrous oxide emissions

Except after fertilisation events, N2O fluxes were generally non-detectable during the experimental period (Fig. 4). After the first application of mineral fertiliser, the highest emission was found in CF-Control (about 0.48 mg N2O per column day−1). After the second and third applications, CF-BC had the highest emissions, ranging from 2.84 to 4 mg N2O per column day−1, respectively (Fig. 4).

Fig. 4
figure 4

Variations in N2O fluxes for the six treatments. Water regime is indicated at the bottom, with saturated soil (hatched rectangle), flooding (grey rectangle) and drying (white rectangle), and (downwards arrow) indicates mineral fertiliser application day. Error bars were removed for better visibility. CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar, MF mineral fertiliser

Full size image

The cumulative emission of N2O is depicted in Fig. 5. In all cases, the treatments under the CF water regime had higher total N2O emissions than the treatments under the AWD water regime. Under the CF water regime, the BC treatment had the highest total N2O emissions, which are 28 and 121 % higher than those of control and RS, respectively. Under the AWD water regime, the highest emissions were found in the control treatments, followed by the BC treatments.

Fig. 5
figure 5

Total N2O fluxes. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05), capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image

As a percentage of the N from mineral fertilisers, the total N2O–N emissions represented about 1.72 % on average, ranging from 0.75 % in AWD-RS to 3.41 % in CF-BC (Fig. 6).

Fig. 6
figure 6

N loss via N2O emissions, proportional to N from mineral fertiliser. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05) and capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image

Yield and yield-scaled global warming potential

Grain yields were slightly higher in treatments under AWD than under the CF water regime (p > 0.05). However, grain yields were very responsive to the mineral fertiliser application rates, most likely due to N limitation in the nutrient-poor soil used in the experiment. Under both water regimes, 87–99 % of the yield variation can be explained by an increased rate of mineral fertiliser application (Fig. 7). The lowest yield under RS-amended treatment could be attributed to two factors: the low amount of mineral N applied compared to the two other treatments, and N immobilisation leading to N limitation for crops. The BC-amended treatment received the same amount of total N (biochar + mineral N) as the control treatment (mineral N only) and BC, being a relatively inert material, is not expected to immobilise N. The lower BC yield can be explained by the low bioavailability of N in BC compared to the N applied through mineral fertiliser. The lower yield in the BC-amended treatment is therefore not due to the reduction effect of BC but due to lower plant available N compared to control treatment.

Fig. 7
figure 7

Grain yield by treatment and yield response to N fertiliser application rates. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05) and capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image

The GWPs were lower in treatments under AWD than in those under the CF water regime except in the control treatment (Fig. 8), and therefore the yield-scaled GWP was also lower in the treatments under the AWD water regime, namely 51 % in RS, 59 % in BC and 17 % in the control treatments (Fig. 9). While control treatments had the lowest GWPs (Fig. 8) but provided the highest yield (Fig. 7), the yield-scaled GWP under these treatments was also lower than the other treatments (Fig. 9). Under the CF water regime, the yield-scaled GWPs in the control treatment were 55 and 83 % lower than those in the BC and RS treatments, respectively. The yield-scaled GWPs in the control treatment under the AWD water regime were reduced by 9 % compared to the BC treatment and by 70 % compared to the RS treatment.

Fig. 8
figure 8

Global warming potentials. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05) and capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image
Fig. 9
figure 9

Yield-scaled global warming potentials. Error bars indicate standard errors of means. Small letters indicate significance of fertiliser treatments (p < 0.05) and capital letters indicate significance of water regime (p < 0.05). CF continuously flooded, AWD alternating wetting and drying, RS rice straw, BC biochar

Full size image

Discussion

Effects of organic amendments and water regime on CH4 emissions

The effect of organic matter application on CH4 emissions depends on the amount of the labile fraction in the applied and inherent soil organic matter at flooding (Yagi and Minami 1990). In the present study, CH4 emissions were found to rise with the increased rates of organic carbon under both water regimes (Fig. 10). However, organic C from fresh organic material (RS) had a greater effect on CH4 emissions than carbonised RS (BC). A unit increase in organic C from BC resulted in an increase of 0.051 units of CH4–C in the CF water regime and 0.014 units of CH4–C in the AWD water regime (data not shown). However, in the RS-amended treatments, CH4–C increased by a factor of 0.078 and 0.033 at a unit increase of organic C in the CF and AWD water regimes respectively (data not shown). The smaller effect of BC on CH4 emissions is due to its recalcitrance to microbial decomposition. Feng et al. (2012) reported that BC amendments increased methanotrophic proteobacterial abundances significantly, and decreased the ratio of methanogenic to methanotrophic abundances greatly. However, under the straw-amended treatment, they found that methanogenic archaeal growth was stimulated to a greater extent than that of methanotrophic proteobacteria, which resulted in the highest ratio of methanogenic archaeal abundance to methanotrophic proteobacterial abundance. As a result, BC amendment significantly decreased CH4 emissions from paddy soils compared with those from the traditional straw return treatment (Feng et al. 2012). In this present study, the results revealed that with the same level of organic C applied, CH4 emissions were reduced by 35 % in the CF water regime and 58 % in the AWD water regime by applying BC instead of RS. Knoblauch et al. (2011) documented significantly faster mineralisation of organic carbon when rice husk was applied untreated compared to charred rice husk, and estimated that field CH4 emissions may be reduced by as much as 80 % by charring of rice residues and adding the black carbon obtained to paddy fields instead of incorporating untreated harvest residues. Dong et al. (2013) reported a reduction in CH4 emissions by 47–86 % from applying BC instead of RS.

Fig. 10
figure 10

Linear regressions between CH4–C emissions and organic carbon. Error bars indicate standard errors of means. CF continuously flooded, AWD alternating wetting and drying

Full size image

Comparing water treatments, there was a lesser effect of organic C on CH4 emissions in the AWD water regime than in the CF water regime for organic amended treatments (a reduction of 44 % in RS and 29 % in the BC treatments). This lesser effect can be attributed to faster C mineralisation (Fig. 2) which could be the result of better aeration during the first 2 weeks, followed by 5 days of soil drying (at 46 days after wetting). This can also be explained by the fact that CH4 emissions in the AWD treatments developed slowly after the soil was flooded (Fig. 1) as a result of there being less carbon available in the system to provide a substrate for methanogenesis when flooding. The same patterns were observed in both the field experiment (Ly et al. 2013) and another laboratory experiment by the authors (unpublished data). In general, the CH4 emissions were associated with low redox potentials. During the water-saturated condition and drying period under AWD treatments, the redox potentials were generally in oxidised state, and CH4 emissions under AWD treatments were lower than those of CF treatments, where redox potentials quickly dropped to about −130 mV after a week of flooding (Fig. 1). This Eh value is close to critical redox potentials for CH4 production in Chinese paddy soil reported by Yu et al. (2001). After flooding, the redox potentials under AWD treatments also gradually declined, and CH4 emissions kept increasing until the drying period during which the redox potentials were at the oxidised state. After re-flooding, the redox potentials quickly reverted to reduced condition, and CH4 emissions started to increase (Fig. 1). The first flooding leads to a slower decline in redox levels, perhaps due to Fe or sulphate reduction, while at the second flooding, the decline is rather controlled by available C (fastest in RS). A delay in CH4 emission after second flooding may partly be caused by a saturation phase where CH4 concentration in soil water increased before emissions became significant.

Effects of organic amendments and water regime on N2O emissions

Nitrous oxide can be produced in soil through several microbial processes: by aerobic nitrification of ammonium (NH4 +) and by anaerobic denitrification of NO3 − (Smith et al. 2003; Yu et al. 2001). Availabilities of C and nitrate–N are among the predominant factors governing N2O emissions (Baggs and Philippot 2010; Bremner 1997; Maljanen et al. 2003; Meng et al. 2005). Microorganisms require readily decomposable C before reduction of NO3 − can occur, whereas limited N availability limits the availability of precursor for nitrification and subsequent denitrification ultimately suppressing N2O emissions (Swerts et al. 1996; Weier et al. 1993). The increased N2O emissions with the increased rate of N applied via mineral fertiliser could be therefore stimulated by high N availability in those treatments. Although high N rate was applied, most of the NO3 − could be likely taken up by the plants before it is denitrified into N2O. This could explain why N2O was observed only after fertilisation event and lasted only for few days.

The exceptionally high total N2O emissions in CF-BC could be attributed to the sampling time. After the second and third applications of mineral fertilisers, N2O emissions were highest with the CF-BC treatment. It is suspected that peak N2O emissions after N fertilisations may have been missed in treatments other than CF-BC. Had sampling been performed half a day earlier, the dynamics might have been different. It appears that the N2O emissions were lower in the RS treatments in both water regimes. This could be explained by the high C/N ratio in this treatment, resulting in greater immobilisation instead of mineralisation.

While N2O was detected after fertilisation events after which soil was flooded (first and third application of mineral fertiliser) and during flooding (second application of mineral fertiliser), the redox potentials during the N2O emission detection were generally in reduced conditions, but it ranged from +329 to −210 mV (Figs. 1 and 4). Yu et al. (2001) reported that maximum N2O emission was at the redox potentials ranging between +120 and +250 mV, and that little N2O emission occurred at redox potential values higher than +250 mV or lower than +120 mV. In this present study, significant N2O emissions (>1 mg N2O per column per day) were found when soil redox potentials were between −123 and −210 mV (Figs. 1 and 4). The result is contradictory to the general assumption that N2O emission is negligible when soil is flooded (Xing 1998). However, Xing et al. (2002) also reported that N2O emissions were observed under continuous flooding where the redox potentials were in reduced condition.

Conclusions

This study showed that total CH4 emissions rose with the increased rates of organic carbon under both water regimes. However, the effects of organic C from fresh organic material (RS) on CH4 emissions were more pronounced than that from carbonised RS (BC), which therefore revealed the benefits of applying BC instead of RS. The water regime also had a remarkable effect on carbon mineralisation. The mineralisation of organic carbon (RS and BC) was much faster when applied in saturated conditions instead of flooding. This resulted in lower total CH4 emissions in treatments under AWD than those under the CF water regime, namely 44 % lower in the RS treatments and 29 % lower in the BC treatments. Nitrous oxide fluxes were generally non-detectable during the experimental period except after fertilisation events. Global warming potentials are lower in the treatments under AWD than those under the CF water regime except in the control treatment with only mineral fertiliser application. Grain yields were slightly higher in treatments under AWD than under the CF water regime, and hence the yield-scaled GWP was also lower in the treatments under the AWD water regime. These results have revealed the potential benefits of using carbonised RS in combination with mineral fertiliser and a modified water regime to reduce GHG emissions without compromising rice yield. However, the potential economic and environmental benefits from using carbonised RS in combination with mineral fertiliser and a modified water regime should be tested at the field scale considering farmers’ specific economic and agro-ecological setting.