Introduction

Eddy covariance (EC) measures the fluxes of greenhouse gas (GHG) that exchange in between the rice-rice ecosystem and atmosphere (Baldocchi 2003). The net ecosystem exchange (NEE) of carbon dioxide (CO2) of the rice-rice system can be measured using the EC data over a large area (Chapin et al. 2006). Carbon dioxide is an important driver controlling carbon cycle. The carbon cycle is governed by photosynthetic fixation of atmospheric CO2 and its release by autotrophic and heterotrophic respiration. Basically, NEE is influenced by several factors like soil organic carbon (SOC) content, plant water supply, leaf area index, type of crop, and management practices (Smith et al. 2010; Reichstein et al. 2005). The SOC accumulates in flooded rice soil due to the oxygen deficiency limited by slow diffusion in flooding water. Thus, waterlogged rice soil acts as a store house of carbon (Joosten and Clarke 2002). During the fallow period, the stored carbon in soil may be oxidized by microorganism (Maljanen et al. 2001). Emission of methane (CH4) from the rice field is gaining attention globally as it is the second largest radiative forcing GHG after CO2 (Forster et al. 2007). Irrigated lowland rice paddies are one of the principal anthropogenic sources of CH4, representing approximately 5–10% of the worldwide CH4 source (Kirk and Bickert 2004; Matthews et al. 2000). In lowland irrigated rice, flooding creates anaerobic soil condition which helps CH4 production from the decomposition of both native and externally added organic materials (Neue 1993; Wassmann et al. 2000). In an estimate, it was observed that methane emission was 500–600 Tg CH4 year−1globally (Dlugokencky et al. 2011). Therefore, diminishment of the global CH4 sources offers possibilities for curtailing the rising trend in the global warming at a short time (Dlugokencky et al. 2011). Also, nitrous oxide (N2O) has higher global warming potential. Nitrogenous fertilizer applied to crops is the predominant source of N2O emission that accounts for approximately 75% of direct emission from agricultural soil (Zheng et al. 2004).

Eddy covariance techniques can be used to measures various components of energy balance like latent heat flux (LE), sensible heat flux (Hs), soil heat flux (G), net radiation (NR), etc. (Kristensen et al. 1997). Because of standing water, the lowland rice field is different from other aerobic crops. Substantial effect of energy partitioning into Hs and LE was noted in rice; for instance, in standing water, LE is around 65–79% of net radiation (NR), while Hs is about 8% of NR (Masseroni et al. 2015). Photosynthetically active radiation (PAR) is mainly the solar radiation energy in between 400 and 700 nm wavelength which is used by plants for photosynthesis (Mitchell et al. 1998; Sinclair and Muchow 1999). The crop duration, growth, and some agronomic characters of rice are influenced by inadequate PAR. For example, prolonged period of log light exposure increases plant height and growth and enlarged the leaf surface (Ren et al. 2002; Liu et al. 2009).

About 90% of the global rice production is contributed by the Asian countries (Masseroni et al. 2015) like India, China, the Philippines, Japan, Bangladesh, etc. In India, rice is grown in 43.95 M-ha areas with an annual production of 106.54 Mt (GOI 2014). Rice-rice cropping system is the dominant system followed in the coastal region of eastern India. This system largely contributes in GHG production which is influenced by various edaphic conditions, agricultural management practices, and soil properties (soil enzymes and carbon pools). For understanding the dynamics of CO2 and CH4 in rice paddies, knowledge of soil enzymes and labile pool of carbon are required to draw a meaningful interrelation of gaseous emission with the chemical and microbiological properties of soil (Ling et al. 2011). For instance, the enzymes like dehydrogenase and β-glucosidase are involved in carbon cycling and reflect the microbial activity in soil organic matter decomposition (Das and Varma 2011). Similarly, fluorescein diacetate activity indicates the active microflora which is involved in producing extracellular enzymes (Adam and Duncan 2001).

For several decades, climate change studies using eddy covariance data focused only on the exchange of carbon dioxide (Alberto et al. 2009; Bhattacharyya et al. 2014), methane (Alberto et al. 2014; Detto et al. 2011), surface energy fluxes, and balance (Masseroni et al. 2015; Tsai et al. 2007) under flooded and aerobic rice ecology throughout the world. All these studies have explained the component of mass and/or energy exchange using eddy covariance, but limited information is available on the factors affecting gaseous exchange and GWP in tropical paddy.

In the earlier work from India, it was demonstrated how NEE, GPP, and RE were governed by net radiation, vapor pressure deficit, soil moisture, soil temperature, soil heat flux, and air temperature (Bhattacharyya et al. 2013a). Seasonal carbon balance was also determined using the data of CO2 and CH4 obtained from EC (Bhattacharyya et al. 2014; Swain et al. 2016). Keeping the above facts in mind, the current study was conducted (i) to study the daily, seasonal, and annual variations in energy component and mass exchange in rice-rice ecology and (ii) to understand the effect of environmental (e.g., Hs, LE, PAR, NR) and biological (e.g., soil enzymes and LC pools) factors on the gaseous exchange and GWP in rice soil.

Materials and methods

Location for measurement

The study site is located in the farm of the National Rice Research Institute (NRRI) Cuttack, Odisha, India (20° 27′ 6″ N latitude, 85° 56′ 25″ E longitude and 24 m above mean sea level). The yearly mean temperature was 27.7 °C and mean annual rainfall was 1300 mm during the period of experimentation. The study site soil is sandy clay loam in texture (24.5% clay, 22.4% silt, 53.1% sand) and characterized as Aeric Endoaquept with bulk density 1.39 Mg m−3, pH (1:2.5 soil:solution ratio) 6.19, electrical conductivity 0.40 dS m−1, total C 11.0 g kg−1, and total N 0.78 g kg−1.

Establishment of crop

Rice was grown in the dry season (DS) and wet season (WS) of 2015 followed by two fallow periods (dry fallow, DF and wet fallow, WF). Twenty-one-day-old rice seedlings (cv. Naveen in DS and cv. Swarna-Sub 1 in WS) were transplanted to a puddled soil with 20 cm × 15 cm spacing. Transplanting was done on the 2nd week of July in WS and 1st week of January of DS and harvested in November and last week of May, respectively. Nitrogen (N) was applied through urea in three equal splits at the basal, maximum tillering (MT), and panicle initiation (PI) stages. About 80 and 100 kg N ha−1 were applied in the WS and DS, respectively. Phosphorus (P) and potassium (K) were added at the rate 40 kg ha−1 each at basal during land preparation in both seasons. However, the observation site was kept flooded with 6 ± 2 cm standing water throughout the crop growth and drained out 15 days before the harvest.

Observation site for CO2 and CH4 flux

The EC was installed at the center of a lowland flooded rice field covering a fetch area of 2 ha. The main components of the EC system were a sonic anemometer (CSAT3, M/s Campbell Scientific Corp.) for 3D wind speed and sonic temperature and an open-path infrared gas analyzer (LI-7500A, M/s LICOR Inc.) for flux of CO2 and water vapor. A data logger (CR3000, M/s Campbell Scientific Corp.) was used to record the flux data. The infrared gas analyzer and sonic anemometer were installed on a tripod aluminum mast at the height of 1.5 m. The LI-7500A was sloped 15° of the head portion from the vertical direction to minimize the entering of precipitation inside the window. The LI-7500A and CSAT3 were directed to the prevalent southwest wind direction (Campbell Scientific, Inc 2009; LICOR, Inc 2011). The NEE was calculated by cumulating the CO2 flux (flux of CO2, Fc) and CO2 storage change (Fs). The Fs was ignored due to the low height of the rice canopy. The mean vertical flux density of CO2 was obtained by using the formula (Webb et al. 1980; Baldocchi 2003):

$$\mathrm{Fz}={\rho}_{\mathrm{a}}\overline{\omega^{\prime }{C}^{\prime }}$$
(1)

Where Fz is the mean vertical flux density of CO2, ρa is the density of air (kg m−3), ω′ is the 30-min covariance between vertical fluctuations in wind speed, and C′ is the CO2 mixing ratio; time averaging was denoted by overbars, whereas the primes presented as fluctuations of mean value. A positive value of the covariance indicated net transfer of CO2 to the atmosphere, while a negative value signified net absorption of CO2 by the photosynthetic flora. Seasonal net exchange was calculated from the half an hour mean NEE data (Tanner and Thurtell 1969). Methane flux was measured using an open-path CH4 analyzer (LI-7700, M/s LICOR Inc., USA) installed at a height of 1.5 m above the soil surface. This sensor contained a diode laser in an open Herriott cell configuration using wavelength modulation spectroscopy (McDermitt et al. 2011; Alberto et al. 2014). Output of the methane sensor was logged into a CR3000 data logger (M/s Campbell Scientific Corp., Canada).

Measurements of heat fluxes and temperature

The Hs and LE were measured by the EC system employing three-axes sonic anemometer and open-path CO2 and H2O infrared gas analyzer. Turbulent fluxes of Hs and LE were calculated from covariance between a respective scalar and the vertical wind (w) as (Aubinet et al. 2000):

$$\mathrm{Hs}={\rho}_{\mathrm{a}}{C}_{\mathrm{pa}}\overline{T^{\prime }w^{\prime }}$$
(2)
$$\mathrm{LE}=L\kern0.5em {\rho}_{\mathrm{a}}\overline{q^{\prime }w^{\prime }}$$
(3)

where ρa (kg m−3) is the density of air, Cpa = (1004 J kg−1 K−1) is the specific heat of air at constant pressure, and L = (2.256 × 106 J kg−1) is the latent heat of vaporization of water. The w′, T′, and q′ are the deviations from time means of vertical wind speed (ms−1), virtual temperature (K), and the water vapor mixing ratio (mmol mol−1), respectively.

The 4-component radiation sensor (CNR4, KIPP, and ZONEN, M/s Campbell Scientific Corp.) was used to take reading of NR. Simultaneously, a soil temperature probe (107 B, Campbell Scientific Corp.) was used to measure soil temperature at a depth of 5 cm.

Quality control, gap filling, and processing of flux data

The raw data collected from EC were checked for quality (Mauder and Foken 2011). The raw flux data was cleaned by removing the spikes (Vickers and Mahrt 1997). All the corrections like planar fit coordinate rotation, conversion of buoyancy, high-frequency response losses, and density fluctuations were done (Wilczak et al. 2001; Liu et al. 2001; Webb et al. 1980; Aubinet et al. 2000). After this, the data were processed using the EdiRe software (University of Edinburgh). The U* filtering method was employed for removal of spikes that arise due to unsuitable environmental conditions (Reichstein et al. 2005; Papale and Valentini 2003; Bhattacharyya et al. 2014). The “look-up” table approach was used for gap filling of lost and discarded data (Falge et al. 2001).

Flux partitioning of NEE

The NEE denotes the net gain or loss of carbon of an ecosystem and it is determined by deducting the loss of carbon by autotrophic (Ra) and heterotrophic (Rh) respiration from GPP of autotrophic organisms. The rectangular hyperbola (Ruimy et al. 1995) method was used for flux partitioning of NEE into GPP and Re:

$${\mathrm{NEE}}_{\mathrm{night}+\mathrm{day}}=-\left[\left(\alpha \beta Q/\alpha Q+\beta \right)\right]+\gamma$$
(4)

Where α is the apparent ecosystem quantum yield, β is the hypothetical maximum GPP uptake, γ is the estimated Re, and Q is the PAR (μmol m−2 s−2). The GPP was calculated by using the equation:

$$\mathrm{NEE}=\mathrm{GPP}-\operatorname{Re}$$
(5)

Nitrous oxide measurement

The EC system used in this study had no separate sensor for measuring N2O; hence, the N2O flux was measured by the manual closed chamber method near the eddy fetch area throughout the year by following the standard procedure (Bhattacharyya et al. 2012; Datta et al. 2009; Chatterjee et al. 2018). Measurements were initiated at the start of the season and continued until the end of whole cropping season including both the fallows. The flux data were taken every 3–7-day intervals; however, during fertilizer application, data were collected 1 day before and 1 day after the fertilizer application. The N2O concentration was measured in a gas chromatograph (Chemito 2000, M/s Thermo Scientific) furnished with an electron capture detector (ECD). The linear interpolation technique was used to fill the gap (Datta et al. 2009). Subsequently, the cumulative and seasonal mean N2O emission was figured graphically.

GWP measurement

The GWP is used to assess a GHG to absorb heat in the atmosphere comparative to CO2. Integrated assessment of GHG emissions articulated as GWP was calculated using the formula: GWP = 24.5 * CH4 + CO2 + 320 * N2O kg CO2 equivalent ha−1 from CH4, CO2, and N2O flux values (Bhattacharyya et al. 2012).

Soil enzymatic activities and soil LC pools

Soil enzymatic activities and LC pools were determined at MT, PI, FL (flowering), and H (harvesting) stages. Soil samples were collected in polythene bags from five random points (0–15 cm depth) using an auger near the fetch area of the EC system. Then, the soil samples were kept in a refrigerator at 4 °C. Dehydrogenase (DHA), fluorescein diacetate (FDA) hydrolysis, and β-glucosidase (β-GLU) activities were determined in the soil samples by the procedure described by Casida et al. (1964), Adam and Duncan (2001), and Eivazi and Tabatabai (1988), respectively. Modified chloroform fumigation-extraction method was used for soil microbial biomass-C (MBC) measurement (Witt et al. 2000). Readily mineralizable carbon (RMC) was extracted by 0.5 M K2SO4 followed by estimation using the wet digestion method (Inubushi et al. 1991; Vance et al. 1987).

Results

Daily, seasonal, and annual variations in NEE and NEME

The mean NEE varied from − 5.71 to 2.29 g C m−2 day−1 during the monitoring period covering both crop seasons and fallow periods (Fig. 1a). The mean NEE was lower in the DF (− 1.49 g C m−2 day−1) than in the WF (− 0.29 g C m−2 day−1). Contrastingly, mean NEE during the DS (− 1.62 g C m−2 day−1) was higher than the mean NEE during the WS (− 1.83 g C m−2 day−1). Annual mean NEE calculated by averaging the crop seasons and fallow periods was recorded − 1.31 g C m−2 day−1 (Table 1). The mean value of NEE was found negative in all the cases. Cumulative NEE values for the DF and WF were − 95.37 and − 14.81 g C m−2, whereas cumulative mean NEE values during the DS and WS were − 199.73 and − 232.55 g C m−2, respectively. The annual cumulative NEE was observed − 542.47 g C m−2 (Table 1). It was also observed from our study that the NEE became more negative with the increase of PAR and NR (Table 2).

Fig. 1
figure 1

Daily average of net ecosystem exchange such as a carbon dioxide (NEE) and b methane (NEME) by eddy covariance technique in the year 2015

Full size image
Table 1 Monitoring of net ecosystem carbon dioxide exchange (NEE), net ecosystem methane exchange (NEME), nitrous oxide, and GWP during the crop growth stage including both the dry and wet seasons and fallow period
Full size table
Table 2 Correlation coefficients between global warming potential (GWP), net radiation (NR), photosynthetic active radiation (PAR), latent heat (LE), and sensible heat flux (Hs)
Full size table

Daily mean NEME varied from 0.12 to 8.85 mg CH4 m−2 h−1 during the study period (Fig. 1b). Mean NEME values recorded during the DF and WF were 1.47 and 1.99 mg CH4 m−2 h−1, respectively. Mean of NEME was 2.86 mg CH4 m−2 h−1 during the DS and it increased to 4.19 mg CH4 m−2 h−1 during the WS. Annual mean NEME during the study period was 2.62 mg CH4 m−2 h−1. The cumulative NEME was 2259.81 mg CH4 m−2 in the DF, while the value was a little higher (2447.87 mg CH4 m−2) during the WF (Table 1). Cumulative NEME values during the DS and WS were 8434.11 and 12,781.96 mg CH4 m−2, respectively. The annual cumulative NEME values during the study period were 25,923.8 mg CH4 m−2 (Table 1). The highest methane flux in a day (5.82 mg m−2 h−1 in the DS and 8.32 mg m−2 h−1 in the WS) was observed during the MT to PI stages of crop development (Fig. 1b). NEME had significant and positive correlation with LE in both seasons (Table 2).

Daily, seasonal, and annual variations in N2O efflux

Daily mean N2O flux varied from 4.17 to 52.96 μg N2O-N m−2 h−1 throughout the crop seasons including the two fallow periods (Fig. 2). Mean N2O-N flux values during the fallow periods after the harvest of the crop were 10.06 μg N2O-N m−2 h−1 (DF) and 5.37 μg N2O-N m−2 h−1 (WF), while the flux values were quite high during the DS (27.51 μg N2O-N m−2 h−1) and WS (23.82 μg N2O-N m−2 h−1). Cumulative N2O fluxes of both fallow (DF and WF) periods were 644.07 and 273.97 μg N2O-N m−2, respectively (Table 1). The annual mean N2O flux was 16.69 μg N2O-N m−2 h−1 (Table 1). Three peaks of N2O-N efflux following the application of urea were observed both in the dry and wet seasons (Fig. 2).

Fig. 2
figure 2

Daily average nitrous oxide (N2O-N) flux by the manual chamber method in the year 2015

Full size image

Seasonal variations in GWP

The cumulative GWP was observed 5803.63 kg CO2 equivalent ha−1 during the DS, which was less as compared to its value (7213.16 kg CO2 equivalent ha−1) during the WS (Table 1). The cumulative GWP in DF and WF was 2201.02 and 1264.24 kg CO2 equivalent ha−1, respectively. The highest GWP (101.96 kg CO2 equivalent ha−1) was observed on the 109th day of the year during the DS which coincides with the H stage of rice crop, whereas in the WS, the highest GWP (111.98 kg CO2 equivalent ha−1) was recorded on the 241st day of the year that falls on the FL stage of rice growth (Fig. 3). Mean GWP at the different growing seasons of rice crop (DS and WS) was correlated with NR, PAR, LE, and Hs (Table 2). Mean GWP exhibited a positive and significant correlation (p < 0.01) with NR, PAR, and LE during both seasons (DS and WS), while a negative significant correlation (p < 0.01) with Hs during the DS (table not shown).

Fig. 3
figure 3

Global warming potential in CO2 equivalent emission in the year 2015

Full size image

Daily and seasonal variation of NEE, GPP, and Re

Over the season, daily variations of mean NEE varied from 2.21 to − 5.51 μmol CO2 m−2 s−1 throughout the cropping season, and on the basis of mean diurnal NEE, maximum CO2 assimilation or uptake by rice crop was found at 12.00 h and maximum emission was observed at 4.30 h (Fig. 4). During the dry season, NEE was lowest (mean − 0.63 μmol CO2 m−2 s−1) in the vegetative stage, then gradually decreased, i.e., MT to PI stages (mean − 1.16 μmol CO2 m−2 s−1) and further decreased value (mean − 2.04 μmol CO2 m−2 s−1) while heading to the FL stage and then it attains its lowest value (mean − 2.33 μmol CO2 m−2 s−1) during the harvest period (Table 3). The GPP and Re varied from − 8.11 to − 18.92 and 7.83 to 15.45 μmol m−2 s−1, respectively, at different growth stages of the DS. The highest GPP (− 18.92 μmol m−2 s−1) and Re (15.45 μmol m−2 s−1) were recorded at the H stage of the DS. Similarly, during the WS, the GPP and Re varied from − 8.1 to − 18.92 μmol m−2 s−1 and 7.83 to 15.45 μmol m−2 s−1, respectively. The highest GPP (− 19.62 μmol m−2 s−1) at the harvest period and Re (16.70 μmol m−2 s−1) at the FL stage were recorded during the WS (Table 3). In both crop seasons, the magnitude of Re increased with the advancement of crop growth stages. The GPP was decreased with the advancement of rice growth during both crop seasons. The GPP of the WF is higher than the GPP of the DF and vice versa for Re (Table 3).

Fig. 4
figure 4

Partitioning of daily average net ecosystem exchange (NEE), gross primary production (GPP), and respiration (RE) in days of the year 2015

Full size image
Table 3 Mean NEE, GPP, and Re of the entire crop growing season including the fallow periods
Full size table

Daily, seasonal, and annual variations in Hs, LE, PAR, and NR

The LE was observed as the dominant component of energy balance in the rice-rice ecosystem. Mean daily LE and Hs ranged from 0.13 to 16.65 MJ m−2 day−1 and from − 4.11 to 4.76 MJ m−2 day−1 during the monitoring period (Fig. 5a, b). The magnitude of mean Hs was in close proximity during the fallow periods (1.45 MJ m−2 day−1 in the DF and 1.64 MJ m−2 day−1 in the WF), while the LE during the DF (8.60 MJ m−2 day−1) was higher than the LE in the WF (2.61 MJ m−2 day−1) (Table 4). The mean Hs was lower than LE during the DS and WS (Table 4). The mean PAR during the DS (743.43 μmol m−2 s−1) was 14.8% higher than the PAR during the WS (647.53 μmol m−2 s−1), whereas NR in the DS (136.97 W m−2) was only 4.6% higher than the mean NR in the WS (130.87 W m−2) (Fig. 5c, d). A negative and highly significant correlation of NEE was observed with LE (− 0.367), NR (− 0.497), and PAR (− 0.456), while a significant and positive correlation was observed with Hs (0.133) throughout the year (table not shown).

Fig. 5
figure 5

Daily heat fluxes of different energy balance such as a latent heat, b sensible heat, c photosynthetically active radiation, and d net radiation

Full size image
Table 4 Mean and integrated Hs (sensible heat) and LE (latent heat) during the monitoring period including both the dry and wet seasons and fallow period
Full size table

Soil enzymatic activities and LC pools at different growth stages

Soil enzymes such as FDA, DHA, and β-GLU showed variation in crop growth stages. The FDA was highest at PI (23.84 μg fluorescein g−1 h−1). A similar trend was observed in the case of DHA and β-GLU. DHA and β-GLU were lowest at the MT stage (161.96 μg TPF g−1 day−1 soil, 19.27 μg p-nitrophenol g−1 day−1) followed by the FL (379.57 μg TPF g−1 day−1, 43.38 μg p-nitrophenol g−1 day−1) and PI (478.82 μg TPF g−1 day−1, 60.20 μg p-nitrophenol g−1 day−1) stages (Table 5). Soil labile C pools are represented by MBC and RMC. The RMC was highest in the PI stage (132 μg C g−1 during the DS and 79 μg C g−1 during the WS) (Table 5). Similarly, the MBC was highest in the PI stage (147.10 μg C−1 g−1).

Table 5 Soil enzymatic activities with soil carbon showing in different rice crop growth stages in lowland tropical rice
Full size table

The GWP and NEME showed a similar trend as soil enzymes and labile carbon pools in both seasons (except GWP at the harvesting stage in the wet season) (Fig. 6a). The N2O flux during the dry season decreased from the PI to H stages. A similar trend was recorded in soil enzymes and labile carbon pools during those stages. The mean NEE exhibited a more negative value with a decrease in labile pools from the PI to H stages in the wet season (Fig. 6b).

Fig. 6
figure 6

Stagewise variation in soil enzymatic activities, LC pools, and gaseous exchange in the a wet season and b dry season (VS: vegetative stage, PI: panicle initiation, FL: flowering and H: harvesting). Unit for DHA: μg TPF g−1 day−1, FDA: μg fluorescein g−1 h−1, β-GLU: μg p-nitrophenol g−1 day−1, RMC and MBC: μg C g−1, NEE: g C m−2 day−1, NEME: mg CH4 m−2 h−1, N2O: μg N2O-N m−2 h−1, N2O: μg N2O-N m−2 h−1, GWP: kg CO2 ha−1

Full size image

Discussion

NEE as influenced by flooded rice ecology

Positive and negative values of NEE were observed during the study period. A positive value of NEE indicated net CO2 emission into the atmosphere, while the negative NEE signified the net CO2 uptake by the rice. During the fallow periods (DF and WF), NEE was slightly higher than the NEE during maturity of crop (Fig. 1a), which indicated slow but steady rates of organic matter decomposition at fallow (McMillan et al. 2007). Once the floodwater receded, the supply of labile fermentation products increased due to the rapid mineralization of leftover organic residues under aerated condition. Higher soil temperatures (as no standing water) during the fallow periods increased in daily CO2 efflux further (Miyata et al. 2000; McMillan et al. 2007). However, NEE in lowland rice ecology was mainly governed by various physical, physiological, and biological parameters (Nair et al. 2011; Saito et al. 2005). It was observed from this study that the daily variation in NEE increased with the progress of growth stages and reached its maximum around the PI to FL stages (Swain et al. 2016). In both the DS and WS, the flooded rice paddy behaved as a net CO2 sink on a daily basis from vegetative to early maturity stage. During the WS, NEE was lowest at the early vegetative stage and shows a peak at the reproductive stage due to greater physiological activity and with elevated photosynthesis rate (Bhattacharyya et al. 2013b). More negative NEE was observed during the day due to higher CO2 assimilation by photosynthesis (Alberto et al. 2009) which suggested that this was because of gas exchange among the leaves and the pattern of light interception at the canopy level (Nair et al. 2011).

Driving factors of methane exchange under flooding condition of rice ecology

The methanogenesis process in general is influenced by redox potential (Eh − 200 mV), LC content, pH, moisture content, and temperature (Guo and Zhou 2007). NEME was found to be higher during active tillering (AT) to PI stage of rice crop in both seasons (Fig. 1b) due to higher methanogenic activity and the presence of a higher quantity of substrate particularly labile pools of soil carbon (Tseng et al. 2010). As the soil remained waterlogged during most of the year in the present experiment, the conditions were favorable for methanogenesis. In a water-saturated condition, high methane fluxes were observed, when favorable temperature prevailed and rice rhizosphere was most active and well-developed (Bhattacharyya et al. 2014). Reduction of NEME was observed during the maturity stage, which continued during the fallow period. This may be due to the release of labile substrate by the root that controls CH4 production up to the flowering of the rice crop and depletes afterwards (Lu et al. 2002). Coupling with less substrate and enhanced CH4 oxidation at the later growth stages of rice limits the CH4 emission (McMillan et al. 2007). At same stage of crop growth, methane emission was higher in high-temperature condition. For example, the methane emission during the PI to FL was higher in the wet season (average methane 4.19 mg CH4 m−2 h−1, average temperature 32.21 °C) than the methane emission during the DS (average methane 2.85 mg CH4 m−2 h−1, average temperature 28.02 °C). The reason of such trend may be due to an increase in plant-mediated CH4 transport with increasing temperature. Moreover, the increase in soil temperature also increases molecular diffusion and ebullition of methane (Alberto et al. 2014). It was observed that methane efflux increased with increase in LE with the season in the study which was well supported with the observation reported by Bhattacharyya et al. (2014). The LE might have an influence on daily variation of CH4 flux in rice paddy because the dissolved CH4 in soil solution could get released to the atmosphere through transpiration flux (Nisbet et al. 2009).

Daily, seasonal, and annual variations in N2O efflux

Application of nitrogenous fertilizer such as urea in the present study significantly increased the N2O flux. The N2O emission from flooded rice soils is mainly attributed to two major microbial processes, i.e., nitrification and denitrification. Both nitrification and denitrification processes are taking place concurrently in lowland rice soils, although the first is aerobic and another is anaerobic process and are considered as a major source of N2O emission (Kyaw and Toyota 2007). Extrapolation of daily N2O-N flux data measured at closed intervals by the manual close chamber method during the entire cropping period was well studied by various researchers (Bhatia et al. 2011; Das et al. 2011; Datta et al. 2009; Bhattacharyya et al. 2012). Emission of N2O is a microbial-mediated process that depends on available C and N sources which are available to microorganisms (Nayak et al. 2007). Distinctly higher N2O emissions were recorded whenever application of urea was done in the field (Fig. 2). Three peaks of N2O efflux were recorded immediately after urea application in both seasons. It was observed that the amplitude of N2O flux was more during the dry season as compared to that during the wet season. This may be due to more PAR and NR required during the DS as compared to the WS (Bhatia et al. 2005; Bhattacharyya et al. 2012).

GWP measurement

The GWP was found higher when methane and nitrous oxide emissions recorded higher values. This is evident from the basic calculation of GWP where higher weightage was given to CH4 and N2O as compared to CO2 (IPCC 2007). Application of N fertilizer resulted in enhanced N2O emission (Fig. 2) which resulted to increased GWP of paddy soil (Fig. 3) (Bhatia et al. 2005). Particularly, the stages of rice crop such as AT to PI where CH4 emission was more, GWP values also increased (Bhattacharyya et al. 2013a, 2014). The GWP exhibited a positive and significant correlation with NR, PAR, and LE. As NR and PAR increase, the available heat energy helped the soil to warm up which in turn enhanced methanogenic activities which might have favored more decomposition of organic matter in soil and produced more CO2 and CH4 that might have increased GWP (Nisbet et al. 2009; Tseng et al. 2010).

Partitioning of NEE into GPP and RE

The GPP and Re which were partitioned from NEE showed a gradual increasing trend from early vegetative to maturity period of the rice crop. Rice behaves as a source of CO2 during nighttime and as a sink of CO2 at daytime. It was due to only respiration in the nighttime as photosynthesis is absent and greater Re than the GPP during early morning (Tseng et al. 2010). In the daytime, CO2 was assimilated by vegetation through photosynthesis and emitted by the soil through microbial respiration (Bhattacharyya et al. 2013b). At nighttime’s lowland rice paddy ecosystem, respiration (Re) led to a net efflux of CO2 to the atmosphere (Swain et al. 2016). Higher GPP during ripening to maturity stage is due to the fact that more CO2 was assimilated into the plant part which is evident from more negative NEE during that stage. It was also observed that with the increase in temperature, Re also increased. It may be due to the warm temperature which promotes photorespiration, and most crops drop their photosynthetic activities at elevated temperature (Crafts-Brandner and Salvucci 2000).

Sensible heat, latent heat, photosynthetically active radiation, and net radiation as influenced by lowland rice ecology

Because of its unique nature like flooded ecology, energy partitioned into LE is an important factor for rice crop. Except in fallow periods, the LE remained higher than Hs irrespective of the seasons. In lowland rice ecology, the amount of energy used mostly for vaporization of water molecules (LE) outweighed Hs which is required for warming up the ambient air temperature (Alberto et al. 2009). During the fallow season as there were no crops except the rice ratoons of the previous crop, Hs dominated over LE due to very low soil moisture content prevailing in the rice paddy field (Alberto et al. 2011). It was observed that the submerged rice fields had significantly lower Hs than LE during the active crop growing season, because of standing water which reduced the soil as well as canopy air temperature, whereas in fallow periods, Hs increased and LE decreased as the soil surface becomes drier (Alberto et al. 2011). Compared with other crops (e.g., mulberry), rice showed less Hs during sunny days under large NR condition in coastal China (Li et al. 2017). The PAR was higher in mid to end of the dry season (from the PI to H of the rice crop) than in the wet season. Similarly, dry fallow recorded higher PAR compared to the wet fallow. This may due to clear sky in the dry season and dry fallow which coincide with the summer season, whereas during the wet season and part of the wet fallow, it is less because of the prevalence of cloudy condition due to the commencement of southeast monsoon through that time period (Cruz et al. 2013). A similar phenomenon was observed in case of NR (Santanello and Friedl 2003; Tsai et al. 2007). The NR radiation represents the total available energy that the rice system can use to perform its physiological processes, and it is the difference between incoming and outgoing radiation. A negative correlation of NEE with LE, PAR, and NR was noted. In a water-sufficient system, GPP is positively related to PAR (Alberto et al. 2009), and as a whole, NEE was negative due to enhanced CO2 uptake during the day through photosynthesis (i.e., increase in GPP). Thus, a negative correlation of NEE with PAR was recorded. Moreover, rice is principally a C3 crop and CO2 is assimilated into a 3-carbon compound by Rubisco, which is a photosynthetic enzyme. The enzyme Rubisco is thermosensitive, which causes photosynthesis to be hampered at higher temperature (Crafts-Brandner and Salvucci 2000). Probably due to the above reasons, a negative correlation of NR with NEE was recorded. As rice has sufficient moisture during its growth stages, the majority of the energy flux is portioned into LE which is linked to high evaporative flux. This may lead to reduction in CO2 uptake due to the stomatal closure (Bhattacharyya et al. 2013a).

Soil enzyme activities and LC pools at various crop growing stages

Soil enzyme activities are often used as an indicator of microbial growth and its activity in soils. β-Glucosidase is a carbon-dependent enzyme and its activity increases with the increase in labile C content like RMC and MBC. Fluorescein diacetate is a hydrolytic enzyme which breaks down the organic components in the soil and represents CO2 evolution as the total microbial activity (Nannipieri et al. 1990). On the other hand, dehydrogenase is considered as the respiratory enzyme where the microbial respiration observed by the CO2 uptake or evolution took place and can be observed in the form of LC pools. Soluble carbon fraction plays an important role in soil organic matter turnover in agricultural soils like puddled rice soil where it acts as a readily decomposable substrate for microorganisms (García-Orenes et al. 2010). Application of organic manure contributes more labile C which acts as a source of energy and nutrients to the microorganisms (Manna et al. 2007). However, the content of labile pools also varied with stage of the crops. The increase in both soil enzyme activities and LC pools in the PI stage may be due to the release of a higher quantity of carbonaceous root exudates that added to the LC pools and also acted as a substrate for soil enzymes (Win et al. 2012).

Soil LC acts as a food source of soil microbes and increased enzymatic activities indicate the presence of a higher microbial population in soil. The NEME and GWP increased when there is availability of a higher amount of LC (substrate) in soil at the PI stage. However, a clear trend was not recorded for N2O across the seasons. The NEE was lowest at the early vegetative stage and reached its peak during the PI to FL due to greater physiological activity and higher photosynthesis CO2-C uptake. With increased soil enzymatic activities, the NEE becomes less negative which indicates lower assimilation of CO2 and this can be explained by linking it to the enhanced Re due to the presence of a higher microbial population in soil as indicated by higher enzymatic activities (e.g., β-GLU, FDA, and DHA) (Frankenberger and Dick 1983). Therefore, soil labile C as well as specific soil enzymes could be used as good indicator of NEE, NEME, and GWP in lowland rice ecology.

Conclusion

The NEE and NEME exhibited a clear daily and seasonal pattern throughout the cropping of the rice-rice season. More negative NEE during the ripening stage and higher GPP during the ripening to maturity stage of rice are due to the assimilation of more CO2 in plant parts. Re increases with temperature rise as warm temperature promotes photorespiration and reduces photosynthetic activities in rice crop. NEME was higher during the AT to PI stages due to higher methanogenesis promoted by the presence of liable substrates (increased RMC and MBC). On the contrary, N2O flux is controlled by nitrogenous fertilizer (urea) application. The GWP increased with the advancement of crop stages and declined at the end of each season and fallow period. It was also observed that LE dominates over Hs during the cropping season (dry and wet season) and it contributes largely to the NR and energy balance. The PAR, NR, and NEE influence GWP positively. Due to overcast situation during the WS and WF, PAR and NR are less as compared to those during the DS and DF. Soil enzyme activities and LC pools increased with the age of the crop and reached its maximum during the PI stage. The NEME and GWP increased with a higher amount of LC pools in soil at the PI stage. Further investigations are still required to improve our understanding of GHG emissions from different rice-rice ecosystems and their potential response to climate change.