Soil carbon dynamics and productivity of rice–rice system under conservation tillage in submerged and unsubmerged ecologies of Eastern Indian Himalaya

Abstract

Intensive tillage practices along with improper residue management in a rice (Oryza sativa)–rice system (RRS) contributed to soil fatigue and declining productivity in South Asia. Therefore, a 3-year (2013–2015) field study was conducted to assess tillage modification effects on productivity and soil C sequestration under RRS at ICAR- RC for North Eastern Hill Region, Tripura, India. The experimental site represented two different ecologies: unsubmerged (ECO 1) and submerged (ECO 2), with three tillage practices: conventional tillage (CT), reduced tillage (RT) and no-till (NT). Results showed that the cultivation of RRS under RT produced significantly higher grain (8.6–9.4 Mg ha⁻¹) and straw (11.8–12.9 Mg ha⁻¹) yields under both ecologies over those under CT and RT, in addition to recycling the maximum biomass. Soil under RT had lower bulk density (ρ_b), the highest soil organic carbon (SOC) concentration, pool, sequestration, accumulation, carbon retention efficiency, soil microbial biomass carbon and dehydrogenase activities under both ecologies as compared to CT. A total amount of 1.34 Mg C ha⁻¹ was accumulated under soils of RT over 3 years. The rate of SOC sequestration ranged from 133.6 kg ha⁻¹ year⁻¹ under soils of CT to 444.7 kg ha⁻¹ year⁻¹ under RT in RRS. Thus, cultivation of RRS under RT with effective residue recycling is recommended for higher system productivity and C sequestration under both rice production ecologies of the NEH region of India.

Keywords:

• Conservation tillage
• rice ecologies
• rice–rice system
• soil carbon sequestration
• system productivity

Introduction

The rice (Oryza sativa L.)–rice system (RRS) is the second most dominant cropping system after the rice–wheat (Triticum aestivum L.) system in India (Mohanty et al., 2013). The employment, income and ultimately livelihood security of millions of rural and urban poor on the Indian sub-continent mainly depend on RRS (Das et al., 2014). Recent reports on long-term RRS experiments revealed that the system productivity of RRS is either stagnating or declining (Cassman and Dobbermann, 2001, Banerjee and Pal, 2009). Poor management practices including intensive tillage, poor residue management, low organic manure application and nutrient supplements, etc. have resulted in declining productivity and soil degradation which will eventually threaten the food security of the region (Ladha et al., 2003) and losses of soil organic carbon (SOC) and nitrogen (N) (Lal, 2015). To keep pace with the demands of the burgeoning population a revolution in global food production is a must, to boost it by about 60% by 2050 (Alexandratos and Bruinsma, 2012). Hence, direct changes in the existing practices toward an integration of technologies and restoring carbon (C) and N pools in soil are prerequisites to enhance rice productivity in a sustainable manner. SOC plays a decisive role in improving crop productivity and reducing C emissions in the agricultural production system; hence, proper management of SOC is crucial (Rajan et al., 2012). Paddy soils are reported to retain higher amounts of C and N compared with all other terrestrial ecosystems (Liu et al., 2006). Higher amounts of carbon added in submerged paddy soils and slower oxidation of soil organic matter (SOM) may lead to higher net C accumulation (Majumdar et al., 2008; Bhattacharyya et al., 2013). Slight alterations in cultivation practices could modify SOC contents (Zhu et al., 2014) to a great extent. Intensive soil disturbances, and improper residue and nutrient management, in rice ecosystems are mainly responsible for depletion of SOC and N pools. Crop residue retention stabilizes SOC and also brings about the sequestration of C, which increases the crop yield (West and Post, 2002; Sun et al., 2007), makes more nutrients available (Kushwah et al., 2016) and reduces the level of greenhouse gas emissions (Liu et al., 2006). A better understanding of SOC and N dynamics in soils of RRS is a must to bring about the desired changes and to increase the productivity of the system (Sainju et al., 2008). Hence, development of appropriate soil and crop management technologies, which can sequester more C and N to enhance the system productivity and soil quality of rice-based ecosystems, is the prime challenge for researchers and policymakers of South Asia.

Conservation agriculture (CA) mainly relies on reduced tillage (RT)/no-till (NT), retention of crop residues and crop rotation to conserve SOC in diverse ecosystems (Gonzalez Sanchez et al., 2012; Prasad et al., 2016). CA-based alternative tillage and crop establishment methods are reported to have positive effects on soil health and crop productivity by curbing production cost, and sustaining crop yields and water productivity (Ladha et al., 2009; Jat et al., 2013). The practice of RRS in Asia is characterized by repeated tillage and soil puddling after manual transplanting coupled with no rice residue incorporation and fertilization. These practices aggravate the soil C and N losses (Singh et al., 2016), disrupt capillary pores (Hobbs, 2007), and increase the soil mechanical resistance by giving rise to an impermeable soil layer that impedes root penetration, leads to poor crop stand and eventually results in poor crop yield (Singh et al., 2015).

The North Eastern Hill (NEH) region of India is characterized by diversified agro-climatic conditions and a topography of hills and valleys, having a rich resource base. Rice is the staple food of the majority of the NEH population and RRS is the most prevalent system; it accounts for about 8% and 6.5% of India’s total rice area and production, respectively. Despite congenial edapho-climatic conditions, the average rice productivity of the region is <2 Mg ha⁻¹ against India’s national average of 3.6 Mg ha⁻¹, and 4.42 Mg ha⁻¹ in neighboring Bangladesh, creating a superlative challenge for researchers (Das et al., 2015). Thus, sustainability of RRS is need of the hour as the food security of NEH India mainly depends on rice. Conservation-effective agro-technologies such as RT, NT and integrated nutrients management (INM) have been proven to increase crop and soil productivity under irrigated conditions on the Indo-Gangetic Plains in the most common cropping systems such as rice–wheat and maize (Zea mays)–wheat (Singh et al., 2016). However, meager information is available on RRS under a CA system (Nagarajan et al., 2013) and scant research work has been so far conducted in the NEH of India for comparative assessment of modified tillage practices in diverse rice production ecologies. Hence, it was hypothesized that conservation tillage practices along with residue retention may enhance the system productivity of RRS, and increase C sequestration by reducing C losses and improving soil health under both ecologies. Therefore, the present study was conducted to assess the effect of various tillage practices in two rice ecologies on system productivity and soil C pools of RRS in the NEH region of India.

Materials and methods

Experimental site

A 3-year study (2013–2015) on RRS was conducted in the two rice ecologies at the Agronomy experimental farm (Cocotilla farm) of the Indian Council of Agricultural Research (ICAR)-Research Complex for NEH Region, Tripura Centre, Lembucherra, Tripura (W), India (23°54’24.02”N, 91°18’58.35”E; altitude 52 m above mean sea level). The soil (Typic Kandihumults) of the experimental field is clay loam, deep and free from gravels and hardpan. Baseline soil samples from 0–20 cm depth were collected and analyzed before initiation of the experiment. The mixed soil of both ecologies had 10.2 g kg⁻¹ of SOC, 1.01 g kg⁻¹ total soil N, 9.5 mg kg⁻¹ available phosphorus (P) and 295.7 mg kg⁻¹ available potassium (K). The soil pH was 5.1 (1:2.5 soil:water ratio) and bulk density (ρ_b) was 1.33 Mg m⁻³.

Weather and crop information

The average annual rainfall at the site was 2200 mm, with >65% received between June and September. The average monthly distribution of rainfall, temperature and relative humidity (RH) of the three-year period (2013–2015) are presented in Figure 1. The wet season rice (WR) was grown in a nursery during June and transplanted in the main field during July and harvested during November. The dry season rice (DR) was transplanted during winter (January) and harvested in summer (May–June).

Figure 1 Average meteorological observations at the experimental site from 2013 to 2015.

Experimental design and crop management

The experiment was laid out in a split plot design with three replications. The gross and net plot size were 6.0 × 5.0 m² and 5.0 × 4.0 m², respectively. Detailed experimental descriptions are as indicated below.

Rice ecologies

ECO 1: This is unsubmerged, rainfed irrigated ecology. WR was grown under rainfed conditions in the lowlands. Drainage channels were provided around the field to drain excess water in the case of heavy rains. Therefore, submerged conditions did not last continuously for more than a week. DR was grown under irrigated conditions in the lowlands. Irrigation was applied as per the needs of the crop.

ECO 2: Both rice crops (WR and DR) were grown under rainfed conditions in the lowlands. The WR crop depended on rainfall and DR depended on seepage water from the surrounding hills. The rice field was submerged throughout the experimental period and no irrigation was applied under this condition.

Tillage systems

Conventional tillage (CT): To plow CT plots, 4 times manual spading was performed and then they were puddled with two more spadings under both rice ecologies.

Reduced tillage (RT): Two manual spadings were given in RT plots and no puddling was done under both rice ecologies. Thirty percent of the rice residue of the previous crop was incorporated to up to 20 cm depth in soil during the spading.

No-till (NT): No spading or puddling was done under NT plots and 30% of the rice residue of previous rice crops was retained on the soil surface under both ecologies and in both seasons. Glyphosate (N-(phosphonomethyl) glycine) at 5 ml l⁻¹ was sprayed on NT plots using a flat fan nozzle (Knapsac sprayer, model AGM/001) to control the weeds a week (7 days) prior to transplanting.

Crop management

Twelve kilograms of rice seed was sown for seedling production in a nursery area of 120 m² for each crop. Rice seedlings, 21 days old, of the popular high-yielding varieties Naveen (DR) and Gomati Dhan (WR) were manually transplanted at 20 cm × 20 cm spacing with two seedlings per hill. DR was raised in a nursery during the first week of January and transplanted during the last week of January. Similarly, WR was nursery-raised during the third week of June and transplanted in the first week of July. The recommended dose of nutrients (i.e. 60:18:33 kg N:P:K ha⁻¹) was applied through urea, single superphosphate (SSP) and muriate of potash (MOP), respectively, in both crops during the two respective seasons. The full amounts of P and K and half of N were applied as basal fertilizer and the other 50% N applied in two equal splits at 30 and 75 days after transplanting (DAT). Weeds in NT plots were controlled by spraying of glyphosate (N-(phosphonomethyl) glycine) @ 5 ml l⁻¹ using a flat fan nozzle (Knapsac sprayer, model AGM/001) 7 days before transplanting of the succeeding rice crop. Irrespective of tillage, pretilachlor (2-Chloro-N-(2,6-diethylphenyl)-N-(2-propoxyethyl) acetamide at 1200 g a.i. (active ingredient ) ha⁻¹ was applied manually 2 DAT to minimize the weed infestation. The DR plots were kept flooded (5 cm of standing water) for the first 2 weeks, followed by irrigation (to 5 cm depth) at the appearance of cracks on the soil surface till maturity. On average, total irrigation water used in DR ranged between 800 and 900 ha-mm across the treatments under ECO 1. In order to avoid moisture stress in crops, irrigation water was applied to the rice at all critical growth stages in the different treatments.

Harvesting, economic yield and biomass measurement

Both WR and DR were harvested at physiological maturity, during the second fortnight of November and the last week of May to the first week of June, respectively, in all years. For recording yield, a net plot area of 5 × 4 m was harvested and kept at the threshing floor to allow the biomass to dry for 4–5 days. The harvested plants were weighed, and threshed manually. Sub-samples of grains and straw were dried in oven at 70 °C to a constant weight. Grain yields were adjusted to 14% moisture content for both WR and DR crops. Biomass of previous WR (grown during the previous rainy season before the initiation of the present study, for treatment stabilization) in the amount of 2.1 Mg ha⁻¹ was considered the initial biomass contribution to the first DR crop (2013).

Root samples were obtained at harvest in both crops at 20 cm depth using a core sampler (5.8 cm height and 5.4 cm diameter). Five hills were randomly selected for sampling from each plot. The core samples with roots and soil were soaked in water for at least 12 hours following the procedure described by Bohm (1979). The roots were washed to clean off the soil and dead organic debris and fresh roots were oven-dried at 70 ± 1 °C to constant weight, and the dry biomass was determined and converted into Mg ha⁻¹.

Soil sampling and analysis

Soil samples were obtained (500 g composite sample, one sample from each plot) from 0–20 cm depth for analyzing the SOC, soil microbial biomass carbon (SMBC) and dehydrogenase activities (DHA) after completion of the 3 years of the experiment. The total C was determined by the dry combustion method (Nelson and Sommers, 2005) using a TOC analyzer (Elementar Vario Select, Germany). Fresh soil samples were stored in freezing temperatures and used for analyzing the SMBC and DHA. SMBC was estimated by the soil fumigation technique (Anderson and Ingram, 1993). Soil DHA was estimated by the procedure described by Tabatabai (1982), by reducing 2, 3, 5-triphenyl tetrazolium chloride (Casida et al., 1964). Soil bulk density (ρ_b) was determined by the core method (Blake and Hartge, 1986) using cores of 5.8 cm height and 5.4 cm diameter at 0–20 cm depth and oven-dried at 105 °C (one sample per plot).

Computation of C pools

The total SOC pool (Mg ha⁻¹) of 0–20 cm depth was calculated using the fixed depth (FD) method according to the following equation (Lee et al., 2009): (1) $${\mathrm{M}}_{\mathrm{C}}={\rho }_{\mathrm{b}}\times {\mathrm{D}}_{\mathrm{f}}\times {\mathrm{C}}_{\mathrm{C}}\times 10$$(1)

where M_C is the SOC mass per unit area (Mg C ha⁻¹), ρ_b is the soil bulk density (Mg m⁻³), C_C is the concentration of SOC (g kg⁻¹), D_f is the depth of the fixed soil layer (m), and 10 is a product of the unit conversion factor (m² ha⁻¹, g Mg⁻¹ and kg Mg⁻¹).

Sequestration of SOC was computed as per Equation (2)(2) $$\mathrm{C}\text{sequestered}\left({\text{Mg}\mathrm{C}\text{ha}}^{-1}\text{soil}\right)=\left[\text{SOC}\text{current}\left({\text{Mg}\text{ha}}^{-1}\right)–\text{SOC}\text{initial}\left({\text{Mgha}}^{-1}\right)\right]÷\text{year}\text{of}\text{experimentation}$$(2) below:

(2) $$\mathrm{C}\text{sequestered}\left({\text{Mg}\mathrm{C}\text{ha}}^{-1}\text{soil}\right)=\left[\text{SOC}\text{current}\left({\text{Mg}\text{ha}}^{-1}\right)–\text{SOC}\text{initial}\left({\text{Mgha}}^{-1}\right)\right]÷\text{year}\text{of}\text{experimentation}$$(2)

Carbon retention efficiency (CRE) was calculated using Equation (3):(3) $$\text{CRE}\left(\%\right)=\left(\text{SOC}\text{final}–\text{SOC}\text{initial}\right)\times 100÷\text{ECI}$$(3) (3) $$\text{CRE}\left(\%\right)=\left(\text{SOC}\text{final}–\text{SOC}\text{initial}\right)\times 100÷\text{ECI}$$(3)

SOC final and SOC initial represent SOC (Mg ha⁻¹) in the final and initial soils, respectively, and ECI is cumulative estimated C input (Mg ha⁻¹) to soil between the initial and final years of experimentation.

Statistical analysis

Statistical analyses were done using the GLM procedure of SAS 9.4 (SAS Institute, 2003) to analyze variance and to determine the statistical significance of the treatment effects. The least significant difference (LSD) at p = 0.05 was used to compare treatment means. $$\text{Model}\text{SOC}=\text{Eco}\text{rep}\text{rep}*\text{Eco}\text{treat}\text{treat}*\text{Eco};$$

Means were compared as follows:

Means Eco/LSD;

Means treat/LSD;

Means Eco*treat/LSD.

In the model, treat indicates tillage, Eco indicates the rice ecology and rep is a replication.

Results and discussion

Wet season rice yield

Grain and straw yields of WR crop were significantly affected by tillage practices and production ecologies (Table 1). Cultivation of WR under RT recorded significantly higher grain yield (4.2–4.7 Mg ha⁻¹) as compared to other treatments across the years. Thus, there was a positive impact of RT practices on the production of the WR crop. Irrespective of tillage practices, the productivity of WR was greater under ECO 1 (4.5–4.69 Mg ha⁻¹) as compared to ECO 2 (3.54–3.86 Mg ha⁻¹) across the years. The data for the three consecutive years indicated that the conservation tillage with 30% residue retention/incorporation produced a higher rice grain yield as compared with that of CT. These trends indicate the opportunity to adopt CA in lowland rice cultivation with some yield advantage. A similar grain yield of rice under puddled and minimum puddled treatments was also reported by Singh et al. (2004). However, some researchers (Hammel, 1995 Haque) observed that conversion from CT to CA did not increase crop yield under humid conditions. However, the soil of the NEH region is clay loam in texture coupled with a high water table, where percolation of water is relatively low, and hence puddling has the least effect on rice productivity. This might be the reason for a good rice yield under RT in the present study. Many previous studies on conservation tillage have also reported crop yield advantages over traditional repeated tillage systems (Ladha et al., 2009). Most previous studies evaluated RT/NT versus CT in wheat and maize (Sharma et al., 2011; Yadav et al., 2015) rather than in transplanted rice under irrigated and rainfed conditions. Sharma et al. (2005) reported similar yields of transplanted rice with intensive puddling and no puddling with RT. Similarly, Haque et al., (2016) observed that RT and unpuddled transplanting of rice produced higher grain yields, besides reducing the cost of production and the time for land preparation and crop establishment across the seasons and years over the CT system. Ladha et al. (2009) also reported that transplanting into unpuddled NT soil increased average rice yields on farmers’ fields by 0.3 Mg ha⁻¹. Therefore, negative effects of RT/NT on crop growth are not observed in wet-seeded flooded-rice production systems (Huang et al., 2012).

Table 1. Effect of tillage practices and production ecologies on grain and straw yield of wet season rice (WR).

Treatment	WR grain yield (Mg/ha)
	2013			2014			2015
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	4.40	3.50	3.95	4.50	3.70	4.10	4.50	3.80	4.15
RT	4.80	3.60	4.20	5.00	4.10	4.55	5.10	4.30	4.70
NT	4.30	3.53	3.92	4.50	3.50	4.00	4.47	3.47	3.97
Mean	4.50	3.54		4.67	3.77		4.69	3.86
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.14	0.17	NS	0.25	0.31	NS	0.17	0.21	NS
	WR straw yield (Mg/ha)
	2013			2014			2015
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	6.23	5.00	5.62	6.27	5.20	5.73	6.30	5.37	5.83
RT	6.80	5.00	5.90	7.00	6.00	6.50	7.30	5.90	6.60
NT	6.00	5.10	5.55	6.30	5.00	5.65	6.40	4.80	5.60
Mean	6.34	5.03		6.52	5.40		6.67	5.36
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.48	NS	NS	0.43	0.53	NS	0.31	0.38	NS

CT: Conventional Tillage; RT: Reduced Tillage, NT: No-Till, ECO: Ecology

Dry season rice yield

Yields of DR crop (grain and straw) were significantly affected by rice ecologies and tillage practices (Table 2). Across the years, RT yielded higher grain (4.38–4.72 Mg ha⁻¹) and straw (5.97–6.37 Mg ha⁻¹) yields of rice over those under CT and NT. The CT and NT treatments did not show a significant difference in yield (grain or straw). However, transplanting of rice under NT systems recorded 7.9 and 17.1% lower grain yield compared with CT. In contrast, RT increased the grain yield of rice by 4.8–5.5% over CT. In general, CT with puddling has greater impact during the dry season, by reducing percolation loss of water and controlling weeds. But these advantages might be achieved through RT because initial plowing may keep the field weed free and incorporate the crop residue into the soil for better decomposition, which may also supply the essential plant nutrients to the crop. However, only a few studies on conservation tillage (RT and NT) under DR have been conducted in India (Pandey and Velasco, 1999; Naresh et al., 2014) to assess the impact of RT in DR. Most of the previous research related to CA involved the rice–wheat system, with special emphasis on wheat (Hobbs, 2007; Jat et al., 2009). However, several researchers in Bangladesh (Haque, 2009; Haque et al., 2016) reported no significant yield differences between RT and CT treatments. In contrast, several experiments in China (Peng et al., 2009; Huang et al., 2015; Mi et al., 2016) showed a declining trend in rice yield during the dry season under conservation tillage, while others (Sun et al., 2010; Xue et al., 2015; Huang et al., 2015, and Ma et al., 2016) reported an increase in rice yield under conservation tillage irrespective of residue removal or retention. These findings suggest that the grain yield responses to CA vary among regions, due to differences in climatic factors (Huang et al., 2015) and soil pH (Huang et al., 2015). Gregory (2012) reported that the NT practice reduces rice grain yield by 2.8% in soils with pH < 6.0 compared with conventional tillage. The soil of the present study also had an acidic soil reaction (pH < 5.5). During the dry season, nutrient uptake by rice roots under NT might have been relatively low, leading to poor biomass production (Huang et al., 2016). These might be the reasons for lower rice yield under NT than the CT. With regards to rice ecologies, ECO 1 produced a higher rice grain yield for DR as compared to ECO 2 (Table 2).

Table 2. Effect of tillage practices and production ecologies on grain and straw yield of dry season rice (DR).

Treatment	DR grain yield (Mg/ha)
	2013			2014			2015
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	4.80	3.50	4.15	5.07	3.70	4.38	5.10	3.90	4.50
RT	5.10	3.67	4.38	4.90	3.80	4.35	5.33	4.10	4.72
NT	4.10	3.53	3.82	4.00	3.60	3.80	4.00	3.47	3.73
Mean	4.67	3.57		4.66	3.70		4.81	3.82
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.16	0.20	NS	0.22	0.27	NS	0.27	0.33	NS
	DR straw yield (Mg/ha)
	2013			2014			2015
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	6.50	4.77	5.63	6.83	5.03	5.93	6.90	5.30	6.10
RT	7.00	4.93	5.97	7.10	5.13	6.12	7.20	5.53	6.37
NT	5.40	4.80	5.10	4.47	4.90	4.68	5.40	4.67	5.03
Mean	6.30	4.83		6.13	5.02		6.50	5.17
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.35	0.43	NS	0.09	0.11	NS	0.29	0.35	NS

CT: Conventional Tillage; RT: Reduced Tillage; NT: No-Till; ECO: Ecology

Effects on system productivity

System yields of grain and straw were also affected by rice ecologies and tillage practices (Table 3). The rice grain yields of the RRS ranged from 7.0 to 10.5 Mg ha⁻¹ across treatments and years. However, impacts of treatments on system productivity were similar to those recorded under individual WR and DR. The system grain and straw yield over three consecutive years were 6.2–9.0% and 5.6–8.8% more, respectively, under RT practices than that under CT. The highest system productivity was recorded when RRS was grown in ECO 1 as compared to ECO 2. Increased productivity of RRS under conservation tillage (RT/NT) has also been reported in other regions of world (Huang et al., 2015), but effects varied across climatic and ecological regions (Huang et al., 2015). The data of the present study demonstrate that cultivation of RRS under RT under both rice ecologies was most suitable for the NEH region. However, further studies are needed to identify a strategy to improve the productivity of RRS under NT under both ecologies for adoption of CA in the NEH region of India.

Table 3. Effect of tillage practices and production ecologies on system productivity of RRS.

Treatment	System grain yield (Mg/ha)
	2013			2014			2015
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	9.20	7.00	8.10	9.53	7.40	8.47	9.60	7.70	8.65
RT	9.93	7.27	8.60	9.90	7.90	8.90	10.47	8.40	9.43
NT	8.43	7.00	7.72	8.50	7.10	7.80	8.43	6.90	7.67
Mean	9.19	7.09		9.31	7.47		9.50	7.67
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.35	0.43	NS	0.41	0.50	NS	0.22	0.27	NS
	System straw yield (Mg/ha)
	2013			2014			2015
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	12.73	9.73	11.23	13.10	10.23	11.67	13.20	10.63	11.92
RT	13.80	9.93	11.87	14.10	11.13	12.62	14.50	11.43	12.97
NT	11.40	9.90	10.65	10.77	9.90	10.33	11.80	9.47	10.63
Mean	12.64	9.86		12.66	10.42		13.17	10.51
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.44	0.54	NS	0.35	0.43	NS	0.49	0.61	NS

CT: Conventional Tillage; RT: Reduced Tillage; NT: No-Till; ECO: Ecology

Effects on biomass and carbon recycling

The amount of biomass (grain + straw + root mass up to 20 cm soil depth) produced and C recycling varied among the treatments (Table 4). Recycling of biomass (straw + root mass) and C was assessed to 20 cm soil depth with retention of residues by 30% along with root biomass, for each crop during the entire experiment. The highest biomass and C were recycled in RT and ECO 1 plots through 30% residue + root biomass (20 cm soil depth) due to higher straw and root mass production. Higher C recycling potential from retention of rice residue and with adoption of RT and NT were also reported by Dobermann and Witt (2000) and rice residue and weed biomass by Das et al., (Das 2014) under RT in lowland rice compared to that grown in a CT system.

Table 4. Biomass and carbon inputs in double cropping of rice (RRS) affected by tillage practices and production ecologies.

Treatment	Biomass input (Mg/ha)			Total carbon input (Mg/ha)			Carbon input (Mg/ha/year)
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	19.0	14.9	17.0	8.2	6.4	7.3	2.7	2.1	2.4
RT	20.7	15.9	18.3	8.9	6.8	7.9	3.0	2.3	2.6
NT	16.6	14.2	15.4	7.2	6.1	6.7	2.4	2.0	2.2
Mean	18.8	15.0		8.1	6.5		2.7	2.2
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.68	0.83	NS	0.29	0.36	NS	0.10	0.12	NS

Where: CT: Conventional Tillage; RT: Reduced Tillage; NT: No-Till; ECO: Ecology, Till: Tillage.

Effects on bulk density and total soil organic carbon

Tillage modifications and rice ecologies significantly affected soil ρ_b at 0–20 cm soil depth (Table 5). By the end of the 3-year study, soil under RT recorded lower ρ_b than that under other treatments (CT and NT). Between the rice ecologies, ECO 2 had the lower soil ρ_b (1.29 Mg m⁻³) as compared to ECO 1 (1.32 Mg m⁻³). The data presented herein suggested that RT in both ECO 1 and ECO 2 may have a relatively positive effect on soil ρ_b. In contrast, other studies have reported higher ρ_b under NT at 0–5 cm depth compared to that under CT in cropping systems other than RRS (Huang et al., 2012; Jat et al., 2013). Management-induced changes in soil ρ_b depend on texture, SOM content, tillage type and intensity, and cropping system (Sharma et al., 2003; Kharub et al., 2004).

Table 5. Effect of tillage practices and production ecologies on bulk density and carbon dynamics of RRS.

Treatment	Bulk density (Mg/m^3)			Soil organic carbon (g/kg)			Soil organic carbon pool (Mg/ha)
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	1.33	1.30	1.32	10.4	10.7	10.6	20.8	20.9	20.8
RT	1.31	1.28	1.30	11.0	11.4	11.2	21.6	21.9	21.8
NT	1.33	1.30	1.32	10.7	11.0	10.9	21.4	21.4	21.4
Mean	1.32	1.29		10.7	11.0		21.3	21.4
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.01	0.02	NS	0.13	0.15	NS	NS	0.5	NS
	Total soil organic carbon accumulation (Mg/ha)			Soil organic carbon sequestration (kg/ha/year)			Carbon retention efficiency (%)
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	0.50	0.37	0.43	144.7	122.4	133.6	5.2	5.7	5.5
RT	1.30	1.38	1.34	427.9	461.4	444.7	11.0	11.9	11.4
NT	1.00	0.98	0.98	336.5	315.0	325.8	13.5	14.7	14.1
Mean	1.07	1.09		355.8	358.7		11.8	12.2
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	NS	0.06	NS	NS	21.63	NS	NS	1.0	NS

CT: Conventional Tillage; RT: Reduced Tillage; NT: No-Till; ECO: Ecology, Till: Tillage; NS: Non-significant

Table 6. Effect of tillage practices and production ecologies on soil pH and selected soil biological parameters in RRS.

Treatment	pH			SMBC (µg g^−1 dry soil)			DHA (µg TPF g^−1 dry soils)
	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean	ECO 1	ECO 2	Mean
CT	5.25	5.18	5.22	234.5	185.1	209.8	6.2	4.1	5.2
RT	5.32	5.25	5.29	314.1	298.8	306.5	10.5	8.3	9.4
NT	5.29	5.22	5.26	304.4	284.6	294.5	9.5	7.0	8.3
Mean	5.29	5.22		284.3	256.2		8.7	6.5
	ECO	Till	ECO × Till	ECO	Till	ECO × Till	ECO	Till	ECO × Till
LSD0.05	0.02	0.02	NS	5.83	7.14		0.32	0.39	NS

CT: Conventional Tillage; RT: Reduced Tillage; NT: No- Till; ECO: Ecology, Till: Tillage; NS = Non-Significant

The SOC concentration at 0–20 cm depth after completing three consecutive cycles of RRS varied significantly among the rice ecologies and tillage systems (Table 5). SOC concentration increased up to 10.9 g kg⁻¹ under NT and to 11.2 g kg⁻¹ under RT as compared to CT (10.6 g kg⁻¹). The data indicate that NT and RT maintained consistently higher SOC concentrations in soil, which were 2.8%and 5.6% higher than the SOC concentration under CT, respectively. Other researchers have also indicated that RT and NT practices can produce comparable higher SOC concentration than CT (Ramesh and Chandrasekaran). In the present study, the maintenance of SOC under RT and NT might be due to retention/incorporation of 30% of rice residues of each crop, and production of more above- and below-ground biomass, in the respective treatments. Several researchers (Mandal et al., 2004; Ramesh and Chandrasekaran; Chen et al., 2017) have also reported that residue incorporation/retention increases the SOC concentration in rice soil. In addition, paddy soils also act as a sink for global C (Pan et al., 2004). Adoption of RRS under ECO 2 (11.0 g kg⁻¹) produced a higher SOC concentration than that in soil under ECO 1 (10.7 g kg⁻¹). The increase in SOC concentration might be due to the slow decomposition of added residues and root mass in ECO 2 as compared to ECO 1 (Zhou et al., 2016). The continuous soil submergence creates anaerobic conditions and reduces the temperature, resulting in a reduction in the rate of decomposition of organic matter and SOC mineralization (Zhou et al., 2016). The results indicate that CA (RT and NT) with residue retention/incorporation enhances the SOC content in the soils of both ecologies under RRS, but the magnitude of the increase was greater under ECO 2 than under ECO 1. In China, after 4 years of tillage and straw retention practices significant differences in SOC concentration (0–40 cm) were observed with different tillage practices (Shang-Qi et al., 2013). Similarly, CA with organic manure and residue retention/incorporation increased the SOC in paddy soils of India (Mandal et al., 2004; Ramesh and Chandrasekaran, Mandal et al., 2008). Conjoint use of organic with inorganic fertilizers increases the productivity and SOC under rainfed-irrigated lowland ecology of the northeastern region of India, as compared to submerged ecology (Nath et al., 2016).

The SOC pool was also increased under RT over CT in both rice ecologies. However, there were no significant differences in SOC under RT compared with NT. A total amount of 1.34 Mg Cha⁻¹ was accumulated under soils of RT over the 3 years. The rate of SOC sequestration ranged from 133.6 kg ha⁻¹ year⁻¹ under CT to 444.7 kg ha⁻¹ year⁻¹ under soils of RT under RRS. Adoption of RRS under RT/NT systems with retention/incorporation of 30% rice residues sequestered 3 times more SOC in the soil systems over CT plots. The variation in SOC pool and sequestration were attributed largely to C addition through recycling of crop residue retention (30% residue), minimal soil disturbances (RT/NT), production of root mass, nutrient-use pattern, soil texture and the prevailing ecosystem (Singh et al., 2015). Enhancement of the SOC pool of paddy soils can improve soil quality and mitigate global warming. Moreover, the CT practice can enhance SOC and N mineralization by incorporating crop residues, disrupting soil aggregates, and increasing aeration, thereby reducing the SOC pool (Xue et al., 2015; Huang et al., 2016). RT/NT practices have been reported to increase the SOC pool compared with CT in paddy-based agro-ecosystems (Sun et al., 2010; Xue et al., 2015), and the retention of rice residues with inorganic fertilizer may also enhance SOC sequestration in the double-rice cropping system in India (Bhattacharyya et al., 2012). However, ECO 2 had a greater SOC pool (21.4 Mg ha⁻¹) due to accumulation of more C (1.09 Mg ha⁻¹) during the 3 years as compared to that under ECO 1. The CRE is an important measure of the amount of C applied and retained in soil (Bhattacharyya et al., 2012). The lowest CRE was observed in CT of ECO 1 (5.2%) and the highest in NT of ECO 2 (14.7%) under RRS. This trend in CRE indicates that the adoption of RT/NT systems under RRS is a better option for C sequestration than the CT-based systems (Bhattacharyya et al., 2012), and might be due to the slow decomposition of applied biomass.

Effects on soil microbial biomass carbon and dehydrogenase activities

Under both rice ecologies, SMBC and DHA ranged from 185.1 to 314.1 µg g⁻¹ dry soil and 4.1 to 10.5 µg TPF (Triphenyl formazan) g⁻¹ dry soil, respectively, across the tillage treatments. Irrespective of the tillage treatments, higher values of SMBC and DHA were recorded under ECO 1 as compared to ECO 2 (Table 6). The SMBC concentration was significantly (p = 0.05) higher under RT (306.5 µg g⁻¹ dry soil) and ECO 1 (284.6 µg g⁻¹ dry soil), than those recorded under CT (209.8 µg g⁻¹ dry soil) and ECO 2 (256.2 µg g⁻¹ dry soil). Similarly, the DHA activity was also higher under RT of ECO 1 (10.5 µg TPF g⁻¹ dry soil) and the minimum was observed under CT of ECO 2 (4.1 µg TPF g⁻¹ dry soil). The SMBC and enzymes are good soil quality indicators because of their relevance to soil biology, rapid response to changes in soil management, and ease of measurement (Dick et al., 1996). In the present study, increasing tillage intensity reduced SMBC and DHA. Soil under RT and NT had significantly higher SBMC and DHA as compared to CT in both ecologies. The SMBC and DHA under RT of ECO 1 were higher than under CT of ECO 2 by 33.9% and 69.6%, respectively. Pandey et al. (2014) concluded that higher DHA in soil under NT and RT was due to larger proportions of SMBC than in soil under CT. Higher activity of soil DHA and SMBC under RT and NT compared to that under CT has been also reported by other researchers (Mina et al., 2008; Ghosh et al., 2010 and Pandey et al., 2014). The accumulation of crop residues on the soil surface is due to the enrichment of SOM in the surface layer and relatively higher microbial activity in soil under NT and RT (Mathew et al., 2012; Chen et al., 2017). Concentration of SMBC is indicative of the soil’s ability to store and cycle nutrients and SOM (Dick, 1992), and plays an important role in physical stabilization of aggregates. Consequently, SMBC is an important indicator of soil quality, and is also closely related to soil fertility. The rate of biomass-C input from plant residues is a predominant factor controlling the amount of SMBC in soil. A continuous and uniform supply of biomass-C is an energy source for microorganisms. Thus, RT/NT with crop residues is necessary for improving soil quality of both soil ecologies (ECO 1 and ECO 2), particularly in the depleted soils of the NEH region of India (Boulal et al., 2008).

Conclusions

The results of the present study demonstrate that RT with 30% residue incorporation under both ecologies enhances system productivity (grain yield) by 6.2–9.0% over CT practices, and also accumulates 1.30 and 1.38 Mg C ha⁻¹ in soils of ECO 1 and ECO 2 of RRS, respectively, over 3 years. Irrespective of the rice production ecology, the rate of SOC sequestration increased from 133.6 kg ha⁻¹ year⁻¹ under CT to 444.7 kg ha⁻¹ year⁻¹ under soils of RT. However, ECO 2 had a higher C sequestration rate than ECO 1. In addition, SOC concentrations, pool, CRE, SMBC and DHA also increased under RT with residue incorporation under both ecologies. Thus, cultivation of RRS under RT with effective residue recycling under both ecologies (ECO 1 and ECO 2) is recommended for enhancing the system productivity and SOC sequestration in paddy soils of the NEH region of India and other, similar agro-ecosystems, in South Asia especially.

Acknowledgements

Authors are very thankful to the Director, ICAR- Research Complex for NEH Region, Umiam, Meghalaya, India for providing the necessary facilities to conduct this research.

Disclosure statement

No potential conflict of interest was reported by the authors.