Effect of Water Saving Irrigation Management Practices on Rice Productivity and Methane Emission from Paddy Field

Irrigation water supply is one of the vital components for sustainable rice farming, which is becoming a limiting resource in the changing climatic condition. An experiment was conducted at the research field of Bangladesh Agricultural University, Mymensingh during dry season from January-June of 2017 to investigate the suitability of Alternate Wet and Dry Irrigation (AWDI) for sustainable rice production and reducing methane emission. The modern rice variety BINA Dhan 10 was used as test crop. There were five irrigation treatments viz. T 1 (saturated condition), T 2 (continuous flooded, 5 cm standing water), T 3 (AWDI-10 cm; irrigated when water level fell 10 cm from surface), T 4 (AWDI-15 cm; irrigated when water level fell 15 cm from surface) and T 5 (AWDI-20 cm; irrigated when water level fell 20 cm from surface). Results of the field trial showed satisfactory grain yield and low seasonal methane emission along with significantly high irrigation water savings (%) in AWDI treated field plots. Among the treatments, T 3 (AWDI-10 cm) and T 4 (AWDI-15 cm) showed higher yield performance (6250 kg∙ha −1 and 5810 kg∙ha −1 , respectively) with lower CH 4 emission (reduced up to 36% and 40%, respectively) compared to continuous flooded treatment (T2, CF 5 cm water). In AWDI field plots less irrigation frequency (6 - 9) was required which significantly saved the amount of irrigation water (12% - 24%). Although T 5 (AWDI-20 cm) showed the highest water savings reproductive stage of rice plant well-developed aerenchyma tissue was observed in root cortex under the continuous irrigated field plot, which indi-cates higher diffusion pathway of methane gas from root rhizosphere to the atmosphere compared to other treatments. Therefore, alternate wet and dry irrigation water management practice may be recommended at farmers’ level for sustainable rice production and reducing methane emission during dry winter Boro season which will reduce the cost of production by water saving as well as energy saving.


Introduction
Global agriculture in 21st century faces the tremendous challenge of providing sufficient and healthy food for growing population under increasing water scarcity, while minimizing environmental consequences. In agriculture sector, rice is a major food grain crop in the world and it will be necessary to produce about 60% more rice than currently produced to meet food demands by 2025. In Bangladesh, rice is the prime food grain crop cultivated in 77% (28.49 million hectares) of the total cropped area that covers 5th largest position of the world in respect to rice cultivation area (BBS, 2012). But rice cultivation especially in dry winter (Rabi) season requires huge amount of irrigation water, which is a great challenge due to energy crisis. From the time immemorial, rice is grown in low land areas under continuous flooding in Asian tropics and subtropics that requires between 700 and 1500 mm water per cropping season depending on soil texture (Bhuiyan, 1992) that leads to high surface runoff, seepage and percolation account for between 50% -80% of total water input. Water availability in Bangladesh is around 90 billion cubic meters (BCM) during dry season against the demand of 147 BCM, a shortage of nearly 40%, resulting in drought situation in large parts of the country (Quassem, 2010). Boro rice in Bangladesh is fully irrigated and the Aman rice is partly irrigated. An estimated 3000 to 5000 litres of water is required to produce one kg rice (BRRI, 2016). So, Bangladesh is going to face severe water crisis within next couple of decades. In future, it will be necessary to produce more food with less water or increase the efficiency of water management. Again, flooded rice cultivation is a major anthropogenic source of methane (CH 4 ) which is one of the important greenhouse gases having 25 times more global warming potential than carbon dioxide (Nieder & Benbi, 2008). These flooded soils characterized by low oxygen and high organic substrates offering an ideal environment for methanogen bacteria generate anaerobic condition that favors CH 4 production as an end product from organic matter degradation (Conrad, 2002). Methane from soil is derived from C mineralization, moisture, pH, soil Eh, etc. and emitted to the atmosphere by molecular diffusion, ebul-lition or plant mediated transport like aerenchyma (Zheng et al., 2007). The CH 4 concentration in the atmosphere has reached 1.77 ppm by volume in 2005, which is more than double that of its preindustrial value (Solomon et al., 2007). Therefore, it is necessary to solve water scarcity and minimizing CH 4 emission from rice paddy without aggravating rice production.
Two main types of water efficient irrigation practiced in Bangladesh are: 1) Alternate Wetting and Drying (AWD) and 2) System of Rice Intensification (SRI). AWDI developed by International Rice Research Institute, is a water saving and methane mitigation technology where rice fields are not kept continuously submerged but are allowed to dry intermittently during growing stages.
But a severe AWD regime in where photosynthesis is severely inhibited and plants cannot rehydrate overnight could markedly decrease grain yield, although it reduces methane emission from paddy fields (Win et al., 2015). Again, SRI is an efficient, resource-saving and productive way to practice the use of younger rice seedlings planted singly at wider spacing, together with intermittent irrigation and organic fertilization adoption. In our country, it is difficult to follow SRI because of some limitations like organic fertilization and more laborious that sum up more cost (Uphoff & Kassam, 2009). On the contrary, alternate wetting and drying irrigation for rice has become popular in Bangladesh than SRI due to economic and environmental feasibility (Price et al., 2013). Although, few studies in Bangladesh showed the effect of AWDI on rice yield, they do not give us a clear idea to sketch out the effect of AWDI on rice yield and CH 4 emission during rice cultivation. Therefore, this research experiment was conducted to evaluate different water saving irrigation management practices in paddy field for mitigating methane emissions and sustaining rice productivity.

Experimental Field Preparation and Rice Cultivation
The experiment was carried out at the research field of Bangladesh Agricultural University in 2017. The soil of experimental site was well drained, loamy texture having a soil pH (6.43) with moderate organic matter (1.78), 0.17% total nitrogen, 13.9 ppm available phosphorous and 16.3 ppm available potassium. The experimental field had three side-by-side blocks for three replications and each plot (20 m 2 ) was laid down in a randomized complete block design where block to block and plot to plot distances were 1.5 m and 1m to prevent interactions.
The field was tilled with a tractor and subsequently ploughed and cross ploughed followed by laddering and cleaning. These plots were surrounded by 25 cm wide and 20 cm high levees separated by 1 m transition zones. The applied fertilizer doses were Urea 220 kg•ha −1 , TSP 115 kg•ha −1 , MoP 100 kg•ha −1 and Gypsum 60 kg•ha −1 . Among them, nitrogenous fertilizer (urea) was applied thrice as basal dose after transplantation, whereas others were applied only once before transplanting. Rice seedlings (21 days old BINA Dhan 10) were transplanted into ex-H. J. Hiya et al. Journal of Geoscience and Environment Protection perimental plots at three seedlings hill −1 with 25 cm × 20 cm spacing between the rows and hills, respectively.

Placement of Perforated Pipes According to Treatments
In this study, five irrigation treatments were followed: T 1 = Saturated, T 2 = Continuous flooding (5 cm standing water), T 3 = AWDI, 5 cm irrigation applied when water level in the pipe fell 10 cm below ground level, T 4 = AWDI, 5 cm irrigation applied when water level in the pipe fell 15 cm below ground level and T 5 = AWDI, 5 cm irrigation applied when water level in the pipe fell 20 cm below ground level. Perforated PVC pipes were installed in the experimental plots 10 days after transplanting (DAT) according to treatments for measuring soil water depletion to follow AWDI techniques. The diameter of pipe was 8 cm and the lengths were 25 cm, 30 cm and 35 cm, respectively. After irrigation had been applied, water entered through perforations and water levels inside the pipe were at same level as that of outside. With the progress of time, water got depleted in AWDI plots, but a close observation revealed that there was water inside the pipe and again 5 cm irrigation was done when the depleting water table inside the pipe reached a certain level from ground soil.

Determination of Effective Rainfall, Crop Water Requirement and Water Savings
Effective rainfall was estimated by USDA Soil Conservation Method (Smith, 1992 2) P effective = 125 + 0.1 × P total (when, P total > 250 win) where, P effective = effective rainfall (mm) and P total = total rainfall (mm).
Total rainfall data of cropping period of that year was collected from the department of Irrigation and Water Management, Bangladesh Agricultural University. Again, water requirements for rice cultivation was computed by adding applied irrigation water, effective rainfall during growing season and water for land preparation (Rashid, 1997). In the study, water saving percentage was calculated as follows:

Gas Sampling, Analysis and Estimation of CH4 Emission
A closed-chamber method    (Rolston, 1986) was used to estimate methane fluxes for every treatment.
Total methane flux for the entire cropping period was computed by the formula (Singh et al., 1999): where, R i = rate of methane flux (g•m −2 •d −1 ) in the ith sampling interval, and n = number of sampling intervals.

Estimation GWP of CH4
To estimate the GWP, CO 2 is typically taken as the reference gas, and an increase or reduction in emission of CH 4 is converted into "CO 2 -equivalents" by means of their GWPs. In this study, we used the IPCC factors to calculate the combined GWP for 100 years (GWP = 25 × CH 4 , kg CO 2 -equivalents ha −1 ) from CH 4 under various agricultural irrigation practices. In addition, the greenhouse gas intensity (GHGI) was calculated by dividing GWP by grain yield for rice (Mosier et al., 2006).

Explorations of Rice Plant Growth and Yield
Rice plant growth parameters like plant height, root length were investigated during different growing period. Again, 4 -5 cm long root samples collected at reproductive stage were conserved in proper tagged glass tubes (with FA solutions) for free hand sectioning to show anatomical view whether aerenchyma developed or not under a scanning electron microscope. Yield components such as panicle number per hill and grain numbers per panicle were determined at harvesting stage. When 95% grains became yellow, harvesting was done. After harvesting, crops were separately collected, bundled and tagged for threshing.
Finally, the grains were sun dried up to14% moisture content, then weighed as kg•ha −1 .

Estimation of Water Productivity Index
Water productivity index is the ratio of crop yield (kg/ha) per unit water (m 3 /ha) supplied (Jaafar et al., 2000) and calculated as follows:

Investigation of Soil Properties
Soil redox potential (Eh) were measured during rice cultivation at certain time intervals by glass electrode Eh meter. At harvesting stage, soil bulk density (BD) was analyzed using cores (volume 100 cm 3 , inner diameter 5 cm), filled with fresh moisture soils. The collected core samples were oven dried at 105˚C for 24 h and then measured the weight of dried core samples. Soil porosity was calculated using BD and particle density (PD, 2.89 Mg•m −3 ) according to the equation: At harvesting stage, chemical properties of the collected soil samples were analyzed for organic carbon by wet oxidation method (Walkley & Black, 1965), organic matter content by multiplying the percent organic carbon with Van Bemmlen factor of 1.73, total nitrogen by Micro-kjeldhal method (Page et al., 1982), available phosphorus by Olsen method (Olsen et al., 1954), exchangeable potassium by Flame photometer (Brown & Lilleland, 1946).

Statistical Analysis
The compiled and tabulated data of rice growth, yield, soil properties and CH 4 emission were statistically analyzed by Analysis of Variance (ANOVA) to examine whether treatment effects were significant or not. The mean differences among treatments were compared by Duncan's Multiple Range Test. The computer software MSTAT-C was used for statistical analysis.

Effect of Irrigation Managements on CH4 Emission during Rice Growing Season
The different irrigation managements significantly influenced on CH 4 emission from paddy field at all stages of growing season ( Figure 1)  The decomposition rate of soil organic matter was more under irrigated field most probably due to the presence of available water. In study, the decomposition rate was more in continuous flooded plots that showed the highest CH 4 emission among all treatments. AWDI plots showed more soil organic carbon due to less decomposition that might be for less water requirements (Figure 3).
After heading stage, CH 4 flux was decreased constantly due to lack of CH 4 producing conditions caused by the end of irrigation and plant senescence, when labile organic C are no longer released due to the lack of water for decomposition.

Effect of Irrigation Managements on Rice Grain Yield and Water Productivity
During first 2 weeks after transplantation, 5 cm standing water was maintained in all plots to avoid weed infestation. Water required for crop establishment was 62.6 cm and for land preparation was 45 cm. Then, plots were irrigated according to treatments. Effective rainfall was 21.7 cm. In study, the maximum 13 irrigations required for continuous flooded plots followed by saturated plots (10 irrigations) in where AWDI treatments T 3 (10 cm), T 4 (15 cm) and T 5 (20 cm) re- respectively. Finally, it was found that AWDI-20 cm could save highest amount (18%) of irrigation water followed by AWDI-15 cm (16%) and AWDI-10 cm (13%) in where saturated plot could save only 8% irrigation water compared to continuous flooded plot.
The grain yield of rice was significantly increased by AWDIs. In present study, AWDI-10 cm gave the highest grain yield (6250 kg•ha −1 ) followed by AWDI-15 cm (5810 kg•ha −1 ) but the lowest yield (4283 kg•ha −1 ) was found in AWDI-20 cm due to more water stress. Again, continuous flooded plots gave 5783 kg•ha −1 grain yield most probably for more waterlogged conditions throughout the growing season.

Effect of Irrigation Managements on Seasonal CH4 Flux, GWP and GHGI
The total GWP of CH 4 significantly decreased with AWDI treatments as compared to continuous flooded plots. The total and highest GWP from the continuous H. J. Hiya et al. irrigated plot was 5.16 Mg CO 2 ha −1 followed by saturated plots (3.82 Mg CO 2 ha −1 ). Conversely, under AWDI treatments the lowest GWP from the AWDI-20 cm was 2.54 Mg CO 2 ha −1 followed by AWDI-15 cm (3.05 Mg CO 2 ha −1 ) and AWDI-10 cm (3.26 Mg CO 2 ha −1 ). In general, intermittent irrigation system (AWDI) significantly reduced the total GWPs (Table 1), even though continuous flooded and saturated plots showed two-fold higher GWPs. The GHGI also decreased remarkably with AWDI treatments compared to other treatments. As a result, AWDI-20 cm reduced highest percentage (50%) of CH 4 emission rate but gave lower yield due to more water stress compared to continuous flooded plots.
AWDI-15 cm could reduce 40% CH 4 emission with satisfactory yield although AWDI-10 cm reduced only 36% CH 4 emission with the highest grain yield production.

Effect of Irrigation Managements on Nutrient Status Prior to Rice Harvesting Stage
In study, Table 2 shows that the intermittent AWDI system increased soil porosity, the content of soil organic matter compared to continuous irrigated control plots in where total N content, available P and K content showed non-significant variation under different irrigation treatments. AWDI did not affect the soil nutrient status in addition, increased organic carbon as like as soil organic matter status. Soil organic matter significantly varied from 2.10% to 1.90% in where the highest % OM was found in AWDI-15 cm and the lowest was in continuous flooded plots.

Correlations Correlation of CH4 Emissions with Other Parameters
There were negative correlations between total seasonal CH 4 flux with grain yield, water productivity index, soil Eh, organic matter, total N, available P and K content except no. of irrigation, total volume of water supplied for rice grain production, soil porosity and soil pH under continuous irrigated treatment (Table   3). On the other hand, CH 4 flux showed a strong positive correlation with no. of irrigations, total volume of water supplied for paddy field, grain yield, water productivity index, the availability of soil organic matter, soil pH, total N, available P and K content, while there were only negative correlation with soil Eh and soil porosity under AWDI systems.

Discussion
In study, Alternate Wetting and Drying Irrigation (AWDI) significantly decreased the seasonal CH 4 emissions during rice cultivation. The lower CH 4 emission from AWDI plots was due to increased aeration, stabilization of soil organic carbon, improved soil redox potential status. Among AWDI treatments, T 3 (10 cm) reduced 36% CH 4 flux followed by T 4 -15 cm (40%) and T 5 -20 cm (50%).
This result revealed that AWDI treatments are the most effective options for reducing CH 4 emission during Boro rice cultivation, which was consistent with the findings of Jain et al. (2013). This result might be also due to highly promoting soil reduced condition that was influenced by different irrigation treatments (Supparattanapan et al., 2009). Ali et al. (2013) also reported that intermittent irrigations significantly reduced total seasonal CH 4 emissions by 27% compared to conventional (124 kg CH 4 /ha) irrigated rice paddy field.
The increased Eh values might have suppressed the methanogens' activities that were responsible for CH 4 emission from rice field, while stimulated the methanotrophs activities, and eventually decreased methane emission rate during rice cultivation (Ali, 2008  emission than AWDI treatments. The seasonal CH 4 emissions gave more or less similar pattern in both cultivation systems, i.e. CH 4 flux was higher during reproductive stage of rice growth in all treatments due to the development of root aerenchyma (Kludge et al., 1993), and decomposition of soil organic materials (Dubey, 2005). The sudden and sharp fall in CH 4 emission rates at grain maturation stage in all might be due to the aging of rice plant, the end of irrigation and plant senescence, when labile organic C are no longer released by organic matter decomposition due to lack of available water that was also supported by Cai et al. (1997). Finally, AWDI treatments significantly could mitigate the rate of CH 4 emission upto 50% (20 cm), 40% (15 cm) and 36% (10 cm) from that of continuous flooding in a cropping season which was supported by Win et al. (2015) and Itoh et al. (2011).
In our experiment, AWDI-20 cm saved upto 18% irrigation water with the lowest CH 4 emission rate compared to continuous flooded treatment but gave lower yield due to more water stress. But AWDI-10 cm and 15 cm saved upto 13% and 16% irrigation water, respectively with minimum rate of CH 4 emission compared to continuous flooded treatment as supported by Chapagain and Yamaji (2010). Yang et al. (2016) also reported that moderate AWDI-in which photosynthesis was not severely inhibited and plants could rehydrate overnight, during soil drying period or when water table was maintained at 10 to 15 cm below soil surface, could increase not only WUE but also grain yield.
AWDI treatments stimulated the more rice yield, which were negatively interlinked with total seasonal CH 4 flux in where more root depth was found most probably for the more soil porosity that supported more tillers and ultimately more yield of rice. This might be due to more depth of rice root support more nutrient uptake from different soil nutrient layer. Rice grain yield was increased by 11%, only in AWDI (10 cm) in where AWDI (15 cm) gave more or less similar grain yield over the control irrigated plots in our experiment. Conversely, rice grain yield was negatively correlated with seasonal CH 4 flux, which was supported by Denier van Der Gon et al. (2002). Among the all treatments, AWDI (10 cm) and AWDI (15 cm) are economically feasible and effective for mitigating seasonal CH 4 flux and increasing rice productivity over the conventional flooded rice cultivation.

Conclusion
This field experiment confirmed that AWDI is a promising method which provides duel benefits of water saving and optimum rice productivity with low CH 4 emission compared to conventional flood irrigation during dry season Boro rice cultivation. Conclusively, suitable alternate wet and dry irrigation water management practices (Irrigate when water level falls below 10 -15 cm from surface) could be implemented in farmers' field based on site specific agro-ecological zone for significant water savings, reducing methane emission and sustaining rice productivity. Further field trials in different agro-ecological zones may provide wider adaptability and farmers' acceptability.