ABSTRACT Water management impacts both methane (CH4) and nitrous oxide (N2O) emissions from rice paddy fields. Although controlled irrigation is one of the most important tools for reducing CH4emission in rice production systems it can also increase N2O emissions and reduce crop yields. Over three years, CH4 and N2O emissions were measured in a rice field in Uruguay under two different irrigation management systems, using static closed chambers: conventional water management (continuous flooding after 30 days of emergence, CF30); and an alternative system (controlled deficit irrigation allowing for wetting and drying, AWDI). AWDI showed mean cumulative CH4 emission values of 98.4 kg CH4 ha−1, 55 % lower compared to CF30, while no differences in nitrous oxide emissions were observed between treatments ( p > 0.05). No yield differences between irrigation systems were observed in two of the rice seasons ( p > 0.05) while AWDI promoted yield reduction in one of the seasons ( p< 0.05). When rice yield and greenhouse gases (GHG) emissions were considered together, the AWDI irrigation system allowed for lower yield-scaled total global warming potential (GWP). Higher irrigation water productivity was achieved under AWDI in two of the three rice seasons. These findings suggest that AWDI could be an option for reducing GHG emissions and increasing irrigation water productivity. However, AWDI may compromise grain yield in certain years, reflecting the importance of the need for fine tuning of this irrigation strategy and an assessment of the overall tradeoff between relationships in order to promote its adoption by farmers.

The bacterial communities in floodwater, from a rice-planted and an unplanted field were characterized at the beginning (flooding stage) and at the end (harvest stage) of the rice cropping cycle. Most probable number estimations and plate counts of aerobic and anaerobic heterotrophic bacteria and of several metabolic bacterial groups (methanogens, sulfate-reducers, anaerobic sulfur and nonsulfur phototrophs, denitrifiers and ammonifiers) were similar in rice and unplanted floodwater at both sampling times. The analysis of denitrifiers and methanogens by fluorescent in situ hybridization revealed a shift in the phylogenetic affiliation only of the former group in the rice-planted floodwater. Terminal restriction fragment length polymorphism of 16S rRNA gene amplicons indicated that the bacterial communities of the rice-planted and unplanted soils were consistently diverse and strongly influenced by the season.