Rice cultivation has been challenged by increasing food demand and water scarcity. We examined the responses of water use, grain yield, and water productivity to various modes of field water managements in Chinese double rice systems. Four treatments were studied in a long-term field experiment (1998–2015): continuous flooding (CF), flooding—midseason drying—flooding (F-D-F), flooding—midseason drying—intermittent irrigation without obvious standing water (F-D-S), and flooding—rain-fed (F-RF). The average precipitation was 483 mm in early-rice season and 397 mm in late-rice season. The irrigated water for CF, F-D-F, F-D-S, and F-RF, respectively, was 263, 340, 279, and 170 mm in early-rice season, and 484, 528, 422, and 206 mm in late-rice season. Grain yield for CF, F-D-F, F-D-S, and F-RF, respectively, was 4,722, 4,597, 4,479, and 4,232 kgha^-1 in early-rice season, and 5,420, 5,402, 5,366, and 4,498 kgha^-1 in late-rice season. Compared with CF, F-D-F consumed more irrigated water, which still decreased grain yield, leading to a decrease in water productivity by 25% in early-rice season and by 8% in late-rice season. Compared with F-D-F, F-D-S saved much irrigated water with a small yield reduction, leading to an increase in water productivity by 22% in early-rice season and by 26% in late-rice season. The results indicate that CF is best for early-rice and FDS is best for late-rice in terms of grain yield and water productivity.

This study provides a complete account of global warming potential (GWP) and greenhouse gas intensity (GHGI) in relation to a long-term water management experiment in Chinese double-rice cropping systems. The three strategies of water management comprised continuous (year-round) flooding (CF), flooding during the rice season but with drainage during the midseason and harvest time (F-D-F), and irrigation only for flooding during transplanting and the tillering stage (F-RF). The CH₄ and N₂O fluxes were measured with the static chamber method. Soil organic carbon (SOC) sequestration rates were estimated based on the changes in the carbon stocks during 1998–2014. Longer periods of soil flooding led to increased CH₄ emissions, reduced N₂O emissions, and enhanced SOC sequestration. The net GWPs were 22,497, 8,895, and 1,646 kg CO₂-equivalent ha⁻¹ yr⁻¹ for the CF, F-D-F, and F-RF, respectively. The annual rice grain yields were comparable between the F-D-F and CF, but were reduced significantly (by 13%) in the F-RF. The GHGIs were 2.07, 0.87, and 0.18 kg CO₂-equivalent kg⁻¹ grain yr⁻¹ for the CF, F-D-F, and F-RF, respectively. These results suggest that F-D-F could be used to maintain the grain yields and simultaneously mitigate the climatic impact of double rice-cropping systems.

Litter decomposition plays a vital role in wetland carbon cycling. However, the contribution of aboveground litter decomposition to the wetland soil organic carbon (SOC) pool has not yet been quantified. Here, we conducted a Carex brevicuspis leaf litter input experiment to clarify the intrinsic factors controlling litter decomposition and quantify its contribution to the SOC pool at different water levels. The Carex genus is ubiquitous in global freshwater wetlands. We sampled this plant leaf litter at −25, 0, and +25 cm relative to the soil surface over 280 d and analysed leaf litter decomposition and its contribution to the SOC pool. The percentage litter dry weight loss and the instantaneous litter dry weight decomposition rate were the highest at +25 cm water level (61.8 %, 0.01307 d⁻¹), followed by the 0 cm water level (49.8 %, 0.00908 d⁻¹), and the lowest at −25 cm water level (32.4 %, 0.00527 d⁻¹). Significant amounts of litter carbon, nitrogen, and phosphorus were released at all three water levels. Litter input significantly increased the soil microbial biomass and fungal density but had nonsignificant impacts on soil bacteria, actinomycetes, and the fungal∕bacterial concentrations at all three water levels. Compared with litter removal, litter addition increased the SOC by 16.93 %, 9.44 %, and 2.51 % at the +25, 0, and −25 cm water levels, respectively. Hence, higher water levels facilitate the release of organic carbon from leaf litter into the soil via water leaching. In this way, they increase the soil carbon pool. At lower water levels, soil carbon is lost due to the slower litter decomposition rate and active microbial (actinomycete) respiration. Our results revealed that the water level in natural wetlands influenced litter decomposition mainly by leaching and microbial activity, by extension, and affected the wetland surface carbon pool.