A meta-analysis on the impact of mid-season drainage on rice greenhouse gas emissions and yields

Context While rice systems are a globally important source of calories, they also contribute disproportionally to greenhouse gas (GHG) emissions, driven by the flooded culture. Introducing non-continuous flooding practices (NCFP) can alleviate this problem but can result in yield loss. Mid-season drainage (MSD) is a form of NCFP that has been practiced in East Asia for centuries with the aim of increasing yield. Today, an MSD can be practiced for several reasons which influences timing; so here we define it as a single intentional drainage event between tillering and flowering. Objective While the MSD practice is sometimes included in NCFP reviews, it has never been focused on. Thus, our objective was to conduct a quantitative review on how an MSD impacts GHG emissions (CH 4 and N 2 O) and rice yields relative to a continuously flooded (CF) system. Methods The final data set with side-by-side comparisons of MSD and CF had 21 studies (71 observations). Uniquely, in this study we also quantified the post-drainage emissions (total post-MSD and end-of-season spike) from figures provided. Using meta-analytical methods, we quantified the overall impacts of an MSD, as well as how various soil properties and agronomic/water management practices influenced MSD outcomes. Results Overall, relative to the CF, MSD reduced seasonal CH 4 emissions by 45%, post-MSD CH 4 emissions by 69% and end-of-season CH 4 spike emissions by 62%. While N 2 O emissions increased significantly on a relative basis, the actual increase was low in terms of GWP. In terms of yield, the MSD yields were non-significantly (0.5%) lower. The primary moderating variable that impacted CH 4 emissions was drainage severity. A severe MSD reduced seasonal CH 4 emissions by 50% compared to 28% for mild drainage. In this analysis, on average, 58% of the total seasonal CH 4 emissions in the CF treatment occurred after the MSD was implemented. For the post-MSD CH 4 emissions, a severe drainage reduced emissions by 73% compared to 48% for mild drainage. Rice yields were not significantly impacted by any of the moderating variables, suggesting broad applicability. Conclusion and significance An MSD can be used for several reasons. Here the focus is on the mitigation of CH 4 emissions. In the first such analysis of its kind, we found that a severe MSD can provide significant reductions in CH 4 emissions and GWP, with low risk of yield loss. Future studies need to better quantify soil moisture conditions during the drainage to provide better recommendations.