How much is soil nitrous oxide emission reduced with biochar application? An evaluation of meta‐analyses

Abstract Nitrous oxide (N 2 O) is the third important long‐lived greenhouse gas next to carbon dioxide and methane and croplands are considered biogeochemical hotspots of soil N 2 O emissions. To reduce soil N 2 O and other greenhouse emissions, climate‐smart agricultural practices including biochar application have been applied. Many studies have been conducted with biochar application but results from these studies are not conclusive. To address this issue, meta‐analysis, a quantitative review that synthesizes results from multiple independent studies, has been widely used. The results from different meta‐analyses also differ but are seldomly evaluated. In this study, we evaluated meta‐analyses on the effects of biochar application on soil N 2 O emissions. A grand mean response ratio (RR) was further proposed to estimate an overall effect and the impacts of experiment setting, properties of biochar and soil, and agricultural practices. We found 18 meta‐analysis papers were published between 2014 and 2022. Sample size (publications or experiments) varied from less than 30 to more than 1000, with a mean sample size of 275. RR was calculated in all studies except one. While four meta‐analyses did not find a significant effect of biochar application on soil N 2 O emissions, all others reported reductions of soil N 2 O emissions, but the magnitude ranged from −10.5% to −54.8%. Synthesizing all results from these meta‐analyses, we found that biochar application overall significantly reduced the soil N 2 O emissions by 38.8%. The impacts increased with experimental duration till one and half years and reduced after that. Biochar application rate and C:N ratio had large influence on the effects of biochar application on soil N 2 O emissions. This study demonstrated that while meta‐analysis provides a more comprehensive and better estimation, the inconsistence among these studies may need to be further evaluated. A grand mean RR based on meta‐analyses could be more accurate and representative than single meta‐analysis.