SIGNATURE WORK
CONFERENCE & EXHIBITION 2022

Anthropogenic greenhouse gas (GHG) emissions are mostly attributed to nitrogen fertilization during agricultural production. However, the interactive effects of various agricultural management practices, climatic information, soil properties, and fertilization on non-CO2 (CH4 and N2O) GHG emissions and gross global warming potential (GGWP) are scarcely discussed. In this study, a meta-analysis of 326 agricultural treatments that were conducted in China from 76 literature was done to elucidate the relationship between the response ratio (RR) of non-CO2 GHG emissions, GWP response ratio (RR), and various environmental variables through redundancy analysis (RDA). With the first and second axes, 38.87% and 3.62% of the variation can be explained by 8 explanatory environmental variables. It is also found that GGWP-RR was closely related to organic application fertilizer rate and initial bulk density while being negatively associated with soil initial total nitrogen. CH4-RR was positively associated with inorganic fertilizer application rate and N2O-RR had a positive association with initial SOC and annual mean precipitation. 10 machine learning models were tested, and the best-fitted regression models for N2O-RR, CH4-RR, and GGWP-RR had an average coefficient of determination of 38.2%, 25.5%, and 50.2%, respectively. In particular, the total fertilizer application rate is the highest contributor to both GGWP-RR and N2O-RR with annual mean precipitation contributing the most to CH4-RR. Besides, sensitivity analysis shows that GGWP-RR was mainly sensitive to synthetic fertilizer application rate and straw application rate which suggests that reducing fertilizer application by 26% and 48% decrease of straw application from the baseline are likely to be the optimum management techniques for alleviating the effect of fertilization on GGWP.