Emissions variability of greenhouse gases within and between sites during late‐season corn production in Mississippi

Abstract Variability of greenhouse gas (GHG) fluxes and emissions among replicates in a study area can have an impact on results depending on the statistical tools used to assess significant differences among fixed effects. The objective of this study was to assess the effect of site and time (i.e., date) and the variability within and between sites on soil GHG (carbon dioxide [CO 2 ], methane [CH 4 ], and nitrous oxide [N 2 O]) fluxes and emissions from two fields under the same management practices during late‐season corn ( Zea mays ) production in Mississippi. Established techniques, such as analysis of variance, variance component analysis, linear regression, and 95%‐confidence intervals, were used to determine potential differences among analytical approaches. Carbon dioxide differed ( p < 0.05) by date, N 2 O differed ( p < 0.05) by site across dates, and CH 4 did not differ ( p > 0.05) among fixed factors, while only N 2 O emissions and yield differed ( p < 0.05) by site. Nitrous oxide emissions and yield were greater at site 2 (1.09 kg N 2 O ha −1 , 11.7 Mg yield ha −1 ) than at Site 1 (0.35 kg N 2 O ha −1 , 6.54 Mg yield ha −1 ). Generally, within‐site variability carried >50% of the variance compared to the between‐site variability for all response variables. The various analytical tools resulted in complementary results. Spatial heterogeneity within each site suggested that replicates in GHG studies within an agricultural field can be considered true replicates and that experimental designs need to account for spatio‐temporal relationships of GHG production and release.