Soil freeze-thaw induces a pulse of nitrous oxide (N2O) emissions fueled by a concomitant increase in available organic carbon (C) and nitrogen (N) substrates. These substrates are hypothesized to originate from the disruption of aggregates and microbial biomass, but experiments designed to falsify these hypotheses have been scarce. We therefore conducted a series of column experiments using intact soil cores of silt loam and loamy sand under different freezing rates and durations, factors previously identified as controlling the magnitude of subsequent N2O emissions. We used a slower freezing rate and shorter freeze duration (6 days) as a Control, to which we compared the same freeze duration but faster freezing rate (Fast, 6 days) and the same freezing rate but longer freeze duration (Long, 21 days). All soils were frozen to ~ -2.5 °C. Upon thaw, only the silt loam emitted N2O; we therefore focused on silt loam C, N, and aggregate dynamics. Control and Fast soils emitted similar amounts of N2O, even though Fast soils maintained aggregate mean weight diameter (MWD). Long soils emitted 1.8 times more N2O than Control, even though Long and Control soils exhibited similar decreases in aggregate MWD. We suggest that because in no instance were differences in N2O emissions between treatments proportional to differences in aggregate disruption, and some soils emitted N2O without any detectable aggregate disruption, that aggregate disruption does not play a significant, universal role in providing substrates for N2O emissions. In contrast, larger N2O emissions due to longer freezing durations were associated with a delayed decrease in microbial biomass. Although we did not assess microbial communities for features other than their biomass, our observations are consistent with longer freezing durations selecting for slower growing microorganisms and freeze-thaw induced N2O emissions being fueled by microbial osmolytes.
