Manure distribution as a predictor of N2O emissions from soil

Abstract

Predicting nitrous oxide (N2O) emissions from manure-amended soil remains a challenge. One reason may be that spatial heterogeneity in distribution of manure is not accounted for in models of N2O emission, but experimental results suggest that both manure and soil properties affect the distribution of manure constituents after field application in a systematic way. Key to predicting the fate of labile carbon (C) and nitrogen (N) in manure is to acknowledge that the liquid phase, and a corresponding fraction of labile C and N, is partly absorbed by the bulk soil in response to the water potential gradient, and partly retained by particulate manure organic matter. Therefore, boundary conditions for subsequent transformations of C and N may be better described as two separate compartments. In this study, N2O emissions were determined in a 42-day experiment that included two soils (7.5% and 17% clay) adjusted to three soil water potentials (–3, –5 and –10 kPa) and amended with surface-applied pig slurry, cattle slurry, digestate or water only, in total 24 treatments. Net emissions of N2O corresponded to between 0.18% and 0.64% of manure N. Experimental results were analysed with a conceptual model of short-term N2O emissions from manure-amended soil, which estimates redistribution of manure constituents and predicts emissions from three sources, i.e. nitrification in bulk soil, and nitrification and denitrification in manure hotspots. Adopting a recent modification, oxygen availability in manure hotspots was related to relative soil gas diffusivity. Model efficiencies were 42% and 12% for the two soil types when using parameters determined by multiple regression of experimental results. With the process-based model Manure-DNDC as reference, the importance of accounting for distribution of manure water and labile C and N is discussed.