This study aimed to assess the feasibility of predicting ranges in N2O emission with a boundary line
approach using a few key driving factors. Intact soil cores (9 cm dia. and ~20 cm in depth) were collected
from pasture, cereal cropping and sugarcane lands and incubated at various temperature and moisture
conditions after addition of different forms of mineral nitrogen (NH4+ and NO3?). The pasture and sugarcane
soils showed greater N2O production capacity than the cropping soils with similar mineral N and organic C
contents or under similar temperature and water filled pore space (WFPS%), and thus different model
parameters need to be used. The N2O emission rates were classified into three ranges: low (< 16 g
N2O/ha/day), medium (16 ?160 g N2O/ha/day) and high (> 160 g N2O/ha/day). The results indicated that
N2O emissions were in the low range when soil mineral N content was below 10 mg N/kg for the cropping
soils and below 2 mg N/kg for the pasture and sugarcane soils. In soils with mineral N content exceeding the
above thresholds, the emission rates were largely regulated by soil temperature and WFPS and the emission
ranges could be estimated using linear boundary line models that incorporated both temperature and WFPS.
Using these key driving factors (land use, temperature, WFPS and mineral N content), the boundary line
models correctly estimated the emission ranges for 85% of the 247 data points for the cropping soils and
59% of the 271 data points for the pasture and sugarcane soils. In view of the fact that N2O emissions from
soil are often very variable and difficult to predict and that the soil and environmental conditions applied in
this study differed substantially, the above results suggested that, in terms of accuracy and feasibility, the
boundary line approach provides a simple and practical alternative to the use of a single emission factor and
more complex process-based models.
