| The microbial process of denitrification has been of great scientific interest, because one of its products, nitrous oxide, is both a greenhouse gas and a contributor to stratospheric ozone depletion. However, current methods do not allow for measuring denitrification at scales relevant to regional and global atmospheric chemistry.; This study explores whether or not soil gases might be used as a proxy for denitrification rate measurements across an agricultural landscape that includes cropland (i.e., cornfield), pasture, and forest. This goal is accomplished by evaluating the use of soil gases, alone or in combination with various soil parameters, as predictors of denitrification rate using a multiple regression modeling approach.; When exploring overall models (which include all three land use types) for the landscape studied, the best predictive model using soil gases as explanatory variables includes: soil nitrous oxide concentration, gravimetric water content, and land use type (R2 = 0.63). Without considering soil gases, the best overall model (R2 = 0.64) includes land use type, ammonium content, and gravimetric water content.; In assessing individual land use models, soil gases were not predictive of denitrification for the cornfield: however, a model R2 of 0.60 is attained using gravimetric water content, ammonium, and total nitrogen content as predictors. For the forest, the best model using soil gases as predictors includes soil nitrous oxide and season as predictors (R 2 = 0.65), although using ammonium content, bulk density and season gave a slightly higher R2 (0.71). For the pasture, the use of soil gases as predictors did not provide much benefit, as volumetric water content alone accounted for 70% of denitrification variability.; The models selected in this study generally account for a fairly high degree of variability in denitrification activity (i.e., 61 to 75%). Moreover, the overall and forest models which include soil gases as predictors performed better in the model validation analysis than the models which did not include soil gases as predictors, and thus may be more transportable to other sites. |
