Landscape‐Scale Variations in Denitrification

Abstract Denitrification from agricultural ecosystems is regarded as a major contributor to atmospheric N levels, but the actual rates of denitrification and the controls on these rates remain poorly understood. This study was conducted to examine landscape‐scale patterns of denitrification and the soil properties that control these patterns. A 110 by 110 m sampling grid was established in an irrigated field in an aridic Boroll (Brown Chernozemic) soil in southern Saskatchewan. The measured soil properties (denitrification rate, respiration rate, volumetric water content, bulk density, soluble organic and inorganic C, total and mineral N, in situ pH, and in situ redox potential) were correlated to the derived landform elements at the site to determine landscapescale patterns and relationships. The soil properties occurred in one of three spatial patterns: (i) a random pattern for mineral N; (ii) a diagonal pattern for pH, soluble organic and inorganic C, and total N; and (iii) a depression‐centered pattern for denitrification, bulk density, moisture, respiration, and redox potential. Denitrification activity displayed a distinct landscape‐scale pattern and statistically distinct rates of denitrification were associated with the different landform elements: rates were lowest in the shoulder elements, intermediate in the footslope and level‐convex elements, and highest in the level‐concave elements. Hot spots of denitrification activity, i.e., sampling sites with denitrification rates identified statistically as outliers, were all associated with the level elements and, predominantly, the level‐concave elements. For the data set as a whole, denitrification was most highly correlated with volumetric water content ( r s = 0.448**) and soil redox potential ( r s = −0.335**). In the level landform elements, volumetric water content was the soil variable most highly correlated with denitrification; in the shoulder and footslope elements, respiration and bulk density, respectively, were the soil variables most highly correlated with denitrification. These data indicate that (i) topography has a strong influence over in situ denitrification, (ii) different controls are at work in the various landform elements, and (iii) landscape analysis should be incorporated in the process of modeling total N losses through denitrification.