Spatial and temporal patterns of nitrous oxide and their relationship to soil water and soil properties

Abstract

Soil N2O flux is sensitive to soil moisture content and soil temperature, which are in turn sensitive to changes in climate and topography. Thus, N2O flux measurements exhibit a high degree of spatial and temporal variability. Knowing how the spatial distribution of soil N2O flux changes over time in a hummocky, agricultural landscape will identify measurement scales appropriate for estimates of N2O emissions from these types of terrains. As well, little is known about N2O emissions from uncultivated, ephemeral wetlands in agricultural landscapes, but this information is needed for accurate inventories of N2O emissions. The objectives of this study were to describe the spatial and temporal distribution of soil N2O flux in a hummocky agricultural landscape, and to understand how soil water and soil temperature control the spatial and temporal patterns of N2O flux. For a hummocky, agricultural landscape in the Dark Brown soil zone of Saskatchewan, N2O flux and related soil variables were measured along a 128-point transect multiple times over two years and concurrently from a 50 point, stratified design over three years. The spatial and temporal variation in N2O flux followed an event-based / background emission pattern. High flux events were triggered by precipitation events and recession of water from wetlands following spring snowmelt. Days with high mean flux were characterized by highly skewed (reverse J-shaped) distributions. High variance and coherency was observed at cultivated wetland elements during emission events. Strong location-dependent positive relationships were found between soil N2O flux and water-filled pore space or soil temperature, related to specific landscape elements. Background emissions were characterized by random variation or cyclic behavior that ranged in scale from 20 to 60 m. Cumulative emissions were highest from cultivated wetlands and basin centers of uncultivated wetlands, although emissions from cultivated wetlands were much more important to total cumulative emissions on an area basis. The results indicate that models intended to estimate N2O flux from these landscapes cannot rely on a single predictive relationship, but will have to incorporate predictive relationships localized at specific landscape elements depending on the time of year. At certain times predictive relationships cannot be used and up-scaled estimates will have to rely on direct measurement of emissions.