Identifying hotspots and representative monitoring locations of field scale N2O emissions from agricultural soils: A time stability analysis

Highlights: • Time stability and multivariate grouping analysis were used to identify N₂O hotspots. • Topographical parameters helped characterize hotspots of N₂O emissions. • Shoulder positions in agricultural landscape acted as N₂O hotspots. • Most of the coldspots and locations with cold moments were located at backslope position. • Representative N₂O monitoring points were located in depressions and leveled terrain. Greenhouse gas sampling from agricultural fields is laborious and time-consuming. Soil and topographical heterogeneity cause spatiotemporal variations, making nitrous oxide (N₂O) estimation and management a challenge. Identification of representative monitoring locations, hotspots, and coldspots could facilitate the mitigation of agricultural N₂O emissions. The objective of this study was to identify and characterize representative monitoring locations, hotspots, and coldspots of N₂O emissions in agricultural fields (Baggs farm; BF and Research North farm; RN) in Cambridge, Ontario, Canada, under humid continental climate. Soil in both fields was classified as Orthic Melanic Brunisol, with some areas categorized as Gleyed Brunisolic Gray Brown Luvisol and Orthic Humic Gleysol. In total, 28 sampling points were selected following conditional Latin hypercube design using topographical parameters (digital elevation, slope, topographical wetness index, and Pennock landform classification). Gas samples were collected over a two-year crop rotation with corn (2019) and soybean (2020). Additional sampling was conducted at BF at spring thaw (2020). Time stability analysis using mean relative difference (MRD) and standard deviation of mean relative difference (SDRD) was performed to test the hypothesis that "simultaneous analysis of spatiotemporal variations in N₂O emissions could help to identify and characterize representative monitoring locations, hotspots, coldspots and areas with few hot and cold moments. Most of the hotspots were located at shoulder positions, coldspots, and cold moments at backslope, and representative monitoring points were located at leveled positions or localized depressions. Time stability analysis coupled with multivariate groping analysis supported our hypothesis and helped successfully identify hotspots, coldspots, and representative locations based on landform classification with few exceptions. However, inclusion of additional topographical (curvature, contributing area, aspect) and morphological parameters (texture, thickness of soil horizon, depth to bedrock, and water table) are suggested for consideration in future research to manage variable-rate fertilizer application and mitigate N₂O hotspots at landscape level.