High nitrous oxide emissions from temporary flooded depressions within croplands

Nitrous oxide (N₂O), a potent greenhouse gas with a global warming potential 298 times that of CO₂, is largely emitted from agricultural activities due to the use of nitrogen-based fertilizers and manure. Its production in soil involves complex microbial processes like nitrification and denitrification, heavily influenced by oxygen availability. Understanding the spatial and temporal variability of N₂O emissions, especially identifying hotspots (locations with exceptionally high emissions) and hot moments (short periods of high emissions), is crucial for effective mitigation strategies. This study focuses on temporarily flooded depressions within croplands as potential hotspots for accelerated denitrification, particularly in the period immediately following fertilization when nutrient-rich water accumulates in these poorly drained areas. Denmark, with its intensive farming practices and relatively flat croplands containing numerous depressions, provides an ideal context to investigate this phenomenon. Previous studies have established links between N₂O hotspots and specific soil types or conditions, but the contribution of flooded depressions in mineral soils within intensively managed croplands has remained largely unquantified. This research aims to address this knowledge gap by combining detailed in-situ N₂O flux measurements from flooded and non-flooded areas with a large-scale assessment of the spatial extent of flooded depressions using remote sensing techniques.

Existing literature highlights the significant contribution of agricultural N₂O emissions to global warming and emphasizes the importance of understanding spatial and temporal variability. Several studies have identified "hot moments" of increased N₂O emissions linked to specific events, such as drought followed by rewetting or spring thaw. Research also points to the role of soil type and management practices in influencing N₂O emissions, with organic soils under warm, well-drained conditions identified as potential hotspots. However, a significant knowledge gap exists regarding the role of temporarily flooded depressions in mineral soils within intensively managed croplands. Previous national greenhouse gas budgets have often overlooked the contribution of these small-scale features. This study builds upon existing knowledge by quantifying N₂O emissions from these previously neglected hotspots and relating them to the wider landscape.

The study involved a multi-faceted approach combining field measurements, laboratory experiments, and remote sensing. **Mapping Flooded Fields:** A deep learning model (U-Net architecture) was trained to segment ephemeral water bodies in high-resolution PlanetScope satellite imagery (3-m spatial resolution) across Zealand, Denmark. The model integrated optical and topographic data (LiDAR-based DEM at 0.4 m resolution) to improve accuracy. A total of 20,063 individual flooded areas were identified, representing 0.5% of the total farmed area. The rim area (5 m buffer around each depression) was identified as a key hotspot for N₂O emissions. **N2O Flux Measurements:** In-situ N₂O fluxes were measured using static chambers at 102 flooded fields. Measurements were taken at three positions: center, rim, and upslope (control). A photo-acoustic gas monitor measured N₂O concentrations for 12 minutes in each chamber, with soil temperature, moisture, pH, and nitrate concentrations also recorded. The non-linear increase in N₂O concentrations was fitted to determine fluxes. **Laboratory Fertilization Experiment:** A controlled experiment used intact soil cores from a typical loamy soil. Cores were subjected to three water content levels (drained, ambient, and flooded) and three nitrate addition levels (low, normal, and high) to simulate different fertilization scenarios. N₂O emissions were measured using an isotopic N₂O analyzer over 75 days at 7°C. This allowed for direct comparison of emission rates under controlled environmental conditions. **Data Analysis:** Statistical analysis was used to compare N₂O fluxes across different field locations and treatments. A conservative estimate of the overall N₂O budget contribution from flooded depressions was calculated by extrapolating the field measurements to the entire study area.

The study revealed exceptionally high N₂O emissions from the rim areas of flooded depressions, significantly exceeding those from the center or surrounding drained areas. Specifically: * **Rim Area Emissions:** Average N₂O fluxes from the rim area were 652 ± 118 µg N m⁻² h⁻¹, more than 80 times higher than adjacent locations (8 ± 2 µg N m⁻² h⁻¹). * **Spatial Variability:** A strong gradient in emissions was observed, with the highest fluxes at the rim, decreasing towards the center of the flooded depression. * **Temporal Persistence:** High N₂O fluxes persisted for more than two months after fertilization. * **Landscape Scale Contribution:** Scaling up the findings to the entire island of Zealand, the 5-m rim of the 20,063 identified flooded areas (representing 0.26% of farmland) accounted for approximately 30 ± 1% of the total N₂O emissions during the two-month study period. * **Laboratory Confirmation:** The laboratory experiment corroborated the field findings, showing significantly higher N₂O emissions under flooded conditions, particularly at normal and high nitrate addition levels. Under normal fertilization, flooded cores released 370 times more N₂O than ambient conditions. * **Emission Factors:** The laboratory incubation experiment indicated that more than 5-8% of added nitrate was released as N₂O. This contrasted with the 1% emission factor typically used in national greenhouse gas inventories.

The results strongly demonstrate that temporarily flooded depressions within croplands are significant N₂O emission hotspots. The high fluxes observed in these areas are primarily driven by the combination of high nitrate concentrations (resulting from fertilization and runoff) and anoxic conditions created by waterlogging. The strong gradient in emissions from the rim to the center of the depressions suggests a rapid nitrate consumption at the rim, with less available nitrate reaching the center. This emphasizes the importance of considering spatial heterogeneity when assessing N₂O emissions. The persistence of high fluxes for extended periods post-fertilization highlights the importance of long-term monitoring. The considerable contribution of these relatively small areas to the overall N₂O budget underscores the potential for targeted mitigation strategies focused on managing these hotspots. This study's findings challenge existing assumptions in national greenhouse gas inventories, highlighting the need for revised emission factors and more refined spatial assessments.

This study provides compelling evidence that temporarily flooded depressions in croplands are substantial N₂O emission hotspots. These small areas contribute disproportionately to the overall N₂O budget, highlighting the need for targeted mitigation strategies. The high emission rates observed are linked to the combination of high nitrate concentrations from fertilization and runoff and the anoxic conditions in these depressions. The findings call for revising N₂O emission factors currently used in national inventories and incorporating spatial heterogeneity in assessments. Further research should focus on the influence of soil type, climate, and management practices on N₂O emissions from these hotspots to improve mitigation strategies.

While the study provides valuable insights into N₂O emissions from flooded depressions, some limitations exist. The study focused on a specific region (Zealand, Denmark) and a limited time period, so generalizability to other regions with different climates and soil types may be limited. The use of a single soil type in the laboratory experiment could also limit the generalizability of those findings. The reliance on remote sensing for identifying flooded areas introduces potential uncertainties associated with image resolution and classification accuracy. Further research incorporating more diverse locations, soil types, and extended time periods is necessary to strengthen the generalizability of the findings.