Soil nitrous oxide (N₂O) emissions exhibit high variability in intensively managed cropping systems, which challenges our ability to understand their complex interactions with controlling factors. We leveraged 17-years (2003-2019) of measurements at the Kellogg Biological Station LTER/LTAR site to better understand controls of N₂O emissions in four corn–soybean–winter wheat rotations employing Conventional, No-till, Reduced input, and Biologically-based/organic inputs. We used a Random Forest machine learning model to predict daily N₂O fluxes, trained separately for each system with 70% of observations, using variables such as crop species, daily air temperature, cumulative 2-day precipitation, water-filled pore space, and soil nitrate and ammonium concentrations. The model explained 29 to 42% of daily N₂O flux variability in test data, with greater predictability for the corn phase in each system. The long-term rotations showed different controlling factors and threshold conditions influencing N₂O emissions. In the Conventional system, the model identified ammonium (>15 kg N ha^-1) and daily temperature (>23 °C) as the most influential variables; in the No-till system, climate variables, precipitation, and temperature were important variables. In low input and organic systems, where red clover (Trifolium repens L.; before corn) and cereal rye (Secale cereale L.; before soybean) cover crops were integrated, nitrate was the predominant variable, followed by precipitation and temperature. In low input and biologically-based systems, red clover residues increased soil nitrogen availability to influence N₂O emissions. Long-term data facilitated machine learning for predicting N₂O emissions in response to differential controls and threshold responses to management, environmental, and biogeochemical drivers.

This study aimed to identify critical management, environmental, and biogeochemical drivers of nitrous oxide (N₂O) emissions and their differential relationships and threshold conditions for emissions under diverse long-term cropping rotations. Historical 17-y (2003 to 2019) data on yearly N₂O emissions, soil properties, agricultural management practices, and weather were obtained from the Main Cropping System Experiment (MCSE) of the Kellogg Biological Station (KBS) Long-Term Ecological Research (LTER) data catalog (https://lter.kbs.msu.edu/datatables). Data were extracted from four treatments, including: (i) a Conventional system with chisel tillage and standard chemical inputs, (ii) a No-till system with standard chemical inputs, (iii) a Reduced input system with chisel tillage, low fertilizer inputs, and cover crops, and (iv) a Biologically-based/organic system managed organically using chisel tillage, cover crops, and no synthetic chemical inputs. Nitrous oxide flux measurements were made using static chambers at weekly to monthly intervals. Details on N₂O flux measurements can be found at https://lter.kbs.msu.edu/protocols/113. During each gas sampling event, gravimetric water content (GWC) was measured from the 0-25 cm depth. Water-filled pore space (WFPS) was determined using the GWC and a consistent bulk density of 1.44 g cm^-3 across all treatments. Soil samples (0 to 25 cm depth) were collected biweekly each year for NH₄⁺ and NO₃^- concentrations. Soils were extracted with 2M KCl and analyzed for NH₄⁺ and NO₃^-concentrations in a continuous flow analyzer (Alpkem 3550, O.I. Analytical, College Station, TX). Soil NH₄⁺, NO₃^-, and WFPS values were linearly interpolated between two sampling dates to match the N₂O sampling dates when soil samples were not collected on the gas sampling days due to management and weather reasons.