High-temporal-resolution ammonia concentration measurements in a naturally ventilated dairy barn to inform RuFaS housing ammonia prediction refinement

Agriculture contributes approximately 80% of ammonia (NH₃) emissions globally and in the United States, with major loss pathways including animal housing, manure storage, and land application of manure and synthetic fertilizers. Although NH₃ is not itself a greenhouse gas, it is a key precursor of nitrous oxide (N₂O), a potent greenhouse gas. Current housing NH₃ emission models, including those implemented in the open-source Ruminant Farm Systems (RuFaS) model, rely on generalized parameters that may not adequately represent region- and management-specific variability, particularly in naturally ventilated barns. The objective of this study was to generate high-temporal resolution housing NH₃ concentration data using IoT-based sensors to inform refinement of housing NH₃ modeling in RuFaS toward more context-specific simulations. Measurements were conducted in a naturally ventilated free-stall dairy barn housing approximately 600 lactating cows with solid flooring and sand bedding in Harford, New York, USA. Manure was mechanically scraped 3 times per day. A total of 7 electrochemical NH₃ sensors (Cynomys, Arenzano, Italy) were deployed evenly throughout the barn at a height of 2 m. Ammonia concentrations were continuously monitored from April 2025 to January 2026 at 10-min intervals. Hourly averages were used to assess diurnal patterns, and monthly averages were calculated to evaluate seasonal trends. Indoor temperature was monitored concurrently. Indoor temperature increased from 12.04±1.52 °C in April to 24.67±0.46 °C in July, before declining to 4.32±0.89 °C in January. Hourly NH₃ concentrations ranged from 0.447±0.497 ppm to 0.714±0.369 ppm, with an overall mean of 0.554±0.088 ppm. Minimum concentrations occurred around 12:00, while maximum concentrations were observed at 23:00. Monthly mean NH₃ concentrations ranged from 0.413±0.090 ppm to 0.752±0.618 ppm, with an overall mean of 0.594±0.133 ppm; the lowest and highest monthly averages occurred in April and August, respectively. These concentration levels are generally consistent with ranges reported in the literature for dairy housing. These measurements provide a high-throughput dataset capturing diurnal and seasonal variability of housing NH₃ concentrations in a naturally ventilated dairy barn. While concentration data alone are insufficient to directly calibrate housing NH₃ emission models, the observed temporal patterns establish essential boundary conditions for subsequent emission estimation when combined with ventilation rates. In this context, the dataset supports future derivation of NH₃ emission fluxes needed to evaluate and calibrate housing NH₃ submodules in whole-farm simulation frameworks such as RuFaS. Ongoing work will integrate ventilation estimates to quantify emission fluxes.