Applying an Inverse Model to Estimate Ammonia Emissions at Cattle Feedlots Using Three Different Observation-Based Approaches

Abstract:

Accurately quantifying emissions of ammonia (NH₃) from confined animal feeding operations (CAFOs) is vital not only to the livestock industry, but essential to understanding nitrogen cycling along the Front Range of Colorado, USA, where intensive agriculture, urban sprawl, and pristine ecosystems (e.g., Rocky Mtn Nat’l Park) lie within 100-km of each other. Most observation-based techniques for estimating NH₃ emissions can be expensive and highly technical. Many methods rely on concentration observations on location, which implicitly depends on weather conditions. A system for sampling NH₃ using on-site weather data was developed to allow remote measurement of NH₃ in a simple, cost-effective way. These systems use passive diffusive cartridges (Radiello, Sigma-Aldrich) that provide time-averaged concentrations representative of a typical two-week deployment. Cartridge exposure is robotically managed so they are only visible when winds are 1.4 m/s or greater from the direction of the CAFO. These concentration data can be coupled with stability parameters (measured on-site) in a simple inverse model to estimate emissions (FIDES, UMR Environnement et Grandes Cultures). Few studies have directly compared emissions estimates of NH₃ using concentration data obtained from multiple measurement systems at different temporal and spatial scales. Therefore, in the summer and autumn of 2014, several conditional sampler systems were deployed at a 25,000-head cattle feedlot concomitant with an open-path infrared laser (GasFinder2, Boreal Laser Inc.) and a Cavity Ring Down Spectrometer (CRDS) (G1103, Picarro Inc.) which each measured instantaneous NH₃ concentrations. This study will test the sampler technology by first comparing concentration data from the three different methods. In livestock research, it is common to estimate NH₃ emissions by using such instantaneous data in a backward Lagrangian stochastic (bLs) model (WindTrax, Thunder Beach Sci.) Considering this, NH₃ fluxes from the inverse model (FIDES) using all three datasets will be compared to emissions from the bLS model (WindTrax) using only high speed data (laser; CRDS). Results may lend further validity to the conditional sampler approach for more easily and accurately monitoring NH₃ fluxes from CAFOs and other strong areal sources.