A process-based model of ammonia emissions from dairy cows: improved temporal and spatial resolution

Abstract

This research has developed an integrated model of a dairy farm that predicts monthly ammonia emission factors based on farming practices and climate conditions, including temperature, wind speed, and precipitation. The model can be used to predict the seasonal and geographic variations in ammonia emission factors, which are important for accurately predicting aerosol nitrate concentrations. The model tracks the volume of manure and mass of ammoniacal nitrogen as the manure moves through the housing, storage, application, and grazing stages of a dairy farm. Most of the processes of ammonia volatilization are modeled explicitly, but poorly understood processes are parameterized and tuned to match empirical data. The tuned model has been compared to independent experimental data and is shown to be robust over the range of experimental conditions. We have characterized the differences in emissions resulting from changes in climate conditions and farming practices and found that both of these factors are significant and should be included when developing a national inventory.

Introduction

Ammonia is an important atmospheric pollutant that plays a key role in several air pollution problems. When combined with nitric acid, ammonia forms aerosol nitrate, which contributes significantly to total particulate matter (PM) (McNaughton and Vet, 1996). These particles have serious impacts on human health (Dockery et al., 1993) and influence climate (Charlson et al., 1992). When deposited in relatively pristine areas, an abundance of ammonia can cause degradation of aquatic (Jenkinson, 2001; Howarth et al., 2002) and terrestrial (Rennenberg and Gessler, 1999) ecosystems. While deposition of other atmospheric pollutants in the United States has been decreasing, concentrations of ammonia in precipitation have increased over the past 20 years (Nilles and Conley, 2001).

The sensitivity of ammonium nitrate aerosol concentrations to ammonia varies seasonally and geographically (West et al., 1999). However, previous inventories of ammonia emissions assume uniform emission factors for livestock across all locations and seasons (Pain et al., 1998; Hutchings et al., 2001). More accurate emission inventories require emission factors that are geographically and temporally resolved.

In both Europe and the United States, the largest source of ammonia emissions is livestock, estimated to account for 70–90% of total emissions, and dairy cows are one of the largest livestock sources (Battye, 1994; USEPA, 2000; Pain et al., 1998; Hutchings et al., 2001). These emissions arise from urine patches on grazed pastures, excreta deposited onto the floor of housing facilities, manure held in storage, and volatilization during the application of manure onto fields (Sommer and Hutchings, 1997).

Variation in ammonia emission factors results from the dependence of ammonia volatilization on meteorological conditions and seasonal and regional differences in farming practices. In field studies, high temperatures and wind speeds have been shown to increase the volatilization of ammonia (Sommer et al., 1991; Demmers et al., 1998). Heavy rains cause emissions to decrease to near zero (Sommer and Olesen, 2000). In cooler climates, cows are confined in housing units for the duration of winter; manure stored during this period is not applied to the fields until spring. In warmer climates, the cows may graze throughout the year. However, the amount of variation we can explain with current scientific understanding is limited. Because of the vast number of farm types and climate conditions, all of the experimental results to date cover only a small subset of possible emission scenarios.

We have addressed this need by developing the Farm Emissions Model (FEM), an integrated model of a dairy farm similar to that presented by Hutchings et al. (1996). The FEM predicts per cow emissions given a specific set of manure management practices and a temporal profile of temperature, precipitation, and wind speed. The FEM can be combined with the geographic distribution of manure management practices, animal populations, and climate data to produce an emission inventory. The results of such an application to the United States are discussed in a future paper.

To determine their model parameters, Hutchings et al. (1996) surveyed the literature and calculated the parameters based on the best available data. A major difference in our approach is that we use a formal technique to estimate model parameters. We apply Bayesian parameter estimation and experimental data to estimate the model parameters and uncertainty. The resulting parameters reflect the range of experimental conditions found in the literature. Where possible, we then validate our model with independent experimental results. As an example application of the FEM, we present monthly emission factors for four different farms in the United States.