Measurement of non-enteric emission fluxes of volatile fatty acids from a California dairy by solid phase micro-extraction with gas chromatography/mass spectrometry

Abstract

Dairies are a major source of volatile organic compounds (VOCs) in California's San Joaquin Valley; a region that experiences high ozone levels during summer. Short-chain carboxylic acids, or volatile fatty acids (VFAs), are believed to make up a large fraction of VOC emissions from these facilities, although there are few studies to substantiate this. In this work, a method using a flux chamber coupled to solid phase micro-extraction (SPME) fibers followed by analysis using gas chromatography/mass spectrometry was developed to quantify emissions of six VFAs (acetic acid, propanoic acid, butanoic acid, pentanoic acid, hexanoic acid and 3-methyl butanoic acid) from non-enteric sources. The technique was then used to quantify VFA fluxes from a small dairy located on the campus of California State University Fresno. Both animal feed and animal waste are found to be major sources of VFAs, with acetic acid contributing 70–90% of emissions from the sources tested. Measured total acid fluxes during spring (with an average temperature of 20 °C) were 1.84 ± 0.01, 1.06 ± 0.08, (1.3 ± 0.5) × 10⁻², (1.7 ± 0.2) × 10⁻² and (1.2 ± 0.5) × 10⁻² g m⁻² h⁻¹ from silage, total mixed rations, flushing lane, open lot and lagoon sources, respectively. VFA emissions from the sources tested total 390 ± 80 g h⁻¹. The data indicate high fluxes of VFAs from dairy facilities, but differences in the design and operation of dairies in the San Joaquin Valley as well as seasonal variations mean that additional measurements must be made to accurately determine emissions inventories for the region.

Introduction

The San Joaquin Valley in central California is widely regarded as having some of the worst air quality in the United States with high mass loadings of particulate matter during the winter, and high ozone levels during the summer. Despite reductions in emissions of volatile organic compounds (VOCs) and oxides of nitrogen (NO_x) which generate ozone in photochemical reactions in the atmosphere, the region violated the national 8-h ozone standard on 86 days during 2006 (CARB, 2007). The region is currently classified as being in non-attainment of the federal ozone standard. The San Joaquin Valley Air Pollution Control District (SJVAPCD) recently requested that the Environmental Protection Agency downgrade the region to being in extreme non-attainment (SJVAPCD, 2007), a classification currently applying only to the Los Angeles air basin.

According to California Air Resources Board estimates, dairy facilities are one of the largest single sources of VOCs in the San Joaquin Valley, accounting for about 40 tons day⁻¹ in 2005 (CARB, 2006). San Joaquin Valley dairies house over 2 million cows (NASS, 2002, CDFA, 2006), and so these emissions translate into about 5.8 kg emitted by each cow per year. In 2004, a change in state law meant that agricultural operations were no longer exempt from emission controls. With significant emission reductions planned from agricultural sources, VOC emission estimates for dairies have started to receive close attention.

The CARB emission estimates are based on a study of total organic gas emissions from Holstein and Jersey cows published in 1938 (Ritzman and Benedict, 1938). Non-methane VOCs are assumed to account for 8% of the total organic compounds (TOCs) measured (EPA, 1980), which are primarily methane. In general, VOC emissions are believed to be sensitive to environmental conditions such as temperature and humidity levels as well as the operational protocols followed by each facility (Crow, 2005). Since the majority of published studies were conducted outside California, these data cannot be directly used to determine emission factors for California dairies.

Dewhurst et al. (2001) analyzed rumen gases from Holstein–Friesian dairy cows using selected ion flow tube mass spectrometry. Hydrogen sulfide, methyl sulfide and dimethyl sulfide were the major gases identified, and their levels decreased with time following feeding. Several alcohols, carbonyls and VFAs were also measured at concentrations that were several orders of magnitude lower than for the sulfides. Cai et al. (2006) also identified about 50 compounds including VFAs in beef cattle rumen gas, but the levels were not quantified.

Hobbs et al. (2004) measured emissions from dairy cattle manure in the UK by sorbent trap-GC/MS. About 20 compounds were identified with dimethyl sulfide and acetic acid being the most prevalent organics. VOC emission rates were then estimated by assuming that VOC and ammonia fluxes from dairy waste are correlated and then using established ammonia emission inventories. Using this approach, total VOC emissions from dairy cow waste are estimated to be 9 kg cow⁻¹ year⁻¹, with over half of this total (4.9 kg cow⁻¹ year⁻¹) consisting of acetic acid. Other C2–C5 VFAs contribute an additional 0.5 kg cow⁻¹ year⁻¹ to the total. In contrast, Spinhirne et al. ( 2004)) reported that VFA emissions from beef cattle pens in Texas are only 0.3 kg cow⁻¹ year⁻¹. A difference between the samples analyzed in these studies was the water content. In the Hobbs et al. (2004) study, dry waste accounted for about 2–8% of the sample by mass whereas the Texas study investigated dried manure. Since the VFAs are generated by bacterial degradation of the cattle slurry, an increase in emissions with water content is not unexpected. A second difference between the studies relates to the diet of the cows. Miller and Varel (2001) measured emissions from manure samples collected at cattle feedlots in Nebraska. The study found a strong anti-correlation between starch content and VFA fluxes indicating that diet plays an important role in determining VFA emissions from animal waste.

Recently, Shaw et al. (2007) used proton transfer reaction mass spectrometry to measure emissions of VOCs from an environmental chamber housing dairy cows and their waste. The authors identify about 20 VOCs emitted by the cows and their waste with a flux corresponding to approximately 1.3 kg cow⁻¹ year⁻¹, assuming a constant annual flux. Methanol, dimethylsulfide, acetone/propanal, 4-methyl-phenol and acetic acid are the dominant species emitted. They conclude that non-methane VOCs make up about 1.3% of the total organics emitted.

A wide range of organic compounds has been detected within and downwind of dairy facilities. Filipy et al. (2006) measured levels of a range of VOCs including alcohols, carbonyls, esters, ethers, aromatics, sulfides and terpenes at a dairy in Washington State using GC/MS following sample collection with sorbent tubes and cryogenic traps. The researchers identified 82 compounds emitted by freestalls with ethanol being the major organic with an emission rate during summer equivalent to 6.4 kg cow⁻¹ year⁻¹. Seasonal variations were observed in these emissions, with smaller fluxes during the winter. Other work has quantified levels of a range of VOCs in samples collected at dairies in northern California (Rabaud et al., 2003) and northern Sweden (Sunesson et al., 2001). Identified compounds included some VFAs such as acetic acid, valeric acid, isovaleric acid and hexanoic acid. A second study carried out in northern Europe (Martensson et al., 1999) measured levels of VFAs within livestock buildings, including some containing dairy cows, using ion exclusion chromatography. Ethanoic acid and lactic acid were the major acids present and smaller amounts of methanoic acid, propanoic acid and butanoic acid were detected.

In 2005, the San Joaquin Valley Air Pollution Control District evaluated the available literature on ROG emissions from dairies in an attempt to improve upon the existing emissions estimate of 5.8 kg cow⁻¹ year⁻¹. Following this review, the district proposed a new emission rate of 8.8 kg cow⁻¹ year⁻¹ (Crow, 2005). The report estimated that VFAs accounted for the majority of these emissions with fluxes of 6.9 kg cow⁻¹ year⁻¹. Non-enteric emissions account for 3.3 kg cow⁻¹ year⁻¹ of this total. The report noted that there are significant uncertainties in these estimates primarily due to the lack of data available for dairies with management practices and conditions that are similar to San Joaquin Valley dairies. In this estimate, only the animal waste was taken into consideration in evaluating non-enteric sources. Dairy cattle are fed total mixed rations (TMR) consisting primarily of silage, which can contain several percent of VFAs by mass. Thus both the TMR and the stored silage are also likely to be significant sources of VFAs at dairy facilities. Given the large uncertainties in VFA fluxes and their potentially high emission rates, improved data that will enable these fluxes to be constrained more accurately should significantly improve our understanding of the impact of dairy facilities on regional air quality.

Analysis of VFAs is complicated by their affinity for adsorption to surfaces. Because of this, VFA losses during sampling and analysis may lead to systematically low measurements and signal carry-over from run-to-run. Quantification is also difficult because accurate calibration standards cannot easily be generated.

The purpose of this research was to develop an analytical method to measure emissions of VFAs from dairy sources. In this work, the analytical method described by Orzechowska and Paulson (2005) was adapted to quantify C2–C6 VFA emissions from non-enteric dairy sources. Samples were collected using a flux chamber coupled to solid phase micro-extraction (SPME) fibers and analyzed by gas chromatography with mass spectrometry. Gas-phase VFA standards were generated by evaporation of the pure liquids within the flux chamber itself. An advantage of this approach is that the sampling apparatus used in calibrations and field sample collection is identical, and so errors resulting from wall losses cancel out. The efficacy of this method was then demonstrated by quantifying non-enteric VFA emissions from a small dairy facility located on the Fresno State campus.