Written for Presentation at the Air & Waste Management Association Symposium
Minneapolis, MN
June 21-24, 2005
Improved, GC-Olfactometry Based, Malodor Assessment of Swine
CAFOs Utilizing Novel Air Sampling Technologies
DONALD W. WRIGHT, David K. Eaton, Lawrence T. Nielsen, Fred W. Kuhrt;
Microanalytics (a MOCON Company), Round Rock TX; Jacek A. Koziel PhD,
(Agriculture & Biosystems Engineering Dept, Iowa State University), Ames, Iowa;
David B. Parker PhD (West Texas A&M University), Canyon, Texas
ABSTRACT
GC-Olfactometry based odor profile work to date indicates that with increasing
distance from high-density swine feeder operations the residual odor is increasingly
defined by a limited number of high priority odorants. These ‘character defining’
odorants appear to be dominated by compounds of relatively low volatility and high
polarity. Recent work suggests that, as previously shown for cattle CAFOs, p-cresol
may carry much of the overall odor impact, especially for the case of increasing
distance from the source. If confirmed, such prioritization questions the
appropriateness of odor sampling protocols which are based on plastic sample bags.
Several studies have shown that p-cresol and other high-impact, semi-volatile
odorants are rapidly lost to the wall of the bag through ‘scalping’ effects. This
project was aimed at exploring alternative whole-air and analytical sampling
strategies which address these limitations of plastic sampling bags for CAFO odor
assessment. This manuscript reports on preliminary design and feasibility study
results for two such alternative approaches. The first is an alternative whole-air
sampling strategy based on at-source volatiles collection utilizing sorbent tubes,
followed by at-instrument thermal desorption and air sample reconstitution. The
second is the concept of a time-indexed, moving film volatiles collecting cassette;
enabling a continuous ‘moving picture’ of the volatiles profile to be collected for any
specific sample point over an extended period of time. The results to date suggest
that the first technique should enable improved precision for Dynamic Dilution
Olfactometry based odor assessment protocols. The second technique offers
potential for off-setting the limitations imposed by point-in-time sampling in
relation to shifting meteorological conditions.
Keywords: malodor analysis, agricultural odor analysis, farm odor, GC-Olfactometry,
GC-O, solid phase microextraction, SPME, multidimensional gas chromatography,
livestock housing
1
INTRODUCTION
A large body of excellent analytical work has been reported during the past three decades
relative to the volatile compounds emitted by high density livestock operations. A variety
of concentrating and analytical techniques (Hutchinson et al., 1982; McGinn et al., 2003;
Mosier et al., 1973; Oehrl et al., 2001) have been utilized in the identification of scores, if
not hundreds, of volatile compounds in these environments. Included among these
volatiles are a large number of compounds which are known to be potent individual
odorants. The challenge relative to the CAFO odor issue is to extract from this large field
of potential odorants, the compounds which actually carry primary responsibility for the
downwind odor complaints relative to these operations.
Over 200 volatile compounds have been identified as potential odor contributors in
agricultural environments. As a result of such complexity much of the odor assessment
work to date has been restricted to qualitative assessment utilizing human detectors in
conjunction with techniques such as Dynamic Dilution Olfactometry. Past and recent
(Wright et al., 2004) GC-Olfactometry work which has been carried out by these authors,
as well as others, suggests that CAFO odor assessment should, in fact, be translatable to
objective, instrument based protocols. This GC-O based work suggests that the key
odorants relative to distance separation from high density CAFOs are dominated by a few
compounds and these are characterized by relatively low volatility, high polarity and high
odor potency (i.e. p-cresol, p-ethyl phenol, isovaleric acid, 2-amino acetophenone, indole
and skatol)(Wright et al., 2004).
The prioritization of individual odorants relative to odor impact can be an extremely
important consideration in the development of odor assessment sampling and analysis
protocols. It is impossible to overstate the importance of sampling quality to the overall
validity of any analytical procedure. There is absolute truth to the adage which says that
‘any analysis is only as good as the sample to which it is applied’. This consideration is
especially pertinent to the question of environmental odor assessment in general and
CAFO odor assessment in particular. For example, much of the odor monitoring work to
date has been carried out utilizing sampling protocols which are based upon Tedlar™ or
other plastic gas sampling bags. Unfortunately, the propensity for plastic films to rapidly
adsorb semi-volatile compounds from contained gas samples has been well documented
(Keener et al., 2002; Koziel et al., 2004). This ‘scalping’ effect coupled with past and
recent GC-O based odorant prioritization results combine to bring into question the
validity of plastic bags for odor sampling relative to these environments. That is, if the
odorant prioritization results presented herein are even close to correct, a sampling
protocol which accepts 70% to 100% (Keener et al., 2002; Koziel et al., 2004) loss of top
priority odorants within the first few minutes or hours after collection should be viewed
with skepticism.
It appears that three major challenges confront on-going efforts to develop objective and
quantitative instrument based odor assessment protocols for CAFO environments. The
first of these is to confirm or disprove the validity of the concept of odorant prioritization
for these environments. If the concept of odorant prioritization is proven to be valid, the
2
second challenge is to correct, refine and expand these initial prioritization rankings. The
third challenge is the development of sampling and analytical protocols which more
closely reflect the ‘consensus’ prioritizations which will emerge from having successfully
addressed the first two of these challenges.
Focusing on the first of the three challenges, a recent project was undertaken to perform a
GC-O based odor profile study for high-density swine CAFOs such as was previously
reported for for high density cattle feedlots (Wright et al., 2004; Wright et al., 2005). The
reported odorant priority rankings for the target swine facility were found to be very
similar with respect to the top two or three priority odorants previously reported for cattle
CAFOs. As previously reported, the top priority was dominated by p-cresol, a potent
semi-volatile odorant which is particularly sensitive to surface adsorption effects. Efforts
are on-going to address the challenges associated with validating and refining the initial
odorant prioritizations emerging from these GC-O based odor profile efforts. In addition,
preliminary efforts are also underway to address the third on the list of challenges
presented above; the development of alternative sampling strategies which, more
accurately, reflect the importance of the high impact semi-volatile odorants. Initial
conceptual and feasibility results relative to two such alternative sampling strategies are
summarized in the paragraphs which follow. The first is an alternative whole-air
sampling approach based on at-site volatiles collection utilizing sorbent tubes followed
by at-instrument thermal desorption and air sample reconstitution. The second is the
concept of a time-indexed, moving film volatiles collecting cassette; enabling a
continuous ‘moving picture’ of the critical volatiles profile to be collected for any
specific sample point over an extended period of time. Presented in the paragraphs which
follow are the authors’ progress, to date, with regard to addressing the three challenges
defined above.
MATERIALS and METHODS
Multidimensional Gas Chromatography-Olfactometry-Mass Sprectrometry
MDGC-O-MS is an integrated approach combining olfactometry and multdimensional
GC separation techniques with conventional GCMS instrumentation. A commercial
integrated AromaTrax™ system from Microanalytics (a MOCON Company) of Round
Rock, Texas was used for the GC-olfactometry profiling work. Details regarding
hardware and operational parameters have been described in detail in past publications
(Wright et al., 2004) and will not be restated here.
Sampling:
Solid Phase Microextraction (i.e. SPME) (Chai and Pawliszyn, 1995; Chai and Tang,
1998) utilizing a 1 cm Carboxen modified PDMS - 75 µm fiber was the headspace
sampling technique which was utilized for this odor profiling study. SPME collections
were carried out by direct fiber exposure of the target swinebarn environment – utilizing
variations in downwind distance for cross-comparison purposes. All SPME collections
were carried out under ambient conditions. In addition, the same SPME fibers were
utilized to carry out the instrumental portion of the comparative odorant recovery studies
with respect to Tedlar bag and thermal reconstitution strategies.
3
Animal Feeding Facilities:
Air samples were collected at a commercial swine finish site in central Iowa in
November, 2004. The site consisted of 4 mechanically-ventilated barns with a deep-pit
manure management system. Each barn was 12.5 m wide by 58.5 m long and carried a
960 head capacity. Air samples were collected for 20 minutes at the exhaust fan and 3
locations downwind from the site. The most distant samples collected were collected
approximately 200 m downwind. Samples were collected with a 5 min offset since @ 5
minutes was required to deploy a SPME sampler and then move to a new location.
Air samples were also collected at and downwind of Bracken cave near San, Antonio,
Texas in September, 2004. Samples were collected near and at-distance relative to a
ventilation shaft which has been bored into the cavern. Air samples were collected for 30
minutes at 2 locations downwind from the ventilation shaft opening. The closest was 15
m while the most distant was approximately 125 m downwind. Samples were collected
with a 10 min offset since @ 10 minutes was required to deploy a SPME sampler and
then move to a new location.
RESULTS and DISCUSSION
The odor issues relative to CAFOs will be very different; depending on the distance of
downwind separation from the source facility (Wright et al., 2004). Results to date,
suggests that p-cresol may represent the preponderence of the odor problem relative to atdistance separation from cattle CAFO sources. As expected, locations at or near these
source facilities are characterized by greater odor complexity; with a greater number and
variety of individual odorants rising above their individulal odor thresholds. The natural
dilution effect associated with increasing distance from these sources has the effect of
simplifying the resulting odor profiles; reducing both the number of individual odorants
detected and the relative intensities of those odorants that are detected. This natural
dilution effect relative to one representative swine CAFO is demonstrated in the
following series of aromagrams (i.e. an odorant profile generated by GC-O utilizing an
experienced, human, odor ‘detector’). Figures 1 and 2 reflect the odor profiles which
were generated near-source and at-distance relative to the targeted swine barn facility.
4
20 Meter
p-cresol
isovaleric acid
p-ethyl phenol
butyric acid
Figure 1. Aromagram for 20 min SPME fiber collection at 20 m downwind (“near” site)
from research swine CAFO.
p-cresol
200 Meter
2-aminoacetophenone
isovaleric acid
diacetyl
Figure 2. Aromagram for 20 min SPME fiber collection at 200 m downwind (“at-distant”
site) from research swine CAFO.
The reduced sample loading resulting from sampling at increasing distance from the
source is clearly reflected in this aromagram series. Key observations which can be
extracted from these comparative profiles are the following:
 Increasing distance from the source results in a significant reduction in the total
number of detectable odors as well as corresponding reductions in odor impact
intensities for those odors that are detectable.

Para-cresol represented the dominant odorant relative to both near-source and atdistance downwind sampling points. This dominance reflects responses of the GC-O
investigator to both perceived odorant intensity as well as perceived odor character.
This priority ranking of p-cresol relative to at-distance separation from the swine
5
CAFO source is in agreement with earlier profiles developed by the same investigator
for cattle CAFOs.

Relative to the near-site collection, only the dimethyl trisulfide homolog of the sulfide
series presented with a significant individual odor response (i.e. distinct 'fecal'). There
were no significant odor responses for H2S or the lower MW organic homologs;
under the selected sampling conditions.

The profile of odorants which were secondary to p-cresol in odor impact ranking
were found to be in good agreement with those previously shown for cattle CAFOs.
These included; isovaleric acid, 2-aminoacetophenone, p-ethyl phenol, butyric acid
and diacetyl.

Surprisingly, in contrast to previous swine CAFO odor profile efforts, skatol and
indole were not shown to be significant secondary odorants relative to this current
series. It is assumed, at this point, that this absence results from the unusually short
exposure times (i.e. 20 minutes) used for these SPME fiber collections. Short
exposure time bias relative to decreasing compound volatilities is a well established
characteristic of the SPME sampling technique (Chai and Tang, 1998).
These odor profile results were shown to be consistant with those previously reported by
these authors for cattle CAFOs. Comparative odorant priority rankings relative to these
two environments are summarized in Tables 1 and 2 below.
Table 1. Approximate odor impact priority rankings for a commercial cattle CAFO.
Odor Priority Ranking
Near Source
Distance From Source
1
2
3
trimethylamine
para-cresol
butyric acid
para-cresol
isovaleric acid
para-ethyl phenol
Table 2. Approximate odor impact priority rankings for a research swine CAFO.
Odor Priority Ranking
Near Source
Distance From Source
1
2
3
para-cresol
isovaleric acid
2-aminoacetophenone
para-cresol
isovaleric acid
guiacol
Although considerable similarity is shown in these comparative odor profiles, there were
also points of significant difference. Particularly noteworthy was an apparent reduction in
the odor impact significance for trimethylamine for the swine CAFO in comparison to the
previous cattle CAFO results. As stated previously, this apparent difference may be
accounted for by the unusually short sample collection time (i.e. 20 minutes) relative to
that of the previous cattle CAFO series (i.e. 1 hour and 4 hour). Perhaps, of greater
importance, is the similarity in the top priority odor impact rankings when these
comparisons are made.
6
Interesting parallels to these odorant ranking results for swine and cattle CAFOs can be
drawn from an analogous odorant profile study which was carried out recently relative to
an unusual high density animal community from nature; the Mexican free-tail bat colony
which inhabits Bracken Cave near San Antonio, Texas. With population estimates
ranging to ~26 million at the peak of the annual season this colony is reportedly the
highest population density of mammals in the world (BCI). In addition, with estimates of
the continuous age of the colony ranging to 900 years (BCI) the colony would also
appear to represent one of the oldest high-density animal sites in the world. Obviously,
with respect to downwind odor impact, the issues with respect to this colony are
analogous to those encountered for cattle and swine CAFOs. Important considerations
relative to these comparisons are as follows:
 The colony is populated by mammals feeding, almost exclusively, on a
high-protein diet.
 An overall volatiles / odorant emission profile which is remarkably similar
to cattle and swine CAFOs; including ammonia, hydrogen sulfide and its
organic homologues, VFAs, and semi-volatiles.
 The downwind odor of the colony carries for great distances, is intense
near the source, is unique in odor character and remarkably dissimilar to
either cattle or swine CAFOs.
 A GC-O based odorant ranking priority which is significantly different
from swine and cattle CAFOs and characterized by differences which are
shown to correlate well with the composite downwind odor of the colony.
To illustrate the last consideration, Table 3 summarizes the equivalent odorant priority
results for a near-source and at-distance collection series which was carried out under
time and distance conditions which were similar to those adopted for the previous swine
series.
Table 3. Approximate odor impact priority rankings for Mexican free-tail bat colony.
Odor Priority Ranking
Near Source
Distance From Source
1
2
3
para-ethyl phenol
2-aminoacetophenone
para-cresol
2-aminoacetophenone
para-ethyl phenol
para-cresol
Whereas p-cresol has consistently been shown to represent the top priority relative to
cattle and swine CAFOs it is shown to occupy a secondary ranking priority relative to 2aminoacetophenone and para-ethyl phenol with respect to the bat colony. In addition,
although 2-aminoacetophenone and para ethyl phenol are consistently identified as
secondary impact odorants relative to both cattle and swine CAFOs they clearly take odor
impact precedence relative to the bat colony. With increasing downwind distance from
the Bracken cave source the unusual and characteristic odor is increasingly dominated by
2-aminoacetophenone; a potent semi-volatile odorant which, coincidentally, is one of the
compounds primarily responsible for the characteristic aroma of taco shells.
7
Clearly, whether cattle, swine or Mexican free-tail bats, there is a repeating theme which
carries across all of the previous odorant priority ranking profiles. That theme is the
relatively high odor impact priority of a few semi-volatile odorants such as p-cresol, 2amino acetophenone and para-ethyl phenol; especially with respect to the case of
increasing downwind separation from the odor sources.
The priority rankings presented above do not purport to represent the definitive
qualitative definition of CAFO odor. Although all of the priority ranking studies
presented herein have been carried out by a highly experienced GC-O investigator, each
of these individual profiles obviously represents the subjective assessment of that
investigator. Correcting, expanding, refining and ultimately validating these initial
priority rankings will be achieved through analogous collaborative efforts incorporating
different investigators, different sample introduction techniques, different facilities and
different experimental parameters. The requirements for arriving at a general agreement
relative to GC-O based odorant priority rankings are, in fact, analogous to those required
for evaluating sensory panels and sensory panel based odor assessments. However,
success in achieving consensus or general agreement relative to odorant prioritization
will ultimately enable the current subjective sensory based assessment protocols to be
augmented or largely replaced by objective instrument based alternatives. In spite of the
subjective nature of these initial odorant prioritization rankings, these assessments are
believed to be sufficiently compelling and consistent to warrant a more comprehensive
GC-O based study. On-going efforts directed at refining these odorant ranking profiles
are being carried out in the laboratories of these authors and others and will be reported in
future meetings.
Regardless of the absolute order which emerges for any such consensus odorant priority
ranking profile there is compelling evidence that sampling strategies reflected in current
CAFO odor assessment protocols should be reassessed. This reassessment should include
a critical review of associated sampling protocols and whether they need to more closely
reflect the constraints of the several odor potent and adsorption prone semi-volatiles.
Toward this end, initial development work has been undertaken relative to two alternative
sampling strategies which address two key issues of CAFO downwind air sampling. The
first strategy addresses the need for an alternative to plastic bag based whole air samplers
in support of laboratory based sensory assessment protocols (i.e. Dynamic Dilution
Olfactometry). The second is a novel time-indexed moving film strategy which addresses
the need for a stabile, continuous, time-indexed volatiles profile record for a specific
sample point with respect to an extended time frame. Whereas the former is designed to
support whole air sensory based assessment protocols the latter is designed to support
instrument based analytical protocols by providing an archivible, continuous ‘moving
picture’ of the volatiles profiles for a specific location. These proposed strategies and
associated development results to date are described in the paragraphs that follow.
8
Alternative whole-air sampling strategy - At-site sorbent tube collection - at-instrument
thermal desorption and air sample reconstitution (Conceptual Drawing).
Olfactometry Sniff Port
Flow Control
Flow Control
Sorbent Tube
Thermal Desorption Block
Diluent Air Supply
Heat Tracing
Inertized Canister
Piston
Inert Gas Supply
Servo Drive Mechanism
The concept behind this sampling strategy is that the actual field air sampling is carried
out utilizing sorbent tubes for collecting the volatiles or odorants from a fixed volume of
air. These sorbent tube collections are then transported to the laboratory for reconstitution
within a heat traced, inertized canister followed by composite odor assessment. The
sample reconstitution process is accomplished by thermally desorbing the collected
odorants into a flowing nitrogen gas stream and making up to a desired final volume with
humidified air to match that of the originally sampled environment. Key design
considerations relative to this proposed strategy are as follows:




Sorbent tube based storage of volatiles has demonstrated improved recovery for
critical semi-volatile odorants relative to plastic bags (Koziel et al, Keener et al).
Passivated stainless steel construction of reconstitution chamber and sample
transfer lines replaces plastic sampling bags and tubing.
Heat tracing of reconstitution chamber and all downstream sample transfer lines
to minimize potential adsorption effects.
Servo feed-back controlled piston delivery of reconstituted sample eliminates the
need for in-line flow control devices, valves or other potential ‘scalping’ sites.
9
Table 4. Comparative Odorant Recovery – Five Day Storage
Odorant
Tedlar Bag
Thermal Reconstitution
para-cresol
2-aminoacetophenone
para-ethyl phenol
To be determined
To be determined
To be determined
To be determined
To be determined
To be determined
Alternative analytical sampling strategy – Time indexed – moving film volatiles
collector cassette (Conceptual Drawing)
Unlike, the sorbent tube based ‘whole air’ sampler approach described above, the second
proposed device is designed to address a different set of challenges presented by CAFO
downwind odor assessment. Whereas the former is designed to deliver a stabile,
transportable, point-in-time whole representation of the originally sampled environment,
the latter is designed to deliver a stabile, continuous, time-indexed volatiles profile record
for a specific sample point with respect to an extended time frame. Whereas the former
primarily targets the need for a robust remote site sampling strategy in conjunction with
whole air sensory assessment (i.e. Dynamic Dilution Olfactometry) the latter targets the
same requirements relative to analytical assessment protocols (i.e. MDGC-MSOlfactometry etc.) and archiving.
10
Ironically, the principles behind this strategy are the same as those forming the basis for
concerns relative to the sampling of semi-volatile odorants with plastic bags; their
tendency to rapidly adsorb critical semi-volatile odorants from the air sample.
Capitalizing on this tendency it becomes possible to utilize a continuous, slowly moving
or ‘scrolling’ plastic film to capture the volatile compounds within a target environment
during a specific time period. Specifically, a specially formulated and preconditioned
plastic film (i.e. or combination of different films) is fed across an environmental
exposure window according to precisely controlled and time indexed format. With an
appropriate time indexed driver mechanism it becomes possible to carry out a long-term
sample collection profile study before any requirement for associated analytical work-up.
In addition, with an appropriate post-exposure sealing process such as the foil
encasement mechanism shown in the drawing above, it becomes possible to collect and
store the sample without the need for immediate analytical work-up. This opens up the
possibility of a method of profiling, documenting and archiving intermittent odor
excursions at target sites downwind from odor sources. Initial design and feasibility study
efforts are on-going relative to this proposed device and the results of these efforts will be
presented in conjunction with future conferences.
CONCLUSIONS
Based upon past and current GC-O odor profile efforts, p-cresol and other semi-volatiles
appear to be critical, ‘character defining’ odorants relative to downwind, distance
separation from both high density cattle and swine CAFOs. Parallel odorant prioritization
efforts relative to the Bracken cave Mexican free tail bat colony demonstrates that GCOlfactometry can be an effective technique for the development of correlations between
downwind odor impact and high priority individual odorants. If the preliminary CAFO
odorant priority rankings are confirmed across a broader representation of investigators,
sampling techniques and CAFO sites there will be increasing impetus for critical review
of current sampling, analytical and odor abatement strategies relative to these
environments. This level of cross-check should yield a preliminary consensus as to the
identity of the top two to four most prominent individual odorants relative to downwind
community odor complaints. Such prioritization efforts appear to be particularly
important with respect to p-cresol and other high impact semi-volatile odorants due to
their well documented sensitivity to adsorption driven loss to the walls of plastic sample
containers. Success in identifying such a minimal critical odorant set from CAFOs
simplifies the challenge of translating current, subjective, human ‘detector’ based odor
assessment protocols to objective, instrument based alternatives. Based upon the results
to date, work has begun to develop alternative sampling strategies which address
concerns presented relative to the currently applied protocols. The status of design and
feasibility efforts relative to two such alternative sampling strategies are summarized
herein. The authors are confident that these efforts can ultimately enable the translation of
current, subjective, sensory-only odor assessment protocols to objective, instrument‘primarily’ alternatives.
11
ACKNOWLEDGEMENTS
The authors would like to express appreciation to Dr. Steven Hoff and the management
of the commercial swine production facility for the invitation to participate in a larger air
quality study and access to the swine finishing site. The authors would also like to
express appreciation to Andy Moore and the management of Bat Conservation
International for logistical assistance and access to Bracken cave.
REFERENCES
Chai, M. & J. Pawliszyn. 1995: Analysis of environmental air samples by solid phase microextraction
and gas chromatography / ion trap mass spectrometry. Environmental Science and Technology. Vol 29
(3), (1995).
Chai, M. & Y. Tang. 1998: Solid phase microextraction (SPME) analysis of whole air samples. Journal
of Environmental and Analytical Chemistry. 72(1):77-82.
Hutchinson, G.L.; A.R. Mosier, and C.E. Andre. 1982: Ammonia and amine emissions from a large
cattle feedlot. Journal of Environmental Quality. Vol 11 (2).
Keener, K.M., J. Zhang, R.W. Bottcher, and R.D. Munilla. 2002. Evaluation of thermal desorption for
the measurement of artificial swine odorants in the vapor phase. Transactions of the ASAE.
45(5):1579-1584.
Koziel, J.A.; J.P. Spinhirne, J.D. Lloyd, D.B. Parker, D.W. Wright, and F.W. Kuhrt. 2005. Evaluation
of sample recovery of malodorous gases from air sampling bags, SPME, air sampling canisters and
sorbent tubes. JAMWA, accepted.
McGinn, S.M., H.H. Jantzen, and T. Coates. 2003. Atmospheric ammonia, volatile fatty acids and
other odorants near beef feedlots. Journal of Environmental Quality. 32:1173-1182.
Mosier, A.R., C.E. Andre, and F.G. Viets, Jr. 1973. Identification of aliphatic amines volatilized from
cattle feedyard. Environmental Science and Technology. 7(7).
Oehrl, L.L, R.W. Keener, R.D. Bottcher, R.D. Munilla, and K.M. Connely. 2001. Characterization of
odor components from swine housing dust using gas chromatography. Applied Engineering in
Agriculture. 17(5):659-661.
Wright, D.W., D.K. Eaton, L.T. Nielsen, F.W. Kuhrt, J.A. Koziel, J.P. Spinhirne, and D.B. Parker.
2004. Multidimensional gas chromatography-olfactometry for identification and prioritization of
malodors from confined animal feeding operations. Proceedings of the ASAE/CSAE International
Conference, Ottawa, August 2004.
Wright, D.W., D.K. Eaton, L.T. Nielsen, F.W. Kuhrt, J.A. Koziel, J.P. and D.B. Parker. 2005.
Improved, GC-Olfactometry based, malodor assessment of swine CAFOs utilizing novel air sampling
technologies. In Proc. of the Symposium of the State of the Science on Animal Manure and Waste
Management, San Antonio, TX, January 5-7, 2005.
12
13