United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-79-004
March 1979
Air
Testing of Hydrocarbon
Emissions from
Vegetation, Leaf Litter
and Aquatic Surfaces,
and Development of a
Methodology for
Compiling Biogenic
Emission Inventories
Final Report
-------
EPA-450/4-79-004
Testing of Hydrocarbon Emissions
from Vegetation, Leaf Litter
and Aquatic Surfaces, and
Development of a Methodology
for Compiling Biogenic Emission
Inventories
Final Report
by
Patrick R. Zimmerman
Washington State University
Pullman, Washington 99164
EPA Project Officer: Thomas F. Lahre
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
March 1979
-------
This report is issued by the U. S. Environmental Protection Agency
to report technical data of interest to a limited number of readers.
Copies are available free of charge to Federal employees, current
contractors and grantees, and nonprofit organizations in limited
quantities from the Library Services Office (MD-35), Research
Triangle Park, North Carolina 27711; or, for a fee, from the
National Technical Information Service, 5285 Port Royal Road,
Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Washington State University, Pullman, Washington 99164, in
fulfillment of a contract. The contents of this report
are reproduced herein as received from Washington State University.
The opinions, findings and conclusions expressed are those of the
author and not necessarily those of the Environmental Protection
Agency. Mention of company or product names is not to be considered
an endorsement by the Environmental Protection Agency.
Publication No. EPA-450/4-79-004
11
-------
Acknowledgements
This project was initiated by Dr. R. A. Rasmussen, now at the Oregon
Graduate Center.
Don Stearns played a key role in the collection of many of the samples.
Revies comments were supplied by Robert R. Arnts through Dr. J. J. Bufalini,
and Harold G. Richter, US EPA, Research Triangle Park, NC; Fred Mowrey,
Research Associate in Forest Meteorology at Duke University; Dr. Dave
Tingey and L. C. Grothaus, US EPA, Corvallis, OR; and Wally Jones, US EPA,
Region IV, Atlanta, GA.
Tom Lahre, US EPA, Research Triangle Park, NC, was especially helpful
in incorporating review comments in the final report.
ill
-------
CONTENTS
INTRODUCTION 1
1 FIELD TESTING FOR EMISSION FACTORS 3
1.1 SAMPLING EQUIPMENT AND PROCEDURES 4
1.1.1 Sampling Equipment 7
1.1.2 Sampling Procedure 14
1.1.2.1 Bag Blank 18
1.1.2.2 Ambient Air Sample 19
1.2 ANALYTICAL INSTRUMENTATION AND PROCEDURE 19
1.2.1 Calibration 20
1.2.2 Sample Enrichment 26
1.2.3 Sample Introduction 27
1.2.4 Analysis 29
1.2.5 Emission Rate Quantitation 29
1.2.6 Biomass Quantitation 29
1.3 SPECIAL FIELD TESTING PRECAUTIONS 31
1.4 EMISSION RATE DETERMINATION 32
2 EMISSION INVENTORY DEVELOPMENT 35
2.1 GENERAL 35
2.2 LEAF BIOMASS FACTORS 36
2.3 EMISSION RATE ALGORITHMS 39
2.4 NATURAL EMISSION DATA 46
2.5 EXAMPLE NATION-WIDE EMISSION INVENTORY 55
2.5.1 Purpose 55
2.5.2 Seasonal Variability. 56
2.5.3 Elements of Annual Emission Inventory 60
2.5.4 Limitations of Emission Estimates 62
2.5.5 Summary of Inventory Procedure 65
-------
2.6 COMPARISON WITH OTHER BIOGENIC EMISSION ESTIMATES 66
REFERENCES 70
APPENDIX A - Experimental Verification of Sample Methodology A- 1
A.I INTRODUCTION A- 1
A.2 PROJECT OBJECTIVES A- 2
A.3 SAMPLING APPROACHES A- 4
A.3.1 UPWIND/DOWNWIND SAMPLING A- 4
A.3.2 ENCLOSURE SAMPLING A- 5
A.3.2.1 Regulated Enclosures A- 6
A.3.2.2 Compensated Chambers A- 7
A.3.2.3 Static Chambers A- 8
A.3.2.4 WSU Static/Dynamic Enclosure A- 8
A.4 ZERO AIR . - A-ll
A.4.1 COMMERCIAL ZERO AIR A-12
A.4.2 MOLECULAR SIEVE FILTERS A-12
A.4.3 HYDROCARBON COMBUSTER A-12
A.4.4 MELOY PURE AIR SOURCE A-12
A.4.5 AADCO PURE AIR GENERATOR A-12
A.4.6 CRYOGENIC COMPRESSION OF ZERO AIR A-13
A.5 SAMPLE METHODOLOGY CHECKS "..... A-15
A.5.1 RELATIVE HUMIDITY A-15
A.5.2 COo A-15
A.5.3 SAMPLE INTEGRITY A-16
A.6 ANALYTICAL METHODOLOGY CHECKS A-17
A.6.1 SAMPLE ENRICHMENT A-17
A.6.2 HYDROCARBON ANALYSIS A-17
A.6.2.1 Oxygenates A-18
A.6.2.2 Analytical Precision A-21
A.6.2.3 Analytical Problems A-22
A.7 EXPERIMENTAL DEAD-ENDS A-23
A.7.1 SOIL LEAF-LITTER SAMPLES A-23
A.7.2 VEGETATION SAMPLES A-24
A.8 SUMMARY A-26
APPENDIX B - Detailed Derivation of Emission Rate Formulas B- 1
B.I INTRODUCTION B- 1
B.I.I DEFINITION OF TERMS B- 3
B.I.2 EMISSION RATE FORMULAS B- 5
VI
-------
FIGURES
1.1-a Vegetation emission sample collection system 5
1.1-b Soil leaf-litter sampling system 6
1.1-c Surface water sampling system 8
1.1-d Portable sample manifold 11
1.1-e Field data format 16
1.2-a Typical analysis of a vegetation sample (Juniper) 22
1.2-b Sample chromatogram, glass capillary column (#5) 23
1.2-c Vacuum system for sample injection 28
2.2-a Major biotic regions of the U.S 40
2.3-a Emission rate algorithms 43
A.6.2-a Stainless-steel sampling can A-25
B.I-2 Laboratory data format B- 2
vif
-------
TABLES
1.2-a VOC ANALYSIS CONDITIONS 21
1.2-b COMPARISON OF EQUIVALENT UNITS OF VOC QUANTITATION 25
2.2-a AREAS OF THE MAJOR BIOTIC REGIONS OF THE CONTINENTAL U.S. . . 41
2.4-a EMISSION PROFILES FOR SELECTED FLORIDA VEGETATION 47
2.4-b EMISSION PROFILES FOR SELECTED SPECIES 48
2.4-c NON-METHANE VOC EMISSION RATES OF BROAD VEGETATION CATEGORIES
FOR VARIOUS SAMPLE SITES 50
2.4-d LOCATION AND SPECIES OF SAMPLES USED IN U.S. INVENTORY .... 52
2.4-e ESTIMATED EMISSION FACTORS FOR BROAD VEGETATION
CLASSIFICATIONS STANDARDIZED TO 30°C 54
2.5-a BIOME EMISSION FACTORS (Standardized to 30°C) 57
2.5-b AVERAGE MONTHLY U.S. TEMPERATURES BY REGION (°F) 61
2.5-c MONTHLY U.S. EMISSION INVENTORY (yg/mo) 63
2.5-d ANNUAL U.S. EMISSION INVENTORY BY LATITUDINAL REGION (yg). . . 64
2.6-a ESTIMATES OF WORLDWIDE EMISSIONS OF NATURAL VOC 69
A.6.2-a LONG-TERM STORAGE OF ALDEHYDES AND KETONES IN STAINLESS
STEEL CANISTERS A-20
Vlll
-------
INTRODUCTION
The regional nature of photochemical oxidant pollutant episodes has
been well documented in the last few years. High oxidant (especially
ozone) levels have been measured in rural areas well away from significant
anthropogenic emission sources. While evidence has accumulated which in-
dicates that oxidant precursors generated in urban centers can be trans-
ported into these rural regions, it has also been shown in smog chamber
tests that biogenic volatile organic compounds (VOC) are photochemically
reactive and can participate in ozone formation.
In order to define the contribution of vegetation, leaf litter, and
water surfaces to the overall atmospheric burden of VOC in a specific re-
gion, an estimate of biogenic VOC emissions is essential. To this end,
Washington State University, under contract to EPA, has developed testing
procedures to measure emissions from vegetation, leaf litter and surface
waters. As part of this effort, measurements have been made on selected
plant species in North Carolina, California and Washington. Some selected
results of a research project conducted by WSU in Florida funded by EPA
Region IV (Contract #68-01-4432) have also been included. The results
of all of these studies have been used to construct an "order of magnitude"
nationwide inventory of biogenic organic emissions.
This report is designed to serve several purposes. First, it des-
cribes in detail, the equipment and methodology required to measure the
-------
organic compound (VOC) emissions from vegetation, leaf litter and surface
water. Second, it presents emission factors for vegetative species, leaf
litter and the surfaces of bays, rivers and marshes that were tested during
the course of this program. Third, an example nationwide annual inventory
of biogenic organic emissions has been prepared.
The reader has several options available to him if he desires to
compile an inventory for a particular geographical area and a particular
time period. If sufficient resources are available, field measurements
should be performed using the techniques described herein, and an inven-
tory compiled by applying the resulting emission factors to the actual veg-
etation mix in the area of concern. In this case, the actual vegetative
species distribution may be known in sufficient detail from existing in-
formation or it may have to be compiled in a separate effort. On the
other hand, if few resources are available extrapolations can be made
from the inventory data given herein as a coarse approximation of local
organic emissions.
A detailed biogenic emission inventory for the Tampa/St. Petersburg
area was compiled for Region IV EPA concurrent with this study. Summary
emission rate data from the Tampa/St. Petersburg study area are included
in this report. The detailed biogenic emission inventory for the Tampa/
St. Petersburg area is included in the final report for EPA Contract No.
68-01-4432.
For those concerned with the development and validity of the metho-
dology, documentation is included in an appendix explaining the evalua-
tion and rationale of the experimental approaches considered for conduct-
ing emission tests.
-------
1. FIELD TESTING FOR EMISSION FACTORS
There are two basic components needed for the compilation of an
inventory of organic emissions from vegetation, leaf litter and water
surfaces: 1) emission factors which relate organic emissions to some
biological indicator (such as dry leaf biomass or surface area) over time,
and 2) biomass density factors, which are measures of how much of the
particular biological indicator is present in a given area. The term
"biomass" as used in this report, refers to the oven dried weight of the
vegetation of interest. Emission rates for vegetation were calculated
in terms of leaf biomass. Emission rates for the surfaces of water and
soil/leaf litter were calculated directly on the basis of the area of
the surface that was sampled.
Five major steps are necessary to develop a detailed area-wide inventory.
1. Identify the major vegetation types and predominant plant
species.
2. Select the representative species to be sampled.
3. Conduct a field program to collect and analyze emission
samples from each of the representative species.
4. Quantify the biomass density of the major species in the
inventory area.
5. Develop emission rate algorithims for the major daily and
seasonal emission variables (light, temperature, moisture),
that affect biogenic emissions.
3
-------
This section deals with the third activity - the actual collection
and analysis of field data for emission factor development. This section
is broken into two major segments: 1) sampling equipment and procedures,
and 2) analytical equipment and procedures.
1.1. SAMPLING EQUIPMENT AND PROCEDURES
The method developed to measure VOC emissions from vegetation combines
the advantages of a static enclosure system and dynamic flow system. The
system developed by WSU to sample vegetation emissions is shown in Figure
1.1-a. Basically, the method involves (1) enclosing a portion of the
vegetation, (2) collecting a background sample of the enclosed air, (3)
filling the enclosure bag with pure hydrocarbon free air, and (4) collecting
a sample of the air in the enclosure. The samples are returned to the
laboratory, the branch is clipped, and the leaves or needles within the
enclosure are removed and dried. VOC emissions are equal to the difference
between the mass of VOC in the bag at the time of the background sample
and the mass of VOC in the bag at the time of the-collection of the emission
rate sample. The resulting emission rates are expressed in terms of the
micrograms of VOC released per gram of vegetation per unit time.
To sample leaf-litter soil emissions, a stainless steel sealing ring
and bag collar are used (Figure 1.1-b.). For collection of a sample the
sealing ring is driven into the soil; the bag collar is then placed in
the center of the sealing ring. Next, moist dirt is used as a filler
between the sealing ring and the bag collar. After the collar and ring
are in place, a Teflon bag vegetation enclosure is attached to the bag
collar. The sample collection procedure is then identical to that for
vegetation. For this type of sampling, vegetation can be clipped near
-------
FIGURE 1.1-a
VEGETATION EMISSION SAMPLE COLLECTION SYSTEM
•r !';-Jfe
<^i»
TEFLON BAG
STAINLESS STEEL
CANNISTER
PORTABLE SAMPLE MANIFOLD
3 WAY VALVE
THERMOMETER
METAL
BELLOWS PUMP
-------
FIGURE 1.1-
SOIL LEAF-LITTER SAMPLING SYSTEM
EVAC.
-COLLAPSIBLE
TEFLON BAG
- BAG COLLAR
- SAMPLE
ZERO AIR INLET
MOIST SOIL SEAL
SOIL
SEALING RING
(2) H"SWAGLOCK BULKHEAD
72'-
k }>
SHARP CUTTING
EDGE
SEALING RING
BAG COLLAR
* all dimensions in centimeters
-------
the ground and leaf litter collected, dried and weighed for emission
rate calculations. Usually, however, it is sufficient to assume that the
bag collar encloses an "average" amount of biomass material. In this
case the emission rate can be directly related to the soil surface area
enclosed to result in emission rate in terms of yg/nr/minute.
Surface waters are sampled by attaching a flotation ring (consisting
of two water-ski belts sewn together) around the bag collar. The bag
collar, with flotation ring and Teflon enclosure bag attached, is then
placed upon the surface of the water (Figure 1.1-c.). The residual air
inside the bag is then quickly removed (thus collapsing the bag). Zero
air is added to the enclosure in the same manner as for vegetation or
soil/leaf litter sampling. -No background sample is necessary, for the
dead volume can be reduced to zero for these samples.
The success of these methods is based upon the short enclosure time
(fifteen minutes or less) and the large amount of diluent zero (VOC free)
air introduced into the enclosure. Both of these factors mitigate static
chamber difficulties such as high chamber temperature and the long-term
accumulation of metabolic CC^ (from soils or from vegetation in the dark)
and/or water vapor that may affect emission rates. At the same time,
samples are concentrated enough to allow good analytical resolution of
the hydrocarbons present.
1.1.1. Sampling Equipment
To collect emission samples from vegetation, leaf litter and water sur-
faces the following basic equipment is needed:
1. A portable source of pure air which is free of VOC and can be
regulated to give a precise flow rate.
-------
Figure i.i-c. SURFACE WATER SAMPLING SYSTEM
FLOATATION BELT
COLLAPSIBLE TEFLON BAG
TEFLON BAG SUPPORT
BAG COLLAR
WATER
SAMPLE
ZERO AIR INLET
FLOATATION DELT
K-3
2.5
BAG COLLAR
*oll dimensions In centimeters
8
-------
2. A vegetation enclosure which does not add to or take away from
organic vegetation emissions.
3. A method to collect an air sample from the enclosure and to move
it to the laboratory for analysis.
4. A method to measure the amount of the air sample which is
collected .
There are primarily two sources of VOC free "zero" air:
1. commercially available cylinders
2. zero-air generating equipment which purifies ambient air
Commercially available cylinders have the advantages of being easy
to obtain and they require a relatively small investment. Their disadvan-
tages are that they may contain relatively high level of VOC (up to 1 ppm)
and unknown quantities of CO, C02, and NOX. Also, the N2/02 ratio is not
constant from tank to tank.
Equipment is available which can manufacture zero air from ambient
air on a day-to-day basis. The Aadco air generator is an example. The
main advantage of the zero air generator is that it can produce a continuous
supply of air of uniform quality characterized by a constant ^702 ratio.
However, the equipment is not portable, so methods must be provided to
get zero air from the laboratory to the field.
The zero air produced by the Aadco pure air generator has no detectable
hydrocarbons or fluorocarbons. It has a C02 level of about 10 ppm. An
ascarite (sodium hydroxide-asbestos) trap can be added to the output
stream to adsorb C02« It may be desirable to add C02 to reach a
concentration of about 365 ppm representative of ambient air. The major
disadvantages of a zero air generator are the higher initial purchase
-------
price than cylinder air and the lack of portability. WSU's method of
providing a portable source of zero air is to cryogenically collect the
output from the zero air generator into empty high pressure medical-grade
oxygen cylinders. The procedure is outlined in the documentation portion
of this manual. Approximately 100 & of zero air is required per vegetation
sample collected.
Regardless of the source of zero air, the result will usually be
packaged in a high pressure cylinder. The cylinder must then be pressure-
regulated and flow-controlled. The apparatus used in this study is shown
in Figure 1.1-d and consists of a 300 a cylinder of zero air connected to
a pressure regulator which contributes no background VOC to the sample.
The regulator is in turn connected to a pressure gauge and to a three-way
valve which is connected to two needle valves. One is pre-set to give a
flow of 10 Vmin and the other is set to give a flow of 2 &/min at a
specified outlet pressure (40 psig). Thus a flow of 10 fc/min or a flow
of 2 Vmin can be selected by positioning the three-way valve. Possible
alternatives to high pressure cylinders include the use of large, clean,
100 i capacity Teflon bags filled with zero air, plus the appropriate
D.C. pump and flow controllers to attain the flow rates specified in the
sampling procedures. Large low pressure (air compressor type) tanks
could also be used to hold the zero air needed for sampling.
The vegetation enclosure which has proven to be the easiest to work
with is a Teflon bag. The bag should be closed on three sides. The dimen-
sions are not critical; however, enclosure bags should be constructed so
that they have a capacity of at least 120 I when fully inflated. The one
used in this program measured 109 cm x 144 cm. These bags can easily be
10
-------
FIGURE 1.1-d
PORTABLE SAMPLE MANIFOLD
0-AIROUT
PRESSURE
GAUGE
SAMPLE' PRESSURE
REGULATOR
0-AIR IN
OFF-
REGULATOR .
EVAC. —
PUMP
SAMPLE
SAMPLE
FLOW
I01-
0-AIR
FLOW
M
0-AIR OUT
•PRESSURE
GAUGE
"^NEEDLE VALVES
n
-------
fabricated in most laboratories. No special fittings for sample probes
are required.
There are a number of methods for removing an air sample from the
Teflon bag enclosure and pressurizing it into a stainless steel canister.
The simplest, most direct method is to use a metal bellows-type pump equipped
with a DC motor and a Teflon valve assembly. (Note: Teflon must be specified
at the time of purchase or a Viton valve assembly will be installed by
the manufacturer. Under heat, Viton has a high bleed rate of organics.)
An alternative is1to use a cryogenic collection technique in which the
sample container is immersed in liquid N2 to "pump" a sample into the
container. Special, large capacity stainless steel wide-mouthed dewars
are required to hold the liquid N£. The least desirable alternative is
to use a 100 ml ground glass or a large gas-tight syringe to remove a
sample and place it in the sample container. The procedure is slow and
chances for contamination are increased. Whatever method is used, it
must provide a means of pressurizing a sample container or filling a
small Teflon sampling bag without adding to or subtracting from the
sample.
Specially treated SUMMA-passivated stainless steel canisters having
an internal volume of 5.5. I were used as sample cont- ners in this study.
The SUMMA process is an electropolishing procedure which removes the
active sites from stainless steel, thus minimizing the adsorption of
organics from the sample. The "cans" are easy to clean by heating and
purging with a clean gas (such as zero air). WSU has tested the storage
characteristics of the electropolished stainless steel canisters and has
found that the total non-methane hydrocarbon (TNMHC) concentrations are
12
-------
stable in the cans; however, the ratios of some reactive hydrocarbons such
as isoprene may change over a period of days in some of the cans. Emission
samples have been stored for up to a week with no noticeable change in the
hydrocarbon composition. Teflon sample bags could be used as sample containers
if they are kept out of direct sunlight after filling and their contents are
analyzed within a matter of hours after collection.
Another possible method of sample collection is to outfit a freeze-
out loop with "quick disconnect" fittings and directly freeze out the
sample. This method has several disadvantages. The loops must be kept
in liquid N£ from sample collection to analysis. Also, only one analysis
per sample is available.
If analysis of a specific compound was desired, a solid absorbent
such as Tenax could be used for sample collection. However, the method
would also allow only one analysis per sample. Additionally, the
adsorption-desorption efficiency of Tenax vaires from compound to com-
pound, and a quantitative broad spectrum analysis would be difficult.
Miscellaneous other sampling equipment is needed in field testing.
An inexpensive indoor-outdoor thermometer can be used to measure bag temp-
eratures and ambient air temperature simultaneously. This works well
because the outdoor temperature sensor can be placed along the branch
inside the sample chamber, and the thermometers can then be hung in a con-
venient location on a tree limb. The temperature sensors should not be
placed in direct sunlight or erroneous reading will result.
If sample canisters are used, a pressure gauge is required. The
pressure gauge should be equipped with a side-port needle. One valve of
the sample canister should be equipped with a Teflon-backed silicon rubber
13
-------
septum. When a pressure reading is required, the pressure gauge needle
is inserted through the septum and the canister valve cracked open.
Relative humidity and other miscellaneous sensors can be added to
the sampling instrumentation. However, the materials of construction of
each of these should be carefully checked to insure that they do not
affect the sample integrity.
Clean copper tubing is acceptable for use as, sample probes and zero
air probes. All tubing should be thoroughly cleaned before use. The
copper tubing can be slowly purged with an inert gas such as clean
nitrogen at ^10 ml/minute, and flamed using a propane torch. The tubing
should be flamed starting at the purge gas inlet and should gradually
proceed to the outlet end. The purge flow should continue until the
tubing has cooled. The tubing can then be purged overnight with "zero
air". Teflon tubing has also been used in the sample train with good
results. However if the Teflon tubing used is suspected of being
contaminated it should be replaced. Attempts at cleaning contaminated
Teflon tubing are usually unsuccessful. A periodic blank analysis should
be run on all components in the sample train to insure cleanliness.
1.1.2 Sampling Procedure
Before a sampling program is begun, the following steps should be
completed:
1. The area of the emission inventory should be defined.
2. The major vegetative species of interest should be determined.
3. Supportive instrumentation such as the portable zero air appara-
tus, sample pump, sample chambers, sample containers and tubing should
should be flushed with zero-air and checked for contamination.
14
-------
The sites to be selected for vegetation sampling will be dictated by
distribution and location of major vegetative species as well as by practical
sampling considerations. Sites which are located away from high anthropogenic
concentrations are preferred though not mandatory. The site for each
vegetation sample should be selected so that it is representative of the
vegetation type for the area defined. The vegetation specimen should be
examined for obvious signs of disease or injury which might cause excessive
resin deposits or other modified metabolic behavior and thus result in a
non-representative sample. After a site has been selected, the vehicle
carrying the sampling equipment should be parked downwind of the site.
Figure 1.1-e. is the field data format which can be used for the col-
lection of vegetation and sofl-leaf litter samples. This data format has
been developed so that all important sample variables are recorded in a
reproducible manner. This will allow the later determination of trends
in emission rates as correlated with site variables and weather patterns.
It also provides an orderly outline for collection of the sample, and
insures that data critical to the calculation of emission rates is recorded.
The following is an instructional outline for collection of an
emission sample from vegetation:
1. Record the site location, sample number, weather and the specimen
description.
2. Place the sample probe, purge and flush tube, and the bag tempera-
ture sensor along the branch. They should be positioned so that they
will not interfere with the bag placement. They should then be
fastened in place. Strips of "Velcro" have been successfully used
as fastening material for probes, sensors and bag sealing by WSU.
15
-------
Figure 1.1-e. Field data format
148 Background
Date 6-13 Sample # RTF Can # 204 Emission
Location Approximately 5 mi SE Umstead Park, NC Barom
Sample type: White Oak
Enclosure: Teflon Bag (new)
Site description: Mixed hardwood in Piedmont area next to powerline
easement (open canopy)
Weather, general: clear, warm, little wind
Weather, site shade Cloud cover 0% Ha (ambient air temp.) 31°C
Wind: direction NW Speed 0-5 Gust 7
Vegetation: describe type, age, physiological state. About 60 years old.
In fair shape. Some insect damage, 8" DBH about 30' tall.
Litter: Type about 5 cm litter and duff
Incorporation Depth
Soil: Moisture damp ph Temp.
Describe
Time at encl. TI 1249 TX End Bkdg sample 1257 start flush, Tg 1258
End flush, T^ 1304 , Start sample, T5 1304 End sample, T6 1307
Sample rate Vmin. 2 Vmin.
Flush flow rate Zf(*/min.) 10 Purge flow rate ZpU/min.) 2_
Enclosure sample temp. 32°C (est.) Can pressure 15 psig
Comments: 30% of leaves have insect damage
Close to road
Estimated dead volume 201
Dimensions 18" X 2" X 36"
16
-------
3. Turn the Teflon bag enclosure inside-out, open the unsealed edge
and alternately fill and deflate it with ambient air. This will insure
that any car exhaust components accumulated during transit will be
flushed out.
4. Gently slide the Teflon bag over the branch. Take care not to
break twigs or crush foliage.
5. Seal the base of the bag around the branch. A 1.2 cm wide by
20 cm strip of Velcro works as a bag sealer. The Velcro is wrapped
tightly around the base of the bag so that the "hook" side and the fuzzy
sides overlap.
6. Record the time at the beginning of enclosure. Pump the residual
air in the bag out until the bag starts to collapse around the branch.
Then collect a sample of the residual air by pumping it into a sample
container. This is the background sample.
7. Record the time at the end of the background sample collection.
8. Measure the rough dimensions of the partially collapsed bag and
record the geometric shape that it approximates. The residual air
left in the bag is the dead-volume (Ve). The dimensional information
is used to aid in the estimation of the dead volume.
9. Start the 10 Vminute flush, record the time. Continue the
flush for 6 minutes.
10. Start the 2 Vnrinute purge, record the time. Immediately begin
collecting the emission sample, record the time. Be sure that the
sample rate is less than ^2 Vnrinute.
11. Record the pressure of the sample cans. This is unnecessary if
the cans have been evacuated. However if they have zero air in them
them the pressure must be measured so that the dilution volume can
be calculated.
17
-------
12. Record the ambient temperature and the bag temperature.
13. If the same branch is to be sampled again at a later time, tag
the branch at the point where the bag was sealed.
14. If the branch is not to be resampled, clip the branch at the
point at which the bag was sealed. Placed the clipped branch in a
paper bag and take to the laboratory.
For leaf-litter/soil samples the sealing ring is driven into the
ground first. Then the bag collar is placed inside the sealing ring and
the bag is attached. For surface-water samples a flotation ring is
strapped to the bag collar before the bag is attached. For both leaf-
litter/soil samples and surface water, evacuation, sample collection and
zero air lines are connected directly to the bag collar. The temperature
sensor is then placed inside the bag collar from the underside. The
Teflon bag enclosure can then strapped to the bag collar. A heavy piece
of elastic belting wrapped around the bag and bag collar provides a good
seal. Steps 6-12 of the vegetation sampling procedure are then followed
with the possible exemptions of steps 6-8, if after the bag is collapsed
the dead volume is equal to zero. This condition normally occurs for
surface-water samples.
1.1.2.1 Bag Blank—If it is suspected that the Teflon bag or other com-
ponent of the sample train is emitting organics which compromise the
sample, a bag blank can be collected. The procedure used is identical to
that for collecting a vegetation emission sample, except that no branch is
enclosed, and no sample is collected before the purge and flush cycles since
the bag can be completely collapsed and the dead volume is zero (Ve=0).
18
-------
1.1.2.2 Ambient Mr Samples—It may be desirable to collect an ambient
air sample at the emission sample site. The ambient air sample should be
collected at the time of the bag enclosure. The collection should be
integrated over a five to ten minute time period. If a local anthropogenic
source is near, such as a highway, integration will minimize the chances
of a non-representative sample due to a passing car, etc. If a bag blank
and an ambient air sample are collected it is usually not necessary to
collect a sample of the enclosed air before the flush and purge cycles
(background sample). However it is still necessary to estimate the dead
volume (Ve) of the enclosure.
1.2 ANALYTICAL INSTRUMENTATION AND PROCEDURE
Analytical instrumentation required for routine samples depends upon
the purpose of the sample program. If total hydrocarbon measurements or
total non-methane organic carbon measurements are desirable, a series of
Flame lonization Detector equipped gas chromatographs (FID, GC) is required.
The FID total hydrocarbon (THC) instruments do not provide the required
accuracy for emission samples. Most commercial THC analyzers have a
lower threshold near 500 ppb. WSU's experience has shown that most emission
samples range between 200 ppb and 2,000 ppb which is on the lower end of
the sensitivity range of THC analyzers. A better alternative to THC
analyzers is the use of temperature programmed FID GC's equipped with
columns to separate the components of the emission sample. One GC is
required to separate ^ to ^6 hydrocarbons, and another for the C/^ to C-^
hydrocarbons. It may be desirable to have a third GC equipped to measure
methane, ethylene, ethane and acetylene.
19
-------
The columns and operating conditions used in this study are given in
Table 1.2-a. Sample chromatograms are shown in Figures 1.2-a and 1.2-b.
The gas chromatograph and mass spectrometer can be combined in a GC-MS
system to positively identify individual VOC peaks and/or as a comparison
with the GC quantisation. Identical GC columns and conditions should be
used in the GC and GC-MS systems so that direct comparisons can be made
with respect to component elution time.
The analysis and quantisation of oxygenates is more difficult than
that for pure (non-oxygenated) hydrocabons. The experiments have shown
that some oxygenates may be partially lost in the sampling train (see
Appendix A). In addition, although oxygenates cause a response on an
FID-GC, their response factors .are not necessarily the same or as great
as the response factors obtained for typical paraffins, olefins and ter-
penes. Also they do not elute from all column types. If their retention
times are known and the compounds can be identified, they can be calibrated
on an individual basis. Otherwise, the oxygenates,which elute will be
quantitated as hydrocarbons and some error due to differing response fac-
tors will result. For the samples collected in this study, oxygenates
do not apear to be a major emittant. Table 1.2-a includes the column
types and GC parameters used for oxygenate analysis. Column 3 also works
well for terpene compounds, although separation is superior with the
glass capillary column.
1.2.1 Calibration
The G£ analyses were calibrated using Scott standard # 54 containing
ethane, ethylene and acetylene and a specially prepared Scott standard
which contains these components plus methane. The area responses for ecch
20
-------
Table 1.2-a. VOC ANALYSIS CONDITIONS
Compound^
Instrument
Operating Conditions
Ethylene
Ethane
Acetylene
P.E. 3920 Iso-
thermal FID GC
Light Hydrocarbon
C2-C6
P.E. 3920 Temp.
Prog. FID GC
Heavy Hydrocarbon
and Oxygenates
C4-C12
Heavy Hydrocarbon
c4-c12
P.E..3920 Temp.
Prog. FID GC
21
Column: 10' x 1/8" OD Porapak Q
Carrier: He 80 pslg, 7 ml/min.
Hydrogen: 22 psig
Compressed Air: 50 psig
Oven: 65° (30°C for CH4)
Total Run Time: 10 min.
Sample Size: 100 ml (5 ml for CH4)
Column: 20' x 1/16" OD Durapak
N-Octane
Carrier: He 90 psig, 6 ml/min
Hydrogen: 40 psig
Compressed Air: 50 psig
Oven: -70°C to 65°C
Delay time: 4 min
Program rate: 16°/min.
Total Run Time: 40 min
Sample size: 500 ml
Column: 10' x 1/18" Durapak Low-K
carbowax 400
Carrier: He 90 psig, 8 rnl/min.
Hydrogen: 40 psig
Compressed air: 50 psig
Oven: -20 to 100#C
Delay Time: 2 min.
Program Rate; 8°/min.
Total Run Time: 20 min
Sample size: 500 ml
Column: 200' SCOT OV-101 with
10' x 1/16" OD Durapak
Low-K, Carbowax 400
precolumn
Carrier: He 90 psig, 5 ml/min
Hydrogen: 40 psig
Compressed Air: 50 psig
Oven: 0°C to 100° Ternp. Prog.
Delay Time: 6 min.
Program rate: 6°/min
Total Run Time: 60 min.
Sample size: 500ml
Column: 30 m SE 30 Glass Capillary
Column
Carrier: He 90 psig, 1 ml/min.
Oven: -30 to 80°C Temp. Prog.
Delay Time: 8 min.
Program Time: 4°/min
Total Run Time: 50 min.
Sample size: 500 ml
-------
Figure 1.2-a. Typical analysis of a vegetation sample (Juniper)
ro
ro
60
40
20 —
SAMPLE 143 JUNIPER
0 2 4 6 8 10 12
-TO -38 -6 26 58 |
TEMPERA' uw es—^- HOLD
100 —
I i i l l l l r I 1
18202224262830323436
MINUTES
SAMPLE 143 JUNIPER
—T—i—i—i—i—i—i—i—i—r—71—i—r
0 2 4 6 8 10 12 14 16 18 20 22 24 26
-3O -22 -14 -6+2 10 18 26 34 42 50 65 ^ HOLD
TFMPFPATLIRF
T 1 1
28 30 32
MINUTES
34 36 38 4O 4?
-------
Figure 1.2-b. Sample chromatogram, glass capillary column (#5)
PONDEROSA PINE
J^AJL
, -»»£
INCREASING TIME a TEMPERATURE
23
-------
unknown peak in the sample are proportional to the standard concentration
on a peak area basis. When an integrator is not available, peak height
proportions can be compared with the peak heights of each compound in the
standard. The concentrations calculated are therefore in the units of
yg/nr for each compound.
The other GC's were routinely calibrated using a neo-hexane standard
(0.209 ppm) prepared by Scott Research, Inc. Five milliliters of this
standard were injected into the freeze-out loop using a pressure-lok® sy-
ringe manufactured by Precision Sampling Corporation. This was followed
by 100 ml of zero air to insure that the sample was frozen out in a
concentrated "slug" in the freeze-out loop. The light hydrocarbon GC
was then run isothermally at 65°C until the neo-hexane peak appeared.
The heavy hydrocarbon GC was temperature-programmed from -30°C to 65°C
until the neo-hexane eluted. The area of the peak was integrated and
the response was calculated with respect to nanograms neo-hexane/area
response.
All concentrations and emission rates in this report are in terms
of ug of compound. Therefore volatile organic compounds (VOC) refers to
volatile organic carbon compounds. Table 1.2-b. illustrates the differ-
ences between some common methods of organic compound quantitation. It
was assumed in this study that all VOC measured had an FID response equal
to the FID response to neo-hexane.
Periodically a qualitative standard was analyzed. This standard was
prepared by adding concentrated microliter amounts of specific compounds
to a volume of zero air or ultrapure helium in a stainless steel canister.
After one or two compounds were added, the standard was analyzed and the
24
-------
Table 1.2-b. COMPARISON OF EQUIVALENT UNITS OF VOC QUANTITATION
Units
Molecular weight (MW)
compound (g/mole)
Weight of carbon (CW)
(g/mole)
Ratio of MW to CW
o
yg/m compound
o
yg/m carbon
*ppb v/v compound
*ppb v/v carbon
+ppb wt/wt compound
+ppb wt/wt carbon
*at 25°C and 760 mm Hg
Density of dry air = 1185
Methane
(CH4)
16
12
0.750
1
0.750
1.528
1.528
0.844
0.633
g/m3
Isoprene
(C5H8)
68
60
0.882
1
0.882
0.360
1.80
0.844
0.744
Terpenes
(C-jgHis)
136
120
0.882
1
0.882
0.180
1.80
0.844
0.744
Note: To convert the compounds listed from yg/m3 to another unit of
measure in the bottom part of this table, multiply by the number
in the appropriate column.
25
-------
peaks were labeled. Gradually a wide range of compounds were introduced,
one or two at a time, and their elution times identified chromatographic-
ally. The procedure can result in a mixture of the full range of com-
pounds expected in the samples to be analyzed. The unknown samples are
then analyzed and tentatively identified by matching elution times with
those of the standard. This tentative identification can be confirmed
using a GC-MS system.
1.2.2 Sample Enrichment
It is necessary to use a sample size of up to 500 ml for biogenic
emission samples and up to 1000 ml for ambient air samples to obtain the
required gas chromatographic response. This sample size is too large to
inject directly into the GC and will overload the columns and destroy
their resolution capabilities. Therefore a standarized sample enrichment
step is necessary. The preferred method is the freeze-out technique.
With this method a 3.2 mm (1/8") o.d. stainless-steel loop filled with
60-80 mesh glass beads is connected to a six-port sampling valve. When
concentrating a sample, the loop is immersed in liquid oxygen. The sample
is then pulled through the loop by a calibrated vacuum system or injected
into the loop with a 100 ml ground glass syringe. All of the organics
are retained on the glass beads while the nitrogen and some of the oxygen
pass through. The valve is then switched. This causes the gas chromato-
graph carrier gas to be introduced from the reverse direction of that
which the sample entered. At the same time, the loop is immersed in hot
water. This procedure flushes the contents of the loop onto the head
of the column in a concentrated slug and results in excellent resolution
and sensitivity.
26
-------
Several solid adsorbents are available as sample concentrators; how-
ever, WSU has found that they are usually satisfactory for only specific
hydrocarbons, sometimes contribute background bleed peaks to the sample,
and require extremely careful sample introductions and temperature control
to get reproducible results. The freeze-out loop method of sample concen-
tration is preferred because of its simplicity, quantitativeness and
proven reliability.
1.2.3 Sample Introduction
Sample introduction into the freeze-out loop can be accomplished by
two methods. One method utilizes a 100 ml syringe and side-port needle.
The syringe is inserted through a Teflon-backed silicon rubber septum
placed over the sample canister valve opening. The 100 ml sample is drawn
out and forced through the liquid oxygen-immersed sample loop. This
procedure is repeated until the desired sample size of 500 to 1000 ml is
reached. The other method involves the use of a vacuum sampling system
(Figure 1.2-c). This system uses a chamber of known internal volume
which has been evacuated. The vacuum gauge is attached to the chamber
and calibrated so that a specified change in vacuum is equivalent to a
specified sample volume. The evacuated chamber is located on the down
stream side of the freeze-out trap, and therefore it cannot contaminate
the sample. The vacuum pulls the sample from the sample canister through
the freeze-out loop where the organics are trapped. If the vacuum
system is used, the sample volume does not require correction to STP.
However, if the standard is injected via syringe, it must be corrected
to STP since the volume of the syringe is affected by the daily barometric
pressure.
27
-------
FIGURE 1.2-c
VACUUM SYSTEM
for sample injection
OFF
^
VACUUM CHAMBER
VACUUM SAMPLE
INLETJ
BACK VIEW
28
-------
1.2.4 Analysis
The samples should be analyzed as soon as possible after collection.
It is desirable to analyze the emission sample first, then the background,
ambient air and bag blank samples (if collected), in that order. Ratios
of the components of a sample may vary with time; however, WSU's experi-
ence during this study showed that the VOC in the sampling canisters
will not vary by more than 15% after a storage of several days.
1.2.5 Emission Rate Quantitation
The light hydrocarbon analysis provides good component separation
for hydrocarbons in the C2 ~ ^6 ran9e* Samples of 500 ml or larger,
which were collected under very humid conditions or wet samples which
are introduced into the freeze-out trap through the potassium carbonate
dryer at too fast a rate, may cause ethylene to elute as an unquantifiable
peak. In those cases ethylene, ethane and acetylene values are obtained
by analyzing the sample on the Porapak "N" column. Methane can also be
quantified on this column by simply filling the sample loop (no sample
concentration step) and injecting isothermally at 30°C.
The heavy hydrocarbon analysis provides quantisation for C^ - C-^2
hydrocarbons. To determine the non-methane VOC total, the C4 - C12 totals
obtained from the heavy VOC analysis are added to the C2 - C4 totals
obtained from the light VOC analysis (and/or C2 analysis).
1.2.6 Biomass Quantitation
After the emission sample is collected the branch should be clipped
and stored in a paper bag. Then the foliage should be placed into an
oven at 70°C Tor 3 to 4 days until it reaches a constant dry weight.
29
-------
If many samples of the same species are to be collected, it may be desir-
able to measure the fresh weight of the sample so that the ratio of
fresh weight to dry weight can be determined. This will aid in estimating
tentative emission rates from branches which will be resampled at a
later date and have not been clipped. After drying it may be desirable
to separate the foliage into its leaf, twig and branch components, and
then to weigh each. Whether leaf, twig and branch biomass, leaf biomass
or some other biological parameter is used will depend upon the biomass
factors available for the study area. For the emission samples collected
by WSU, dry leaf biomass was used in the calculation of vegetation emission
rates.
Emission testing by WSU has shown that there are no significant dif-
ferences between emission rate estimates developed on the basis of leaf
biomass or leaf twig and small branch (jC 4 cm O.D.) biomass. The
emission rates are smaller when calculated in terms of leaf twig and
small branch biomass than when calculated for leaf biomass only. But
since the leaf twig and small branch biomass/unit area is larger than
leaf biomass alone, the total emission factor is the same as when the
emission rates are calculated in terms of leaf bimass only and multiplied
by the smaller leaf biomass/unit area factor. As long as the allometric
relationship between the chosen sample parameter (leaf biomass, leaf
surface area, Chlorophyll content, etc.) remains constant for the species
being sampled (on a ground area basis), the parameter used to relate the
emission rate of the enclosed portion of a plant on a sample basis, to
the emission rate on an area basis, is unimportant. The emissions from
tree trunks have not been considered in this report. It was assumed
30
-------
that tree trunk emissions would be insignificant compared to the emissions
from leaves.
1.3 SPECIAL FIELD TESTING PRECAUTIONS
1. Be sure all sampling equipment is free of contamination before
an extensive field testing program is begun. Special care should be
taken to slowly purge and to heat sample loops and columns to at
least 100°C between each heavy hydrocarbon analysis. Otherwise
residual hydrocarbons may contaminate the next sample. This may be
evidenced as large broad "ghost peaks" with peaks from the next
analysis superimposed. If a syringe is used to introduce the sample
into the freeze-out loop, it should be purged with zero air between
G.C. runs.
2. If samples are suspected of having large amounts of moisture, use
the smallest sample volume which will allow quantisation of the VOC
present. Water may interfere with the light hydrocarbon (LHC) analysis.
This problem can be minimized by passing the sample slowly ( 25 ml/min)
through a 1/4" x 12" copper tube packed with potassium carbonate.
The potassium carbonate will dry the LHC sample without affect-
ing its composition or quantisation. Excessive water or an excessive
rate of sample introduction through the potassium carbonate dryer
will be evidenced by a poor peak shape for ethylene. The fy column
is relatively insensitive to water. Therefore, ethylene quantisation
can be obtained from the G£ analysis if necessary.
The glass capillary column for heavy hydrocarbon analysis is
also relatively insensitive to water. However, large amounts of
water could strip some of the capillary coating material and result
31
-------
in large rounded peaks. To date, no samples have been collected
which cause this problem with the sample sizes introduced (100-500 ml.)
3. Extreme care must be taken when enclosing a branch not to break
twigs or leaves. Broken surfaces tend to emit a large amount of VOC
immediately and then emissions rapidly taper off. Since the back-
ground sample is collected after enclosure it will partly account for
these excess emissions. However, if broken twigs are noted, it is
best to abandon the selected branch and sample another.
4. After the initial flush of 60 *, of zero air is put in the bag
(10 «,/min for 6 min) and the purge flow-rate is started (2 Vmin),
the contents of the bag can be mixed by gently grasping each side
of the bag and using a gentle kneading motion to stir the air inside.
5. If two consecutive samples are to be collected in short succes-
sion using the same bag, it may be desirable to turn the bag inside
out for the second sample.
6. When sampling leaf-litter, it is not necessary to firmly compress
the dirt between the bag collar and the sealing ring. Also, as in
branch sampling, care should be taken to minimize vegetation damage.
1.4 EMISSION RATE DETERMINATION
Emission rates cannot be calculated until the following have been
completed:
1. Field data recorded.
2. Each emission sample, background sample and bag blank and/or am-
bient air sample analyzed.
3. VOC in each sample calculated.
32
-------
4. The biomass for each sample determined by drying the leaves
and weighing or estimating allometrically from the dry weights
of other samples.
Basically, the emission rate is equal to the mass of VOC emitted
per unit time per unit biomass or surface area.
The emission rate formula used for most of the samples collected during
the course of the studies reported here was:
C-- (Zv + Ve) - C Ve
CD-
(Sa) (A^)
where:
Ccc: (yg/nP) equals the TNMOC measured for the emission sample
o ^
Csb: (yg/m3) equals the TNMOC measured for the background sample
Zv: (m) equals the total volume of zero air put into the enclosure
Sa: (g) equals the chosen biomass component of the sample
T-J: (min) equals the total emission time = Tg - T-j
Ve: (m3) equals the dead volume of the bag when collapsed around the
branch. It can be calculated on the basis of the dilution of a
marker component. If, for example c^ is equal to the concentration
of acetylene in the background sample and cs is the concentration in
the emission sample, then:
M ZV
Ve=
The resulting emission rate (ER) developed is in terms of micro-
grams of non-methane VOC emitted per minute per gram of vegetation or
per square meter of litter. (Note that any set of units, either on a
mass or area basis, can be used if the appropriate values of Sa are
substituted in the equation for ER.)
33
-------
The emission rate formula described in the main text (Section 1.4)
applies only to samples collected using a collapsable chamber (Teflon
bag) and sample containers which have a very small dead volume such as
evacuated stainless steel canisters, Teflon bags, and adsorbent traps.
Appendix B shows the derivation of the emission rate equation and illus-
trates the formulas used to calculate emission rates when the basic
sampling procedure is not followed.
It should be noted that the above emission factor development is
based on the following assumptions:
1. The enclosure contents are evenly mixed prior to sampling.
2. Enclosure volume and concentration are constant over the short
sampling time.
3. The enclosure is at ambient pressure.
34
-------
2. EMISSION INVENTORY DEVELOPMENT
2.1 GENERAL
To be useful, the emission factors developed through use of the
procedures outlined in Chapter 1 should be incorporated into an inventory.
Such an inventory can be used to estimate spatial and temporal emission
patterns and emission densities that can, together with other kinds of
information, be used in oxidant control strategy development.
Basically, three components are necessary for inventory development:
(1) a set of representative emission factors for the vegetative species,
leaf litter and water surfaces in the area, (2) an indication of the
conditions (season, temperature, etc.) that prevail for the time interval
of interest and (3) biomass density factors, which are a measure of the
quantity of vegetation or litter present in the the area. In this sense,
inventorying vegetation is analogous to any other area source category,
i.e., an appropriate emission factor is multiplied by some source activity
level in order to estimate emissions.
As with any other source category, the same limitations must be
recognized when preparing and using the vegetation emission inventory.
First, most emission factors are generally only estimates of emissions
for a particular set of conditions, i.e., those that prevailed during
the actual source testing. In this study, most of the testing was done
during the summer and fall. Only limited testing was conducted at cold
35
-------
temperatures and on dormant or dead plants. Second, the biomass factors
themselves (derived from the literature or developed from actual field
surveys) are generally only approximations, the variability of which
depends on the method used to derive them and the area to which they
are applied. Third, various types of errors are always unavoidable in
any program involving physical and statistical sampling. These include
the imprecision and inaccuracy that are associated with the sampling and
analytical procedures as well as the sampling error that results from
using limited data bases to estimate the characteristics of entire popu-
lations.
2.2 LEAF BIOMASS DENSITY FACTORS
All of the vegetation emission rates used in this report are calcu-
lated in terms of leaf biomass. In other words, it was assumed that the
emission rates for each vegetation sample would be proportional to the
dry weight of the leaves present. It was further assumed that this
emission rate relationship would not change when extrapolating the
sample emission rates to larger area emission factors. To calculate
vegetative emissions over some area of interest, these sample emission
rates (pg/g»hr) must be multipled times approximate indicators of the
amount of leaf biomass present per some unit of area, i.e., leaf biomass
density factors (g/nr). The result is an area wide emission factor
(yg/m^-hr). Leaf biomass density was selected as the most appropriate
indicator for the following reasons:
1. A broad and well-defined data base exists for leaf biomass
density factors which can be applied to regional or area specific
emission inventories (Lieth and Whittaker, 1975; NAS, 1975).
36
-------
2. Leaf biomass is easily measured in the field for emmision rate
samples. The measurements require no special equipment or training
to complete. The averages of sample emission rates can then be
directly multiplied by leaf biomass densities to result in area
emission factors. (The emission rates for leaf litter or water
surfaces are simply multiplied times the area covered to estimate
emissions).
3. Leaf biomass densities for many forest types tend to be uniform
over broad areas and are relatively insensitive to site-specific
variables (Satoo, 1967; Bazilevich, _££. al_., 1968).
Much of the data concerning the dynamics of biomass distribution
within various ecosystems has been generated as the result of investiga-
tions carried out under auspices of the International Biological Program
(IBP). The IBP was a multinational, multidisciplinary program with the
ultimate goals of developing ecosystem models by reducing the interac-
tions of environmental factors into a series of mathematical relation-
ships. These models could then be used to predict the impact of various
land use options upon specific ecosystems. One of the achievements of
the IBP program was the formation of various models to predict the primary
productivity or the amount of carbon formed for various land and vegeta-
tion types. In order to develop these models, fieldwork was conducted
to harvest tree species for various forest types and to quantify each
component.
The results of these studies, conducted in countries throughout the
world and at 21 research sites in the United States, were used in this
report to estimate leaf biomass densities so that an example VOC emission
inventory for the U.S. could be compiled.
37
-------
The estimation of leaf biomass density factors is made easier because
the amount of leaves over a given area of ground is fairly independent
of the type of vegetation and the age of vegetation (Lieth and Whittaker,
1975). Many authors have also reported that stand density, after canopy
closure, does not affect the leaf biomass per ground area (Satoo, 1967).
It has further been reported that the leaf biomass density of many dissim-
ilar types of vegetation from grassy savannas and subtropical rain forests
P
to northern and middle Tiaga coniferous forests range from 600 to 1250 g/rrr
(Bazilevich, _et_.al_., 1969). Leaf biomass density factors for the United
States seem to range from 100 g/nr for extreme desert conditions to
O
1100 g/rrr for the most productive rain forests (NAS, 1975).
An additional advantage of using leaf biomass density in compiling
an emission inventory is the availability of locally or regionally specific
information for detailed emission inventories. Local sources of informa-
tion which may be useful include agricultural crop yield reports, local
forest surveys, land use planning studies, and doctoral dissertations on
local ecosystem types. IBP publications such as Primary Productivity of the
Biosphere, (Lieth and Whittaker, 1975) supplement this information with
conversion factors which can be used to arrive at leaf biomass densities.
The IBP classification system for specific land forms and vegetation
types incoporates the concept of "biomes". Biome classifications are
based upon the composition of the potential vegetation (before disturbance
by man), the physiography and physiognomy of the land and on the climate
of the area. Many of the biome names correspond to types of vegetation
(i.e., Deciduous Forest Biome or Coniferous Forest Biome), however,
their classification is independent of transient vegetation types such
as hardwoods or conifers. For this reason, although Pine plantations
38
-------
have rapidly replaced many hardwood stajids in the Southeast United States,
the area is still classified due to its potential vegetation, land form,
and climate as a Deciduous Forest Biome. Figure 2.2-a illustrates a
generalized biome composition for the Continental U.S. The numbers on
the map identify each biome so that their areas can be estimated (Table
2.2-a).
2.3 EMISSION RATE ALGORITHMS
Early in the research program it was recognized that certain sample
variables seemed to affect emission rates. Terpene type emissions tended
to be greater at higher temperatures, low elevations and early in the
growing season than emissions collected at low temperatures, high eleva-
tions or late in the fall. Rasmussen (1970) indicated that isoprene
was only emitted by certain plants and only in the light. The field
measurements in this study confirmed this light dependency for isoprene
emissions. The variability of the field data indicates that although
trends are present, many other sample factors could affect emission
rates. These variables include site specific variables such as soil
fertility, plant moisture, weather, individual variability, location of
the sample on the tree, various pathologic conditions such as disease or
injury and the age of the vegetation.
In order to more clearly estimate the effects of temperature and
light on emission rates, a laboratory research program headed by Dr. D.
Tingey, EPA Corvallis, has been conducted utilizing specially designed
environmentally controlled chambers (Tingey, ^t jl_., 1978 a,b). Whole
plants were place inside the chambers and the selected variable of temp-
erature or light was changed while other conditions remained coastant.
39
-------
Figure 2.2-a. MAJOR BIOTIC REGIONS OF THE US
GRASSLAND
SCLEROPHYLL SCRUB
TEMPCRATE RAIN FOREST
DECIDUOUS FOREST
CONIFEROUS FOREST
DESERT
TUNDRA, ALPINE FIELDS
40
-------
Table 2.2-a. AREAS OF THE MAJOR BIOTIC REGIONS OF THE CONTINENTAL U.S.
Biotic Region
Temperate Grassland
Sclerophyll Scrub
Temperate Rain Forest
Deciduous Forest
Coniferous Forest
Desert Scrub
Alpine Fields or Tundra
Total Continental U.S.
Map No.
3
5
19
Total Grassland
14
1
17
2
4
6
7
8
9
13
16
Total Coniferous Forest
15
18
Area (km2)
133,577
2,441,864
31,375
2,606,816
287,162
158,846
3,105,709
214,925
228,922
14,278
108,966
97,881
27,054
257,290
243,892
1,192,892
1,654,681
33,159
9.04 x 106
41
-------
These experiments, completed for Live Oak, (an isoprene emitter) and
Slash Pine, (a terpene emitter), indicate that there is a relationship
between temperature and emission rates. For terpene emissions no light
dependency could be detected. Isoprene emissions varied logarithmically
with temperature. A strong light dependency was also noted for isoprene
emission, however, maximum isoprene emissions were reached at fairly low
intensities and increasing light intensity above this point of saturation
did not increase isoprene emission rates. The study quantified the
relationships between temperature and terpene emissions at any light
level, between isoprene emissions and temperature at various light levels
and between isoprene emissions and light at various temperatures.
For purposes of this emission inventory WSU assumed that the change
in VOC emission rates with temperature for all vegetation types (except
for isoprene emitters) would vary similarly to the Slash Pine studied by
Tingey. WSU also assumed that changes in emission rates with temperature
for isoprene emitting species would vary similarly to the Live Oak in the
EPA-Corvallis study. An additional assumption was made that all isoprene
emitting species were light saturated during the daylight hours. This
means that emission rates might be over-estimated for leaves deep within
the canopy or in very shady locations during some part of the day. For
nighttime, all isoprene emission rates were assumed to be zero. Finally;
WSU assumed that leaf temperature would approximate the temperature meas-
ured inside the bag during the sampling. Figure 2.3-a shows the emission
rate algorithms used to calculate the respective VOC emissions. Since
the field data were collected under a wide range of temperatures, a
correction factor was used to standardize the VOC emissions to specified
conditions of saturating light and a leaf temperature of 30°C. The
42
-------
Figure 2.3-a. Emission rate algorithms
Isoprene
(Er) = - i^§ - +0.11
1 + exp [-0.18 (Ta - 25.26)]
Isoprene Temperature correction factor to 30°C:
34.194
Er = Er
*
exp
, _ 1 + exp [-0.18 (Ta - 25.26)] _
where: Er* = Isoprene emission rate (measured)
Er = Isoprene emission rate (std. to 30°C)
Ta = Leaf temperature
exp is an exponential
Terpenes
++Er = exp [-0.332 + 0.0729 (Ta)]
Terpene correction factor to 30°C:
6.392
Er = Er*
exp [-0.332 + 0.0729 (Ta)]
where: Er* = Terpene emission rate (measured)
Er = Terpene emission rate (standardize to 30°C)
Ta = Leaf temperature
exp is an exponential
+From Tingey et al., 1978a.
"""From Tingey et al., 1978b.
43
-------
correction factors take the form of the ratio of Tingey's (et al., 1978a,b)
predicted emission rate at 30°C to his predicted emission rate at the
sampling temperature, times the emission rate which WSU measured in the
field. According to the equation in Figure 2.3-a, temperature correction
must take into account the temperature of the leaves themselves. Since
leaf temperatures, per se, were not routinely measured in the study, an
assumption had to be made concering the relationship between leaf temper-
atures and the measured bag and and air temperatures. From energy balance
calculations it is apparent that leaf temperature and air temperature
inside our enclosure during sampling are very close (Gates, 1971). The
relationship between air temperature and bag temperature and leaf temper-
ature, however, is difficult to estimate. The primary factors that affect
this relationship are the size of the leaf, the energy absorption by the
leaf, wind speed and transpiration rate (Gates, 1965). From our field
measurements, it appears that in the morning or afternoon hours or if
the sunlight is filtered through foliage or shaded by clouds, bag temper-
atures are within 5°C of ambient air temperatures.
Because bag temperature more accurately reflects leaf surface temper-
ature, a probable controlling factor for emissions, the raw emission rates
were specified in terms of bag temperature. When the emission rates are
standardized to an ambient temperature of 30°C, the possibility of under-
estimating emission factors is enhanced. For instance, the emission
rate measured at a bag temperature of 35°C is necessarily lowered when
standardized to prevailing ambient conditions of 30°C. Under these condi-
tions leaf surface temperatures of unenclosed as well as the enclosed vege-
tation are probably closer to the bag temperature than to the ambient air
44
-------
temperature (Gates, 1971). Therefore, when the emission rate is standard-
ized to an ambient air temperature of 30°C, the effect is a lower emission
estimate than would be expected at a corresponding leaf surface tempera-
ture of 35°C.
Although leaf temperatures may be higher than ambient temperatures
for some leaves during some period of the day, it is much more difficult
to estimate average diurnal leaf temperature cycles than average diurnal
air temperature cycles. For this reason, in this inventory WSU has
assumed the bag temperatures equaled air temperature. It was recognized
that this assumption could lead to underestimation of emission rates.
This potential underestimation of emission estimates would be moderated
somewhat for isoprene emitters because during periods of direct sunlight
temperatures of some leaves may exceed 44°C and the leaf would then begin
to physiologically shut down (Tingey, _et_. al., 1978a). Since isoprene
emissions seem to be tied to photosynthesis (Sanadze and Kalandadze, 1966)
the isoprene emission rate would be reduced for the over-heated leaves.
In other words, in bright sun, leaf temperatures of some of the leaves
for some broadleafed plants tend to be wanner than ambient air during some
hours of the day, causing emission rates based only upon bag temperature
and standardized to ambient air temperature to be too low. However,
some of the leaves of a canopy may exceed temperature of 44°C, causing a
sharp decrease in isoprene emission rates. These factors, therefore,
may tend to balance.
The ranges, as well as the magnitudes, of emission rates from leaf
litter, soil and water surfaces were much smaller than for vegetation.
Their temperatures remained more uniform and no exmperimental work on
emission rate algorithms had been completed. Therefore, no attempt was
45
-------
made to standardize the emission rates for these categories. Additionally,
for this study it was assumed that the emission rate of an enclosed
branch at a specific bag temperature would be representative of the
emission rate of the whole plant if the ambient temperature was equal to
the bag temperature.
2.4 NATURAL EMISSION DATA
Each vegetation species exhibits a characteristic spectrum of emis-
sion components. Table 2.4-a lists the emission profiles for some species
sampled in Florida. Table 2.4-b exhibits typical emission ratios of sel-
ected species sampled near Santa Barbara, CA, Pullman, WA and RTF, NC.
The percentages listed are averages. The specific sample ratios vary
with sampling conditions and with the individual samples; however, the
major components are consistenly emitted for each vegetation type. As
Table 2.4-a and 2.4-b show, the major emission components do not add up
to 100%. For some samples the remainder of the total emission consisted
of many small component peaks. For many samples these peaks typically
eluted in the portion of the chromatogram where paraffin type compounds
elute. For many of the California samples such as Mesquite and Manzanita,
the emissions eluted from the GC column in the area of the chromatogram
where paraffins normally elute or later in the chromatogram where some
aromatic compounds and oxygenates elute. In these samples, emission
components could not be matched with known standard compounds as GC-MS
analysis was unavailable at the time that the samples were collected.
The emission "fingerprint" of these samples was dissimilar to any other
vegetation type sampled previously.
46
-------
Table 2.4-a. EMISSION PROFILES FOR SELECTED FLORIDA VEGETATION
Vegetation Type
All oaks (Quercus spp.)
Long Leaf Pine
fPinus palustris)
Major Emissions
Approximate
Percent of non-methane VOC
Slash Pine
(Pinus elliotti)
Sand Pine
(PijTus clausa)
Austrailian Pine
(Casuarina eQuisetifolia)
Saw Palmetto
(Serenqa repens)
Sabal Palmetto
(Sabal Palmetto)
Cypress
(Taxodium distrlchum)
Sweet gum
flJQuidumbar stvraciflua)
Isoprene(daytime only)
a-Pinene
B-Pinene
drLimonene
A -Carene
Myrcene
Propane
a-Pinene
8-Pinene
dsLimonene
A -Carene
a-Pinene
6-Pinene
Isoprene
Isoprene (daytime only)
Isoprene
a-Pinene
AJ-Carene
Isoprene
a-Pinene
3-Pinene
d-Limonene
AJ-Carene
3-Phellandrene
90-99
30
30
2
2
1
8
27
16
19
12
35
44
92
85
90
46
21
20
13
02
02
06
51
47
-------
Table 2.4-b. EMISSION PROFILES FOR SELECTED SPECIES
Vegetation Type
Pullman, Washington
Ponderosa Pine
Lombardi Poplar
Douglas Fir
California
Manzanita
Chemise
Eucalyptus
RTP, NC
Dogwood
Yellow Poplar
American Sycamore
Eastern Red Cedar
Loblolly Pine
Shortleaf Pine
Virginia Pine
Major Emissions
0
A°-Carene
3-Pinene
a-Pinene
Isoprene
a-Pinene
d-Limonene
$-Pinene
unknown 8A
unknown 4A
Isoprene
n
AJ-Carene
d-Limonene
unknown #28
Terpinolene
Isoprene
0
A°-Carene
a-Pinene
a-Pinene
d-Limonene
3-Pinene
A^-Carene
a-Pinene
d-Limonene
3-Pinene
A3-Carene
unknown #26
a-Pinene
3-Pinene
d-Limonene
A -Carene
Approximate
Percent of non-methane VOC
36
25
14
99
24
7
5
18
25
40
10
10
8
5
74
52
26
35
26
19
11
53
17
12
11
26
22
10
38
2
48
-------
In general, most of the species tested exhibited definite emission
patterns. Oaks emitted primarily isoprene. Conifers emitted primarily
o
terpene type compounds such as a-pinene, 8-pinene and A°-carene.
Table 2.4-c presents a summary of the VOC emissions for broad cate-
gories of vegetation grouped together on the basis of similarities in
emission components and in emission rates. All oaks are grouped together.
All conifers are also averaged. From examination of emission rate data it
appears that one of the primary factors influencing emission rates is
whether or not the plant is a prolific isoprene emitter. For this reason
all of the non-oak isoprene emitters (where isoprene is greater than 50%
of the VOC) are averaged and all of the non-oak non-isoprene emitters are
averaged. Additionally, groupings were made of emission rates for leaf
litter and pasture samples, marine samples, and fresh water samples.
(Note that the emissions for these latter categories are expressed in
terms of yg/nr'hr, while the emission from vegetation are expressed in
terms of ug/g'hr.) The values in Table 2.4-c represent the collection
of samples over a broad range of temperatures and seasons. The first
column of data in Table 2.4-c shows the "measured" means and standard
errors for the samples at actual ambient temperatures. The "standardized"
column uses the same data corrected to 30°C. The "measured" Florida
data for Oaks and Non-Oak Isoprene Emitters does not include samples
collected at night when isoprene emissions do not occur. RTP data could
not be standardized for temperature, as the probe which monitored bag
temperatures malfunctioned during sampling. No sweetgum samples were
included. Sweetgums have relatively large delicate leaves and it was
very difficult to collect samples without causing vegetation damage.
Unfortunately, this problem was not apparent until the emission rates
49
-------
Table 2.4-c. NON-METHANE VOC EMISSION RATES OF BROAD VEGETATION
CATEGORIES FOR VARIOUS SAMPLE SITES
Measured Standardized to 30°C
Classification (units)
Confiers (yg/g-hr.)
Pullman, WA -76
Pullman, -77
North Carolina
Florida
Oaks (yg/g'hr.)
Pullman-76 (Dark)
North Carolina
California
Florida
Non Conifer, Non Isoprene (yg/g
Pullman-76
Pullman-77 (Sagebrush)
North Carolina
California
Florida
Non Oaks-Isoprene (yg/g-hr.)
Pullman-76
Pullman-77
North Carolina
California
Florida
N
21
5
7
70
2
3
3
47
•hr.)
9
2
8
17
143
1
1
2
1
55
X
2.42
2.65
8.35
9.36
2.55
26.8
8.60
22.93
2.92
36.0
4.68
3.48
4.28
38.7
37.1
5.02
10.6
17.45
SX
0.40
0.93
3.93
1.31
0.53
12.26
1.25
3.35
1.43
31.89
1.72
0.97
0.48
...
—
0.90
3.51
X
4.97
3.64
_ —
8.80
3.00
28.7
21.93
7.76
23.6
4.13
4.74
27.7
....
9.16
SX
0.74
0.93
1.27
0.61
9.76
4.14
4.24
17.18
....
1.01
0.67
____
Leaf Litter-Pasture (yg/m2-hr.)
Pullman 5 105.8 35.5
California 2 262.0 80.6
Florida 101 162.0 17.6
Marine (yg/nr.hr.).
Florida 141 128.9 10.2
n
Aquatic (yg/m-hr.)
Florida 11 102.4 13.4
50
-------
had been calculated. The extremely high VOC content of the background
samples made it necessary to estimate the dead volumes of the vegetation
samples to within 0.1 liter, which was not possible. Therefore, emission
rates for sweetgum could only be estimated to range from 0 to 120 yg/g-hr.
Each sample can be considered to represent an average emission rate.
Since the means of each species were averaged, the variability is best
expressed in terms of a standard error of the means, rather than as the
standard deviation. The standard error is equal to the standard deviation
divided by the square root of the number of observations.
Although there is a range of variation between emission rates of
different species collected in different locations, some broad generali-
zations based upon the data available can be made. For instance, it is
apparent that the oak emission rates for summer samples collected in
Florida are similar to those for California. Emission rates for conifers
in Florida and North Carolina also appear to be similar. Conifer emis-
sions from samples collected during the fall in Pullman WA appear to be
much lower. It is not presently known whether this latter difference
represents species, site, or seasonal effects upon emission rates. It was
assumed for this study that the lower emission rates for Pullman conifers
represented seasonal effects. Based upon the limited data collected in Pullman,
California and RTP, North Carolina, and the more extensive sampling conducted
in Florida, generalized emission factors were determined that were estimated
to be representative of broad categories of vegetation in the U.S.
Table 2.4-d shows the vegetation species composition of the samples
used to. determine the emission rates for the broad vegetation categories.
Generalized emission factors for broad vegetation categories are shown in
Table 2.4-e.
51
-------
Table 2.4.d. LOCATION AND SPECIES OF SAMPLES USED IN U.S. INVENTORY
Species Number of Samples
Non-Conifer Non-Isoprene Emitters
Species
Number of Samples
Non~Conifer Non-Isoprene Emitters
Pullman
Maple
American Elm
Weeping Willow
Winter Wheat
Dry Peas
Sage Brush
North Carolina
Pignut Hickory
Mockernut Hickory
Dogwood
Yellow Poplar
American Beech
American Hornbeam
Sourwood
California
Adenostema Fasiculatum
Manzanita
Chemise
Big Pod Ceanothus
Eucalyptus
Florida
Mangrove
Wax Myrtle
Elderberry
Groundsel Bush
Persimmon
Dahoon Holly
Red Mulberry
Sweet Acacia
Viburnum
Oleander
Oranges
Grapefruit
3
1
1
1
1
4
1
1
2
2
1
1
1
5
3
4
4
37
9
5
2
17
3
3
3
1
1
29
16
Florida (cont.)
American Elm
Carolina Ash
Red Maple
Hickory
Non-Oak Isoprene Emitters
Pullman
Lombardy Poplar
North Carolina
American Sycomore
California
Eriopictyon Traskii
Florida
Saw Palmetto
Sabal Palmetto
Australian Pine
Willow
Leaf Litter-Pasture
Pullman
Conifer Duff
California
Eucalyptus-Oak
Florida
Pasture
Aquatic
Florida
Man" ne
Florida
1
1
9
4
2
2
1
35
12
1
7
5
2
101
11
141
52
-------
Table 2.4.d. LOCATION AND SPECIES OF SAMPLES USED IN U.S. INVENTORY (continue
Species
Conifers
Pullman
Ponderosa Pine
Douglas Fir
Juniper
Norway Spruce
Blue Spruce
Sub Alpine Fir
Mugo Pine
Port Orford Cedar
Grand Fir
Western Larch
North Carolina
Shortleaf Pine
Virginia Pine
Loblolly Pine
Eastern Red Cedar
Florida
Slash Pine
Longleaf Pine
Sand Pine
Southern Red Pine
Cypress
Number of Samples
7
8
3
2
1
1
2
1
1
1
2
1
2
2
16
29
4
1
20
Species
Oaks
Number of Samples
North Carolina
White Oak
Southern Red Oak
Blackjack Oak
California
Coast Live Oak
Florida
Laurel Oak
Water Oak
Turkey Oak
Live Oak
Blue Jack Oak
Myrtle Oak
Willow Oak
10
3
7
18
7
1
1
53
-------
Table 2.4-e ESTIMATED EMISSION FACTORS FOR BROAD VEGETATION
CLASSIFICATIONS (STANDARIZED TO 30°)
Classification (Emission rate units)
Conifers (yg/g-hr)
Oaks (yg/g-hr)
Non-Conifer, Non-Isoprene ( yg/g -hr )
Non-Oak, Isoprene (yg/g-hr)
Leaf-Litter, pasture (yg/nr-hr.)
*Marine (ug/nr'hr.)
*Aquatic (yg/nr'hr.)
Day
8.9
24.7
4.3
10.3
162
129
102
Night
8.9
4.7
4.3
2.4
162
129
102
Winter
3.5
0
0
0
0
0
0
*Not included in U.S. Inventory
54
-------
Also, as Table 2.4-e shows, the emissions from soil leaf litter and
from all deciduous vegetation types were assumed to be zero during the
winter months. No field samples have been collected to confirm this
assumption.
2.5 EXAMPLE NATIONWIDE EMISSION INVENTORY
2.5.1 Purpose
Although the biogenic emission rate data gathered to date are limited,
it is useful to apply this data to the construction of a nation wide
emission inventory. The example emission inventory serves the following
purposes:
1. It illustrates the procedure involved in preparing a biogenic
emission inventory.
2. It provides a raw data base subject to further refinement.
3. It illustrates the general magnitude of biogenic VOC emissions
within the current limitations of state-of-the-art technology
and data.
Caution should be exercised when extrapolating the emission rates
determined in the study to other areas of the country. Unique species
composition, environmental influences, and meteorological conditions may
result in local or regional biogenic emissions that vary substantially
from estimates made in this report.
Of the approximately 780 samples collected, 630 were collected between
April and mid-August in the Tampa/St. Petersburg area of Florida. Twenty-
five samples were collected near Santa Barbara, California in September,
and approximately 100 samples were collected near Pullman, Washington
during the fall months of 1976 and 1977.
55
-------
Table 2.5-a gives the leaf biomass density factor for each of the
broad vegetation emission rate classifications in each bioine type (Figure
2.2-a). These leaf biomass density factors were determined from literature
values for each biome type (Leith and Whittaker, 1975; NAS, 1975; AAAS, 1967).
The approximate composition of each biome was then estimated (Rasmussen, 1972;
Dasmann, 1976; Dix and Beidleman, 1969; Franklin and Dyrness, 1969; French, 1971;
Preston, 1948; Society of American Foresters, 1954). The biome biomass density
factor was multiplied by its relative percent composition to get the leaf biomass
density factors for each broad vegetation classification in each biome.
2.5.2 Seasonal Variability
Since it is known that season, temperature and daylight affect emission
rates, it is important to consider these factors when preparing an emission
inventory. Upon examining monthly average weather maps in A Climatic Atlas
of the U.S. (U.S. Dept. of Commerce, 1978) and noting the regularity that
temperature isopleths corresponded with latitude, the U.S. was divided into
four latitudinal regions as follows: Region I = 45° - 50°; II = 40° - 45°;
HI = 35° _ 40°, IV = 25° - 35°. These are shown on the map in Figure 2.2-a.
The monthly maximum, minimum and average temperature for each region were
then estimated visually from monthly minimum, maximum and average tempera-
ture isopleths. In addition, based upon the climate information available
in the atlas (including temperatures, solar incidence, number of frost
free days, etc.), the month was designated as a summer month or a winter
month for each region (Table 2.5-b). It should be emphasized that these
are only very rough approximations of average temperature conditions and
that site specific information, especially in mountainous terrain might
vary significantly from the figures reported here.
56
-------
Table 2.5-a. BIOME EMISSION FACTORS
(Standardized to 30°C)
Grassland
Conifer
Oaks
!NC-I
2NO-I
3LL
Sclerophyll
Conifer
Oaks
NC-NI
NO-I
LL
1 Non-Coni
2 Non-Oak,
3 I _ ... p I ,' 4.
Leaf Biomass Density
(g/m2)
5
2.5
3.75
3.75
-.« — _
250
Scrub
15
30
210
45
—
300
fer, Non-Isoprene Emitters
Isoprene Emitters
Emi
Day
yg/nr hr.
44.5
61.75
16.13
38.6
162
32.98
133.5
141
903
463
162
2402.50
ssion Factor (ER)
Night
ug/nr hr.
44.5
11.71
16.13
9.00
162
243.38
133.5
141
903
24.72
162
1364.22
Winter
ug/m hr.
17.5
0
0
0
0
17.5
52.5
0
0
0
0
52.5
57
-------
Table 2.5-a. (continued). BIOME EMISSION FACTORS
(Standardized to 30°C)
" "
Emission Rate (ER)
Leaf
Temperate Rain Forest
Conifer
Oak
NC-NI
NO-I
LL
Deciduous Forest
Confier
Oaks
NC-NI
NO-I
LL
Coniferous Forest
Conifer
Oak
NC-NI
NC-I
LL
Biomass Density
(g/m2)
990
55
22
22
_ _ ..
1100
135
180
90
45
— _ _ —
450
559
39
26
26
«K — _
650
Day
p
ug/m .hr.
8811
1385
94.6
226.6
162
10679.20
1201.5
4446.0
387
463.5
162
6660
4975.10
963.30
111.80
267.80
162
6480.00
Night
p
pg/m-hr.
8811
258.50
94.6
52.8
162
9378.9
1201.5
846.0
387
108
162
2704.5
4975.10
183.30
111.80
62.40
162
5494.60
Winter
pg/m2-hr
3465
0
0
0
0
3465
473
0
0
0
0
473
1957
0
0
0
1957
58
-------
Table 2.5-a. (continued). BIOME EMISSION FACTORS
(Standardized to 30°C)
Leaf
Desert
Conifer
Oaks
NC-NI
NC-I
LL
Tundra, Alpine
Conifer
Oaks
NC-NI
NC-I
LL
Emission Rate (ER)
Biomass
(g/m2)
25
25
40
10
100
Fields
18
0
9
9
_-__
180
Density Day
Mg/nr hr.
222.5
617.5
172.0
103.0
162
1277
160.2
0
38.7
92.7
162
453.6
Night
yg/nr hr.
222.5
117.5
172
24
162
698
160.2
0
38.7
21.60
162
384.3
Winter
ug/nr hr.
88
0
0
0
0
88
63
0
0
0
0
63
59
-------
2.5.3 Elements of Annual Emission Inventory
The example annual U.S. emission inventory compiled in this study is
based on the following data:
1. A set of emission rates for broad categories of vegetation, leaf
litter, and water surfaces. Each category has an emission factor
for day, night, and winter. Since isoprene is not emitted in the
dark, the difference between the day and the night emission factor
is equal to the isoprene emissions. The winter emission factors
assume that only conifers have emissions for winter months, and that
their emissions approximate the winter emission rates of Pullman, WA
conifers. All vegetation emission rates were standardized to 30°C
using the terpene and isoprene emission rate algorithims shown in
Figure 2-3-a.
2. The approximate areas and leaf biomass density factors for biotic
regions in the U.S. were apportioned among the broad vegetation class-
ifications, according to the abundance of each classification in each
biome. Biome emission factors for each biotic region were then calcu-
lated by multiplying emission rates by leaf biomass density factors
(Table 2.5-a). For "flat samples", emission factors were related
directly in yg/nr-hr. Emission factors for "marine" and "fresh water"
categories were not used in this inventory. The addition of these
factors would not significantly affect the inventory results.
3. Monthly minimum, maximum and average temperatures by latitudinal
region were estimated for the U.S. by visually evaluating temperature
isopleths. Each month was also designated as a summer or a winter
month for each region based upon the temperature record. Generally
if the daily minimum fell below 40°F and the average was below 50°F,
the month was classified as a "winter" month. However, May was
classified as a "summer" month in Region I on the basis of the solar
incidence and growing season data in the Atlas.
To complete the emission inventory the biome composition of each lat-
itudinal temperature region was estimated and hourly emission factors (in
terms of terpenes and isoprene for summer and winter) were estimated for
each region. Next, for each summer month the standardized emission fac-
tors were adjusted to the actual regional monthly temperature using the
following scheme:
1. For isoprene emissions (only during "summer" months), the Tingey,
et al., (1978a) isoprene emission rate algorithim was used to correct
the emission factors from 30°C to the average monthly temperature
assuming that each isoprene emission day consisted of four hours of
maximum temperature and eight hours of average temperature. This
daily isoprene emission rate was then multiplied by the number of
days in the month. Q
-------
Table 2.5-b. *AVERAGE MONTHLY U.S. TEMPERATURES BY REGION (°F)
January
Region
I
II
III
IV
Class Max.
W1
W
w
w
„
30
35
60
Min.
0
16
20
40
Ave.
22
20
30
50
Region
I
II
III
IV
July
Class Max. Min.
S
S
S
S
February
I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
w
w
w
s2
March
W
W
W
S
April
W
W
S
S
May
S
S
S
S
June
S
S
S
S
35
35
45
65
35
45
55
70
50
55
65
75
60
70
75
85
70
80
85
90
10
15
20
45
20
25
30
50
30
35
40
55
40
45
55
60
50
55
60
65
20
25
35
55
25
35
45
60
40
45
55
70
50
60
65
75
60
65
75
80
I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
S
S
S
S
S
S
S
S
W
s
s
s
w
w
w
w
w
VJ
w
w
80 50
80 60
90 60
90 70
August
80 50
80 60
85 70
95 75
September
65 45
70 50
80 55
85 65
October
55 30
60 35
70 45
90 60
November
40 20
45 25
55 30
70 50
December
30 10
35 15
45 25
65 40
Ave.
65
70
75
80
65
65
75
80
55
60
70
80
45
50
60
70
30
35
40
60
20
25
35
50
*Estimated from A Climatic Atlas of the U.S. (U.S. Dept. Comm., 1968)
W = Winter
S = Summer
61
-------
2. For terpene emissions, the terpene emission rates were corrected
from 30°C to the regional monthly temperature using the Tingey,
et al., (1978b) terpene emission rate algorithim. A terpene
emission day was assumed to consist of four hours of maximum, 16
hours of average, and four hours of minimum temperature. The
daily emission rates were then muliplied by the number of days
in the month. If the month was designated as a "winter month"
the winter terpene emission factor was used. Where temperatures
below 25°C were encountered, it was assumed that the Tingey
emission rate algorithms would still be valid.
The monthly emission estimates by latitudinal region (See Figure
2.2-a) are shown in Table 2.5-c. Annual emission estimates by region
are summarized in Table 2.5-d.
From these tables it appears that natural emisions are not evenly
distributed by region or by season. Approximately 43% of the total
emissions occur during the summer months of June, July and August, and 45%
of the annual emissions occur in the Southern United States. Of the
emission in the Southern U.S., 34% is isoprene.
It is important to note that this emission inventory can only be
considered a very rough estimate of U.S. biogenic emissions, for the
data are too few to enable completion of a comprehensive nation wide
annual inventory.
2.5.4 Limitations of Emission Estimates
There are a number of areas where more data is needed:
1. It is presently not known how the emission rate from a single
tree varies over a long period. Only limited, repetitive sampling
of specific trees has been done and has never continued throughout
a year.
2. No samples have been collected from northern forests during the
spring and summer. It is not known if summer emission estimates
for northern species ever approach those of southern species.
3. The EPA-Corvallis studies were conducted over a small tempera-
ture range. Extrapolation of these emission algorithms to lower
temperatures may not be valid. Emission rate algorithms probably
only apply within the temperature range for which the plant was
adapted. That is, cold adapted plants would probably have higher
62
-------
Table 2.5-c. MOIITillY U.S. EMISSION IflVINTOKY (i;-.,.-tric to;is/iiio)
Isoprene
Jan. I
II
III
IV
Total
Mar I
II
III
IV
Total
May I
II
III
IV
Total
July I
II
III
IV
Total
Sept I
II
III
IV
Total
Nov I
II
III
IV
Total
0
0
0
0
0
0
0
0
2.2 x
2.2 x
1.3 x
8.1 x
3.3 x
1.1 x
1.5 x
6.0 x
3.5 x
1.4 x
1.9 x
3.7 x
1.7 x
5.0 x
5.4 x
1.4 x
2.0 x
0
0
0
0
0
Annual
105
IO5
IO4
104
10*
106
106
104
105
106
106
106
104
104
105
106
106
Total
Tcrpcnos
3.7 x 104
8.1 x 104
1.0 x 105
1.3 x 105
3.5 x 105
6.8 x 104
1.4 x 105
1.9 x 105
4.6 x 105
8.6 x 105
5.6 x 105
1.3 x 106
1.9 x 106
2.1 x 106
5.9 x 106
1.1 x 106
2.1 x 106
3.0 x 106
5.2 x 106
1.1 x 107
6.5 x 105
9.0 x 105
2.3 x 106
2.4 x 106
6.3 x 106
7.8 x 104
1.4 x 104
1.5 x 105
1.4 x 105
3.8 x 105
1.5 x
Total
3.7 x
8.1 x
1.0 x
1.3 x
3.5 x
6.8 x
1.4 x
1.9 x
6.8 x
10.8 x
5.7 x
1.4 x
2.3 x
3.2 x
7.4 x
1.1 x
2.4 x
1.4 x
7.1 x
1.5 x
6.7 x
9.5 x
2.8 x
3.8 x
8.3 x
7.8 x
1.4 x
1.5 x
1.4 x
3.8 x
107
10*
IO4
10*
105
105
IO5
IO5
10*
105
105
105
106
106
106
106
106
IO6
106
106
107
105
105
106
10°
io6
104
IO4
IO5
IO5
IO5
1.5
Teb I
II
III
IV
Total
Apr I
II
III
IV
Total
Juno I
II
III
IV
Total
Aug I
II
III
IV
Total
Oct I
II
III
IV
Dec I
II
111
IV
Total
x IO7
Isopronc Torpcncs
0
0
0
0
0
0
0
4.5 x 103
4.3 x IO5
4.3 x IO5
2.4 x IO4
2.0 x IO5
6.9 x IO5
1.8 x IO6
2.7 x IO6
6.0 x 104
2.0 x IO5
9.9 x IO5
2.4 x IO6
3.7 x 106
1.1 x IO4
4.4 x IO4
2.0 x IO5
6.7 x IO5
9.3 x IO5
0
0
0
0
0
6.5 x IO7
4.9 x
1.2 x
1.1 x
1.5 x
4.3 x
1.2 x
2.1 x
1.2 x
1.6 x
3.1 x
7.9 x
1.7 x
2.8 x
2.5 x
7.8 x
1.0 x
1.8 x
3.0 x
2.8 x
8.6 x
1.4 x
8.8 x
1.6 x
1.7 x
4.3 x
5.1 x
1.4 x
1.2 x
1.4 x
4.5 x
IO4
IO5
IO5
1U5
11)5
IO5
IO5
IO6
IO6
IO6
105
IO6
IO6
IO6
IO5
IO6
IO6
IO6
106
IO6
105
IO5
IO6
IO6
IO6
IO4
IO5
IO5
IO5
IO5
4.9
1.2
1.1
1.5
4.3
1.2
2.1
1.2
2.0
3.5
8.1
1.9
3.5
4.3
1.1
1.1
2.0
4.0
5.2
1.2
1.5
9.2
1.8
2.4
5.2
5.1
1.4
1.2
1.4
4.5
Total
x TO4
x IO5
x IO5
x 1U5
x IO5
x IO5
x IO5
x IO6
x IO6
x IO6
x IO5
x IO6
x IO6
x IO6
x IO7
x IO6
x IO6
x IO6
x IO6
x IO7
x IO5
x IO5
x IO6
x IO6
x IO6
x IO4
x IO5
x IO5
x IO5
x IO5
(7.2 xlO7 short tons)
*Regions outlined in Figure 2.2-9.
63
-------
Table 2.5-d. ANNUAL U.S. EMISSION INVENTORY BY LATITUDINAL REGION*
(Metric Tons)
I
II
III
IV
Annual Total
Isoprene
1.9 x 105
9.3 x 105
4.2 x 106
9.9 x 106
1.5 x 107
Terpene
4.5 x 106
9.4 x 106
1.7 x 107
1.9 x 107
5.0 x 107
Total
4.8 x 106
1.0 x 107
2.1 x 107
2.9 x 107
6.5 x 107
**1.7 x 107 **5.5 x 107 **7.2 x 107
* Region oulined in Figure 2.2-a.
** Short tons = metric tons x 1.102
64
-------
emission rates at lower temperatures than predicted by the algorithms
for Florida species.
4. More sampling should be completed to determine the emission rate
variability with vertical profile for a single tree. Present emission
estimates are based upon extrapolation of emission estimates from
branches.
5. Emission estimates for deciduous species were assumed to be zero
during winter months. Sampling should be done to confirm this.
6. Sample programs should be conducted throughout the country to
see if emission estimates corroborate those reported here.
7. The relationship between leaf temperature, bag temperature and
ambient air temperature and emission rates should be investigated
further. For this inventory it was assumed that the emission rate as
measured at a specific bag air temperature was equal to the emission
rate at an equivalent air temperature and that bag air temperature
was equal to leaf temperature.
8. The leaf biomass data used in this report is very generalized.
More specific data about leaf biomass composition, temporal leaf
biomass fluctuation due to annual climate variables such as drought or
flood and the variation in leaf biomass throughout the season for each
vegetation type are needed.
2.5.5 Summary of Inventory Procedure
Because of limitation noted above, the emission inventory reported
here can only be considered as a "first cut" approximation of actual
biogenic emissions in the U.S. However, the procedure illustrated by
this inventory could be applied to smaller areas where more specific
data could be determined. For this reason it is useful to summarize
the inventory procedure used:
1. The U.S. was divided into biotic regions.
2. The total leaf biomass for each biotic region was estimated.
3. The compostion of each biotic region, in terms of leaf biomass
for each of the emission rate categories of Table 2.4-d,was estimated.
T-he results are emission rates for the major biotic regions of the
U.S. in terms of yg emission/m • hr (Table 2.5-a).
4. The U.S. was divided into latitudinal regions as follows:
Region I = 45° - 50°; Region II = 40°- 45°; Region III = 35°-
40°; Region IV = 25° - 35°.
65
-------
5. The average, the minimum, and the maximum latitudinal tempera-
tures were estimated by month for each latitudinal region. The
month was then classified as a winter or a summer month (Table
2.5-b).
6. The area of each latitudinal region and its respective biorne
composition was estimated.
7. The hourly emission rates for each biome were converted into
monthly emission rates and into emission rates by latitudinal
region using the appropriate emission rate algorithms for isoprene
and terpenes.
8. For isoprene emissions, each day was assumed to consist of
four hours maximum, and eight hours of average temperature for
each summer month.
9. For non-isoprene emissions, each day was assumed to consist
of four hours of maximum, four hours of minimum, and 16 hours
of average temperature.
If a detailed inventory were desired for a specific area, similar
steps would be taken:
1. Determine the major vegetation types and quantify the leaf
biomass for each.
2. Develop a temperature and light regime which would describe
the variability in emission rates with sufficient detail for the
time period of interest.
3. Develop species specific emission factors based upon field
sampling or use the broad vegetation emission factors given in
Table 2.4-e.
4. Standardize the vegetation emission rates with respect to
temperature, light and species composition and multiply the
emission rates by their respective leaf biomass factors. Add
the soil-leaf litter and water surface emission factors to the
vegetation emission factors. The sum of the components is equal
to an hourly emission factor for the entire area.
5. Modify the hourly area emission factor using the appropriate
time scale for temperature and light conditions to arrive at
total daily, seasonal or annual actual VOC emission.
2.6. Comparison With Other Biogem'c Emission Estimates
Tingey et. al. (1978 a,b) reported an average emission rate from
Slash Pine of 9.38 ug carbon/g-hr (10.63 yg compound/g«hr) at a leaf
66
-------
temperature of 35°C. Their reported isoprene emission rates for live
oak reached a maximum of 120 yg carbon/g dry weight at a leaf temperature
of 44°C (136 yg compound/g-hr). At a leaf temperature of 30°C and
saturating light the predicted emission rates ranged from 28.35 to 46.1
yg/g-hr. (32.14 to 52.27 yg/ compound/g-hr). These laboratory results
obtained under laboratory conditions are fairly close to the emission
rates of 8.9 yg/g-hr for conifers and 24.7 yg/g-hr for oaks reported in
Table 2.4-e of this report. It should be expected that laboratory studies
might yield slightly higher emission rates since the plants are grown
under ideal conditions.
Arnts, et. al., described a micrometeorological technique to estimate
the emission flux of a-pinene over an even aged loblolly pine plantation.
These estimates were made by comparing vertical profile measurements of
a-pinene with net radiation and vertical gradients of water vapor using an
energy balance approach. Their flux estimates ranged from 13 to 119
yg/nr min"^ with a mean value of 52.8 yg/m^min"^ at an average temperature
of 29.7°C (Arnts, et al., 1978).
If the leaf biomass for the loblolly pine plantation is 750 g/m2
(Arnts personal communication), then this results in an emission rate of
from 1.0 to 9.5 yg/g-hr of a-pinene. If a-pinene is roughly 35% of the
non-methane VOC (Table 2.4-b), then these figures are equal to an emission
rate range of 1.7 to 16 yg/g-hr with an average emission rate of 6.9 yg/g-hr
for loblolly pine.
The results of the studies by Tingey, et. al., and Arnts, et. al., sug-
gest a good agreement betwen different sampling, methodologies, analytical
techniques and vegetation species and also corroborate some of the emission
rates used in this report.
67
-------
Information available in the literature concerning the area-wide
emission rates of natural hydrocarbons is very scarce. Most estimations
have been based upon early work by Went (1960) and Rasmussen and Went
(1965). Westberg (1977) points out that these previous emission rates
are inadequate for they deal only with terpenes and their derivatives.
Also, in general, it is difficult to extrapolate to national or regional
biogenic emission rates from these worldwide estimates.
In order to compare the average yearly U.S. emission rate calculated
in this study to the previous worldwide emissions rates, a common basis
had to be found. Numerous estimates of the net amount of carbon fixed by
the terrestrial vegetation of the world (net primary productivity) have
been made in the course of IBP studies.. These estimates, included in the
The Primary Productivity of the Biosphere (Lieth and Whittaker, 1976)
are based upon computer modeling maps which relate the Net Primary Produc-
tivity (NPP), measured at different points around the world, to climato-
logical factors. The relationship between NNP and climate is then used
to construct NPP isopleths. The areas of the isopleths drawn on computer
generated maps can then be summed to give regional or worldwide estimates
of NPP. If NPP is related to the average VOC emission rate for the U.S.,
and this factor (the ratio of the weight of emissions to the weight of carbon
fixed) is assumed to be representative of the worldwide emission rate/NPP
ratio, the annual worldwide hydrocarbon burden can be estimated. Using
these assumptions an estimated total non-methane VOC emission of 8.3 x
108 MT/yr can be calculated based upon an ER/NPP ratio for the U.S. of
approximately 0.7%. This annual worldwide estimate is roughly 10 times
Rasmussen's and Went's previous estimates which included terpene compounds
only, and is near the lower limit of the 1974 RTI estimate (Table 2.6-a).
68
-------
Table 2.6-a. ESTIMATES OF WORLDWIDE EMISSIONS OF NATURAL VOC *
Investigator
Went (1960)
Rasmussen & Went (1965)
Ripperton, White & Jeffries (1967)
Robinson & Robbins (1968)
RTI (1974)
CTl
Zimmerman (1978)c
Method
Sum of sagebrush emission
and terpenes as percentage
of plant tissues
Direct _ir^ situ ambient cone.
Reaction Rae 03/pinene
Based largely on Rasmussen
and Went's data
Lower limit = 10 x Rasmussen
and Went's figure
Upper limit = .10 x total
carbon fixed by green plants
Annual US Natural Emissions
Annual US NPP
w ,,
Emission Rate (tons/yr)
1.75 x 108a
4.32 x 108a
2 to 10 x previous
estimates3
*Adapted from Westberg 1977
aBiogenic hydrocarbn emissions, i.e. terpenoid vapors
^Biogenic organic emissions
cThis study (extraplating nationwide estimated to the world)
4.80 x 10
,8b
0.4 - 1.2 x 10
8.3 x 108
lOb
-------
REFERENCES
Alexander, J. T., "Emission Inventory of Petroleum Storage and Handling
Losses" in: Emission Inventory/Factor Workshop, Vol II. EPA Office of
Air Quality Planning and Standards, 1977.
American Association for the Advancement of Science, Primary Productivity
and Mineral Cycling In Natural Ecosystems. A Symposium. 13th Annual Meeting,
New York City, December 27, 1967.
Arnts, R. R., Seila, R. L., Kuntz, R. L., Mowry, F. L., Knoerr, K. R. and
Dudgeon, A. C., "Measurements of a-Pinene Fluxes from a Loblolly Pine
Forest." Fourth Joint Conference on Sensing of Environmental Pollutants,
American Chemical Society, Washington, D.C., pp. 829-833, March 1978.
Dasmann, R. F., Environmental Conservation (4th Ed.), Int. Union for the
Conservation of Nature and Natural Resources, Merges, Switzerland. 1976.
Duce, R. A., "Speculations on the Budget of Particulate and Vapor Phase
Non-Methane Organic Carbon in the Global Troposphere," Phageoph, Vol.
116, 1978.
Gates, D. M., "Heat Transfer in Plants," Sci. Am. 213:76-84, 1965.
Holdren, M., "Letter Report to Members of the CRC-APRAC, CAPA-II Project
Group," October 17, 1978.
Kamiyama, K., T. Takai, and Y. Yamanaka, "Correlation Between Volatile Substances
Released from Plants and Meteorological Conditions" in: Proceedings of the
International Clean Air Conference, Brisbane Australia, 1978.
Lieth, H., and R. H. Whittaker, (eds.), Primary Productivity of the Biosphere,
Springer-Verlag, New York, 1975.
Miller, P. R., and R. M. Yoshiyama, "Self Ventilated Chambers for Identification
of Oxidant Damage to Vegetation at Remote Sites." Environmental Science and
Technology, 62490, Kb 10, Jan. 1973.
National Academy of Science, Productivity of World Ecosystems. Proceedings of a
Symposium presented Aug. 31 - Sept. 1, 1972 in Seattle, WA. Wash. D.C., 1975.
Rasmussen, R. A., "Isoprene: Identified as a Forest-Type Emission to the
Atmosphere." ES & T, Vol. 4, 8, pp. 667-671, 1970.
Rasmussen, R. A., and Went, F. W., "Volatile Organic Material of Plant
Origin in the Atmosphere," Proc. N.A.S., vol. 53, pp. 215-220, 1965.
Rasmussen, R. A., "What do the Hydrocarbons from Trees Contribute to Air
Pollution?", J. of Air Pollution Control Association, Vol. 22, No. 7,
pp. 537-543, 1972.
Research Triangle Institute, "Natural Emissions of Gaseous Organic Compounds
and Oxides of Nitrogen in Ohio and Surrounding States," Final Report, EPA
Contract 68-02-1096, 1974.
70
-------
Ripperton, L. A., White, 0., and Jefferies, H. E., "Gas Phase Ozone-Pinene
Reactions," Div. of Water, Air, and Waste Chemistry, 147th National Meeting
American Chemical Society, Chicago, IL., pp. 54-56, September, 1967.
Robinson, E., and R. C. Robbins, "Sources, Abundance, and Fate of Gaseous
Atmospheric Pollutants," SRI Final Report, PR-6756, 1968.
Rodin, L. E., N. I. Bazilevich, N. N. Rozov, "Productivity of the Worlds
Main Ecosystems," in: Productivity of World Ecosystems, N.A.S., 1975.
Sanadze, G. A., Kalandadze, A. N., "Light and Temperature Curves of the
Evolution of CcHg," Fiziologiya Rastenii, Vol. 13, No. 3, ppb 458-461,
May - June, 1966.
Sanadze, G. A., "Absorbion of Molecular Hydrogen by Illuminated Leaves,"
Fiziologiya Rastenii, Vol. 8, No. 5, pp. 555-559, 1961.
Satoo, T., "Primary Production Relations in Woodlands of Pinus Densiflora,"
in: Primary Productivity and Mineral Cycling in Terrestrial Ecosystems.
Symposium, Ecological Society of America, American Association for the
Advancement of Science, 13th Annual Meeting, Dec. 27, 1967, New York City.
Tingey, D. T., H. C. Ratsch, M. Manning, L. C. Grothaus, W. F. Burns, and
E. W. Peterson, "Isoprene Emission Rates from Live Oak," Final Report, EPA
CERL-040, May, 1978.
Tingey, D. T., M. Manning, H. C. Ratsch, W. F. Burns, L. C. Grothaus, and
R. W. Field, "Monoterpene Emission Rates from Slash Pine," Final Report,
EPA CERL-045, August, 1978.
U.S. Dept. of Commerce: A Climatic Atlas of the U.S., 1968.
Went, F. W., "Blue Haze in the Atmosphere," Nature, pp. 641-643, August, 1960,
Westberg, H. H., "The Issue of Natural Organic Emission" in: International
Conference on Oxidants 1976 Analysis of Evidence and Viewpoints, EPA-600/3-
77-116, October 1977.
71
-------
APPENDIX A
EXPERIMENTAL VERIFICATION OF SAMPLE METHODOLOGY
A.I. INTRODUCTION
This appendix is intended to explain the rationale for the development
of the sampling methodology outlined in the main report. It details the
evolution of the field testing methodology and includes experimental
verification procedures, and "dead ends." It is designed to aid researchers
who wish to refine the sampling or analytical techniques used in this
study. The explanation of WSU's technique development should also
minimize experimental back-tracking.
A-l
-------
A.2. PROJECT OBJECTIVES
The objectives of this study were as follows:
1. To develop a standardized sampling and analytical methodology
2. To -develop emission factors for a limited number of species
3. To develop standardized emission inventory techniques
4. To develop an example nation wide emission inventory of
biogenic emissions
A standardized sampling and analytical methodology was desired so that
emission estimates could be" easily generated by different laboratories and
could be directly compared. Thus, the techniques developed were to be
fairly simple and were to use non-specialized, commercially available equip-
ment where possible. It was desirable that the experimental procedures
allow for alternate funding levels and equipment availability which might
occur among laboratories.
The development of emission factors for a limited number of vegetation
species, leaf litter and water surfaces was meant to serve the dual
purposes of testing the methodology and of generating a data base which
could be used in preliminary emission inventories. The data generated
would also provide a historical comparison for past and future emission
rate data.
The emission factors developed were to be applied to standardized
emission inventory procedures. A nationwide emission inventory using the
A-2
-------
standardized methodology would then serve to demonstrate the procedures in-
volved. Thus a nationwide inventory was to be developed in a format so
that the data could be easily applied as a preliminary estimate for a
specific area, or could be easily refined as additional locale-specific
data was generated.
A-3
-------
A.3. SAMPLING APPROACHES
There are a number of possible sampling approaches which might be ap-
plied to fulfill the project objective. The following were considered
to be the most likely alternatives for this project:
A.3.1. UPWIND/DOWNWIND OR VERTICAL GRADIENT SAMPLING
The upwind/downwind method involves the collection of a sample up-
wind of the vegetative area of interest and simultaneously collecting a
sample downwind of the area. The difference in total non-methane organic
emissions between the two samples is then assumed to be due to the vegeta-
tion effects. If desirable, the biomass unit of interest (leaf, twig,
leaf and branch; or surface area) for the vegetation stand can be calcu-
lated and an emission rate can be estimated.
The upwind/downwind procedure requires a very carefully selected site.
The air entering the vegetation must be relatively clean. It is also im-
portant that the incoming air be uniform with respect to composition and
concentration.
The advantages of this sampling methodology are that the sampling pro-
cedure itself is "outside the system" and thus does not affect emission
rates. Ths methodology also results in a net emission flux from the
vegetation, so sources, as well as sinks, are automatically included.
The upwind/downwind approach has several profound disadvantages.
First, the logistics of operating two simultaneous sampling sites presents
A-4
-------
problems. These problems include sample timing, instrument calibration,
personnel limitations and site selection.
Secondly, it is nearly impossible to find an ideal site. Most sites
represent compromises in that they are affected by the wind shifts due
to local terrain effects or weather patterns, mixing of the incoming air,
and the vegetation emissions are never uniform. Moreover, anthropogenic
emissions invariably impact on the vegetation study area. The end result
is that a.large number of samples are required so that trends can be
determined. This procedure has the additional disadvantage in that
biogenic hydrocarbon emissions reacting between the sampling points are
not accounted for. The upwind/downwind scheme, therefore, invariably
requires complicated and highly speculative modeling to estimate emission
rates. Additionally, rural ambient VOC levels are usually quite low
and, therefore, problems of analysis and accurate quantisation are
extensive. Also these programs are very expensive by nature. Finally,
the upwind/downwind approach can not readily isolate any of the variables
which might affect emission rates. These variables include heat, light,
and relative humidity.
The vertical gradient approach, such as that used by Arnts, et al_.,
(1978), which attempts to establish a vertical profile of selected hydro-
carbons within a forested canopy, is essentially a variation of the upwind/
downwind approach and shares similar advantages and disadvantages.
A.3.2 ENCLOSURE SAMPLING
An alternative to upwind/downwind and vertical gradient sampling is
enclosure sampling. The basic method involves placement of an enclosure
chamber around the selected vegetation to isolate it from some of the
variables of the surrounding environment.
A-5
-------
This technique has several inherent advantages:
1. Higher concentrations of emissions can usually be obtained, which
allows one to use simpler and more accurate analytical techiques.
2. No model is needed to determine source strengths. Source emis-
sions are directly measured.
3. Better control can be attained over variables which might
affect emission rates. If these variables are better known it is
easier to accurately extrapolate emission rates spatially and
temporally.
4. The sample site selection is not so critical. Usually this ap-
proach allows emission testing even in areas of high anthropogenic
concentrations.
5. As a result of the above advantages, the sampling programs
tend to be less expensive and more portable.
There are a number of enclosure techniques which have been used previ-
ously for vegetation emission/uptake studies and were considered for
use in this project.
A.3.2.1 Regulated Enclosures
The most sophisticated type of enclosures can be classified as regu-
lated enclosures. This type of enclosure has often been used to evaluate
seasonal fluctuations of C02 emissions from a specific portion of a branch.
The regulated chambers provide control of temperature, relative humidity,
and often even light. They are designed to maintain long-term environmental
equilibrium and control over the portion of vegetation or whole plant of
interest. This type of chamber has inherent disadvantages for use in the
measurement of hydrocarbon emissions from vegetation:
A-6
-------
1. The most common materials of construction include plexiglass,
neoprene gaskets, and other components which may emit VOC.
2. Long-term equilibrium invariably requires the circulation of air
through a conditoning system which cools and controls the humdity of
the environment. VOC emissions could conceivably be lost in such a
system. An experiment conducted in our laboratory to try to remove
some of the moisture from an air sample illustrated this point. When
the sample was passed through a 3 foot length of coiled 1/16 inch
(o.d.) stainless-steel tubing cooled to 0°C or lower, all of the hy-
drocarbons above ^^ were lost.
3. Regulated enclosures are usually meant to be semi-permanent,
therefore, a wide variety of species could not be sampled in an
emission inventory program.
4. These enclosures require external 110-120 volt power or elaborate
battery systems which limit their portability and also put a con-
straint on possible sample sites.
A.3.2.2. Compensated Chambers
Compensated chambers are a class of enclosures which attempt to com-
pensate for the inescapable enclosure effects of temperature and moisture
buildup. The open-top chambers used for the study of ambient pollutant
effects on vegetation are an example. This type of chamber is often quite
large and encloses a whole tree. It relies upon a fast exchange rate (up
to eight exchanges per minute) of the air inside the chamber to minimize
heat and metabolic by-product buildup. Usually the chamber has a filter
on the inlet and a collecting device on the outlet. Often the tops of
the chambers are painted white to minimize heating. They are usually
provided with self ventilating wind powered stacks to enhance air flow and
to minimize the need for electric power. However, the temperature within
the chamber may reach temperatures above 44°C ( 6°C above ambient air
temperature) even at an air exchange rate of 3.5 times per minute (Miller
and Yoshiyaroa, 1973).
A-7
-------
The disadvantages of these types of chambers for the measurement of
VOC emissions from vegetation include:
1. The high flow rates required result in low sample concentrations
which, in turn, necessitate a high degree of analytical sensitivity.
2. The filtering and sample collection techniques required would be
quite complicated if it was desirable to quantify the full range of
organic emissions.
3. Compensated chambers typically are semi-portable at best and are
meant for long-term installation.
4. The cost of this type of approach would be very high for a wide-
spread, short-term project.
A.3.2.3 Static Chambers
The simplest and least expensive enclosure approach is the static
chamber approach such as that used by Rasmussen (1972) and Sanadze (1961).
In this method a portion of vegetation is enclosed for a specified length
of time and a sample is collected from the chamber.
This method has the advantages of being relatively inexpensive and
portable. In addition, high VOC levels can be generated that enhance
analysis and quantisation.
The static chamber method has the following disadvantages:
1. There is no control over the buildup of heat, water vapor, C02 or
other metabolic products, which can affect emission rates.
2. If the enclosure is rigid, anthropogenic VOC emissions will
"invade" the chamber as the sample is withdrawn. Also, the chamber
is full of ambient air at the time of enclosure. If anthropogenic
VOC emissions are high they may obscure vegetation emissions.
A.3.2.4 WSU Static/Dynamic Enclosure
After evaluation of the preceeding techniques and preliminary collec-
tion of samples using the static system, a static dynamic sampling procedure
was developed. Basically. the method which evolved involves enclosing a
portion of the vegetation of interest with a Teflon bag, sucking most of
A-8
-------
the ambient air out of the bag and thus collapsing the bag around the
branch, and collecting a sample of the air from the collapsed bag. The bag
is then quickly filled with a known volume of air which is free of all VOC
and has a controlled level of CC^. After the bag is partially inflated,
another sample is collected while zero air continues to flow into the
enclosure at a rate slightly higher than the sample rate. The difference
between the hydrocarbon content of the bag, divided by the time interval
between the two samples, is the emission rate of the vegetation. This
emission rate can then be divided by a foliage unit such as biomass or sur-
face area to result in an emission rate per unit vegetation per unit time.
Total enclosure time is on the order of 15 minutes. The method has several
advantages:
1. The short enclosure time and large volume of zero-air minimizes heat
buildup and water vapor accumulation, which ensures that the emission
rates are not disturbed.
2. The ultraclean air entering the enclosure and the organic emission
concentrations which accumulate in the zero air enhance analytical sensi
tivity and precision.
3. The equipment is fully portable. A wide variety of emission samples
can be collected in a relatively short field sampling period.
4. Emission samples can be collected practically anywhere, the procedure
is not site-selective since anthropogenic contributions are minimized.
The method requires the following resources:
1. An enclosure which does not contribute to or "scrub" hydrocarbon emis-
sions.
2. A source of clean "zero air"
3. A method to take a sample from the enclosure and transport it to the
laboratory for analysis while preserving its integrity.
The static/dynamic approach has disadvantages in that:
1. A portable source of clean "zero air" may not be readily available.
A-9
-------
2. Each major vegetation-type must be sampled (a disadvantage common among
all enclosure techniques).
3. There is a possible disturbance of the vegetation emission due to the
sample collection procedure. This disadvantage is also common among any
methodology which does not remain "outside the system".
A-10
-------
A.4. ZERO AIR
The sampling methodology is dependent upon source of clean "zero"
air. The air is required to be uniform in composition and preferably
should have a total VOC concentration of 10 yg/m3 or less. Sources of
zero air which have high levels of hydrocarbons could be used; however,
care must be taken to "fingerprint" this air source so that amounts and
species of hydrocarbons present can be quantitated and subtracted from
emission rate values. This adds to the expense and complexity of perform-
ing all analyses. During the developement of the sampling procedure, WSU
evaluated a number of possible zero air sources. These included commer-
cial zero grade air from Matheson and Scott, molecular sieve filters,
hydrocarbon combusters, and two commercial brands of zero air generators:
The Meloy pure air source and Aadco zero air source.
A.4.1 COMMERCIAL ZERO AIR
Our evaluations revealed that Ng and 02 levels in commercially avail-
able cylinders were not uniform from cylinder to cylinder. It is not known
what effect this might have on vegetative emission rates, however it was
felt that an air supply of uniform composition would help to limit some of
the variables which might affect emission rates. Also, most tanks had var-
iable levels of methane and various higher hydrocarbons.
A-ll
-------
A.4.2 MOLECULAR SIEVE FILTERS
Molecular sieve filters work well in removing most higher hydrocarbons
above C2; however, breakthrough of the C2 hydrocarbons can occur quite rap-
idly.
A.4.3 HYDROCARBON COMBUSTER
The hydrocarbon combuster consists of a platimum wire coil in a glass
tube. The wire is heated until it is red hot and the air to be cleaned
has passed over it. The hydrocarbon combuster could only process air at
the rate of 1 liter/minute. However the outlet air was hydrocarbon free.
The construction of the combuster did not allow pressures to exceed 5 psig.
The outlet also had variable C02 and moisture levels.
A.4.4 MELOY PURE AIR SOURCE
The Meloy pure air source is essentially a series of molecular sieve
filters. In operation, one filter is on-line and a portion of its filtered
outlet air is used to back-flush the other filter. Periodically the flow
through the filters is switched so that a clean filter is always on line.
The Meloy we tested allowed passage of light (C2) hydrocarbons. Also, the
materials of construction emitted hydrocarbons, and concentrations less than
1 ppmC could not be obtained.
A.4.5 AADCO PURE AIR GENERATOR
The other commercially available unit, the Aadco pure air generator,
also uses switching molecular sieve filter units. However, it uses a very
short cycle time (30 sec) and two sets of molecular sieve filters. After
passing through the molecular sieve filters, the outlet air enters a cata-
lytic combuster which is maintained at MOO°C and eliminates methane and
A-12
-------
C2 hydrocarbons which may pass through the molecular sieve filters before
back-flushing. The unit can supply pressures up to 80 psig at a rate of
up to 10 liters/minute. The air leaving the combuster can be routed through
a Mallcosorb® (indicating Ascarite) scrubber to eliminate C02. The result is
air which has the the same N2/02 ratio as ambient air and a dew point of minus
60°F. WSU has concentrated up to 10 liters of this zero air using the freeze-
out loop method (previously described as a sample concentration step prior
to analysis), and have found no hydrocarbons. WSU has also checked the
output for fluorocarbons and has found none. The Aadco pure air source
was thus selected for use on the project and has been used continually
throughout the year. It has been found to be a reliable source of zero
air.
A.4.6 CRYGENIC COMPRESSION OF ZERO AIR
Since the zero air generator required an external source of 110 volt
power, WSU had to develop a method to get large quantities of zero air into
the field. The method developed consists of immersing the base of a clean,
empty medical grade oxygen cylinder in liquid nitrogen and connecting the
valve opening to the outlet of the zero-air source. Air is cooled as it
enters the cylinder and is condensed. The inlet flow rate and the cylinder
capacity are known. Enough air is placed in the cylinder so that when the
cylinder is removed from the liquid nitrogen and equilibrated to ambient
temperature, the cylinder is filled to approximately one half of its rated
capacity. Typically, cylinders are used that are rated at 1800 - 2000 psig
and which have been burst checked to 6000 psig. WSU then cryogenically com-
presses approximately 900 psig of zero air in them. Medical grade oxygen
cylinders are used because they are required to be oil-free and therefore,
usually require no cleaning. The cryogenic compressing procedure should be
A-13
-------
used only under close supervision of trained personnel using the proper
safety equipment and precautions. Alternatives to the cryogenic compres-
sion method are to put the zero air for use in the emission sampling pro-
cedure into large clean Teflon bags, or to compress the zero air output
directly into large low-pressure (80 psia) tanks.
A-14
-------
A.5 SAMPLE METHODOLOGY CHECKS
It is important that the procedure used for sampling emissions min-
imizes the disturbance of the enclosed vegetation. To insure that the
procedure meets with this criterion, the C02 content of the bag as well
as its temperature are routinely monitored. Periodic tests are also
made to monitor the relative humidity of the bag. WSU has found that if
the bag is in direct sunlight on a cool day (e.g. 10°C), the bag tempera-
tures might be up to 10°C warmer than ambient temperatures. Also, the
bag air tends to heat up rapidly if the transpiration rate of the enclosed
vegetation is low. For most samples collected to date, the enclosure
temperature is usually the same as the ambient temperature.
A.5.1. RELATIVE HUMIDITY
The zero air used in the procedure is dry (i.e., the dew point is
-65°C). However, we have monitored the relative humidity in the enclosure
and have found that it rapidly reaches a relative humidity very close
to that of ambient air at the time that the emission sample is collected.
If desired, the relative humidity of the incoming air could be adjusted
by use of a bubbler system to approximate the moisture content of ambient
air.
A.5.2 C02
C02 concentrations in the bag in sunlight were also periodically mon-
itored. WSU found that the levels of C02 in the bag in sunlight were lower
A-15
-------
than typical levels in ambient air by 5 to 10 percent. For some samples
that were collected in the dark, C02-free zero air was used. Under these
conditions, CC^ levels in the sample were approximately equal to or
slightly higher than those of ambient air at the time of sampling. No
changes in emission rates with C02 concentration have been detected to
date.
Finally, WSU's sampling procedure has been evaluated by a plant
pathologist and a plant physiologist (personal communication with Dr. M.
Pack, plant physiologist, WSU, Air Pollution and Resources, and Dr. R.
Rasmussen, Professor Environmental Technology, Oregon Graduate Center,
respectively.) They feel that due to the very short enclosure time
( 15 min.), the sampling procedure should not significantly affect the
vegetation emission rates.
A.5.3. SAMPLE INTEGRITY
Sample integrity in the stainless steel sampling canisters has been
checked by placing known standards in the cans and periodically analyzing
the contents. WSU has also collected vegetation emissions in a stainless
steel canister and periodically analysed its contents. The results confirm
that measureable amounts of hydrocarbons are not lost on the container
walls between the time of sampling and analyses. Repeated analyses over
a time period of one week do show some shift in the component peak concen-
trations of the sample for some canisters; however, the total non-methane
organic carbon content of the cans tested has remained stable to within
15% over a one-week time period. Usually samples were analysed within
24 hours of sample collection.
A-16
-------
A.6. ANALYTICAL METHODOLOGY CHECKS
A.6.1. SAMPLE ENRICHMENT
The sample enrichment procedure is explained thoroughly in the main
report. It is a procedure which has been duplicated in many laboratories.
It involves slowly putting a sample through a 1/8" stainless steel loop
that is filled with glass beads and is immersed in liquid 02. Organics
in the sample are frozen out on the glass beads. WSU has checked the
effluent from the freeze-out loop for breakthrough of C2~C^2 hydrocarbons
and has also compared peak heights obtained by directly injecting a
quantitative standard. WSU has then made dilutions of this standard and
frozen out an equivalent amount using the freeze-out technique. The
results are identical.
A.6.2. HYDROCARBON ANALYSIS
Columns and conditions are explained in the instruction manual. We
used neohexane as a quantitative standard. Qualitative standards were
prepared and the tentative identification of sample components was made
on the basis of retention time. The identification of the emissions was
confirmed using a G.C. Mass Spectrometer system. The G.C. Mass Spectrometer
also provided an independent means of comparing the quantisation of samples
analysed via FID. The results were in good agreement.
A-17
-------
A.6.2.1 Oxygenates
WSU conducted an experiment to determine the possibilty that oxygenates
may be lost in the sample train between sampling and analysis. For these
experiments a closed 1m x 1.5m Teflon bag equipped with Teflon-brass fittings
and a Teflon-backed Silicon-rubber septum was used. The bag had a capacity
of approximately 400 «,. Three hundred liters of zero air was placed in
the bag and 1 to 2 yl (liquid) each of propionaldehyde, butyraldehyde,
crontonadldehyde, valeradehyde, iso-valeraldehyde and benzaldehyde was
injected via a micro liter syringe through the septum. These aldehydes
were chosen because they were readily available and were thought to be
represent!tive of some of the oxygenated species which might be emitted
from vegetation. Another 50 liters of zero air was then added to facili-
tate mixing. Concentrations ranged from 0.5 to 2 ppm (v/v) compared for
each component. The field sampling train (pump, tubing, portable sampling
manifold) was then connected to a bag fitting and a stainless steel
canister was filled. Samples were transferred from the bag through the
sample manifold into the canister using Teflon tubing and copper tubing.
The Teflon bag was connected directly to the G.C. sample loop and sampled.
The G.C. sample loop was also filled by connecting the bag to the loop
with about 1m of 1/8" OD Teflon tubing and with 1m of 1/8" OD stainless steel
tubing. The bag and its contents were stored for 24 hours and the can
filling and sampling procedures were repeated.
The results indicated that on the first day there was an initial loss
of about 30% in the sample train regardless of whether copper and Teflon
sample lines were used. This could be accounted for if the bag had not
equilibrated properly before sampling. The experiment was then repeated
the second day after the bag contents had time to equilibrate. The
A-18
-------
concentrations measured in the bag were 50% lower. This loss could
represent better equilibration of the contents, as well as overnight
storage loss in the Teflon bag. When the bag was sampled through the
portable sample manifold, no significant differences in G.C. response
could be noted among direct bag injection, collection of the sample through
the manifold with copper tubing or with teflon tubing. When the bag con-
tents were introduced directly into the sample loop and compared with the
results obtained when the bag sample was transferred into the loop using
Teflon tubing and stainless steel tubing, the G.C. response indicated a
slight loss due to stainless steel tubing (o3%) and a greater loss due
to the Teflon tubing (a!0%).
Experiments have also been conducted by Mike Holdren (Assistant Chemist,
WSU) to test the storage efficiency of the stainless steel canisters for
aldehydes and ketones as part of a CRC-APRAC CAPA-11 project. These tests
were accomplished by adding nanogram amounts of an aldehyde-ketone mix to
canisters filled with rural ambient air. Addition of an internal standard
(neohexane) provided for better quantitative analysis. It was also felt
that the addition of rural air would represent the best sampling matrix
for these stability tests. The construction dates of the canisters chosen
for the experiment varied from four years to less than six months.
During an earlier study with the compound a-pinene, it had been discovered
that the age of the container could be important, i.e. accelerated loss
of a-pinene tended to occur in the older canisters.
During the three-week test period gas chromatographic analyses were
completed on the canisters. Table A.6.2-a indicates that the majority of
the compounds remained within 15% of the original concentration level (10-20
ppb). However, in two of the newer containers there were losses of furfural
A-19
-------
Table A.6.2-a. LONG-TERM STORAGE OF ALDEHYDES AND KETONES
IN STAINLESS STEEL CANISTERS**
Compound
Acetone
Neohexane
Butyraldehyde
Methyl -ethyl -ketone
Crotonal dehyde
Valeraldehyde
Methyl -i sobutyl -ketone
Furfural
2-Heptanone
Heptal dehyde
Benzal dehyde
2-Octanone
% Change
Can 1
+10%
0%
- 7%
+ 9%
+ 9%
- 2%
+10%
- 2%
+15%
+ 6%
+ 6%
+10%
Over Three-Week
Can 2
+13%
0%
-17%
-14%
- 6%
-31%
- 1%
-22%
-13%
-50%
-14%
-46%
Period
Can 3
+12%
0%
-27%
-19%
- 9%
-35%
+ 2%
-49%
- 7%
-76%
-13%
+24%
*From Oct. 17, 1978 letter report to CRC-APRAC CAPA-II Project Group, by
Mike Holdren, WSU, Air Pollution and Resources.
A-20
-------
and heptanal. Part of these apparent losses were later thought to be
due to a malfunction of the peak integration equipment. For this reason
the results shown in Table A.6.2-a are "worst case".
The possibility that oxygenates may be sampled but not "seen" by the
GC-FID's was examined two ways. First, a column (Durapak-low K) was made
and oxygenate standards were run to determine response factors and elution
times. Quantisations of oxygenates is difficult, for standards are diff-
icult to make because they are so "sticky." Their FID response varies
with each species so calibration using a representative hydrocarbon would
result in inaccuracies. The G.C. Mass-Spectrometer was also used to check
for the presence of aldehydes in the samples. To date no oxygenates have
been detected in the emission samples analyzed via G.C. Mass-Spectrometer.
At this time, however, WSU is just developing techniques to insure accurate
quantisation of low levels of heavy oxygenated hydrocarbons in ambient air.
These techniques are not presently routinely available to all laboratories.
A.6.2.2. Analytical Precision
The accuracy and repeatability of the analytical technique has been
thoroughly tested. We have determined that the error associated with
quantitation of the total hydrocarbons in a sample is a fixed amount rather
than a percent of the total. This fixed error is 5 - 10 ppb and represents
the error associated with the integration of very small peaks, background
bleed from the column substrate and inaccuracies in sample introduction.
To determine the error in the sample introduction technique, WSU
diluted a standard concentration of neohexane in a 5.5 liter stainless
steel can using zero air as a diluent. Then the sample was run five con-
secutive times to determine the range and standard deviation. It was
A-21
-------
subsequently determined that using the vacuum introduction system, the
range was 0.08 yg/m3 and the standard deviation was 0.04 yg/m3. The
average concentration of the neohexane standard was 1.35 yg/m3.
4.6.2.3. Analytical Problems
Water introduced into the freezeout loop because of samples which
have extremely high humidity, can cause problems when samples are analyzed
using SCOT columns. The water evidently creates a stripping action on the
column and results in broad extraneous peaks. A 10' (3.1m) durapak-low K
precolumn was added to minimize these effects.
Water also affects the ^2 - C6 analysis. Specifically, it causes
ethylene to elute as a broad unquantifiable peak. The use of a potassium
carbonate trap on the sample inlet minimizes this effect. However, if the
sample is introduced into the freeze-out loop too quickly the sample may
still not be dry enough. WSU, therefore, chose to duplicate the analysis
for the Cg hydrocarbons on a Porapak Q column. This column substrate is
not affected by samples of high relative humidity and the analysis can be
performed quickly. Comparisons of Cg hydrocarbons from the isothermal G.C.
and from the temperature-programmed light hydrocarbon G.C. was favorable
for low-humidity samples.
High concentrations of COg may interfere with the analysis of fy hydro-
carbons using the Porapak Q column. This only occurred for some of the
early soil plug experiments when a soil plug was left enclosed in a static
chamber for a number of hours. The effect is noted as a negative peak
which may interfere with ethylene quantisation.
A-22
-------
A.7. EXPERIMENTAL DEAD ENDS
A.7.1. SOIL/LEAF LITTER SAMPLES
The first procedure used by WSU to try to estimate soil/leaf litter
emissions involved cutting a plug of soil and placing it in an air tight
stainless steel bucket. The bucket was taken to the lab and pressurized.
A head space sample was then removed for analysis. The procedure resulted
in emission rates which varied widely between samples. The enclosure time
was usually 1 to 3 hours. During this time period C02 and water vapor
accumulated in the chamber which adversely affected the analytical proce-
dures. Next, WSU tried driving a cylinder (open on one end) into the ground
4 to 8 cm. A pump was put on the inlet side that pumped air through the
cylinder into a sample collection chamber. This procedure also failed
due to contamination of the sample by auto exhaust components. Also very
large emission rates for forest soils were sometimes observed due to the
cutting of needles and leaves along the edge of the sampler.
To alleviate the above problems, the stainless steel bucket (which
had been used earlier for soil plug enclosure experiments) was inverted
over the soil and its edges sealed with moist soil. The bucket was then
flushed with zero air and slowly purged at a faster rate than the sample
was removed. This procedure worked fairly well; however, difficulty
remained in maintaining a seal around the chamber base. Also, all samples
were collected in the "dark" using this method.
A-23
-------
The method WSU settled upon is outlined in the main report. Basic-
ally, a soil sealing ring is driven into the ground and a concentric bag
collar (which is open at both ends) is placed inside the ring. The inter-
space between the two rings is filled with dirt. A Teflon bag is then
placed over the inner "bag collar" and the procedure used for vegetation
samples is duplicated. This procedure can also be used to collect samples
over ponds and marshes by inserting the bag collar in a flotation ring of
closed cell polyethylene (two water-ski belts sewn together).
4.7.2. VEGETATION SAMPLES
The initial vegetation enclosure used in this study is shown in
Figure A.7.2-a. It was constructed of stainless steel, measured 30 cm
in diameter by 90 cm in length, and weighed about 16 kg. It had a plate
at its base (the end that faced the tree trunk) which had a number of
inserts with holes of different sizes to fit snugly around different
sized branches. It was equipped with an inlet zero air tube with small
holes along its side to promote even mixing. It also had numerous avail-
able sample ports and fittings for support brackets. This enclosure
apparatus was abandoned because it was difficult to handle and did not
allow samples to be taken under daytime conditions since the chamber
kept out all light. WSU found that Teflon bags were easier to clean
than expected, much easier to use, and allowed light to enter. Hence,
Teflon was adapted for use in this study.
A-24
-------
Figure A.6.2-a
STAINLESS STEEL SAMPLING CAN
INLET
SUPPORT
CONNECTOR
HANDLE-
OUTLET^ ___
SAMPLE PORT
SUPPORT CONNECTOR
OUTLET
INLET
PURGE
TUBE.
HANDLE
SAMPLE PORT
-SAMPLE
PORT
BRANCH INCLOSURE FLANGE
A-25
-------
A.8. SUMMARY
The emission measurement procedure developed for this project maxi-
mizes hydrocarbon concentrations while minimizing enclosure effects on
vegetation, such as heat build-up and vegetation damage. Each step of
the procedure has been carefully checked to insure that nothing is added
to the sample and that nothing is subtracted from it. WSU has checked the
analytical procedures for reproducibility and accuracy. WSU believes that
the result of this method is a realistic emission rate for each vegetation
species tested, and that each emission rate approximates the emission rate
of unenclosed vegetation at the time of sampling.
A-26
-------
APPENDIX B
DETAILED DERIVATION OF EMISSION RATE FORMULAS
B.I. INTRODUCTION
This appendix details the derivation of the emission rate formula. It
also shows how to calculate emission rates for samples collected using various
types of equipment and techniques.
The laboratory data format (Figure B.l-a.) lists and defines the emission
variables needed and contains the information necessary for the calculation
of emission rates. Depending upon the sampling procedure used, not all of
the blanks will be filled. For example, if a background sample is collected
and a bag blank and/or ambient air sample are not, the spaces for ambient
air data and bag blank data will remain empty. Empty spaces should always
be marked with a dash to indicate that the data were not collected. Questionable
data should be marked with an asterisk and a note at the bottom referring to
a laboratory notebook or to the field notes for a full explanation. The
sample illustrated in the figure utilized a bag blank sample, an ambient air
sample and an emission rate sample to determine emission rates. The sample
canisters had not been evacuated prior to sampling. Instead they contained
zero air at ambient pressure. This procedure was modified for later samples
(background and emission sample, evacuated canisters) and the simplified emission
rate formula was used. The old procedure included here as an illustration
of another emission rate collection technique and calculation procedure
which might be used if, for example, an evacuation system was not available.
B-l
-------
T
T
T2
T3
6
H
ZP
Vc
Figure B.l-a. Laboratory data format
Date collected 9/2'C____ Sample * Yf7. _
Date collectod__9/28_ Species Pondenjsj^f inc
Sample Blank
Time* at sample enclosure 1618 1549
Time at end of background sample
Time at start of flush 1620 1551
Time at end of flush 1626 1557
Time at start of purge 1626 1557
Time at start of sample collection 1627 1558
Time at end of sample collection 1633 1603
(°C) Ambient Temperature at Time of Sampling 28"C (cst.)
(t/min) Zero air flush flow rate (10 i/min) 10
(t/min) Zero air purge flow rate (2 z/min) 2
(m ) Dead volume of sample.enclosure at STP
(760 mm, 22/C) for bags = residual volume
after enclosure, before flush. 0.025
(m3) Dead volume of enclosure hank at STP
(« zero for bags) 0
Pcs: (psig) Pressure of sample can after sample
collection 15
Pca: (psig) Pressure of ambient air can after sairple
collection 15
Pc£: (psig) Pressure of enclosure blank can after sample
collection 15
(psig) Pressure of background can after sample
collection
Pb : (psia) Residual pressure of zero air in the sample
cans 13.19
- Barometric pressure nm Hg ^ for cans flushed with zero air
51.7 mm/psig
(ambient air can flushed, no dilution)
Zero for fully evacuated cans
CJS: (Mg/m3) TMMOC measured in vegetation sample 1248
Cjijt (iig/m3) THMOC measured in background sample —
Csa: (ug/m3) TK'HOC measured In ambient air sample 137
Cse: (wg/m3) TNMOC measured in enclosure blank sample 38
SA: (cm3 or g) Area or bionwss of sample vegetation 273.9Sg
*T1me refers to the hour and minute of the day. From this the length of
enclosure tinu1 (in minutes) can be determined.
B-2
-------
B.I.I DEFINITION OF TERMS
An explanation of each term used in Figure B.l-a and which might be
used to calculate emission factors is shown below:
AT^: This is the total length of time over which the emissions occured.
If a background sample is collected, the emission time is equal
to the time interval between the end of the background sample
collection and the end of the emission sample collection. If no
background sample is collected, the emission time is equal to the
total sample enclosure time (T6 - T0).
AT2: This is the flush time. It is standarized at 6 minutes.
AT3: This is the purge time. This purge flow is set a t 2 2,/min , the
2 &/min flow is begun at the end of the flush period and prior to
the beginning of the collection of the emission sample. The
purge flow is continued until the end of the sampling period.
AT^: This is the time that it takes to fill the sample canister to
the desired pressure. If the sample collection rate is known, the
volume of sample can be estimated. In practice, however, pressure
differentials are used to calculate the sample volume because the
sample rate is not constant. However, (AT^ x sample rate) is useful
to check the sample volume determined by pressure differentials.
The sample volume is only important when non-evacuated sample con-
tainers are used.
Zv: This is the total volume of air passed into the sampling enclosure
(Teflon bag), in cubic meters. It is equal to the flush flow
rate times the flush time plus the purge flow rate times the
purge time.
CT: This is the corrected VOC concentration of the sample canister.
For samples collected in evacuated stainless steel canisters or
Teflon sample bags, there is no dilution. For example, the
evacuated cans used by WSU in this study were pumped down to
approximately 30 microns. This yields a residual dead volume
of less than 0.21 ml. For samples collected in cans which have
been purged with a zero gas but not evacuated, a dilution factor
should be used that is equal to the absolute pressure gain of
the cans. The dilution factor Pc+Pb times the TNMOC of the can
Pea
(C,-s, Csjj, Csa and Cse in Figure B.l-a) is equal to the corrected
TNMOC for each sample.
Subscripts e, a, b and s designate enclosure blank, ambient air, back-
ground and emission rate samples respectively. Therefore, CTS is equal to
the concentration of the emission rate sample in yg/m .
B-3
-------
The contribution of the various components which can be included in
the emission sample can be calculated as shown below:
MGa: This component is the small contribution to the emission sample of
the residual ambient air in the enclosure. If a background sample
is collected, it will include MGa. When the bag is placed over a
branch and collapsed, some residual dead volume (Ve) will remain.
The ambient air in this residual volume is assumed to have a VOC
burden similar to ambient air collected in the proximity of the
vegetation chamber at the time of enclosure. Therefore, the dead
volume of the bag is estimated and then multiplied times the cor-
rected VOC for the ambient air sample, or MGa = (CTa) (Ve). This
gives the amount of VOC contributed to the emission sample due to
ambient air.
MGe: This component'is the contribution from the enclosure walls to
the emission sample. In some cases the enclosure may be difficult
to clean, or after many consecutive sampling times it may have
accumulated pitch or other emission residues. If this is suspected,
a bag blank should be collected to determine the enclosure emis-
sion rate. This sample is collected in the same manner as a
branch sample. The formula used to calculate the "emission rate
of the chamber walls" is identical to the emission rate formula.
Primes indicate that the variables apply to the "blank" sample.
Ambient air can also be an additive component to the blank total
for rigid enclosures. The ambient air contribution to rigid
enclosures is equal to MGa (Ve1). The dead volume (Ve') of
Teflon bag enclosures when collapsed during collection of the
enclosure blank equals zero, hence the ambient air contribution
to a Teflon bag blank is zero. The enclosure emission rate
(yg/nr) is then multiplied times the emission sample enclosure
time to give an estimate of the amount of VOC "bled" from the
enclosure wall (yg) during the vegetation emission period. The
equation is written as:
MGe = CTe (Ve1 + Zv1) - CTa (Ve1) A^
Ve1, Zv1 and ATi' represent the dead volume, the purge plus flush
volume and the length of enclosure time respectively, for a blank
sample. MGe here is equal to the TNMOC concentration of the bag sample
(yg/nr) x the volume enclosed by the bag (Ve1 + Zv1) minus the contri-
bution of ambient air CTa (Ve1) divided by the enclosure time ATi.
For Teflon bag enclosures enclosures, Ve1 = 0 since the bag can be
completely collapsed when a blank is collected. For Teflon bag en-
closure the equation therefore becomes:
MGe = CTe v^v ; AT}
B-4
-------
If A"TI = ATj'd'f the vegetation enclosure time is equal to the bag
blank enclosure time), which is often the case, the equation is simply:
MGe = CTe (Zv1)
If proper care is taken to clean enclosure chambers by purging and
heating slightly (they can be propped open and left in the sun), MGe
becomes very small. When a background sample is collected, AT^ is
shorter and MGe is even smaller and can therefore usually be omitted.
MGb: This is equal to the contribution from ambient air, from the
vegetation enclosure and from the vegetation enclosed until ATi,
as determined from a "background" sample. The background sample
is collected after a branch is enclosed but before any air has
been added to the enclosure, and is therefore equal to the VOC
concentration in the dead volume (Ve). For Teflon bag enclosures
this is the time at which the bag is collapsed around the vegeta-
tion. MGb is therefore equal to the VOC of the background sample
CTb times the dead volume of the bag (Ve). The dead volume is
estimated using the dimension measured at the time the background
sample is collected. The dead volume can also be calculated if
the concentration of a relatively inert tracer such as acetylene
or a halocarbon in the background sample is ratioed to its concen-
tration in the emission rate sample using the following formula:
Zv
"c " (Csb/Css^1
Where Cs^ and C$s are the concentrations of the tracer in the
background sample and emission rate sample respectively.
ER: This is the resulting emission factor for the vegetation. It is
equal to the VOC contribution from the emission sample (CTe)
(Zv+Ve) minus the VOC contribution of the background (MGb) minus
the VOC of the bag blank, divided by the chosen biomass components
of the branch and the length of emission time (SaMATj). It is
also equal to the VOC contribution of the emission sample minus
the contribution of ambient air and minus the contribution from
the enclosure walls.
B-5
-------
B.I. 2. EMISSION RATE FORMULAS
The emission rate formula when a background sample is collected is:
ER Ug/SaMrtn)- CT$ (Ve + Zv) " MGb " MGe where AT, = Tg-T, (1)
(Sa) (AT^
The emission rate formula when no background sample is collected is:
GTS (Ve + Zv) - MGa-MGe
_ =
" (Sa) (ATj) ' b U
If a background sample is collected, evacuated canisters or Teflon
sample bags are used and the bag blank MGe is small, the equation
is reduced to:
Css (ve + Zv) - Csb ye
(Sa)
where:
Ccc: equals the TWIOC measured for the emission sample (yg/m3)
5> 5>
C^: equals the TNMOC measured for the background sample (yg/nr)
Ve: equals the dead volume of the bag when collapsed around the
branch (nr)
Zv: equals the total volume of zero air put into the enclosure (m3)
Sa: equals the chosen biomass components of the sample (g) or leaf
litter or water surface area (nr)
ATj^: equals the total emmission time (min) = T6 - Tj.
B-6
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA 450/4-79-004
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Testing of Hydrocarbon Emissions from
Vegetation, Leaf Litter and Aquatic Surfaces, and
Development of a Methodology for Compiling Biogenic
5. REPORT DATE
February 1979
6. PERFORMING ORGANIZATION CODE
-Eroi
SS10
(OR iS)
on
7. AUTHOR
8. PERFORMING ORGANIZATION REPORT NO.
Patrick R. Zimmerman
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Washington State University
Air Pollution and Resources Section
Pullman, Washington 99164
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
14. SPONSORING AGENCY CODE
200/4
15. SUPPLEMENTARY NOTES
EPA Project Officer: Thomas F. Lahre
16. ABSTRACT
This report outlines a general methodology for estimating emission rates of volatile
organic compounds (VOC) from vegetation, soil/leaf litter, and surfaces of bodies of
water. Techniques are prescribed for sample collection and analysis as well as for
extrapolating the emission rates determined to estimate biogenic VOC emissions^over
any area. Emission factors are presented for broad classes of vegetation and for
the major biotic regions in the conterminous U.S. Emission inventory procedures
are illustrated by using emission factors to develop an example annual VOC emission
inventory for the U.S. This nationwide emission inventory indicates that most
biogenic VOC emissions can be expected during the summer months. In general, vege-
tation emits much larger quantities of VOC than does either leaf litter or water
surfaces. Isoprene and various terpenes comprise the bulk of all vegetation
emissions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Natural Emissions
Plants
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Biomes
Gas Chromatograph Soil
Hydrocarbons
Inventories
Isoprene
Leaf Litter
Terpenes
Trees
Vegetation
.Volatile Organic
Compojnds
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report/
21. NO. OF PAGES
20. SECURITY CLASS (This page)
-Hi
22. PRICE
EPA Form 2220-1 (9-73)
------- |