EPA/600/A-94/073
Assessing Ozone Effects on Plants Native to
the Southeastern United States
Neufeld, Howard S.1, James R. Renfro2, Songqiao Huang1,3,
W. David Hacker213, Deborah Mangis2,0, Arthur H. Chappelka3,
William E. Hogsett4, Andrew A. Herstrom5, John A. Laurence6,
E. Henry Lee5, James E. Weber4, and David T. Tingey4
1Dept. of Biology, Appalachian State Univ., Boone, NC USA 28608
2Great Smoky Mountains National Park, 1314 Cherokee Orchard
Rd., Gatlinburg, TN USA 37738
3School of Forestry, Auburn University, Auburn, AL USA 36849
4U.S. EPA, Environmental Research Laboratory - Corvallis, 200 SW
35th St., Corvallis, OR USA 97333
5ManTech Environmental Technology, Inc., 200 SW 35th St.,
Corvallis, OR USA 97333
6Boyce Thompson Institute for Plant Research, Tower Road, Ithaca,
NY USA 14853
Current Addresses:
aDept. of Botany, University of Wyoming, Laramie, WY USA 82071
Environmental Division, Texas Dept. of Transportation, 125 E. 11th
St., Austin, TX USA 78701-2483
CU.S. EPA, MD 80A, Research Triangle Park, NC USA 27711

-------
Abstract
For the past six years, the U.S. National Park Service, U.S. Environmental
Protection Agency, and University researchers have been documenting the effects of
ozone on a large number of woody and herbaceous species native to the
southeastern United States. In Great Smoky Mountains National Park (GRSM), ozone
levels exhibit diel patterns at low elevations, where concentrations are low in the
morning and high in the afternoon. At high elevations (> 800 m), morning
concentrations are much higher, and the total daily exposure is approximately twice
that at the lower elevations. Putative ozone injury has been observed in the field in
GRSM on 90 species, representing approximately 6% of the known flora in the Park.
Surveys of foliar injury on several tree species show a general pattern of increasing
frequency and amount of stipple with increasing elevation in GRSM, and in nearby
Shenandoah National Park. Tall milkweed (Asclepias exaltata L.) and black cherry
(Prunus serotina Ehrh.) appear to be among the most sensitive herb and woody
species, respectively, in GRSM. Exposure-response studies were carried out in open-
top chambers for six years with 46 species. Foliar symptoms seen in the field, were
reproduced on 30 species, providing evidence that the foliar injury found in the field
was probably due to ozone exposure. The exposure-response studies indicated that
early-successional and shade intolerant species were less resistant to high ozone
levels than late-successional species and dry-site adapted conifers. Geographic
information systems (GIS) techniques, coupled with growth models, were used to
scale-up responses from the seedling to the geographic level. Exposure-response
data were coupled with species' distributions and geographic patterns of ozone
exposure to predict growth losses for several woody species. These data suggest
that species such as black cherry may suffer yearly growth losses of up to 32%,
whereas less sensitive species, such as red maple (Acer rubrum L.) may exhibit
losses of only 2-4%, even though they share a similar geographic range. In addition,
some species, such as trembling aspen (Populus tremuloides Michx.) show genetic
variation in tolerance to ozone. When growth data for species are projected over a
typical 70 year life span, small yearly losses can be compounded, resulting in over
50% growth losses, as is the case predicted for loblolly pine (Pinus taeda L.).
Because the modeling work is totally dependent on the quality of the exposure-
response data, future work must concentrate on more realistic exposure systems (i.e.
chamberless systems, plants rooted in the ground), genetic variation within species,
and refinement of the modeling procedures used in scaling up.

-------
introduction
Ozone is regarded as the most widespread phytotoxic air pollutant in the
United States, and an assessment of its impacts on plant growth and ecosystem
functioning is necessary for setting air quality standards. Currently, the National
Ambient Air Quality Standard (NAAQS), as set by the Clean Air Act of 1977 (U.S. EPA
1986), mandates that hourly ozone levels in an airshed not exceed 120 ppb more
than three times in three years. Many urban areas often exceed this level, and must
take measures to bring the levels down. However, recent observations in natural
areas throughout North America, where ozone concentrations are below NAAQS, are
showing many plants exhibiting foliar symptoms suggestive of ozone injury (Neufeld
et al. 1992). This suggests that the primary NAAQS may be inadequate for protecting
certain plant species, and that perhaps either the primary or secondary standards
could be modified. Currently, the secondary standard is being modified to reflect
recent studies of the effects of ozone on woody plants (W. Hogsett, U.S.
Environmental Protection Agency, pers. comm.).
Nearly two-thirds of the population of the United States lives east of the
Mississippi River (U.S. Bureau of the Census, 1991). A night view of North America
shows the high density of lights in the eastern half of the nation, which is highly
correlated with population estimates (Figure 1). Not only do these lights reflect
population density, they also act as surrogate indicators of energy consumption.
Since fossil fuels are the primary energy source, these lights are also good visual
estimators of the production of air pollutants, particularly mobile sources, such as
cars, and includes nitrogen oxides, SOx and volatile organic carbon compounds.
Most Federal natural areas in the United States are located in the western half
of the nation (Figure 2). The Clean Air Act (CAA) amendents of 1977 define
wilderness areas and national parks as Class I areas if they were formed prior to the
act and are over 6,000 acres in size. The CAA specifies that the ozone in these areas
must remain below the 120 ppb standard to maintain a high degree of air quality.
These areas are afforded the greatest degree of protection under the CAA. Although
relatively few key Class I areas are located in the eastern United States, they are
potentially at great risk due to high pollutant loadings. The most notable areas are
Acadia National Park in Maine, Shenandoah National Park in Virginia, Great Smoky
Mountains National Park (GRSM) in Tennessee and North Carolina, Okefenokee
Wildlife Refuge in Georgia, and Everglades National Park in southern Florida. Great
Smoky Mountains National Park is the most visited park in the nation, drawing some
8-10 million visits each year (Peine and Renfro 1988), and has some of the highest
levels of biological diversity in the southeast (White, 1982). In addition, there are
numerous rare and endangered species that occur only within the boundaries of
1Some of the research described herein was developed by Deborah Mangis while an employee of
the National Park Service. It was conducted independent of EPA employment and has not been
subjected to the agency's peer and administrative review. Therefore, the conclusions and opinions
drawn are solely those of the author and should not be construed to reflect the views of the EPA.

-------
GRSM which might be at risk either to pollutants directly, or secondarily because of
ecosystem degradation due to pollutants. These Federal areas are of concern to the
U.S. National Park Service and U.S. Fish and Wildlife Service because of their
potential vulnerability to air pollutants.
The success of any air quality program depends not only on determining
pollutant loadings in critical areas, but also on determining the source locations for
these pollutants. Identification of pollution sources enables regulators to apply
control measures specified in the Clean Air Act for maintaining good air quality. Air
pollutants may drift hundreds of miles before they reach natural areas. For example,
when ozone levels in GRSM are high, the air containing that ozone can be shown to
have originated either over the industrial Ohio Valley, or to have traveled northward
from industrial areas in Louisiana, Alabama, Georgia and Tennessee (Dattore et al.
1991). Even in fairly remote areas, it is common to see pollutants accumulate at the
inversion layer, such as the one shown in Figure 3, which is on Whitetop Mountain in
southwestern Virginia, one of the most rural areas in the eastern United States.
This paper deals primarily with the effects of ozone on native plants in GRSM,
but because many of these species are geographically widespread, has important
implications for plants throughout the southeastern United States. We concentrate on
this particular geographic area because it is subject to chronically high ozone levels
(60-100 ppb) during the growing season, and because of the importance of forests
and native plants to this region. The southeast is prone to high ozone because of
several reasons. First, as mentioned above, polluted air drifts into the region from
other areas. Second, the region has abundant sunshine, high temperatures, and a
large number of days with stagnating air masses. Together, these factors contribute
to the rapid and abundant formation of ozone (Lefohn 1992). Thus, despite relatively
low population figures in comparison to the more densely settled northeast, the
southeast is at a relatively greater threat from ozone than would be predicted from
population estimates alone.
For the last six years, the Air Quality Division of the U.S. National Park Service
has been funding research into the effects of ozone pollution on plants native to
GRSM. This research had several goals: 1) to survey plants in the field for signs of
putative ozone injury, 2) to determine relative sensitivities of plants to ozone injury,
based on both foliar symptoms and growth reductions, and 3) to document whether
the injury observed in the field was indeed due to ozone exposure. For this, an
ozone exposure facilty was set up at the Uplands Field Research Laboratory, located
at Twin Creeks in GRSM. Plants were exposed to varying levels of ozone and foliar
symptoms, if any, were noted. Additionally, the facility was used to generate an
extensive set of exposure response curves for biomass accumulation, which are
currently being used in conjunction with the foliar data to rate species' sensitivities to
ozone. Finally, the exposure data, along with the field data, are being used to
estimate the potential threat of ozone to the plants and ecosystems of the Park.
2

-------
Models are being developed to predict ozone in remote locations, where monitoring
equipment is not available, and then used in conjunction with plant growth models to
scale up responses from seedling level to large-scale geographic patterns.
The intent of this paper is to present a synopsis of our work in the
southeastern United States regarding ozone effects on native plants, and to suggest
possible future avenues of research in this area. In addition, we outline a preliminary
attempt to scale up responses from seedlings, to mature trees, to stands, and finally
to the regional geographic level.
Methodologies
Plants in GRSM have been observed during the summers of 1987-1993 for
foliar symptoms consistent with those known to be caused by ozone. For the past
three years, additional effort has been made using trend plots to survey dominant
canopy trees at high and low elevations. Trend plots have no fixed boundaries,
rather, the criteria are that sample trees be located within 152 meters elevation and
3.3 km of stationary ozone monitors. These monitors are located at Cove Mountain
(1264 m elevation), Look Rock (863 m elevation), and Twin Creeks (594 m elevation).
Sampling large trees has entailed the use of tree climbers to collect foliage from the
crown. Data from these plots allows researchers to identify geographic spatial
variation in ozone injury within GRSM. In addition to the trend plots, observers have
walked or driven along trails and roads looking for injury symptoms. These visual
surveys are always carried out from mid- to late-summer when symptoms of ozone
injury are most likely to be seen. Observers look for any species that shows
symptoms that might be due to ozone, and only presence or absence of injury is
recorded. Slightly greater than 50% of the trail system in GRSM has been surveyed.
Because observers change from year to year, extensive training sessions and quality
control exercises are done to minimize human error in the assessments. Voucher
specimens showing putative injury are located in the herbarium at the Park.
The ozone exposure system started out as a nine chamber system in 1987,
and with additional funding from the U.S. Environmental Protection Agency (U.S.
EPA), expanded to 15 chambers in 1990. Frustums and raincaps were also added in
1990 to better distribute the ozone within the chambers, and to prevent rain damage
from thunderstorms.
Ozone was generated by an electric spark discharge generator. Ambient air
was used to generate ozone in 1987 and 1988, but liquid oxygen was used thereafter
to prevent formation of oxides of nitrogen (Brown and Roberts 1988). Ozone was
dispensed 24 hours per day, seven days per week. This was done because high
ozone levels have occassionally been observed at night, and without a 24 hour
fumigation, these peak events would have been missed.
3

-------
During 1987 and 1988, the ozone treatment levels consisted of 7-day exposure
profiles (developed from data at the Look Rock site), with one simulating ambient
conditions at Look Rock, and the other 2.Ox ambient. Control chambers had
charcoal filters to reduce the ozone levels below ambient. In 1989, modified ambient
exposures were used, and the treatments were: charcoal-filtered, 1 .Ox, 1,5x, and 2.Ox
ambient, as well as open ambient plots. In 1990, a 0.5x treatment was added.
Ozone levels were continually adjusted by a datalogger throughout the day
approximately five times per hour so as to more closely track diurnal patterns. There
were three replicate chambers for each ozone treatment for 1987 and 1988. In 1989
treatment replications were reduced to two, except for the 2.Ox and open plots.
Beginning in 1990, all treatments were replicated three times. Generally, about 10
plants per species were placed in each chamber, but for difficult-to-grow species,
there were sometimes fewer than this number. Plants were arranged within the
chambers to minimize shading by taller species.
Ozone concentrations in the chambers were monitored by TECO Model 49
analyzers. Teflon tubing and filters were used throughout the system, and line losses
were consistently less than 5%. All monitoring equipment was checked quarterly by
state and national auditors, and in all cases the monitors were within compliance for
precision and accuracy. Data were stored in a Campbell 21 x datalogger, and
downloaded to a computer for final storage twice daily.
Plants were raised from seed collected in the field within GRSM, and were
generally first year or one year old seedlings at the time of exposure. Seedlings were
grown in pots of varying sizes, depending on species, in Pro-mix soil-less media,
watered to excess daily, and fertilized once weekly with a 20-20-20 water soluble
fertilizer. For certain tree species, slow-release Osmocote fertilizer was used instead.
Where possible, height, diameter, biomass, leaf count and area were obtained
at the end of the exposure period. Most species were exposed for a single growing
season (typically from some time in May to late August or September), but some
woody and herbaceous species were exposed for more than one growing season.
Table mountain pine (Pinus pungens Lambert), and Canadian hemlock (Tsuga
canadensis L. Carr.) for example, were exposed for three consecutive seasons.
The data from the exposure experiments were subjected to regression
analysis, covariance analysis, and Chi-square analysis depending on the variable
being analyzed. The Weibull function was used to model responses to ozone
exposure if the analysis of variance indicated a significant non-linear trend. Details of
the statistical treatments can be found in Neufeld et al. (1992) and Neufeld and
Renfro (1993).
4

-------
Results and Discussion
Ozone Dynamics
Diel (day and night) concentrations at low elevation sites show a typical pattern
of low concentrations in the morning hours, with higher concentrations in the
afternoon (Figure 4). This happens because ozone is formed through the interactions
of light, nitrogen oxides, and volatile organic compounds (Chameides and Lodge,
1992). Because of the diurnal patterns of light, and the time constants involved in the
reactions, it takes several hours for the ozone to build up each day. During the night,
ozone concentrations decrease because the inherently unstable molecule breaks
down and is scavenged by various sources, mainly NOx. In contrast, at high
elevations, early morning ozone concentrations are high, and there is much less
contrast between afternoon maxima and morning minima. This results when low
elevation inversions trap ozone at higher elevations, when there is a lack of NOx, and
if stratospheric intrusions occur (Wolff et at. 1987). In GRSM, on average, ozone
exposures at high elevations (> 800 m) (obtained by integrating under the daily
concentration curves) run about twice that at low elevations (<. 800 m) in the Park.
Thus plants at high elevations are potentially at greater risk due to the higher
exposures.
In addition to the different diel patterns, higher ozone concentrations are more
frequent at the upper elevations in the Park. The majority of ozone concentrations
are between 40-70 ppb at high elevations, whereas at low elevations, they are
between 20-40 ppb (Figure 4). Much of the difference between the two elevations is
due to the low nighttime values at lower elevations.
Because ozone is highly reactive, it can be scavenged out of the air by coming
into contact with surfaces, such as soil, tree trunks and leaves. Fully 75% of GRSM is
closed canopy hardwood forest. Using a portable ozone monitoring station (Neufeld
et at. 1992), we found that at the ridge-tops, where the stationary ozone monitors are
located, and where forests are of smaller stature and lower basal area, depletion of
ozone through the canopy is at most only about 20% at 1 meter from the forest floor
(Figure 5). In contrast, ozone 1 meter above the ground in closed-canopy cove
hardwood forests can be depleted to less than 50% of that above the canopy. Thus,
even though the upper canopy trees in cove hardwood forests may be experiencing
high ozone, plants in the understory are relatively protected. These microsite and
topographic influences make it difficult to model ozone distributions in areas with
complex relief and diverse vegetation types.
5

-------
Foliar Symptoms
To date, 90 species of plants have been observed to exhibit putative ozone
symptoms in the field in GRSM. A table of species and their foliar sensitivities can be
found in the appendix at the end of this paper (from Neufeld et al. 1992). This
represents approximately 6% of the total known flora in GRSM, and given the small
amount of land area covered by the surveys, suggests that there may remain other
species which are sensitive to ozone. Tall milkweed, (Asclepias exaltata L.) appears
to be the most sensitive herbaceous species. Complete defoliation has been
observed in the field, and has occurred in areas where the maximum concentrations
have not exceeded 70 ppb. The symptomatology prior to defoliation is suggestive of
ozone injury. The most sensitive tree species appears to be black cherry (Prunus
serotina Ehrh.), and foliar symptoms on this species have been found in the field
throughout the Park (Chappelka et al. 1992).
In our surveys of mature trees, using the tree climbers, foliar injury in 1991 and
1992 followed a general pattern of increasing frequency and amount with increasing
elevation in GRSM (Chappelka et al. 1992). A similar pattern was noted in
Shenandoah National Park (J.M. Skelly, Dept. of Plant Pathology, Pennsylvania State
University, pers. comm.), which is about 500 km to the north. Since higher elevation
sites were sampled earlier than low elevation ones, this trend seems due more to
ozone than to seasonality.
Exposure-Response Results
We have fumigated a total of 46 species to date in the chambers. A list of the
species tested can be found in Neufeld et al. (1992). Of these 46, 35 had shown
foliar injury in the field. Among these 35, 30 showed symptoms in the chambers like
those found in the field, lending credence to our hypothesis that the injury symptoms
seen in the field were due to ozone.
As mentioned earlier, tall milkweed appeared hypersensitive in the field, and
proved to be so in the exposure chambers as well. After four weeks in the 2.0x
treatment, there was premature sensescence of the leaves, and after six weeks, many
of the plants had died. Only plants grown in charcoal-filtered air retained all their
leaves throughout the exposure period (Figure 6). The large loss of leaves in the
open plots relative to the 1 .Ox chambers is most likely a result of strong winds and
rain in the former.
Black cherry exhibited large, statistically significant biomass responses to
increasing ozone. As with tall milkweed, the most obvious response was
discoloration, and stipple on older leaves, followed by premature senescence. This
led to more than 50% losses in leaf number and area in the 2.Ox treatment (Figure 7).
6

-------
These losses in leaf material contributed significantly to large biomass losses,
particularly for leaves and roots. At 2.0x, total biomass was reduced by up to 32% in
comparison to the charcoal-filtered treatment, as estimated by the Weibull function
(Figure 7). For sycamore, (Platanus occidentalis L.) there was severe defoliation of
the lower leaves at 2.0x (Figure 8), but interestingly, no change in final leaf count,
presumably because this species compensated for the lower leaf loss by producing
more new leaves. In addition, there were no effects on height or diameter, and only
a marginal effect on total biomass. Finally, the most resistant species appeared to be
late successional species, and dry-site adapted conifers. Table mountain pine {Pinus
pungens Lambert), which was exposed for three consecutive seasons, showed
severe foliar injury on two year old needles at 2.Ox, but did not exhibit a significant
biomass response. This illustrates an important point, namely, that foliar injury and
biomass responses are not highly correlated. Growth reductions can occur in the
absence of injury (Hogsett et al. 1985), and vice-versa (Neufeld et al. 1992). In
addition, it points to an important modifier of ozone susceptibility, namely drought.
Because ozone enters leaves primarily through the stomata, any environmental stress
that causes stomatal closure is likely to reduce ozone uptake and injury. For plants
growing in the field, drought is the most likely cause of reduced stomatal aperture.
Thus to accurately rate the sensitivity of a species requires documentation of both
foliar and growth responses, as well as the modifications induced by environmental
change, particularly drought.
Spatial Risk Characterization2
Characterizing the spatial risk that ozone poses for forested areas requires
knowledge of the ozone exposure characteristics over the appropriate geographic
area (watershed, region, national), quantification of the phytotoxic effects at the
appropriate biological level (species, population, community), and the types of
uncertainty associated with these effects. It also includes recognizing the
perspectives by which cultural and societal needs dictate the definition of "risk". For
the purposes of this paper, risk is associated only with development of damage,
where damage means an impairment of growth. It is implicitly assumed that damage
is correlated with altered community and ecosystem functioning, and that such
changes represent a cost to society as a whole. Therefore, such risk assessments
can be useful in setting guidelines for secondary air quality guidelines within the
criteria of the Clean Air Act for protecting forest resources.
A first attempt was made several years ago to estimate ozone risks for crops
under the NCLAN program (National Crop Loss Assessment Program) and the results
2The spatial characterization section in this paper is adapted from Hogsett et al. (1993) and
appropriate authorship should be attributed to them for the content contained therein.
7

-------
suggest that current ambient levels of ozone cause anywhere from 11-14% reductions
in crop yields (Tingey et al. 1993). However, similar risk assessments for natural
areas, and forests in particular, are lacking due to the difficulty of working with large
trees, and organisms with long lifespans.
The risk characterization presented in this paper uses GIS (Geographic
Information Systems) to integrate empirical growth data from ozone-exposure
response studies and model simulations of long-term growth trends with spatial
estimates of ozone exposure and species distributions in order to generate a spatial
assessment of ozone impact. In this preliminary assessment, we use the eastern
United States to illustrate the approach, mainly because there are more data available
for species in this geographic region, and the ozone exposures are higher. The
eastern United States contains 10 different forest types (Eyre 1980) and a
substantially large number of species (over 100 in GRSM alone). The assessment is
based on the adverse effect on productivity and species assemblages. We have
used only limited empirical data on seedling exposure-responses of eight different
species.
The spatial distribution of ozone exposures was interpolated from stationary
monitoring sites located in the eastern United States, using a model for the formation
and degradation of ozone (Hogsett et al. 1993). The model uses environnmental
variables such as temperature, wind direction, NOx formation, stagnation and
radiation values to estimate ozone formation, drift, and degradation. This results in a
contour map showing regions of similar ozone exposure. For model simulations, the
exposures are routinely standardized to a three month summer exposure period. GIS
technology is used to generate the maps, using grids 20 km on a side. A total of 140
north-south cells and 135 east-west cells (18,900 total) were used (Figure 9).
Seedling ozone exposure response data are then used to generate species
specific responses to different ozone exposures. These can be overlain on the GIS
maps of ozone exposure to generate spatial patterns in predicted biomass
responses. A tree growth model, in this case TREGRO (Weinstein et al. 1992) is
used to "grow" species over standardized lengths of time to generate the necessary
data for the spatial modeling effort. Eventually, it should be possible to pass these
data on to a stand growth model, i.e., ZEUG (Urban 1990), to simulate biomass
responses at the next higher level of organization and to estimate multispecies
responses. Finally, the stands can be "grown" over long time intervals to estimate the
potential effects of ozone on forests themselves. The decreases in biomass over
these time intervals can then be used as one of several inputs for spatial risk
characterization. Details of the modeling efforts can be obtained from Hogsett et al.
(1993).
8

-------
Results of the Modeling Efforts
Exposure-response data were obtained for eight different tree species (see
Figure 12 for the list). The SUM06 exposure index was used to calculate the
response curves. This is simply the sum of the concentrations for those hours when
the ozone exceeded 60 ppb (Hogsett et al. 1988). The Weibull function was used to
model the growth responses (Rawlings and Cure 1985), and maps of the spatial
distribution of each species were generated using the GIS. The maps of ozone
exposure and distribution were then superimposed, and the response functions used
to predict percent biomass losses relative to charcoal-filtered air at each point on the
map. Interpolation techniques allowed predictions for areas where there were no
ozone data available.
Black cherry is predicted to have the greatest losses due to its high sensitivity
(Figure 10). In contrast, red maple (Acer rubrum L.), which has a very similar
distribution, is predicted to have losses of less than 2% in 1988, a high ozone year
(Hogsett et al. 1993). This is because it is inherently less sensitive to ozone than
black cherry (Neufeld et al. 1992). There are both year to year variations in losses,
due to different amounts of ozone each year, and there are genotypic variations
within species. For example, losses of greater than 30% are predicted for black
cherry over 90% of its range in 1988 (Figure 10), but in the low ozone year of 1989,
less than 10% of the area was predicted to have losses over 20%. In Michigan,
studies were carried out in a similar fashion to those in GRSM, and for one species of
aspen (Populus tremuloides Michx.) different genotypes were compared for their
response to ozone. Based on the results of these studies, one clone is predicted to
have much greater losses than another clone, even though they can be found
growing over the same geographic range (Figure 11). Looking at all the species
surveyed, the predicted area-weighted biomass losses vary widely, ranging from
0-35% depending on species and year (Figure 12).
When NCLAN data are used to predict annual agricultural crop losses over a
standardized three month growing season, and using the SUM06 as the index, 50%
of the crops would suffer little or no loss up to 26.4 ppm*hrs exposure (Tingey ef al.
1991). For trees, the data suggest that at 20 ppm*hrs, 50% of the species would
suffer annual losses of 4%. This is a small number, but if the losses are
compounded over a portion of the lifespan of a tree (40-80 years), the cumulative
losses become substantial. In fact, over this time span, losses may exceed 50%,
even in relatively insensitive species such as loblolly pine (Pinus taeda L.) (Figure 13).
9

-------
Conclusions
This risk assessment characterization is still in the preliminary stages. More
data are needed on environmental interactions, and how they affect exposure
response curves. We need to study temperature, drought, light, and humidity
interactions, as well as soil factors such as fertility and type (Pye 1988). Furthermore,
we must ascertain whether exposure response curves from studies using potted
seedlings are useful for scaling up to sapling and mature tree responses.
Increasingly, global climate change must be factored in, particularly the potential
influence of rising C02 levels. And finally, the genetic and physiological bases of
resistance to ozone require much more study (Roose 1991).
Taken in context, the data suggest potential risk at current levels of ozone to
plants in natural areas of the southeastern United States. Only further intensive study
can determine what the ecological, sociological, economic and cultural risks are to
the United States.
10

-------
References
Brown, K.A. and T.M. Roberts. 1988. Effects of ozone on foliar leaching in Norway
spruce [Picea abies (L.) Karst]: Confounding factors due to NOx production
during ozone generation. Environmental Pollution 55:55-73.
Chameides, W.L. and J.P. Lodge. 1992. Tropospheric ozone: formation and fate, In
"Surface Level Ozone Exposures and Their Effects on Vegetation", A.S. Lefohn,
ed., Lewis Publishers, Inc., Chelsea, Ml. pp. 5-30.
Chappelka, A.H., E. Hildebrand, J.M. Skelly, D. Mangis and J.R. Renfro. 1992. Effects
of ambient ozone concentrations on mature eastern hardwood trees growing in
Great Smoky Mountains National Park and Shenandoah National Park. Proc.
Air and Waste Management Association Meetings, Kansas City, MO,
June 21-26, 1992.
Cooper, A.W. 1981. Above-ground biomass accumulation and net primary production
during the first 70 years of succession in Populus grandidentata stands on
poor sites in northern lower Michigan, In "Forest Succession: Concepts and
Applications", D.C. West, H.H. Shugart, and D.B. Botkin, eds., Springer-Verlag,
NY. pp. 339-360.
Dattore, R.E., K.A. Gebhart, W.C. Malm, and M. Flores. 1991. Use of an atmospheric
trajectory model to explore the source regions affecting ozone concentrations
at five eastern U.S. National Parks. Paper 91-66.7. 84th Annual Meeting of the
Air & Waste Management Association, Vancouver, British Columbia, Canada,
June 16-21, 1991. Pgs. 1-25.
Eyre, F.H. 1980. Forest Cover Types of the United States and Canada; Society of
American Foresters.
Hogsett, W.E., M. Plocher, V. Wildman, D.T. Tingey, and J.P. Bennett. 1985. Growth
response of two varieties of slash pine seedlings to chronic ozone exposure.
Canadian Journal of Botany 63:2369-2376.
Hogsett, W.E., D.T. Tingey, and E.H. Lee. 1988. Ozone exposure indices: Concepts
for development and evaluation of their use, In "Assessment of Crop Loss from
Air Pollutants: Proceedings of an International Conference", W.W. Heck, O.C.
Taylor and D.T. Tingey, eds., Elsevier Applied Science, London, pp. 107-138.
11

-------
Hogsett, W.E., A.A. Herstrom, J.A. Laurence, E.H. Lee, J.E. Weber, and D.T. Tingey.
1993. Risk characterization of tropospheric ozone to forests, In "Proceedings of
the 4th U.S./Dutch International Symposium: Comparative Risk Analysis and
Priority Setting for Air Pollution Issues. Air and Waste Management Association,
Pittsburgh, PA. In press.
Lefohn, A.S. 1992. Tropospheric ozone: Formation and fate. In "Surface Level Ozone
Exposures and Their Effects on Vegetation", A.S. Lefohn, ed. Lewis Publishers,
Inc., Chelsea, Ml. Pgs. 5-30.
Neufeld, H.S. and J.R. Renfro. 1993. Sensitivity of black cherry seedlings (Prunus
serotina Ehrh.) to ozone in Great Smoky Mountains National Park: The 1989
seedling set. Natural Resources Report NPS/NRTR-93/112, Air Quality Division,
U.S. National Park Service, pgs. 1-26.
Neufeld, H.S., J.R. Renfro, W.D. Hacker, and D. Silsbee. 1992. Ozone in Great Smoky
Mountains National Park: Dynamics and effects on plants. In "Tropospheric
Ozone and the Environment II: Effects, Modeling and Control", edited by
Ronald L. Berglund. Air & Waste Management Society, Publisher, Pittsburgh,
PA USA. pgs. 594-617.
Peine, J.D. and J.R. Renfro. 1988. Visitor Use Patterns in Great Smoky Mountains
National Park. Research/Resources Management Report SER-90, Southeast
Regional Office, U.S. National Park Service, Atlanta, GA. 93 pgs.
Rawlings, J.O. and W.W. Cure. 1985. The Weibull function as a dose-response model
to describe ozone effects on crop yields. Crop Science 25:807-814.
Roose, M.L. 1991. Genetics of response to atmospheric pollutants, In "Ecological
Genetics and Air Pollution", G.E. Taylor, L.F. Pitelka, and M.T. Clegg, eds.,
Springer-Verlag, NY. pp. 111-126.
Schumacher, F.X. and T.S. Coile. 1960. Growth and yields of natural stands of the
southern pines, T.S. Coile, Inc., Durham, NC.
Tingey, D.T., W.E. Hogsett, E.H. Lee, etal. 1991. An evaluation of various alternative
ambient ozone standards based on crop yield loss data, In "Transactions:
Tropospheric Ozone and the Environment", R.L. Berglund, D.R. Lawson, and
D.J. McKee, eds., Air and Waste Management Association, Pittsburgh, PA.
pp. 272-288.
Tingey, D.T., D.M. Olszyk, A.A. Herstrom, etal. 1993. Effects of ozone on crops, In
"Ozone - Public Health and Welfare", D.J. McKee, ed., Lewis Publisher, Boca
Raton, FL. (in press).
12

-------
Urban, D.L. 1990. A Versatile Model to Simulate Forest Pattern, A User's Guide to
ZELIG Version 1.0. Environmental Sciences Department, The University of
Virginia, Charlottesville, VA.
U.S. Bureau of the Census. 1991.
U.S. Environmental Protection Agency. 1986. Air Quality Criteria for Ozone and Other
Photochemical Oxidants, Vol. Ill, EPA-600/8-84-020cF.
Weinstein, D.A., R.M. Beloin, R.D. Yanai, and C.G. Zollweg. 1992. The Response of
Plants to Interacting Stresses: TREGRO Version 1.74. Description and
Parameter Requirements. Electric Power Research Institute TR-101061.
White, P.S. 1982. The flora of Great Smoky Mountains National Park: An annotated
checklist of the vascular plants and a review of previous floristic work.
Research/Resources Management Report SER-55. U.S. National Park Service,
Southeast Regional Office, Atlanta, GA. 218 pgs.
Wolff, G.T., P.J. Lioy, and R.S. Taylor. 1987. The diurnal variations of ozone at
different altitudes on a rural mountain in the eastern United States. Journal of
the Air Pollution Control Association 37:45-48.
13

-------
Figure Legends
1.	A night view of North America, showing the lights from major urban areas. Photo
courtesy of NASA.
2.	Mandantory Class 1 Areas in the United States.
3.	View from Whitetop Mountain, Virginia, showing accumulation of pollutants in the
inversion layer. Photo by H. Neufeld.
4.	Diel ozone patterns for one day in Great Smoky Mountains National Park. Twin
Creeks is a low elevation site (594 m), while Look Rock (823 m) and Cove
Mountain (1264 m) are ridge top sites.
5.	Mean ozone concentrations at 3 heights above the forest floor for several sites in
GRSM expressed as a percentage of the value above the canopy at Look
Rock, a mid-elevation (823 m), ridge-top site. The first three stations on the
left side of the graph are ridge sites, and the remaining three stations are
closed canopy forests.
6.	Final leaf number at the end of the season for tall milkweed (Asclepias exaltata
L.) as a function of ozone treatment. Treatments (along with their SUMO
exposures in ppm*hrs) are: CF (charcoal-filtered, 7.4), 1.0x ambient (20.5),
1.5x ambient (30.2), 2.Ox ambient (44.0), and open plots (non-chambered,
26.3).
7.	Response of black cherry seedlings to ozone fumigation in 1989: a) leaf area, b)
height (O) and diameter (•), and c) dry weight accumulation for various plant
parts: (V) - total dry weight, (•) - root dry weight, (O) - leaf dry weight. Solid
lines are best fits obtained using a Weibull function. Points are chamber
means. Ozone exposure calculated as a 24 hour seasonal sum.
8.	Sycamore seedlings (Platanus occidentalis L.) grown in (A) charcoal-filtered
chambers, and (B) in the 2.Ox chambers. Notice the lack of leaves on the
lower stem of the plants from the 2.0x chambers.
9.	Ozone monitoring site locations for 1988 and calculated 3-month SUM06 at each
site (A). Ozone exposure potential surface "EPS" (B). Increasing potential for
high ozone exposure is indicated with increasing degree of shading. EPS
derived from factors given in Table 1 of Hogsett et al. (1993).
14

-------
10. Predicted biomass loss (PRBL) for black cherry (Prunus serotina Ehrh.) (A) and
red maple (Acer rubrum L.) (B) with 1988 ozone exposure. PRBL calculated •
for each 20 km cell based on estimated ozone exposure value (3-month
SUM06) and Weibull parameters for each species response function. Note the
different scales for each species. The PRBL ranges from <20 to >30% in
black cherry, but in red maple, PRBL is <2 to >4% over its entire range.
11.	Variation in biomass loss with genotype with estimated 1988 ozone exposure.
Aspen (Populus tremuloides Michx.) clone 259 (A) and clone 271 (B). PRBL
calculated for each 20 km cell based on estimated ozone exposure value
(3-month SUM06) and Weibull parameters for each genotype response
function.
12.	Box-plots of annual area-weighted biomass loss for the 8 tree species with
estimated ozone exposure in 1988 (A) and 1989 (B). The predicted biomass
loss is taken from each 20 km cell in the species' distribution, weighted for the
area and the distribution plotted showing the 10th (bracket), 25th (lower
shaded box), 50th (clear bar in shaded box), 75th (upper shaded box), and
90th (bracket) percentile. The single values outside these percentiles are also
plotted (•). The percentiles represent the area of the species exhibiting that
level or less of biomass loss. OR = Oregon site, Ml = Michigan site, SMNP =
Smoky Mountain National Park, and AL = Alabama.
13.	Effect of ozone exposure on productivity of aspen (Populus tremuloides Michx.)
(A) and loblolly pine (Pinus taeda L.) (B). Predictions of loss were made for
aspen using allometric equations from Cooper (1981) and for loblolly using
equations of Schumacher and Coil (1960). The effect of ozone was
incorporated by multiplying the yearly incremental growth times experimentally
determined reductions in growth from Table 3 in Hogsett et al. (1993). The
reductions were further modified by assuming either a simple interest or
compound interest model for the reduction in growth over years.
15

-------
1.
• • -* v
-.'ii
Ft^vJv^ £L

-------
-O
Q_
Q_
1 00
90
80
70
60
50
40
30
20
1 0
0
Aug. 16, 1988

o"°
/ /T-r\ o
T Q.O \0'° o
/Q'°	\
O-O,
O-O
••••••• ,
• \
J	|	|	1	|	I	|	|	I	L

• Twin Creeks
O Look Rock
~ Cove Mt.
J	L
8 10 12 14 16 18 20 22 24
Time of Day (hrs)
FIGURE f:

-------
		 I	1 '	-v	1	I
Noland Sugarlands Pines Elkmont
Look Rock Sugarlands Hard Twin Creeks
Sample Sites
I | Mid-Canopy W//A 3 Meters EE 0 5
I
FIGURE j]

-------
b
CHANGE IN LEAF COUNT
_l^
CJl
o
I
cn
cn
H
O
Nl
O
m
o
7Q
5
m
o
x
CO cn
x
3
<
<
<
rn
m
0
1
CD
oo
CD

-------
Dry Weight (g) ^ (cm) Leaf Area (dm2)
—• N) K	W Oi	M	CD CD O	NJ
OOOOOO	O O O O OlO N> cn 00 —
o
Diameter (mm)

-------
Figure 8

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
1. REPORT NO. 2.
EPA/600/A-94/073
3. RECIi
4. TITLE AND SUBTITLE
Assessing Ozone Effects on Plants Native
to the Southeastern United States
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7 AiifingRi^ufeld, 2J.R.Renfro, 3S.Huang, 4W. D. Hacker,
5D.Mangis, 6A.H.Chappelka, 7W.E.Hogsett, SA.A.Herstrom,
9J.A.Laurence, 8E.H.Lee, 7J.E.Weber, 7D.T.Tingey
8. PERFORMING ORGANIZATION REPORT NC.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
'Appalachian State Univ.Boone,NC,2Gatlinburg, TN,
*Univ of WY, Laramie , 4TX DOT, Austin, TX,5EPA RTP, NC,
'Auburn Univ,Auburn, AL,7EPA,ERL-Corvallis, OR,
8ManTech,ERL-Corvallis, OR, 'Boyce Inst.,Ithaca,NY
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
US EPA ENVIRONMENTAL RESEARCH LABORATORY
200 SW 35th Street
Coivaila, OR 97333
13. TYPE OF REPORT V»D fie-RlOO COVERED
Symposium Paper
14. SPONSORING AGENCY CODE
£PA/6&0/j z
15. SUPPLEMENTARY NOTES
1993 In: Proceedings of "International Association of Botanical Gardens"
16- R,¥tTe last six years, the U.S. National Park Service, U. S. Environmental
Protection Agency, and University researchers have been documenting the effects
of ozone on a large number of woody and herbaceous species native to the
southeastern United States. In Great Smoky Mountains National Park, (GRSM) , ozone
levels exhibit diel patterns at low elevations, where concentrations are low in
the morning and high in the afternoon. At high elevations (>800 m) , morning
concentrations are much higher, and the total daily exposure is approximately
twice that at the lower elevations. Putative ozone injury has been observed in
the field in GRSM on 90 species, representing approximately 6% of the known flora
in the Park. Surveys of foliar injury on several tree species show a general
pattern of increasing frequency and amount of stipple with increasing elevation
in GRSM, and in nearby Shenandoah National Park. Exposure - response studies were
carried out in opentop chambers for six years with 46 species. Foliar symptoms
seen in the field, were reproduced on 30 species, providing evidence that the
foliar injury found in the field was probably due to- ozone exposure. The
exposure - response studies indicated that early-successional and shade intolerant
species were less resistant to high ozone levels than late-successional species
and dry-site adapted conifers. Geographic informations systems (GIS) techniques,
coupled with growth models, were used to scale-up responses from the seedling to
the geographic level. Because the modeling work is totally dependent on the
quality of the exposure-response data, future work must concentrate on more
realistic exposure systems, genetic variation within species, and refinement of
the modeling procedures used in scaling up.
17 KEV WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDEDTERMS
c. cosati Field/Group
ozone, plants, southwestern U.S.


18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21. NO. OF PAGES
23
20. SECURITY CLASS This pagei
22. PRICE
EPA Form 2220-1 (Rav. 4-77) previous edition is obsolete

-------