What are the effects of elevated tropospheric ozone?

The NAAQS are intended to:

1. Protect public health and

2. Protect public welfare from any "adverse effects"

The latter is particularly problematic, in that both "public welfare" and "adverse effects" are defined only loosely. For example, if ecosystem services are considered important for public welfare (which they surely should be!), how are they quantified, and what degree of injury constitutes adverse effects? This is a problem that EPA struggles with for O3 as well as for other pollutants. There is much debate at present about whether the current NAAQS for O3 is sufficiently stringent (EPA feels that it isn't -- there are, as of 2006, more violations in the US of the NAAQS for O3 than for any other criteria pollutants); we'll explore that later, but below we'll look at adverse effects caused by O3 exposure.


Ozone is unambiguously injurious to human health. Predominant among problems attributable to O3 exposure are respiratory problems. These may become chronic following long exposures -- sad to think about children growing up in high ozone areas (such as LA) and the possibility (some would say likelihood) that they will suffer chronic respiratory problems!

It is estimated that, as of 2004 when the new ozone NAAQS for ozone went into effect, about 159 million people were exposed to levels of ozone that exceed that standard. That is almost half of the US population! Even under the old standard, EPA estimated in 1991 that about 67 mill people were routinely being exposed to O3 at concentrations higher than the NAAQS. In fact, in 1993, and again in 2007, the American Lung Association sued EPA to increase stringency of the standard, which EPA has now done (not only as a result of this suit, as we'll see!).

Ozone exposure also causes headaches and reduced physical performance. In fact, scientists first became aware of O3-related health effects in 1967 in CA, when it was noted that high school athletes had decreased performance on high exposure days. Exposure can also cause reduced resistance to infections.

The standard is supposed to be set to protect even the most sensitive persons -- with a margin of safety -- but the current version of the standard probably doesn't do that.

The first detectable effects often occur at about 0.10 ppm, and prolonged, cumulative exposure seems to be more related to health effects than is the peak hourly concentration. (Recall that in many areas ozone occurs over prolonged peaks of 6-8 hr rather than shorter steeper peaks.) The 0.12 ppm for one hour standard was based on the assumption that peaks are sharp and short, which isn't necessarily so. Thus, the degree of protection afforded by that NAAQS was inadequate for this reason as well.

"Adequate protection" is actually a very sticky issue with regard to ozone and human health. When the Clean Air Act was first written, mandating that standards be set to protect public health, and with an adequate margin of safety, it was assumed that health responses to O3 would show a threshold. That is, there would be no detectable effect up to some certain O3 level, after which effects would become apparent. That threshold would be used to inform the standard. However, it turns out that human health responses to O3 are basically linear, rather than exhibiting a threshold. Linear, and with an intercept at zero -- that is, some people are sensitive to even the lowest background levels of ozone.

SO, how does the EPA set a standard that is protective of human health, other than saying "0 ppm?" This is an example of science making the discovery (that health responses to O3 are linear) but then being unable to make a policy recommendation based on that -- values re human health must be considered, and the decision can't be based simply on the science!


Ozone is hard on certain materials, particularly those based on rubber. It causes accelerated cracking of rubber products, including latex-based house paints.


We will discuss these more when we talk about global climate change and the "greenhouse effect," but for now, recognize please that O3 is a "greenhouse gas," as is CO2. Ozone is very effective at trapping outgoing heat radiation and radiating it back to earth.


Plants are generally more sensitive to O3 than are humans, with damage to some sensitive species occurring at concentrations as low as 0.04 ppm.

Ozone injury can take several forms:

1. Alterations in physiology, particularly decreased rates of photosynthesis in some species and altered carbohydrate allocation patterns. A common alteration is decreased allocation to roots and increased allocation to shoots, which makes the plant more vulnerable to drought.

2. Injury to membranes. Once inside the plant, ozone causes the production of secondary products, such as OH- radicals, organic free radicals, H2O2, and singlet O. All of these are oxidizing agents which appear to be responsible for its toxic effects.

Secondary consequences of ozone exposure then result somehow from #1 or 2, above. These secondary consequences include:

As elaborated below, we know from field and laboratory investigations that ambient levels of ozone in the US (and elsewhere too) are injuring both native and crop plants.

Scientists have known for quite some time -- since 1958 -- that O3 is phytotoxic. As early as 1944, there was a report of visible injury to vegetation following a smog episode in the Los Angeles basin, but it wasn't known what compound or compounds in the smog were injurious. However, in 1958 it was reported that ozone --known to be a part of smog -- caused injury to grape leaves in CA, and the next year it was reported that the mysterious speckling that ruined tobacco leaves in the Southeastern US after some thunder storms was caused by ozone (resulting from stratospheric intrusions). There followed a spate of studies demonstrating that O3 was responsible for injury to many plants, both crops and natives.


Ozone is by far the most important air pollutant in terms of damage to crop plants. Globally, ~ 35% of cereal crops are grown in areas where O3 levels are high enough to decrease yields!! In the US, O3 accounts for 90% of all pollution-induced crop yield losses.

Based on experiments described below and on knowledge of ambient O3 concentrations, the US Office of Technology Assessment (OTA) estimates the following yield reductions attributable to O3, all expressed as averages across major growing areas. Losses are higher in hotspot areas, with yield decreases ranging from 0-56% depending on crop, location and exposure. Average losses (compared to yields expected if O3 were present at natural background levels) across major growing regions for the US are:

Again, losses are variable across the US, depending on O3 concentrations, growing conditions, and crop varieties being planted. For example, certain cotton varieties in areas with 7 hour average concentrations during the growing season of 0.09 ppm (such as portions of southern CA) suffer 62% yield losses, with losses decreasing to 31% at 0.06 ppm. Some varieties of wheat and soybeans are nearly as sensitive as these cotton varieties.

The US Office of Technology Assessment (OTA) estimated that O3-induced yield losses summed to about $3 - $5 Billion per year in 1997. Ozone concentrations can now be better mointored in rural areas because of satellite technology, rather than relying on gound-based sensors. Such monitoring in concert with crop yield assessments indicates that soybean yield losses attributable to ozone exposure alone top $1 billion per year [FEE Oct. 2011]! Such agricultural losses are extremely important and tragic, given what we now know about the ecological degradation associated with trying to feed people!

How are crop loss data like these derived?


Before 1976, no research had been done that provided reliable data relating a range of pollutant exposures to crop yields. Studies prior to that time had numerous weaknesses:

There was, however, a real need to estimate crop losses from air pollutants to provide a basis for setting air quality standards. Thus, the government established NCLAN (National Crop Loss Assessment Network) in 1980 to accomplish this mission. (Funding for NCLAN ended in 1987.) It was funded primarily by the EPA, and was a major, carefully planned program. NCLAN included four major components or goals:

1. To carry out experiments that reliably related doses of pollutants to yields for those crops that are economically important in several major crop growing areas in the US. The idea was to develop dose-response curves that would allow estimation of yield as a function of pollutant dose. (It turns out that there are difficulties with these, because the curve varies for each variety of each crop and with weather and other growing conditions, nevertheless approximations were possible.) (A technicality; there are also problems with determining a biologically meaningful definition of "dose." Dose is usually reported as the product of air concentration X time, but not all of that pollutant actually gets inside the plant tissues. Ideally, one would report "internal dose" but that is very difficult to measure.)

2. To integrate these crop losses over entire regions using the dose-response information gained in the experiments, the acreage devoted to the crop, and the pollutant levels in each county.

3. To assess dollar losses each year from these pollutant effects.

4. To create models that relate yields to level of pollutant, water stress, stage of crop development, and temperature, using results to decide what the NAAQS should be, based on injury thresholds. Such models could be built basically as "black box models," in which the modeler would only know empirically what went in and what came out, then using those numbers in equations. Preferable would be the creation of true mechanistic models, where you understand why what happened happened. These mechanistic models would depend on detailed physiological studies done on each aspect of the interaction, and would allow safer extrapolation beyond the range of the data. However, the amount of research required to build a true mechanistic model is typically much greater than that required to build an empirical "black box" model.

The first aim of NCLAN, carrying out experiments that reliably related dose to yield, received the most funding. Experiments were done at six sites across the US, each representing a major crop growing region: New York; North Carolina; Maryland; Illinois; Livermore, CA; and Riverside, CA. The sites were chosen because they had current or potential air quality problems and also because a group of researchers in air pollution effects were located nearby. You will note that entire areas of the US were excluded from the studies.

NCLAN researchers used a single, standardized approach. Crops were grown in open top chambers to improve the realism of the experiments. Perhaps you have seen the open top chambers behind the EPA lab on 35th Street? Basically, they are large plastic cylinders without tops (or with moveable rain-exclusion tops, depending on the experiment). They are located outdoors, and, for NCLAN, plants were planted in them using agricultural practices that are most common for the crop and the region. Air is provided to the chamber via fans located near the base of the chamber, and pollutant can be added to the incoming air, or filtered out of it (for control chambers) in cases where ambient air contains significant quantities of pollutant. Thus, there is a positive air pressure and flow within the chambers.

Why use open top chambers rather than treating plants with pollutant in a greenhouse? The chambers are more similar to the real world than is a greenhouse; they seem to have little in the way of "chamber effects" associated with them (chamber effects are effects on response that are attributable simply to the chamber rather than to the pollutant).

To test for the existence of chamber effects, NCLAN used chambers with nonfiltered air and compared responses of plants grown in these to responses of plants grown in open air near the chambers.

Researchers would then add O3 to the chambers at levels actually or potentially encountered by crops in the region. Technically, they added a constant incremental amount of O3 to the variable amounts of O3 present in ambient air for 7 hr/day, giving concentration curves that followed the changes over time in ambient concentrations, with curves for different treatments paralleling each other, differing only in the amount of O3 being added over ambient. (The site at Riverside, CA was so polluted that it was necessary to filter the air to achieve the two lowest treatment levels [and the controls of course!]).

I mentioned that O3 was added for 7 hrs/day. These additions took place during the hours of the day when pollutant exposure is normally highest; 9:00 am - 4:00 pm. These are also the hours when plants are most active physiologically.

Researchers also monitored temperature, wind, O3 concentration, concentration of other pollutants (where relevant; in some sites, for example, SO2 was also of concern), precipitation, and so forth within the chambers.

The primary response measured was final marketable yield per plant.


Nearly all the O3 levels above background significantly reduced yields below those that would be found if O3 levels were reduced to normal background levels. As mentioned, the dose-response curves were quite variable depending on the crop variety and growing conditions.

Importantly, yields nearly always differed between charcoal-filtered air (charcoal filtering removes O3) and ambient air, meaning that ambient air even in a region thought of as moderately polluted (such as upstate NY) were injurious. These levels were considerably below the NAAQS of 0.12 ppm for one hour that prevailed at the time these experiments were going on; most of the areas used in NCLAN were well within compliance with that standard.

An important side note is that some of the studies looked at combined effects of two pollutants, such as SO3 and O3. This is important because often they are elevated together. This allowed assessment of potential interactive effects.

Overall, NCLAN represented a nice model of careful science, coordinated among different research groups so that results could be fully intercompared and experiments could be replicated. It is, however, impossible to test every possible combination of crop varieties, growing conditions, and ozone doses. It is difficult enough to develop a dose-response relationship for one crop (given interactions with other growing conditions), much less for all combinations of crops, growing conditions, and associated pollutants. At some point, scientists and policy makers will simply have to be satisfied that ozone at levels lower than the current standard is having injurious effects on crops. The problems are what constitutes "sufficient proof" and what O3 level is "sufficiently protective of public welfare?"

Questions also still remain about what numerical formulation the NAAQS standard for O3 should take: should it be based on a 1-hr high as it was in the past, or should it be integrated over a longer period of time (e.g., 8 hours as is presently the case)? EPA's new standard tries to accommodate information on O3 effects on plants, by lowering the maximum concentration allowed and using a longer period of integration, but there are still questions about whether or not it is sufficiently protective (more on that later.)


Ozone also has effects on plants in natural ecosystems, as you might expect. These effects often ramify throughout the entire ecosystem, given that plants are the producers and often the dominant organisms in many ecosystems. These effects occur in many regions of the country where O3 levels are well below the current NAAQS for O3.

Tree growth rates are often slowed by O3 exposure, with some species being more sensitive than others. Reduced growth rates are observed even in areas like N. Michigan and upstate NY, where O3 levels are well below even the old, relatively lax NAAQS. For example, in N. Michigan, responses as extreme as 20% decreases in aspen growth have been documented (without any visible foliar injury). Effects on trees are studied in various ways, often through field observations leading to generation of hypotheses, followed with experiments in controlled environments.


Observations of problems in ponderosa and Jeffrey pine in the San Bernardino/San Gabriel Mts. began in the late 1950's and early 60's. Problems involving loss of vigor, yellowing foliage, needle dropping, and reduced growth were noted as far as 120 km from the Los Angeles Basin, and were initially called "x-disease." While distant from LA, these areas receive abundant O3 (and O3 precursors) from the LA Basin. In the Los Angeles National Forest over 20% of these trees were reported to be injured, and in the San Bernadino National Forest alone, over 100,000 acres were reported to be affected as early as 1970.

The problem has continued, and is now observed to a lesser degree as far north as Sequoia and Yosemite National Parks, resulting from transport of O3 and its precursors. The forests have many sick and dying trees, and there is evidence that ponderosa and Jeffrey pine will diminish in importance in these forests to be replaced by other, more O3-tolerant species. With shifts in dominant tree species are likely to come shifts in properties of the entire ecosystem. (For example, increased fir and incense cedar in the understory are likely to change many aspects of the ecosystem; they provide habitat for different species, influence nutrient, light and water regimes differently than do the pines, and so forth.)

Evidence that ozone is causal?

(1) The spatial pattern of injury coincides with O3 exposure. Injury increases with elevation (as does O3) and is worst on west-facing slopes, which are directly in the path of O3-laden winds from Los Angeles). There is also a sharp west-to-east geographic gradient in injury. In the westernmost regions, growth of ponderosa pine is down by as much as 50% and mortality over 1973-1978 reached 10%.

(2) No "natural" causes seemed to match spatially or temporally – drought, disease, etc.

(3) The temporal pattern coincided with the growth of precursor emissions from Los Angeles (that is, problems with the pines were noticed only after LA's production of pollutants increased greatly)

(4) Visible symptoms on pine needles match those produced in lab by controlled fumigations with O3, including chlorotic mottle, tip necrosis, and premature senescence.

(5) Species known from laboratory work to be most sensitive to O3 are declining the most, including ponderosa and Jeffrey pine.

(6) Known physiological mechanisms are capable of producing the observed effects (disruptions of photosynthesis and altered carbohydrate allocation patterns).

Thus, all criteria needed to establish causation for air pollution injury are actually met in this situation (a rare case when all criteria can be met!) You might call this a "text book example...".

What is actually killing many of the trees is bark beetles (western pine bark beetles), who are able to attack the O3-weakend trees. That is, beetles are the proximate cause of death, while O3 is ultimate (or is the ultimate factor high population density and use of fossil fuels??). In addition, the trees' weakened roots are vulnerable to attack by root rotting fungi which can cause death (recall that O3 decreases plant allocation of carbohydrate to roots...).

Click "ozone regulation" to jump to information on that topic. To return to the index of information on tropospheric ozone pollution, click "Tropospheric ozone," or to return to the master Table of Contents for these BI 301 note, click "Contents." Click "Navigate " for reminders on how to move about within and among these pages.

Page maintained by Patricia Muir at Oregon State University. Last updated Dec 4, 2012