The last step in collective interpretation of SOS scientific findings is development of a series of general lessons that are available to be learned from SOS' research, assessment, outreach, and extension efforts during the past 17 years. Major differences in air-quality have been found among geographical regions within the continental United States. These important regional differences include differences in:

Most of these regional differences are related to differences in meteorology, biogenic emissions, and/or industrial emissions. For these reasons, selection and implementation of optimum ozone and PM management strategies in any region requires comprehensive understanding of- (or specific studies to determine) the chemical climatology, biogenic and industrial emissions patterns, and the meteorological climatology of the region.

LAL1. Observation-Based Models and methods for understanding ozone and PM pollution problems are very effective - and, in many cases, they are also more reliable than currently available Emissions Based Models and methods - especially for determining the relative efficiency and cost effectiveness of VOC and NOx controls in decreasing ambient ozone exposures.

LAL2. The databases produced during SOS field measurement campaigns are very useful for comprehensive and reliable evaluation of Emissions Based Models, and, also, for supporting research on new or emerging scientific issues. This is true for two principal reasons:

LAL3. In cases of unusually large spatial and temporal variability of emissions, it is preferable that day-specific and location-specific emission inventory data be used in Emissions-Based Models of ozone and particulate matter pollution problems - especially in selecting optimal emission-control strategies of ozone and/or PM.

LAL4. Many ozone and PM exceedances in urban non-attainment areas are not caused by the non-attainment area's emissions alone. Contributions of precursors from the surrounding region (and at least occasionally from far-distant regions such as wild fires in Canada, Mexico, Louisiana, Idaho, or other places) can also be important. Thus, to achieve attainment in many non-attainment areas, emission controls may have to be extended to include regional background sources outside the borders of the designated non-attainment area.

LAL5. Sources of regional ozone and PM pollution are mainly urban plumes traversing the region and regional emission sources (the free troposphere is another possible source but is relatively rare and usually of minor importance). Control of regional ozone or PM pollution, therefore, may require control of both urban sources and regional sources. Furthermore, and more importantly, the optimum control approach for local and the regional sources may have to be different - that is, VOC control for local source and NOx control for regional sources, or vice versa, that is, NOx control for certain local sources and VOC controls for regional sources.

LAL6. Control of NOx from point sources should be viewed with great caution. If ozone and/or PM formation conditions in the area impacted by an NOx source are VOC-limited, then control of the NOx sources may cause increased concentrations of ozone and/or PM in that area. This sometimes unexpected phenomenon is often referred to as "NOx disbenefit."

LAL7. Ozone accumulation is almost always VOC-limited in urban-core areas of large cities (population greater than about one million people), almost invariably NOx-limited in rural areas, and certainly less VOC-limited and more NOx-limited in downwind suburban areas.

LAL8. Biogenic VOC emissions are extremely important in that they tend to create NOx-limited conditions under which the effectiveness of VOC controls for decreasing ozone exposures is diminished. NOx-limited conditions also are created within "aged" urban plumes (due to rapid depletion of NOx), and by unusually large (or unusually reactive) VOC emissions from industrial sources such as the Houston Ship Channel adjacent to the Houston, Texas metropolitan area.

LAL9. Even when modest amounts of biogenic VOC are being emitted, for example in city parks, the remarkably high reactivity of isoprene and other biogenic VOC makes it imperative that these emissions be taken into account in using Emissions-Based Models for State Implementation Plans (SIPs). This raises the important requirement that reliable biogenic VOC measurements be made in essentially all ozone and PM non-attainment areas. Such measurements are not simple. They should include detailed information on land use, vegetation species composition, and emission-rate determinations for each important species of vegetation as a function of ambient temperature and wind speed conditions that affect biogenic VOC emission rates. Also, calculated biogenic emission estimates should be checked against direct measurements of biogenic loading rates (see next item, below).

LAL10. Given the extremely fast decreases of biogenic VOC concentrations in ambient air (caused by rapid chemical reactions with peroxy radicals (OH) and ozone in air, the most reliable method for measuring atmospheric loading rates of biogenic VOC is radiocarbon (14C) measurement applied on both gas-phase and aerosol-filter samples. This method produces estimates of both biogenic and anthropogenic VOC in air.

LAL11. Well-fertilized crop and pasture lands are significant "semi-anthropogenic" sources of NOx emissions in rural areas. These biogenic sources of NOx emissions in rural areas can be as large as (and sometimes even exceed) the NOx emissions from motor vehicles and combustion sources in rural areas.

LAL12. In emission trade-off deliberations, consideration should be given to the fact that ozone production potentials of VOC and NOx emissions vary, depending on the environment within which the emissions are discharged (for example, ozone production efficiency of NOx is much greater in regions with large vs small total emissions of biogenic VOC) and on other factors (for example, the size of the NOx source).

LAL13. Ozone production efficiency per unit of NOx emissions from power plants often is inversely proportional to the magnitude of the NOx source. Thus, smaller power plants produce more ozone per unit of NOx emitted than large power plants.

LAL14. The photochemical processes that lead to formation of the secondary aerosol fraction of ambient PM2.5 from SO2, NOx, and VOC emissions are very similar to the photochemical processes that lead to ozone formation from NOx and VOC in air. For this reason, the concentrations of secondary aerosols in PM2.5 non-attainment areas, as in the case of ozone in ozone non-attainment areas, depend more on the ratios among the reactants and their relative reactivity than they do on the amounts of these three reactants in the non-attainment atmosphere. Unfortunately, however, the optimum NOx:VOC ratio for decreasing ozone concentrations in ambient air are not always the same as those for decreasing secondary aerosol concentrations in ambient air. Thus, optimum control strategies designed to decrease secondary aerosol formation in PM2.5 may make the ozone problem worse. Air quality managers need to be aware of these interactions and possible interferences between ozone-control strategies and secondary-aerosol-control strategies.

LAL15. A significant part of PM2.5 is ammonium sulfate and nitrate, and may include also semi-volatile organic compounds. This has two implications. First, because of partial on-filter volatilization of the ammonium nitrate and semi-volatile organic components of PM2.5, care should be exercised in preserving the integrity of filter samples. Second, besides SO2, NOx, and VOC, NH3 also is a significant PM2.5 precursor, the emissions of which need to be measured and considered in developing optimum PM2.5 control strategies.