Page 20
4
Current Capabilities for Observing Global Air Quality Changes
As discussed in Chapter 1 , improving our understanding of issues such as climate/chemistry interactions and intercontinental pollution transport will require a comprehensive research framework that integrates observations covering a wide range of temporal and spatial scales with modeling and process studies. Some components of this research framework are in a more mature state than others. For instance, significant progress is being made in the development of regional/global chemical transport models and their integration with global climate models. Likewise, in recent years the atmospheric chemistry community has organized many successful field campaigns that combine model analyses with intensive observations from a variety of platforms. These studies have enhanced our understanding of many complex chemical and radiative processes that play a role in global air quality change.
In contrast, current observational programs are not adequate to detect many of the important medium- and long-term atmospheric chemical changes discussed in the previous chapters. This weakness greatly limits our ability to document the evolution of the atmosphere in the coming decades. If improvements in observational capabilities are not made, this record could be lost forever. It also limits the value of the developments cited above, since a strong observational base is needed to test and improve model predictions, and to provide a longer-term context for the observational “snapshots” obtained through intensive field campaigns.
Current tropospheric observational programs include global and regional networks designed to measure background atmospheric composition at selected remote sites; regulatory monitoring networks that analyze day-to-day air quality
Page 21
changes at numerous sites located primarily in urban areas; remote-sensing (satellite) instruments that provide global-scale observations of selected atmospheric species; and a variety of balloon- and aircraft-borne instruments used for in situ measurement campaigns. These different observational platforms vary widely in their scope and degree of analytical rigor. The following sections describe the capabilities and limitations of these different observational systems for addressing the global air quality issues discussed in this report. Gas-phase species and aerosols are discussed separately since each presents unique observational challenges.
Gas-phase Species: In Situ Observations
The temporal and spatial sampling requirements for observations of a gas-phase chemical species are closely linked to the atmospheric lifetime of that species: the required geographical density of sampling locations increases as the atmospheric lifetime of the species decreases. The necessary sampling density also increases with the need for higher levels of measurement precision, a requirement that is often related to the nature of the individual measurement sites. For instance, the more a sampling site is affected by local sources or local meteorology (e.g., upslope/downslope diurnal winds, passing fronts, land/sea breezes), the more frequent the sampling must be in order to resolve these effects. At many sites, hourly or more frequent sampling is often required to resolve measurable changes from background values for some species.
There are a few networks of surface observational sites that are far removed from local sources and sinks and thus can be used to obtain a regionally representative “baseline” of atmospheric composition. This includes U.S.-supported networks operated by the National Oceanic and Atmospheric Administration's Climate Monitoring and Diagnostics Laboratory (NOAA/CMDL) and by the Advanced Global Atmospheric Gases Experiment (AGAGE), as well as a number of non-U.S. national and regional efforts (e.g., in Australia, the European Community, Japan, and New Zealand). These networks employ in situ automated measurements as well as flask sampling, which permits centralized analyses of samples from different sites and helps verify the relative calibrations of the in situ measurements.
These types of networks have generally proven successful in characterizing the overall distribution and temporal trends of long-lived gases including CO2, CO, CH4, N2O, and CFCs 1 . In contrast, the atmospheric concentrations of shorter-lived (i.e., more reactive) species including NOx, VOCs, O3, and PM exhibit large spatial and temporal variations, and hence measurements at select baseline sites around the globe are generally far too sparse to characterize their overall
1 Quantifying regional sources/sinks of these compounds requires more intensive measurement strategies, but this issue is beyond the scope of this report.
Page 22
distribution and temporal trends. Integrating the observations from different networks could provide fairly dense coverage in some regions, but this would require substantially improved cooperation in matters such as calibration. Independent calibration standards are made by only a few research laboratories (e.g., AGAGE/SIO, NOAA/CMDL, Tohoku/Nippon Sanso), and comparisons of the calibration standards among these different laboratories have revealed significant disagreements for several important chemical species. This issue needs to be addressed in order to make optimal use of existing networks.
In the United States and many other countries, air quality monitoring networks provide extensive ambient measurements of pollutant species. 2 However, there are at present serious limitations to the use of these data in the study of large-scale atmospheric chemical changes. Most air quality monitoring sites are focused on heavily populated areas, and are designed simply to determine whether or not the area is in compliance with current air quality standards. There are very few air quality monitoring sites that allow meaningful examination of long-range transport and trends in background concentrations. Ozone trends in particular are difficult to assess because changes due to anthropogenic forcing are often confounded by natural variations. Finally, techniques used to monitor ozone precursors, NOx and VOCs, do not yet have sufficient sensitivity or accuracy to contribute to trend analysis (Demerjian, 2000; NARSTO, 2000; Fiore, 1998; Wolff et al., 2001; Porter et al., 2001).
These types of urban and regional air quality monitoring networks rely entirely upon ground-based sites that sample within the boundary layer (the lowermost atmosphere). Addressing the global air quality issues highlighted in this report, however, will require observations that extend over a larger altitude range. For instance, the transport of pollution between Asia and the United States occurs primarily through the middle and upper troposphere, and because of the highly episodic nature of this transport, there can be significant inhomogeneity in the air masses reaching the continental United States. Thus, ground-based networks that only sample air masses within the boundary layer would not allow a quantitative determination of long-range pollutant fluxes. Likewise, the radiative forcing exerted by ozone and aerosols is highly sensitive to altitude, and thus it is important to have the capability to observe changes in chemistry and climate over a broad altitude range.
There are a variety of techniques to measure vertical profiles of atmospheric composition. Ozone profiles are commonly measured with ozonesondes, lightweight instruments carried aloft on balloons (the most common form is an elec-
2 A report entitled “The Role of Monitoring Networks in the Management of the Nation's Air Quality" by the U.S. National Science and Technology Council provides a detailed summary of the different air quality and deposition monitoring networks operating in the United States. This report is available at: http://www.nnic.noaa.gov/CENR/cenr.html
Page 23
Box 4-1Meteorological Data Needs
Meteorological and dynamical processes operating over a wide range of spatial scales play a central role in the air quality changes discussed in this report. As described earlier, the emissions, chemical transformations, and deposition of air pollutants are affected by parameters such as temperature, cloud cover, humidity, and wind speed and direction. In addition, dynamical forces related to vertical temperature profiles control the dispersion of pollution within the urban/local boundary layer and the release of pollutants from the boundary layer into the free troposphere. Finally, the long-range transport of pollutants is influenced by atmospheric high- and low-pressure systems occurring on scales of hundreds to thousands of kilometers. Thus, the meteorological context of atmospheric chemical measurements must be established in order to accurately assess implications for air quality and to develop the models used for studying long-range pollution transport and climate-air quality linkages |
trochemical concentration cell). In a comprehensive analysis of the various ozonesonde observations carried out worldwide, it was found that these data can be valuable for developing a global climatology of tropospheric ozone and for testing 3-D tropospheric chemistry models. However, it was also found that there are currently insufficient data for large regions of the earth, particularly in the tropics and subtropics (Logan, 1999). Ozone profile measurements can also be made with differential absorption lidar, a technology that can potentially be implemented on ground-based, airborne, and spaceborne platforms. Lidar provides high-resolution vertical profiles and thus may be particularly useful for identification of pollution plumes aloft. Lidar systems are also being developed for measurements of water vapor, aerosol, and cloud profiles.
Finally, aircraft provide a valuable platform for studying atmospheric chemical composition over a wide range of altitudes. The use of aircraft equipped with a large array of analytical instruments are becoming common in intensive field campaigns for measuring a comprehensive suite of trace gases and reactive species in the upper troposphere and lower stratosphere. More routine aircraft-based sampling is being carried out by the European MOZAIC program (Measurement of OZone by Airbus In-service airCraft), where measurements of ozone, water vapor, and temperature are collected by automated instruments mounted on passenger airliners (Marenco et al., 1998; Thouret et al., 2000). This approach of using sensors onboard commercial aircraft holds great potential as a tool for obtaining vertical profiles near airports and frequent measurements of atmospheric composition at flight altitudes. In order to play a major role in studies of global air quality change, however, it will be necessary to expand the geographic and temporal coverage of these types of measurements. Also, as
Page 24
with the surface-based observations, developing more uniform standards for calibration and quality control will greatly facilitate efforts to integrate the observations obtained from different measurement programs.
Gas-phase Species: Remote Sensing (Satellite3) Observations
A considerable effort has been mounted to develop satellite remote sensing techniques for making global measurements in the troposphere. There are now about 20 instruments in orbit or under development that have substantial tropospheric monitoring capabilities. Table 4-1 lists some of these instruments, with an emphasis on nearer-term measurements. Measurements of tropospheric O3, H2O, NO2, CO, CH4, and HCHO have been demonstrated for potential future use, and there have also been significant efforts to extract tropospheric information from instruments currently in operation, for instance, extracting tropospheric ozone information from the Total Ozone Mapping Spectrometer (TOMS).
Information on the chemical state of the troposphere from satellite measurements is currently very limited because tropospheric measurements from space are inherently difficult. For example, nadir-looking measurements must be able to discriminate between stratospheric and tropospheric contributions of the species of interest. Limb-sounding measurements are often susceptible to interference from clouds and aerosols, and in some cases from H2O absorption or emission. Another limitation is that most current and planned instruments are on polar orbiting platforms, which can only observe any particular location once every few days. In contrast, geostationary satellites allow instruments to observe one location for an extended period of time and thus would generally be more useful for studying pollution transport patterns.
Satellite instruments that could potentially be used for determination of long-term trends in tropospheric chemistry include the Global Ozone Monitoring Experiment (GOME) instruments on the European Space Agency's ERS-2 satellite and the Eumetsat Metop satellites (ca. 1995-2012 time series); the Infrared Atmospheric Sounding Interferometer (IASI) instruments on the Metop satellites (ca. 2006-2012 time series); the Ozone Mapping and Profiler Suite (OMPS) instruments on the U.S. NPOESS satellites (ca. 2008-2017 time series); and potentially, the U.S. TOMS instruments (1978 and continuing time series). Although these instruments are not optimized for tropospheric measurements, it is worthwhile to plan strategies for optimizing the interpretation of the collected tropospheric data and integrating these data with shorter-term satellite measurement programs (e.g., Earth Observing System).
3 Most all of the current and planned remote sensing measurement programs reviewed here are satellite-based. It should be noted however, that the potential exists for using other remote sensing platforms such as the space shuttle.
Page 25
Although much development work is still needed, the possibilities for using satellite-based measurements to obtain trend-quality tropospheric measurements of species including O3, NO2, and HCHO should eventually be realized. However, major challenges will need to be addressed as tropospheric measurements from satellites move from initial characterization to trend determination. Monitoring tropospheric trends using multiple instruments (often operated by different countries) will require tying together measurement records with intensive radiative transfer modeling, photochemical and transport modeling, and instrument characterization. Extensive ground-based and other in situ measurements will continue to be a critical component of remote sensing observational systems, as they are needed for evaluating and interpreting satellite data.
Aerosols: In Situ Observations
Particulate matter presents a “multi-dimensional” observational challenge, because it is important to characterize not only total PM concentration, but also size, composition, mass, and phase; and these parameters can evolve as the air mass containing the aerosol undergoes changes in humidity and temperature and exposure to new condensable emissions. Real-time measurements of particle size are possible but routine measurements of aerosol mass and chemical composition rely primarily on the accumulation of particles over extended sampling times with subsequent laboratory analyses (although techniques for real-time chemical analysis of aerosols are rapidly developing4).
In situ measurements can provide useful information about which aerosol types predominate in a given region, but there are serious gaps in existing aerosol observational programs. For example, the NOAA/CMDL sites are gathering useful aerosol optical data, but these measurements are very limited in time and space. Urban areas in the United States are monitored as part of the EPA's air quality regulatory framework (Demerjian, 2000), but large spatial gaps exist in many rural areas, where background aerosol plays a major role in determining local air quality. The IMPROVE network (Interagency Monitoring of Protected Visual Environments; Malm et al., 1994), collaboratively operated by several U.S. federal agencies, obtains information on aerosols in national parks and wilderness areas, and has recently been expanding to fill some of the observational gaps in sparsely populated areas. These data, however, cannot yet yield a picture of global or even regional aerosol distributions. There are opportunities for obtaining more complete information through better integration of existing measurement programs. For example, aerosol optical depths are measured by sun photometers at a number of sites (Aerosol Robotic Network-AERONET;
4 A special issue of the journal Aerosol Science and Technology (v33, 2000) provides a detailed overview of real-time single-particle analysis techniques.
Page 26
Sensors |
TOMS a |
GOME |
MOPITT |
MODIS |
SCIAMACHY |
MIPAS |
TES b |
HRDLS |
OMI a |
MLS |
SAGE III |
TRIANA |
Launch |
1978 |
1995 |
1999 |
1999 |
2001 |
2001 |
2002 |
2002 |
2003 |
2002 |
2003 |
2002 |
O3 |
column |
column+ |
column+ Îz=3-4km limb |
z>5 km, limb |
Îz=2/4km limb/nadir |
Îz=1km (UT) |
column |
UT |
Îz=1km (UT) |
column |
||
H2O |
column |
column+ |
z>5 km, limb |
Îz=2/4km limb/nadir |
UT |
Îz=1km (UT) |
||||||
CO |
3-4 levels |
column+ Îz=3-4km limb |
z>5 km, limb |
Îz=2/4km limb/nadir |
UT |
|||||||
NO |
tropical UT Îz=2km |
|||||||||||
NO2 |
column |
column+ |
column |
Îz=1km (UT) |
||||||||
HNO3 |
UT |
Îz=2km (UT) |
Îz=1km (UT) |
|||||||||
CH4 |
column |
column+ Îz=3-4km limb |
column |
|||||||||
HCHO |
column |
column+ Îz=3-4km limb |
column |
|||||||||
SO2 |
column |
column |
column |
column |
column |
|||||||
BrO |
column |
column+ Îz=3-4km limb |
column |
|||||||||
Aerosol |
column |
column |
column |
column/profiles |
column |
Îz=1km |
column |
a TOMS has been in operation since 1978. Last launch was in 1996 and data continues to be collected at this time. OMI will take over some TOMS functions in year 2003. b A number of additional derived chemical products such as acetone, methanol, H2O2, HCRN, NH3, HNO4, SO2, and PAN are possible.
Definitions: UT = upper troposphere; Îz is to the instrument vertical resolution; ‘column' is the vertically integrated abundance of a species; ‘column +' means that some information on vertical distribution is provided in addition to total column abundance; ‘limb' and ‘nadir' refer to observations made (respectively) along horizontal and vertical viewing paths.
Holben et al., 1998), but few are co-located with chemical composition measurements.
Measuring vertical aerosol distributions is also critically important. Models that can reproduce surface aerosol observations quite well often differ significantly in their estimates of total column burden, and therefore the transport, of aerosols. Sporadic data from airborne field campaigns have been used to help understand the vertical distribution of aerosols, and balloon-borne measurement programs have provided valuable information on mid- and upper-tropospheric
Page 27
NOTE: Table on previous page.
aerosol size distributions (Hofmann, 1998). Lidars and lidar-in-space technology also offer new opportunities for filling this important observational need. At present, however, there is no clear plan for routine observations of aerosol vertical distributions.
As with gas-phase observations, obtaining accurate and precise aerosol measurements on a global basis requires the development of uniform, primary standards for calibration. Neither primary standards nor standard operating procedures are readily available for many research instruments. This lack of
Page 28
consistency in operational measures presents a major obstacle to integrating the observations from different programs.
There are few long-term records of tropospheric aerosols, making it difficult to identify trends and to relate trends to anthropogenic activities. Even though some monitoring programs have been established over the last several decades, it can be difficult to interpret trends at individual sites when changes in the background air reaching that location are not characterized. For example, sulfate levels near coastal sites can be affected by ocean fluxes of reduced sulfur compounds, transport of natural precursors and aerosols, and anthropogenic emissions; and these different sources cannot be readily distinguished or quantified.
Surface observations of sulfate aerosols in remote regions are currently very difficult to reproduce in models. The agreement of model predictions with observations is much worse for other aerosol types, such as black carbon and organic species, for which we have fewer observations and less fully characterized atmospheric transformation and removal processes. Progress in improving these models is hindered by the lack of vertically-resolved aerosol data and by significant inaccuracies and poor spatial resolution in emissions inventories, particularly of biogenic and other organic emissions that condense in the atmosphere to form PM.
Aerosols: Remote Sensing (Satellite) Observations
Current satellite instruments (pre-1999) were not designed specifically to detect tropospheric aerosols, however they can provide a qualitative picture of the atmospheric distribution of some aerosol types. For instance, data from the GOES-8 satellite have been used for detection of smoke/haze from fires (e.g., Prins et al., 1998). Maps of “aerosol optical thickness” (AOT) over the oceans, showing widespread plumes attributed to dust, biomass burning, and pollution aerosols, have been produced using data from the polar-orbiting AVHRR instrument (e.g., Husar et al., 1997). The TOMS instrument has been shown to be sensitive to absorbing aerosols such as dust and smoke (Herman et al., 1997) and has been used to investigate long-range transport of African dust. By combining these data with observations from ground-based monitoring networks, much has been learned about global sources and transport of some important aerosol types. However, it is difficult to obtain quantitative information on aerosol type and mass flux from such images, largely because of the complex and poorly characterized radiative properties of most aerosols.
Recognition of the important role played by aerosols in climate forcing and in air quality has led to significant international support for satellite missions aimed at improving current observational capabilities. Newly-launched and proposed instruments (e.g., MODIS, MISR, and PICASSO-CENA) have been specifically designed for detection of aerosols and clouds (King et al., 2000). Nevertheless, it remains difficult to retrieve aerosol properties over land, particularly
Page 29
for regions with low AOT; and unfortunately, these regions are of particular interest because they are most susceptible to anthropogenically-driven changes in aerosol loadings. Tropospheric aerosol determination is also limited by our incomplete understanding of aerosol light scattering physics.
Validation plans for MODIS and MISR include intercomparison with ground-based monitoring networks and intensive campaigns involving measurements from aircraft and other platforms. The international AERONET network of sunphotometers (Holben et al., 1998) plays a prominent role in evaluating the accuracy of AOT data. The vertical profiles of aerosol properties produced by the PICASSO-CENA lidar measurements will also need to be validated against ground-based and airborne in situ measurements. The value to these satellite missions of data from a long-term, ground-based network measuring aerosol chemical composition and physical properties is apparent. The network data help place the satellite measurements in context by defining characteristics such as typical aerosol loadings for particular locations. Ground-based observations can also provide information that cannot yet be obtained from space-borne instruments, such as the chemical composition (and hence the sources) of particulate matter. And finally, since these aerosol-specific satellite missions have a relatively short lifetime and thus limited applicability for trend analyses, continuous ground-based monitoring data are required.
Careful consideration needs to be given to making optimal use of the aerosol data being provided by current and planned satellite missions. Aerosol optical thickness, a common satellite data product, has limited utility as input to models for validation purposes; it is likely that combining satellite information with routine and intensive surface observations will be more useful. Continued investigations into coordination of efforts—for example, co-locating satellite validation monitoring sites with existing observational network sites, and ensuring continuity of the longer-term network data—should be undertaken and supported.
An Integrated Research Strategy
It is clear that our understanding of global air quality would be greatly aided by strengthening the infrastructure for obtaining high-quality atmospheric chemical measurements from ground-based stations and platforms such as balloon sondes and aircraft. However, in situ observations alone cannot provide sufficient spatial coverage to assess the global distribution and trends of short-lived gases and aerosols. Ideally, remote sensing instruments could provide global coverage, but the capability of satellites to provide accurate measurements of global atmospheric composition and long-term trends for tropospheric species lies years in the future. Even when comprehensive satellite observations do become available, correlative in situ measurements will still be needed to test the remote sensor inversion algorithms. The most feasible strategy then, is to use
Page 30
Box 4-2Key Factors in Maintaining High Quality Observations
Over the past several decades, beginning with the pioneering studies of atmospheric CO2, a great deal has been learned about what is required to create and sustain effective, reliable networks for observing the chemical state of the atmosphere. Some of the key factors include the following:
|
Page 31
the limited available remote sensing data in combination with measurements from the range of available in situ platforms.
These integrated measurement programs form a core component of the overall research framework needed for the study of global air quality change. Models comprise another core component of this research framework, since they are needed to fulfill at least three important roles: (i) helping to determine the optimal locations for long-term observational sites and short-term process studies, in order to maximize the usefulness of the data collected; (ii) assimilating the observations acquired from different platforms and helping to place isolated measurements in a larger context; and (iii) providing a prognostic capability for predicting future air quality trends, for estimating transboundary pollution fluxes, and for assessing the impacts of coupling between climate and air quality changes. Fulfilling all of these roles will require careful integration of various types of modeling tools. For instance, process models can be embedded in regional models to provide detailed representation of certain critical processes; and likewise, regional models can be embedded in global models to expand the spatial and temporal resolution in areas of particular interest.
Coordinating measurements among various remote sensing and in situ platforms, and integrating these measurements with detailed modeling studies presents an immensely complex research challenge. This approach is being developed in focused process studies such as the Southern Hemisphere ADditional OZonesondes project (SHADOZ, which coordinates a network of balloon-borne ozonesondes with TOMS remote sensing measurements), and the INDOEX, TRACE-P, and ACE-Asia field campaigns (studies aimed at understanding the impacts of pollution outflow from the Asian continent). This level of integration, however, does not yet exist in an ongoing, operational sense or extend over the range of scales needed to accomplish the research objectives highlighted in this report.