Cox

Atmospheric Environment 80 (2013) 584e590 Contents lists available at ScienceDirect Atmospheric Environment journal ho...

0 downloads 92 Views 283KB Size
Atmospheric Environment 80 (2013) 584e590

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Toward the next generation of air quality monitoring: Particulate Matter Jill Engel-Cox a, *, Nguyen Thi Kim Oanh b, Aaron van Donkelaar c, Randall V. Martin c, e, Erica Zell d a

Battelle Memorial Institute, 505 King Ave, Columbus, OH 43201, USA Asian Institute of Technology, P.O. Box 4, Klong Luang, Pathumthani 12120, Thailand Dalhousie University, Halifax, Nova Scotia B3H 4R2, Canada d Battelle Memorial Institute, 2111 Wilson Boulevard, Suite 900, Arlington, VA 22201, USA e Harvard-Smithsonian Center for Astrophysics, Cambridge, MA 02138, USA b c

h i g h l i g h t s  Summary of particulate matter (PM) air pollution monitors, models, and indicators.  Data variability makes global comparison of PM concentrations difficult.  Cheaper, more durable, personal, crowd-sourced PM monitoring technologies needed.  Improved data sharing, standards, models needed for global indicators.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 February 2013 Received in revised form 4 August 2013 Accepted 7 August 2013

Fine particulate matter is one of the key global pollutants affecting human health. Satellite and groundbased monitoring technologies as well as chemical transport models have advanced significantly in the past 50 years, enabling improved understanding of the sources of fine particles, their chemical composition, and their effect on human and environmental health. The ability of air pollution to travel across country and geographic boundaries makes particulate matter a global problem. However, the variability in monitoring technologies and programs and poor data availability make global comparison difficult. This paper summarizes fine particle monitoring, models that integrate ground-based and satellite-based data, and communications, then recommends steps for policymakers and scientists to take to expand and improve local and global indicators of particulate matter air pollution. One of the key set of recommendations to improving global indicators is to improve data collection by basing particulate matter monitoring design and stakeholder communications on the individual country, its priorities, and its level of development, while at the same time creating global data standards for inter-country comparisons. When there are good national networks that produce consistent quality data that is shared openly, they serve as the foundation for better global understanding through data analysis, modeling, health impact studies, and communication. Additionally, new technologies and systems should be developed to expand personal air quality monitoring and participation of non-specialists in crowdsourced data collections. Finally, support to the development and improvement of global multipollutant indicators of the health and economic effects of air pollution is essential to addressing improvement of air quality around the world. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Particulate matter Aerosols Monitoring Policy Indicators

1. Background and objective Over the past century, atmospheric scientists and environmental regulators have focused on particulate matter (PM) as one of the

* Corresponding author. Tel.: þ1 614 424 4946; fax: þ1 614 458 4946. E-mail address: [email protected] (J. Engel-Cox). 1352-2310/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atmosenv.2013.08.016

major areas of air pollution study and pollution control. PM includes both primary particles such as soot and dust from combustion sources and agricultural practices, and secondary particles such as sulfate and nitrate that form though chemical reactions in the atmosphere from

J. Engel-Cox et al. / Atmospheric Environment 80 (2013) 584e590

sulfur dioxide, nitrogen oxide, and ammonia emitted from power plants, industries, automobiles, and agriculture. The current regulatory focus is on subsets of fine particulate matter, specifically PM10 and PM2.5, particles less than 10 microns and 2.5 microns in diameter, respectively. Epidemiologic research on long-term exposure to ambient fine particulate air pollution has documented serious adverse health effects, including increased mortality from chronic cardiovascular and respiratory disease, lung cancer, and adverse reproductive outcomes, with outdoor PM2.5 estimates at causing approximately 3.1% of all disability adjusted life years worldwide (e.g., Pope et al., 2009; Lim et al., 2012). Increasingly, researchers are finding that PM chemical composition is a significant variable in its health impacts, but supporting datasets are limited (Lippmann, 2012). Regulation and control of emissions has been enabled by advancements in PM monitoring and modeling. The ability to separate ambient PM levels into different particle sizes as well as to speciate the chemical components were major developments in monitoring technology that serve as the regulatory foundation today. These data have also been used to support chemical transport models (CTMs) that describe the formation of secondary particles and help estimate and forecast PM concentrations based on known emissions and meteorological conditions. In the past 15 years, remote sensors on satellites have expanded understanding of the spatial distribution and movement of PM, primarily by calculation of aerosol optical depth (AOD), which can serve as a surrogate for tropospheric pollution, as summarized by Hoff and Christopher (2009). Air pollution information is communicated to the public and decisionmakers through air quality indices. These are typically based on health studies for both long-term (annual) and short-term (daily) exposure, converted from concentration to a simple unitless numerical scale and color-coded for visualization. The public may use these data to modify their behavior to reduce exposure to pollution and the regulatory decisionmakers use them to make changes in regulatory controls. Exceedance of national standards is the most common indicator communicated to policymakers. While there have been considerable advancements, challenges remain: PM monitoring technologies and models require experienced users, satellite and ground-based data measure related but different phenomena, and data collection and indicators are inconsistent globally and does not represent actual personal exposure. The ability for air pollution to travel across geographic boundaries makes PM a global problem that would benefit from consistent indicators for global intercomparisons. Therefore, the objective of this paper is to review the current status of PM monitoring at a high-level and make recommendations to improve the links between PM monitoring, modeling, and communication in ways that better enables global participation. This paper is one of four review papers and a synthesis paper describing key aspects of air pollution monitoring and proposed research and policy topics to support improved global indicators. 2. Overview of existing monitoring and indicators The existing infrastructure and processes for monitoring, modeling, and communicating information about particulate matter is well documented. Therefore, this section highlights only key information related to each area. 2.1. Ground-based measurements Ground-based PM monitoring is commonly performed using either filter-based manual sampling or semi-continuous measurement using a wide range of PM monitors. Filter-based sampling is normally performed over a sampling period (usually 24 h), followed by gravimetric mass determination, to provide the mass concentration measurements for different particle size ranges such

585

as total suspended particles (TSP), PM10 and PM2.5. The mass is reported for “dry” PM, commonly at relative humidity of 35e50%. In many developing countries TSP mass is still regulated and monitored (Maggiora and López-Silva, 2006; Kim Oanh et al., 2012); while in most of developed countries TSP is collected for lead content analysis only. Monitoring for PM10 requires less human and laboratory resources and skills than for PM2.5, which, combined with the lack of PM2.5 standards in many countries, make PM2.5 data relatively scarce outside of the U.S. and Europe. PM chemical speciation monitoring requires significant laboratory resources, thus is available in only limited networks or in short timeframe campaigns. Real-time PM measurements provide a better insight into the temporal variations of contributing sources and secondary particle formation. Automated monitoring techniques are available to measure online mass/mass equivalent, e.g., beta attenuation monitors (BAM or b-gauge), tapered element oscillating microbalance (TEOM), among others (Chow et al., 2008). Several particle properties, i.e., light scattering and light absorption, can also be monitored online and used as the PM mass surrogate. Accordingly, simpler PM monitors such as a nephelometer can be used to estimate PM mass while an Aethalometer can be used to measure or report light absorption on a filter tape reported as black carbon mass. National monitoring networks vary in their emphasis on TSP, PM10, and PM2.5. Where a standard permanent automatic monitoring network is available, PM10 (and PM2.5 in some cases) data are generated routinely. Such networks commonly use automated PM monitoring techniques to produce online hourly PM data in many cases with co-located manual PM samplers (U.S. EPA, 2012; ECWG, 2002). In addition to urban data, monitoring at remote sites is particularly important in assessment of regional and long-range transport. However, remote sites and countries with inadequate infrastructure have significant problems related to the accessibility to the sites, to electricity and water, and to technical capacity, which results in a scarcity of data. All ground-based measures only represent a single location, which, combined with a few other monitors in a city or region, are assumed to represent a typical exposure to pollutants. However, exposure is highly personalized, dependent on how and where an individual travels, works, cooks, and lives (O’Neill et al., 2003). Personal monitors have been used in only very small studies, with an emphasis on indoor air quality (e.g., Williams et al., 2000). Additionally, no significant effort has been made to improve spatial resolution by expanding monitoring locations by orders of magnitude to hundreds or thousands of monitors. 2.2. Satellite measurements To provide a better overview of air pollution over large geographical areas, satellite observations can provide valuable information relevant to ground-level PM2.5 concentrations (Martin, 2008; Hoff and Christopher, 2009). The most commonly used instruments for estimating global ground-level PM2.5 concentrations are the Moderate Resolution Imaging Spectroradiometer (MODIS) and Multi-angle Imaging Spectroradiometer (MISR). Both instruments are onboard the Terra satellite in a sun-synchronous orbit with a 10:30 am equator crossing time. A second MODIS instrument is onboard Aqua satellite with a 1:30 pm equator crossing time. Both instruments offer retrievals of aerosol optical depth (AOD; a measure of extinction of radiation by aerosols in the entire atmospheric column) for cloud-free and snow-free conditions (Levy et al., 2013). The spatial resolution of operational AOD retrievals is typically about 10 km  10 km for MODIS and 18 km  18 km for MISR. A 3 km MODIS AOD product recently

586

J. Engel-Cox et al. / Atmospheric Environment 80 (2013) 584e590

became available (Remer et al., 2013). As valuable as the MODIS and MISR instruments have been, they are well past their 6 year design life as research instruments and have already experienced deterioration of monitoring capacity. Satellite-based estimates of PM2.5, such as AOD, are starting to be applied to produce indicators with policy implications (e.g., Cooper et al., 2012). Since 2004, the relationship between AOD and PM2.5 has been intensely studied (Engel-Cox et al., 2004; Hoff and Christopher, 2009). A major challenge continues to be the conversion the satellite retrieval of AOD into ground-level PM2.5 mass concentrations at all locations across the globe. Potential factors that affect the relation of AOD to “dry” PM2.5 include the surface reflectance variability, aerosol vertical profile, the aerosol hygroscopicity (ability of the aerosol to swell with relative humidity), and sampling methods (Drury et al., 2010). Fundamentally, an exact correlation between AOD and PM2.5 may not be possible, since AOD is a snapshot visibility-based measure over a total column of the atmosphere averaged over a spatial area whereas PM2.5 is a physical dry massbased measure at a single ground-based location averaged over time. A powerful synergy exists between satellite and ground-based measures. Satellites enable measurements in areas without monitors, in remote locations, and over oceans, which, when combined and validated with accurate ground-based data, present a more complete and consistent air pollution at larger scales. 2.3. Models Models are the logical next step to integrating disparate datasets to estimate and predict air pollution. Models make use of auxiliary datasets to represent the true state of atmospheric aerosols, expanding the spatial and/or temporal coverage of observations. The simplest of models relate in situ measurements to unmonitored locations through geostatistical interpolation. Kriging, which uses spatiotemporal correlations between in situ measurements to produce a continuous pollutant field, is a common interpolation technique. Land use regression (LUR) models, by contrast, expand the spatial coverage of in situ concentrations by relating geographic factors such as land cover, traffic and topography to changes in pollution level (Ryan and LeMasters, 2007). Both LUR and kriging are powerful techniques on intra-urban scales (e.g., Clougherty et al., 2008; Jerrett et al., 2005). Their dependence on in situ networks, however, limits their applicability on a global scale. Chemical transport models (CTMs), a subclass of integrated meteorological-emission models, use meteorological datasets, emission inventories, and equations of atmospheric physics and chemistry to predict the temporal and spatial variation of atmospheric aerosol and other trace gases. CTM resolution and uncertainty is directly related to the availability and accuracy of auxiliary datasets and computational expense. Typical global resolution is on the order of hundreds of kilometers, with regional or nested simulations possible at about an order of magnitude less. Auxiliary data, such as emissions inventories, and the level of understanding of relevant chemical reactions, both of which may vary with time and region, strongly impact the accuracy of CTM results. For example, global GEOS-Chem simulations of PM2.5 perform better inside North America than the rest of the world, likely due to increased uncertainty in emissions, meteorology, and sub-grid variation in aerosol concentration (van Donkelaar et al., 2010). The complete spatial, temporal, and speciated coverage of CTMs can aid in the interpretation of observed values, providing insight into their composition and impact of spatial or temporal sampling. One such application is the use of CTM-simulated aerosol vertical profiles to relate satellite retrievals of AOD to surface PM2.5, which has been used in exposure and epidemiological studies (e.g., Brauer et al., 2012; Hystad et al., 2011).

Statistically-based methods of relating AOD to PM2.5 are also effective on a local scale. These approaches typically develop empirical relationships based upon a training dataset of in situ observations. More recently, meteorological observations, land cover, lidar measurements, and other relevant values have also been incorporated, improving the accuracy of these methods (e.g., Liu et al., 2011). In some cases, where sufficient ground-based monitors are present, empirical relationships are determined on a daily basis allowing for calibration against daily variability in both satellite retrieval error and aerosol vertical profile (Hu et al., 2013). 2.4. Measures for policy communication Communication of air quality information covers a range of time regimes spanning real time for immediate action to annual for setting policy. Communication of air quality information to the public and policymakers is largely done through the use of indicators and indices. As described in detail in the synthesis paper (Hsu et al., 2013), an indicator is a specialized statistic using complex scientific data for the purpose of gaining and sharing information with stakeholders in understandable ways so they can make decisions or establish policy. Since PM has a direct impact on human health, PM indicators are typically developed using healthbased thresholds of pollutant concentrations. As an example, the World Health Organization (WHO) sets annual average concentration guidelines for PM2.5 of 10 mg m3 and a 24-h daily average limit for PM2.5 of 25 mg m3; PM10 is 20 mg m3 annually and 50 mg m3 daily (WHO, 2006). A city or a country is either above or below the threshold. Most country-level regulatory agencies have set their own health-based thresholds that are similar but often not the same as the WHO guidelines. An index is an aggregation of indicators, which is sometimes simplified in order to be easier to communicate to the public than multiple concentration values; indices are often unitless and can provide consistency to a communicated value (e.g., 0e100) for multiple air pollutants with varying units of measure. For public communication of forecasts and in real-time, the limits are broken down into multiple levels of severity, depending on the level of response required by different affected groups. The concentration levels are given color codes or simple general numbers and often combine multiple air pollutant components (PM, ozone, SO2, etc.). For example, the U.S. Environmental Protection Agency uses an air quality index (AQI) with numerical values of 0e300þ and 6 primary colors (U.S. EPA, 2009). The AQI uses the highest value of the multiple pollutants in the index, including PM and ozone. AQI values over 100 (which for PM2.5 represent daily average concentrations greater than 40 mg m3) are considered unhealthy for at least a subgroup of the population. Communications are done using both forecasts and real-time data, with the policy objective to enable the public to potentially minimize their exposure and for government decisionmakers to take action to reduce emissions from industry or transportation (U.S. EPA, 2013a). Using a simple numerical value and color scale means stakeholders do not need to understand different pollutants and their varying units and thresholds. These indicators link together complex scientific measurements, healthbased policy decisions, and clear communication methods. Many other countries have implemented versions of the AQI with concentrations and colors specific to their regulations and culture (U.S. EPA, 2013b). The simplicity of many indicators can be a disadvantage to more knowledgeable stakeholders and data users, since the method of calculation of the indicator is not always clearly described and the raw or underlying data is not made available. While indicators are proven at the city scale, there are few effective indicators at both smaller and larger scales. Small spatial scale differences in PM, such as street level pollution or local

J. Engel-Cox et al. / Atmospheric Environment 80 (2013) 584e590

industry emissions, are not well measured by indicators. For example, a city overall may have low levels of PM, but residents in a house close to a busy road may experience high PM levels. There are also very few good indicators of indoor air pollution, which is most important where cooking and heating still come from biomass burning. At the large scale, global PM indicators are challenging since PM consists of a wide variety of potential components (e.g., soot, dust, sulfate) measured at varying sizes (e.g., PM10, PM2.5) with multiple measuring devices (e.g., TEOMs, FRMs, remote sensors) with a range of accuracies. Each country’s measurements are different and cannot be easily combined. This is sometimes addressed by selecting representative or surrogate measures and averaging them over time to reduce variability. For example, the global Environmental Performance Index (EPI) used an integrated model with satellite based AOD for global coverage and groundbased measures of PM2.5 for accuracy to obtain a policy-focused measure for comparisons between countries and regions (Emerson et al., 2012). In general, however, there has been relatively little work done to develop quantitative indicators of PM for comparison at global scales. 3. Consideration of regional aspects Particulate matter concentration and composition is highly dependent on the types of sources, geography, and atmospheric conditions that can influence the formation of secondary particles and transport varying distances from its source. Monitoring methods and communication also varies significantly by country and region. 3.1. Regional variability of sources and pollutant transport across boundaries While the overall sources of PM pollution are similar globally, the dominance of some sources over others varies by region. Additionally, long-range transport of PM from one region to another is a challenge to reducing PM concentrations in any one region. For example, biomass burning and Saharan dust are issues of concern for Africa, as well as transportation pollutants from rapid urbanization. Both the dust and smoke from Africa also contribute to high particulate matter air pollution events in the Mediterranean and the Middle East (Katsoulis, 1999; Draxler et al., 2001; Rodriguez et al., 2001). In Asia, rapid urbanization with associated growth in energy consumption and transportation contribute to high concentration levels in Asian countries. These are exacerbated by dust storms that occur over arid desert areas of East Asia, such as the Gobi Desert, which contribute to deterioration of the air quality in China, Hong Kong, Taiwan, Korea, Japan, and Mongolia (Ma et al., 2004). Under certain conditions, this mix of dust and urban pollutants can be transported across the Pacific to North America (Heald et al., 2006). Industrial, transportation, and agricultural emissions are major sources in Europe and the Americas, with transport events across multiple borders. 3.2. Monitoring variability by country and region National and regional monitoring capacity also varies widely. North America has largely transitioned from ambient PM10 to PM2.5 monitoring, including total mass and speciation, using established QA/QC procedures and made available in several databases (IMPROVE, 2012; SLAMS, 2012; STN, 2012). Europe still primarily focuses on PM10 although with the PM2.5 standard is being introduced under the new Directive (EU, 2008), routine PM2.5 monitoring is expanding. In Asia, the available national monitoring networks in developing countries in the past focused more on TSP but recently PM10 is now monitored routinely; With the exception a

587

few countries such as China, PM2.5 data in Asia are still fragmented and generally have been generated through various coordinated international research studies (Hopke et al., 2008; Kim Oanh et al., 2006). This situation is expected to change as some Asian countries now have set PM2.5 standards. It is a similar situation in Latin American countries (Maggiora and López-Silva, 2006), with Central American countries of Panama, El Salvador, Guatemala, and Costa Rica in the early stages of implementing PM2.5 monitoring programs (Zell et al., 2008). Fewer PM data have been collected for Africa and the reported data are largely limited to PM10 levels in some countries (e.g., Naidoo et al., 2006). The air quality monitoring capabilities in developing countries have been improving and many countries/cities have automatic air quality monitoring networks equipped to provide routine PM10 and/or PM2.5 measurements. However, the availability of modern monitoring equipment does not always guarantee the generation of data of acceptable quality. In some cases, the monitoring networks do not function properly due to unreliable power supplies, lack of spare parts, poor capacity for calibration and data management, and low commitment; hence either no data or no reliable data are produced. In other cases, the air quality data that are generated are not publicly shared because of data telemetry issues and/or political sensitivities. 4. ‘State of the art’ of monitors, models, and indicators The monitoring for PM data has expanded and improved over the past 50 years. Generalized methods have been developed to communicate these data at the local and national scale. However, there continues to be gaps and inconsistencies in data and technology, which can hinder their full use in improving air quality impacts on human and environmental health. 4.1. Improving ground-based and satellite measurements of air quality Advancements in ground-based measurements continue incrementally as more monitoring networks are installed and the monitors can increasingly differentiate the size fractions and chemical species of PM. Only very limited work has been done to improve the simplicity and robustness of the monitors and to globally standardize sampling, chemical analyses, and data analysis and management. Alternative and innovative new monitors in the early stages of development focus on personal and low-cost monitors such as tiny air pollution monitors installed in mobile phones that upload data (Shah, 2012) or surrogates such as using photographs from mobile phones that can be converted into visibility measurement (Mobile Sensing, 2012), although there has been very little use of these instruments. Current limitations in satellite-based PM data availability include the inability to retrieve accurate AOD over bright surfaces and under clouds, inability to identify chemical species of PM, and spatial and temporal gaps due to the mechanics of current satellites and sensors. The MODIS and MISR sensors are aging and were designed as research, not operational, instruments. However, recent developments in satellite remote sensing bode well for addressing some of these limitations and improving their use for air pollution indicators. This includes developments to renew and enhance the existing fleet of satellites and to develop retrieval algorithms to improve their accuracy and precision for air pollution. The new Visible Infrared Imaging Radiometer Suite (VIIRS) instrument launched in October 2011 onboard the Suomi National Polarorbiting Partnership (NPP) satellite offers an important dataset for future applications. The development and launch of geostationary missions designed for aerosol measurements will improve issues related to sampling by increasing the frequency of observations. For

588

J. Engel-Cox et al. / Atmospheric Environment 80 (2013) 584e590

example, the Advanced Baseline Imager (ABI) is expected to be launched onboard GOES-R in 2015, and will provide continuous aerosol observations for daytime cloud-free conditions. The development of new satellite instrumentation, such as multiangle spectro-polarimetric imagers could offer more information to identify the size and chemical species of aerosols, which are both of relevance for air quality indicators. 4.2. Advancements in modeling Integration of models and remote sensing through radiative transfer models have the potential to improve AOD retrieval algorithms and ultimately PM2.5 estimates (e.g., Drury et al., 2008; Wang et al., 2010). Further integration requires an understanding of the relative accuracy of each set of measurements. Globally, current methodologies provide incompatible metrics of error, if any metrics at all, limiting the ability with which global PM2.5 estimates can be accurately combined. A flexible global repository of in situ PM2.5 measurements would be an important step towards such a consistent, global error evaluation. An accurate inventory of emissions of precursor pollutants is essential to CTM performance of both aerosol concentration and speciation. For example, high uncertainties in the emissions of organic aerosols cause substantial underestimates in simulated PM2.5 for many regions of the world (e.g., Hodzic et al., 2009; Volkamer et al., 2006). The impact of emissions inventories on simulated AOD to PM2.5 relationships has not been studied in detail, but errors in these inventories have the potential to impact the accuracy of satellite-based estimates through these relationships. Infrequent regional updates to anthropogenic emission inventories and differences in local methodology to estimate emission inventories inhibits model accuracy in many regions and globally. So, while the models have advanced significantly, they are currently limited by availability of accurate global datasets as inputs. 4.3. Better indicators for communication of health impacts Over the past decade, PM indicators have changed relatively little, still relying on ground-based monitoring networks and spatial algorithms. Integrated global indices have been limited by the lack of consistent harmonized datasets. Development of new algorithms is needed to better integrate datasets together, not just AOD and PM2.5 concentrations, but also emission sources, landuse, buildings, and weather. Even if the indicators and indices have not changed, the communication of them has greatly expanded as global digital communication has undergone a rapid and phenomenal transition toward increased accessibility and decentralization. For air quality, new means of communication has enabled data, maps, and notifications to be provided more easily to the public and decisionmakers via websites, social media, and mobile devices. Increased processing capacity and bandwidth have resulted in access to an increased range of online PM and AOD data and maps. The greatest amount of available PM data is in North America and Europe, so the expansion and innovation in communications will most likely grow in other regions with current significant air quality problems and an increasingly aware public. In addition to expansion of communications, a new trend is to also make the data more interactive and personal. The principles of open science and crowdsourcing could make PM data more relevant to users, but the tools to do so, such as the personal monitors described in Section 4.1, need to be developed, provided, and promoted to the interested public. New methodologies to ensure data quality of new widely distributed data collection systems also needs to be developed. These systems should be an affordable cost

because the information is of most interest in places where pollution levels are high, such as less developed countries that cannot easily afford or maintain expensive networks. 5. Recommendations for next generation PM indicators The analysis of the information in this paper as well as a workshop on indicators for multiple pollutants including described in Hsu et al. (2013) resulted in key recommendations, summarized below. The sections of this paper with information related to each recommendation are in parentheses. Increase and tailor sensor networks to country and region, while producing globally harmonized data (see Sections 2.1, 3.1). PM sources, concentrations, and composition vary by region and country, as does the institutional capacity to support air quality monitoring. Therefore, the next generation air quality monitors, supporting systems, and indicators should consider a country’s priorities and capacity when developing ground-based monitor networks, institutional arrangements for monitor operation, application of satellite and modeled information, and stakeholder communications. At the same time, country-specific networks need to use globally consistent unified data monitoring standards, including metadata that define how the data is collected and its potential error, to make global comparison easier. Then, even with differences in regulations and communication at a country level, both ground-based and satellite data can be aggregated at a global scale into a harmonized global indicator. Support development of new sensors that are cheaper and more durable (see Section 3.2). The technology development for PM monitors has focused on improvement and expansion of the size fractions and speciation that can be measured. However, current monitoring equipment requires significant resources to operate and maintain that may not be available in all regions of the world. Research and development funding targeted on developing monitors that are lower cost, require less maintenance, and are energy self-sufficient could enable an expansion of monitoring into remote regions and locations without strong infrastructure. Conduct targeted campaigns to enable better use of satellite data and improve models where networks are not possible (see Sections 2.2, 2.3, 4.2). A country’s level of development also impacts the appropriate institutional arrangements and timeframes for monitoring. Countries at the lowest level of development are unlikely to have the institutional capacity, financial resources, or political will to sustain a ground-based monitoring program. However, satellite data can be a surrogate with appropriate validation with ground data. With additional data collection, models could be created that are focused at the country-level. Short (e.g., 1year) campaigns, funded and technically led by an external research organization in partnership with local scientists and institutions, focused on the data collection needed to develop targeted PM models from satellite and available meteorology data would be an interim solution. The data from such a campaign would improve application of satellite data and models for ongoing understanding of air quality in a country. Promote data sharing of networked data (see Sections 2.4, 3.2, 4.2). Global monitoring and data standards would help ensure data quality, but there is also the need for governments and scientific organizations to openly share their data. Open sharing of data is not universal and, in some cases where it is shared, it is not easily accessible. When there are good national networks that produce consistent quality data that is shared openly, they serve as the foundation for better global understanding and the development of strong research communities through data analysis, modeling, health impact studies, and communication. Membership of nations in organizations such as the Group on Earth Observations

J. Engel-Cox et al. / Atmospheric Environment 80 (2013) 584e590

(GEO) could address data quality issues and encourage countries to openly share data. Develop new personal sensors that link to social media as citizen science (see Sections 4.1, 4.3). The growth of mobile phones, the Internet, and social media has greatly changed the global information landscape. Air quality monitoring could leverage this but requires the development of smaller monitors or similar surrogates (e.g., visibility via photographs) plus the QA processes and the electronic tools and motivations for people to contribute the data as citizens participating in open science. A publiceprivate partnership or challenge grants to develop these types of new mobile technologies could result in a leap in innovation. Expand and improve a set of multi-pollutant, multi-purpose indicators to compare monitoring data across countries (see Sections 2.4, 3.2). While air pollution monitoring and assessment are an essential element of air quality improvement, they are not an objective in itself. New and improved indicators should be developed that aim to emphasize the health and economic benefits of good air quality, not only for PM but also for all air pollutants. While there may not be a single AQI due to the diversity of country level requirements, common standardized source datasets and translation between them should be developed through research and analysis in multiple countries. Indicators that include estimates for anthropogenic and transported pollution are particularly important. Acknowledgments The authors would like to thank NASA, Columbia University, and Yale University for their organization of the Next Generation of Air Quality project and support for workshops to develop the recommendations in this paper. Special personal thanks also for comments and input from Dr. Michal Krzyzanowski. References Brauer, M., Amman, M., Burnett, R., Cohen, A., Dentener, F., Ezzatti, M., Henderson, S., Krzyzanowski, M., Martin, R., van Dingenen, R., van Doneklaar, A., Thurston, G., 2012. Exposure assessment for the estimation of the global burden of disease attributable to outdoor air pollution. Environmental Science and Technology 46, 652e660. Chow, J., Doraiswamy, P., Watson, J., Chen, L., Ho, S., Sodeman, D., 2008. Advances in integrated and continuous measurements for particle mass and chemical, composition. Journal of Air and Waste Management Association 58, 141e163. Clougherty, J., Wright, R., Baxter, L., Levy, J., 2008. Land use regression modeling of intra-urban residential variability in multiple traffic-related air pollutants. Environmental Health 7 (17). Cooper, M., Martin, R., van Donkelaar, A., Lamsal, L., Brauer, M., Brook, J., 2012. A satellite-based multi-pollutant index of global air quality. Environmental Science and Technology 46 (16), 8523e8524. Draxler, R., Gillette, D., Kirkpatrick, J., Heller, J., 2001. Estimating PM10 air concentrations from dust storms in Iraq, Kuwait and Saudi Arabia. Atmospheric Environment 35, 4315e4330. Drury, E., Jacob, D., Spurr, R., Wang, J., Shinozuka, Y., Anderson, B., Clarke, A., Dibb, J., McNaughton, C., Weber, R., 2010. Synthesis of satellite (MODIS), aircraft (ICARTT), and surface (IMPROVE, EPA-AQS, AERONET) aerosol observations over eastern North America to improve MODIS aerosol retrievals and constrain surface aerosol concentrations and sources. Journal of Geophysical Research: Atmospheres (1984e2012) 115. Drury, E., Jacob, D., Wang, J., Spurr, R., Chance, K., 2008. Improved algorithm for MODIS satellite retrievals of aerosol optical depths over western North America. Journal of Geophysical Research 113 (D16204), 1e11. ECWG, 2002. Guidance to Member States on PM10 Monitoring and Intercomparisons with the Reference Method. European Commission Working Group on Particulate Matter. http://ec.europa.eu/environment/air/quality/ legislation/pdf/finalwgreporten.pdf. Emerson, J., Hsu, A., Levy, M., de Sherbinin, A., Mara, V., Esty, D., Jaiteh, M., 2012. 2012 Environmental Performance Index and Pilot Trend Environmental Performance Index. Yale Center for Environmental Law and Policy, New Haven. http://epi.yale.edu. Engel-Cox, J., Hoff, R., Haymet, A., 2004. Recommendations on the use of satellite remote-sensing data for urban air quality. Journal of the Air & Waste Management Association 54, 1360e1371. EU, 2008. The Directive 2008/50/EC. http://ec.europa.eu/environment/air/quality/ standards.htm.

589

Heald, C., Jacob, D., Park, R., Alexander, B., Fairlie, T., Yantosca, R., Chu, D., 2006. Transpacific transport of Asian anthropogenic aerosols and its impact on surface air quality in the United States. Journal of Geophysical Research 111, D14310. http://dx.doi.org/10.1029/2005JD006847. Hodzic, A., Jimenez, J., Madronich, S., Aiken, A., Bessagnet, B., Curci, G., Fast, J., Lamarque, J., Onasch, T., Roux, G., Schauer, J., Stone, E., Ulbrich, I., 2009. Modeling organic aerosols during MILAGRO: importance of biogenic secondary organic aerosols. Atmospheric Chemistry and Physics 9, 6949e6982. Hoff, R., Christopher, S., 2009. Critical reviewdremote sensing of particulate pollution from space: have we reached the promised land? A critical review. Journal of Air & Waste Management Association 59, 645e675. Hopke, P., Cohen, D., Begum, B., Biswas, S., Ni, B., Pandit, G., Santoso, M., Chung, Y., Davy, P., Markwitz, A., Waheed, S., Siddique, N., Santos, F., Pabroa, P., Seneviratne, M., Wimolwattanapun, W., Bunprapob, S., Vuong, T., Duy Hien, P., Markowicz, A., 2008. Urban air quality in the Asian region. Science of the Total Environment 404, 1103e1112. Hsu, A., Reuben, A., Shindell, D., de Sherbinin, A., Levy, M., 2013. Towards the next generation of air quality monitoring indicators. Atmospheric Environment 80, 561e570. Hu, X., Waller, L., Al-Hamdan, M., Crosson, W., Estes, M., Estes, S., Quattrochi, D., Sarnat, J., Liu, Y., 2013. Estimating ground-level PM2.5 concentrations in the southeastern U.S. using geographically weighted regression. Environmental Research 121, 1e10. Hystad, P., Setton, E., Cervantes, A., Poplawski, K., Deschenes, S., Brauer, M., van Donkelaar, A., Lamsal, L., Martin, R., Jerrett, M., Demers, P., 2011. Creating national air pollution models for population exposure assessment in Canada. Environmental Health Perspectives 118, 1123e1129. IMPROVE, 2012. Interagency Monitoring of Protected Visual Environments Network. http://vista.cira.colostate.edu/improve/. Jerrett, M., Arain, A., Kanaroglou, P., Beckerman, B., Potoglou, D., Sahsuvaroglu, T., Morrison, J., Glovis, C., 2005. A review and evaluation of intraurban air pollution exposure models. Journal of Exposure Analysis and Environmental Epidemiology 15, 185e204. Katsoulis, B.,1999. The potential for long-range transport of air-pollutants into Greece: a climatological analysis. The Science of the Total Environment 231, 101e113. Kim Oanh, N., Permadi, D., Phuc, N., Zhuang, Y., 2012. Air Quality Status and Management Practices in Asian Developing Countries. In: Integrated air quality management: Asian Case Studies. CRC Press. Taylor & Francis Group, ISBN 9781-4398-6225-4 (Chapter 1). Kim Oanh, N., Upadhyay, N., Zhuang, Y., Hao, Z., Murthy, D., Lestari, P., Villarine, J., Chengchua, K., Co, H., Dung, N., Lindgren, E., 2006. Particulate air pollution in six Asian cities: spatial and temporal distributions, and associated sources. Atmospheric Environment 40, 3367e3380. Levy, R., Mattoo, S., Munchak, L., Remer, L., Sayer, A., Hsu, N., 2013. The Collection 6 MODIS aerosol products over land and ocean. Atmospheric Measurement Technologies 6, 159e259. Lim, S., et al., 2012. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990e2010: a systematic analysis for the Global Burden of Disease Study 2010. The Lancet 380, 2224e2260. Lippmann, M., 2012. Particulate matter (PM) air pollution and health: regulatory and policy implications. Air Quality, Atmosphere & Health 5, 237e241. Liu, Y., Wang, Z., Wang, J., Ferrare, R., Newsom, R., Welton, E., 2011. The effect of aerosol vertical profiles on satellite-estimated surface particle sulfate concentrations. Remote Sensing of Environment 115, 508e513. Ma, C., Tohno, S., Kasahara, M., Hayakawa, S., 2004. Properties of individual Asian dust storm particles collected at Kosan, Korea during ACE-Asia. Atmospheric Environment 38, 1133e1143. Maggiora, C., López-Silva, J., 2006. Vulnerability to Air Pollution in Latin America and the Caribbean. The World Bank. Latin America and Caribbean Region Sustainable Development Working Paper 28. Martin, R., 2008. Satellite remote sensing of surface air quality. Atmospheric Environment 42, 7823e7843. Mobile Sensing, 2012. Air Visibility Monitoring. Robotic embedded Systems Laboratory. http://robotics.usc.edu/wmobilesensing/Projects/AirVisibilityMonitoring (accessed 02.08.12.). Naidoo, M., Zunckel, M., John, J., Taviv, R., 2006. Overview of ambient air quality monitoring in South Africa. In: Presented at NACA Conference 2006, 18e20 October. East London, South Africa. Available at: http://researchspace.csir.co.za/ dspace/handle/10204/1710. O’Neill, M., Jerrett, M., Kawachi, I., Levy, J., Cohen, A., Gouveia, N., Wilkinson, P., Fletcher, T., Cifuentes, L., Schwartz, J., 2003. Health, wealth, and air pollution: advancing theory and methods. Environmental Health Perspectives 111,1861e1870. Pope, C., Ezzati, M., Dockery, D., 2009. Fine particulate air pollution and life expectancy in the United States. The New England Journal of Medicine 360, 376e386. Remer, L., Mattoo, S., Levy, R., Munchak, L., 2013. MODIS 3 km aerosol product: algorithm and global perspective. Atmospheric Measurement Techniques 6, 69e112. Rodriguez, S., Querol, X., Alastuey, A., Kallos, G., Kakaliagou, O., 2001. Saharan dust contributions to PM10 and TSP levels in Southern and Eastern Spain. Atmospheric Environment 35, 2433e2447. Ryan, P., LeMasters, G., 2007. A review of land-use regression models for characterizing intraurban air pollution exposure. Inhalation Toxicology 19, 127e133. Shah, A., 2012. Intel Researchers Plot a Smarter, Personalized Cloud, IDG News, PC World, 26 April 2012. http://www.pcworld.com/article/254545/intel_ researchers_plot_a_smarter_personalized_cloud.html.

590

J. Engel-Cox et al. / Atmospheric Environment 80 (2013) 584e590

SLAMS, 2012. State and Local Air Monitoring Stations (SLAMS). http://www.epa. gov/ttn/amtic/pmqainf.html. STN, 2012. Speciation Trends Network (STN). http://www.epa.gov/ttn/amtic/ speciepg.html. U.S. EPA, 2009. Technical Assistance Document for the Reporting of Daily Air Quality e the Air Quality Index (AQI). EPA-454/B-09-001, February 2009. U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina. Available at: http://www.epa.gov/airnow/aqi_ tech_assistance.pdf. U.S. EPA, 2012. List of Designated Reference and Equivalent Methods. Issue Update June 2012. Available at: http://www.epa.gov/ttn/amtic/criteria.html (accessed August 2012). U.S. EPA, 2013a. AIRNow. http://www.airnow.gov (accessed July 2013). U.S. EPA, 2013b. AIRNow International Air Quality. http://www.airnow.gov/index. cfm?action¼airnow.international (accessed July 2013). van Donkelaar, A., Martin, R., Brauer, M., Kahn, R., Levy, R., Verduzco, C., et al., 2010. Global estimates of ambient fine particulate matter concentrations from satellite-based aerosol optical depth: development and application. Environmental Health Perspectives 118, 847e855.

Volkamer, R., Jimenez, J., San Martini, F., Dzepina, K., Zhang, Q., Salcedo, D., et al., 2006. Secondary organic aerosol formation from anthropogenic air pollution: rapid and higher than expected. Geophysical Research Letters 33, L17811. Wang, J., Xu, X., Spurr, R., Wang, Y., Drury, E., 2010. Improved algorithm for MODIS satellite retrievals of aerosol optical thickness over land in dusty atmosphere: implications for air quality monitoring in China. Remote Sensing of Environment 114, 2575e2583. WHO, 2006. Air Quality Guidelines. Global Update 2005. Particulate Matter, Ozone, Nitrogen Dioxide and Sulfur Dioxide. World Health Organization, Geneva. Williams, R., Creason, J., Zweidinger, R., Watts, R., Sheldon, L., Shy, C., 2000. Indoor, outdoor, and personal exposure monitoring of particulate air pollution: the Baltimore elderly epidemiology-exposure pilot study. Atmospheric Environment 34, 4193e4204. Zell, E., Cherrington, E., Friedl, L., Duke, V., Gonzales, O., Delgado, R., Hoff, R., Jordan, N., Irwin, D., 2008. An integrated approach to measuring air quality in Central America. In: Proceedings of AWMA Symposium on Air Quality Measurement Methods and Technology, Research Triangle Park, North Carolina.