Grantee Research Project Results
Final Report: Coarse PM Emissions Model Development and Inventory Validation
EPA Grant Number: R834552Title: Coarse PM Emissions Model Development and Inventory Validation
Investigators: Hannigan, Michael P. , Wiedinmyer, Christine , Fierer, Noah
Institution: University of Colorado at Boulder , National Center for Atmospheric Research
EPA Project Officer: Chung, Serena
Project Period: June 1, 2010 through May 31, 2013
Project Amount: $500,000
RFA: Novel Approaches to Improving Air Pollution Emissions Information (2009) RFA Text | Recipients Lists
Research Category: Air Quality and Air Toxics , Air
Objective:
This research project was designed to ultimately provide new information about emissions from PM10-2.5 and PM10. Existing measurements of PM10-2.5 would be used to evaluate existing inventories and their use in CMAQ. This evaluation process would result in a set of recommendations for updating existing emissions modules to more accurately simulate PM10-2.5 emissions. Further, a new emissions model that would simulate the emissions of primary biological particles would be developed.
Summary/Accomplishments (Outputs/Outcomes):
Below are some highlights of each of the study sub-components, but we wanted to lead with a brief synthesis of our findings. Emissions of PM10-2.5 are rarely linked with emissions of PM2.5, even when the general source (for example, roadways) is the same. As such, any emissions estimations that are tied to PM2.5 emissions are not wise. Emissions of PM10-2.5 are intimately linked with meteorological/environmental conditions (for example, wind speed and soil moisture) in addition to human activity (for example, vehicle braking). Thus, emissions estimates for PM10- 2.5 need to be generated ‘on-line’ in an atmospheric model, similar to biogenic VOC emissions. This link also impacts spatial variability. Ambient PM10-2.5 has a shorter atmospheric lifetime than PM2.5 so it exhibits more temporal variability than PM2.5 at any given site and it can have a large impact with significant spatial variability near to big sources. However, we often observe relatively spatially homogeneous PM10-2.5 in a region, when we are not monitoring near to and downwind of a source. This lack of spatial variability can be attributed to the broad reach of weather systems; in other words, when it is dry and windy at one location in the Front Range, it is dry and windy at most locations in the Front Range. We hypothesize that this also is the reason that we saw little variability in PM10-2.5 with height at the BOA tower, while observing a decreasing gradient for PM2.5 going up the tower. Emissions of PM10-2.5 contain significant organic content (~15-50%), and the make-up of those organics is different than the organics in the PM2.5. The PM10-2.5 organic material is less water soluble than PM2.5; similarly, when the high pH water is used, PM2.5 water soluble components increase more than PM10-2.5. Ambient concentrations of carbonaceous PM10-2.5 and PM2.5 exhibit seasonal differences but not the same seasonality patterns. Summer increases in carbonaceous PM2.5, with increasing water solubility, are likely due to increased atmospheric processing of gas-phase organics and not an increase in a primary PM source. Spring/summer/fall increases in carbonaceous PM10-2.5 are likely due to increases in primary sources, potentially biological particles but also likely are meteorological/environmental conditions that are more conducive to emissions.
Model-Measurement Comparison
We characterized coarse particulate matter (PM10-2.5) spanning the western United States based on the analysis of measurements from 50 sites reporting in the U.S. EPA Air Quality System (AQS) and two state agencies. We found that the observed PM10-2.5 concentrations show significant spatial variability and distinct spatial patterns, associated with the distributions of land use/land cover and soil moisture. The highest concentrations were observed in the southwestern United States, where sparse vegetation, barren or shrublands dominate with lower soil moistures, whereas the lowest concentrations were observed in areas dominated by grasslands, forest, or croplands with higher surface soil moistures. The observed PM10-2.5 concentrations also show variable seasonal, weekly, and diurnal patterns, indicating a variety of sources and their relative importance at different locations. To obtain insights for regional PM10-2.5 modeling, the observed results also were compared to modeled PM10-2.5 concentrations from an annual simulation using the Community Multiscale Air Quality (CMAQ) modeling system that has been designed for regulatory or policy assessments of a variety of pollutants including PM10, which consists of PM10-2.5 and fine particulate matter (PM2.5). The model under-predicts PM10-2.5 observations at 49 of 50 sites, among which 14 sites have annual observation means that are at least 5 times greater than model means. Below, we show this comparison for two sites with different characteristics, Seattle and Fargo.
(a) Seattle (b) Fargo
Model results also fail to reproduce their spatial patterns. Important sources were not included in the emission inventory used and/or the applied emissions were greatly under-estimated. Unlike observations, the modeled concentrations show similar monthly, weekly, and diurnal patterns across the entire domain. CMAQ does not include organics in PM10-2.5, which recent measurements show to be a significant component. The results of the analysis improve our understanding of sources and behavior of PM10-2.5 and suggest avenues for future improvements to models that simulate PM10-2.5 emissions, transport and fate. More details are provided in our publications (Bowers, et al., 2011).
Conclusions:
Temporal and Spatial Analysis of PM10, PM10-2.5, and PM2.5 Measurements
The characteristics of concentrations of PM10-2.5, PM2.5, and PM10 in the United States using measurements at 77 sites were evaluated. PM10 concentrations show strong spatial variability, and the highest concentrations were observed in the southwestern United States, with the maximum annual average concentration (50.3 μg/m3) at El Paso North in New Mexico being six times higher than that at the Badlands site (8.5 μg/m3) in South Dakota. At about half of the sites, PM10-2.5 is the major fraction of PM10. PM10-2.5 and PM2.5 concentrations show different spatial patterns. The highest concentrations of PM10-2.5 were observed at sites in the southwestern United States, leading to the highest PM10 concentrations there. The PM2.5 concentrations are the major contributor to the PM10 concentrations at many sites in the eastern United States. Poor correlations were generally found between PM10-2.5 and PM2.5, suggesting that PM10-2.5 and PM2.5 are influenced by different sources. Only 4 of 77 sites possessed correlation coefficients R2 between PM10-2.5 and PM2.5 that are higher than 0.5, where there are significant soil dust or marine influences. PM10-2.5 is generally more variable than PM2.5; the average ratio of PM10-2.5 CV (coefficient of variation) over PM2.5 CV is 1.7. This is likely because PM10-2.5 has a shorter lifetime in the atmosphere (higher deposition velocity) and is primarily emitted from mechanical processes such as agricultural harvest and construction that are more influenced by factors including human operation and wind speed leading to a strong episodic nature whereas PM2.5 has a lower deposition velocity, and is primarily emitted by combustion processes or formed by chemical reactions, and therefore is more affected by regional transport sources. As a result of its higher relative temporal variability, PM10-2.5 acts as the major driver for PM10 extremes. PM10-2.5 is significantly correlated with PM10 at all investigated sites, with the average correlation value R2 = 0.8. Correlations of PM2.5 with PM10 (average of 0.37) were overall considerably lower than those between PM10-2.5 and PM10. Different seasonal, weekly, and diurnal patterns were observed between PM10-2.5 and PM2.5 at agricultural, on-road traffic, quarrying, airport, and marine sites. Below, we show that even in urban areas near roadways, the diurnal patterns of PM10-2.5 and PM2.5 are different; likely indicating that even though the roadway is a significant source in both, different emission and transport processes are impacting the different PM size modes.
(a) Urban/Near Road PM10-2.5 (b) Urban/Near Road PM2.5
At investigated agricultural sites, PM10-2.5 concentrations are lower in winter months than in summer and autumn months, with highest levels corresponding to harvest and planting, while the concentrations of PM2.5 are higher in winter when there were few agricultural activities. The harvest and planting signatures were not observed in PM2.5 concentrations at any of these sites, suggesting that agricultural activities do not have a strong influence on PM2.5 concentrations. Weekly and diurnal patterns reveal a stronger local anthropogenic influence on PM10-2.5 than PM2.5 at all sites except the marine sites, which have a relatively less obvious anthropogenic influence on both PM10-2.5 and PM2.5. More details are provide in our publications (Bowers, et al., 2012).
Temporal, Spatial and Vertical Analysis of PM10-2.5 and PM2.5 and Components in the Colorado Front Range
As mentioned above, we were able to leverage the EPA STAR funded C-CRUSH study as well as a new collaboration with F. Rosario-Ortiz to further this aspect of our research. Here, we highlight the carbonaceous analysis of the C-CRUSH study filters, then biological analysis of the same filters and finally the analysis of the BAO tower filters. Additionally, our team also led an effort to synthesize the current state of knowledge about natural carbonaceous material in the atmosphere; see our publications for more details. Carbonaceous analysis of C-CRUSH filters. In this portion of the study, the organic matter from PM10-2.5 and PM2.5 filter samples collected every sixth day for over a year was characterized using bulk characterization methods including: total organic and elemental carbon, organic carbon peak fractions, water-soluble carbon and nitrogen, UV-vis absorbance, fluorescence, and endotoxin content. Filters were collected at four sites, two located in urban Denver and two in relatively rural Greeley, Colorado, as part of the Colorado Coarse Rural-Urban Sources and Health (C-CRUSH) study. Of the total particulate mass, organic matter contributed 35% of the PM10-2.5 and 58% of the PM2.5 total mass. PM10-2.5 organic matter was 32% water soluble. In contrast, PM2.5 was 75% water soluble. Organic carbon that volatilized at low temperatures during bulk carbon analysis dominated fine aerosols, while higher temperature peak fractions contributed substantially to PM10-2.5. High correlation was found for the lowest organic peak fraction (PK340) across size fractions and space and indicates a similar source or atmospheric process is contributing to low molecular weight organic compounds in both size fractions. Specific ultraviolet absorbance at 254 nm (SUVA254) values, indicative of aromaticity, were elevated in both size fractions at a traffic-influenced site relative to rural and residential-urban sites. SUVA254 values in Greeley were similar to those for agricultural soils and fell between values measured at the two sites in Denver. Fulvic and humic-like fluorescence peaks dominated PM2.5 samples but were not measured in the coarse mode. Instead, PM10-2.5 fluorescence peaked in the tryptophan and tyrosine-like regions. SUVA254 and Humification Index values peaked in winter for PM2.5, a trend attributed to residential woodsmoke emissions. As shown in the figure below, endotoxin concentrations were significantly higher in Greeley and in PM10-2.5, a trend that is likely due to cattle feedlot and agricultural emissions in the region. For more details about this portion of the research see our publications (Duhl, et al, 2014).
(a) Seasonal PM10-2.5 endotoxin (b) Seasonal PM2.5 endotoxin
Biological Analysis of C-CRUSH Filters
Bacteria and fungi are ubiquitous throughout the Earth’s lower atmosphere where they often represent a major component of atmospheric aerosols with potentially important impacts on human health and atmospheric dynamics. However, the diversity, composition, and spatiotemporal dynamics of these airborne microbes remain poorly understood. We performed a comprehensive analysis of airborne microbes across two aerosol size fractions at an urban and rural site in the Colorado Front Range over a 14-month period. PM10-2.5 and PM2.5 samples were collected at weekly intervals with both bacterial and fungal diversity assessed via high-throughput sequencing. The diversity and composition of the airborne communities varied across urban and rural sites, between the two size fractions, and over time. Bacteria were the dominant type of bioaerosol in the collected air samples, while fungi and plants (pollen) made up the remainder. Considering bacteria made up the majority of bioaerosol particles, we analyzed the bacterial communities in greater detail using a bacterial-specific sequencing approach. Overall, bacterial taxonomic richness and the relative abundances of specific bacterial taxa exhibited significant patterns of seasonality. Likewise, airborne bacterial communities varied significantly between urban and rural sites and across aerosol size fractions. Source-tracking analyses indicate that soils and leaves represented important sources of bacteria to the near-surface atmosphere across all locations with cow fecal bacteria also representing an important source of bioaerosols at rural sites during early fall and early spring. We highlight this result in the figure below. Colors represent the three dominant sources: soils = red, leaf-surfaces = green, and cow fecal material = orange, and intensity of the color represents the relative contribution. Numbers represent the summed source contribution for each month, city (Denver or Greeley) and particle size fraction.
The data from this study suggest that a complex set of environmental factors, including changes in atmospheric conditions and shifts in the relative importance of available microbial sources act to control the composition of microbial bioaerosols in rural and urban environments. More details about these results can be found in our publications (Li, et al., 2013b).
Analysis of BAO Tower Filters
To characterize the vertical and seasonal variability of PM10-2.5 and PM2.5 in the lower atmosphere, filter samples were collected simultaneously at ground level and at a height of 250 m at the Boulder Atmospheric Observatory tower in Erie, CO, starting in late October 2012. At both sampling heights, the total mass, elemental carbon (EC), organic carbon (OC), and water soluble organic carbon (WSOC) concentrations of both PM10-2.5 and PM2.5 were measured over a 7-month period. The ground level PM2.5 concentration was always higher than the 250 m PM2.5 concentration, while the ground and 250 m levels were not significantly different for the PM10-2.5. The same trends were observed for EC and OC. As for seasonality, there was little to no trend in the PM2.5; whereas, the PM10-2.5 showed lower concentrations during the winter. This seasonality pattern also was observed in the C-CRUSH sample. Additionally, we observed significant differences in the proportion of the carbonaceous material that was water soluble between size modes, PM2.5 is approximately twice as soluble as PM10-2.5, and between seasons, increases in summer as compared to winter. We characterized the optical properties (fluorescence and ultraviolet/visible spectroscopic analyses) of the water-soluble and base-soluble aerosol fractions to investigate the contribution of soluble organic matter, including humic and humic-like substances (HULIS), to the carbonaceous fraction of bulk aerosol. Differences were observed in the optical properties of the WSOC between the two size modes as well as between the two heights, especially in fine-mode samples, with seasonal variability observed in both size fractions. The figure below shows the fluorescence excitation and emission matrices for typical samples, and highlights these differences and also shows comparison to terrestrial natural organic matter. Notice the strong similarity between the summer PM2.5 and the surface water sample, both showing a strong humic signature. More details about this tower study can be found in our publications (Bowers, et al., 2013).
Biological PM Emissions Module Development
Finally, we have developed a continental-scale biological particle emission inventory and algorithms describing the timing and magnitude of biological particle release into the atmosphere. Three types of primary biological particles are considered; fungi, bacteria, and pollen with aerodynamic diameters ≤ 10 μm. Emission capacities are Plant Functional Group (PFT)-specific (for fungi and bacteria) or species-specific (for pollens), and are modified by hourly as well as 24-hour cumulative relative humidity, air temperature, and precipitation. The modeling framework utilizes a dispersal module to simulate conditions that lead to passive release of bioparticles into the air. The figure below shows the general framework employed for fungi. A similar approached is employed for bacteria but without the precipitation initiated release mechanism. Maximum emission rates for each fungal PFT have been compiled by other researchers and these rates serve as our starting point. Other algorithms that represent the temperature and relative humidity impacts are drawn from the literature. Currently, we are seeking additional funding to continue on this pathway, by first attempting validation of the module and subsequently refining based on results from the validation.
Journal Articles on this Report : 5 Displayed | Download in RIS Format
Other project views: | All 7 publications | 6 publications in selected types | All 5 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Bowers RM, Sullivan AP, Costello EK, Collett Jr. JL, Knight R, Fierer N. Sources of bacteria in outdoor air across cities in the Midwestern United States. Applied and Environmental Microbiology 2011;77(18):6350-6356. |
R834552 (Final) |
Exit Exit Exit |
|
Bowers RM, Clements N, Emerson JB, Wiedinmyer C, Hannigan MP, Fierer N. Seasonal variability in the bacterial and fungal diversity of the near-surface atmosphere. Environmental Science & Technology 2013;47(21):12097-12106. |
R834552 (Final) R833744 (Final) |
Exit Exit Exit |
|
Li R, Wiedinmyer C, Baker KR, Hannigan MP. Characterization of coarse particulate matter in the western United States: a comparison between observation and modeling. Atmospheric Chemistry and Physics 2013;13(3):1311-1327. |
R834552 (2011) R834552 (Final) |
Exit Exit |
|
Bowers RM, McCubbin IB, Hallar AG, Fierer N. Seasonal variability in airborne bacterial communities at a high-elevation site. Atmospheric Environment 2012;50:41-49. |
R834552 (Final) |
Exit Exit Exit |
|
Li R, Wiedinmyer C, Hannigan MP. Contrast and correlations between coarse and fine particulate matter in the United States. Science of the Total Environment 2013;456-457:346-358. |
R834552 (Final) |
Exit Exit Exit |
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.