Grantee Research Project Results
2002 Progress Report: Eddy-Correlation Measurement of Size-Segregated and Composition-Resolved Aerosol Depositional Flux Using an Aerosol Mass Spectrometer
EPA Grant Number: R828172Title: Eddy-Correlation Measurement of Size-Segregated and Composition-Resolved Aerosol Depositional Flux Using an Aerosol Mass Spectrometer
Investigators: Smith, Kenneth A.
Institution: Massachusetts Institute of Technology
EPA Project Officer: Hahn, Intaek
Project Period: July 24, 2000 through July 23, 2002 (Extended to July 23, 2004)
Project Period Covered by this Report: July 24, 2001 through July 23, 2002
Project Amount: $225,000
RFA: Exploratory Research - Engineering, Chemistry, and Physics) (1999) RFA Text | Recipients Lists
Research Category: Safer Chemicals , Water , Land and Waste Management , Air
Objective:
The overall objective of this research project is to improve our understanding of atmospheric aerosol behavior by directly measuring the flux of individual species associated with fine particles in the atmosphere. These measurements are made by coupling a novel aerosol mass spectrometry (AMS) instrument with the established eddy-correlation technique. An Aerodyne AMS is used to quantitatively measure the concentration of individual species in a size-segregated aerosol. This is the first measurement of the deposition flux of individual species in a size-resolved aerosol. The deposition velocities developed here will be published for use in urban-, regional-, and global-scale models of aerosol formation and fate.
This research was proposed to be done in seven tasks. The next sections of this progress report discuss the progress made on these tasks in Years 1 and 2 of the project. The tasks are listed in the order completed. Tasks 1-3 were planned for Year 1 and Tasks 4-7 were planned for Year 2 of the project. We have requested and received a no-cost extension for this project; the project will end February 28, 2003.
Progress Summary:
Progress in Project Year 1
Task 1: Couple AMS and Sonic Anemometer. For eddy-correlation measurement, aerosol concentration measurements by the AMS are saved at a rate of 10 Hz. We analyzed the data stream produced by the original AMS data acquisition software and determined that 30 minutes of data would occupy 1,800 files and 2 gigabytes of storage. Thus, rapid sampling was impractical. We revised the data acquisition software to only save essential data and to save data in the National Center for Supercomputer Applications Hierarchical Data Format . The revised AMS data acquisition software can save 30 minutes of 10 Hz data in 16 megabytes, reducing storage needs by more than 100 times. In addition, data saving requires approximately 10 percent of the data acquisition computer resources, ensuring minimal data loss. The AMS and the sonic anemometer are coupled electronically. The AMS data acquisition software was modified to send a triggering pulse to the sonic anemometer at approximately 10 Hz. Upon receiving the triggering pulse, the anemometer reports data averaged over the interval between two triggering pulses.
These revisions to the AMS data acquisition software were done at Aerodyne between September 2000 and August 2001. This task is complete and AMS data acquisition software is available to measure the flux of a single aerosol species. As discussed below, we used this software in August 2001 to successfully acquire 750 MB of synchronized AMS data during 8 days of continuous operation.
Task 2: Conduct Roof-Top Sampling Experiment. This task was eliminated in favor of a field experiment in Year 1 (see Task 5).
Task 5: Conduct Field Sampling. As part of the PROPHET Summer 2001 Study, we measured aerosol deposition flux at a tower above a temperate forest. This work was undertaken in collaboration with Jonathan Allen at Arizona State University, whose work was funded separately by the National Science Foundation.
The flux measurements were highly synergistic with the PROPHET study, an integrated set of intensive measurements combined with transport and chemical modeling. PROPHET facilities include a 31 m research tower and laboratory funded by the University of Michigan and the National Science Foundation. Research activities in summer 2002 included continuous measurements of O3, CO, PAN, and meteorological parameters, along with simultaneous measurements of ROx, NOx, NOy, volatile organic compounds (VOCs), organic nitrates, and peroxides. In addition, a second Aerodyne AMS was deployed by Darin Toohey's group from the University of Colorado, Boulder (UC). The UC instrument will be used to measure total aerosol composition continuously and as a function of position in the forest canopy. This combination of measurement and modeling studies will allow us to place our deposition flux measurements in the context of well-characterized air mass.
The AMS, associated electronics, and data acquisition computers were housed in the PROPHET laboratory at the base of the tower. A sampling tube (0.5 in OD) was run from the top of the tower to the AMS. A laminar flow regime was chosen to minimize wall losses. Flow through the sampling line was laminar with a flow rate of 10 L min-1; the center of the main flow was sampled isokinetically into the AMS at a flow rate of 1.5 cm3 s-1. This inlet was designed to minimize spreading of the sampled aerosol because of diffusion and dispersion, so that particle spreading within the sampled volume was the equivalent of approximately 0.03 s; this is sufficient for approximately 10 Hz correlated concentration measurements. Sample flow rates were controlled using calibrated critical orifices.
The sonic anemometer was deployed from a boom on the top of the 31 m research tower facing west, in the direction of predominant wind flow (see Figure 1). A Li-Cor infrared hygrometer (Model 7500), purchased as part of this project, was deployed near the anemometer. Aerosol was sampled through a cyclone mounted near the anemometer and ducted to the AMS deployed at the base of the tower.
Figure 1. Photos of Sampling Equipment at PROPHET. Left: Research tower; sampling boom extends to the left from the top of the tower. Right: Sampling assembly with infrared hygrometer, sonic anemometer, and aerosol inlet (cyclone).
We deployed aerosol generation equipment in the PROPHET laboratory and generated known aerosols to calibrate the AMS instrument. We also introduced short bursts calibration aerosols at the sampling inlet using solenoid values to determine the behavior of aerosols in the sampling line.
The AMS was operated to measure the concentration of ions characteristic of the most important aerosol species, such as SO42-, NO3-, and organic matter (see Figure 2). These species were chosen for their environmental relevance and because they are present in high enough concentrations that characteristic ion fragments can be detected with high time resolution.
Task 3: Determine Aerosol Deposition Fluxes From AMS and Anemometer Measurements. We have completed preliminary analysis of the synchronized AMS-anemometer data using the well-known eddy-correlation technique (Businger, 1986) to determine the flux of aerosol species. The flux of species i, Fi, was calculated as the covariance of the vertical wind speed, w, and the concentration, Ci,
- Fi = <w' Ci'>
where primes designate fluctuations from the mean and the angled brackets represent an average over a 30-minute sampling period. The deposition velocity, v, for species i was calculated as
- vi = -Fi / Ci
where by convention a positive vi is toward the surface.
Figure 2. Ions Monitored by AMS During Flux Measurements. Mass-to-charge (m/z) 30 is mainly NO+, characteristic of nitrate; 43 is mainly C3H7+, characteristic of organic matter; 64 is mainly SO2+, characteristic of sulfate.
The deposition velocities for sulfate aerosol calculated for each 30-minute period on August 4, 2001, are shown in Figure 3. These results seem reasonable because the deposition velocity is nearly 0 during stable atmospheric conditions at night; deposition velocity is in the range 0 to 0.6 cm s-1, in the range of experimental values reported by others for fine particle flux to a forest canopy (Gallagher, et al., 1997).
Progress in Project Year 2. The main accomplishments toward completing the proposed work in Year 2 of the project have focused on analyzing the large data set collected in Year 1. The eddy-correlation and related data are in approximately 3,000 files, which occupy approximately 5.5 gigabytes. This data analysis work was completed by our collaborator, Jonathan Allen, and his students at Arizona State University (ASU); their work was not funded by this project.
Task 6: Determine Aerosol Deposition Fluxes From AMS and Anemometer Measurements. The ASU group has made substantial progress on the analysis of the PROPHET 2001 data. They have: (1) created an online data catalog; (2) applied AMS field calibration data; (3) filtered anemometer data to remove invalid data; (4) filtered AMS data to remove infrequent single particle events; (5) identified and excluded periods of erroneous anemometer measurements (less than 2 percent of total measurements); (6) identified and excluded periods of erroneous infrared hygrometer measurements (less than 1 percent of total measurements); (7) determined momentum and sensible heat fluxes from anemometer data; (8) determined latent heat and CO2 fluxes from anemometer and infrared hygrometer data; and (9) evaluated quality of infrared hygrometer data by cospectral analyses.
The data analysis tasks listed above have been implemented as procedures written in the Matlab programming environment; data analysis for subsequent research projects (including the Phoenix deposition study) will be able to reuse these programs for more rapid data analysis.
Preliminary quality checks of the anemometer data have been completed and a small portion of data was eliminated (< 2 percent of the total). Wind speed, wind direction, and coordinate rotations also have been completed using the sonic data. Momentum and sensible heat fluxes then were calculated using the eddy-correlation technique. The momentum flux was calculated as the covariance of the vertical wind speed, w, and the horizontal wind speed, u, multiplied by the air density, ,
F = <w'u'>
and the sensible heat flux was calculated as the covariance of the vertical wind speed, w, and the sonic anemometer temperature, T, multiplied by the air density, , and the heat capacity of the air, Cp,
QH = Cp<w'T'>.
In both equations, primes designate fluctuation from the mean and the angled brackets represent an average.
Momentum flux data for August 4, 2001, are shown in Figure 3. The circles are fluxes over 30-minute averaging periods and the solid line is an exponential, running average of the 30-minute fluxes over 2-hour periods. As expected, the momentum flux shows a strong diurnal pattern, with night time fluxes near 0 and average day time fluxes near -1.0 kg (m s-1)/m2 s. These values are in agreement with the range of data reported by others (McMillen, 1988).
Figure 3. Momentum Flux Data for 30-Minute Averaging Periods on August 4, 2001
Thirty minute averaged sensible heat flux data are shown in Figure 4 for the same time period. The circles are fluxes over 30-minute averaging periods and the solid line is an exponential, running average of the 30-minute fluxes over 2-hour periods. These results also show a strong diurnal pattern as expected; with nighttime fluxes fluctuating near 0 and average daytime fluxes around 300 J/m2 s. This is in good agreement with reported values (Schmid, et al., 2000).
Figure 4. Sensible Heat Flux Data for 30-minute Averaging Periods on August 4, 2001
The deposition velocities for sulfate aerosol calculated for each 30-minute period on August 4, 2001, are shown in Figure 5. The spike at about 5:00 a.m. is still being analyzed, but a trace amount of rain on that morning may have been the cause.
Figure 5. Deposition Velocities for Sulfate Aerosol for 30-Minute Sampling Periods on August 4, 2001
Preliminary analysis of aerosol deposition velocities has been completed for the entire 8-day period of synchronized AMS-sonic anemometer data. The results are shown in Figure 6 for sulfate, organic, and nitrate aerosol deposition velocities. Values are within the range reported by others; however, additional analysis of the data is needed to ensure data quality.
Figure 6. Deposition Velocities for Sulfate, Organic, and Nitrate Aerosols for 30-Minute Sampling Periods During PROPHET 2001
Future Activities:
Future activities for this research project include the following:
Task 4: Modify Instrument Based on the First Field Sampling Experiment. The AMS and anemometer instrument performed well in the first field experiment. We improved the AMS sensitivity by installing a modified pumping system in the AMS instrument. The pumping modification developed by researchers at Aerodyne has been shown to reduce background ion counts, which currently limit chemical detection. This upgrade has been delayed because the first model of the hybrid turbo-drag pumps manufactured by Alcatel have proven unreliable. Researchers at Aerodyne are working with the manufacturer to address pump quality issues.
Task 5: Conduct Second Field Sampling. We will conduct a second field sampling study in collaboration with Jonathan Allen at Arizona State University. The study site will be an agricultural or desert area within or downwind of metropolitan Phoenix, AZ, near the ASU campus. We have planned a 6-week deposition study in the spring of 2003. This includes 1 week of setup, 4 weeks of measurements, and 1 week of break down. (Note: We originally proposed two 1-week field experiments.)
The deposition site will be a topologically simple site, so that analysis of wind and concentration covariance are not complicated by local emission sources or unstable atmospheric turbulence. We propose 4 weeks of measurement so that we can monitor the deposition of a range of chemical species over a range of meteorological conditions. We expect to collect sufficient data to determine size-resolved deposition rates (see Task 6).
Task 6: Determine Aerosol Deposition Fluxes From AMS and Anemometer Measurements. We will complete the analysis of field calibrations made at PROPHET, as well as the Tempe Deposition study, and incorporate these into our calculations of deposition fluxes. We will improve the data analysis programs to follow the recommendations of experts in boundary layer meteorology (Businger, 1986). A partial list of the planned data analysis improvements include: (1) scale AMS data to use measured inlet line losses; (2) use measured inlet line hold-up times and temporal spreading; (3) exclude periods of nonstationary meteorology; (4) exclude periods of nonstationary aerosol concentration; and (5) assure quality of deposition flux measurements by comparing covariance spectra of temperature and aerosol composition.
We also will attempt to calculate particle-size dependent deposition fluxes as:
- Fi,j = <w' Ci,j'>
where the subscript j indicates a particle size range. The deposition velocity for species i in size range j will be calculated as:
- vi,j = -Fi,j / Ci,j.
For this analysis, we will classify sampling periods based on meteorological conditions and average deposition velocities under similar conditions as a function of chemical species and particle size.
Task 7: Compare Aerosol Deposition Fluxes With Model Predictions. The measured deposition velocities will be compared with other measurements (Gallagher, et al., 1997) and with predictions of aerosol deposition models (Slinn, 1982; Williams, 1982). We will present recommended deposition velocities for individual species as a function of particle size, which may be incorporated into existing atmospheric aerosol models.
Journal Articles:
No journal articles submitted with this report: View all 8 publications for this projectSupplemental Keywords:
PM2.5, air toxics, sulfate, nitrate, polycyclic aromatic hydrocarbons, PAHs., RFA, Scientific Discipline, Air, Toxics, Water, Nutrients, particulate matter, Environmental Chemistry, climate change, HAPS, Environmental Monitoring, Engineering, Chemistry, & Physics, nutrient supply, aerosol particles, fine particles, mass spectrometry, toxicology, hydrocarbon, PM 2.5, air sampling, chemical composition, PAH, atmospheric nitrogen deposits, circulation model, airborne aerosols, PM2.5, monitoring of organic particulate matter, nitrogen removal, meterology, acid rain, climatology, atmospheric deposition, eddy-correlation, humidity, vegetative surfacesProgress 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.