Excerpted from Galena-Park-Monitoring-Report-FINAL.pdf
III. Community Monitoring Project
A year-long air monitoring project began in Galena Park in November 2011. Together, Air Alliance Houston, Global Community Monitor, and the people of Galena Park developed a plan for a community based participatory research project to monitor for particulate matter pollution. Galena Park residents were trained in the use of air monitoring equipment and invited to join Air Alliance Houston in the deployment of our monitors. Sites for monitoring were selected based on information gathered during the community health survey and community mapping workshop.
(not reproduced on this wiki, see PDF linked at top of page for this section)
B. Monitoring Equipment and Analysis (Page 13 of the PDF)
Particulate matter samples were gathered using two MiniVol11 Tactical Air Samplers (TAS).12 Our partners Global Community Monitor have much experience using MiniVols. They are manufactured by Airmetrics, a Eugene, OR based company providing innovative air sampling equipment. Although the MiniVol is not a reference method sampler, its manufacturer states that “the MiniVol™ TAS gives results that closely approximate data from Federal Reference Method samplers.”13
The MiniVol uses a system that includes a pump, impactors, and a mechanical filter. Each MiniVol is calibrated by Airmetrics before it is sold. The calibration ensures that the pump draws 5.0 liters of ambient air per minute. The monitor is equipped with a flow meter that allows operators to check the flow rate before and after each use. With a flow-rate of 5 liters per minute (approximately the same as that of the human lungs), each MiniVol collected 7.2 cubic meters of air during a 24-hour sample.
The MiniVol can be configured using a variety of impactors to sample particulate matter in the ranges of 10 microns (PM-10), 2.5 microns (PM2.5), and total suspended particles (TSP). For this project. Air Alliance Houston collected only PM2.5 samples for 24-hour periods. Samples are collected in a 47 millimeter (mm) mechanical filter. Filters were purchased and prepared for use by CHESTER LabNet (CLN), a Tigard, OR based specialty laboratory that focuses on inorganic air quality analysis. Samples were collected using CLN’s procedures for proper handling, storage, and shipment of filters. Samples were returned to CLN for analysis.14
Two types of filters were used: Teflon and quartz. Samples taken with Teflon filters were analyzed using two analytical protocols. A gravimetric analysis yields a total PM2.5 mass. CLN weighs each filter after conditioning at a constant temperature and relative humidity prior to shipment using a balance accurate to one microgram. After sampling, the filter is reconditioned and then CLN reweighs the filter to determine the total weight of material present. Dividing this weight by the total volume of air sampled (7.2 m3 for a 24-hour sample) gives an average concentration of PM2.5 in the ambient air during the 24-hour sample collection. Teflon filters were also analyzed using an x-ray fluorescence (XRF) analysis that determined concentrations of forty different elements.
Quartz filters were used to sample for elemental and organic carbon. Quartz filters must be refrigerated before and after use, and during shipping, in order to limit volatilization of collected material. Quartz filters were analyzed using the National Institute for Occupational Safety and Health (NIOSH) Method 5040 for diesel particulates as elemental carbon. Method 5040 determines total carbon as organic carbon and elemental carbon. Elemental carbon was then used to calculate the concentration of diesel particulates.
C. Monitoring Protocol and Quality Control (Page 14 of the PDF)
All participants in the study handling the MiniVols were trained in their use by Denny Larson of Global Community Monitor. Two monitors were deployed side-by-side throughout the project. Each monitor includes an on-board programmable timer. Timers are programmed to sample for twenty-four hours and independently verified. Monitors are powered by on-board battery packs. Battery packs are exchanged after each run for freshly charged batteries. Monitors were placed in secure locations on the roofs of one or two story buildings. Each monitor was positioned according to manufacturer specifications and GCM training to avoid nearby objects or conditions that could impact sample collection and accuracy. Monitor flow rates were checked before and after each sample collection. Programmable timers were also checked to ensure samples were collected for twenty-four hours. Procedures were documented on field data sheets required by CLN. A detailed protocol checklist developed and used by Air Alliance Houston throughout the project is included as Appendix D.
1. Duplicate Samples
Typically, one monitor used a Teflon filter and the other a quartz filter. However, on five occasions, the same filter type was placed in each monitor in order to perform a duplicate sample analysis. Three duplicates were performed using Teflon filters; two with quartz filters. Results of duplicate samples are reported as the average of the two samples.
A regression analysis of the three Teflon duplicates is included in the independent analysis of this project performed by Laura Campos, a graduate student with the Rice University Department of Statistics.15 Two of those duplicates showed good precision. A third showed poor precision, with a relative error of 91%. Possible causes for this error were discussed, but no conclusions were reached. Mark Chernaik, Ph.D., Science for Citizens, offered this explanation:
First, the two samples collected at 2000 Clinton Drive (City Hall) on 22-23 January 2013 were co-located. However, the PM2.5 levels are not comparable, differing by almost 100% (19 versus 36.2 ug/m3. Interestingly, the XRF levels of these two collated samples are comparable (for example, iron levels of 0.157 versus 0.161 ug/m3). Since the XRF levels are comparable, my best guess is that some error with the gravimetry analysis may have produced the anomalous result.16
The two duplicate samples performed on quartz filters were not included in Ms. Campos’ regression analysis. Those duplicates were found to have relative errors of 4% and -31%, respectively.
2. Field Blanks
Several field blanks were also included in the project. Each field blank was transported to the monitoring site, removed from its package, placed in the monitor while it was turned off, immediately removed from the monitor, returned to its package, and placed in the closed casing of the monitor during its 24-hour run. Three field blanks were included using this procedure: one quartz and two Teflon. The results of their analysis are below:
Table 1: Field Blanks
|Lab ID||Start Date||Total PM2.5 (μg)||Total Organic Carbon (μg)||Total Elemental Carbon (μg)||Total Carbon (μg)|
|13-U164||1/20/2013||4.309 ± 1.668||0.0000 ± 1.390||4.309 ± 3.058|
|12-T3958||5/21/2013||130 ± 10|
|12-T4053||8/7/2013||8 ± 10|
The two Teflon filter blanks contained 130 μg and 8 μg of total PM2.5, respectively. The Teflon blank handled on 5/21/2013, Lab ID 12-T3958, was found to be damaged by CLN. The filter had an indentation that may have been caused by it being bumped during handling. This sampling error was a likely source of contamination for this blank. The second Teflon blank, Lab ID 12-T4053, yielded 8 μg of total PM2.5 and did not exceed the margin of error. In addition to these planned field blanks, two filters were placed in monitors that, due to user error, did not turn on and drew no air through their filters during the 24-hour sampling period. Rather than discard these filters, they were also sent to CLN for analysis as a sort of field blank. The result of this analysis is below:
Table 2: Additional Field Blanks
|Lab ID||Start Date||Total PM2.5 (μg)||Total Organic Carbon (μg)||Total Elemental Carbon (μg)||Total Carbon (μg)|
|12-T3952||3/5/2013||3 ± 10|
|13-U396||3/5/2013||15.29 ± 2.085||0.0000 ± 1.390||15.29 ± 3.475|
For the Teflon filter, the total PM2.5 mass of 3 μg does not exceed the margin of error.
For the quartz filter, the total organic carbon mass was 15.29 μg. Again, this exceeds the margin of error and can be compared to the average organic carbon mass of all quartz filters of 63.98 μg. This quartz filter was also not found to contain elemental carbon.
Given the limited size of this study, and of the number of duplicate and blank samples included, it is difficult to draw conclusions from these results. Several duplicate samples did show good precision, but others—and one in particular—did not. One Teflon field blank showed large amounts of total PM2.5, but others did not. Some quartz field blanks indicated organic carbon, but none showed any elemental carbon, which was used in this study to estimate diesel pollution. A larger study would be needed to more fully understand these findings.
3. Data Translated or Removed from Analysis
On two occasions, monitors did not run for the standard 24 hours. Sample 12-T4096, collected on 1/21/2013 at Early Head Start, ran for 25 hours due to a timer programming error. The resulting sample volume was 7.5 m3, rather than the usual sample volume of 7.2 m3. The net mass for this sample was 81 μg, resulting in a concentration of 25.4 μg/m3. This concentration was correctly calculated and reported to us by CLN.
Sample 12-T3951, collected on 3/4/2013 at the Resource Center, ran for 21.8 hours due to a dead battery. The resulting sample volume was therefore 6.54 m3. The net mass for this sample was 166 μg, resulting in a concentration of 25.4 μg/m3. This sample was mistakenly calculated by CLN using the typical sample volume of 7.2 m3, resulting in an incorrect report of 23.06 μg/m3. This was corrected in our analysis.
Two samples were removed from analysis. A sample collected at Galena Manor on 3/5/2013 used the wrong type of filter. A sample collected on 8/8/2013 was discovered to have a hole in the filter.
D. Analysis of Results
Data gathered during the community monitoring project were independently analyzed by several individuals and groups. Mark Chernaik, Ph.D., with Science for Citizens, providing periodic analysis and interpretation. A table of his analysis is included in this report as Appendix B. His narrative interpretations are excerpted in the sections below.
An independent analysis was performed and a report completed by Laura Campos, a graduate student in the Rice University Department of Statistics. Ms. Campos was assisted by Loren Raun, Ph.D., Department of Statistics, Rice University. Some suggestions and minor edits were provided by Air Alliance Houston board member Thomas H. Stock, PhD, MPH, Program in Environmental and Occupational Health Sciences, Division of Epidemiology, Human Genetics and Environmental Sciences, University of Texas School of Public Health. The Campos Report is included in this report at Appendix A.
This report was produced by Adrian Shelley, Executive Director of Air Alliance Houston, with extensive help as noted in the Acknowledgements.
- 11 MiniVol is a trademark of Airmetrics.
- 12 Initially, the project included sampling with a DustTrakII aerosol meter. Early difficulties with the DustTrakII led to its data being discarded. It has since been repaired and recalibrated and will be included in future studies.
- 13 See http://www.airmetrics.com/products/minivol/index.html.
- 14 Full data reports, field data sheets, and chain of custody forms for this project are available upon request.
- 15 See Campos, Laura, “Are the current TCEQ PM2.5 Monitors representative of the Galena Park Community?” (June 2014). See Appendix A.
- 16 Email from Mark Chernaik to Adrian Shelley (8 Feb. 2013).