97-MP6.07
Y. J. Lin
Bechtel Power Corporation
Gaithersburg, Maryland
Alex Kasprak
Bechtel/Parsons Brinckerhoff
Boston, Massachusetts
INTRODUCTION
The Massachusetts Highway Department's Central Artery/Tunnel (CA/T) Project in Boston, Massachusetts, is one of the largest ongoing public works projects in the country today. When completed, the CA/T Project will consist of a new third harbor tunnel (named the Ted Williams Tunnel) which will link downtown Boston to Logan Airport in East Boston and a new 3 mile underground artery which will replace Boston's current elevated north-south expressway.
The results of the Final Supplemental Environmental Impact Statement/Report for the CA/T Project concluded that the Project, when completed, should meet all applicable National and State Ambient Air Quality Standards (AAQS), and will result in beneficial long-term air quality impacts for the Boston area. However, during the peak construction period of the Project, fine particulate matter with a diameter of less than 10 microns (i.e., PM10) generated from CA/T construction activities, may result in potential exceedances of the 24-hour PM10 standard. In order to mitigate and minimize potential dust impacts, the CA/T Project has implemented PM10 monitoring to evaluate construction related fugitive particulate emissions. The objectives of the monitoring programs have been to 1) characterize background and construction phase PM10 levels along the Project's alignment; 2) verify design-phase dispersion modeling analyses; 3) assess the need for construction-related dust control measures. PM10 data collection for these programs have been obtained using a continuous real-time PM10 monitor approved by the Environmental Protection Agency (EPA) along with portable PM10 saturation samplers originally developed by the Lane Regional Air Pollution Authority (EPA Region 10). In addition, the equipment used to assess ambient PM10 impacts from construction related activities, have also been used to assess PM10 emissions from the ventilation system of completed sections of the CA/T's Ted Williams Tunnel.
This paper provides a comparison of ambient PM10 data that have been collected during the summer of 1994 using portable saturation samplers and comparing the data to coincident data collected by the Project's continuous reference PM10 air monitor (EPA approved). The comparison is used to justify the use of portable air samplers to record in-tunnel PM10 concentrations which are also presented to assess vehicle emissions rates from the CA/T's ventilation system.
PM10 AMBIENT BASELINE MONITORING
PM10 monitoring was conducted during 2 separate data collection periods along the existing elevated Central Artery alignment during the summer of 1994. The primary purpose of the monitoring program was to collect representative background PM10 levels, prior to the commencement of mainline construction contracts, and other construction contracts that might in the future cause significant increases in PM10 levels. Monitoring was carried out in two areas, Boston's downtown and south end areas. PM10 monitoring in the downtown area commenced 13 July 1994 and was completed on 4 August 1994. PM10 concentrations were collected at eight locations (Figure 1) during the weekdays for a total of fourteen days. PM10 monitoring in Boston's south end area commenced 15 August 1994 and was completed on 4 September 1994. Similar to the downtown area, data were collected at seven locations (Figure 2) during the weekdays for a total of fourteen days.
Monitoring Equipment
24-hour samples were collected at each monitoring location as described above during the two study periods. The equipment used to carry out the monitoring consisted of the following:
· Portable MiniVOL air samplers manufactured by Air Metrics of Springfield, Oregon;
· A Cahn C-33 microbalance scale which has an accuracy of ±1 mg;
· A TEOM series 1400 high volume continuous PM10 sampler manufactured by Rupprecht & Patashnick of Albany, New York; and
· A DSM 3260 data acquisition unit manufactured by Odessa Engineering of Austin, Texas.
The portable air samplers were used for all baseline sampling conducted in the two study areas. In the portable air samplers, air is drawn through a particle pre-selector (10 mg nominal particle size cut) and then through a filter which is weighed on the C-33 microbalance both before and after exposure. A control unit which forms the bulk of each sampler ensures that a volumetric flow rate of five liters per minute at ambient conditions is maintained throughout the pre-programmed sampling period (i.e., 24-hours). A rechargeable battery pack that powers the unit is attached to its base (Figure 3).
The following summarizes the sampling protocol used for the portable MiniVOL samples:
· The flow rate of each sampling unit was calibrated prior to the start of the monitoring study.
· Clean sample filters were weighed in batches using the calibrated Cahn C-33 microbalance scale. Three control filters that had equilibrated for at least 24 hours in a temperature and humidity controlled environment were weighed at the beginning and end of each clean filter weighing session.
· The clean weighed filters were placed in filter holder assemblies and attached to the portable samplers.
· The portable samplers were mounted to utility poles and hung approximately 10 feet above the ground.
· Each sampler remained in the field from Monday to Friday.
· Each sampler battery and filter holder assembly was changed daily, Tuesday through Thursday.
· The exposed filters were allowed to equilibrate for at least 24 hours and were then re-weighed. The same three control filters used to weigh the clean filters were weighed at the beginning and end of the exposed filter weighing session.
· A PM10 concentration derived from the pre and post-sampling difference in filter weight, corrected for any changes in average control filter weight, was calculated for each sample.
The Project's TEOM continuous air sampler is housed in an environmentally controlled shelter located along the project alignment in South Boston and has been in operation since 1 July 1992. It has been designated as the Project's PM10 reference station. The TEOM uses a gravimetric, EPA-designated equivalent method (No. EQPM-1090-079) for the determination of 24-hour average PM10 concentrations. Data from the CA/T's TEOM unit are transferred automatically to a DSM 3260 data acquisition system, which process and stores the data as both hourly and daily averages.
Quality Control
For quality control purposes, a portable sampler was collocated with the TEOM reference station to assess the accuracy of the portable sampling method against an accepted method. In addition, at least one sampling location in each study, also contained a collocated portable sampler mounted along side a sampler to assess the precision of the portable sampling method.
As a validity check for the portable sampling method, data recovery rates were calculated for each monitoring area. The EPA guideline for ambient air monitoring specifies a data recovery rate of at least 80 percent for a sample to be considered valid (EPA, 1987).
DATA ANALYSIS
PM10 concentrations for each monitoring location were summarized for both the downtown and south end areas.
A Kendall concordance coefficient was calculated to evaluate the correspondence of PM10 concentrations determined simultaneously at all sampling locations. The Kendall coefficient is a non-parametric measurement that represents the intensity of agreement among sample groups (i.e., locations) using ranked values (i.e., PM10 concentrations). It is analogous to average rank order correlation between two variables. A high level of concordance indicates that for any high or low PM10 concentration recorded at one location, a similar high or low concentration would be recorded at all locations. Separate concordance analyses were performed for the downtown and south end areas.
In order to perform a Kendall concordance analysis, PM10 concentrations from all corresponding sampling locations were required. Therefore, on dates when PM10 concentrations from one of the corresponding samplers were not available, the entire set of concentrations for that day was excluded from the analysis.
It was assumed that significant concordance among PM10 concentrations recorded at various locations might indicate a strong regional influence on PM10 concentrations, such as weather conditions. Thus, PM10 concentrations from the Project's continuous reference station in South Boston were included in each analysis to help test whether the day-to-day PM10 concentrations observed in the downtown and south end areas were also observed at the reference site; then assessing whether the pattern was local or regional.
A single-factor Analysis of Variance (ANOVA) was performed on the same data set used for the Kendall's analysis, with "location" as the factor, to test whether PM10 concentrations were significantly different among sampling locations. Data were evaluated to ensure that statistical requirements for ANOVA were met including equality of variances, normal distribution of data, and no correlation between means and variances. Statistical requirements of data were tested using Levene's tests (i.e., p less than 0.01), normal probability plots, and a computation of correlation coefficient (i.e., p less than 0.05), respectively. If the ANOVA results indicated there were differences among the locations, then a Tukey's Honest Significant Difference Test was used to determine which locations differed from one another.
PM10 concentrations recorded in the downtown area were then compared to concentrations recorded in the south end area using a t-test for independent samples. This test was used to assess whether there were differences in baseline PM10 concentrations that were not attributable to changes in meteorological conditions during the two sampling periods. In performing the t-test, all concentrations were "standardized" by subtracting the corresponding concentration measured at the reference station assuming that regional meteorological conditions would influence PM10 concentrations in South Boston and the two study areas in a similar manner.
RESULTS
Data representing PM10 concentrations in two study areas were evaluated to: 1) demonstrate sampling proficiency; 2) determine trends in data among sampling stations; and 3) compare concentrations monitored at locations within and between study areas.
During the three week period when monitoring was performed in the downtown and south end areas, the overall data recovery rates achieved for the two areas were 94 and 84 percent, respectively. These data recovery rates are considered extremely good for establishing background concentrations in these areas.
PM10 Concentrations
The maximum background PM10 concentration recorded in the downtown area was 91 mg/m3 at Site 8 (Table 1), whereas, the maximum background PM10 concentration recorded in the south end area was 54 mg/m3 at Site 5 (Table 2). When compared to the maximum PM10 concentration recorded by the reference station, the 91 mg/m3 concentration was found to be slightly above the maximum reference station concentration of 88 mg/m3 (Table 1). Figure 3 and 4 show the frequency distribution of PM10 concentrations for the downtown and south end areas, respectively.
Among the downtown sampling locations, the mean PM10 value was 61 mg/m3. The mean PM10 value for the south end was 33 mg/m3. Some of the differences between these mean values may be attributable to varying weather conditions such as temperature, humidity, and precipitation that were occurring at the time the samples were taken. As a further examination, a t-test analysis described below was performed for comparing mean values among sampling locations with standardized data.
Data Trends: Kendall's Concordance Analysis
Kendall's coefficient (W) of concordance was calculated for PM10 concentration data from all sampling locations (including the reference station in South Boston) for each baseline study area. To assess the significance of the association represented by the Kendall coefficient, an equivalent chi-square value was derived using the following formula (Zar, 1984):
X2 = M(n-1)W
where:
X2 = chi square value (with n-1 degrees of freedom)
M = number of variables being correlated (sampling locations)
n = number of data per variable
A chi-square table was then used to determine the probability of an association among PM10 concentrations measured at various sampling locations.
This analysis showed a high level of concordance (i.e., Kendall's p less than 0.05) among sampling locations (including the reference station) for both study locations (Table 3). This indicates that PM10 concentrations measured at all sampling locations tracked well with each other and with concentrations measured at the Project's reference station in South Boston. The high degree of concordance suggests that day-to-day variation in baseline PM10 concentrations is largely affected by regional factors, such as temperature, humidity, and precipitation.
Comparison of Maximum PM10 Concentrations
The sampling design involved two sequential monitoring studies, the downtown area followed by the south end area. With sampling design, data collected could be directly compared among monitoring locations within a sampling area; however, comparisons between study areas required standardization of the data using samples from the reference station in South Boston. This standardized approach was considered viable due to the fact that regional transport and diffusion conditions would similarly affect regional air quality over the entire study area.
Comparison within Study Areas
A one-way ANOVA test, followed by a Tukey's Honest Significant Difference Test, was performed on every combination of PM10 sampling locations within the downtown and south end areas (Table 4). For the downtown area a significant difference (i.e., Tukey's p less than 0.05) was found in PM10 between monitoring Site 4 and Site 6. The concentrations from Site 4 were significantly lower than the concentrations recorded at Site 6 (Figure 5). This may be attributed to Site 6 being in an area that experiences much higher traffic volumes than Site 4, which is located on a side street. No significant differences were observed among concentrations measured at all other PM10 monitoring sites in the downtown area (i.e., Tukey's p greater than 0.05).
Within the south end area, a significant difference was found in PM10 concentrations between monitoring Site 3, monitoring Site 5, and monitoring Site 7. The concentrations recorded at Site 3 were significantly lower than the concentrations recorded at both Site 5 and Site 7 (Figure 6). Again, this may be attributed to Site 5 and 7 being in locations that have different traffic configurations and experience much higher traffic volumes than Site 3, which is located in the middle of a parking lot. No significant differences were observed among concentrations measured at all other PM10 monitoring sites in the south end area (i.e., Tukey's p greater than 0.05).
Comparison between Downtown and South End Areas
As mentioned above, a t-test was performed to compare the standardized concentrations for PM10 recorded in the downtown area to those recorded in the south end area. The results of the t-test (Table 5) revealed a minor, yet statistically significant difference between the average PM10 concentrations recorded in the downtown area versus those recorded in the south end (p less than 0.05). These results, using standardized data, suggest that if air quality were simultaneously measured in the downtown and south end areas, the downtown concentrations would, on average, be only 3.5 mg/m3 higher than the concentrations recorded in the south end area. PM10 concentrations from the downtown and south end portable samplers compared well with the reference station PM10 concentrations (Figure 7).
DISCUSSION AND CONCLUSION
As demonstrated by the statistical analyses presented above, PM10 concentrations recorded using portable samplers compared well with coincident PM10 concentrations recorded using the EPA's TEOM series 1400 PM10 monitor. In order to determine in-tunnel PM10 emission rates from vehicles using the Ted Williams Tunnel, portable PM10 monitors are being used to collect PM10 concentrations inside the CA/T's ventilation system. Along with data collected from the portable samplers, daily traffic volume, traffic mix and exhaust flow rates from the TWT's ventilation buildings are also being obtained. These data, along with the in-tunnel PM10 concentrations recorded in the tunnel's exhaust plenum using the portable samplers, should provide reliable emissions rates from the ventilation building.
Click here to download the graphics and charts associated with this article.
REFERENCES
AirMetrics, Springfield, Oregon: Operation Manual for the MiniVOL Portable Sampler.
CAHN Instruments, Inc., Cerritos, CA., Instruction Manual for the C-33 Microbalance Scale.
Environmental Protection Agency: Ambient Monitoring Guideline for Prevention of Significant Deterioration, EPA-450/4-87-007, 1987
Lane Regional Air Pollution Authority (LRAPA): Portable Sampler Operation Manual, November 1992.
Odessa Engineering, Inc.: DSM 3260 User's Manual, August 1991.
Rupprecht & Patashnick, Inc.: Operating Manual for the TEOM Series 1400 Ambient Particulate (PM10) Monitor, Revision 1.18, May 1992.
Zar, Jerrold H., Biostatistical Analysis, 2nd Edition, Prentice-Hall, Inc., Englewood Cliffs, NJ, 1984.