The Effects of Scrubber Installation at the Navajo Generating Station ...

TECHNICAL PAPER

ISSN 1047-3289 J. Air & Waste Manage. Assoc. 55:1675–1682 Copyright 2005 Air & Waste Management Association

The Effects of Scrubber Installation at the Navajo Generating Station on Particulate Sulfur and Visibility Levels in the Grand Canyon Mark Green Desert Research Institute, Las Vegas, NV Rob Farber and Nghi Lien Southern California Edison, Rosemead, CA Kristi Gebhart National Park Service, Fort Collins, CO John Molenar Air Resource Specialists, Fort Collins, CO Hari Iyer Colorado State University, Fort Collins, CO Delbert Eatough Brigham Young University, Provo, UT

ABSTRACT Grand Canyon National Park (GCNP) is a mandatory Class I federal area that is afforded visibility protection under the Federal Clean Air Act. In this paper, we have examined the effects on visibility and particulate sulfur (Sp) at GCNP as a result of reducing sulfur dioxide (SO2) emissions by 90% from the Navajo Generating Station (NGS). Scrubbers were retrofitted to each of the three units at NGS during 1997, 1998, and 1999. The Interagency Monitoring of Protected Visual Environments aerosol network database affords us an opportunity to examine trends in Sp and extinction both prescrubber

IMPLICATIONS This paper examined changes in Sp and light extinction at the Grand Canyon after installation of SO2 scrubbers at the NGS. The analysis showed a statistically significant reduction in wintertime sulfur concentrations and a decrease in the difference between in-canyon and south rim Sp and light extinction, in particular, at the high extremes. The expenditure of major financial resources for the scrubbers has resulted in identifiable benefits to wintertime visibility in the Grand Canyon.

Volume 55 November 2005

and postscrubber. The NGS impacts GCNP primarily during the winter (December to February). During winter, at times, there are fogs, stratus, and high-relative humidity in the Grand Canyon. When the NGS plume interacts with these fogs and stratus, rapid conversion of SO2 to Sp can occur. A variety of analytical techniques were used, including cumulative frequency plots of Sp and extinction, and chemical mass balance and tracer source apportionment analysis. We also deployed P value statistical analysis of “extreme” Sp values. Before scrubbers were installed, values of Sp approaching 2 ␮g/m3 were occasionally observed. Because scrubbers have been installed, high levels of Sp have been markedly reduced. Statistical P value analysis suggests that these reductions were significant. Furthermore, we have also observed that Sp has been reduced throughout the cumulative frequency curve during winter by ⬃33% since scrubbers were installed. By contrast, during summer when the NGS impact on the Canyon is minimal, there has been only a relatively small decrease in Sp. INTRODUCTION In 1977, the Clean Air Act was amended to include a section 169A recognizing the importance of visibility in most of the U.S. national parks and wilderness areas. In Journal of the Air & Waste Management Association 1675

Green et al. all, 156 areas were deemed to have important visibility attributes. The Grand Canyon National Park (GCNP) was one of these designated visibility protected areas. Congress set a lofty goal of no made-made visibility impairment at these areas. U.S. Environmental Protection Agency (EPA) decided to address this challenging issue in phases with phase I focusing on point sources and Phase II focusing on regional/area sources. EPA has now determined that the visibility goal should be attained by 2064. Furthermore, demonstrable progress toward this goal needs to be assessed periodically. One of the point sources examined during phase I was the Navajo Generating Station (NGS) located at the eastern end of the Grand Canyon ⬃6 mi east of Page, AZ. This 2250 MW coal-fired power plant is operated by Salt River Project (SRP). After several research studies to determine the impact of the plant on Class I areas, in 1990, SRP and its five co-owners agreed to reduce SO2 from the combustion of its nominal low-sulfur (0.5%) coal by 90%. Before reductions occurred, NGS, a base-loaded plant, typically emitted ⬃200 t/day of SO2. Scrubbers were installed on three generating units at NGS in compliance with EPA requirements. The first unit was placed in service in November 1997 and the third and last unit in August 1999. This paper examines changes in haze and the particulate sulfur (Sp) in the nearby Grand Canyon associated with these SO2 emission reductions. Using a variety of descriptive and statistical analytical techniques, this paper illustrates the decreases in Sp and light extinction (increased visibility) in the GCNP, presumably in part from NGS SO2 reductions. The Geographical and Meteorological Setting Figure 1 illustrates the geographical setting of the NGS and other Class I visibility protected areas in this portion of the Colorado Plateau. The NGS is located at the eastern end of GCNP. In a straight line, the main visitor center at GCNP is ⬃100 km southwest of the plant. However, actual distances traveled through this meandering portion of the Colorado River in the Grand Canyon, to Indian Gardens, inside the canyon at 1166 m mean sea level (MSL), and Hopi Point, on the south rim at 2164 m MSL, is ⬃160 km. At the western end of the canyon is a sampling site at Meadview, which is 250 km straight-line from NGS but ⬃400 km along the river. The meteorology during the winter (December to February) is often quite consistent. Approximately 80% of the time during these months, the weather is fair. This fair weather results from a surface high-pressure center over the Colorado Plateau. The Colorado Plateau is both higher in elevation and colder than surrounding areas resulting in a persistent dome of surface high pressure for extended periods. Flow within the canyon 1676 Journal of the Air & Waste Management Association

Figure 1. Regional map of the Colorado Plateau showing the location of the NGS in relation to the GCNP and other Class I visibility protected areas. The aerosol, transmissometer, and photographic ambient monitoring sites are also identified.

is from high to low pressure, which results in downriver flow (generally east to west) under these conditions. Although the shallow surface flow is predominantly northeasterly, flows aloft such as at the effective plume height of the NGS (⬃600 m above ground level) are often lighter and more variable. The other notable feature during the winter is the presence of low-lying valley fogs in such places as within the Grand Canyon, which can persist for days once formed when these cold high-pressure domes are present. During the summer, May to September, the meteorology is also quite consistent. High pressure is the persistent feature in the upper atmosphere. This is associated with fair warm weather in the Western United States. However, during July and August, because of southerly flow aloft advecting moisture from the Gulf of Mexico and/or California, monsoonal moisture is common. This results in scattered afternoon thunderstorms throughout the region. The wind direction in the typically ⱖ3000-m deep afternoon transport/mixed layer is nearly always from the west to east. These thermal winds result from the intense heating of the deserts compared with the relatively cool Pacific Ocean. In addition, summer heating of the Colorado Plateau results in warmer temperatures than at equivalent elevations above mean sea level over lower elevation regions away from the plateau. Thus, during the summer, winds in the Grand Canyon are nearly always up canyon (generally from the west). Volume 55 November 2005

Green et al. Historical Studies NGS became fully operational in 1976. Multiple studies have been conducted to understand the fate of NGS emissions on nearby national parks and wilderness areas. Visibility Impairment because of Sulfur Transport and Transformation in the Atmosphere1,2 found that in dry conditions (the plume not interacting with clouds) sulfur gas-to-particle conversion rates ranged from 0 to 1% per hour. The Winter Haze Intensive Tracer Experiment (WHITEX)3 was conducted in January to February 1987. WHITEX results indicated that for favorable transport conditions, ⬃70 – 80% of the Sp on the rim at Hopi Point was attributed to NGS and that NGS emissions were present at Hopi Point ⬎70% of the time. SRP commissioned the NGS Visibility Study (NGSVS), which was conducted from January 10 to March 31,1990.4 At Hopi Point, ⬃9% of the Sp was attributed to NGS. However, during a single moist, cloudy (fog/stratus) condition, NGS was responsible for ⬃66% of the total Sp. This program estimated that for this entire period, emission products from NGS were present at Hopi Point ⬃33% of the time. Finally, it was estimated that 4 –7% of the sampling periods had high-relative humidity (⬎70%) with transport of emissions from NGS toward Hopi Point; these conditions would be conducive to impacts from NGS on Sp and visibility levels at Hopi Point. This study also noted that for certain meteorological conditions, Sp concentrations could be considerably higher inside the canyon at Indian Gardens than on the rim. The WHITEX and NGSVS studies highlight the meteorological differences that can occur among winters. For example, during NGSVS, there were many more clear, dry, stratus-free days with more westerly winds transporting NGS emissions away from the canyon compared with WHITEX. In 1992, Project Measurements of Haze and Visual Effects was conducted to examine the contributions of Mohave Power Project to haze in the Grand Canyon.5 Mohave Power Project is a 1580 MW coal-fired power plant located ⬃100 km south-southwest of and down river from the western end of the Grand Canyon. The winter-intensive study occurred from mid-January through mid-February. During this period, a perfluorocarbon tracer perfluoromethylcyclopentane (PMCP) was released continuously ⬃25 km northeast of NGS on Lake Powell at Dangling Rope (Figure 1). Transport winds were down river for the first 3 weeks or so of the 30-day study period. Above background, PMCP and, by inference, NGS plume products were often at Hopi Point, Indian Gardens, and Meadview. As an indication of the strong channeling of flow within the canyon, the PMCP tracer was observed Volume 55 November 2005

above background more frequently and with higher average concentration at Meadview than on the rim at Hopi Point.6 This occurred in spite of Meadview being about twice as distant from the tracer release point as Hopi Point. The Tracer-Aerosol Gradient Interpretive Technique (TAGIT)7 receptor model used the Dangling Rope tracer (PMCP) for Sp attribution to NGS. At Marble Canyon, TAGIT attributed 59 ⫾ 12% of the Sp (an average of 0.35 ⫾ 0.06 ␮g/m3) to NGS. This monitoring site inside the canyon is at the east entrance to the GCNP (Figure 1). The tracer was strongly channeled throughout the entire ⬎400 km length of the Grand Canyon. Additionally, a hybrid chemical mass balance model8 was used, which accounted and corrected for SO2 and Sp deposition. The sampling period was from January 27 to February 9, 1992. The NGS contribution to Sp averaged ⬃0.21 ␮g/m3 during these 2 weeks at Indian Gardens. At Meadview, NGS accounted for 59% (0.10 ␮g/m3) of the modeled Sp. The Current Study The Database. The Interagency Monitoring of Protected Visual Environments (IMPROVE) network9 affords an opportunity to examine particulate and visibility trends in Class I areas. IMPROVE began monitoring at Hopi Point in the winter of 1987 and at Indian Gardens in 1989. This long-term monitoring program has several objectives, including the routine monitoring and identification of aerosols, gases, and visibility conditions in Class I areas, such as the Grand Canyon. The aerosol data will be used to assess long-term trends and progress toward the national visibility goal. Particulate sulfur was formerly collected twice weekly and currently every third day on a Teflon filter and analyzed formerly by particle-induced X-Ray emission and currently by X-ray fluorescence. In this study, Sp was collected at Indian Gardens about halfway down inside the canyon and nearly directly below Hopi Point. Light extinction (bext) has been measured by a long-path transmissometer since 1986. This is a recording instrument measuring continuously for 10 min in each hour. The system consists of a constant output light source transmitter and a computer-controlled photometer receiver.10,11 Data from two transmissometer systems were used in this study. One is located on the Grand Canyon rim between Grandview and Moran Points and has a path length of 5.8 km. The other transmissometer has the receiver on the rim at the Yavapai museum at 2200 m MSL and the transmitter at Phantom Ranch in the canyon bottom at ⬃700 m MSL. The path length is 5.1 km. These rim locations are in the proximity of Hopi Point. The other data used in this analysis are 35-mm photographs shot automatically three times daily from Journal of the Air & Waste Management Association 1677

Green et al. Desert View looking west down the canyon toward Mt. Trumbull (Figure 1). Photos taken from 1990 through 2002 were manually examined for humanly perceptible hazes inside the canyon. Rationale and Approach for Descriptive Statistical Analysis of This Data. From the review of the previous field studies, it is quite apparent that in a dry, cloud-free environment, NGS converts sulfur oxides (SOx) to Sp at ⬍1%/hr. From the chemical mass balance apportionment analysis, even for travel times of ⱖ10 hr to Hopi Point and Indian Gardens, ⬍0.4 ␮g/m3 of Sp is produced. Given this, it is prudent to focus on and investigate those days and events when the air mass is stable and there are boundary layer clouds, stratus, and fogs between NGS and the receptor sites at the Grand Canyon. These meteorological conditions do occur as described above. An important question is how frequently and to what extent these atmospheric conditions occur. When these conditions exist, much faster aqueous-phase SOx conversion chemistry can occur. To address these questions, we have approached the data analyses in the following framework. Human and photographic observations clearly record the presence of these hazes, clouds, and fogs inside the canyon during the winter. During these periods, the rim often remains haze free and above this haze layer. Thus, it makes sense to examine the visibility data (bext) and Sp data as the difference between the in-canyon and the rim values. These differences, denoted as ⌬Tr and ⌬Sp, are then plotted as cumulative frequency. We have divided the calendar year into four seasons: winter (December to February), spring (March to April), summer (May to September), and fall (October to November). We have also divided the data into three periods, from 1989 to 1997 before the first scrubbers were installed, the transition period while the three scrubbers were being installed from 1997 to 1999, and then the postscrubber period from 1999 through 2003. We also consider cumulative frequency distributions of Sp and bext for the periods before and after scrubbers were installed, for both the rim and within-canyon measurements. This allows a simple assessment of Sp and Tr (transmissometer bext) before and after the scrubbers. The ⌬Tr and ⌬Sp analyses are done based on an assumption that within-canyon bext and Sp are more affected by NGS than are the rim values. Looking at Tr and Sp rather than the changes avoids this assumption. However, temporal changes in Tr and Sp, especially on the rim, that are because of NGS alone may be difficult to distinguish from changes because of regional emissions. RESULTS Figures 2-5 display the key results for ⌬Tr and ⌬Sp. We have shown that winters before and after scrubbers 1678 Journal of the Air & Waste Management Association

Figure 2. Cumulative frequency plots of ⌬Tr (the in-canyon transmissometer minus the rim transmissometer) for the winter (December to February). The period “before” is from 1989 until scrubbers were installed beginning in 1997. The “after” period is from completion of scrubber installation in 1999 through 2003. The 2-yr period of scrubber installation is not plotted, because of too few data points for more than 90th percentile cumulative frequency.

were installed, and, for comparison purposes, summers before and after scrubbers were installed for both ⌬Tr and ⌬Sp. We have shown summers so that we can compare the high values before and after the scrubbers were installed to the analogous winter periods. As discussed earlier, during the summer, the NGS plume is transported away or up-canyon ⬃90% of the time. In each figure, we have plotted the entire data set and then, separately, the highest values representing the last few percentiles. Figure 2 displays the ⌬Tr data during winters before and after scrubber installation. The greatest difference between these two periods is observed in the extremes. For example, before the scrubber installation, above the ninety-sixth percentile, ⌬Tr ranged from 30 to 91 and

Figure 3. Cumulative frequency plots of ⌬Tr for the summer (May to September) for the same years as in Figure 2. Volume 55 November 2005

Green et al.

Figure 4. Cumulative frequency plots of ⌬Sp (the in-canyon Sp minus the rim Sp) for the winter (December to February) for the same period as in Figure 2.

after from 12 to 30. Obviously, this is a marked decrease, and this decrease is observed less noticeably but well down the cumulative frequency plot to the eightieth percentile. Figure 3 illustrates the summer for the analogous period. In the bottom graphs, the few high points have been eliminated. These extreme high values are from various kinds of summer brush and forest fires common throughout the Western United States. Here we observe little difference between the more than ninety-sixth percentile plots for the two periods with ⌬Tr ranging from ⬃18 to 52 before and 18 to 32 after. Additionally, there is only a small difference between the ⌬Trs well down the cumulative frequency plot to the fiftieth percentile. Figure 4 displays the ⌬Sp data during the winter before and after scrubber installation. The greatest difference between these two periods is observed in the extremes. Before scrubber installation, above the ninetysixth percentile, ⌬Sp ranged from 0.4 to 1.46 and after from 0.2 to 0.46. This marked decrease is observed but less noticeable down to the eightieth percentile where the

Figure 5. Cumulative frequency plots of ⌬Sp for the summer (May to September) for the same period as in Figure 2. Volume 55 November 2005

Figure 6. Cumulative frequency plots of wintertime Tr for the rim to in-canyon transmissometer before and after scrubber installation.

difference is double. By the 70th percentile, the ⌬Sps are comparable before and after. Finally, Figure 5 illustrates the ⌬Sp data during the summer before and after scrubber installation. Once again, for the summer there is little difference between these two periods. For example, before scrubber installation, above the ninety-sixth percentile, ⌬Sp ranged from 0.1 to 0.6 and after from just under 0.1 to 0.32. As was the case for the ⌬Trs, one has to surmise that sulfur emissions are slowly decreasing in this part of the United States. However, in both winter and summer, the trends are subtle over time.12 From 1990 to 1999, Sp decreased ⬃0.13 ␮g/m3. Figures 6 and 7 show the cumulative frequency of winter daily averaged bext in the canyon and on the rim before and after scrubbers were installed. For the rim to in-canyon transmissometer, a shift in the frequency distribution to lower bext can be noted for the scrubbed period. For the rim transmissometer, there is little difference for the before and after scrubber installation periods.

Figure 7. Cumulative frequency plots of wintertime Tr for the rim transmissometer before and after scrubber installation. Journal of the Air & Waste Management Association 1679

Green et al.

Figure 8. Cumulative frequency plots of wintertime Sp for the in-canyon site before and after scrubber installation.

Figures 8 and 9 show the cumulative frequency of winter Sp in the canyon and on the rim before and after scrubbers were installed. The frequency distribution of Sp both on the rim and in the canyon is shifted toward lower values for the scrubbed period, with the in-canyon site showing a larger decrease than the rim site. The Tr and Sp plots suggest that the decreased emissions at NGS provided more benefits in-canyon than on the rim as we had expected. Statistical Significance of this Analysis An important question arises from the above analysis. Are these decreases observed during the winter statistically significant? We have performed a statistical analysis to address this question. The database was divided into four time periods based on the number of scrubbers operating: zero, one, two, or three. Only winter months were examined from December to February. Because we are interested in the “extreme” values, we calculated the number of days within each “scrubber” period when a given Sp

Figure 9. Cumulative frequency plots of wintertime Sp for the rim site before and after scrubber installation. 1680 Journal of the Air & Waste Management Association

was exceeded. This was accomplished for several threshold values, starting from a threshold value close to the median (⬃0.1 ␮g/m3) to a high value of 0.45 ␮g/m3 of Sp. In Tables 1 (Indian Gardens) and 2 (Hopi Point/ Hance Camp), the number and percent of observations exceeding several threshold values of Sp by number of scrubbers operating are shown. The P value gives the statistical significance for testing the null hypothesis that the proportion of days when the threshold value is exceeded is “independent” of the time period. If the P value is small (typically a value ⬍0.05), then the null hypothesis is rejected, and we conclude that there is, indeed, a difference. The direction of the difference is observed by looking at the percent of days for each time period that exceeded the threshold. If the scrubbers were effective at reducing Sp inside the canyon, then as scrubbers were added, we would expect the percent of days exceeding the threshold to decrease, particularly for high thresholds. This is usually true, although for the cases of 1 and 2 scrubbers, the number of observations is very small. We will use the threshold value of 0.4 ␮g/m3 in Table 1 as an illustrative example. The total number of days in each row is the same, regardless of the Sp threshold. Thus, before scrubbers were installed from 1989 until 1987, there were 164 data points. After all three of the scrubbers were installed, or from 1989 through 2003, there were 74 values. For the other two rows, during scrubber installation, there were only 23 and 8 values for each period. Before scrubbers were installed, 16% of the values exceeded the threshold, whereas the remaining 84% did not. After installation of all three of the scrubbers, only 3% of the sampled days exceeded the threshold implying that the remaining 97% did not. According to the P value, this decrease was statistically significant. In most cases, the P values are very small, indicating there is a demonstrable difference among time periods with respect to the proportion of days exceeding the threshold. The proportions are higher before any scrubbers were installed (0 scrubbers) and lower afterwards (three scrubbers). The statistical test is the “␹2 test for independence.” The statistical significance is not as strong at Hopi Point/Hance Camp compared with Indian Gardens, particularly for the higher threshold values of Sp. With our meteorological understanding, this is an expected result. The meteorological periods conducive to the highest yields of Sp occur when the atmosphere is stably stratified and there is haze, fog, or stratus inside the canyon, but not on top on the rim. CONCLUSIONS We have examined Sp and extinction levels at various key locations in GCNP. The data was stratified into four seasons: winter (December to February), spring (March Volume 55 November 2005

Green et al. Table 1. Statistical information (P values) for varying Sp threshold values and varying numbers of installed scrubbers on NGS. Threshold Value of Sp 0.10 ␮g/m3 0.15 ␮g/m3 0.20 ␮g/m3 0.25 ␮g/m3 0.30 ␮g/m3 0.35 ␮g/m3 0.4 ␮g/m3 0.45 ␮g/m3

P Value 0.00081 0.00119 0.00519 0.00519 0.00815 0.01272 0.00561 0.01937

(significant) (significant) (significant) (significant) (significant) (significant) (significant) (significant)

0 Scrubbers Days > Threshold nⴝ164 (%)

1 Scrubber Days > Threshold nⴝ23 (%)



130 (79) 93 (57) 70 (43) 53 (32) 39 (24) 31 (19) 27 (16) 22 (13)

13 (59) 7 (32) 6 (27) 4 (18) 3 (14) 2 (9) 0 (0) 0 (0)

2 (25) 2 (25) 1 (13) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)

48 (65) 25 (34) 16 (22) 8 (11) 5 (7) 3 (4) 2 (3) 2 (3)

Note: The S data is from Indian Gardens for the winter months December through February from 1989 to 2003.

to April), summer (May to September), and fall (October to November). We focused on the winter and summer periods for “before” and “after” scrubber installation. Our analyses included cumulative frequency plots of ⌬Sp and ⌬Tr, as well as on-rim and in-canyon Tr and Sp. We also performed a P value statistical analysis of the number of days of high Sp both inside the canyon and on the rim. After scrubbers were installed, the data clearly suggest that the frequency of hazy days decreased in-canyon and that the extreme Sp levels decreased by a factor of two to three at both Indian Gardens and Hopi Point. Sp levels have also decreased throughout the cumulative frequency distribution. Between the 50th and 70th percentiles, Sp has decreased ⬃0.1 ␮g/m3 (⬃33%) from before to after scrubber installation. During summer, Sp has decreased little, ⬍0.1 ␮g/m3, from before to after. Also, the frequency and magnitude of extreme values during the summer is lower. The patterns of winter compared with summer confirm our hypothesis of NGS emissions having their greatest impact on

the GCNP during winter and minimal impact during summer. Our analyses confirm that reducing NGS emissions has had a beneficial effect in decreasing winter haze and improving visibility in the GCNP. There has been a decrease in both Sp concentrations and extinction during the winter. The haziest days with highest Sp have markedly decreased since the reduction of NGS emissions. Because of interannual meteorological variability, NGS impact is expected to be quite variable. However, we have suggested that the frequency of extreme values of Sp, particularly in the Grand Canyon, has markedly decreased since the installation of scrubbers at NGS. Future studies will continue to examine this postscrubber trend and determine whether our initial conclusions continue to be validated. ACKNOWLEDGMENTS A special thanks to Don Blumenthal of Sonoma Technology, Inc., for his insightful discussions and reference material on the history and pertinent findings of the NGS research studies during the past 25 yr. A special thanks to

Table 2. Statistical information (P values) for varying Sp threshold values and varying numbers of installed scrubbers on NGS. Threshold Value of Sp 0.05 ␮g/m3 0.10 ␮g/m3 0.15 ␮g/m3 0.20 ␮g/m3 0.25 ␮g/m3 0.30 ␮g/m3 0.35 ␮g/m3 0.4 ␮g/m3 0.45 ␮g/m3

P Value 0.30601 0.00016 0.00003 0.00002 0.00165 0.02150 0.07621 0.12756 0.17378

(not significant) (significant) (significant) (significant) (significant) (significant) (borderline) (not significant) (not significant)


1 Scrubber Days > Threshold nⴝ22 (%)



150 (91) 108 (66) 78 (48) 54 (33) 34 (21) 19 (12) 13 (8) 8 (5) 7 (4)

19 (86) 8 (36) 6 (27) 5 (23) 4 (18) 3 (14) 2 (9) 2 (9) 0 (0)

8 (100) 2 (25) 2 (26) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)

63 (85) 31 (42) 12 (16) 3 (4) 1 (1) 0 (0) 0 (0) 0 (0) 0 (0)

Note: The S data comes from the south rim of Hopi Point and Hance Camp for winter months December through February from 1989 to 2003. Volume 55 November 2005

Journal of the Air & Waste Management Association 1681

Green et al. Marc Pitchford of National Oceanic and Atmospheric Administration for his many helpful and creative suggestions for data analysis. REFERENCES 1. Richards, L.W.; Anderson, J.A.; Blumenthal, D.L.; Brandt, A.A.; McDonald, J.A.; Waters, N.; Macias, E.S.; Bhardwaja, P.S. The Chemistry, Aerosol Physics, and Optical Properties of a Western Coal-Fired Power Plant Plume; Atmos. Environ. 1981, 15, 2111-2134. 2. Blumenthal, D.L.; Richards, L.W.; Macias, E.S.; Bergstrom, R.W.; Wilson, W.E.; Bhardwaja, P.S. Effects of a Coal-Fired Power Plant and Other Sources on Southwestern Visibility (Interim Summary of EPA’s Project VISTTA); Atmos. Environ. 1981, 15, 1955-1969. 3. Malm, W.; Gebhart, K.; Latimer, D.; Cahill, T.; Eldred, R.; Pielke, R.; Stocker, R.; Watson, J. Winter Haze Intensive Tracer Experiment (WHITEX); National Park Service: Ft. Collins, CO, 1989; available at: http://vista.cira.colostate.edu/improve/Studies/WHITEX/Studywhitex. htm (accessed September 27, 2005). 4. Richards, L.W.; Blanchard, C.L.; Blumenthal, D.L., Eds. Navajo Generating Station Visibility Study-Final Report; STI,-90200-1124-FR; Sonoma Technology Inc.: Santa Rosa, CA, 1991. 5. Pitchford, M.; Green, M.; Tombach, I.; Malm, W.; Farber, R.; Mirabella, V. Measurement of Haze and Visual Effects (MOHAVE)-Project MOHAVE Final Report; 1999; available at http://vista.cira.colostate.edu/ improve/Studies/MOHAVE/Studymohave.htm (accessed September 27, 2005). 6. Pitchford, M.; Green, M.; Kuhns, H.; Farber, R. Characterization of Regional Transport and Dispersion Using Project MOHAVE Tracer Data; J. Air & Waste Manage. Assoc. 2000, 50, 733-745. 7. Kuhns, H.; Green, M.; Pitchford, M.; Vasconcelos, L.; White, W.; Mirabella, V. Attribution of Particulate Sulfur in the Grand Canyon to a Specific Point Source using Tracer-Aerosol Gradient Interpretive Technique (TAGIT); J. Air & Waste Manage. Assoc. 1999, 49, 906-915. 8. Eatough, D.J.; Farber, R.J.; Watson, J.G. Second Generation Chemical Mass Balance Source Apportionment of Sulfur Oxides and Sulfate at the Grand Canyon during the Project MOHAVE Summer Intensive; J. Air & Waste Manage. Assoc. 2000, 50, 759-774.

1682 Journal of the Air & Waste Management Association

9. Malm, W.C.; Schichtel, B.A.; Pitchford, M.L.; Ashbaugh, L.L.; Eldred, R.A. Spatial and Monthly Trends in Speciated Fine Particle Concentration in the United States; J. Geophys. Res. 2004, 109, D03306. 10. Molenar, J.V.; Dietrich, D.L.; Tree, R.M. Application of a Long Range Transmissometer to Measure the Ambient Extinction Coefficient in Remote Pristine Environments; In Visibilty and Fine Particles; Mathai, C.V., Ed.; Air & Waste Management Association: Pittsburgh, PA, 1989; 305-317. 11. Molenar, J.V.; Persha, G.C.; Malm, W.C. Long-Path Transmissometer for Measuring Ambient Extinction; In Environment and Pollution Measurement Sensors and Systems; Nielson, H.O., Ed.; The International Society for Optical Engineering: Bellingham, WA, 1990; 1269, 37-55. 12. Malm, W.C.; Schichtel, B.A.; Ames, R.B.; Gebhart, K.A.; A 10-Year Spatial and Temportal Trend of Sulfate across the United States; J. Geophys. Res. 2002, 107, 4627.

About the Authors Mark Green is a research professor with the Desert Research Institute. Rob Farber, a senior research scientist, and Nghi Lien, a research scientist, are with Southern California Edison. Kristi Gebhart is a physical scientist with the National Park Service. John Molenar is with Air Resource Specialists. Hari Iyer is a professor with Colorado State University. Delbert Eatough is a professor with Brigham Young University. Address correspondence to: Mark Green, Division of Atmospheric Sciences, Desert Research Institute, 755 E. Flamingo Rd., Las Vegas, NV 89119; phone: ⫹1-702-862-5445; fax: ⫹1-702-862-5507; e-mail: [email protected].

Volume 55 November 2005