Secondary Pollutant Formation in the Lake Tahoe Basin...Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop .....5-65. 6 . 1. Introduction and Statement

Secondary Pollutant Formation in the Lake Tahoe Basin

Project No. P075

Final Report

Prepared for:

U.S. Department of Agriculture Forest Service

Tahoe Science Program

Southern Nevada Public Lands Management Act (SNPLMA)

Prepared by:

Barbara Zielinska (PI)1, Andrzej Bytnerowicz (co-PI)2, Alan Gertler (co-PI) 1

Wendy Goliff3 , Sandra Theiss1, Joel Burley4, Mark McDaniel1 and Dave Campbell1

1Desert Research Institute, Reno, NV 89512 2US Forest Service, Pacific Southwest Research Station, Riverside, CA 92507

3University of California, Riverside, CA 4St. Mary’s College, Moraga, CA 94556

January 30, 2015

Contents 1. Introduction and Statement of the Problem .......................................................................................... 7

1.1 Current understanding of secondary pollutant formation ............................................................. 7

1.2 Atmospheric air quality models .................................................................................................. 10

2. Objectives and Hypotheses ................................................................................................................. 11

3. Methodology ....................................................................................................................................... 12

3.1 Monitoring sites and measurement periods................................................................................. 12

3.2 Monitoring methods .................................................................................................................... 14

3.2.1 Continuous measurements .................................................................................................. 14

3.2.2 Time-integrated measurements ........................................................................................... 14

3.3 Analytical methods ..................................................................................................................... 15

3.3.1 Laboratory analysis of canister samples ............................................................................. 15

3.3.2 Laboratory analysis of carbonyl compounds ...................................................................... 16

3.3.3 Laboratory analysis of time-integrated filter samples ......................................................... 16

3.3.4 Passive and honeycomb denuder samples ........................................................................... 17

3.3.5 Quality assurance and data validation ................................................................................. 18

3.4 Modeling approach ..................................................................................................................... 18

3.4.1 Modeling approach by CE-CERT ....................................................................................... 18

3.4.2 Modeling approach at DRI .................................................................................................. 19

4. Results ................................................................................................................................................. 21

4.1 Meteorological conditions during the measurement period ........................................................ 21

4.2 Ozone and oxides of nitrogen ..................................................................................................... 23

4.2.1 Continuous ozone and NOx instrumental inter-comparisons ............................................. 23

4.2.2 Ozone data during the field study ....................................................................................... 25

4.2.3 Comparison of ozone data with other sites in the Basin and Central Valley ...................... 26

4.2.4 Oxides of nitrogen data ....................................................................................................... 28

4.3 Gaseous organic compounds ....................................................................................................... 30

4.3.1 Volatile organic compounds from canisters ........................................................................ 30

4.3.2 Carbonyl compounds .......................................................................................................... 35

4.4 Particulate species ....................................................................................................................... 37

4.4.1 Organic and elemental carbon ............................................................................................. 37

2

4.4.2 Speciated particulate polar organic compounds ............................................................... 4-39

4.5 Gas- and particle phase nitrogenous and sulfur species ........................................................... 4-42

4.5.1 Passive samples ................................................................................................................ 4-42

4.5.2 Honeycomb denuder/filter pack systems ......................................................................... 4-43

4.6 Hysplit trajectories ................................................................................................................... 4-56

4.7 Modeling results ....................................................................................................................... 4-58

4.7.1 CE-CERT ......................................................................................................................... 4-58

4.7.2 DRI ................................................................................................................................... 4-61

5. Discussion ........................................................................................................................................ 5-63

6. Acknowledgements: ......................................................................................................................... 6-66

7. Literature cited ................................................................................................................................. 7-66

3

Table of Figures

Figure 1-1.The Central Valley of California, with the San Joaquin Valley in the southern sub-region, and the Sacramento Valley in the northern sub-region, and the Lake Tahoe Basin. ........................................... 7

Figure 3-1. Map of selected sampling locations: UH – Upper Homewood, top of Homewood Ski Resort, 7880’ (2402 m); LH – Lower Homewood, near lake shore, 6225’ (1897 m); HV – Heavenly Ski Resort Sky Deck, 8548’ (2605 m); TRPA – roof of Tahoe Regional Planning Agency building in South Lake Tahoe, 6411’ (1954 m) ............................................................................................................................... 12

Figure 3-2 Monitoring sites: (a) UH and LH, (b) HV. Stars denote sampling locations. .................. 13

Figure 3-3.Sampling sites at: TRPA rooftop (a), Heavenly (b), Upper Homewood (c) and Lower Homewood (d). VOC and carbonyl sampling equipment at the Lower Homewood was located inside the DRI van. ...................................................................................................................................................... 14

Figure 3-4. Modeling domains .................................................................................................................... 19

Figure 3-5. The domains covered by the Weather Research and Forecasting Model (WRF) .................... 20

Figure 4-1. Maximum, average and minimum temperatures recorded in Tahoe City during the study period. ......................................................................................................................................................... 21

Figure 4-2. Wind speed and direction recorded in Tahoe City during the study period ............................. 22

Figure 4-3. Maximum, average and minimum temperatures recorded in South Lake Tahoe during the study period. ................................................................................................................................................ 22

Figure 4-4. Wind speed and direction recorded in South Lake Tahoe during the study period .................. 23

Figure 4-5. Ozone data comparisons between 2B monitors and TRPA Thermo Scientific monitor .......... 24

Figure 4-6. NO and NO2 data comparisons between 2B monitor and TRPA Thermo Scientific monitor. 24

Figure 4-7. Hourly ozone data for lower (LH) and upper (UH) Homewood (a), Heavenly (HV) and TRPA rooftop (b). .................................................................................................................................................. 25

Figure 4-8. Average diurnal ozone concentrations (July 21 - 26, 2012) ..................................................... 26

Figure 4-9. Peak hourly ozone concentrations at Central Valley and Lake Tahoe Basin monitoring sites during the high ozone day (a) and mean daily ozone on the same day (b) ................................................. 27

Figure 4-10. Average diurnal ozone concentrations (July 21 - 26, 2012) at Central Valley and Tahoe Basin monitoring sites ................................................................................................................................. 28

Figure 4-11. NO2 (a) and NO (b) hourly data for all four sites: LH, UP, HV and TRPA. .......................... 29

Figure 4-12 Averaged diurnal NO/NO2 concentrations at all four monitoring sites: UH, LH, HV and TRPA .......................................................................................................................................................... 29

Figure 4-13 Canister VOC measured during the July 20-26 period at four Lake Tahoe Basin locations: LH, UH, HV and TRPA .............................................................................................................................. 33

Figure 4-14 Volatile Organic Compounds (VOC) averaged over 3-daily monitoring periods (a) in ppbv; and in percentage contribution to total VOC (b) ......................................................................................... 34

4

Figure 4-15 Concentrations of biogenic species, isoprene and α-pinene (a) and anthropogenic species benzene, toluene, ethylbenzene and xylenes (BTEX) (b) averaged over 3-daily monitoring periods. ...... 34

Figure 4-16 Benzene concentrations at all sampling sites and sampling periods. ...................................... 35

Figure 4-17 Concentrations of carbonyl compounds measured at all sites and sampling periods (a); (b) with acetone and methyl-ethyl-ketone (MEK) removed ............................................................................. 36

Figure 4-18 Concentrations of aldehydes averaged over monitoring periods. ........................................... 37

Figure 4-19 PM2.5 mass, OC and EC concentrations measured during July 21-26 sampling period at four Lake Tahoe Basin locations (a); percentage of total carbonation aerosol (TC) in the PM2.5 mass at these locations. ..................................................................................................................................................... 38

Figure 4-20 OC and EC concentrations for combined night and day samples from four Lake Tahoe Basin locations: LH, UH, HV and TRPA. Night samples from July 22-23 were kept separately. ................... 4-39

Figure 4-21 Concentrations of polar organic particulate matter species for LH = Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop ...................................................... 4-41

Figure 4-22. Concentrations of 2-methyltetrols and cis-pinonic acid in all sampling locations: LH = Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop .............. 4-42

Figure 4-23. Nitric oxide (a) and nitrogen dioxide (b) concentrations during the July 21-26 period at four Lake Tahoe Basin locations: TRPA rooftop (TRPA), Heavenly Sky Deck (HV), Lover Homewood (LH), and Upper Homewood (UH). ................................................................................................................... 4-43

Figure 4-24. Nitric acid (a) and sulfur dioxide (b) concentrations during the July 21-26 period at four Lake Tahoe Basin locations: TRPA, HV, LH, and UH. .......................................................................... 4-43

Figure 4-25. Ozone concentrations during the July 21-26 period at four Lake Tahoe Basin locations: TRPA, HV, LH, and UH .......................................................................................................................... 4-43

Figure 4-26. Ammonia concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. ............................................................. 4-45

Figure 4-27. Nitrous acid concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. ..................................... 4-47

Figure 4-28. Nitric acid concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. ............................................................. 4-48

Figure 4-29. Particulate ammonium concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. .............. 4-50

5

Figure 4-30 Particulate nitrate concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. ..................................... 4-52

Figure 4-31. Concentrations of all measured nitrogenous pollutants at four study sites. ........................ 4-52

Figure 4-32. Apportionment of nitrogenous air pollutants (%) at individual sites: a. TRPA; b. HV; c. LH; and d. UH. ................................................................................................................................................ 4-53

Figure 4-33. SO2 concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. ...................................................................................................................... 4-54

Figure 4-34. . Particulate sulfate concentrations determined TRPA, HV, LH and UH during July 21 – 26, 2012. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment. ..................................................................................................... 4-56

Figure 4-35. HYSPLIT trajectories for Homewood sites ....................................................................... 4-57

Figure 4-36. HYSPLIT trajectories for TRPA and Heavenly sites .......................................................... 4-58

Figure 4-37. Comparison of WRF temperature output to observations at Lake Tahoe airport. .............. 4-59

Figure 4-38. Comparison of WRF humidity output to observations at Lake Tahoe airport. ................... 4-59

Figure 4-39. SOA formation by source category. SOA1 and SOA2 are generated by aromatic compounds, SOA3 and SOA4 by isoprene, SOA5 and SOA6 from terpenes, and SOA7 from sesquiterpenes .......................................................................................................................................... 4-60

Figure 4-40. Nitrogen apportionment predicted by CAMx. .................................................................... 4-60

Figure 4-41. Comparison of measured and modeled temperatures in Sacramento (a) and Reno (b) ...... 4-61

Figure 4-42. Comparison of measured and modeled ozone concentrations at T-Street in Sacramento. .. 4-62

Figure 4-43. Comparison of measured and modeled NO concentrations for Upper Homewood (a), Lower Homewood (b), TRPA (c) and Heavenly (d) ........................................................................................... 4-63

Figure 5-1. Comparing observed (blue diamonds) and modeled ozone levels with all emissions (red squares), no anthropogenic emissions (green diamonds), and no biogenic emissions (purple x) for the TRPA site on July 20, 2012. .................................................................................................................... 5-64

Figure 5-2. Toluene/benzene and m,p-xylene/benzene ratios for all sites and measuring periods. LH = Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop .............. 5-65

6

1. Introduction and Statement of the Problem

The Lake Tahoe Basin is situated between the states of California and Nevada, east of the Central Valley of California, which includes San Joaquin and Sacramento Valleys and the metropolitan areas of Sacramento, Fresno, and San Francisco (Figure 1-1). Lake Tahoe is a high altitude lake at 1900 m (6200 ft), and is separated from the Central Valley by the Sierra Nevada divide, ranging from 2200 m (7200 ft) at the passes to 3050 m (10000 ft) at the summit of the Crystal Range. The prevailing westerly winds in this region have a potential to transport polluted air masses from the Central Valley area. However, the Sierra Nevada Mountain Range acts as a barrier to reduce the potential impact of transported primary emissions and their reaction products, but the complex topography of the Sierra Nevada and the Tahoe Basin make evaluations of pollution transport from the Central Valley and the western slopes of the Sierra Nevada to the Tahoe Basin difficult.

Figure 1-1.The Central Valley of California, with the San Joaquin Valley in the southern sub-region, and the Sacramento Valley in the northern sub-region, and the Lake Tahoe Basin.

1.1 Current understanding of secondary pollutant formation Secondary pollutants are formed by chemical reactions in the atmosphere from precursors that are

directly emitted from sources. Thus, the development of effective control strategies requires detailed knowledge of the nature of the precursors, their sources and the processes that lead to their formation. Ozone is one of the most important secondary criteria pollutants, and O3 concentrations sometimes exceed California’s 8-hr ambient air quality standards in the Lake Tahoe Basin (0.070 ppm). PM2.5, another criteria pollutant, is composed mostly of secondary pollutants including ammonium nitrate, ammonium sulfate, and secondary organic aerosol (SOA).

Ozone. O3 is not directly emitted from sources, but formed by chemical reactions in the atmosphere. In order to control O3 levels in the Lake Tahoe Basin, it is necessary to understand the underlying chemistry of O3 formation. The factors to consider are nitrogen oxides (NOx = NO + NO2) and volatile organic compounds (VOCs). During the daylight hours NO2 decomposes photochemically to

7

produce NO and atomic oxygen (equation 1), which reacts with O2 to form O3 (eq. 2). NO reacts with O3 to regenerate NO2 (eq. 3) in a process that is only limited by the availability of O3 . The ratio between NO and NO2 is governed by the photolytic rate J{NO2}. Since O3 is removed by the reaction with NO and is produced via the photochemical reaction, it seems that the net production of O3 would be NO2 -limited (eq. 4). However, through a series of reactions involving NO, VOCs and the hydroxyl radical (OH), additional NO2 is formed (eq. 5 through 9), which can subsequently photolyze to generate O3 via a chain reaction mechanism. This chain reaction is eventually terminated by a process that yields nitric acid (HNO3). Thus, the chemical processes involving NOx and VOCs that lead to O3 formation in the Lake Tahoe Basin can also result in the formation of HNO3, an important contributor to both overall nitrogen deposition and the formation of NH4NO3.

NO2 + hν O + NO (1)

O + O2 O3 (2)

O3 + NO NO2 + O2 (3)

[O3] = (4)

RH + HO R + H2O (5)

R + O2 RO2 (6)

RO2 + NO RO + NO2 (7)

RO + O2 HO2 + CARB (8)

HO2 + NO HO + NO2 (9)

Schematic 1. Processes leading to ozone formation. R=volatile organic compound (VOC); CARB=carbonyl compounds

Different types of hydrocarbons react at different rates and with a variety of oxidants based on their structure and composition. Each VOC species has its own unique reaction rate in the atmosphere. Some VOCs have an atmospheric lifetime of a few days (e.g., benzene), and contribute little to O3 formation on a local scale. Others, such as isoprene, have a lifetime of a few hours, and can lead to significant O3 formation on a local scale. The mechanism of degradation for each VOC is also important. VOCs such as benzaldehyde (formed from the degradation of toluene) consume radical species in the atmosphere, thereby reducing O3 formation. Conversely, VOCs such as formaldehyde and glyoxal form large numbers of radicals during their degradation in the atmosphere, increasing the rate of O3 formation. For these reasons, it is important for O3 modeling to have as much information regarding the composition of VOCs in the atmosphere as possible.

j [NO2]

k [NO]

8

Secondary Organic Aerosol (SOA). Organic carbon constitutes a large portion of PM2.5 in the basin (Engelbrecht et al., 2009) which is an important contributor to visibility degradation. Organic aerosols are complex mixtures of directly emitted or primary organic aerosol (POA) and secondary organic aerosols (SOA) derived from chemical reactions and gas-to-particle conversion of VOC emitted by both anthropogenic and natural sources. Recent field studies indicate that SOA is significantly more abundant than state-of-the-art SOA models predict (de Gouw et al., 2005; Johnson et al., 2006). Hence, the atmospheric relevance and contributions of the different SOA formation pathways and associated chemical reactions mechanisms still remain to be clarified (Fuzzi et al., 2005). Oxidation of biogenic emissions is believed to the dominant source of SOA globally, mostly by ozonolysis of terpenes (Jenkin, 2005). However, Volkamer et al (2006) showed that in the real urban atmosphere reactive anthropogenic VOCs produce significant amounts of SOA.

The contribution of biogenic and anthropogenic hydrocarbons to SOA in the Lake Tahoe Basin is presently unknown. Although we measured selected VOC with passive samplers during our summer 2010 field study at the 34 sites in the Lake Tahoe Basin (Bytnerowicz et al., 2013), the measurements were averaged over 2-week periods and the list of compounds was limited. For modeling SOA and ozone concentrations the more complete VOC species list and much better time resolution is needed.

Ammonium nitrate and sulfate. Ammonium nitrate (NH4NO3) and sulfate [(NH4)2SO4] are rather minor constituents of PM2.5 in the Lake Tahoe Basin (Engelbrecht et al., 2009), especially sulfates. However, due to the importance of nitrogen deposition to the Lake and potential artifact connected with nitrate measurements, we monitored precursors of NH4NO3 including ammonia, NOx and HNO3, using honeycomb denuder & filter pack systems (Koutrakis et al., 1993). We measured concentrations of NH3, HONO, HNO3, SO2 and fine particulate NH4, NO3 and SO4 with a resolution of several hours.

Ammonium nitrate aerosols (s) are formed from an equilibrium reaction involving gaseous (g) ammonia and nitric acid:

NH3(g) + HNO3(g) ↔ NH4NO3(s) (10)

Reaction (10) is dependent on temperature and as temperature decreases, the equilibrium shifts toward NH4NO3(s). Therefore, NH4NO3(s), or nitrate particle, formation occurs overnight when temperatures are lower. NH3 is directly emitted by motor vehicles (Fraser and Cass, 1998) and livestock operations in the eastern Basin, while HNO3 is a secondary pollutant. Nitric acid can be formed through a homogeneous gas phase reaction with nitrogen dioxide and hydroxyl radicals:

NO2 + HO → HNO3. (11)

Formation of nitric acid with reaction (11) occurs only during the daytime because HO radicals are predominantly present only during the day. HO radical production is driven by photochemistry. Two very important formation processes of HO in a polluted atmosphere such as California Central Valley are the photodissociation of HONO and photodissociation of H2O2.

HNO3 can also form in a series of reactions involving NO2, H2O, and O3. First ozone reacts with nitrogen dioxide to form NO3:

NO2 + O3 → NO3 + O2, (12)

9

and then the newly formed NO3 reacts with NO2 to produce N2O5:

NO3 + NO2 ↔ N2O5. (13)

N2O5 then reacts with liquid water on aerosol particles such as fog or cloud water droplets to form nitric acid:

N2O5 + H2O → 2HNO3. (14)

The production of HNO3 by the heterogeneous gas phase reaction of N2O5 and H2O is minimal during the day because NO3 quickly photodissociates soon after it is formed (Stockwell et al., 1997):

NO3 + hυ → NO + O2. (15)

However, during the nighttime, NO3 accumulates as NO2 reacts with O3 (Calvert and Stockwell, 1983).

1.2 Atmospheric air quality models Air quality models are developed for two purposes: 1) to determine the cause (or causes) of air

pollution events in the past, and 2) to predict air quality in future possible scenarios. Once previous air pollution events are fully explained and understood, one may predict future air quality with increased confidence. There are multiple components to air quality models that are needed to fully understand air pollution events: emissions, meteorological conditions, topographical features, and chemistry.

Emissions come from a variety of sources, which are divided into four categories: area (including airports, dry cleaners, and agricultural activities), point (e.g., smoke stacks), biogenic and mobile (onroad and offroad vehicles). Constructing and maintaining an accurate emissions database is a continuing challenge for the modeling community, requiring partnerships with a variety of research groups who collect emissions information in the field, (such as emissions from tailpipes from mobile sources or leaf-level emissions from varieties of oak or pine).

Air quality models require input from meteorological models such as the Weather Research and Forecasting (WRF) Model, a next-generation mesoscale numerical weather prediction system designed to serve both atmospheric research and operational forecasting needs. The WRF model provides vital input into air quality models such as temperature, relative humidity, wind speed and direction, inversion height, all of which influence air pollution events.

The chemistry portion of air quality models originates with laboratory data, which generates kinetic data, and field measurements which provide information on which chemical species are present in the atmosphere at a given time and location. Because explicit chemical mechanisms are not used in air quality models due to the amount of computer time required, lumped chemistry mechanisms such as the SAPRC Atmospheric Chemical Mechanism, Carbon Bond, and the Regional Atmospheric Chemistry Mechanism, version 2 (RACM2) provide the link between laboratory and field data and the air quality models. Chemical mechanisms are tested extensively, first test with chamber studies and later with field studies, before they are included in air quality models.

10

Air quality models currently use a grid-based, or Eulerian, model. The model area is divided into grids or boxes in the horizontal and vertical, and pollutant concentrations are calculated at fixed geographical locations at specified times based on their initial concentrations, new emissions, transport into and out of the box, dilution, and chemical reactions. Most Eulerian models assume that emissions are immediately mixed within the grid into which they are emitted. The two most common air quality models in use today are the Comprehensive Air Quality Model with Extensions (CAMx), which is maintained by Environ, and the Community Multi-scale Air Quality (CMAQ) modeling system, which is maintained by the Community Modeling and Analysis System (CMAS) under contract with the US EPA.

2. Objectives and Hypotheses

Goals, objectives, and hypotheses to be tested were as follows:

Main Goals:

1. Identify the precursors and pathways leading to the formation of secondary pollutants, including ozone, NH4NO3 and SOA

2. To employ the air quality model CAMx to predict the formation of O3, SOA and NH4NO3. Model output will be compared to observations made during the field campaign portion of this project to assess the model’s capabilities and potential biases.

3. Provide information for important policy decisions designed to reduce air and water impacts of atmospheric pollutants.

Specific Objectives:

1. Based on the results of the 2010 summer study (Bytnerowicz et al., 2010) select up to four sites in the Lake Tahoe Basin for the 5-days intensive air quality study

2. Conduct the air quality measurements in the Lake Tahoe Basin to quantify VOC, carbonyl compounds, NH3, HNO3, particulate NH4, NO3 & SO4, and PM2.5 with sub-daily resolution

3. Determine real-time concentrations of O3 and NO/NO2 in these sites with portable UV absorption monitors.

4. Employ the air quality model CAMx to predict O3 for the Lake Tahoe Basin.

5. Employ the air quality model CAMx to predict SOA and NH4NO3 for the Lake Tahoe Basin.

6. Communicate the results of this study to the managers and the public

Hypotheses to be tested:

1. The majority of precursors for O3 formation come from in-basin sources.

2. The majority of precursors for SOA formation come from in-basin sources.

11

3. Out-of-basin contribution to observed O3 and SOA levels is limited.

4. To control O3 formation, one needs to develop policies to limit NOx emissions.

5. Reduction of O3 levels will reduce the amount of SOA.

3. Methodology This study was planned for two years. The first year (2012/2013) was devoted to field measurements

and chemical analyses and the second year (2013/2014) to data processing, model development/validation and writing reports and scientific papers.

3.1 Monitoring sites and measurement periods Four sampling sites were established in the Lake Tahoe Basin (Figure 3-1). Two of these sites were

located at high elevation (one each on the western and eastern sides of the Basin) and two were positioned near Lake level. The two western sites were situated at the Homewood Ski Resort (Figure 3-2a); the low site was located near the south parking lot (elevation 6225’), and the high site was located at the top of the resort (elevation 7880’). The high site on the eastern side of the Basin was situated at the Heavenly Ski Resort Sky Deck at 8548’ elevation (Figure 3-2b), and the lower eastern site was the roof of the TRPA building (128 Market Street, Stateline, NV; 6411’ elevation). Figure 3-3 shows each of the four sites: (a) TRPA rooftop; (b) Heavenly; (c) upper Homewood; (d) lower Homewood. Volatile Organic Compounds (VOC) and carbonyl sampling equipment at the lower Homewood site was located inside the DRI van.

Technical equipment was deployed at field monitoring sites on July 19 and 20, 2012. The field study started at 0600 PDT on July 21 (Saturday) and concluded at 1730 PDT, July 26 (Thursday). We collected three samples per day: one in the morning during the period of rapid ozone accumulation (0600 to 0930), one during the period when maximum ozone concentrations typically occur (1000 to 1730) and one overnight (1800-0530).

Figure 3-1 Map of selected sampling locations: UH – Upper Homewood, top of Homewood Ski Resort, 7880’ (2402 m); LH – Lower Homewood, near lake shore, 6225’ (1897 m); HV – Heavenly Ski Resort Sky Deck, 8548’ (2605 m); TRPA – roof of Tahoe Regional Planning Agency building in South Lake Tahoe, 6411’ (1954 m)

12

a b

Figure 3-2 Monitoring sites: (a) UH and LH, (b) HV. Stars denote sampling locations.

The weather during the field monitoring period was generally very good, warm and sunny, with exception of Sunday night and Monday (July 22-23), when thunderstorms and intense rainfall occurred. One of our canister samplers on the TRPA roof got damaged during the second sampling period (1000 – 1730) on Monday and we had to replace it. Since we could not install a replacement before early Tuesday morning, and the heavy rain continued into Monday night, we decided to stop sampling at all sites for the third sampling period (1800 – 0530). We resumed the sampling at all 4 sites on Tuesday morning (0600). This gap in sampling was the reason we extended the overall length of the sampling period to 6 days instead of the initially planned 5 days.

The monitoring equipment was taken out of the sites on Thursday evening and Friday morning (July 26-27). All samples were delivered to the labs and stored according to established protocols.

13

Figure 3-3.Sampling sites at: TRPA rooftop (a), Heavenly (b), Upper Homewood (c) and Lower Homewood (d). VOC and carbonyl sampling equipment at the Lower Homewood was located inside the DRI van.

3.2 Monitoring methods

3.2.1 Continuous measurements Ozone and NO/NO2 were monitored continuously at all four sites. For ozone, portable UV absorption

2B Technologies Model 202 monitors were used at all sites except the TRPA building, where a Thermo Scientific Model 49 ozone analyzer was employed. For NOx, a 2B Technologies Model 410 / 401 NOx monitor was used at the lower Homewood site, Horiba APNA-360CE monitor at the upper Homewood site, TEI-17C NH3/NOx/NO analyzer at the Heavenly site (the ammonia channel was not operated), and Thermo Scientific Model 42C analyzer at the TRPA site. Prior to the commencement of the field study, a 24-hour inter-comparison of continuous ozone and NOx instruments used in this study was carried out on the TRPA rooftop on July 19-20, 2012.

3.2.2 Time-integrated measurements Hydrocarbons in the C2 to C12 range were collected using 6 L passivated stainless steel SUMMA

canisters, according to the EPA Method TO-14A. Carbonyl compounds were collected using Sep-Pak cartridges which had been impregnated with an acidified 2,4-dinitrophenylhydrazine (DNPH) reagent (Waters, Inc.) according to the EPA Method TO-11A. PM2.5 were collected using medium volume 2-channel filter samplers (113 Lpm sampling rate) with 47-mm Teflon (for PM2.5 mass) and quartz filters for organic and elemental carbon (OC/EC) measurements.

Honeycomb denuder/filter pack systems were used for collection of NH3, HONO, HNO3, SO2, and particulate NH4, NO3, and SO4. Acidic gases (HONO, HNO3 and SO2) were collected on a honeycomb denuder coated with carbonate, glycerin and methanol solution; NH3 on citric acid, glycerin and methanol

a b

c d

14

solution; and fine (<2.5 µm) particulate NH4, NO3 and SO4 on a filter pack consisting of Teflon, Nylasorb nylon and glass filters coated with citric acid, glycerin and methanol solution. Coarse particles (>2.5 µm diameter) were removed on an impaction plate. These systems were connected to steady flow pumps moving air through the assembly at 10 L/min (Koutrakis et al., 1993). Collection of samples with the honeycomb denuder/ filter pack systems was coordinated with all other air quality measurements (sampling times: morning [from 0600 to 0930 PDT], day [from 1000 to 1730 PDT] and night [1800 to 0530 PDT]. Due to the heavy rain during the night of July 23-24 sampling at all four sites was stopped for the 1800-0530 time period.

In addition, passive Ogawa samplers were used for monitoring ozone ambient concentrations (Koutrakis et al., 1993). Each sampler contained two cellulose filters coated with nitrite (NO2

-) which was oxidized by O3 to nitrate (NO3

-) and analyzed with ion chromatography. The rate of NO3- formation

(amount of NO3- formed on a filter over time of exposure) served as a measure of O3 concentration.

Ogawa passive samplers were also used for monitoring nitrogen oxides (NOx) and nitrogen dioxide (NO2). Each NO2 sampler contained two filters coated with triethylene amine (TEA) and each NOx sampler consisted of two filters coated with TEA and oxidizer. Both NOx and NO2 were collected as NO2

- . Passive samplers developed by the US Forest Service (Bytnerowicz et al., 2005) were used for nitric acid (HNO3) measurements. In the HNO3 sampler, ambient air passed through a Teflon membrane and gaseous HNO3 was absorbed on a Nylasorb nylon filter as NO3

–. Three replicate HNO3 samplers were exposed at each site.

3.3 Analytical methods

3.3.1 Laboratory analysis of canister samples Canister samples were analyzed for volatile organic compounds (VOC) using gas

chromatography/flame ionization detection/mass spectrometry (GC-FID/MS) system, according to EPA Method TO-15. The GC-FID/MS system includes a Lotus Consulting Ultra-Trace Toxics sample preconcentration system built into a Varian 3800 gas chromatograph with flame ionization detector (FID) coupled to a Varian Saturn 2000 ion trap mass spectrometer. The Lotus preconcentration system consists of three traps. Mid- and heavier weight hydrocarbons are trapped on the front trap consisting of 1/8” nickel tubing packed with multiple adsorbents. Trapping is performed at 55 ºC and eluting is performed at 200 ºC. The rear traps consist of two traps: empty 0.040” ID nickel tubing for trapping light hydrocarbons and a cryo-focusing trap for mid and higher weight hydrocarbons isolated in the front trap. The cryo-focusing trap is built from 6’ x 1/8” nickel tubing filled with glass beads. Trapping of both rear traps occurs at -180 ºC and eluting at 200 ºC. Light hydrocarbons are deposited to a Varian CP-Sil5 column (15m x 0.32mm x 1μm) plumbed to a column-switching valve in the GC oven, then to a Chrompack Al

2O

3/KCl column (25m x 0.53mm x 10μm) leading to the flame ionization detector for quantitation of

light hydrocarbons. The mid-range and heavier hydrocarbons cryo-focused in the rear trap are deposited to a J&W DB-1 column (60m x 0.32mm x 1μm) connected to the ion trap mass spectrometer. The GC initial temperature is 5 ºC held for approximately 9.5 minutes, then ramps at 3 ºC/min to 200 ºC for a total run time of 80 minutes.

Calibration of the system is conducted with a mixture that contained commonly found hydrocarbons (75 compounds from ethane to n-undecane, purchased from Air Environmental) in the range of 0.2 to 10 ppbv. Three point external calibrations are run prior to analysis, and one calibration check is run every 24 hours. If the response of an individual compound is more than 10% off, the system is recalibrated.

15

Replicate analysis is conducted at least 24 hours after the initial analysis to allow re-equilibration of the compounds within the canister. Minimum detection limit (MDL) is approximately 0.01 ppbv.

3.3.2 Laboratory analysis of carbonyl compounds After sampling, the DNPH-impregnated Sep-Pak cartridges are eluted with acetonitrile. The

hydrazones are separated and quantified per EPA Method TO-11A using a high performance liquid chromatograph (Waters 2690 Alliance HPLC System with 996 Photodiode Array Detector). An aliquot of the eluent is transferred into a 2-ml septum vial and injected with an autosampler into a Polaris C18-A 3µm 100 x 2.0 mm HPLC column. The following HPLC program was used for carbonyl compound analysis: solvent A: water (H2O); solvent B: acetonitrile (ACN), flow 0.2 mL/min, time: 0 - 10 min, isocratic 50% A, 50% B; 10-18 min gradient to 70% B; 18-20 min gradient to 100% B; 21 min 50% A, 50% B. Total run time 31 min. To correct for the acrolein rearrangement products (Tejada, 1986), a post-analysis reprocessing of the HPLC spectra was used, as described before (Fujita et al., 2011). Since DRI HPLC system is equipped with the photodiode array detector, the identification of carbonyl compounds is much more accurate than with standard UV/VIS detector. Also, the sensitivity of the analysis is enhanced by using the photodiode array detector.

3.3.3 Laboratory analysis of time-integrated filter samples Gravimetric Analysis. Unexposed and exposed Teflon-membrane filters are equilibrated at a temperature of 20±5 °C and a relative humidity of 30±5% for a minimum of 24 hours prior to weighing. Weighing is performed on a Cahn 31 electro microbalance with ±0.001 mg sensitivity. The charge on each filter is neutralized by exposure to a polonium source for 30 seconds prior to the filter being placed on the balance pan. The balance is calibrated with a 20 mg Class M weight and the tare is set prior to weighing each batch of filters. After every 10 filters are weighed, the calibration and tare are re-checked. If the results of these performance tests deviate from specifications by more than ±5 mg, the balance is re-calibrated. If the difference exceeds ±15 mg, the balance is recalibrated and the previous 10 samples are re-weighed. At least 30% of the weights are checked by an independent technician and samples are re-weighed if these check weights do not agree with the original weights within ±0.015 mg. Pre- and post-weights, check weights, and re-weights (if required) are recorded on data sheets as well as being directly entered into a database via an RS232 connection.

Elemental and Organic Carbon. Elemental carbon (EC) and organic carbon (OC) were measured by thermal optical reflectance (TOR) method using the IMPROVE (Interagency Monitoring of Protected Visual Environments) temperature/oxygen cycle (IMPROVE TOR) (Chow et al., 2005). Samples are collected in this method on quartz filters. A punch (0.5 cm2) from the filter sample is placed in the carbon analyzer oven such that the optical reflectance or transmittance of He-Ne laser light (632.8 nm) can be monitored during the analysis process. The filter is first heated under oxygen-free helium purge gas. The volatilized or pyrolyzed carbonaceous gases are carried by the purge gas to the oxidizer catalyst where all carbon compounds are converted to carbon dioxide. The CO2 is then reduced to methane, which is quantified by a flame ionization detector (FID). The carbon evolved during the oxygen-free heating stage is defined as “organic carbon”. The sample is then heated in the presence of helium gas containing 2 percent of oxygen and the carbon evolved during this stage is defined as “elemental carbon”. Some organic compounds pyrolyze when heated during the oxygen-free stage of the analysis and produce additional EC, which is defined as pyrolyzed carbon (PC). The formation of PC is monitored during the analysis by the sample reflectance. EC and OC are thus distinguished based upon the refractory properties

16

of EC using a thermal evolution carbon analyzer with optical correction to compensate for the pyrolysis (charring) of OC. Carbon fractions in the IMPROVE method correspond to temperature steps of 120oC (OC1), 250oC (OC2), 450oC (OC3), and 550oC (OC4) in a nonoxidizing helium atmosphere, and at 550oC (EC1), 700oC (EC2), and 850oC (EC3) in an oxidizing atmosphere. The IMPROVE method uses variable hold times of 150-580 seconds at each heating stage so that carbon responses return to baseline values.

Particulate organic compounds. After taking punches from quartz filters for OC/EC analysis, the remaining of the filters were extracted for polar organic compound analysis. Prior to extraction, the following internal standards were added to each sample: cholesterol-2,2,3,4,4,6-d6, levoglucosan-d7,

hexanoic-d11 acid, benzoic-d3 acid, adipic-d10 acid, suberic-d12 acid, homovanilic-2,2-d2 acid, tetradecanoic-d24 acid, eicosanoic-d39 acid, myristic-d27 acid, succinic-d4 acid, and phthalic 3,4,5,6-d4 acid. Filters were extracted with 150 mL acetone using the Dionex ASE for 15 min/cell at 1500 psi and 80°C. Extracts were concentrated to ~1ml by rotary evaporation at 35 °C under gentle vacuum, and filtered through a 0.2 µm PTFE disposable filter device (Whatman Pura discTM 25TF), rinsing the flask 3 times with 1 ml acetone each time. Filtrate is collected in a 4 mL amber glass vial for a total volume of ~4 mL. Approximately 200 µl of acetonitrile was added and the extract was then concentrated using a Pierce Reacti-Therm under a gentle stream of ultra-high purity (UHP) nitrogen with a water trap (Chrompack CP-Gas-Clean moisture filter 17971) to 100-200 µL. The fraction was derivatized using a mixture of bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% trimethylchlorosilane (TMS) and pyridine to convert the polar compounds into their trimethylsilyl derivatives for analysis of species of interest (Mazzoleni et al., 2007). The samples were analyzed by the electron impact (EI) GC/MS technique, using a Varian CP-3800 gas chromatograph (GC) equipped with a CP-8400 Autosampler and interfaced to a Vairan 4000 Ion Trap Mass Spectrometer (MS). Injections were 1 µl in size in the splitless mode onto a 30m long 5% phenylmethylsilicone fused silica capillary column (J&W Scientific type DB-5ms) 30m x 0.25mm x 0.25 mm. Identification and quantification of the analytes were made by Selected Ion Storage (SIS), by monitoring the molecular ions of each analyte and each deuterated analyte. Calibration curves were made by the molecular ion peaks of the analytes using the corresponding deuterated species as internal standards. If there is no corresponding deuterated species, the one most closely matching in volatility and retention characteristics is used. Six concentration levels for each analyte of interest were employed. One replicate analysis and one calibration check wad performed for every 10 injections of samples. If difference between true and measured concentrations exceeds ±20%, the system is recalibrated. During batch processing, calibration is performed before each batch.

3.3.4 Passive and honeycomb denuder samples Ogawa passive samples (O3, NOx and NO2) were extracted with water and analyzed by ion

chromatography, as described before (Bytnerowicz et al., 2007). Concentrations of all species were calculated using appropriate calibration curves. Concentrations of NO were calculated from a difference between NOx and NO2 concentrations. Nitric acid concentrations were determined by extracting nylon filter with water and analyzing the extracts by ion chromatography. Concentrations of HNO3 were calculated using calibration curves developed against the honeycomb denuder systems (Koutrakis et al., 1993).

Honeycomb denuders used for collection of HNO3, HONO and SO2 were extracted with a mixture of 0.002M Na2CO3 + 0.002M NaHCO3 solution and concentration of NO3

- , NO2- and SO4

2- in extracts were determined with ion chromatography (Dionex Model ICS-2000). Honeycombs used for collecting NH3

17

were extracted with ultrapure water and the extracts were analyzed for NH4+ colorimetrically (TRAACS

2000 Autoanalyzer). Teflon and nylon filters were extracted with the 0.002M Na2CO3 + 0.002M NaHCO3 solution and concentrations of NO3

- and SO42- were determined with ion chromatography (Dionex Model

ICS-2000). Glass filters were extracted with ultrapure water and extract concentrations of NH4+ were

determined colorimetrically (TRAACS 2000 Autoanalyzer).

3.3.5 Quality assurance and data validation DRI field operators conducted flowmeter calibration and system leak checks on the active filter samplers. The accuracy of the continuous gas monitoring instruments used in this project was checked at DRI prior to and following field use using certified gas standards. During the field study, the zero levels and air flow rates were checked daily and adjusted when necessary. Active sampling was carried out using widely accepted methods and following established SOPs (available upon request). All analytical data was reviewed and validated prior to calculation of summary values presented in this paper. Problems that occurred during sample collection or sample analysis were indicated using a system of coded data flags, and data from flagged samples were subjected to additional review and excluded where appropriate.

All analytical results were evaluated in terms of their associated measurement errors according to the following equation:

Uncertainty (Unc) = �(analyte concentration ∗ replicate precision)2 + (detection limit)2

Replicate precision for each analyte is determined by multiple injections (replicates) of at least ten per cent of all of the analyzed samples. Precision is then determined as:

𝑃𝑃 = C1−C2(C1+C2)/2

By this approach the analytical minimum detection limit (MDL) will determine the analyte uncertainty when sample concentrations approach zero. Similarly, the MDL will have little impact on the uncertainty of a higher concentration sample, where the concentration is many times the detection limit. In addition to this, the uncertainty in the volume flow is incorporated into the final concentration uncertainty by a similar sum of squares method. In this way the uncertainty represents the true uncertainty of each sample. Also, all samples are corrected for lot-specific sampling media blank values prior to the final concentration calculations. Data processing and reporting functions have been automated, using software developed by DRI, to reduce the incidence of calculation errors.

3.4 Modeling approach Two independent modeling approaches were utilized for this study. One approach, which was

outlined in our proposal and funded by SNPLMA program, was performed by CE-CERT, University of California, Riverside (Dr. Wendy Goliff, Principal Investigator). The other approach was undertaken by DRI as part of the Ph.D. Thesis of Sandra Theiss. Both approached are described below.

3.4.1 Modeling approach by CE-CERT The Weather Research and Forecasting Model (WRF) 3.4.1 was run at CE-CERT. Four domains

were used, with the parent domain at 54km, the first nested at 18km, the second nested at 6km and the innermost at 2km grid size resolution. The innermost domain covers the Lake Tahoe Basin and the

18

surrounding counties containing the suburbs east of Sacramento upwind and Reno and Carson City downwind (Figure 3-4). This domain covers 7 counties in California and 5 counties in Nevada.

Subsequently, the program Meteorology-Chemistry Interface Processor 4.2 (MCIP) was run in order to generate meteorological input files for the Sparse Matrix Operator Kernel Emissions version 3.51 (SMOKE) model and the Comprehensive Air Quality Model with Extensions (CAMx). In order to obtain emission input files for CAMx,, the Motor Vehicle Emission Simulator 2010b (MOVES) must be run for the mobile sources (onroad and offroad) ,then SMOKE must be run for all other sources (biogenic, area, point). The BEIS3 program (a SMOKE core program) was run to process biogenic emissions.

SMOKE was then run on all categories: stationary area, biogenic, non-road mobile, on-road mobile, and point emissions. The emissions inventory data, NEI 2011, was obtained from the EPA data clearinghouse (http://www.epa.gov/ttn/chief/net/2011inventory.html). After all the sources were completed, they were merged into a single model-ready file. Because SMOKE output is generated in netCDF format, the preprocessor for CAMx cmaq2camx must be run to convert the files to Fortran binary.

Once the emissions were completed, the preprocessors for CAMx which process WRF output, ozone column information, solar radiation, and initial and boundary conditions were run for the innermost nested domain simulated by WRF. Once these were completed, CAMx was run, beginning with three spin-up days to eliminate any effects from initial conditions. After CAMx was run, the post-processor CAMxPOST was run to evaluate of model performance. It is used to combine observations and predictions.

3.4.2 Modeling approach at DRI The Weather Research and Forecasting Model (WRF) 3.6.1 was run at DRI. Four domains were used

with the outermost at 27km, then 9km, 3km and the innermost at 1km grid size resolution. The innermost

Figure 3-4. The modeling domain is contained in the red box.

19

domain covers the area of Lake Tahoe and includes the cities of San Francisco, Sacramento, Reno and Carson City (Figure 3-5). This domain covers 34 counties in CA and 7 counties in NV.

Subsequently, the program Meteorology-Chemistry Interface Processor 4.1 (MCIP) was run in order to generate meteorological input files for the Sparse Matrix Operator Kernel Emissions 3.6 (SMOKE) model and the Community Multi-scale Air Quality 5.0.1 (CMAQ) model. In order to obtain emission input files for CMAQ, the Motor Vehicle Emission Simulator 2010b (MOVES) must be run for the mobile sources then SMOKE must be run for all other sources. MOVES can only be run on a single processing windows machine. It takes 12 hours per county, so we used the EMFAC emissions guide from the California Air Resource Board in order to compare similar counties and picking out a “representative county”, reducing our counties for MOVES to 7 on the CA side (representative counties are: Amador, Marin, San Mateo, San Francisco, San Juaquin, Fresno and Almeda) and 5 on the Nevada side (representative counties are: Washoe, Douglas, Carson City, Storey and Churchill).

SMOKE was then run on all categories: stationary area, non-road mobile, biogenic, on-road mobile (which includes rate per distance with refueling and non-refueling, rate per profile and rate per vehicle), and point emissions. The emissions inventory used was obtained from the EPA. The most recent emissions data, NEI 2011, was used. Then CMAQ was run.

Figure 3-5. The domains covered by the Weather Research and Forecasting Model (WRF)

20

4. Results

4.1 Meteorological conditions during the measurement period The closest meteorological station for the Homewood sampling sites was situated on the western

shore in Tahoe City and for TRPA and Heavenly sites in South Lake Tahoe (all data/graphs obtained from Western Regional Climate Center at DRI). At Tahoe City, the high temperatures during the measurement period ranged from 70 degrees Fahrenheit on 7/23 to 81 degrees Fahrenheit on 7/22. The low temperatures ranged from 42 degrees Fahrenheit on 7/25 to 51 degrees Fahrenheit on 7/22 and 7/23 (Figure 4-1). A line of thunderstorms associated with a cold front came through on 7/23 giving measureable precipitation for the two days of 7/23 (.5 inches) and 7/24 (.02 inches). This is what also helped reduce the daytime high temperature and increase the nighttime low temperatures.

Figure 4-1. Maximum, average and minimum temperatures recorded in Tahoe City during the study period.

The average wind direction for Tahoe City was from the north for the measurement period with wind speeds from 1.8 – 3.6 m/s. However the faster winds were typically from the west at 3.6 -5.8 m/s (Figure 4-2).

21

Figure 4-2. Wind speed and direction recorded in Tahoe City during the study period

At South Lake Tahoe, the high temperature during the measurement period ranged from 68 degrees Fahrenheit on 7/23 to 83 degrees Fahrenheit on 7/22. The low temperatures ranged from 37 degrees Fahrenheit on 7/25 to 52 degrees Fahrenheit on 7/23 (Figure 4-3). The cold front and line of thunderstorms came through on 7/23 also providing measureable precipitation for South Lake Tahoe for that day (.22 inches).

Figure 4-3. Maximum, average and minimum temperatures recorded in South Lake Tahoe during the study period.

The winds for South Lake Tahoe were primarily from the south and southwest with the strongest winds between 5.85 – 8.5 m/s (Fgure 4-4).

22

Figure 4-4. Wind speed and direction recorded in South Lake Tahoe during the study period

4.2 Ozone and oxides of nitrogen

4.2.1 Continuous ozone and NOx instrumental inter-comparisons Prior to the commencement of the field study, a 24-hour inter-comparison of the continuous ozone

and NOx instruments used in this study was conducted on the TRPA rooftop (July 19-20, 2012). The continuous ozone instruments included the 2B Technologies Model 202 monitors that were used at the Homewood and Heavenly sites, and a Thermo- Scientific Model 49 ozone analyzer operated by the TRPA. The NOx instruments consisted of the 2B Technologies Model 410/401 NOx monitor that was later deployed to the lower Homewood site and the Thermo Scientific Model 42 analyzer, operated by the TRPA.

The ozone results from the three 2B monitors are shown in Figure 4-5. These results are in excellent agreement with the TRPA ozone monitor. The NOx data from the 2B NOx monitor are shown in Figure 4-6, along with the TRPA NOx data. The nominal agreement between the two sets of NOx data (Figure 4-6) is not as good as the agreement observed for ozone (Figure 4-5), which reflects the fact that the 2B NOx monitor was operating near it’s detection limit of 3-4 ppb. The TRPA NO values remained at about 2-3 ppb until day-of-year 202.20 (~ 5 am PST on July 20), which is just below the effective detection limit for the 2B NO instrument. Thus, the 2B correctly measured baseline noise during this period. The spike in NO that occurred at day-of-year 202.25 (~6 am PST on Friday, July 20) was observed on both the TRPA monitor and the 2B. For NO2, the ambient concentrations were lower – they typically remained below the detection limit for the 2B. There may have been a small spike in NO2 at day-of-year = 201.83 that was observed on both instruments, but the 2.6 ppb value measured by the 2B monitor could also have been random baseline noise.

23

Figure 4-5. Ozone data comparisons between 2B monitors and TRPA Thermo Scientific monitor

Figure 4-6. NO and NO2 data comparisons between 2B monitor and TRPA Thermo Scientific monitor.

24

4.2.2 Ozone data during the field study Figure 4-7a shows the hourly ozone data for Upper and Lower Homewood and Figure 4-7b for

Heavenly and TRPA sites.

Figure 4-7. Hourly ozone data for lower (LH) and upper (UH) Homewood (a), Heavenly (HV) and TRPA rooftop (b).

As Figure 4-7 indicates, all sites show maximum ozone concentrations in the approximate range of 60 ppb during the daylight hours. The lowest ozone values were recorded on Sunday night and Monday (July 22-23), when thunderstorms and intense rainfall occurred. Figure 4-8 shows the diurnal ozone concentrations averaged over the monitoring period (July 21-26, 2012). Consistent with the observations from the previous study (Bytnerowicz et al., 2013; Burley et al, 2015) all four locations experience similar mid-day ozone maxima of ~ 50 to 60 ppb, which suggests that the Basin is well mixed during daytime hours. During the night there are large site-to-site variations; higher elevation sites (UH, HV) experience high ozone concentrations, while lower elevation sites experience much lower nocturnal ozone concentrations (LH, TRPA).

0

10

20

30

40

50

60

70

Jul-2

1 3 6 9 12 15 18 21Ju

l-22 3 6 9 12 15 18 21

Jul-2

3 3 6 9 12 15 18 21Ju

l-24 3 6 9 12 15 18 21

Jul-2

5 3 6 9 12 15 18 21Ju

l-26 3 6 9 12 15

ppb

(a)

LH

UH

Thunderstorm & Rainfall

0

10

20

30

40

50

60

70

Jul-2

1 3 6 9 12 15 18 21Ju

l-22 3 6 9 12 15 18 21

Jul-2

3 3 6 9 12 15 18 21Ju

l-24 3 6 9 12 15 18 21

Jul-2

5 3 6 9 12 15 18 21Ju

l-26 3 6 9 12 15

ppb

(b)

TRPA

HV

Thunderstorm &Rainfall

25

Figure 4-8. Average diurnal ozone concentrations (July 21 - 26, 2012)

Accordingly, the 24-hr average ozone concentrations are the highest at UH (53 ± 2.5 ppb), followed by HV (47 ± 6.7 ppb), LH (42 ± 8 ppb) and TRPA (40 ± 12 ppb).

4.2.3 Comparison of ozone data with other sites in the Basin and Central Valley The hourly ozone concentrations at our four monitoring sites were compared with those recorded in

other ozone monitoring sites in the Basin and in the Central Valley for the same monitoring period. The Central Valley sites included Sacramento at elevation 49’ (15 m); Folsom at 98’ (30 m); Placerville at 583’ (178 m); and Cool at 1551’ (473 m). The Tahoe Basin sites included Echo Summit at 2250’ (686 m); Bliss State Park at 6942’ (2116 m); Incline Village at 6420’ (1957 m); Tahoe City at 6288’ (1916 m) and Kings Beach at 6266’ (1919 m).

Figure 4-9a shows peak hourly ozone concentrations observed at Central Valley and Tahoe Basin monitoring sites during the high ozone day (July 26) and Figure 4-9b shows the mean daily ozone on the same day. As Figure 4-9a shows, the peak hourly ozone concentrations are lower in the metropolitan Sacramento area and getting higher in the downwind Folsom, Placerville and Cool areas. The sites in the Tahoe Basin area show peak ozone concentrations that are higher than Sacramento, but lower than Placerville or Cool. Mean daily ozone concentrations on July 26 are the highest at the Upper Homewood and Bliss State Park, consistent with the little changes in ozone concentrations during the whole 24-hr period at these high elevation sites.

26

Figure 4-9. Peak hourly ozone concentrations at Central Valley and Lake Tahoe Basin monitoring sites during the high ozone day (a) and mean daily ozone on the same day (b)

Figure 4-10 shows the diurnal ozone concentrations averaged over the monitoring period (July 21-26, 2012) for all of these sites. Placerville, Cool and Folsom show higher maximum hourly ozone concentrations than Basin sites, however this peak occurs later at the day and has much shorter duration;

a

b

27

the Tahoe Basin sites experience broad ozone maximum concentrations, lasting from approximately 1000 to 1800.

Figure 4-10. Average diurnal ozone concentrations (July 21 - 26, 2012) at Central Valley and Tahoe Basin monitoring sites

4.2.4 Oxides of nitrogen data Figure 4-11 shows the NO2 (a) and NO (b) hourly data for all four sites: LH, UP, HV and TRPA

rooftop.

28

Figure 4-11. NO2 (a) and NO (b) hourly data for all four sites: LH, UP, HV and TRPA.

Figure 4-12 shows the averaged diurnal concentrations of NO and NO2 at four sites.

Figure 4-12 Averaged diurnal NO/NO2 concentrations at all four monitoring sites: UH, LH, HV and TRPA

NO2 concentrations were generally low, with the highest hourly values observed for the TRPA rooftop, approaching 16 ppb. The NO2 and NO diurnal pattern for this site is somewhat different for

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

Jul-2

1 3 6 9 12 15 18 21

Jul-2

2 3 6 9 12 15 18 21

Jul-2

3 3 6 9 12 15 18 21

Jul-2

4 3 6 9 12 15 18 21

Jul-2

5 3 6 9 12 15 18 21

Jul-2

6 3 6 9 12 15

NO2_LH

NO2_UH

NO2_HV

NO2_TRPA

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

Jul-2

1 3 6 9 12 15 18 21

Jul-2

2 3 6 9 12 15 18 21

Jul-2

3 3 6 9 12 15 18 21

Jul-2

4 3 6 9 12 15 18 21

Jul-2

5 3 6 9 12 15 18 21

Jul-2

6 3 6 9 12 15

NO_LH

NO_UH

NO_HV

NO_TRPA

b

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

ppbv

NO_UH

NO2_UH

0

0.5

1

1.5

2

2.5

3

3.5

4

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

ppbv

NO_HV

0.0

0.5

1.0

1.5

2.0

2.5

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

ppbv

NO_LH

NO2_LH

0

1

2

3

4

5

6

7

8

9

10

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

ppbv

NO_TRPA

NO2_TRPA

a

29

weekend (July 21-22) and weekday (July 23-26) intervals. In addition, NO2 concentrations were influenced by the forest fire that started on the afternoon on July 22 (Sunday) in the Gardnerville area (NV), and continued through Monday, July 23. The maximum NO/NO2 concentrations at the TRPA site during weekdays usually occurred between 0600 -0900 PST, which coincides with rush hour traffic. There is also second NO peak observed during the evening hours, but NO concentrations are close to the detection limit of the NOx instrument. The upper Homewood site shows higher NO2 than NO concentrations, occurring during the daylight hours. The lower Homewood NO/NO2 concentrations are low and show an irregular pattern. NO2 data for Heavenly are most probably invalid, due to a converter malfunction in the instrument. The NO concentrations at Heavenly were generally below 10 ppb (with one high value of 25 ppb on July 21 at 1000 PST removed as a likely artifact) and showed a regular diurnal pattern – starting rising at around 0600, reaching maximum around noon and dropping to zero around 2000 PST.

4.3 Gaseous organic compounds

4.3.1 Volatile organic compounds from canisters Table 1 lists the canister VOC compounds quantified for this study together with their abbreviated

names. Figure 11 shows the concentrations of canister VOC species measured at four sites: Lower Homewood (LH), Upper Homewood (UH), Heavenly Sky Deck (HV) and TRPA roof (TRPA). The site names are followed by the sampling dates and sampling start times, for example LH7/21_06 means Lower Homewood, 7/21/2012, sampling period 0600 – 0930.

Table 1. List of the canister VOC quantified for this study together with their abbreviated names.

Abbreviation Compound Abbreviation Compound Abbreviation Compound

acetyl acetylene bu23dm 2,3-dimethylbutane hep2me 2-methylheptane

ethene ethene pena2m 2-methylpentane hep4me 4-methylheptane

ethane ethane pena3m 3-methylpentane hep3me 3-methylheptane

lprope propene p1e2me 2-methyl-1-pentene n_oct n-octane

lpropa propane n_hex n-hexane etbz ethylbenzene

lbud13 1,3-butadiene t2hexe t-2-hexene mp_xyl m&p-xylene

lbut1e 1-butene c2hexe c-2-hexene styr styrene

lc2but c-2-butene hxdi13 1,3-hexadiene (trans) o_xyl o-xylene

libute isobutylene mcypna methylcyclopentane n_non n-nonane

lt2but t-2-butene pen24m 2,4-dimethylpentane iprbz isopropylbenzene

lbutan n-butane benze benzene n_prbz n-propylbenzene

30

Abbreviation Compound Abbreviation Compound Abbreviation Compound

libuta iso-butane cyhexa cyclohexane a_pine alpha-pinene

lipent iso-pentane hexa2m 2-methylhexane m_etol 3-ethyltoluene

lnpent n-pentane pen23m 2,3-dimethylpentane p_etol 4-ethyltoluene

bud12 1,2-butadiene cyhexe cyclohexene bz135m 1,3,5-trimethylbenzene

pente1 1-pentene hexa3m 3-methylhexane o_etol o-ethyltoluene

b1e2m 2-methyl-1-butene

cpa13m 1,3-dimethylcyclopentane

n_dec n-decane

i_pren isoprene hep1e 1-heptene bz123m 1,2,3-trimethylbenzene

t2pene t-2-pentene pa224m 2,2,4-trimethylpentane

indan indan

c2pene c-2-pentene n_hept n-heptane detbz13 1,3-diethylbenzene

b2e2m 2-methyl-2-butene

p2e23m 2,3-dimethyl-2-pentene

detbz14 1,4-diethylbenzene

bu22dm 2,2-dimethylbutane

mecyhx methylcyclohexane n_bubz n-butylbenzene

cpente cyclopentene pa234m 2,3,4-trimethylpentane

n_unde n-undecane

cpenta cyclopentane tolue toluene

Canister VOC sample from Heavenly Sky Deck for July 22-23, 1800-0530 sampling period was void, since the canister sampler stopped during the night, due to the thunderstorm. Also, due to the heavy rain during the night of July 23-24 we stopped sampling at all four sites for 1800-0530 time periods.

As can be seen from Figure 4-13, the highest VOC concentrations are generally observed at the TRPA site, followed by the Lower Homewood site. Upper Homewood and Heavenly Sky Deck sites show much lower VOC concentrations, with highest concentrations observed on July 22-23. This may be due to the meteorological conditions - low clouds and some rain. Also, a wild fire started at the late afternoon on July 22nd around Minden - Gardnerville area, east of the Heavenly and TRPA sites. The smell of the wood smoke was evident at South Lake Tahoe at the evening of July 22nd. The canister sample from the TRPA roof site collected during this sampling period (1800 – 0530) shows the highest VOC concentrations.

31

Figure 4-14a shows the VOC concentrations averaged over 3-daily monitoring periods : 0600-0930, 1000 – 1730 and 1800 – 0530. For simplicity, the VOC data are combined into groups: C2 – C3 group contains acetylene, ethane, ethene, propane and propene; Aliphatic C4-C10 group contains all aliphatic hydrocarbons in the range of C4 to C10; Aromatic C6-C19 contains benzene, toluene, ethylbenzene, xylenes (BTEX) and higher alkyl-benzenes. Two biogenic hydrocarbons, isoprene and α-pinene are shown as separate species. Figure 4-14b shows the percentage contribution to total VOC for each of these groups. The sites are organized from West to East and from lower to higher elevation site.

Figure 4-15a shows isoprene and α-pinene concentrations and Figure 4-15b benzene, toluene, ethylbenzene and xylenes (BTEX) concentrations, respectively, averaged over 3-daily monitoring periods, as in Figure 4-14. The anthropogenic species, BTEX concentrations are lower than biogenic species concentrations.

32

Figure 4-13 Canister VOC measured during the July 20-26 period at four Lake Tahoe Basin locations: LH, UH, HV and TRPA

Lower Homewood

Upper Homewood

Heavenly

TRPA

33

Figure 4-14 Volatile Organic Compounds (VOC) averaged over 3-daily monitoring periods (a) in ppbv; and in percentage contribution to total VOC (b)

Figure 4-15 Concentrations of biogenic species, isoprene and α-pinene (a) and anthropogenic species benzene, toluene, ethylbenzene and xylenes (BTEX) (b) averaged over 3-daily monitoring periods.

As mentioned above, a wild fire started at the late afternoon on July 22nd around Minden - Gardnerville area, east of the Heavenly and TRPA sites. The smell of the wood smoke was evident at South Lake Tahoe at the evening of July 22nd. The concentrations of benzene, a species that is emitted in wood smoke was higher during this sampling period in all sampling sites, as shown in Figure 4-16 (note that the Heavenly VOC data are missing for this sampling period and TRPA data are missing for July 23, 1000-1800, due to samplers malfunction). However, benzene is also emitted in automotive exhaust (Zielinska et al., 2012).

a

b

a b

34

Figure 4-16 Benzene concentrations at all sampling sites and sampling periods.

4.3.2 Carbonyl compounds Table 2 lists carbonyl compounds quantified for this study together with their abbreviated names and

Figure 4-17a shows the concentrations of these species measured at all sites and sampling periods. The concentration of acetone is high at all sampling sites, which may be attributed to biogenic emissions of acetone and its photochemical production from VOC atmospheric reactions (terminal product). However, the very high concentrations of acetone in Heavenly site are probably due to proximity of this site to a ski area maintenance station, where acetone could be used as a cleaning solvent. When acetone is removed (Figure 4-17b) the most abundant carbonyls at all sites are formaldehyde and acetaldehyde.

Table 2. Carbonyl compounds quantified for this study together with their abbreviated names.

Abbreviation Compound Abbreviation Compound formal formaldehyde macrol methacrolein acetal acetaldehyde butal n-butyraldehyde acroln acrolein mek 2-butanone (MEK) glyoxl glyoxal valal valeraldehyde aceto acetone hexal hexaldehyde proal propionaldehyde benzal benzaldehyde croton crotonaldehyde tolual m-tolualdehyde

35

Figure 4-17 Concentrations of carbonyl compounds measured at all sites and sampling periods (a); (b) with acetone and methyl-ethyl-ketone (MEK) removed

a

b

Lower Homewood Upper Homewood

Heavenly

TRPA

36

Figure 4-18 shows the concentrations of aldehydes only, averaged over all monitoring periods. Among aldehydes, the highest concentrations were observed for formaldehyde and acetaldehyde. Methacroleine, an atmospheric photo-oxidation product of isoprene, was present in 0.1 – 0.2 ppbv concentration range. In addition, glyoxal, a di-aldehyde that may be present in biogenic emissions and a degradation product of aromatic compounds, was also measured.

Figure 4-18 Concentrations of aldehydes averaged over monitoring periods.

In general, the concentrations of carbonyl species are more uniform across sampling sites, with slightly higher concentrations at the lower sites, especially TRPA site. Formaldehyde is produced from the photochemical reactions of VOCs (see eq. 8), but it is also present in motor vehicle emissions (Zielinska et al., 2012).

4.4 Particulate species

4.4.1 Organic and elemental carbon Figure 4-19a shows PM2.5 mass, organic and elemental carbon (OC and EC) concentrations measured

at four sampling sites. As this figure shows, elemental carbon concentrations are very low, much lower than organic carbon concentrations. The highest OC and EC concentrations are observed on July 22 -23 nights sampling periods, especially for Heavenly and TRPA roof sites. This may be due to a wild fire that started at the late afternoon on July 22nd around Minden - Gardnerville area, east of the Heavenly and TRPA sites. Figure 4-19b shows the percentage of total carbonaceous aerosol (TC) in the PM2.5 mass. This percentage is quite variable, from 20% to up to 80%. Although samples from Heavenly and TRPA collected on 7/24 at 0600 – 930 PDT show higher than 100% contents of TC, this may be due to a very low PM2.5 mass for these two samples (below 5 µg/m3) and a possible sampling artifact due to VOC adsorption on quartz filters.

37

Figure 4-19 PM2.5 mass, OC and EC concentrations measured during July 21-26 sampling period at four Lake Tahoe Basin locations (a); percentage of total carbonation aerosol (TC) in the PM2.5 mass at these locations.

a

b

38

4.4.2 Speciated particulate polar organic compounds In order to have enough organic carbon mass for organic compound speciation, all daytime quartz

filter samples from each sampling location were combined (after removing 0.5 cm2 for OC/EC analyses). Likewise, all nighttime samples were combined for each site except for the night samples from July 22 -23 that were affected by the wildfire, which were analyzed separately. Figure 4-20 shows the average OC and EC data for all combined samples. As can be seen from this figure, the night samples from July 22-23 (collected from 1800 to 0530 PDT) show the highest OC and EC concentrations at each sampling location.

Figure 4-20. OC and EC concentrations for combined night and day samples from four Lake Tahoe Basin locations: LH, UH, HV and TRPA. Night samples from July 22-23 were kept separately.

The combined samples were extracted and analyzed by GC/MS, as described above in Section 3.3.3. Table 3 lists the target compounds, together with their abbreviated names.

Table 3. List of the polar particulate compounds quantified for this study together with their abbreviated names.

Abbrev. Compound Abbrev. Compound Abbrev. Compound Memalon me-malonic (d-c3)a Dimeo24 2,4-dimethoxybenzoic

acid Tricosac tricosanoic acid

Maleac maleic acid Dimeo25 2,5-dimethoxybenzoic acid

Tetraco tetracosanoic acid (c24)

Mesucac me-succinic acid (d-c4) Dimeo26 2,6-dimethoxybenzoic acid

Chol cholesterol

Sucac succinic acid (d-c4) Dimeo34 3,4-dimethoxybenzoic acid

Ergo ergosterol

Guai guaiacol Dimeo35 3,5-dimethoxybenzoic acid

Stigma stigmasterol

Guac glutaric acid (d-c5) Docosac docosanoic acid (c22) Bsit b-sitosterol Salcyl salicylic acid Homov homovanillic acid Sitost Sitosterol Megua4 4-me-guaiacol Syrald syringaldehyde Pentac pentanoic acid Meglu2 2-methylglutaric (d-c5) Cpinac cis-pinonic acid Hexac hexanoic acid (c6) Meglu3 3-methylglutaric acid (d-

c5) Azeac azelaic acid (d-c9) Heptac heptanoic acid (c7)

Dime235 2,3-and 3,5- Syrgac syringic acid Benac benzoic acid

4-39

Abbrev. Compound Abbrev. Compound Abbrev. Compound dimethylbenzoic acid

Dimeb24 2,4-dimethylbenzoic acid Sebac sebacic acid (d-c10) Octanac octanoic acid (c8) Dimeb25 2,5-dimethylbenzoic acid Undecdi undecanedioic acid (d-

c11) Otoluic o-toluic

Dimeb26 2,6-dimethylbenzoic acid Myrol myristoleic acid Mtoluic m-toluic Dimeb34 3,4-dimethylbenzoic acid Traum traumatic acid Nonac nonanoic acid (c9) Fguai4 4-formyl-guaiacol

(vanillin) Dodecd dodecanedioic acid (d-

c12) Ptoluic p-toluic

Etgua4 4-ethyl-guaiacol Undd111 1,11-undecanedicarboxylic acid

Decac decanoic acid (c10)

Syri syringol Palol palmitoleic acid Undecac undecanoic acid (c11) Hepdac heptanedioic (pimelic)

acid (d-c7) Dodd112 1,12-

dodecanedicarboxylic acid

Lauac dodecanoic (lauric) acid (c12)

Meadip3 3-methyladipic acid (d-c6)

Elac elaidic acid Tdecac tridecanoic acid (c13)

Mann mannosan Isster isostearic acid Myrac myristic acid (c14) Levg levoglucosan Hexdac hexanedioic (adipic)

acid (d-c6) Pdecac pentadecanoic acid

(c15) Alguai4 4-allyl-guaiacol

(eugenol) Dhabac dehydroabietic acid Palac palmitic acid (c16)

Iseug isoeugenol Paulust palustric acid Heptadac heptadecanoic acid (c17)

Isphac isophthalic acid Pimara pimaric acid Olac oleic acid Phthac phthalic acid Sandpim sandaracopimaric acid Steac stearic acid (c18) Acvan acetovanillone Abac abietic acid Ndecac nonadecanoic acid

(c19) Vanil vanillic acid Isopim isopimaric acid Ecosac eicosanoic acid (c20) Mesyr4 4-methyl-syringol Levopim levopimaric acid Octdecdi octadecanedioic acid Tdecen2 trans-2-decenoic acid Dhydpim dihydroisopimaric

acid Metrol 2-methylthreitol

Suber suberic acid (d-c8) Oxodeh7 7-oxodehydroabietic acid

Mytrol 2-methylerythritol

Dimeo23 2,3-dimethoxybenzoic acid

Hcosac heneicosanoic acid (c21)

ad=diacid (two carbonyl groups); c(x): number of carbon atoms in an acid

Figure 4-21 shows the more abundant species quantified in combined samples. Samples from the night of July 22-23 show the highest concentrations of particulate polar compounds, due to the very high abundance of biomass burning markers, namely levoglucosan, mannosan, and dehydroabietic acid (Mazzoleni et al., 2007).

4-40

Figure 4-21 Concentrations of polar organic particulate matter species for LH = Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop

Figure 4-22 shows the concentration of the secondary organic aerosol (SOA) tracers for volatile biogenic hydrocarbons, isoprene and α-pinene. Isoprene oxidation products include 2-methylerythritol and 2-methylthreitol, whereas cis-pinonic acid is formed from the reaction of α-pinene with ozone. As Figure 4-22 shows, all these species are present in the particulate matter samples collected at all sampling sites.

Dehydroabietic acid

Levoglucosan

4-41

Figure 4-22. Concentrations of 2-methyltetrols and cis-pinonic acid in all sampling locations: LH = Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop

4.5 Gas- and particle phase nitrogenous and sulfur species

4.5.1 Passive samples In order to compare the continuous monitoring data with average passive sampling results, passive

samplers for monitoring NO2, NOx, HNO3, O3 and SO2 were installed in all four monitoring sites and exposed during 6 days of sampling (July 21-26, 2012). Generally the determined pollutant concentrations were within the ranges determined at various Lake Tahoe Basin locations in the 2010 study (Bytnerowicz et al., 2012). Average NO and NO2 concentrations from passive samplers were much higher at TRPA (13.2 and 9.3 µg/m3, respectively) than in the three remaining sites (<8.3 and <5.7 µg/m3, respectively (Figures 4-23a and 4-23b). Nitric acid concentrations at the TRPA, Lower Homewood (LH), Upper Homewood (UH) sites were very low (1.2 – 1.75 µg/m3). The Heavenly Sky Deck (HV) HNO3 sampler was contaminated and therefore results from that site are not included (Figure 2a). Concentrations of SO2 were very low (<0.55 µg/m3) at all sites confirming typical background concentrations of this pollutant occurring in remote areas of the Sierra Nevada (Figure 4-24b). Ozone concentrations (Figure 4-25) were low at all sites (34.2-43.6 ppb) with the two higher elevation locations (UH and HV) showing higher values than the lower elevation sites (TRPA and LH). The O3 passive sampler results confirm the 2010 Tahoe study in which O3 concentrations were increasing with elevation.

4-42

Figure 4-23. Nitric oxide (a) and nitrogen dioxide (b) concentrations during the July 21-26 period at four Lake Tahoe Basin locations: TRPA rooftop (TRPA), Heavenly Sky Deck (HV), Lover Homewood (LH), and Upper Homewood (UH).

Figure 4-24. Nitric acid (a) and sulfur dioxide (b) concentrations during the July 21-26 period at four Lake Tahoe Basin locations: TRPA, HV, LH, and UH.

Figure 4-25. Ozone concentrations during the July 21-26 period at four Lake Tahoe Basin locations: TRPA, HV, LH, and UH

4.5.2 Honeycomb denuder/filter pack systems Gaseous ammonia, nitrous acid, nitric acid and particulate ammonium and nitrate. Measurements of the nitrogenous gaseous and aerosol pollutants were performed at four sampling sites in the Lake Tahoe Basin: TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH) and Upper Homewood (UH) during the July 21 – 26, 2012 period. The sampling times were: morning (from 0600 to 0930 PDT),

4-43

day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT). Due to the heavy rain during the night of July 23-24 (1800-0530) sampling was not conducted.

Ammonia concentrations were highly variable between the monitoring sites and periods of collection. The lowest concentrations occurred in the beginning of the study (July 21 – 22), with the remaining periods characterized by occasionally high values. The highest average concentration for the duration of the study was at HV (6.06 µg/m3), followed by UH (4.45 µg/m3), LH (2.92 µg/m3) and TRPA (2.85 µg/m3) (Table 4 and Figure 4-26a). In general NH3 concentrations at night were lower than those occurring in the morning and day. The only exception was a high night-time NH3 values at TRPA on July 22/23, which could be caused by emissions from the Garnerville Fire that affected the South Lake Tahoe area from the afternoon of July 22 until noon of July 23 (Figure 4-26 b).

Table 4. Ammonia concentrations (as N-NH3) determined at the TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH), and Upper Homewood (UH) during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods.

Ammonia (NH3) as µg N/m3 Date on Day part TRPA HV LH UH 21-Jul morning 1.57 3.40 3.55 3.00 21-Jul day 1.26 1.44 1.70 1.60 21-Jul night 0.78 1.13 1.11 1.60 22-Jul morning 1.88 2.95 2.86 2.76 22-Jul day 1.63 6.48 2.92 4.41 22-Jul night 7.94 1.51 1.35 4.74 23-Jul morning 3.78 12.00 3.27 12.93 23-Jul day 2.99 5.93 6.57 6.30 24-Jul morning 0.34 3.25 2.79 10.70 24-Jul day 3.03 10.24 3.79 5.27 24-Jul night 1.46 2.67 1.33 1.52 25-Jul morning 2.15 21.49 3.98 4.10 25-Jul day 2.20 3.95 2.28 3.10 25-Jul night 3.19 3.33 1.74 2.02 26-Jul morning 6.76 6.83 4.47 2.64 26-Jul day 4.65 10.31 2.90 4.55 Average 2.85 6.06 2.91 4.45

4-44

Figure 4-26. Ammonia concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

Nitrous acid (HONO) concentrations varied between the monitoring sites and periods of collection. At the TRPA elevated concentration’s occurred during the night of July 22/23 and July 23 morning period (0.298 and 0.281 µg/m3, respectively). Most likely those values as well as those elevated in the morning of July 23 at the remaining sites were caused by emissions from the Gardnerville Fire. Similarly, a high value at HV in the morning of July 25 could be related to a local fire observed in that area (Table 5, Figure 4-27 a). Generally, the lowest concentrations of HONO occurred in the daytime with the morning periods the highest, followed by the nighttime ones (Figure 4-27 b).

4-45

Table 5. Nitrous acid concentrations determined at the TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH), and Upper Homewood (UH) during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods.

Nitrous acid (HONO) as µg N/m3 Date on Day part TRPA HV LH UH 21-Jul morning 0.114 0.070 0.090 21-Jul day 0.035 0.029 0.006 0.036 21-Jul night 0.054 0.021 0.032 0.038 22-Jul morning 0.122 0.077 0.039 0.006 22-Jul day 0.049 0.045 0.055 0.073 22-Jul night 0.298 0.009 0.124 0.140 23-Jul morning 0.281 0.163 0.205 0.320 23-Jul day 0.080 0.064 0.131 0.107 24-Jul morning 0.019 0.087 0.257 0.234 24-Jul day 0.034 0.062 0.090 0.059 24-Jul night 0.114 0.031 0.038 0.065 25-Jul morning 0.132 0.614 0.274 0.174 25-Jul day 0.047 0.028 0.075 0.041 25-Jul night 0.089 0.027 0.038 0.041 26-Jul morning 0.183 0.081 0.226 0.065 26-Jul day 0.043 0.046 0.108 0.068 Average 0.106 0.091 0.113 0.097

4-46

Figure 4-27. Nitrous acid concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

Concentrations of HNO3 were low and highly variable between the monitoring sites and periods. The high value of 2.39 µg/m3 determined at LH on July 25 could be related to the local fire observed in this area (similarly to the high HONO value determined at HV at that period) (Table 6, Figure 4-28a). The highest HNO3 concentrations were generally observed in the morning and daytime periods, and the lowest at night (Figure 4-28b).

Table 6: Nitric acid concentrations determined at the TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH), and Upper Homewood (UH) during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods.

Nitric acid (HNO3) as µg N/m3 Date on Day part TRPA HV LH UH 21-Jul morning 0.33 0.84 0.86 21-Jul day 0.31 0.46 0.65 0.60 21-Jul night 0.11 0.11 0.16 0.47 22-Jul morning 0.21 0.37 0.34 0.67 22-Jul day 0.29 0.55 0.39 0.81 22-Jul night 0.20 0.00 0.57 0.32 23-Jul morning 0.22 0.19 1.33 0.67 23-Jul day 0.15 0.21 0.27 0.43 24-Jul morning 0.02 0.77 0.60 0.75 24-Jul day 0.19 0.82 0.50 0.40 24-Jul night 0.06 0.12 0.34 0.43 25-Jul morning 0.59 0.25 2.39 1.01 25-Jul day 0.09 0.38 0.25 0.31

4-47

25-Jul night 0.04 0.01 0.10 0.20 26-Jul morning 0.14 0.08 0.25 0.22 26-Jul day 0.24 0.13 0.58 0.34 Average 0.20 0.33 0.58 0.53

Figure 4-28. Nitric acid concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

4-48

Concentrations of particulate ammonium (NH4+) were scattered in time and space with most of the

measured values below ~5 µg N-NH4+/m3 (Table 7 and Figure 4-29 a). However, on several occasions the

determined values were much higher, and in the morning of July 22 exceeded 18 µg N-NH4+/m3 at the

TRPA site. This high reading, as well as the high value of 17.2 at HV during the daytime period of July 21, could be related to the Gardnerville Fire. No clear effects of the period of collection on concentrations of particulate NH4

+ were seen (Figure 4-29 b).

Table 7. Ammonium ion concentrations determined at the TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH), and Upper Homewood (UH) during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods.

Particulate ammonium (NH4+) as µg N/m3

Date on Day part TRPA HV LH UH 21-Jul morning 1.412 na 11.177 5.821 21-Jul day 0.817 17.201 2.715 8.198 21-Jul night 1.056 2.155 4.757 2.001 22-Jul morning 18.473 5.383 3.796 6.820 22-Jul day 4.000 na 4.194 2.414 22-Jul night 2.795 na 2.279 1.530 23-Jul morning 10.220 2.541 4.322 4.479 23-Jul day 4.381 2.406 2.407 1.849 24-Jul morning 1.647 2.934 9.951 3.079 24-Jul day 3.918 1.497 3.185 1.619 24-Jul night 2.924 0.706 16.124 0.705 25-Jul morning 5.158 4.287 2.878 2.135 25-Jul day na 1.593 1.184 0.226 25-Jul night na 0.806 2.319 0.545 26-Jul morning 8.810 3.707 3.331 4.653 26-Jul day 3.453 1.474 1.543 0.410 Average 4.933 3.592 4.760 2.905

na= missing data

4-49

Figure 4-29. Particulate ammonium concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

Concentrations of fine particulate nitrate (NO3-) measured at all four sites (Table 8 and Figure 4-30 a)

showed similar spatial and temporal distribution. However, the nitrate concentrations were slightly lower than those determined for ammonium. This is reasonable considering that most of inorganic nitrogen contained in fine particles is in a form of ammonium nitrate (NH4NO3) but sulfate and organic anions could also be bound to it. No clear effects of the period of collection on concentrations of particulate NO3- were determined (Figure 4-30b).

4-50

Table 8. Particulate nitrate concentrations determined at the TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH), and Upper Homewood (UH) during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods.

Particulate nitrate (NO3-) as µg N/m3

Date on Day part TRPA HV LH UH

7/21/2012 morning 0.068 na 1.095 1.568 7/21/2012 day 0.098 15.075 1.363 6.290 7/21/2012 night 0.032 0.726 3.533 0.816 7/22/2012 morning 14.685 1.088 0.617 3.076 7/22/2012 day 1.936 0.763 2.860 0.172 7/22/2012 night 1.529 0.013 1.419 0.247 7/23/2012 morning 10.051 0.392 1.891 0.525 7/23/2012 day 3.087 0.415 1.346 0.245 7/24/2012 morning 1.447 0.385 5.271 0.144 7/24/2012 day 2.716 0.426 1.674 0.170 7/24/2012 night 2.129 0.061 14.602 0.119 7/25/2012 morning 3.841 1.292 1.468 0.246 7/25/2012 day 4.751 0.097 1.340 0.069 7/25/2012 night 1.429 0.054 2.310 0.069 7/26/2012 morning 8.275 0.168 2.635 0.213 7/26/2012 day 3.397 0.092 1.371 0.111 Average 3.717 1.403 2.800 0.880

4-51

Figure 4-30 Particulate nitrate concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

All measured nitrogenous pollutants and their apportionment. Sum of all gaseous and fine particulate nitrogenous compounds was determined for the entire period of the study based on NO and NO2 values from passive samplers and NH3, HONO, HNO3, particulate NH4

+ and particulate NO3- from honeycomb

denuder/ filter pack systems. The highest amount of all measured nitrogenous species was determined at TRPA (20.82 µg N/m3). Total N was similar at HV (15.84 µg N/m3) and LH (16.38 µg N/m3), while at UH the lowest value was determined (12.82 µg N/m3) (Figure 4-31).

Figure 4-31. Concentrations of all measured nitrogenous pollutants at four study sites.

4-52

Two low elevation sites (TRPA and LH), and two high elevation sites (HV and UH) had significantly different apportionment of the N species. At the TRPA and LH sites NO (29.6 and 23.6%, respectively) and NH4

+ (23.7 and 29.1%, respectively) were dominating, followed by NH3 (13.7 and 17.8%, respectively). At the HV and UH sites NH3 (38.3 and 34.7%, respectively) had the highest participation in total N. Participation of NO at those sites was also high, especially at UH (25.7%). Presence of NO2 in all sites (5.1 – 13.6%) and NO3

- (6.9 – 17.9%) were also high, while those of HNO3 (1.0 – 4.1%, and HONO (0.5 – 0.8%) were much lower (Figure 4-32).

Figure 4-32. Apportionment of nitrogenous air pollutants (%) at individual sites: a. TRPA; b. HV; c. LH; and d. UH.

Sulfur dioxide and particulate sulfate. Sulfur dioxide concentrations were very low (<1 µg/m3) (Table 9, Figure 4-33 a) without any clear temporal patterns (Figure 4-33 b). Highest SO2 concentrations were at the UH which may indicate an influence of the long-range transport of SO2 from the California Central Valley area to the highest elevation west-located site.

Table 9. Sulfur dioxide (SO2) concentrations at four sampling sites.

Sulfur dioxide (SO2) as µg/m3 Date On Day Part TRPA HV LH UH 7/21 morning 0.263 0.493 0.349 0.483 7/21 day 0.141 0.214 0.304 0.297 7/21 night 0.309 0.144 0.226 0.566 7/22 morning 0.264 0.388 0.103 0.157 7/22 day 0.218 0.291 0.236 0.487 7/22 night 0.275 0.081 0.305 0.509 7/23 morning 0.172 0.318 0.417 0.356

4-53

7/23 day 0.093 0.058 0.154 0.225 7/24 morning 0.015 0.440 0.345 0.388 7/24 day 0.407 0.348 0.777 0.883 7/24 night 0.346 0.148 0.305 0.489 7/25 morning 0.428 0.808 0.592 0.374 7/25 day 0.222 0.344 0.174 0.496 7/25 night 0.269 0.241 0.219 0.637 7/26 morning 0.204 0.233 0.000 0.937 7/26 day 0.344 0.398 0.464 0.699 Average 0.248 0.309 0.310 0.499

Figure 4-33. SO2 concentrations determined at TRPA, HV, LH and UH during July 21 – 26, 2012. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

4-54

Particulate sulfate concentrations were generally low, which is typical for the western United States (Table 10). However, during the July 22/23 Garnerville Fire, sulfate concentrations at all sites were elevated, especially at the Upper Homewood site (Table 10 and Figure 4-34 a). No clear effects of period of a day on particulate sulfate concentrations were seen (Figure 4-34 b).

Table 7. Particulate sulfate concentrations determined at the TRPA roof (TRPA), Heavenly Sky Deck (HV), Lower Homewood (LH), and Upper Homewood (UH) during July 21 – 26, 2012. Samples were collected during the morning (from 0600 to 0930 PDT), day (from 1000 to 1730 PDT) and night (1800 to 0530 PDT) periods.

Particulate sulfate (SO42-) as µg/m3

Date on Day part TRPA HV LH UH

7/21/2012 morning 0.000 0.620 0.646 0.966 7/21/2012 day 0.081 0.540 0.851 1.285 7/21/2012 night 0.716 0.683 0.894 1.985 7/22/2012 morning 0.653 0.516 0.693 1.299 7/22/2012 day 1.100 0.471 1.158 2.287 7/22/2012 night 1.439 0.000 1.242 2.090 7/23/2012 morning 0.785 0.332 0.346 1.545 7/23/2012 day 0.764 0.322 0.272 0.601 7/24/2012 morning 0.091 0.000 0.216 0.709 7/24/2012 day 0.641 0.784 0.738 0.083 7/24/2012 night 0.865 0.071 0.592 0.294 7/25/2012 morning 0.335 0.886 0.123 0.149 7/25/2012 day 0.337 0.157 0.026 0.240 7/25/2012 night 0.687 0.452 0.163 1.143 7/26/2012 morning 0.513 0.000 0.082 0.342 7/26/2012 day 0.505 0.096 0.161 0.322 Average 0.574 0.371 0.513 0.959

4-55

Figure 4-34. . Particulate sulfate concentrations determined TRPA, HV, LH and UH during July 21 – 26, 2012. Results are presented: (a) in consecutive chronological order; and (b): sorted by period of a day according throughout the experiment.

4.6 Hysplit trajectories The Hybrid Single Particle Lagrangian Integrated Trajectory Model was run for the field study for the

days of July 21st – July 26th. The backward trajectories were run for 24 hr time periods and the ending time of day for each run was chosen for 5 pm PDT which corresponds to 00 UTC the day after for the HYSPLIT model (i.e. 0000 UTC 22 Jul corresponds to 1700 PDT 21 Jul) . Trajectories were obtained from the Air Resources Laboratory website: http://ready.arl.noaa.gov/HYSPLIT.php. The red line represents 100-m above ground level (agl) while the blue line represents 1500-m agl and the green line

4-56

represents 2500-m agl (Figure 4-35 and 4-36). The red line is within the boundary layer, the blue line is the top of the boundary layer and the green line is above the boundary layer.

The two days before the cold frontal passage look very similar for all sites. Within the boundary layer, the air mass is not moving very far or very fast and primarily comes from the direct west, while the air masses at the top of the boundary layer and above are coming from a more southerly direction. During the cold frontal passage on July 23rd, the air mass within the boundary layer doesn’t move at all, it mostly stays over the Lake Tahoe Basin. The air mass at the top of the boundary layer takes on a more easterly component while the air mass above the boundary layer is mostly from the south. After the cold frontal passage, the air mass within the boundary layer starts moving again and takes on a westerly component. The air masses at top and above the boundary layer come primarily from the south with a southwesterly component to them. On July 26th the air mass within the boundary layer keeps the westerly flow while the air masses at the top and above the boundary layer have a primarily southerly flow with an easterly component to them.

Figure 4-35. HYSPLIT trajectories for Homewood sites

4-57

Figure 4-36. HYSPLIT trajectories for TRPA and Heavenly sites

4.7 Modeling results

4.7.1 CE-CERT The WRF model output was evaluated by comparison to the observations at the Lake Tahoe Airport

(KTVL station). The difference between modeled and observed temperatures varied between -10 K on the night of July 19, 2012 (spin up period) and +10 K on the night July 25 (see Figure 4-37). Most of the modeling period was within +/-5 K when compared to observations. WRF model output for humidity was also compared to observations at the Lake Tahoe airport. Predicted and observed humidity were in poor agreement during the three spin up days, and in good agreement during the time of the field campaign (Figure 4-38).

4-58

Figure 4-37. Comparison of WRF temperature output to observations at Lake Tahoe airport.

Figure 4-38. Comparison of WRF humidity output to observations at Lake Tahoe airport.

Hourly CAMx output is generated for ozone, particulate ammonia (PNH4) and particulate sulfate, as well as SOA produced by emitted hydrocarbons. CAMx output for secondary organic aerosol (SOA) has the following format: SOA1 and SOA2 are generated by aromatic compounds, SOA3 and SOA4 by isoprene, SOA5 and SOA6 from terpenes, and SOA7 from sesquiterpenes. Although SMOKE predicts sesquiterpene emissions for input into CAMx, these species were not measured in the field campaign, so this portion of the model output cannot be compared to observations. The current version of CAMx (version 6) does not generate gas phase products from sesquiterpene oxidation, only SOA.

Secondary organic aerosol species exist in equilibrium with condensable gasses (CG) that can be produced by VOC oxidation:

VOC + oxidant → CG ↔ SOA

CG formation from VOC oxidation reactions is handled within the SOA module rather than the main gas-phase chemistry in CAMx.

There are two areas in SOA formation where CAMx does not reflect real-world scenarios with respect to mobile emissions. First is that it does not aggregate benzene into aromatic species, although it’s known that SOA is formed from benzene oxidation (Ng et al. 2007). Second, CAMx does not contain a model species for long-chain alkanes (emitted mainly from diesel exhaust) which are also known to form SOA upon oxidation.

4-59

CAMx predicts significant SOA formation due to aromatic oxidation (SOA1 and SOA2), with sesquiterpenes also playing a significant role (SOA7) (Figure 4-39). Overall, CAMx predicts anthropogenic and biogenic sources to be equally important to SOA formation.

Figure 4-39. SOA formation by source category. SOA1 and SOA2 are generated by aromatic compounds, SOA3 and SOA4 by isoprene, SOA5 and SOA6 from terpenes, and SOA7 from sesquiterpenes

CAMx output for the afternoon of July 21, 2012 was compared to observations at the TRPA site. While the model reproduced the ozone levels to within 10%, the predicted apportionment of the N species did not match observations (see Figure 4-40). The model predicted almost half of the nitrogen to be in the form of NO2, followed by NH3 and NO. We plan to revisit NO2 emissions predicted by SMOKE for input into the model.

Figure 4-40. Nitrogen apportionment predicted by CAMx.

Future plans also include re-running MOVES due to some missing emissions data for high and low temperatures which were input manually after the run was performed and before merging with the other

4-60

emission sources. We would also like to process biogenic emissions with MEGAN and compare the predictions to the BEIS3 program.

4.7.2 DRI Modeling effort at DRI has not produced conclusive results so far. We plan to continue the modeling

work, even after submitting this final report, so we would be able to publish reliable results in a peer-review literature in a future. The following is a summary of the results obtained so far.

The DRI modeling effort has focused on getting the meteorology component more accurate. The model had issues with the terrain height, which have now been resolved. The WRF Model temperature at Sacramento and Reno is very close to the station temperature (Figure 4- 41). Although the model doesn’t hit every point exactly, it is important to note that it is getting the overall larger scale meteorology with peaks and troughs occurring in the same locations. The other important thing to note is the planetary boundary layer height. At Reno, the model has a good handle on the height during the daytime hours.

Figure 4-41. Comparison of measured and modeled temperatures in Sacramento (a) and Reno (b)

After obtaining better meteorology, CMAQ was run again. The ozone was off again, however it was hitting the diurnal variability more correctly. It was not hitting the height of the peaks during the daytime and the lows of the troughs at night. Even in Sacramento we had this problem, Figure 4- 42 shows the ozone from the T- Street station in Sacramento. All of our sites look relatively the same as Sacramento.

a b

4-61

Figure 4-42. Comparison of measured and modeled ozone concentrations at T-Street in Sacramento.

CMAQ output for NOx is not correct either. The CMAQ NO2 at our sites is basically zero for all four sites where the NO has some minor improvement but is still inaccurate (Figure 4- 43).

a b

c d

4-62

Figure 4-43. Comparison of measured and modeled NO concentrations for Upper Homewood (a), Lower Homewood (b), TRPA (c) and Heavenly (d)

We believe the emissions inventory is incorrect and will be running the CMAQ model with increased and decreased amounts of NO and NO2 in order to see if that fixes the ozone issue. We are also looking into the emissions inventory itself to determine why it may be incorrect.

5. Discussion The following hypotheses were tested during this study:

1. The majority of precursors for O3 formation come from in-basin sources.

The results of this and the previous (Bytnerowicz et al., 2013, Burley et al., 2015) study do not support this hypothesis. On the contrary, our data indicate that the majority of O3 precursors (VOC and NOx) are emitted in the urbanized areas of the Central Valley and possibly the San Francisco Bay Area.

The concentrations of ozone precursors, VOC and NOx at all four sites are very low (Section 4.2). Therefore, substantial in-situ photochemical production of ozone is unlikely at most sites. The only site which showed slightly higher NOx and VOC concentrations is TRPA, which is located in South Lake Tahoe near the Kingsbury Grade Highway. South Lake Tahoe is the biggest city in the Tahoe Basin and experiences higher local traffic. It is a major tourist attraction in the California – Nevada area and is home to a number of summer outdoor recreations. The Nevada side also includes large casinos. Thus photochemical production of ozone may be non-negligible in this area.

The Tahoe sites display great site-to-site variability in O3 concentrations during the evening and pre-dawn hours of 1900 to 0700 PST, but consistent maxima of 50 to 60 ppb from 1000 to 1700 PST (Figures 4-8 and 4-10). Upper Homewood (UH) differs from other Tahoe sites because of its very small diurnal cycle magnitudes and the observation of increasing ozone between the hours of 1700 to 2200 PST (Figure 4-8). This latter result – an increase in O3 despite the shutdown of photochemical production pathways - is consistent with the observations from the previous Lake Tahoe study in 2010 (Bytnerowicz et al., 2013; Burley et al., 2015) that measured similar increase in late evening ozone concentrations in high altitude sites (Genoa Peak at 9000’ and Angora Lookout at 7277’).

The observation that all of the Tahoe sites tend to rapidly reach the same middle day maximum between the hours of 0600 and 1000 PST suggests that this rise in surface ozone is due primarily to the vigorous vertical mixing that occurs in the hours immediately after sunrise, when ozone-depleted air near the surface is warmed by solar heating of the surface, while cooler, ozone-rich air from aloft is simultaneously mixed downwards. These well-mixed conditions persist throughout the Tahoe Basin until approximately 1700 PST, when there is marked decrease in solar insolation. During the evening hours, those sites that have good exposure to ozone-rich air from aloft, which is being transported in via regional-scale and long-range transport, continue to exhibit high ozone concentrations. Sites that are more conducive to the formation of nocturnal temperature inversions see a decrease in surface level ozone, primarily via dry deposition (Burley et al., 2015).

Average ozone concentrations measured at most of the Tahoe sites are roughly similar to those recorded at the comparison sites Folsom, Cool and Placerville – in the range of 40-60 ppb – and higher than those recorded in downtown Sacramento (20-40 ppb, see Figure 4-9b), a heavily polluted source

5-63

region. The afternoon maxima at Sacramento, Folsom, Cool and Placerville are different from those observed in Tahoe sites; rather than the flat-topped maxima from 1000 to 1700 PST observed at Tahoe, the afternoon maxima in the comparison sites do not arrive there until approximately 1400-1600 and they last for only 1-3 hr (see Figure 4-10). This indicates that the direct transport of intact ozone-rich air masses from upwind air basins to Lake Tahoe is infrequent; rather emissions from upwind regions are raising background concentrations of pollutants that are subsequently transported into Tahoe Basin (Dolislager et al., 2012).

The 24-hr HYSPLIT back-trajectories indicate that predominant air flow within boundary layer comes from direct west, while the air masses at the top of the boundary layer and above are coming from a more southerly direction for all four sites. During the cold frontal passage on July 23rd, which resulted in lower O3 concentrations, the air mass at the top of the boundary layer takes on a more easterly component while the air mass above the boundary layer is mostly from the south.

These observations likely indicate that much of the ozone measured at Lake Tahoe sites results from the transport of polluted background air into the Basin from upwind pollution source regions (e.g., Sacramento, San Joaquin Valley and possibly San Francisco Bay Area). This conclusion is consistent with Dolislager et al. (2012) study, who assessed the relative impacts of transport versus local photochemical production by making continuous measurements during the summer of 2003 along the axis of predominant airflow (i.e., roughly southwest to northeast). They utilized two transport assessment sites at Big Hill and Echo Summit, along with other monitoring sites at various locations on the western slope of the Sierra Nevada, plus four in-Basin monitoring sites and incorporated aircraft data. They concluded that pollutants from upwind regions act to raise background concentrations entering the Tahoe Basin to the extent that local contributions do not need to be large to cause exceedances of air quality standards.

Although the CAMx model predicted ozone levels at the TRPA site for the afternoon of July 20, 2012 to within 10%, the predicted ozone profile did not match the observation (Figure 5-1). This indicates that the model did not take into account the regional transport of polluted air masses into the Basin.

Figure 5-1. Comparing observed (blue diamonds) and modeled ozone levels with all emissions (red squares), no anthropogenic emissions (green diamonds), and no biogenic emissions (purple x) for the TRPA site on July 20, 2012.

5-64

2. The majority of precursors for SOA formation come from in-basin sources.

The results of this study do not support this hypothesis either. The concentrations of measured precursors of SOA (biogenic hydrocarbons, isoprene and α-pinene, and anthropogenic aromatic hydrocarbons) are low and they do not explain the concentrations of SOA tracers, two 2-methyltetrols and cis-pinionic acid. Isoprene oxidation products include 2-methylerythritol and 2-methylthreitol, whereas cis-pinonic acid is formed from the reaction of α-pinene with ozone. 2-methyltetrols constituted 0.3% - 3% and cis-pinonic acid 0.1% - 0.7% of total organic mass (OM, assuming OM/OC=2). Methacrolein, an atmospheric photooxidation product of isoprene, was observed in gas phase samples.

However, isoprene concentrations measured at four sites are lower than α-pinene concentrations and they cannot account for in-situ production of 2-methyltetrols. Also, the higher concentrations of both 2-methyltetrols during the night time than the day time (opposite to cis-pinionic acid, see Figure 4-22) argue against their in-situ formation in the Basin. This indicates that the isoprene photochemical transformation products, methyltetrols and methacrolein are most likely transported from out-of-basin sources or formed during transport from the Central Valley area, which is the source of significant isoprene emissions.

Toluene/benzene and xylenes/benzene ratios provide information regarding the photochemical age of the emissions (de Gouw et al., 2005). For aromatic VOC, the day time reaction with OH radicals is the dominant loss mechanism during transport through the atmosphere. The rate coefficients for the reaction of toluene and benzene with OH radicals are equal to 5.63 x 10-12 cm3 molecule-1 s-1 and 1.22 x 10-12 cm3 molecule-1 s-1, respectively (Atkinson and Arey, 2003), thus toluene reacts approximately 5 times faster with OH radicals than benzene. Xylene isomers react even faster than toluene. The ratio of toluene/benzene in fresh emissions was estimated to be 3.7 ± 0.3 (De Gouw et al., 2005). The lower the ratio is, the more aged the air plume is. Figure 5-2 shows these ratios, which are the lowest in the western high site (UH), indicating aged air masses. Low eastern site (TRPA) shows high morning ratios, which indicates fresh morning rush hour emissions.

Figure 5-2. Toluene/benzene and m,p-xylene/benzene ratios for all sites and measuring periods. LH = Lower Homewood , UH = Upper Homewood , HE = Heavenly , and TRPA = TRPA rooftop

5-65

Model simulations indicate the majority of SOA precursors are emitted upwind of the Lake Tahoe Basin. Ozone is transported from aloft upwind as well, influencing ozone levels in-basin significantly.

3. Out-of-basin contribution to observed O3 and SOA levels is limited.

In contrary, in-basin contribution to observed O3 and SOA level is limited. See the discussion of hypotheses #1 and #2 above.

4. To control O3 formation, one needs to develop policies to limit NOx emissions.

In-basin NOx concentrations are very low. As noted above, the only exception is TRPA which showed higher NOx concentrations, although concentrations at this site are still low when compared to sites outside the basin with similar ozone levels.

5. Reduction of O3 levels will reduce the amount of SOA.

Reduction in ozone levels will reduce the amount of SOA by a modest amount. Ozone reaction rates with SOA precursors are slow compared with other oxidants, such as hydroxyl radical.

In addition, as discussed above, since the majority of SOA precursors originate from out-of-basin sources, reduction of O3 level in the Tahoe Basin area would not influence significantly the amount of SOA in the Basin.

6. Acknowledgements: This study was supported by the grant from the U.S. Department of Agriculture Forest Service through the Tahoe Science Program funded by the Southern Nevada Public Lands Management Act (SNPLMA). The authors of the report would like to thank: Tiffany Van Huysen of the USDA Forest Service PSW Research Station for administrative support for the study; Heavenly and Homewood Ski Resorts for providing access to their areas; Joey Keely of the USDA Forest Service Lake Tahoe Basin Management Unit for logistical help; and Chad Praul of Environmental Incentives for his help in facilitating meetings with program managers and other stakeholders.

7. Literature cited Atkinson, R., and Arey, J. 2003. Atmospheric degradation of volatile organic compounds, Chem.

Rev., 103, 4605– 4638.

Burley, J.D., Theiss S., Bytnerowicz A., Gertler A., Schilling S, and Zielinska B. (2015). Surface ozone in the Lake Tahoe Basin, submitted to Atmospheric Environment.

Bytnerowicz, A., Fenn, M., Gertler, A., Preisler, H., Zielinska, B., Ahuja, S., Cisneros, R., Shaw, G., Schweizer, D., Procter, T., Burley, J., Schilling, S., Nanus, L., McDaniel, M., Theiss, S., Sickman, J., Bell, M., Rorig, M., 2013. Distribution of ozone, ozone precursors and gaseous components of atmospheric nitrogen deposition in the Lake Tahoe Basin. Final Technical Report for SNPLMA Contract P063, USDA Forest Service, PSW Research Station, 86 pp.

7-66

Bytnerowicz A., Arbaugh M., Schilling S., Fraczek W., Alexander D., Dawson P. 2007.Air pollution distribution patterns in the San Bernardino Mountains of southern California: a 40-year perspective. The Scientific World Journal 7(S1), 98–109. DOI 10.1100/tsw.2007.57.

Bytnerowicz, A., Sanz, M. J. Arbaugh, M. J. Padgett, P. E. Jones, D. P., and Davila, A. 2005. Passive sampler for monitoring ambient nitric acid (HNO3) and nitrous acid (HNO2) concentrations. Atmospheric Environment, 39, 2655-2660.

Calvert, J.G., Stockwell, W.R., 1983. Acid generation in the troposphere by gas-phase chemistry. Environ. Sci. Technol. 17, 428A-443A.

Chow, J.C.; Watson, J.G.; Chen, L.W.A.; Paredes-Miranda, G.; Chang, M.C.; Trimble, D.; Fung, K.K.; Zhang, J.; Yu, J.Z. Refining Temperature Measures in Thermal/Optical Carbon Analysis. Atmos. Chem. Physics Discussion 2005, 5, 4477-4505.

de Gouw, J. A.; Middlebrook, A. M.; Warneke, C.; Goldan, P. D.; Kuster, W. C.; Roberts, J. M.; Fehsenfeld, F. C.; Worsnop, D. R.; Canagaratna, M. R.; Pszenny, A. A. P.; Keene, W. C.; Marchewka, M.; Bertman, S. B.; Bates, T. S. (2005). Budget of organic carbon in a polluted atmosphere: Results from the New England air quality study in 2002, J. Geophys. Res. 110, D16305

Dolislager, L. J., VanCuren, R., Pederson, J. R., Lashgari, A., McCauley, E., 2012. Anassessment of ozone concentrations within and near the Lake Tahoe Air Basin. Atmospheric Environment 46, 645-654.

Engelbrecht J., A Gertler and T. VanCuren (2009). Lake Tahoe Source Attribution Study (LTSAS): Receptor Modeling Study to Determine the Sources of Observed Ambient Particulate Matter in the Lake Tahoe Basin, Final Report to USDA Forest Service Pacific Southwest Research Station, September 2009

Fraser, M. and G.R. Cass (1998). Detection of excess ammonia emissions from in-use vehicles and the implications for fine particle control. Environ. Sci. Technol. 32, 1053-1057.

Fuzzi S., M.O. Andreae, B.J. Hubert, M. Kulmala, T.C. Bond, M. Boy, S.J. Doherty, A. Gunther, M. Kanakidou, K. Kawamura, V-M Kerminen, U. Lohmann, L.M. Russel and U. Pooschi (2005). Critical assessment of the current state of scientific knowledge, terminology, and research needs concerning the role of organic aerosols in the atmosphere, climate, and global changes. Atmos. Chemistry and Physics Discussions, 5, 11729-11780

Jenkin, M.E., (2004). Modeling the formation and composition of secondary organic aerosol from α- and β - pinene ozonolysis using MCMv3. Atmos. Chem. Phys. 4, 1741-1757

Johnson, D.; S.R. Utembe, M.E. Jenkin, R.G. Derwent, G.D.Hayman, M.R. Alfarra, H.Coe, and G. McFiggans (2006). Simulating regional scale secondary organic aerosol formation during the TORCH 2003 campaign in the Southern UK., Atmos.Chem Phys., 6, 402-418

Jordan, C. E., et al. (2008). Modeling SOA formation from OH reactions with C8–C17 n-alkanes. Atmospheric Environment 42.34, 8015-8026.

7-67

Koutrakis, P., C. Sioutas, S. T. Ferguson, J. M. Wolfson, J. D. Mulik, R. M. Burton (1993). Development and evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles. Environ. Sci. Technol., 27, 2497-2501.

Mazzoleni, L. R., B. Zielinska, and H. Moosmüller (2007). Emissions of Levoglucosan, Methoxy Phenols, and Organic Acids from Prescribed Burns, Laboratory Combustion of Wildland Fuels, and Residential Wood Combustion. Environ. Sci. Technol., 41 (7), 2115-2122

Ng, N. L., et al. (2007). Secondary organic aerosol formation from m-xylene, toluene, and benzene. Atmospheric Chemistry and Physics 7.14, 3909-3922.

Stockwell, W.R., Kirchner, F., Kuhn, M., Seefeld, S., 1997. A new mechanism for regional atmospheric chemistry modeling. Journal of Geophysical Research 102, 25847-25879.

Volkamer, R., J.L. Jimenez, F.S. Martini, K. Dzepina, Q. Zhang, D. Salcedo, L.T. Molina, D.R. Worsnop, M.J. Molina (2006). Secondary organic aerosol formation from anthropogenic air pollution: rapid and higher than expected. Geophys. Res. Letters, 33, L17811, doi: 10.1029/2006GL026899

Zielinska, B., Fujita, E., Ollison, W., Campbell, D., and Sagebiel. J. (2012). Quantification of Personal Exposure Concentrations to Gasoline Vehicle Emissions in High-End Exposure Microenvironments: Effects of Fuel and Season, J. Air Waste Manage. Assoc., 62, 1346-1357.

7-68