Weekend/Weekday Ozone Observations in the South Coast Air Basin Sponsored by National Renewable Energy Laboratory and Coordinating Research Council Eric Fujita, Robert Keislar, and William Stockwell Desert Research Institute University and Community College System of Nevada Reno, Nevada Paul Roberts, Hilary Main, and Lyle Chinkin Sonoma Technology, Inc. Petaluma, CA Weekend/Weekday Ozone Effect Workshop Sacramento, CA November 16, 1999

Overview Conceptual explanation of the weekend/weekday ozone effect. Preliminary hypotheses NREL weekend/weekday study Other related studies Desert Research Institute, 11/16/99

What do we know about the weekend/weekday ozone effect in the South Coast Air Basin?  During , ozone episodes occurred significantly more often on Saturdays than on Sundays through Wednesdays (Blier and Winer, 1996).  During , large increases in ozone from Friday to Saturday (~30%) in many sites in central SoCAB, no change or slight decrease from Saturday to Sunday (Austin and Tran, 1999).  Many sites show a “Sunday effect” in the period (Austin and Tran, 1999).  Weekend effect is least pronounced at transport sites further downwind (e.g., Lake Gregory, Banning, Hemet, Perris, and Santa Clarita). Coastal sites (Hawthorne and West Los Angeles) also exhibit a mild weekend effect (Austin and Tran, 1999).  Decreases in peak ozone levels from the mid-1980 to mid-1990 were greatest in western and central portions of the SoCAB. Greater reductions on weekdays than on weekends and hence the differences in WD vs. WE ozone maxima are larger now than the 1980s (Blier and Winer, 1996).  Similar WE/WD effect in San Francisco Bay Area and cities in northeastern U.S., no effect in Sacramento, reserve effect in Atlanta. Desert Research Institute, 11/16/99

What do we know about the weekend/weekday differences in VOC, NOx and PM in the South Coast Air Basin?  VOC, NOx and PM are all higher during weekdays.  During , average early morning NO 2 and NOx were lower by 20-25% and %, respectively on weekend days in the Coastal/Metropolitan region of the SoCAB. (Blier and Winer, 1996)  Morning NOx is highest on weekdays, followed by Saturday and lowest on Sunday.  Saturday afternoon levels are comparable to or slightly lower than weekday levels.  Saturday evening levels tend to be lower than on Friday and roughly equal to or higher than the mean weekday evening levels.  NOx mixing ratios are lower on Sunday than other days for all hours except at midnight to 4 a.m. when they are comparable to weekdays.  The reactivity of the ambient hydrocarbon mixture has dropped between 1995 and Reactivity appears slightly lower on weekends (Franzwa and Pasek, 1999).  6 to 9 a.m.VOC/NOx ratios have decreased from 8 to 10 in 1987 (SCAQS) to 4 to 7 in Desert Research Institute, 11/16/99

Historic Ozone Air Quality Trends South Coast Air Basin ( ) ) Desert Research Institute, 11/16/99

Historic Ozone Air Quality Trends South Coast Air Basin ( ) - WESTERN Desert Research Institute, 11/16/99

Historic Ozone Air Quality Trends South Coast Air Basin ( ) - CENTRAL Desert Research Institute, 11/16/99

Historic Ozone Air Quality Trends South Coast Air Basin ( ) - EASTERN Desert Research Institute, 11/16/99

Historic Ozone Air Quality Trends South Coast Air Basin ( ) Desert Research Institute, 11/16/99

Ozone Formation Chemistry NO 2 h HCHO, RCHO HO 2 RO 2 CO, VOC H 2 O 2 ROOH O 3 HNO 3 NO 2 PAN from RO 2 h O 3 from HO 2

Desert Research Institute, 11/16/99

Factors Affecting the Magnitude and Spatial Extent of the WE/WD Ozone Effect Ozone formation depends on VOC, NOx and VOC/NOx ratios (O 3 Potential). –For VOC/NOx < 5.5, OH reacts more with NO 2, removing radicals and NOx to retard O 3 formation. Under these conditions, a decrease in NOx favors O 3 formation. –At low NOx mixing ratios, or sufficiently high VOC/NOx, decrease in NOx favors peroxy-peroxy reactions, which retard O 3 formation by removing free radicals from the system. –At a given level of VOC, there exists an optimum VOC/NOx ratio at which a maximum amount of ozone is produced. For ratios less than this optimum ratio, increasing NOx decreases ozone. This situation occurs more commonly in urban centers and is the case for most of the central SoCAB. WE/WD differences in VOC and NOx emissions patterns (Diurnal and Spatial Distribution). Transport and ventilation (Meteorological Effects). –Sea breeze limits ozone accumulation in the western portion of SoCAB. –Increasing mixing height due to surface heating reduces [VOC] and [NOx]. –Horizontal transport increases VOC/NOx ratios due to more rapid removal of NOx than VOC. The observed "weekend effect" in the South Coast Air Basin arises from differences in ozone forming potential due to day-of-the-week changes in ROG and NOx emissions. Variations in meteorology affect the magnitude and spatial extent of the WE/WD ozone effect within the basin. Desert Research Institute, 11/16/99

Preliminary Hypotheses 1. Ozone formation in SoCAB, particularly the western and central portions of the basin, is VOC-sensitive with respect to ozone formation. VOC/NOx ratios are higher on weekends due to WE/WD changes in emissions resulting in greater ozone forming potential despite lower [VOC] and [NOx] on weekends. –The weekend effect is greater where  [O 3 ] max /  [VOC] is greater during weekdays than during weekend days. –WE/WD effect is most pronounced in area of the basin with the greatest NOx disbenefit (i.e., most VOC-limited on weekdays). 2.The magnitude of the weekend effect is a function of the ozone forming potential and the time available for ozone formation before dilution offsets ozone formation. 3.The "weekend effect" is less pronounced in the eastern portion of the SoCAB where WE/WD differences in VOC and NOx emissions are masked by emission transport. Transport causes higher VOC/NOx ratios due to more rapid removal of NOx versus VOC as the emissions are transported toward the eastern side of the Basin. 4.Overnight carry-over of ozone, VOC and NOx from Friday and Saturday nights are greater than during other days of the week. Increased carryover is greater for VOC than for NOx. This affects the ozone forming potential of the ambient air. 5.A number of changes in emissions by day-of-week, time-of-day, and by location in the SoCAB can be postulated. Desert Research Institute, 11/16/99

Scope of Work PHASE I: Retrospective Analysis of Ambient and Emissions Data and Refinement of Hypotheses –Task 1: Review available emissions data. (STI) –Task 2: Analyze retrospective ozone and ozone precursors and ozone episodes (DRI) –Task 3: Review source apportionment analyses (DRI) –Task 4: Analyze SCOS97-NARSTO meteorological and 3-D ozone data (STI) –Task 5: Synthesize phase I data analysis and prepare Phase 1 Report (DRI and STI) PHASE II: Summer 2000 Field Measurements Program –Task 6: Conduct field measurements (DRI) –Task 7: Update and improve source composition profiles (DRI) –Task 8: Update and improve temporally and spatially-resolved activity factors (STI) –Task 9: Compile and validate data (STI) PHASE III: Data Analysis and Final Report –Task 10: Analyze temporal and spatial variations in O 3, VOC, NOx and related air quality and meteorological parameters (DRI) –Task 11: Analyze PAMS upper-air meteorological data (STI) –Task 12: Update source apportionment analysis (DRI) –Task 13: Analyze activity data (STI) –Task 14: Update EKMA analysis (DRI) –Task 15: Evaluate SCOS97-NARTSO model sensitivity results (STI) –Task 16: Synthesize results and prepare final report (DRI and STI Desert Research Institute, 11/16/99

PHASE I: Retrospective Analysis of Ambient and Emissions Data and Refinement of Hypotheses Task 1: Review available emissions data. (STI) –Based on available emission inventory data, identify VOC and NOx sources with potential to be different on weekends than on weekdays. –Summarize diurnal variations in daily ROG and NOx emissions by day-of-the-week for these sources. –Review the method(s) used to determine temporal variations and evaluate uncertainties and identify alternative methods or additional data that are available to update and improve existing temporal allocation of ROG and NOx emissions. Task 2: Analyze retrospective ozone and ozone precursors and ozone episodes. (DRI) –Characterize and classify evolution of temporal and spatial patterns of O 3, CO, total NMHC, carbonyl compounds, NOx, and NMHC/NOx ratios from Thursday to Monday during the summers of by meteorological conditions. Task 3: Review source apportionment analyses. (DRI) –Review the source apportionment analysis conducted by the Desert Research Institute for SoCAB PAMS data ( ) for weekend days and weekdays. –Review available source composition profiles and identify source for which updated profiles are needed. Desert Research Institute, 11/16/99

PHASE I: Retrospective Analysis of Ambient and Emissions Data and Refinement of Hypotheses Task 4: Analyze SCOS97-NARSTO meteorological and 3-D ozone data. (STI) –Evaluate meteorological conditions during SCOS97-NARSTO IOPs to determine applicability of weekend IOPs for assessments of the weekend effect. –Characterize the surface and aloft spatial and temporal patterns of ozone and ozone precursors during weekend intensive operational periods. –Analyze the data from the SCOS97 upper-air meteorological network and evaluate the regional representativeness of the temporal and spatial variations in wind and mixing heights that can be obtained from the two PAMS profilers (at LAX and Ontario) alone. Task 5: Synthesize phase I data analysis and prepare Phase 1 Report (DRI and STI) –Summarize results of phase I data analysis, revise conceptual model, update hypotheses, and finalize field measurement program. –Submit draft report for Phase I in April Desert Research Institute, 11/16/99

PHASE II: Summer 2000 Field Measurements Program Task 6: Conduct field measurements (DRI) –Continuous NOy and NOy* (NOy-HNO3 at Pico Rivera, Azusa, and Upland. –Continuous total NMHC by TEI 55 at Azusa and Upland. –Continuous CO by TEI 48C-TL (0.4 ppb) at Pico Rivera, Azusa, and Upland. The District typically reports CO to the nearest ppm. –Continuous light absorption by aethalometer at Pico Rivera, Azusa and Upland. –DRI comparison with speciated NMHC from the SCAQMD auto-GC and TEI 55. Optional –EC and OC by R&P carbon analyzer or automated Thermal Optical Reflectance at Pico Rivera and Upland. –Continuous PM mass by at Pico Rivera, Azusa and Upland. –Supplemental canister and DNPH samples at Upland during weekends (Friday-Sunday) to fill in the PAMS every third day sampling. Collect a total of up to 72 canister and 72 DNPH cartridge samples. –NO2 and PAN –Continuous HCHO Desert Research Institute, 11/16/99

PHASE II: Summer 2000 Field Measurements Program Task 7: Update and improve source composition profiles (DRI) –Collect additional VOC source composition profiles identified in Task 1 and 4. –Conduct saturation monitoring near epicenter of non-mobile VOC source to determine source composition and zone of influence. Task 8: Update and improve temporally and spatially-resolved activity factors (STI) –Gather and compile existing information and new data that will support weekend-weekday comparisons of emissions as determined in the plan developed in Phase I. Task 9: Compile and validate data (STI) –Compile and validate the SCAQMD’s PAMS VOC data; NO x, CO, and ozone data; and upper- air data collected during the ozone seasons of 1999 and Desert Research Institute, 11/16/99

PHASE III: Data Analysis and Final Report Task 10: Analyze temporal and spatial variations in O 3, VOC, NOx and related air quality and meteorological parameters (DRI) Task 11: Analyze PAMS upper-air meteorological data (STI) Task 12: Update source apportionment analysis (DRI) Task 13: Analyze activity data (STI) Task 14: Update EKMA analysis (DRI) Task 15: Evaluate SCOS97-NARTSO model sensitivity results (STI) Task 16: Synthesize results and prepare final report (DRI and STI Desert Research Institute, 11/16/99