1

How is model predicted O3 sensitive to day type emission variability and morning Planetary Boundary Layer rise? Hypothesis 2

Outline Introduction Modeling Datasets Results Conclusions Future work How is model predicted O 3 sensitive to day type emission variability and Planetary Boundary Layer rise? 3

Ozone - A Secondary Criteria Pollutant Health effect Environmental Problems Factors in O3 Production: Emissions NOx VOC Meteorology Wind Profile Rise of Planetary Boundary Layer 4

Houston is Non-attainment Area Ship Channel 5

The O3 Standard National Ambient Air Quality Standard (NAAQS) for O3 8-hour standard: 0.08 ppm State Implantation Plan (SIP) Current SIP (8-hour standard) Attainment Test based on Relative Model predictions 6

How to develop Attainment Test for the SIP EPA Guidance: 8-hr Ozone Attainment Test Monitor by Monitor Test Based on Observations Maximum 8-hr Averages Based on Air Quality Models (AQMs) Base case -- used in simulation performance not in Attainment Test Base line case -- ‘typical’ emission inventory Future case -- ‘controlled’ emission inventory 7

The 8-hr Ozone Attainment Test DV F = RRF x DV B Future design value If it is below the standard the monitor is in compliance of NAAQS Baseline design value based on observations Relative Reduction Factor based on model predictions 8

DV F = RRF x DV B Year st nd rd th th Avg. of DVs DV B Ozone Data for Highest Daily 8-hr max (ppm) 9

DV F = RRF x DV B Average of predicted Future case 8-hr daily max “near” monitor Average of predicted Base line case 8-hr daily max “near” monitor RRF = O 3 Future O 3 Base line = Day 1 F +Day 2 F …. Day N F N RRF days Day 1 B +Day 2 B …. Day N B N RRF days = RRF M = 10

An RRF Example represent a reduction of 11% of DV B Lower RRF values indicate larger relative decreases in future predicted ozone concentrations When calculating an RRF remember : EPA recommends a threshold concentration of 85 ppb for days used in RRF calculations, but allows concentrations as low as 70 ppb. EPA recommends using at least 10 days in calculating RRFs, but allows as few as 5 days RRF M ==

Day Type Emissions How is model predicted O 3 sensitive to day type emission variability? NOx 12

Changing “The Box” changes Emission Concentrations 13

Different Simulated Planetary Boundary Layer Rises How is model predicted O 3 sensitive to Planetary Boundary Layer rise? 14

Outline Introduction Modeling Datasets Results Conclusions Future work How is model predicted O 3 sensitive to day type emission variability and Planetary Boundary Layer rise? 15

Modeling Standard Attainment Dataset The Texas Commission of Environmental Quality (TCEQ) developed Attainment Test for the SIP 8-hr O3 Attainment Test for 25 Surface Monitors Calculating 25 DV B, RRFs, and DV F 4 monitors failed attainment test 18 AQMs Simulation Episodes using CAMx 2005 Base case 120 modeling days from 2005 and 2006 episodes 2005 Base line case 120 modeling days from 2005 and 2006 episodes 2018 Future case 120 modeling Days projected to

Across the Board Emissions Controls Applied with Growth 600 TPD364 TPD 992 TPD1011 TPD Base Line Case 2005Future Case 2018 NOxNOx VOCVOC 17

Outline Introduction Modeling Datasets Results Conclusions Future work How is model predicted O 3 sensitive to day type emission variability and Planetary Boundary Layer rise? 18

Results Weekday Weekend Analysis Results Meteorological Analysis Results Process Analysis Results 19

Results Weekday Weekend Results Meteorological Analysis Results Process Analysis Results Do different type of day emissions affect Houston’s predicted O 3 ? Day Type Emissions Variability 20

Central Houston 1,250 km 2 21

Different Concentrations for Weekdays and Weekends HRVOC = Highly Reactive VOCs, Weekday Weekend 22

Spread of Weekdays to Weekends in the Attainment Dataset

Weekday Weekend O 3 Base line Predicted 8-hr max O 3 24

Weekday Weekend O 3 Future Predicted 8-hr max O 3 25

Weekday Weekend RRF 0.03 – 0.08 <

Weekday Weekend DV F DV B influence > RRF 2-6 ppb < 2 ppb O3O3 27

Weekday Weekend Summary and Conclusions NOx emissions on the weekends constitute a reduction on average of 13% Eastern Houston monitors affected by industrial emissions have higher weekend O3 Western Houston monitors affected by mobile emissions have lower weekend O3 Model response to changes in day type emission is sensitive to location O3 concentrations can vary by as much as 35 ppb and RRFs can vary by up to 0.08 between weekdays and weekends. Arbitrary averaging different day type Emission introduces a margin of error that may vary attainment 28

Results Weekday-Weekend Results Meteorological Analysis Results Process Analysis Results How does the model simulate the PBL ? The Simulated Planetary boundary or Model Mixing Volume (MMV) is produced by the AQM by adjusting the vertical mixing (k v ) between layers in the Eularian grid structure. Simulated Planetary Boundary Layer Height How is O 3 sensitive to MMV ? 29

MMV with focus on Central Houston 1,250 km 2 30

2 Distinct MMV Rises Slow Riser = MMV change less than 700 m/h between 6 to 11 LST Fast Riser = MMV change more than 700 m/h between 6 to 11 LST 31

MMV of Slow Riser & Fast Riser “Slow Riser” August 1, am 1- Hour O ppb “Fast Riser” August 6, Hour O ppb 32

Fast Riser higher than Slow Riser 7 am “Fast Riser” August 6, LST“Slow Riser” August 1, LST 33

Spread of Slow Risers to Fast Risers in the Attainment Dataset

Slow Riser Fast Riser O 3 Base line Predicted 8-hr max O 3 35

Slow Riser Fast Riser O 3 Future Predicted 8-hr max O 3 36

Slow Riser Fast Riser RRF Fast Riser respond better to 2018 controls x the difference of day type

Slow Riser Fast Riser DV F DVf per Monitor, Type of RiserDVf per Monitor, Type of Riser DV B > Riser influence > Type of day 3-6 ppb 2-4 ppb O3O3 38

Meteorological Analysis Summary and Conclusions O3 concentrations can vary by 35 ppb and RRFs as much as 0.07 between Slow Risers and Fast Risers. Model response to changes in MMV Rise is sensitive to location Fast Risers are more responsive to 2018 Controls than slow risers Reponses to Controls varies with MMV Rise Type Rise in MMVs can influence DV F s enough to bring them into attainment or out of attainment 39

Results Emission Inventory Results Meteorological Analysis Results Process Analysis Results Too much NOx = NOx-inhibited Too much VOC = NOx-limited Why is it important how O 3 is produced in the studied phenomena ? 40

Process Analysis: Central Houston 1,250 km 2 41

4 new-modeled days Type of Day MMV Rise WeekdaySlow WeekdayFast WeekendSlow WeekendFast 42

2 Emission Inventories Weekday Weekend 43

2 Meteorological Days Slow Riser Fast Riser 44

Physical Processes : NOx Slow Riser Fast Riser 06/21/05 Weekday 45

Physical Processes : VOC Slow Riser Fast Riser 06/21/05 Weekday 46

Physical Processes : O3 Slow Riser Fast Riser 06/21/05 Weekday 47

A Very Simple Intro to O3 Chemistry OH. + VOC HO 2 RO 2 RO. OH. NONO 2 O2O2 O3O3 O2O2.. HNO 3 NOx-limited H2O2H2O2 NO 2 HO 2 NOx-inhibited 48

Fast Riser NOx-limited Earlier, Longer Weekday OH + NO 2  HNO 3 HO 2 + HO 2  H 2 O 2 + O 2 NOx-limited NOx-inhibited - P(H 2 O 2 ) : P(HNO 3 ) 49

Process Analysis Summary and Conclusions 2 distinct rising MMV patterns show the same behavior in both emission types with different magnitudes. Slow Riser Fast Riser Entrainment of VOCs that bring in new VOCs 5x more Dilution of NOx and VOCs Steeper O3 production rate mainly due to entrainment NOx-limited much earlier in day than Slow Riser Restricts O3 formation to NOx availability Lower & Earlier Peak O3 Same set of EI show distinct O3 producing regimes Affect the type of controls needed to reduce O3 50

Outline Introduction Modeling Datasets Results Conclusions Future work How is model predicted O 3 sensitive to day type emission variability and Planetary Boundary Layer rise? 51

Conclusions RRF Modeled data are “averaged” over weekday and weekend emissions and over recurring meteorological phenomena. Averaging of these phenomena results in an artificial response in many cases is less responsive than actual conditions Houston Complex Environment with different response to across the board controls Several ways to produce ozone require different controls 52

Outline Introduction Modeling Datasets Results Conclusions Future work How is model predicted O 3 sensitive to day type emission variability and Planetary Boundary Layer rise? 53

Future Work Develop DV B as a function of weekend and weekday. Calculate only weekend and only weekday DV F Compare Slow Riser and Fast Riser phenomena with Observed data. Use Process Analysis to compare eastern clusters of monitors with western monitors Design optimum control strategies that consider variability due to geographic location, MMV rise, and day type emissions of Houston Expand criteria for days selected when using the RRF averaging metric 54

Acknowledgements Dr. Vizuete MAQ Lab CHAQ Lab Dr. Jeffries Dr. Arnold Barron Henderson This project was funded by the HARC under Project H97. Image Sources Liz Christoph Sources Evan Couzo Barron Henderson Dick Karp. Initial 2018 hgb modeling results. 55

Questions 56

Appendix : Process Analysis Results H2O2 and HNO3 Production 57

Appendix : Process Analysis Results H2O2 and HNO3 Production OH + NO 2  HNO 3 ________________ HO 2 + HO 2  H 2 O 2 + O 2 58

Weekend 59