Reducing Air Pollution In Los Angeles

University of California at Riverside Review of N2O5 Hydrolysis in Urban/Regional Scale Modeling Gail Tonnesen, Chao-Jung Chien, Zion Wang, Mohammad Omary University of California, Riverside Bourns College of Engineering Center for Environmental Research and Technology October 22, 2003, AAAR Air Quality Modeling Group University of California at Riverside Motivation • Chamber and flow reactor experiments suggest that N2O5 hydrolysis may occur by homogeneous and heterogeneous pathways. • Global modeling showed 20% (summer) to 80% (winter) of HNO3 production via N2O5. (D&C,93) • Until recently little attention to N2O5 hydrolysis in urban/regional scale modeling – Gas phase modeling uses a pseudo gas phase reaction to account for heterogeneous rxns. – Simple parameterizations used in regional modeling. Air Quality Modeling Group University of California at Riverside Key Questions on N2O5 Hydrolysis • • • • Important for Summer as well as Winter? How important for 1-hr or 8-hr O3? Effects on aerosol nitrate? Other PM species? Is homogeneous or heterogeneous hydrolysis more important? • Can modeling & comparison to ambient monitoring and be used to bound the uncertainty in N2O5? Air Quality Modeling Group University of California at Riverside Models and Episodes • Models-3/CMAQ – – – – WRAP western US 1996 (gas and PM) VISTAS Continental US (gas and PM) CCOS 2000 (gas) SCOS 1997 (gas) • CAMx – CCOS 2000 (gas) – SCOS 1997 – Houston 1993 (gas) Air Quality Modeling Group University of California at Riverside Previous N2O5 Hydrolysis Approaches • Ozone modeling CAMx/UAM/CMAQ: – CB4 use an pseudo gas rate constant – SAPRC99 uses Mentel et al upper limit gas rate constant • Regional Acid Deposition Modeling (RADM): – kN2O5 = 0.2 min-1 if RH > 70% • Recent CMAQ Versions (beginning July 2002) use heterogeneous hydrolysis – Specify the Reaction Probability (gamma). – Simple parameterizations have been used in regional modeling. Air Quality Modeling Group University of California at Riverside Gas Phase Reaction • • • • • CB4 N2O5+H2O k = 1.3E-21 molec-1 s-1 SAPRC99 N2O5+H2O k = 2.6E-22 molec-1 s-1 RADM N2O5+H2O k = 2.0E-21 molec-1 s-1 RACM: no gas reaction CB4-2002 N2O5+H2O k = 2.5E-22 molec-1 s-1 N2O5+H2O+H2O k = 1.8E-39 molec-2 s-1 Air Quality Modeling Group University of California at Riverside Western US 1996 PM Modeling • CMAQ v2 (2001): • CMAQ v3 (2002): • CMAQ v4.2: • CMAQ vUCR: gas only, k = 1.3E-21 gas k = 2.6E-22,  = 0.1 0.002 <  < 0.02; f(SO4/NO3) 0.01 <  < 0.06; f(RH) Air Quality Modeling Group University of California at Riverside January time-series plots production of HNO3 for CMAQ v2 P(HNO3) for Old CMAQ from January 24-27 NO3+VOC 6.0E+07 N2O5+H2O OH+NO2 6.00E+07 5.0E+07 5.00E+07 4.0E+07 4.00E+07 Total moles 3.0E+07 3.00E+07 2.0E+07 2.00E+07 1.0E+07 1.00E+07 0.0E+00 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 0.00E+00 GMT Time (hr) Air Quality Modeling Group University of California at Riverside January time-series plots production of HNO3 for CMAQ v3 P(HNO3) for New CMAQ from January 24-27 NO3+VOC 6.0E+07 N2O5+H2O OH+NO2 N2O5_aero 6.00E+07 5.0E+07 5.00E+07 4.0E+07 4.00E+07 Total moles 3.0E+07 3.00E+07 2.0E+07 2.00E+07 1.0E+07 1.00E+07 0.0E+00 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 0.00E+00 GMT Time (hr) Air Quality Modeling Group University of California at Riverside July time-series plots production of HNO3 for CMAQ v2 P(HNO3) for Old CMAQ from July 24-27 NO3+VOC 6.0E+07 N2O5+H2O OH+NO2 6.00E+07 5.0E+07 5.00E+07 4.0E+07 4.00E+07 Total moles 3.0E+07 3.00E+07 2.0E+07 2.00E+07 1.0E+07 1.00E+07 0.0E+00 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 0.00E+00 GMT Time (hr) Air Quality Modeling Group University of California at Riverside July time-series plots production of HNO3 for CMAQ v3 P(HNO3) for New CMAQ from July 24-27 NO3+VOC 6.0E+07 N2O5+H2O OH+NO2 N2O5_aero 6.00E+07 5.0E+07 5.00E+07 4.0E+07 4.00E+07 Total moles 3.0E+07 3.00E+07 2.0E+07 2.00E+07 1.0E+07 1.00E+07 0.0E+00 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 0.00E+00 GMT Time (hr) Air Quality Modeling Group University of California at Riverside IRR Analysis of HNO3 Production in January 1.20E+10 1.00E+10 Total moles 8.00E+09 6.00E+09 4.00E+09 2.00E+09 0.00E+00 Jan-New N2O5 (aero) --> HNO3 N2O5+H2O --> 2HNO3 NO3+VOC --> HNO3 OH+NO2--> HNO3 8.6265E+09 1.7237E+08 2.3232E+08 2.2461E+09 Jan-Old 2.0705E+09 2.4606E+08 2.3856E+09 Air Quality Modeling Group University of California at Riverside Totals for HNO3 Production in July 1.20E+10 1.00E+10 Total moles 8.00E+09 6.00E+09 4.00E+09 2.00E+09 0.00E+00 July-New N2O5 (aero) --> HNO3 N2O5+H2O --> 2HNO3 NO3+VOC --> HNO3 OH+NO2--> HNO3 3.0583E+09 1.8122E+08 2.8134E+08 6.2058E+09 July-Old 1.0748E+09 3.1598E+08 6.2463E+09 Air Quality Modeling Group University of California at Riverside WRAP NO3 Model Evaluation Summary 1996 IMPROVE data Mean Normalized Bias January July 810% 66% 1214% 428% 534% 3% Normalized Mean Bias 98% -33% Air Quality Modeling Group • v2 • v3 • v4.2 • v4.3 University of California at Riverside VISTAS NO3 Model Evaluation Summary 1996 IMPROVE data Normalized Mean Bias Jan 02 July 99 180% -50% -44% • v4.3 • vUCR Air Quality Modeling Group University of California at Riverside SCOS Aug, 1997, LA, Vertical Slice Air Quality Modeling Group University of California at Riverside SCOS Aug 3-7, 1997 CMAQ 215 t/d CAMx 629 t/d Air Quality Modeling Group University of California at Riverside Houston 1993 SIP Modeling Air Quality Modeling Group University of California at Riverside Houston 1993 SIP Modeling Air Quality Modeling Group University of California at Riverside CAMx Sensitivity in Houston • Yarwood and Mansell (2002) investigated sensitivity of ozone to uncertainties in N2O5 chemistry in the 1993 CAMx O3 SIP modeling: – Found small increases of O3 (4 ppb) when reducing k to the Mentel value. – Concluded effects are not very important for the Houston area . – Did not evaluate possible changes in O3 sensitivity to VOC & NOx Air Quality Modeling Group University of California at Riverside CCOS 2000 7/30-8/2 • xx Air Quality Modeling Group University of California at Riverside CCOS 2000 O3 under predicted Air Quality Modeling Group University of California at Riverside Conclusions • Counter to intuition, N2O5 chemistry appears to have little effect on peak O3. • Factors other than rate constant affect N2O5 hydrolysis. Vertical mixing, h-diffusion? • Can not use ambient modeling to bound uncertainty in rate constant. • Need more experimental testing: – Chamber experiments – Requires NO3, N2O5 & sensitive HNO3 methods Air Quality Modeling Group