Comparison of PM Source Apportionment and Sensitivity Analysis

Comparison of PM Source Apportionment and Sensitivity Analysis in CAMx Bonyoung Koo, Gary Wilson, Ralph Morris, Greg Yarwood ENVIRON Alan Dunker General Motors R&D Center 8 th Annual CMAS Conference October 19 -21, 2009 Chapel Hill, North Carolina Template

Probing Tools in CAMx • Source Apportionment – Ozone Source Apportionment Technology (OSAT) – Particulate Source Apportionment Technology (PSAT) – Reactive Tracer Source Apportionment (RTRAC) • Sensitivity Analysis – Decoupled Direct Method (DDM) for gas and particulate species – Higher-order DDM (HDDM) for gas-phase species • Process Analysis – Integrated Process Rate (IPR) – Integrated Reaction Rate (IRR) – Chemical Process Analysis 2009 CMAS Conference 2

Probing Tools in CAMx • Source Apportionment – Tagged Species – Ozone Source Apportionment Technology (OSAT) – Particulate Source Apportionment Technology (PSAT) – Reactive Tracer Source Apportionment (RTRAC) • Sensitivity Analysis – Decoupled Direct Method (DDM) for gas and particulate species – Higher-order DDM (HDDM) for gas-phase species • Process Analysis – Integrated Process Rate (IPR) – Integrated Reaction Rate (IRR) 2009 CMAS Conference 3

Probing Tools in CAMx • Source Apportionment – Ozone Source Apportionment Technology (OSAT) – Particulate Source Apportionment Technology (PSAT) – Reactive Tracer Source Apportionment (RTRAC) • Sensitivity Analysis – Decoupled Direct Method (DDM) for gas and particulate species – Higher-order DDM (HDDM) for gas-phase species • Process Analysis – Integrated Process Rate (IPR) – Integrated Reaction Rate (IRR) – Chemical Process Analysis 2009 CMAS Conference 4

Probing Tools in CAMx • Source Apportionment – Ozone Source Apportionment Technology (OSAT) – Particulate Source Apportionment Technology (PSAT) – Reactive Tracer Source Apportionment (RTRAC) • Sensitivity Analysis – Decoupled Direct Method (DDM) for gas and particulate species – Higher-order DDM (HDDM) for gas-phase species • Process Analysis – Integrated Process Rate (IPR) – Integrated Reaction Rate (IRR) – Chemical Process Analysis 2009 CMAS Conference 5

Brute-Force Method Pollutant Concentration DCBFM 0 2009 CMAS Conference E 1 E 0 Emission 6

First-Order Sensitivity Pollutant Concentration DDM DCBFM 0 2009 CMAS Conference DCDDM E 1 E 0 Emission 7

Source Apportionment Pollutant Concentration DDM DCBFM DCPSAT BFM PSAT 0 2009 CMAS Conference DCDDM E 1 E 0 Emission 8

Zero-Out Contribution Pollutant Concentration BFM DCPSAT = DCBFM PSAT 0 2009 CMAS Conference DCDDM E 0 Emission 9

PM Modeling Episode • February & July from the St. Louis 36 -/12 km 2002 PM 2. 5 SIP modeling • Urban & rural receptors: – 2 PM 2. 5 NAAs – 6 Federal Class-I areas • BFM reductions of 20% and 100% in various emission Chicago PM NAA (CNAA), species from. St. Louis PM NAA (SNAA), Mingo wilderness area (MING), Herculesanthropogenic Glades wilderness area (HEGL), Upper Buffalo wilderness area (UPBU), Caney Creek wilderness sources area (CACR), Mammoth Cave national park 2. 5 (MACA), and Sipsey wilderness area (SIPS) 2009 CMAS Conference 10

Contributions of Point-Source SO 2 to PM 2. 5 Sulfate February July 2009 CMAS Conference 11

Contributions of Point-Source SO 2 to PM 2. 5 Sulfate February Oxidant-limiting effects July 2009 CMAS Conference 12

PM 2. 5 Sulfate Changes due to Point SO 2 Emiss Reductions 2009 CMAS Conference 13

PM 2. 5 Sulfate Changes due to Point SO 2 Emiss Reductions Oxidant-limiting effect 2009 CMAS Conference 14

PM 2. 5 Sulfate Changes due to Point SO 2 Emiss Reductions Non-linear responses 2009 CMAS Conference 15

PM 2. 5 Sulfate Changes due to On-road MV Emiss Reductions 2009 CMAS Conference 16

PM 2. 5 Sulfate Changes due to On-road MV Emiss Reductions Indirect effect February: Reducing NOx emissions Lower acidity of the aqueous phase More SO 2 dissolves in the aqueous phase More sulfate produced Negative Sensitivity 2009 CMAS Conference 17

PM 2. 5 Sulfate Changes due to On-road MV Emiss Reductions Indirect effect July: Reducing NOx emissions Less oxidant available to oxidize SO 2 Further reduction in sulfate Positive Sensitivity 2009 CMAS Conference 18

PM 2. 5 Ammonium Changes due to Area NH 3 Emiss Reductions 2009 CMAS Conference 19

PM 2. 5 Nitrate Changes due to Area NOx Emiss Reductions 2009 CMAS Conference 20

PM 2. 5 Nitrate Changes due to On-road MV Emiss Reductions Less indirect effec because NOx dominates on-roa MV emission 2009 CMAS Conference 21

PM 2. 5 SOA Changes due to Area VOC Emiss Reductions 2009 CMAS Conference 22

Primary PM 2. 5 Changes due to On-road MV Emiss Reductions 2009 CMAS Conference 23

Summary • 1 st-order DDM sensitivities agree well with the BFM model responses to small emission changes (20%) – With large emission changes, non-linearity comes into play – For SOA and primary PM 2. 5, the DDM works relatively well even with 100% emission reductions • PSAT and zero-out are nearly equivalent in cases with no indirect effect – PSAT starts to deviate from the zero-out contribution as indirect effects from limiting 2009 CMAS Conference 24

Summary (cont. ) • Source sensitivity and source apportionment are equivalent for pollutants that are linearly related to emissions; However, when they are different: – PSAT is best at apportioning PM pollutants to sources emitting their primary precursors (e. g. , sulfate to SO 2, nitrate to NOx) – DDM sensitivities are more accurate than PSAT in determining the impact of emissions that have indirect effects on secondary PM – PSAT works better at estimating the impact of zeroing-out a source while DDM does generally better when a fraction of emissions are eliminated 2009 CMAS Conference 25

Summary (cont. ) • BFM (zero-out) also has limitations: – Computationally expensive and subject to numerical noises – Sum of the BFM source contributions will not always equal the simulated concentrations in the base case 2009 CMAS Conference 26

Acknowledgement • Funded by the Coordinating Research Council For more details… Koo, B. , G. M. Wilson, R. E. Morris, A. M. Dunker and G. Yarwood. 2009. “Comparison of Source Apportionment and Sensitivity Analysis in a Particulate Matter Air Quality Model. ” Environ. Sci. Technol. , 43 (17), pp 6669 -6675. doi: 10. 1021/es 9008129 2009 CMAS Conference 27