Download presentation
Presentation is loading. Please wait.
Published byStanley Randall Modified over 9 years ago
1
Three-State Air Quality Study (3SAQS) Three-State Data Warehouse (3SDW) 3SAQS 2011 Modeling Update University of North Carolina (UNC-IE) ENVIRON International Corporation (ENVIRON) April 29, 2015
2
2 Summary 3SAQS Base 2011 version A (Base11a) MPE Status Additions to the Base11a evaluation suite 2011 MPEv2 Base11a Platform release Schedule
3
3 3SAQS Pilot Project Timeline 20122013 2014 OCT 2012 Pilot Project Start JAN 2013 EI Improvement Meetings With CO, UT, WY NOV-DEC 2012 CO, UT, WY 2008 EI Analysis and Evaluation FEB 2014 3SAQS Monitoring Network Report OCT 2013 WRF 2011 Sensitivities AUG 2014 CAMx 2008b, WRF 2011, 2011 Oil & Gas EI, 2011a Emissions JUL 2014 Final 2008 Emissions OCT 2014 CAMx 2011a MPE SEP 2014 SMOKE 2011a NOV 2014- APR 2015 CAMx 2011 Sensitivities AUG 2012 CAMx 2008a JUN 2012 Emissions VOC Reactivity Analysis FEB-MAY 2012 2008 EI Modeling MAR-JUN 2014 2011 EI Modeling Dec 2014 – Apr 2015 Additional MPE for Simulation Base11a
4
4 3SAQS Base 2011a MPE Draft 1 of the 2011a MPE report released November 2014 – O 3, NO 2, CO, SO 2, PM 2.5, wet deposition Draft 2 in progress – VOCs, More PM 2.5, GEOS-Chem boundary conditions, CMAQ, dry deposition, visibility
5
5 3SAQS Base 2011a MPE PM 2.5 Performance Significant PM performance Issues – Total PM 2.5 overestimates related to boundary condition dust – Urban OC overestimates related to residential wood combustion (RWC) – Persistent summer NO 3 negative biases and winter SO 4 positive biases.
6
6 3SAQS Base 2011a MPE PM 2.5 Performance Boundary Conditions MOZART BCs found to be contributing too much dust to the CAMx simulation Replacing MOZART with GEOS-Chem BCs drastically improves Total PM 2.5 model performance 4-km domain IMPROVE spring NMB: MZBC = 155% GCBC = -8.5% 3SAQS BC Sensitivity Wiki* 3SAQS BC Sensitivity Wiki *http://vibe.cira.colostate.edu/wiki/wiki/1043/bc-sensitivity-modeling-results
7
7 3SAQS Base 2011a MPE PM 2.5 Performance Winter Urban Organic Carbon (OC) Systematic urban monitor (CSN) winter OC overbias from local/regional sources in 2011 -Not seen in 2008 WestJump 4-km domain average plots indicative of performance at individual monitors
8
8 3SAQS Base 2011a MPE PM 2.5 Performance Winter Urban Organic Carbon (OC) Statewide Colorado 2011 residential wood combustion PM 2.5 inventory decreased, primarily driven by El Paso County Increases in urban counties from 2008 to 2011 RWC significant anthropogenic PM2.5 source in urban counties 2008 2011 Colorado 2011 PM2.5 Emissions Colorado 2011 POC Emissions RWC was by far the largest source of urban POC in the CO 2011 inventory 2011 daily 4-km domain total RWC POC emissions
9
9 Recommendations If using MOZART BCs, apply cap on dust and remove all sea salt -MOZART BCs slightly outperform GEOS-Chem for O 3 and speciated (non-dust) PM 2.5 Review the 2011 RWC inventory with states -Option to either cut urban county inventories by 50% or revert to 2008 levels 3SAQS Base 2011a MPE PM 2.5 Performance
10
10 3SAQS Base 2011a MPE CMAQ vs. CAMx CMAQv5.0.2 run for 4-km domain – Annual simulation with 2011a emissions and met – SMOKE routine developed (Firemtx) to spatially distribute fire emissions based on acreage, grid cell size, and land use – BCs extracted from CAMx GEOS-Chem BC 12-km simulation Model performance shown relative to both observations and CAMx Base11a_GCBC 4-km simulation
11
11 3SAQS Base 2011a MPE CMAQ vs. CAMx: AQS MDA8 CMAQ (NMB: -0.2%, NME: 12.2%) CAMx (NMB: -0.76%, NME: 14.5% ) 4-km domain-wide annual performance between CMAQ and CAMx is comparable
12
12 3SAQS Base 2011a MPE CMAQ vs. CAMx: CASTNET MDA8 CMAQ (NMB: -3.17%, NME: 9.39%) CAMx (NMB: -1.25%, NME: 9.25% ) 4-km domain-wide annual performance between CMAQ and CAMx is comparable
13
13 3SAQS Base 2011a MPE CMAQ vs. CAMx: AQS NO 2 CMAQ (NMB: 32.3%, NME: 79.1%) CAMx (NMB: 43.5%, NME: 82.7% ) CMAQ estimates consistently lower NO 2, but still overestimating in all months
14
14 3SAQS Base 2011a MPE CMAQ vs. CAMx: AMON NH 3 CMAQ (NMB: -63.4%, NME: 69.2%) CAMx (NMB: -66.2%, NME: 69.0% ) 4-km domain-wide annual performance between CMAQ and CAMx is comparable
15
15 3SAQS Base 2011a MPE CMAQ vs. CAMx: IMPROVE Total PM2.5 CMAQ (NMB: -26.2%, NME: 56.0%) CAMx (NMB: -8.7%, NME: 57.0% ) Moderate performance improvements with CAMx relative to CMAQ
16
16 3SAQS Base 2011a MPE CMAQ vs. CAMx: IMPROVE PM2.5 Species NMB (%) SpeciesCMAQCAMx SO4- 2.97 8.03 NO3- 22.2 - 31.2 EC- 32.2 - 15.3 OC- 37.6 -8.5 NH4- 34.9 - 31.7
17
17 3SAQS Base 2011a MPE CMAQ vs. CAMx: CSN Total PM2.5 CMAQ (NMB: 64.5%, NME: 95.9%) CAMx (NMB: 79.8%, NME: 109.0% ) Moderate performance improvements with CMAQ relative to CAMx; systematic winter biases
18
18 3SAQS Base 2011a MPE CMAQ vs. CAMx: CSN PM2.5 Species NMB (%) SpeciesCMAQCAMx SO414.519.0 NO3- 24.5 - 46.9 EC118147 OC344270 NH4-1.2- 17.5
19
19 3SAQS Base 2011a MPE CMAQ vs. CAMx: CSN PM2.5 Species NMB (%) SpeciesCMAQCAMx SO414.519.0 NO3- 24.5 - 46.9 EC118147 OC344270 NH4-1.2- 17.5
20
20 CMAQ and CAMx give comparable results for most species Biggest differences in carbonaceous aerosols at rural sites Both models impacted by systematic deficiencies in the input data and model formulations For basic simulations (i.e. no diagnostic or source apportionment) the models are fairly interchangeable Differences in the models will be highlighted when exercising unique features (e.g. PSAT in CAMx or NH3 BiDi in CMAQ) 3SAQS Base 2011a MPE CMAQ vs. CAMx
21
21 3SAQS Base 2011a MPE Dry Deposition CASTNET reports weekly dry deposition for SO 4, NO 3, Total NO 3, NH 4, and SO 2 Results for Base11a 4-km simulation
22
22 3SAQS Base 2011a MPE SO 4 Dry Deposition 1.Modeled DryDep-to- CASTNET Obs 2.CASTNET DryDep 3.DD quartile 4.CASTNET concentrations 5.4-km average DryDep 6.4-km average concentrations
23
23 3SAQS Base 2011a MPE NO 3 Dry Deposition 1.Modeled DryDep-to- CASTNET Obs 2.CASTNET DryDep 3.DD quartile 4.CASTNET concentrations 5.4-km average DryDep 6.4-km average concentrations
24
24 3SAQS Base 2011a MPE NH 4 Dry Deposition 1.Modeled DryDep-to- CASTNET Obs 2.CASTNET DryDep 3.DD quartile 4.CASTNET concentrations 5.4-km average DryDep 6.4-km average concentrations
25
25 3SAQS Base 2011a MPE Dry Deposition Concerns about the CASTNET measurements – CASTNET dry deposition estimates use a climatological average deposition velocity by species and site – CAMx and CMAQ calculate the dry deposition velocity using simulated meteorology and aerosols CASTNET may not be reliable for evaluating dry deposition model performance. – Effectively comparing a model to a model, not a model to obs
26
26 3SAQS Base 2011a MPE Visibility Species and total extinctions calculated at IMPROVE sites with revised IMPROVE equation MOZART BCs create big problems for visibility calculations Base11a MZBC Base11a GCBC Species NMBNMENMBNME Units %% Total 4.429.3-0.432.6 SO4 0.951.210.250.4 NO3 -9.583.7-28.299.9 OC -37.369.65.892.2 EC -2.269.6-16.664.9 CM -43.683.0-24.875.9 SS 570679240348 Soil 25932024.494.5 4-km domain average species extinction MPE
27
27 3SAQS Base 2011a MPE Visibility: 20% Best Observed Days MOZART BCGEOS-Chem BC
28
28 3SAQS Base 2011a MPE Visibility: 20% Worst Observed Days MOZART BCGEOS-Chem BC
29
29 3SAQS Base 2011a Attainment Test 8-hour Ozone Attainment Test Modeled Attainment Test Software (MATS) configuration – Version 2.6.1 – 5-year weighted average (2009-2013) – Maximum concentration in a 3x3 array of model grid cells around the monitor using inverse distance weighting – No spatial gradient adjustment – Sites need to have a concentration >= 60 ppb for 5 or more days in the year to be plotted
30
30 3SAQS Base 2011a Attainment Test Base Year DVs (Base2011a) Future Year DVs (Base2020_11a)RRFs Max DV = 80.7ppbMax DV = 78.7ppb
31
31 3SAQS Base 2011a Attainment Test No future year simulation with GEOS-Chem BCs, so can’t calculate DVF and RRFs EPA has not yet released the 2013 PM2.5 design values – MATS isn’t working correctly using a straight 3- year average (2010-2012) – Waiting for 2013 to complete the 5-year weighted average and calculate PM2.5 and visibility DVs and RRFs
32
32 3SAQS 2011a MPE version 2 Additions from v1 (November 2014) – PM level plots – Dry deposition – AMON NH3 – CMAQ – GEOS-Chem BC sensitivity results – Garfield County, Mesa County, and Upper Green River VOCs Release to collaborators in May
33
33 WSAQS Project Timeline 2015 DEC 2014- APR 2015 Base11a MPEv2 JAN-MAR CMAQ Base11a JAN-APR WRF Winter Modeling Configuration MAR-JUN 2011b Emissions Development AUG 2011b CAMx MPE JUN-JUL 2011b CAMx Modeling Early SEP 2014 Technical Comm Meeting SEP 2014 Release CAMx 2011b NOV 2014a Platform Release April 29-30 Tech Comm Call MAY Base2011a Platform Release; MOVES and Winter O3 CAMx Complete APR Phase II O&G EI Release MAY-AUG 2014a WRF and SMOKE JAN 21 Governing Board Meeting FEB 25, 2015 Technical Comm Meeting JUL 2014 Technical Comm Call
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.