Emissions Modeling For Photochemical Modelers Mark Janssen LADCO Photochemical Modelers Training August 3-4 th 2010
Overview Emissions Modeling (EM) Issues Every Photochemical Modeler Needs To understand What to look for in Emissions to identify quality Changes coming to Emissions Modeling Obstacles to the next round of SIPS
Base J Base K Improved NH3 Dust Trans Beta of Megan Base M2 Rain Problem Base M1 Cars/Trucks
Issues Every Photochemical Modeler needs to know The change to Multi-pollutant Modeling has changed the way emissions modeling is done. Traditional Ozone Inventories had 5 sectors(point, Nonroad, Area, Onroad, Biogenics) New inventories added(fires, agricultural NH3, 2X area source categories, Complex EGU, ) Annual Month Specific Inventories need better temporal and new categories(residential wood and snowmobiles)
How Has This Changed EM? Need more than 1 emissions modeler to run all the sectors for all the days necessary for a SIP. Emissions modelers have less understanding of each individual sector. Regional Groups produce some sectors(fires, Dust, Rail, Marine) More contractor support to accomplish goals which leads to misunderstandings between groups.
3 Major Parts of Emissions Modeling Spatial Allocation Emissions modelers take inventories developed at different spatial scales to create modeling grid level inventories Surrogates can be a poor for certain categories. Housing = Lawn and Garden, water = Recreational Marine. Result in concentration or dilution of emissions.
Temporal Allocation Turn annual inventories into day and hour specific emissions estimates. Most complex job of emissions modelers. Motorboats, Agricultural Ammonia, EGUs. Parts: Year to Month, Month to Day of Week, Day of Week to Hour of day. Each is a function of the one before. Most obvious errors/differences come from this activity.
Chemical Speciation/Lumping Inventories are by pollutant but transport models need lumped groups. Default speciation acceptable for most categories but local fuels can influence results. VOC is generally stable but PM is changing all the time. Least likely to cause problems and difficult for modelers to detect problems. (How much ALD2 is too much for a given cell?)
What Modelers look for in emissions to identify quality No odd sources in lakes, recognizable activity between urban areas(population is not the only surrogate) Should see recognizable seasonal variation Weekdays, Saturdays, and Sundays differ Hourly variation by sector differs: non-egu point is mostly flat EGU has noticeable daytime peak, summer is lower than winter(sipcall) Dust, Biogenics should change with Meteorology
What to look for in Emissions to identify quality(2) Everyone involved sees how the top 15 categories of PM are speciated. Back calculated emissions from CMAQ/CAMX files matches emissions inventory reports. Remember apply correction for difference between actual average molecular weight and the photochemical models assumed molecular weight.
Future of Emissions Modeling 1KM Resolution Grids: As models like AERMOD and Benmap become more integrated we will need finer spatial resolution to inventories. More Meterological effects and days specific emissions on the way. 5 years ago(onroad, Biogenics) Now(Fires, limited EGU) 5 years( Complex EGU, Agricultural NH3, Nonroad)
Future of Emissions Modeling(2) Fires will likely all be satellite derived and days/hour specific.(southeast may be an exception) Agricultural NH3 will use process based models Meteorology, Local Farming Practices have strong influence on emissions(cow in Texas 76lbs/year, Cow in Wisconsin 14 lbs/year) Must Capture Winter NH3 spikes during episodes.
Ammonia Emissions from a Commercial Broiler House Worley, J.W. 2002. ASAE Paper 024118
Kg Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Predicted monthly NH3-N emission, kg/cow-mo Predicted monthly NH3-N emission, kg/cow-mo 4.5 Predicted NH 3 Emissions from Dairy Lagoon under Different ph (H=25ft and TAN=450mg/L, Fresno) 3.0 Predicted NH 3 Emissions from Dairy Lagoons in Fresno and San Joaquin 4.0 3.5 3.0 ph = 7.0 ph = 7.4 ph = 7.8 2.5 2.0 Fresno San Joaquin 2.5 2.0 1.5 1.5 1.0 1.0 0.5 0.5 0.0 0.0 1.8 Month of the year Month of the Year 1.6 1.4 1.2 1 0.8 7/2/2002 7/3/2002 0.6 0.4 0.2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (GMT)
Future of Emissions Modeling Onroad Mobile Link Networks Built on Travel Demand Models or Counts Only way to build high spatial resolution inventory Starts and Stop emissions geography use different activity than road traffic. (15 story parking ramps at 5PM) How do you validate speeds, vehicle mix, etc in countywide data? High Temporal Resolution Not just VMT but other activities(vehicle mix, speed) While VMT is down 20% on weekends, Weekend NOX is down >50% because of HDDV decreases. If there is congestion it occurs at different times and is likely not bi-modal. Congestion modeling of speed to capture 5 MPH to 75 MPH swings in speed overlayed on complex VMT Organic Carbon - Semi-Volatile Organic carbon(c8-c40)
Interstate Vehicle Mix Percent by State and Urban/Rural Vehicle mix 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 U-IL U-IN U-MI U-MN R-IL R-IN R-MI R-MN 0 1 2 3 4 5 6 7 Day Of Week
Need to Improve On-Road OC (Primary) Emissions Nonroad Point-EGU SOA/Sulfate 10% Biogenic 6% Burning 12% Point-NonEGU On-road (diesel) Industrial 20% On-road (gas) Area Burning Mobile 52% LADCO Regional Emissions Typical Monitor-Based Source Apportionment Result
Obstacles To The Next Round Of SIPS. 2008 inventory not complete until 9/2010, Might need modeling done by 1/2011. EPA s EIS system cause state invenory folks problems. EPA will not deliver NEI until fall 2010 No framework to share data between groups(mjos, States, EPA) SOLUTION Collaberative Data Sharing Group EGU Growth Methods(IPM Not likely) SOLUTION: ERTAC
Obstacles To The Next Round Of SIPS(2). PM25 Condensable as a new pollutant. MOVES not released until December 2009 slow600+ CPU days to run a Year) New, LADCO won t have state inputs until DEC 2010. Winter PM OC Increases NH3 model may not be ready until OCT/NOV Biogenics from MEGAN may be problematic.
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec tons/month Impact of MOVES emissions on onroad gasoline vs. NMIM PM2.5 national onroad gasoline 14,000 12,000 10,000 8,000 6,000 4,000 2,000 NMIM MOVES @72ºF MOVES w/temp 0 20 Courtesy: Marc Houyoux, USEPA
Priority ERTAC Projects Eastern Regional Technical Advisory Committee Rail emissions Mobile source PM emissions Agricultural ammonia Area source comparability Electric Utility Temporal and Growth