1. Development of Wildland Fire Emission Inventories with the BlueSky Smoke Modeling Framework Sean Raffuse, Erin Gilliland, Dana Sullivan, Neil Wheeler, and Lyle Chinkin Sonoma Technology, Inc., Petaluma, California Sim Larkin, Robert Solomon, Tara Strand U.S. Forest Service AirFIRE Team Thompson Pace U.S. EPA OAQPS Presented at the 7th Annual Community Modeling and Analysis System (CMAS) Conference Chapel Hill, North Carolina October 7, 2008 STI-3457

2. 2 Background NASA Decision Support NASA-STI-USFS Partnership SMARTFIRE * – Fire Information BlueSky Framework – Fuels, Consumption, and Emissions EPA Work Assignment Emission Inventory for Wildland Fires * Satellite Mapping Automatic Reanalysis Tool for Fire Incident Reconciliation

3. 3 Emission Inventory Processing Scope  Contiguous United States  August 2002 through 2006  Wildfires, Wildfire Use, prescribed burning  Agricultural burning excluded BlueSky Pathway  Fire information: SMARTFIRE  Fuel loading: FCCS (Fuel Characteristic Classification System)  Fuel consumption: Consume 3.0  Emissions: FEPS (Fire Emission Production Simulator)

4. 4 BlueSky Pathway ICS-209 SMARTFIRE Other Rx Sys FCCS NFDRS Hardy LANDFIRE Ag Other CONSUME 3 FEPS FOFEM ClearSky (Ag) Satellite Other FEPS WRAP FOFEM Rx / WF Idealized Manual Other EPM FEPS FOFEM Literature Other Briggs FEPS WRAP Multi-core Daysmoke Other CALPUFF CMAQ HYSPLIT BlueSky Framework 3.0 Fire Information Fuels Total Consumption Time Rate Emissions Plume Rise Dispersion / Trajectories

5. 5 Data Sources ICS-209 reports: AirFIRE, Fire and Aviation Management Web Applications (FAMWEB) web site, and Tom Pace HMS fire detects: NOAA Hazard Mapping System (Mark Ruminski) MODIS * fire detects: USFS Remote Sensing Applications Center (used to fill gaps in HMS data) Fuel moisture: USFS Wildland Fire Assessment System * Moderate Resolution Imaging Spectroradiometer

6. 6 0 50 100 150 200 250 300 350 400 JanFebMarApr May JunJulAugSepOctNovDec PM 2.5 (ktons) 2003200420052006 Annual PM 2.5 Primary Emissions Annual Total 1.67 1.10 0.90 1.28 0.00.51.01.52.0 2006 2005 2004 2003 PM 2.5 (million tons) Monthly Totals (2003 - 2006, Lower 48 States)

7. 7 Annual Average PM 2.5 Wildland Fire Emission Density (2003 – 2006)

8. 8 Average Monthly Wildland Fire PM 2.5 Emissions (2003-2006) Month

9. 9 2003 0 50 100 150 200 250 1/12/13/44/45/56/57/68/69/610/711/712/8 Area (thousand acres) 0 5 10 15 20 25 30 35 PM2.5 (ktons) AreaPM2.5 2004 0 50 100 150 200 250 1/12/13/34/35/46/47/58/59/510/611/612/7 Area (thousand acres) 0 5 10 15 20 25 30 35 PM2.5 (ktons) AreaPM2.5 2005 0 50 100 150 200 250 1/12/13/44/45/56/57/68/69/610/711/712/8 Area (thousand acres) 0 5 10 15 20 25 30 35 PM2.5 (ktons) AreaPM2.5 2006 0 50 100 150 200 250 1/12/13/44/45/56/57/68/69/610/711/712/8 Area (thousand acres) 0 5 10 15 20 25 30 35 PM2.5 (ktons) AreaPM2.5 2002 0 50 100 150 200 250 1/12/13/34/35/46/47/58/59/510/611/612/7 Area (thousand acres) 0 5 10 15 20 25 30 35 PM2.5 (ktons) AreaPM2.5 Area Burned vs. PM 2.5 Emissions

10. 10 Effect of Different Fire Information Sources y = 104x R 2 = 0.66 0 50 100 150 200 250 300 0100020003000 Pixel Count Acres Burned (Thousands) ICS-209s only  This was BlueSky’s previous data feed.  ICS-209s report cumulative area burned.  We subtracted the previous day’s area from today’s to get the daily area burned MODIS Terra and Aqua only  This is the most commonly used satellite-derived fire data set  We developed acres burned per pixel relationship by examining 30 wildfires  Used 100 acres per pixel

11. 11 SMARTFIRE vs. MODIS vs. ICS-209 Area Burned

12. 12 Differences Between MODIS and HMS Because HMS includes GOES and AVHRR derived fire pixels in addition to MODIS, it detects more fires overall. This is especially true in the southeast, where fires are often small and/or short lived. In addition to the increased coverage, HMS provides human quality control.

13. 13 MODIS consumption rate is much higher than the other two. Why? Yearly Totals Overall Consumption Rate (tons/acre) ICS-20910.0 MODIS17.6 SMARTFIRE11.9 MODIS consumption rate is much higher than the other two. Why? Area Burned Fuel Consumed 0 2 4 6 8 10 12 14 Million Acres ICS-209MODISSMARTFIRE 0 50 100 150 200 Million Tons 0.0 0.5 1.0 1.5 2.0 2003200420052006 Million Tons Fuel Consumed Area Burned PM 2.5 Emitted

14. 14 SMARTFIRE vs. MODIS vs. ICS-209 PM 2.5 Emissions

15. 15 Aspen Fire, Arizona – 2003

16. 16 Acres Burned - Aspen Fire, Arizona 0 2 4 6 8 10 12 14 6/176/206/236/266/297/27/57/87/11 Thousand Acres ICS209 SMARTFIRE Total ICS-209 84,750 SMARTFIRE 80,896 Helicopter 84,733 ICS-209 8,281 SMARTFIRE 3,659 Total Aspen Fire – Acres Burned vs. Emissions

17. 17 B & B Complex, Oregon – 2003

18. 18 Acres Burned - B&B Complex, Oregon 0 5 10 15 20 25 30 8/198/248/299/39/8 9/13 Thousand Acres ICS-209 MODIS SMARTFIRE Total ICS-209 91,500 SMARTFIRE 117,100 Helicopter 90,800 MODIS 189,800 Total ICS-209 10,700 SMARTFIRE 44,200 MODIS 73,100 B & B Complex, Oregon – 2003

19. 19 Future Work Continued validation and improvement of fire size parameters Further exploration of high consumption rate when using MODIS Differences with high resolution fuel loading data (LANDFIRE) –30-m spatial resolution is finer scale than the satellite fire information –Will use helicopter-flown perimeters or high resolution satellite burn scars to determine area burned Smoke and Emissions Model Intercomparison Project (SEMIP)

20. 20 Acknowledgments Funding from National Fire Plan, USFS, Joint Fire Science Program, EPA, DOI, and NASA ROSES DSS Our many collaborators and partners  Tom Pace (EPA)  Amber Soja (NIA)  Mark Ruminski and John Simko (NOAA NESDIS)  Roger Ottmar and Susan Pritchard (USFS-FERA) Thank you! Contact: sraffuse@sonomatech.com Information: www.getBlueSky.org