Community Smoke Emissions Model

1
Community Smoke Emissions Model

• WRAP Fire Emissions Forum Meeting
• December 2002
• Douglas Fox
• Cooperative Institute for Research in the
  Atmosphere
• Colorado State University
• Ft. Collins, CO 80524-1375
• fox_at_cira.colostate.edu

2

• Based on work presentations developed in
  cooperation with Dr. Mike Sestak (an independent
  consultant), Dr. Susan ONeill (research
  scientist in PNW Seattle FERA group), Dr. Sue
  Ferguson (PNW FERA), Dr. Jason Ching (EPA/NOAA
  research) Dr. Al Riebau (FS research)

3
History

• Technically Advanced Smoke Estimation Tools
  (TASET)
• JFSP project (Fox Riebau, 1998-2000)
• Establish FCAMMS Cooperative Agreement
  (2000-continuing)
• NPS Air Quality Division/CIRA (1999)
• Evaluate applicability of Models3/CMAQ in Western
  US (WRAP.)
• RD on fire emissions

4
SETS -- Evaluation
SETS -- Operational
SETS -- Tactical
SETS -- Strategic
SETS -- Strategic
TASET suggested Smoke Estimation Tool Sets need
to be populated with some different tools at each
level of activity but tools must be able to
interact between the activity levels.
5
USFS/Fire Consortia for Advanced Modeling of
Smoke and Meteorology (FCAMMS) will implement
smoke management using BlueSky by 2003.
6
BlueSky Smoke Modeling Framework

• Regional application
• Automated, centralized processing
• Emission tracking
• Prediction of surface concentrations
• Quantitative verification
• Community model development
• Web-access control and output products
• www.BlueSkyRAINS.org/

7
Needs for a Community Smoke Emissions Model

• Fire smoke is significant
• Current emissions inventories are labor intensive
• GCVTC, WRAP
• Potential Applications
• Regional Air Quality Modeling
• Regional Haze planning SIP development
• PM2.5
• Smoke Management
• Blue Sky Framework
• State based regulatory programs
• Land Manager inventory evaluations

8
Monitoring Data

• IMPROVE program measures visibility speciated
  aerosol data representing Class I areas relates
  them to each other for the regional haze rule
• Majority of fine particle species emitted from
  fires are organic and elemental carbon
  secondary organic aerosol formation is poorly
  understood.
• Wildland fire contributes to the 20 worst
  visibility days, especially in the west

9
Fire effects visibility

• Monthly OC contribution to total fine mass
  reaches 80 in some western US locations, longer
  term 10-30
• IMPROVE monitoring suggests a range of 10-40 of
  OC (organic carbon contribution to PM2.5) on the
  high mass days (20 worst visibility) may be from
  wild fires.

10
On-going research is attempting to quantify
fires contribution to organic aerosols p
Organic Carbon contribution to total extinction
Elemental Carbon contribution to total
extinction
11
Preliminary research results from the NPS
Yosemite field study August 2002, on particle
chemical composition

• Organics accounted for about 80 of the non-soil
  fine mass during the summer of 2002.
• The summertime organics in 1996-98 account for
  about 60 of the non-soil fine mass.
• Organics come from biomass fossil fuels

DRAFT not for publication
IMPROVE Data 1996-98
12
Why a Community Smoke Emissions Model?

• Common fire data
• Inputs not readily available
• Common modeling heritage
• Fire Behavior - BEHAVE
• Fuel Consumption - CONSUME
• Emissions Production EPM
• Emissions Factors
• Variety of applications
• Different objectives drive different accuracy
  resolution needs

13
Community Smoke Emissions Modeling
Identify Fires
Identify Fuels
Meteorology
Plume Rise
Input for Regional AQ model
14
What we CSEM istrying to do

• Goal to build a tool to generate emissions from
  forest burning for use in regional air quality
  modeling with the following characteristics
• Scale is regional to national with resolution
  ranging from 1 km to 36 km
• Temporal resolution from hourly to multi-year
• Chemical species including all NAAQS visibility
  components their precursors
• Accuracy equivalent to other emissions estimates.

15
Identify vegetation cover fuel loadings (1 km
resolution) Read from NFDR fuel model coverage
Modify with National FCC coverage
MM5 Meteorology 2pm local time Temperature
Relative humidity Cloud cover Wind
speed Daily Temperature range Relative humidity
range Past 7 days Precipitation Same as above
Generate species Emissions Plume Rise (hourly,
regional model resolution) Develop emissions
profiles to scale species from EPM generated
emissions to generate hourly emissions
distributions. Estimate plume rise based on
Briggs at appropriate resolution for the spatial
scale of emissions.
Calculate Fuel Moisture Content (daily, weekly,
regional model resolution)   NFDR calculations
based on MM5 input for range of variables at 36
km resolution
Calculate Fuel Consumption (daily, regional model
resolution) Utilize CONSUME to generate fuel
consumption and EPM to estimate emissions heat
release rate for each fire.
16
Assumptions about our approach

• Build a 1st order tool capable of estimating
  needed information from existing data
  information sources
• Accuracy scale needed are compatible with the
  National Fire Danger Rating System (NFDR)
• Based on historical fire data
• Meteorological data generated from MM5 /or
  higher resolution diagnostic models.

17
Approach outline

1. Identify fire boundaries
2. Identify vegetation fuels involved
3. Calculate fuel moisture content
4. Calculate fuel consumption
5. Calculate fire emissions
6. Speciate fire emissions calculate plume rise.

18
Identify fire boundaries

• Time, location, size of fires determined from
  National Fire Occurrence Database (Hardy, et.al.
  Missoula Fire Lab.)
• Federal most State fires, from 1986-1996, at
  1km resolution in a daily GIS database .

19
(No Transcript)
20
Identify vegetation fuels

• Identify NFDR fuel model at 1 km resolution from
  Bergen, et.al., 1998
• Modify fuel loading, if necessary, using fuel
  National Current Condition Class coverage
  (Hardy, et.al. Missoula Fire Lab.)

Identify vegetation cover fuel loadings (1 km
resolution) Read from NFDR fuel model coverage
Modify with National FCC coverage
21
(No Transcript)
22
Optional modifier for NFDR fuel loadings, if
needed to replicate WRAP 96 fire emissions
23
Calculate fuel moisture content

• Use NFDR equations based on data from MM5
  including daily temperature RH range, wind
  speed, cloud cover, precip.
• Drought indices from MM5
• Resolution from MM5

• Calculate Fuel Moisture Content
• (daily, weekly, regional model resolution)
•  
• NFDR calculations based on MM5 input for range of
  variables at 36 km resolution

24
(No Transcript)
25
Calculate fuel emissions

• Use CONSUME with NFDR model estimates of fuel
  loading moisture content.
• Use EPM to generate PM10, PM2.5, CO heat
  release rate.

Calculate Fuel Consumption (daily, regional model
resolution) Utilize CONSUME to generate fuel
consumption and EPM to estimate emissions heat
release rate for each fire.
26
Speciate emissions calculate plume rise

• Develop emissions profiles from ratios of species
  to calculated CO emissions from current research
  results.
• Calculate plume rise using Briggs per SASEM

• Generate species Emissions Plume Rise
• (hourly, regional model resolution)
• Develop emissions profiles to scale species from
  EPM generated emissions to generate hourly
  emissions distributions.
• Estimate plume rise based on Briggs at
  appropriate resolution for the spatial scale of
  emissions.

27
Emissions speciation
CO2 1521 g/Kg 1833CE g/Kg
CO 144 961 - (984CE)
CH4 6.8 42.7 (43.2CE)
PM2.5 12 67.4 (66.8CE)
PM10 14 1.18PM2.5
EC 0.7 (0.7) 0.072PM2.5
OC 5.8 (5.8) 0.54PM2.5
NOx 3.1 (2.0) 16.8MCE-13.1
NH4 0.6 0.012CE
VOC 6.8 (5.3) 0.085CE
SO2 0.8 (0.8)
Etc, etc.
CE DCO2 / DCODCO2DCH4DCother MCE
0.15.86CE
28
Preliminary Results

• Comparative data inputs from 2002 Oregon fire
  (actual vs. 1996 met)
• 
  BlueSky/FASTRACS CSEM
• Area of Burnsite acre
  500 500
• 0 - 0.25 inch fuel tons/acre
  1.0 2.9
• 0.25 - 1 inch fuel tons/acre
  2.2 2.3
• 1 - 3 inch fuel tons/acre
  1.6 5.6
• 3 - 9 inch fuel tons/acre
  5.4 13.2
• 9 - 20 inch fuel tons/acre
  24.6 0
• 20 inch fuel tons/acre
  0.1 0
• Duff
  8.0
  2.5
• Burn-site slope percent
  50 50
• Ignition time HHMM
  1400 1400
• 10-hr fuel moisture
  9 13.5
• Surface wind speed (mph)
  6 5.5

29
Preliminary Results

• Comparative emissions from 2002 Oregon fire
  (actual vs. 1996 met)
• Bluesky
  CIRA
• Time Heat Rel PM-10 Heat
  Rel PM-10
• 60 1.448E07 9079.0
  1.521E07 9054.1
• 120 1.495E07 9416.7
  1.533E07 9144.7
• 180 1.497E07 9429.6
  1.533E07 9145.7
• 240 1.497E07 9430.0
  1.533E07 9145.8
• 300 1.497E07 9430.1
  1.533E07 9145.8
• 360 1.497E07 9430.1
  1.533E07 9145.8
• 420 4.890E05 351.0
  116867.1 91.6
• 480 1.868E04 13.4
  1287.9 1.0
• 540 7.137E02 .5
  14.2 0.0
• 600 2.727E01 .0
  0.2 0.0

30
(No Transcript)
31
(No Transcript)
32
Preliminary Results
33
CSEM Summary

• A rational approach to generating forest fire
  emissions for regional scale modeling has been
  developed.
• Results appear to be consistent with site
  specific emissions estimates (BlueSky) but more
  testing is needed.
• Plans exist to incorporate CSEM into the SMOKE
  processor.

34
Challenges remaining

• Coding CSEM into appropriate emissions
  processors, i.e. SMOKE
• Testing sensitivities simulating WRAP 96 fire
  emissions
• Compare simulated emissions with WRAP 96 Fire
  Emissions results
• Adding smoke emissions into regional modeling
  (REMSAD CMAQ)
• Finding adequate input data for years since 1996.