Transcript - Atmospheric Chemistry Modeling Group

Importance of peat burning and injection heights in boreal fire emissions:
evaluation with MOPITT satellite observations for the summer 2004
Abstract Reference Number: 7046
S. Turquety (1,2), J. A. Logan (1), D. J. Jacob (1), R. C. Hudman (1), R. M. Yevich (1), F. Y. Leung (1), R. M. Yantosca (1), C. L. Heald (1),
L. K. Emmons (3), D. P. Edwards (3), G. W Sachse (4), J. Holloway(5), and the INTEX Science Team.
(1) Division of Engineering and Applied Science, Harvard University
(2) Now at Service d'Aéronomie, IPSL, Paris ([email protected])
(3) Atmospheric Chemistry Division, National Center for Atmospheric Research
1. Daily Biomass burning emissions inventory
Daily area burned maps were generated
by combining daily reports from fire
agencies (US National Interagency Fire
Center) with hot spots detected from
space by the MODIS instrument.
Area burned (ha)
Alaska-Yukon [165-125W]
4 106 ha burned
(ha)
North-Central Canada [125-90W]
MOPITT Level 2 measurements of CO [Deeter et al., 2003]: ~ 1 piece of
information on the vertical profile at extra-tropical latitudes, weighted towards the
middle and upper troposphere.
Daytime, column CO retrievals are compared to GEOS-Chem simulations, with
application of the MOPITT averaging kernels to the GEOS-Chem profiles [Heald et
al., 2004].
Average CO within regions
MOPITT Total CO, June–Aug. 2004
2. Simulation of atmospheric CO
Day since 20040601
5.6 106 hectares burned in North America during the summer of 2004, mainly
in Alaska-Yukon and Central Canada.
CO emissions for June – August 2004
CO emissions = area burned × fuel consumption(1) × CO emission factors(2)
(1) Amiro et al. [2001] for Canada and Alaska, Yevich et al. [2006] for contiguous
USA. The Amiro et al. estimates include contributions from surface burning and
crown fires, but are generally regarded as conservative.
(2) Duncan et al. [2003]; Goode et al. [2000]; Kajii et al. [2002]; Kasischke and
Bruhwiler [2003]; Yokelson et al. [1997].
 Current inventories do not account specifically for the burning of peat, although it
is expected to make a large contribution in boreal regions, especially under warm
and dry conditions [Zoltai et al., 1998]. We included this contribution based on
distributions of the areal fraction of peat for Canada [Hall et al., 2001] and on soil
drainage maps for Alaska [Harden et al., 2003], with fuel consumption from
Turetsky et al. [2002].
Contribution from peat
Potential fuel
consumption
Above ground biomass
The summer of 2004 was one of the strongest fire seasons on record for Alaska and western Canada. We
present a daily fire emission inventory for that season, including consideration of peat burning and highaltitude (buoyant) injection, and evaluate it in a global chemical transport model (GEOS-Chem CTM)
simulation of CO through comparison with ICARTT aircraft, and MOPITT satellite observations. Our
simulation shows that including emissions from peat burning improves the agreement between simulated
and observed CO. Model comparisons to observations are very sensitive to the altitude of injection of the
fire emissions in the CTM, highlighting the importance of considering pyro-convective events when
simulating fire influences or using atmospheric observations of trace gases as top-down constraints on fire
emissions.
1.1 106 ha burned
Day since 20040601
MOPITT satellite observations
Abstract.
Total area burned – June-August 2004
Area burned
(4) NASA Langley Research Center
(5) Aeronomy Laboratory, NOAA
Atmospheric CO distributions were simulated using the GEOS-Chem 3-D model (resolution: 2º×2.5º):
Simplified, “tagged” simulation: track CO sources from individual regions, using archived OH fields
from a non-linear O3-NOx-VOC-aerosols simulation.
Anthropogenic emissions for the USA: EPA NEI 1999 version 1, but with a 50% decrease in the on-road
mobile sources based on the ICARTT aircraft campaign [Hudman et al., 2006]; rest of the world (including
Canada): as described by Bey et al. [2001].
Biomass burning emissions for North America: this study; rest of the world: Logan and Yevich monthly
climatology distributed according to MODIS fire counts [L. Giglio, personal communication].
CO emission
factor
Total CO emissions, June – August 2004
Surface and crown fires
Contribution from peat burning
(Tg CO)
Including peat burning
Injection height
Indication that strong events occurred during the summer of 2004 from aircraft observations [e.g. De
Gouw et al., 2006; Kittaka et al., 2006] and the TOMS aerosol index.
We assume that 40% of biomass burning emissions for North America are injected in the modeldiagnosed boundary layer (typically up to 800 hPa), 55% in the free troposphere up to 400 hPa, and 5% in
the upper troposphere (400–200 hPa).
3. Atmospheric observations of CO as constraints
on biomass burning emissions
Comparison GEOS-Chem – ICARTT data
In situ observations
Measurements from the ICARTT campaign over
eastern North America and western North Atlantic:
• NASA DC-8 aircraft: [27N-53N;139W-36W]
• NOAA WP3-D: [28N-53N;59W-85W]
Data compared to GEOS-Chem simulations sampled
along the flight tracks
WP3-D
DC8
Data
Model, BB no peat burning
Model, BB with peat burning
Several factors could explain the observed discrepancy:
 Over-estimate of the U.S. anthropogenic CO
 Under-estimate of the CO background in the model
 Too low a contribution from injection of the fire
emissions into the upper troposphere and too high a
contribution in the boundary layer.
Our estimate of total CO fire emissions in North America during the summer 2004
is ~ 26 Tg CO, with ~ 30% from the burning of peat:
• Alaska-Yukon ~ 20.5 Tg CO (25% peat);
• North-Central Canada ~ 5 Tg CO (50% peat).
GEOS-Chem total CO (x MOPITT AK)
Surface and crown fires only
MOPITT avg. total CO
Model: BB no peat burning
Model: BB with peat burning
Model: BB with peat burning, linear increase
btw. June 1st and August 31st
Including peat burning in the inventory globally improves the agreement between
model and observations. However, this results in an overestimate of the model CO
over the Alaska-Yukon region in July.
The discrepancy between
model and observations
could be explained by many
factors of uncertainty in the
inventory: area, location of
the fires, fuel consumption.
Transport errors also impact
the
simulations.
In
particular, the comparisons
are very sensitive to the
injection height of the fire
emissions, as shown on a
transport event in mid-July.
MOPITT CO, 15-18/07/2004
GEOS-Chem: 100% BL
G-C: 40% BL + 55% FT + 5% UT G-C: 30% BL + 40% FT + 30%
UT
Individual contributions to the model CO
Contributions to the model CO plotted for:
– US anthropogenic sources (dots, grey line),
– biomass burning over Alaska and Canada, assuming:
• 100% of the emissions injected into the boundary
layer (black triangles, dotted line);
• 100% injected into the free troposphere (black
crosses, dashed line);
• 100% injected into the upper troposphere (black
stars, dashed-dotted line).
4. Conclusions and future directions
We estimate that the total CO emissions from biomass burning in North America
during the summer of 2004 was ~ 26 Tg CO, with 8 Tg CO from peat burning;
This total is consistent with the top-down estimate of 30 ± 5 Tg CO derived by
Pfister et al. [2005] from an inverse modeling of the MOPITT observations;
Including peat burning improves the agreement between model and observations;
Results are sensitive to the assumed distribution of injection heights of the fire
emissions;
We plan to constrain the magnitude of the emissions and the injection height
distribution in parallel using an inverse modeling approach with MOPITT data.