Developing Daily Biomass Burning Inventories from Satellite Observations and MOPITT Observations of CO during TRACE P Colette Heald Advisor: Daniel Jacob IDS Meeting: Duke University April 26, 2002
Satellite Observation of Fires and Biomass Burning Emission Inventories MOTIVATION Specific: Improve the forward and inverse GEOS-CHEM simulation of CO General: Improve Temporal Resolution of BB Emissions Use satellite observations to constrain emission features
Concept Annual BB CO Emission Budget (Logan & Yevich) Monthly BB CO Emission Budget (Martin & Duncan) Observed Daily Satellite FireCounts Daily FireCounts (after correct for coverage) Daily BB CO Emission Budget Constrain Total EmissionsAdd Temporal Variability
AVHRR Fire Data Why AVHRR? Need: global, daily coverage during Spring 2001 ATSR, MODIS and TRMM not suitable AVHRR Observations: 13:40 local cross-over time 1 km resolution at nadir World Fire Web: 22 ground stations Gridded product: 0.5°x0.5° #pixels on fire, #cloudy pixels, total #pixels observed WFW: 10 day composite
AVHRR Coverage Limitations Coverage limited by: 1. Polar Orbit 2. Ground Station Data Submission 3. Clouds WFW: 1 day coverage Percentage of days observed in Spring 2001
Accounting for Clouds Threshold Box > 90% cloudy pixels = No Information Defining the Fraction on Fire (FOF) fi = # pixels on fire ci = # cloudy pixels ti = # total pixels Average Cloud Cover during Spring 2001
Correlation Scales Local temporal correlation = persistence of fire patterns Local spatial correlation = cohesion of fires
Application of Correlation Scales to Daily Fire Data Correlation analysis supplements daily measurements with weighted information from neighbouring gridboxes, in either space or time.
Application of Fire Data to CO Emissions Magnitude of FOF used to partition monthly BB CO budget
Implementing in GEOS-CHEM Boundary Layer: Difference strongest over source regions Mid-troposphere: Difference strongest In outflow (W. Pacific) CO: Standard SimulationStandard Simulation - Daily Emissions Simulation
CO Sources: Integrating MOPITT and Aircraft MOPITT Data Evaluation Comparing to aircraft CO, GEOS-CHEM CO and fire activity Forward Model Evaluation of Aircraft Observations Comparing flight data with GEOS-CHEM fields Inverse Modeling of CO Sources using MOPITT and aircraft observations Characterize regional (Asian) emission sources via inversion of combined observational set
Background: TRACE-P and GEOS-CHEM TRACE-P Feb-Apr Characterize evolution and composition of outflow from Asia GEOS-CHEM tagged CO “tag” CO by emission type (biomass burning, fossil fuel, etc.) and source region using linear OH chemistry
GEOS-CHEM during TRACE-P GEOS-CHEM UNDERESTIMATES THE OBSERVED CO BY 5-10%
Attributing Source Type to Observations Multivariable fit to aircraft CO: PCE (C 2 Cl 4 ) = fossil fuel HCN = biomass burning, biofuel Background term = chemical production BLACK=OBSERVATIONS RED=FIT
Observation and Model Location of Source Influenced CO PCE=FFHCN=BB/BFTOTAL COMODEL MISSING A BB/BF SOURCE?
Observation and Model Location of Source Influenced CO PCE= FF HCN= BB/BF MISSING A BB/BF SOURCE? TOTAL CO MODEL
Use of MOPITT Integrated Analysis of global troposphere: MOPITT CO GEOS-CHEM Aircraft CO Observations CMDL CO Understanding Tropospheric Processes (CO) Emission Inventories
MOPITT Averaging Kernels Retrieved CO:
Comparing MOPITT and GEOS-CHEM MOPITT adjusted for bias
…the Next Day ( ) MOPITT adjusted for bias
…and the Next ( ) MOPITT adjusted for bias
TRACE-P Validation Profiles Courtesy: Louisa Emmons (NCAR) BLACK=AIRCRAFT RED=AIRCRAFTxAVG KERNELS BLUE=MOPITT V2 Retrieval: ~20% bias Preliminary V3 Retrieval: Better agreement EMBARGO’ED! PRELIMINARY FIGURES PROVIDED BY LOUISA EMMONS
CO Source Inversion: Aircraft + Satellite A posteriori CO emissions: CO Inversion from aircraft and satellite observations: Goal: Refine regional (Asian) sources Collaborate with those working on global inversions = a priori Initially: exploit TRACE-P aircraft data and MOPITT Associated error covariance:
Future Observations of CO: SCIAMACHY Interests: Compare and evaluate MOPITT and SCIAMACHY observations of CO Exploit SCIAMACHY observations in future CO source inversion