Estimation of black carbon emissions from Siberian fires using satellite observations of absorption and extinction optical depths

I. B. Konovalov, Daria A. Lvova, Matthias Beekmann, Hiren Jethva, Eugene Mikhailov, Jean-Daniel Paris, B. D. Belan, В. С. Козлов, Philippe Ciais, Meinrat O. Andreae

TL;DR

This paper presents a novel method to estimate black carbon (BC) emissions from Siberian wildfires by combining satellite observations of aerosol optical properties (AAOD and AOD) with a chemistry transport model. The study reveals that existing emission inventories likely underestimate BC emissions from Siberian fires, with significant implications for Arctic climate change.

What Problem Does This Paper Address?

Black carbon (BC) from biomass burning (BB) significantly impacts the Earth's radiative budget, especially in the Arctic. However, accurately quantifying these emissions, particularly from remote regions like Siberia, has been challenging due to a lack of robust observational constraints in emission inventories. This paper aims to address this gap by developing a method to "top-down" estimate BC emissions using satellite data, providing a more accurate picture of their contribution to climate forcing.

Methods

The authors developed an inverse modeling approach to constrain BC emissions. Key steps include:

  • Empirical Parameterization: They established a relationship between the ratio of aerosol absorption optical depth (AAOD) and extinction optical depth (AOD) and the elemental carbon (EC) to organic carbon (OC) ratio in biomass burning aerosols. This parameterization was derived using ground-based AERONET data and laboratory experiments (FLAME-4).
  • CHIMERE Model Simulations: The CHIMERE chemistry transport model was used to simulate BC and OC concentrations and optical properties.
  • Optimization Procedure: Using satellite-derived AAOD and AOD, along with fire radiative power (FRP) measurements from MODIS, the model's BC and OC emission parameters were optimized. This involved an iterative process to minimize the differences between simulated and observed optical depths.
  • Validation: The optimized emissions were validated against independent datasets, including satellite data withheld from the optimization, in-situ measurements at the Zotino Tall Tower Observatory (ZOTTO), and aircraft measurements from the YAK-AEROSIB campaign.

Datasets

  • Satellite Data:
    • Ozone Monitoring Instrument (OMI) Aerosol Absorption Optical Depth (AAOD) retrievals (OMAERUV dataset).
  • Moderate Resolution Imaging Spectroradiometer (MODIS) Aerosol Optical Depth (AOD) retrievals (MYD04-L2 Collection 6).
  • MODIS Fire Radiative Power (FRP) data (MYD14/MOD14 Collection 6).
  • Ground-based Data:
    • AErosol Robotic NETwork (AERONET) data from Siberian sites (Tomsk and Yakutsk) for AAOD, AOD, and SSA.
  • In-situ EC and OC measurements from the Zotino Tall Tower Observatory (ZOTTO).
  • Aircraft Data:
    • YAK-AEROSIB campaign data for equivalent black carbon (eBC), PM2.5, and CO mixing ratios.
  • Ancillary Data:
    • USGS land use data.
  • Meteorological data from the WRF model.
  • Emission Inventories for Comparison:
    • Global Fire Emissions Database (GFED4.1s).
  • Fire Emission Inventory–Northern Eurasia (FEI-NE).

Key Insights and Ideas

  • Satellite Synergy: The study demonstrates the power of combining different satellite sensors (OMI for AAOD, MODIS for AOD and FRP) to constrain emissions in regions with limited ground-based networks.
  • AAOD as a BC Proxy: AAOD, particularly from OMI, proves to be a valuable tool for constraining BC emissions, even with the complexities of brown carbon absorption.
  • Underestimation in Inventories: The findings strongly suggest that existing emission inventories like GFED4 significantly underestimate BC emissions from Siberian wildfires, particularly from agricultural and grass fires.
  • Arctic Impact: A substantial portion (about a quarter) of the emitted BC from Siberian fires is transported to the Arctic, highlighting the critical role of these emissions in Arctic climate change.
  • Empirical Parameterization: The development of an empirical relationship between optical depths and EC/OC ratios offers a practical way to link satellite observations to aerosol composition.

Key Results

  • Total BC Emissions: Siberian fires emitted an estimated 405 ± 135 Gg of BC during May-September 2012.
  • Comparison with Inventories: This estimate is approximately twice that of GFED4 (198 Gg) and 1.5 times smaller than FEI-NE (614 Gg).
  • Monthly Variations: BC emissions were highest in July (139 ± 49 Gg) and lowest in September (20 ± 9 Gg).
  • Arctic Transport: Approximately 27% of the emitted BC was transported into the Arctic.
  • Model Performance: The optimized model simulations showed good agreement with satellite observations (correlation coefficients of ~0.79 for AAOD and ~0.84 for AOD) and in-situ measurements, significantly improving upon initial-guess simulations.

Limitations and Future Work

  • AAOD Retrieval Uncertainties: The authors acknowledge that OMI AAOD retrievals can have larger uncertainties than AOD retrievals, partly due to the influence of brown carbon and assumptions about aerosol layer height.
  • Model Representation of Aging: The CHIMERE model's representation of aerosol aging processes (e.g., SOA formation, wet deposition) might not fully capture the complexities of biomass burning aerosols, potentially affecting emission estimates. Sensitivity tests were conducted to assess these impacts.
  • Spatial Resolution: The coarse resolution of the model simulations compared to satellite pixels can introduce aggregation errors.
  • Data Selection Criteria: The choice of criteria for selecting satellite data for optimization can influence the results.
  • Future Work: The authors suggest the need for more data from remote sensing and in-situ measurements of aerosol optical properties and composition in northern Eurasia. They also highlight the importance of improving the representation of the vertical distribution of BB aerosols in models, potentially using data from instruments like CALIPSO. Extending the analysis to a multi-annual period is also recommended to better understand the role of Siberian fires in Arctic warming.

Summary updated

Abstract

Black carbon (BC) emissions from open biomass burning (BB) are known to have a considerable impact on the radiative budget of the atmosphere at both global and regional scales; however, these emissions are poorly constrained in models by atmospheric observations, especially in remote regions. Here, we investigate the feasibility of constraining BC emissions from BB using satellite observations of the aerosol absorption optical depth (AAOD) and the aerosol extinction optical depth (AOD) retrieved from OMI (Ozone Monitoring Instrument) and MODIS (Moderate Resolution Imaging Spectroradiometer) measurements, respectively. We consider the case of Siberian BB BC emissions, which have the strong potential to impact the Arctic climate system. Using aerosol remote sensing data collected at Siberian sites of the AErosol RObotic NETwork (AERONET) along with the results of the fourth Fire Lab at Missoula Experiment (FLAME-4), we establish an empirical parameterization relating the ratio of the elemental carbon (EC) and organic carbon (OC) contents in BB aerosol to the ratio of AAOD and AOD at the wavelengths of the satellite observations. Applying this parameterization to the BC and OC column amounts simulated using the CHIMERE chemistry transport model, we optimize the parameters of the BB emission model based on MODIS measurements of the fire radiative power (FRP); we then obtain top-down optimized estimates of the total monthly BB BC amounts emitted from intense Siberian fires that occurred from May to September 2012. The top-down estimates are compared to the corresponding values obtained using the Global Fire Emissions Database (GFED4) and the Fire Emission Inventory–northern Eurasia (FEI-NE). Our simulations using the optimized BB aerosol emissions are verified against AAOD and AOD data that were withheld from the estimation procedure. The simulations are further evaluated against in situ EC and OC measurements at the Zotino Tall Tower Observatory (ZOTTO) and also against aircraft aerosol measurement data collected in the framework of the Airborne Extensive Regional Observations in SIBeria (YAK-AEROSIB) experiments. We conclude that our BC and OC emission estimates, considered with their confidence intervals, are consistent with the ensemble of the measurement data analyzed in this study. Siberian fires are found to emit 0.41±0.14 Tg of BC over the whole 5-month period considered; this estimate is a factor of 2 larger and a factor of 1.5 smaller than the corresponding estimates based on the GFED4 (0.20 Tg) and FEI-NE (0.61 Tg) data, respectively. Our estimates of monthly BC emissions are also found to be larger than the BC amounts calculated using the GFED4 data and smaller than those calculated using the FEI-NE data for any of the 5 months. Particularly large positive differences of our monthly BC emission estimates with respect to the GFED4 data are found in May and September. This finding indicates that the GFED4 database is likely to strongly underestimate BC emissions from agricultural burns and grass fires in Siberia. All of these differences have important implications for climate change in the Arctic, as it is found that about a quarter of the huge BB BC mass emitted in Siberia during the fire season of 2012 was transported across the polar circle into the Arctic. Overall, the results of our analysis indicate that a combination of the available satellite observations of AAOD and AOD can provide the necessary constraints on BB BC emissions.
BibTeX
@inproceedings{konovalov2018estimation,
  author = {I. B. Konovalov and Daria A. Lvova and Matthias Beekmann and Hiren Jethva and Eugene Mikhailov and Jean-Daniel Paris and B. D. Belan and В. С. Козлов and Philippe Ciais and Meinrat O. Andreae},
  title = {{Estimation of black carbon emissions from Siberian fires using satellite observations of absorption and extinction optical depths}},
  booktitle = {Atmospheric},
  year = {2018},
  doi = {10.5194/acp-18-14889-2018},
}
Figure