11:30 AM - 11:45 AM
[AAS11-15] Spatiotemporal Analysis of SO2 Distribution from Forest Fires Using Satellite Imagery
Keywords:forest fire, sulfur dioxide, satellite remote sensing
The extent of burned areas caused by forest fires has been increasing at an annual rate of 5.4% from 2001 to 2019, with significant atmospheric consequences. Forest fires emit pollutants such as PM2.5, which poses health risks, and NOx and SOx, which contribute to environmental degradation. Particularly, SO2 is a key precursor to acid rain and is emitted in amounts comparable to NO2. However, due to spectral overlap with other gases, research on post-fire SO2 distribution remains limited compared to NO2 studies.
Efforts to estimate SO2 concentrations from forest fires have used remote sensing techniques, including aircraft-based laser-induced fluorescence in Kansas, USA. However, such methods provide only localized measurements. More recently, a 2022 study utilizing TROPOMI satellite data estimated SO2 distribution from a forest fire in Turkey. While informative, it did not distinguish pre-existing SO2 from fire-emitted SO2, limiting its ability to assess fire-related emissions. As a result, uncertainties persist regarding total SO2 emissions, their temporal dynamics, and the influence of meteorological and vegetation factors on SO2 distribution.
This study utilizes the Sentinel-5P SO2 L2 product to estimate atmospheric SO2 concentrations. Employing Differential Optical Absorption Spectroscopy (DOAS), this product analyzes multiple SO2 absorption peaks, offering higher accuracy than conventional single-band methods. However, a single satellite image cannot distinguish pre-existing SO2 from fire-emitted SO2. To address this, we quantify SO2 emissions and their temporal variations by calculating differences between fire and non-fire conditions.
We analyzed seven Sentinel-5P images, including one from a 2024 forest fire in Valparaiso and six from 2020-2025 non-fire days. SO2 concentrations were derived from the Sentinel-5 Precursor Level 2 Sulphur Dioxide dataset, based on Band 3 (310-405 nm). The study focused on a 68 km x 17 km area surrounding the fire hotspot. Baseline SO2 concentrations were calculated from non-fire days, and differences were used to estimate fire-induced SO2 distribution.
Results showed a high SO2 concentration zone extending ~50 km north to south, with levels 0.002 mol/m2 above normal. Additionally, part of the SO2 plume traveled ~35 km north from the ignition point. By comparing SO2 distributions over three days post-fire, we analyzed temporal variations. Further, we visualized SO2 distributions from other forest fires, examining aerosol effects and their implications for air pollution and acid rain risk assessments.
This study was partially supported by a grant from the Telecommunication Advancement Foundation for the project ‘Development of an Education and Training Method for Solving SDG Challenges Using Both ICT and Hands-on Approaches’ (Shiga University of Medical Science, FY 2024), as well as by support from the NPO Super Scientist Program Plus.
Efforts to estimate SO2 concentrations from forest fires have used remote sensing techniques, including aircraft-based laser-induced fluorescence in Kansas, USA. However, such methods provide only localized measurements. More recently, a 2022 study utilizing TROPOMI satellite data estimated SO2 distribution from a forest fire in Turkey. While informative, it did not distinguish pre-existing SO2 from fire-emitted SO2, limiting its ability to assess fire-related emissions. As a result, uncertainties persist regarding total SO2 emissions, their temporal dynamics, and the influence of meteorological and vegetation factors on SO2 distribution.
This study utilizes the Sentinel-5P SO2 L2 product to estimate atmospheric SO2 concentrations. Employing Differential Optical Absorption Spectroscopy (DOAS), this product analyzes multiple SO2 absorption peaks, offering higher accuracy than conventional single-band methods. However, a single satellite image cannot distinguish pre-existing SO2 from fire-emitted SO2. To address this, we quantify SO2 emissions and their temporal variations by calculating differences between fire and non-fire conditions.
We analyzed seven Sentinel-5P images, including one from a 2024 forest fire in Valparaiso and six from 2020-2025 non-fire days. SO2 concentrations were derived from the Sentinel-5 Precursor Level 2 Sulphur Dioxide dataset, based on Band 3 (310-405 nm). The study focused on a 68 km x 17 km area surrounding the fire hotspot. Baseline SO2 concentrations were calculated from non-fire days, and differences were used to estimate fire-induced SO2 distribution.
Results showed a high SO2 concentration zone extending ~50 km north to south, with levels 0.002 mol/m2 above normal. Additionally, part of the SO2 plume traveled ~35 km north from the ignition point. By comparing SO2 distributions over three days post-fire, we analyzed temporal variations. Further, we visualized SO2 distributions from other forest fires, examining aerosol effects and their implications for air pollution and acid rain risk assessments.
This study was partially supported by a grant from the Telecommunication Advancement Foundation for the project ‘Development of an Education and Training Method for Solving SDG Challenges Using Both ICT and Hands-on Approaches’ (Shiga University of Medical Science, FY 2024), as well as by support from the NPO Super Scientist Program Plus.