5:15 PM - 6:45 PM
[HCG26-P04] Challenges in Crop Residue Burning Detection from Satellite: Insights from Observations in Northwestern India
Keywords:crop residue burning, air pollution, fire detection, satellite observation
Ecological burning significantly impacts the physical and social systems of the Earth. Understanding the causes and effects of burning activities is crucial for developing effective measures to prevent future fires. Satellite observations of biomass burning include the identification of Burned Area (BA) and the observation of the thermal energy of Active Fire (AF), which is burning at high temperatures. While the sensor board on geostationary satellites like VAS and those on polar-orbiting satellites such as NOAA AVHRR have been instrumental since 1980, and newer sensors like MODIS and VIIRS provide advanced capabilities.
However, crop residue burning (CRB) detection remains challenging due to various factors. The spatial scale of Active Fires is small, making radiation signal detection difficult. Additionally, the short duration of each burning event, typically lasting 1–2 hours, increases the likelihood of missed detections. Cloud cover, smoke, and other atmospheric factors further impede observations. Burned Area detection faces challenges with post-burning changes of the land due to agricultural activities such as plowing and sowing. Besides, it also faces the difficulty in distinguishing between post-harvest period and burning because a lot of vegetation is lost by harvest.
As part of the Aakash project, we focused on analyzing fire detection data (AF) in the Punjab region of northwestern India. Numerous long-term analyses of AF in this region have been conducted, interpreting fire detection trends in relation to social conditions. However, the emissions estimated from satellite-observed AF remains considerably underestimated compared to actual emissions (Liu et al. 2020).
Our analysis of two VIIRS observations on board NOAA-20 (observation time at 12:40) and Suomi (13:30), with only a 50-minute time difference, revealed minimal agreement of the fires (Kurogi and Hayashida, JPGU2022). This suggests that the short duration of each fire contributes to a significant number of undetected fires in instantaneous observations by orbiting satellites. Despite this limitation, why is it believed that the long-term trend of fire detection remains representative as shown by many precedent studies? If the time of rice straw burning is concentrated at the time of MODIS and VIIRS observations, then we can expect a certain degree of representativeness. On the other hand, an "apparent decrease" would be observed if behavioral changes occur, such as farmers burning in the evening for fear of regulations. In the local media in Punjab, many cases of farmers intentionally burning in the evening have been reported, some of which were witnessed and photographed by the presenter.
In this presentation, we discuss the reliability of fire detection through a comparative analysis of NOAA-20 and Suomi observations, supplemented by insights from geostationary satellite data. We aim to shed light on the challenges in accurately capturing CRB trends and emphasize the importance of refining satellite-based detection methodologies for a comprehensive understanding of biomass burning dynamics.
However, crop residue burning (CRB) detection remains challenging due to various factors. The spatial scale of Active Fires is small, making radiation signal detection difficult. Additionally, the short duration of each burning event, typically lasting 1–2 hours, increases the likelihood of missed detections. Cloud cover, smoke, and other atmospheric factors further impede observations. Burned Area detection faces challenges with post-burning changes of the land due to agricultural activities such as plowing and sowing. Besides, it also faces the difficulty in distinguishing between post-harvest period and burning because a lot of vegetation is lost by harvest.
As part of the Aakash project, we focused on analyzing fire detection data (AF) in the Punjab region of northwestern India. Numerous long-term analyses of AF in this region have been conducted, interpreting fire detection trends in relation to social conditions. However, the emissions estimated from satellite-observed AF remains considerably underestimated compared to actual emissions (Liu et al. 2020).
Our analysis of two VIIRS observations on board NOAA-20 (observation time at 12:40) and Suomi (13:30), with only a 50-minute time difference, revealed minimal agreement of the fires (Kurogi and Hayashida, JPGU2022). This suggests that the short duration of each fire contributes to a significant number of undetected fires in instantaneous observations by orbiting satellites. Despite this limitation, why is it believed that the long-term trend of fire detection remains representative as shown by many precedent studies? If the time of rice straw burning is concentrated at the time of MODIS and VIIRS observations, then we can expect a certain degree of representativeness. On the other hand, an "apparent decrease" would be observed if behavioral changes occur, such as farmers burning in the evening for fear of regulations. In the local media in Punjab, many cases of farmers intentionally burning in the evening have been reported, some of which were witnessed and photographed by the presenter.
In this presentation, we discuss the reliability of fire detection through a comparative analysis of NOAA-20 and Suomi observations, supplemented by insights from geostationary satellite data. We aim to shed light on the challenges in accurately capturing CRB trends and emphasize the importance of refining satellite-based detection methodologies for a comprehensive understanding of biomass burning dynamics.