Quantitative imaging of carbon dioxide plumes using a ground-based shortwave infrared spectral camera

• Verfasst von: Knapp, Marvin [VerfasserIn]
• Verfasst von: Kleinschek, Ralph [VerfasserIn]
• Verfasst von: Vardag, Sanam Noreen [VerfasserIn]
• Verfasst von: Külheim, Felix [VerfasserIn]
• Verfasst von: Haveresch, Helge [VerfasserIn]
• Verfasst von: Sindram, Moritz [VerfasserIn]
• Verfasst von: Siegel, Tim [VerfasserIn]
• Verfasst von: Burger, Bruno [VerfasserIn]
• Verfasst von: Butz, André [VerfasserIn]
• Titel: Quantitative imaging of carbon dioxide plumes using a ground-based shortwave infrared spectral camera
• Verf.angabe: Marvin Knapp, Ralph Kleinschek, Sanam N. Vardag, Felix Külheim, Helge Haveresch, Moritz Sindram, Tim Siegel, Bruno Burger, and André Butz
• Jahr: 2024
• Umfang: 19 S.
• Illustrationen: Illustrationen, Diagramme
• Fussnoten: Online veröffentlicht: 18. April 2024 ; Gesehen am 06.09.2024
• Titel Quelle: Enthalten in: Atmospheric measurement techniques
• Ort Quelle: Katlenburg-Lindau : Copernicus, 2008
• Jahr Quelle: 2024
• Band/Heft Quelle: 17(2024), 8, Seite 2257-2275
• ISSN Quelle: 1867-8548
• DOI: doi:10.5194/amt-17-2257-2024
• Datenträger: Online-Ressource
• Sprache: eng
• K10plus-PPN: 1902118502

Abstract

We present the first results of a ground-based imaging experiment using a shortwave infrared spectral camera to quantify carbon dioxide (CO2) emissions from a coal-fired power plant in Mannheim, Germany. The power plant emits more than 4.9 Mt CO2 yr−1 and is a validation opportunity for the emission estimation technique. The camera is a hyperspectral imaging spectrometer that covers the spectral range from 900 to 2500 nm with a spectral resolution of 7 nm. We identify CO2 enhancements from hourly averaged images using an iterative matched filter retrieval using the 2000 nm absorption band of CO2. We present 11 plume images from 5 d in 2021 and 2022 covering a variety of ambient conditions. We design a forward model based on a three-dimensional, bent-over Gaussian plume rise simulation and compare our observed emission plumes with the forward model. The model depends on the parameters ambient wind velocity, wind direction, plume dispersion, and emission rate. We retrieve the emission rate by minimizing the least-squares difference between the measured and the simulated images. We find an overall reasonable agreement between the retrieved and expected emissions for power plant emission rates between 223 and 587 t CO2 h−1. The retrieved emissions average 84 % of the expected emissions and have a mean relative uncertainty of 24 %. The technique works at wind speeds down to 1.4 m s−1 and can follow diurnal emission dynamics. We also include observations with unfavorable ambient conditions, such as background heterogeneity and acute observation angles. These conditions are shown to produce considerable biases in the retrieved emission rates, yet they can be filtered out reliably in most cases. Thus, this emission estimation technique is a promising tool for independently verifying reported emissions from large point sources and provides complementary information to existing monitoring techniques.