Quantification of SO2 and CO2 emission rates from coal-fired power plants in the Korean peninsula via airborne measurements

Abstract

A reliable ensemble averaging method was developed to quantify sulfur dioxide (SO₂) and carbon dioxide (CO₂) emission rates from the Taean and Dangjin power plants in South Korea. This method integrated mass balance and Gaussian footprint approaches while accounting for individual uncertainties and deviations arising from distinct modeling assumptions and measurement variability. Eighteen representative spiral flights in 2022 (9 cases) and 2023 (9 cases) were conducted to evaluate emission rates and revealed several optimal conditions for achieving accurate quantification, including a small spiral radius with a fine vertical resolution under unstable atmospheric conditions. Validation of the estimated SO₂ emission rate revealed comparable correlation coefficients (R > 0.72) between the two methods and the real time automatic telemonitoring system (CleanSYS). The ensemble averaging method mitigated the sensitivity of the Gaussian footprint to the effects of meteorological conditions and high uncertainty in the mass balance, which resulted in an improved correlation of the estimated SO₂ emission rate with that measured by the CleanSYS (R > 0.78). When the same approach was applied, the CO₂ emission estimates from both methods showed a high correlation (R > 0.78) and confirmed the robustness of our ensemble averaging method. Although there was no significant difference between monthly electricity production in 2022 (October and November) and 2023 (May, October and November), the SO₂ emission rates decreased by 37 % and 29 % compared with the ensemble averaging method and CleanSYS, respectively; however, CO₂ emission rates increased by approximately 62 % at Taean and 83 % at Dangjin. This could be attributed to the use of carbon-intensive fuel sources, more intensive operations during research flight, and the desulfurization process, which aimed to reduce SO₂ emissions and release CO₂ as a byproduct. This study highlights the broad application of our ensemble averaging method for emission monitoring and regulatory compliance, particularly for CO₂, when real-time emission monitoring systems are absent.