A Method for Assessing Background Concentrations near Sources of Strong CO2 Emissions

Abstract

In the quantification model of emission intensity of emission sources, the estimation of the background concentration of greenhouse gases near an emission source is an important problem. The traditional method of estimating the background concentration of greenhouse gases through statistical information often results in a certain deviation. In order to solve this problem, we propose an adaptive estimation method of CO2 background concentrations near emission sources in this work, which takes full advantage of robust local regression and a Gaussian mixture model to achieve accurate estimations of greenhouse gas background concentrations. It is proved by experiments that when the measurement error is 0.2 ppm, the background concentration estimation error is only 0.08 mg/m3, and even when the measurement error is 1.2 ppm, the background concentration estimation error is less than 0.4 mg/m3. The CO2 concentration measurement data all show a good background concentration assessment effect, and the accuracy of top-down carbon emission quantification based on actual measurements should be effectively improved in the future.