Comparative Study of Source Inversion Under Multiple Atmospheric Pollutant Emission Scenarios

• Published in: Frontiers in environmental science Vol. 10
• Main Authors: Mao, Shushuai, Lang, Jianlei, Chen, Tian, Cheng, Shuiyuan, Hu, Feng
• Format: Journal Article
• Language: English
• Published: Frontiers Media S.A 17-03-2022
• Subjects: air pollution; atmospheric dispersion conditions; inversion modeling; multiple pollutant emission scenarios; small-scale regions; source parameter estimation
• ISSN: 2296-665X; 2296-665X
• DOI: 10.3389/fenvs.2022.857701

Source inversion is an effective approach for estimating air pollutant source parameters (e.g., source emission or source strength [Q
0
], source horizontal location [x
0
, y
0
], and release height [z
0
]) in industrial activities or accidents. Air pollution events in the real world generally correspond to complex application scenarios arising from unknown source parameters (i.e., Q
0
, [Q
0
, z
0
], [Q
0
, x
0
, y
0
], and [Q
0
, x
0
, y
0
, z
0
]) and atmospheric dispersion conditions. However, the source inversion characteristic law of these complex practical scenarios and the interaction mechanism between source location prior information and source strength inversion have not been revealed. In this study, the source inversion performance (accuracy and robustness) under the aforementioned scenarios was evaluated based on the Prairie Grass field experiments. Results indicated that the estimation accuracy of source strength was worse with an increase in the number of unknown source parameters with absolute relative deviations of 34.4, 46.0, 80.1, and 83.6% for a single parameter and double, triple, and quadruple parameters, respectively. Source strength inversion performance was obviously affected by location parameters; robustness was markedly reduced when source height was unknown, whereas accuracy was obviously reduced when source horizontal locations were unknown. Impacts of atmospheric conditions on different source parameters were distinct. Extreme atmospheric conditions (stability A and F) can obviously reduce the estimation accuracy of source strength for single and double parameter inversion scenarios, whereas unstable conditions (stability A, B, and C) can reduce the estimation accuracy of source strength for triple and quadruple parameter scenarios. Source inversion accuracy and robustness were generally poor under extremely stable conditions. This study can fill the knowledge gap in characteristic laws of source inversion under complex application scenarios and the interaction relationship between different unknown source parameters. The results of the influence law of location prior information on source strength inversion have important guiding significance to further improve the inversion accuracy of source strength in practical environmental managements.