Statistical source analysis of recurring sulfur dioxide pollution events in a chemical industrial park

The whole process of a retrospective source analysis for a historical SO2 pollution scenario in a chemical in-dustrial park occurred during 2016-2017 is introduced. The case was a recurring SO2 pollution event at one of the monitoring sites (M1) in the park. A total of 68 samples for the SO2 pollution scenario were identified from the measurements at M1. The pollution scenario was identified as a single-source event based on pollution characteristics analysis. The main challenge was that the first-hand support data was only the measurements at one monitor while having many more sources. Statistical source analysis based on Kalman filter algorithm and CALPUFF model together with emission rate constraints and tracer species for source verification were incor-porated in the final analysis strategy. Results show that the source of SO2 pollution belonged to a catalytic cracking production unit. The cumulative matching rate of the real source was about 37%, and the average emission rate of SO2 was 47.94 +/- 73.27 g/s. It is evident that the comprehensive source analysis strategy established in this study was effective for improving certainty of source localization. The effectiveness of this strategy and the feasibility of simplified source simulation in the absence of detailed source data are discussed in detail.