Application of the Inverse Dispersion Model to Estimate Sulfur dioxide Emission from the Offshore Industrial Park

Abstract

In the last 20 years, the Yunlin offshore industrial park has significantly contributed to the economic development of Taiwan. Its annual production value has reached almost 12 % of Taiwan’s GDP in 2012. However, the offshore industrial park is considered the major source of air pollution to nearby counties, especially, the emission of sulfate dioxide (SO2). Although studies have found that exposures to high level of some SO2 have caused adverse health effects on both human and ecosystem, it is a critical issue in estimating SO2 emissions. Nowadays emission estimation techniques are usually used emissions factors in calculation. Because the methodology considered totality of equipment activities based on statistical assumptions, it would encounter great uncertainty between these coefficients. The methodology of this study attempts to estimate SO2 emission of the Yunlin Offshore Industrial Park using an inverse atmospheric dispersion model which is applied to the combination of CALPUFF dispersion model adopted by the United States Environmental Protection Agency (U.S. EPA) as a preferred model and observation data of SO2 at monitoring site in Yunlin district. After that, comparing the solution with observation data collected for the time before industrial operation in 1999 and after 2010 by the Taiwanese Environmental Protection Administration (TW EPA). This study work in group on 4 kinds of dispersion coefficient around 7 monitoring sites. It shows well simulation performance on Tai-si, Mai-liao, Lun-bei Villages. In contract, the results in Er-lin Township is not well associated with the simulation and monitoring where might suffer from other pollution source. Estimated SO2 emission in the study area from July 2, 2012 to July 29, 2012 is around 1612-880 ton which is a little lower than Environmental Impact Assessment’s approve but a little bit higher than industrial park’s self-announced. Despite of that, the study result already have lots of challenge. There are so much uncertainty about land use data, terrain data, upper air data, parameter, even model itself, the standard deviation shows high uncertainty.