Haze Influencing Factors: A Data Envelopment Analysis Approach

Abstract

This paper investigates the meteorological factors and human activities that influence PM_2.5 pollution by employing the data envelopment analysis (DEA) approach to a chance constrained stochastic optimization problem. This approach has the two advantages of admitting random input and output, and allowing the evaluation unit to exceed the front edge under the given probability constraint. Furthermore, by utilizing the meteorological observation data incorporated with the economic and social data for Jiangsu Province, the chance constrained stochastic DEA model was solved to explore the relationship between the meteorological elements and human activities and PM_2.5 pollution. The results are summarized by the following: (1) Among all five primary indexes, social progress, energy use and transportation are the most significant for PM_2.5 pollution. (2) Among our selected 14 secondary indexes, coal consumption, population density and civil car ownership account for a major portion of PM_2.5 pollution. (3) Human activities are the main factor producing PM_2.5 pollution. While some meteorological elements generate PM_2.5 pollution, some act as influencing factors on the migration of PM_2.5 pollution. These findings can provide a reference for the government to formulate appropriate policies to reduce PM_2.5 emissions and for the communities to develop effective strategies to eliminate PM_2.5 pollution.