Modeling the relationship between noise pollution and landscape metrics of urban structures and green covers (case study: Esfahan City)

Population growth and the increase in urban migration in the past decades have led to an increase in population density and size of major cities. Unfortunately, this kind of pollution has mostly gone under-noticed. The purpose of this research was to model the correlation between noise pollution level and landscape metrics of urban structures and vegetation. To do so, 67 stations were selected in different parts of Isfahan and noise parameters were measured at peak traffic hours (16 to 19) during the fall season. Sampling stations were located through a systematic-random method based on the amount of construction, green spaces and structural diversity. There were 27 types of landscapes and three stations were randomly selected in each. In most stations, the noise level was above the permitted level(Residential 45-55, Residential-Commercial 50-60). The advanced regression method of random forest was used for the analysis. Through this method, the most effective metrics identified in different buffers were IJI index, FRAC_MN index, CLUMPY index, CONTIG_MN index, SHAPE_MN index, ENN_MN index. Also, checking of the first six metrics in each of the buffers and land uses showed the importance of the metrics is different. Identification of important metrics in each buffer and land use helps better design urban blocks and their arrangement.