Predicting the vulnerability of oil spill in the Persian Gulf using artificial intelligence methods in the GIS

The Persian Gulf contains a significant part of the world's oil reserves. Generally, oil spill is one of the main pollutions in this region. Determining the degree of sensitivity of coastal areas to this type of pollution is the first step to control and prevent oil pollution. The aim of this research is to predict the vulnerability of oil spill in the Persian Gulf. Therefore, in this research the criteria of oil pipelines, oil platforms, shipping lanes, ports, heavy metals, water level fluctuations, rainfall, sea currents, air pollution and monsoons were used to determine the high-risk areas vulnerable to oil in the Persian Gulf. The innovation of the current research is to provide a new hybrid approach to determine the effective vulnerability criteria of the Persian Gulf. In this regard, the combination of geographic weighted regression (Gaussian and triple cubic kernels) and particle swarm optimization algorithm were used. The proposed hybrid method is suitable for spatial regression problems because it is compatible with two unique properties of spatial data, namely spatial autocorrelation and spatial non-stationary. The values of R2 and RMSE obtained from the GWR method with the triple cube kernel were 0.9971 and 0.2142, respectively, which indicates the high consistency of the triple cube kernel compared to the Gaussian kernel. Also, the obtained results showed that oil transfer pipes, oil platforms and the passage of oil tankers have a significant impact on the vulnerability of the Persian Gulf.