The neural network predictive model for heat island intensity in Seoul

The heat island effect in cities becomes intensified due to rapid urbanization and industrialization. This urban heat island has negative effects such as increase in cooling energy use and impairment of urban air quality. This study aims to develop a predictive model for heat island in Seoul, Korea using neural networks. To create the neural network predictive model, air temperatures of 28 locations in Seoul for a year have been collected from automatic weather stations operated by the Korea Meteorological Administration. The neural network model was created and tested for estimating the urban heat island intensity according to albedo, building coverage, green area, building area, water area, road area, temperature, humidity, wind speed and direction, and precipitation. Finally, prediction results from the neural network model were compared with the measured data. The coefficients of correlation of the developed models range from 0.95 to 0.99. The analysis also indicates that the neural network model has better predictive performance compared to the multiple regression model. (C) 2015 Elsevier B.V. All rights reserved.