Investigating the Relationship between Land Surface Temperature and Landscape Spatial Pattern by Using Regression Models and Landscape Metrics

Due to the negative effects of built up area temperature on energy consumption and human welfare, it is important to mitigate these negative effects by controlling the spatial pattern of green cover and its cooling effect. Therefore, in order to reveal the type and shape of relationship between landscape spatial patterns with land surface temperature (LST) in Rasht, the statistical modeling approach and landscape metrics were used. Land use- land cover and LST mapping were performed using Landsat sensor imagery classification and finally calculated landscape metrics of entered into the modeling process. Validation results showed that although all four linear, logarithmic, exponential and power regression models have good performance in predicting LST through land cover metrics, the highest performance was related to power model. In addition, this models indicate LST is strongly correlated with near distance to green patch and near distance to water patch metrics.