Predicting land use dynamics, surface temperature and urban thermal field variance index in mild cold climate urban area of Pakistan

Abstract

Rapid urbanization attributed to population growth is affecting the built environment's thermal and landscape dynamics. Using Landsat satellite datasets, this study investigated the complex interplay between urban Land Cover (LC) modification, fluctuation in Land Surface Temperature (LST) and severity of Urban Heat Island (UHI) from 1990 to 2020 in Peshawar City, Pakistan. Thermal bands were used to calculate LST and severity of UHI using the Urban Thermal Field Variance Index (UTFVI). Furthermore, through Cellular Automata (CA), Logistic Regression (LR), and Artificial Neural Network (ANN), future predictions on thermal characteristics associated with land use changes were made. The results showed that the urban areas expanded by ∼25 % from 1990 to 2020, while a ∼10 % decrease occurred in urban vegetation. The city is projected to expand by ∼45 % and ∼56 % in 2035 and 2050, respectively. Notably, the results also demonstrated that urban hotspots were found the warmest with the strongest UHI severity (∼34 °C), followed by the barren land (∼32 °C), and vegetation. The results further predicted an increase of LST (∼55 % and ∼82 %) and UTFVI (∼62 % and ∼83 %) in 2035 and 2050, respectively. These findings provide useful insights for policymakers and city planners to mitigate heat stress and create a sustainable urban environment through the development of effective urban land use policies and urban greening.

Keywords

• Artificial neural network
• Cellular automata
• Land cover modification
• Land surface temperature
• Urban heat island
• Urban thermal field variance index