Addressing the impact of land use land cover changes on land surface temperature using machine learning algorithms

Area of Science:

• Environmental Science
• Urban Planning
• Remote Sensing

Background:

• Rapid urbanization in Kabul province over 25 years has driven significant land use and land cover (LULC) changes.
• Understanding these LULC dynamics is crucial for assessing their impact on the urban environment, particularly land surface temperature (LST).

Purpose of the Study:

• To assess the impact of LULC changes on LST in Kabul province from 1998 to 2022.
• To predict future LULC and LST patterns using the Cellular Automata-Logistic Regression (CA-LR) model for 2034 and 2046.

Main Methods:

• Support Vector Machine (SVM) algorithm applied to Landsat satellite images (1998-2022) to classify LULC into four categories.
• LST assessment using Landsat thermal band data.
• Cellular Automata-Logistic Regression (CA-LR) model for future LULC and LST prediction.

Main Results:

• Built-up areas increased by 9.37%, while bare soil and vegetation cover decreased by 7.20% and 2.35% respectively (1998-2022).
• Built-up areas exhibited the highest mean LST, followed by bare soil and vegetation.
• Future projections indicate a continued increase in built-up areas and a significant rise in areas with LST ≥ 32°C.

Conclusions:

• A direct correlation exists between increasing built-up areas, decreasing vegetation cover, and rising land surface temperatures in Kabul.
• Urbanization is a primary driver of LST increase, necessitating urban planning strategies to mitigate heat effects.