Investigating the effects of local climate zones on land surface temperature using spectral indices via linear regression model: a seasonal study of Sapanca Lake.

The formation of urban heat islands is a widespread issue in cities. However, the impact of spectral indices on land surface temperature (LST) with various urban forms, climates, and functions has not been sufficiently examined. Currently, the prevalent method for analyzing complex urban areas is the classification of local climate zones (LCZs). In this study, we aim to explore the urban thermal environment by utilizing GIS-based spatial analyses and statistical methods. We also examine LCZs and the temporal-spatial changes of LSTs in Sapanca Lake and its surroundings. A comparative analysis was conducted on the relationship between the normalized difference vegetation index (NDVI), the modified normalized difference water index (MNDWI), and the normalized difference built-up index (NDBI) spectral indices in different LCZs and LST using a linear regression model. The results showed that all LCZs experienced a warming effect with an increase in NDBI, while they exhibited a cooling effect with the influence of NDVI and MNDWI. Notably, NDVI demonstrated a strong cooling effect in LCZ A (Dense trees) during the summer season, with an R coefficient of 0.73. Similarly, MNDWI had an R coefficient of 0.73 in LCZ A during spring. Values calculated as a result of regression are found as MAE:0.72 and MSE:0.75. These findings indicate the cooling effect of urban areas characterized by dense trees and water surfaces, highlighting their role in reducing LST. As a result, the research revealed the role of urban green systems and water surfaces in reducing the heat island effect, which is a problem, especially in urban centers. Overall, the study's results contribute to a better understanding of the thermal environmental characteristics in complex urban settings.