Trend Detection and Forecasting of LST in Tabriz City using the Non-parametric Mann-Kendall and NNAR

Aim: This study aims to analyze and forecast the LST during the summer season in Tabriz by 2030. Material & Method: The LST data were extracted from the MODIS satellite images for both day and night. These images were processed using the GEE platform and obtained for each summer season from 2002 to 2023 (minimum, average, and maximum). The Mann-Kendall test was used to assess the linear trend in the LST. An autoregressive neural network was employed to forecast the trend by 2030. Finding: During the daytime, the highest positive departure and the highest negative departure from the overall mean were recorded in 2006 (4.74°C) and 2023 (4.29°C), respectively. During the nighttime, the highest positive and negative departures from the overall mean were observed in 2006 (2.8°C) and 2009 (-2.77°C), respectively. Based on the trend analysis, the trends of the minimum (0.031°C), average (0.037°C), and maximum (0.065°C) LSTs during the daytime are significant. At night, only the trend of the minimum LST (0.034°C) is significant. Additionally, the rate of increase in the maximum daytime LST is higher than the minimum nighttime LST. The forecasting findings indicate that the model performed better during the nighttime than the daytime. Furthermore, the maximum daytime LST and the minimum nighttime LST in the summer of 2030 are expected to deviate by 1.12°C and 1.28°C from the overall mean of the period, respectively. Conclusion: The trends in the maximum daytime and minimum nighttime LSTs are increasing. Also, the upward trend will continue until 2030. Consequently, the thermal comfort in Tabriz is expected to decrease over time, leading to an increased demand for cooling energy. Innovation: This study provides insight into the trends of LST, which can be useful for urban planners in adopting mitigative and adaptive strategies to cope with climate change.