Evaluation and Separation between Urban Heat Island and Global Warming in Mashhad Region

Abstract

Introduction: In arid and semi-arid regions such as Iran, water is the most important limiting factor in economic development, and its management is of high importance. In recent years, due to irrigation expansion, low productivity in agricultural sector, and the rainfall shortage, water resources have been adversely affected in Iran. Undoubtedly, global warming in arid and semi-arid countries has increased the need for aquatic plants and the severity of drinking water shortages, making it more difficult to plan for limited resources. Studying the spatial and temporal changes of evapotranspiration is essential for the comprehensive planning of water management in Mashhad and providing an appropriate solution for optimal use of available water resources. However, spatiotemporal analysis of evapotranspiration regardless of the phenomenon of global warming and thermal island leads to unrealistic results. Therefore, the aim of this study was to address these shortcomings in previous studies in Mashhad. The specific objectives were: temporal analysis of evapotranspiration in the existing statistical period and estimation of annual evapotranspiration volume with respect to global warming, investigating the effect of global warming factors and thermal island on evapotranspiration and eventually water resources management in Mashhad. Materials and Methods: This study was carried out in Mashhad, city of Khorasan Razavi province with an area of 204 square kilometers, in northeastern Iran. Satellite imagery used for this research was a time series from Landsat 5 (TM sensor), Landsat 7 (ETM +) and Landsat 8 (OLI and TIRS sensors) from 1996 to 2016. The selected images for 2016 consisted of a time series of 13 images and a 16-day interval. After receiving satellite imagery, the performance of atmospheric corrections was evaluated based on FLAASH and TAC methods for reflective and thermal bands, respectively. The radiometric correction of images and reflection calculation of reflection was also conducted for bands 4 and 5 (values of ρ) and radiations of thermal bands10 and 11 (Lsen values) in the ILWIS software environment. Then, the temperature of the vegetation was calculated using different methods of determining the surface temperature (LST). Result and Discussion: The results showed that, on average, NDVI values in urban, mountainous and agricultural classes were 0.39, 0.37, and 0.4, respectively. However, the lowest and largest absolute value of NDVI were, respectively, 0.29 and 0.82, both of which are obtained in agricultural lands. The mean land surface temperature (LST) was 34.2 °C during days, which had a time and spatial variation between 17.9 to 49.4 °C in different regions. The highest and lowest mean LST was observed in urban and mountainous applications, respectively. Urban areas also had a significant difference in LST compared to other land uses due to the type of land cover in urban areas (mainly asphalt, stone, brick, cement, iron, etc.) and activities such as vehicle traffic, smoke and heat from factories and industries. The Split-Window (SW) method gave higher LST values compared with other methods. Then, the single-channel (SC), Improved Mono-Window (IMW) and single-window (MW) methods provided lower amounts of LST. The same trend was observed in almost all land use classes in the study area. It was also found that in urban areas, the strongest correlation between air temperature and LST was calculated by applying SC (R2 = 0.937). In mountainous regions, the highest correlation between air temperature and computed LST was found for the IMW (R2 = 0.951). Similarly, in the agro-rangeland areas, the highest correlation between air temperature and computed LST was obtained by IMW (R2 = 0.953). Conclusion: In the study area, the general trend of NDVI index was declining between 1996 and 2016. Reducing the percentage of vegetation cover in different sectors such as agriculture and rangeland, changing the type of vegetation (crop pattern) in agricultural sector and urban green spaces are the reasons for decreasing NDVI index in Mashhad region. The average LST was 34.2 °C in the days, which had a time and spatial variation between 17.9 to 49 °C in different regions. The maximum and minimum average LST was observed in urban and mountainous regions, respectively. The SW provided higher LST values compared to other methods. The SC, IMW and MW methods, however, provided lower LST values. The same trend was observed in almost all land use classes in the study area. It was also found that in urban areas, the highest correlation between air temperature and LST was found by using SC (R2=0.937). In mountainous regions, the strongest correlations between air temperature and LST was observed by using the Split Window Algorithm (SW) Improved Mono-Window (IMW) (R2=0.951). Similarly, in the agricultural and rangeland areas, the highest correlation between air temperature and LST was observed using the Split Window (SW) Improved Mono-Window (IMW) (R2 =0.953).