Urban heat island (UHI) refers to the occurrence of very high temperature patches in urban areas with respect to surrounding rural areas. The behaviour of Land Surface Temperature (LST) with different land covers varies for day and night time in a given urban area. Therefore, an attempt has been made for the comparative analysis of the day and night LST in two semi-arid cities in India i.e. Ahmedabad and Gandhinagar using two Landsat 8 images (one for day and one for night time) for each city. The relations of day and night LST with three independent variables such as Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI) and impervious surface fraction (ISF) were established using Landsat 8 data. Based on the relation of LST with indices, the combinations of indices were used for developing the models to improve the correlation of day and night LST with the land cover. The developed models improved the correlation of day and night LST with the combinations of indices as compared to the relation of LST with individual variables in semi-arid urban landscapes with reasonably good coefficient of determination (R) values of 0.63 in Ahmedabad and 0.79 in Gandhinagar for day time. The reasonably good values of R were found to be 0.81 and 0.75 for Ahmedabad and Gandhinagar, respectively, for night time. The developed model was also evaluated using 12 MODIS satellite images i.e. six images of each city for the summer, post-monsoon and winter seasons (one for day and one for night time in each season). The results of the developed models showed high capability to explain LST in heterogeneous semi-arid urban area for both day and night time in the summer, post-monsoon and winter seasons in a year. However, the relations established in this study can further be investigated for other months and years to understand its significance for other dates in semi-arid urban areas.

中文翻译：

---

利用不同季节卫星影像对两个半干旱城市昼夜地表温度进行对比分析


城市热岛（UHI）是指城市地区相对于周边农村地区出现非常高的温度斑块。在给定的城市地区，不同土地覆盖的地表温度 (LST) 在白天和夜间的表现各不相同。因此，我们尝试使用每个城市的两张 Landsat 8 图像（一张白天一张夜间）对印度两个半干旱城市艾哈迈达巴德和甘地讷格尔的昼夜地表温度进行比较分析。利用Landsat 8数据建立了昼夜地表温度与归一化植被指数（NDVI）、归一化植被指数（NDBI）和不透水表面分数（ISF）三个自变量的关系。基于LST与指数的关系，利用指数组合建立模型，以提高昼夜LST与土地覆盖的相关性。与半干旱城市景观中 LST 与单个变量的关系相比，所开发的模型改善了昼夜 LST 与指数组合的相关性，艾哈迈达巴德的决定系数 (R) 值为 0.63，甘地讷格尔的决定系数 (R) 值为 0.79。白天。研究发现，艾哈迈达巴德和甘地讷格尔夜间的 R 值相当不错，分别为 0.81 和 0.75。开发的模型还使用 12 个 MODIS 卫星图像进行了评估，即每个城市夏季、季风后和冬季的六幅图像（每个季节一幅用于白天，一幅用于夜间）。所开发模型的结果表明，能够很好地解释一年中夏季、季风后和冬季的白天和夜间异质半干旱城市地区的地表温度。 然而，本研究中建立的关系可以进一步研究其他月份和年份，以了解其对于半干旱城市地区其他日期的意义。