URBAN HEAT ISLAND STUDY AND MODELLING OPTIMAL URBANIZATION PATTERN IN SELECTED INDIAN CITIES

Abstract

The rise of urbanization has led to many critical issues like increased air pollution, emission of more greenhouse gases, sudden climate change, and a rise in temperature of an urban area compared to surrounding rural regions know as urban heat islands (UHI). The conversion of natural surfaces into impervious surfaces of cities is increased manifold in the last decades, which is a major cause of UHI. The migration of people from rural areas to urban areas for better job opportunities, good livelihood, and health services has been a major cause of UHI. Cities occupuing just 3% of the Earth’s land surface has to accommodate more than half of the world’s urban population. It is projected that urban areas have to accommodate around 64% of the world's population by 2050. In India, around 34% of the population lives in urban areas, which is projected to increase to 50% by the end of 2050. So, it becomes crucial to understand the effect of UHI on urban climatic conditions for better sustainability in near future. In this present research, we examine the behavior of surface UHI over major Indian cities using both coarse and fine resolution data. The main novelty of this research is that all of the discussion is presented in terms of Koppen climatic zones. Before estimating SUHI, it is essential to know the air temperature and precipitation variation over major Indian cities situated over different climatic zones. This, the trend and magnitude of air temperature and precipitation over 139 major Indian cities with respect to different Koppen climatic zones are estimated using Climatic Research Unit (CRU) datasets of the last 115 years (1901–2015). The results show that the annual and seasonal temperature trend was significantly decreasing over the cities of the northwestern region, whereas an increasing trend in the southeastern cities of India. A significant relationship was observed between temperature and precipitation in the hot steppe (BSh), dry winter, humid subtropical (Cwa) and tropical wet and dry (Aw) climatic zone. The distribution of precipitation trends is highly heterogeneous and uneven as compared to temperature. The eastern part of India shows decreasing precipitation trend in comparison with the western part. After that, the methodology of SUHI estimation with long term coarse resolution datasets and their different associated driving variables were developed. Firstly, the entire process of the method was tested in a single Ahmedabad city as a case study before automating the entire process for all Indian cities. So, the seasonal and diurnal behavior of land surface temperature (LST) and surface urban heat island (SUHI) is examined from MODIS data over Ahmedabad city, Gujarat state (India), from 2003 to 2018. The investigation focuses on the SUHI variations due to warming or cooling trends of urban and rural areas, providing quantitative interpretations and understanding of the phenomenon. Different drivers like land cover maps, normalized differential vegetation index, surface albedo, evapotranspiration, urban population, and groundwater level were analyzed across the years to assess their impact on SUHI variations. Moreover, a field campaign was carried out in summer 2018 to measure LST in several rural and urban sites. The field study's primary purpose is to collect the dense point measurement of whether parameters and collection of LST over different LULC classes distributed throughout the city and the surrounding rural area. The rural zone exhibits a higher average LST during summer daytime than the urban area, resulting in a mean negative SUHI, typical of arid cities. In contrast, a slight positive SUHI (mean intensity of 0.4 °C) during winter daytime is present. An evident positive SUHI is found during summer (1.8 °C) and winter nighttime (3.2 °C). The negative SUHI intensity is due to the low vegetation present in the rural area, dominated by croplands turning into bare land surfaces during the pre-monsoon summer season. The field campaign also confirmed higher LST values in the rural area than in the urban area, with an average difference of about 5 °C. After the development of the complete methodology, the whole process is automated for 150 Indian cities. The objective is to quantify the diurnal, seasonal, and inter-annual variation of SUHI intensity (SUHII) over 150 major Indian cities situated over different climatic zones using MODIS data from 2003 to 2018. The results reveal urban cool islands occurrence over the hot desert, hot steppe, and tropical monsoon climatic zone during daytime in both summer (-0.25 to -0.17 °C) and winter (-0.33 to 0.17 °C) season. In contrast, nighttime SUHII shows clear evidence of positive urban heat island irrespective of climatic region and seasonal variation of 0.48 to 1 °C in summer and 0.46 to 1.32 °C in winter is seen. The Mann-Kendall and Sen's slope estimator tests are used to detect the trend of the SUHII during the study period, which suggests a higher percentage of cities showing an increasing trend of SUHII for urban heat islands than the cities of the urban cool island. Pearson's correlation and stepwise multiple linear regression model determine the possible SUHII controlling variable over different climatic zones. During the daytime, the SUHII's distribution is controlled by vegetation, evapotranspiration, and thermal inertia in the summer/winter season. It is linked tightly to built-up intensity, white sky albedo, and thermal inertia in both seasons at night. Overall, we found that the stepwise multiple linear regression model can explain the SUHII variance more in the daytime (>0.8) than in nighttime (>0.7, except for tropical cities) and more in understanding the SUHII behaviors for cool cities as compared to hot cities. Next is to see the detailed variation of SUHI and LST over selected cities of different climatic zones using high-resolution data like Landsat. In detail analysis, first, the relationship of LST and SUHI with the degree of impervious surface (IS) and green spaces (GS) along an urban-rural gradient in four rapidly growing Indian cities is presented using Landsat OLI/TIRS data. The results signify a strong negative correlation of LST with the IS for Ahmedabad, Jodhpur, and Nagpur, while a positive correlation is seen over Guwahati. The negative correlation is the manifestation of the urban cool island, pertaining to higher LST over rural areas. On the other hand, Guwahati is surrounded by green vegetation, which provides natural cooling and thus lowers the LST, resulting in positive SUHI. The density of GS is a significant contributor to SUHI in Guwahati city, whereas, in the other three cities, its impact is insignificant due to its presence in very little amount in rural surroundings. Lastly, it is essential to know the urban growth modelling and prediction of LST to know the future scenario of SUHI. For this purpose, this thesis's last analysis emphasizes the prediction of LULC, seasonal LST, and urban thermal field variance (UTFVI) over the Ahmedabad and Guwahati city, India using multi-date Landsat data. Artificial Neural Network (ANN) based Cellular Automata (CA) model is used to predict the LULC. In contrast, the XGB Regressor model predicts seasonal LST with input data for 2010, 2015, and 2020 to predict future scenarios for 2025 and 2030. The result of LULC suggests an expansion of urban built-up area at the expense of deduction in cropland and vegetation area in both the cities. The urban built-up area increased from 189.16 km2 in 1995 to 238.12 km2 in 2020 in Ahmedabad city, and 59.23 km2 in 1995 to 146.57 km2 in 2020 in Guwahati city. On the other hand, vegetation and cropland decrease their aerial extent from 139.07 km2 and 193.12 km2 in 1995 to 88.19 km2 and 185.98 km2 in 2020, respectively, in Ahmedabad city. Whereas a decreasing trend of about 179.13 km2 and 103.34 km2 in 1995 to 124.46 km2 and 54.87 km2 in 2020 in Guwahati city. This excessive urban growth in the cities will cause to face higher LST ranges of greater than 45 °C in summer and 35 °C in winter in Ahmedabad city, while 20 °C-<25 °C in summer and 15 °C-<20 °C in winter over Guwahati city. The concentration of higher LST zones is seen in the rural areas than urban areas, witnessing cool urban islands in Ahmedabad. Whereas a positive heat island is seen over the Guwahati city with warmer urban areas than rural areas. The predicted LST analysis suggests a dominant occurrence of none UTFVI zone in the city area and the strongest UTFVI zone in the surrounding rural area during both seasons in Ahmedabad city. While the dominance of the strongest UTFVI in the central urban area and none UTFVI zone in a rural area is evident in Guwahati city. In future scenarios, an increase in green space area and avoidance of non-impermeable surfaces is suggested to mitigate UTFVI. The research outcome would be very helpful for urban planners and policymakers while formulating urban heat island-related mitigation strategies in the near-future scenario.