CHARACTERIZATION OF URBAN CANOPY PARAMETERS AND THEIR RELATIONSHIP WITH SPATIALLY VARIABLE URBAN CLIMATE INDICATORS

Abstract

Urban areas occupy 3% of ice free land area but are cumulatively responsible for emission of 75% of Green House Gases (GHGs), therefore contribute significantly in global warming. Developing regions are expected to house nearly 90% of future growth by 2050 with 35% of the growth concentrated in just three countries – India, China and Nigeria. Apart from the high pressure of urban growth in these regions, inadequate infrastructure and lack of implementation of planned urban development is going to exacerbate a number of climatic and environmental problems such as heat stress, air pollution, global warming, and increased use of artificial energy, storm water runoff, and overall environmental degradation in these regions. Increased frequency of natural disasters and extreme weather events in these regions calls for scientific understanding of impact of urban areas on climate and vice versa to ensure sustainable development and to mitigate climate change impacts. Lack of information on Urban Canopy Parameters (UCPs) is considered as one of the major reason behind limited urban climate research and modeling in developing regions. Availability of detailed information on UCPs is critical for urban climate studies as well as for implementation of recent Urban Canopy Models (UCMs) for scientific understanding of climatic phenomenon in these regions. Urban Canopy parameters define those characteristics of urban built form which has direct or indirect bearing on urban climate. Most of the studies in developed world have utilised 3D GIS database either developed from ground survey or remotely sensed data such as Aerial Photographs, Airborne LiDAR and high resolution InSAR data for retrieval of UCPs. However, nonexistence of 3D GIS database and non-availability of above RS datasets in developing regions necessitates to employ widely available alternative datasets for retrieval of UCPs. Hence, this study focuses on retrieval of UCPs by employing Very High Resolution Satellite (VHRS) optical stereo data in highly dense and complex urban environment of Delhi, India. Not many studies have explored the use of VHRS data for extraction of UCPs, however, the repeat availability, extensive coverage and low cost makes this data much suitable for generation of UCPs. It further dwells into the demonstration of retrieved UCPs for urban climatic applications in the study region. Although, the impact of 3D UCPs on urban climate is significant and it is considered as one of the main contributor to UHI phenomenon. However, a very few studies have analyzed relationship of 3D UCPs such as building height, frontal area index, floor area ratio and sky view factor with spatially variable climate indicators such as RS derived Land Surface Temperature Characterization of urban canopy parameters and their relationship with spatially variable urban climate indicators Kshama Gupta, 12910002, Ph.D. (2018-2019), IITR vi (LST) and it remained largely unexplored even for the regions having highly planned urban infrastructure Hence, the study also aims to analyze the relationship of generated UCPs especially 3D UCPs with spatially variable climate indicators. The Delhi Urban Agglomeration (UA) , which is the third largest UA of the world and largest UA of the India, has been selected as the study area due to its sprawled, highly heterogeneous and complex development characteristics of urban built form, high air pollution levels, high anthropogenic pressure and challenging composite climate. A novel step-by step methodology have been developed in this study to extract key UCPs such as building height, building surface fraction and wall area ratio from VHRS optical stereo data in complex urban environment of study region. Photogrammetric processing of VHRS optical stereo data (Pleiades1A/1B) have been carried out to obtain Digital Surface Model (DSM), Digital Terrain Model (DTM), normalized DSM (nDSM) and ortho images which have been further employed to retrieve key UCPs . The nDSM contains the height of all above ground objects that includes vegetation, building and all elevated objects. The key UCPs thus retrieved have been employed to compute other UCPs in complex urban environment of Delhi. The validation of key UCPs derived through VHR optical stereo images have shown good accuracy with ground measurements. The Mean error, RMSE and MAE for building heights has been found to be less than 1 m and Cumulative Random Error (CRE) ranged from 2.5% to 9.9% in high rise to low rise development respectively. The other key UCPs such as Land Use Land cover (Accuracy ~ 85%), Building Surface Fraction (BSF) (Accuracy ~84.27%) and SVF (RMSE- 0.046 and correlation-0.94) also displayed reasonable accuracy. It renders VHR optical stereo data a good choice for generation of UCPs especially in a highly heterogeneous urban built-up environment. Characterization of UCPs in the study region revealed highly dense, heterogeneous and sprawled character which has significant impact on urban climate. High Building Surface Fraction value (>0.6) in more than 35% of built-up area has shown high building density while distribution of Mean building height in study area revealed a highly sprawled character of the study region as nearly 96% of buildings falls in the height range of 3 -21m. Only 4% of buildings have height more than 21m and building with more than 30 m in height are very few and are mainly found in the peripheral region. Nearly 50% of built-up area has standard deviation of building height more than 2 which is indicative of highly heterogeneous and complex development in the study region. Characterization of urban canopy parameters and their relationship with spatially variable urban climate indicators Kshama Gupta, 12910002, Ph.D. (2018-2019), IITR vii Generated UCPs were further utilized to demonstrate few applications of UCPs such as Ventilation assessment and GIS based Local Climate Zone (LCZ) map which can be utilized to understand and characterize the urban climate phenomenon as well as for urban climate research and modeling. The pixel based classification methodology proposed by World Urban database and Access Portal Tool (WUDAPT), which have been widely applied to collect data on urban form and function by utilizing free RS datasets and GIS software, provided poor accuracy (overall accuracy -49.43% and kappa ~0.46) of LCZ classification in the complex urban environment of study area. However, GIS based LCZ maps generated from application of detailed UCPs displayed high accuracy of classification (overall Accuracy >85% and kappa ~0.86) not only for the entire classification but even for each LCZ class. Analysis of ventilation path maps showed merely 17% of built-up area as ventilated area. More than 45% area falls under weak or blocked ventilation and nearly 38% area is partially ventilated. The ventilation map of Delhi clearly brings out the lack of adequate ventilation in the study area which makes the city prone to severe air pollution in winter season and high UHI conditions. The unavailability of proper ventilation corridors retards the air flow within the built-up area resulted in very weak circulation and thereby restricts the continuous flow and exchange of fresh air in these regions. The spatially variable climate indicators (LST, temperature at 2m and wind speed at 10m) for the assessment of relationship with UCPs, ventilation assessment and LCZ map have been obtained from Landsat 8 and WRF simulations. The primary results of UCPs-LST relationship revealed strong correlation with 2D UCPS while 3D UCPs other than building height and surface roughness length did not show strong correlation. For 3D UCPs, complete temperature data cube in horizontal as well as vertical direction may help in analyzing the relationship. The key UCPs which exhibited strong relationship with LST were utilized to analyze the variability of Surface Urban Heat Island (SUHI) across all four seasons in a year. SUHI intensities was found to be maximum during winter season while lowest during post-monsoon season across all UCPs. The ventilation–LST relationship in study area revealed an interesting finding that well ventilated area has mean lower temperature and higher mean wind speed as compared to weak ventilation area. The difference in the mean temperature of both the classes were highest in monsoon season which raises serious concerns regarding the thermal comfort in built-up area. Although, the study area is mostly dry during the year, but, in monsoon season humidity level is more than 80% and slight increase in temperature leads to substantial increase in thermal stress. Analysis of GIS Characterization of urban canopy parameters and their relationship with spatially variable urban climate indicators Kshama Gupta, 12910002, Ph.D. (2018-2019), IITR viii based LCZ map with LST of all four seasons revealed that all the built-up LCZ classes other than open low rise and sparsely built LCZ exhibited more temperature than mean LST is all seasons. However, out of all built-up classes’ heavy industry, large low rise and compact low rise exhibited maximum deviations from mean LST for most of the seasons. The study recommends the use of VHRS optical stereo data for retrieval of 2D as well as 3D UCPs in complex and heterogeneous urban environment of developing regions. This is the next best suitable alternative available in the absence of Airborne LiDAR and aerial images data in the developing regions. Similarly, study also recommends use of retrieved UCPs for ventilation assessment and GIS based LCZ map in developing regions. Availability of information on UCPs has substantial potential for understanding the climate characteristics of urban areas in developing regions, carrying out urban climate research, improved modeling of urban climate phenomenon and climate oriented urban planning to mitigate climate change impacts and to ensure sustainable development.