Intra-urban morphological heterogeneity and its impact on the micro-climatic variations in the Kolkata Metropolitan region of India

Kolkata, one of the oldest and largest urban centres in Asia echoes all the major characteristics of cities from developing countries experiencing rapid urbanization and unplanned development. This study focuses on understanding the impact of intra-urban variations within a morphologically complex metropolis and its capability to modify the enveloping atmospheric conditions at the meso to micro-scale. The Weather Research and Forecasting (WRF) model was configured using a 3-tier nested domain to conduct high-resolution simulations incorporating improved land surface and urban parameterization and appropriate physical parameterization. Local Climate Zone (LCZ) map representing the land use land cover (LULC) was prepared for the innermost domain covering the city and surroundings using Planet-Scope datasets for 2019, adopting Artificial Neural Network (ANN) approach and validated using ground information. This was then integrated into the model with redefined values of specific urban parameters for a better representation of the city’s morphology. It was observed that LCZ-2 (Compact Mid-rise) and LCZ-3 (Compact Low-rise) cover almost the entire core city and around 70 % of the total built-up extent of the study area, which also consists of LCZ-5 (Open Mid-Rise) and LCZ-6 (Open Low-Rise) regions. Thus, the model was calibrated according to the surface and atmospheric conditions of the region and its performance was evaluated in comparison with ground observations. Simulations were conducted at hourly intervals for a 10-day period (August 2019) during the peak summers coinciding with the south-west monsoon period receiving heavy rain spells, to analyse the impact of heterogenous urban form on the micro-climatic variations within the city as well as its surroundings. The modelled results obtained for 2m air temperature (T_air), surface temperature (T_skin), 10m Wind Speed (WS) and Rainfall (RF) indicated a significant influence of the different LCZ classes and their spatial variations over the city. The average daytime T_air and T_skin values in the LCZ-2 is about 1℃ and 1.5℃ higher than LCZ-3, which is again 0.7℃ and 1.5℃ higher compared to the other urban LULC classes where the internal variation is relatively less. Further, the average temperature differences between the compact and open built-up structures increase significantly during night (2℃ – 3.5℃), further increasing when compared to the peri-urban (around 5℃). The inter-urban heterogeneity however, has a reverse effect on the average WS even during the typical monsoon period, with lowest speed observed in the compact core (2ms^-1 – 5ms^-1) due to highest surface drag (d), increasing (3ms^-1 – 7ms^-1) along LCZ-5 and LCZ-6 with reduction in d; which further increases substantially in the peri-urban areas (10ms^-1 – 15ms^-1) with lowest value of d. The variations in total RF received from the complex towards peripheral urban also depicts a similar pattern, as average RF intensity is the lowest within LCZ-2 and LCZ-3 (7mm/hr – 15 mm/hr), moderate in LCZ-5 and LCZ-6 (10mm/hr – 20mm/hr) and highest along the peri-urban areas (15mm/hr – 35 mm/hr). Thus, urban structural and morphological complexity can have a substantial effect on the local-scale climatic variations even along small horizontal distances within a city.