Understanding physical processes and feedback mechanisms between the urban surface and the overlying atmosphere

The surface energy balance is different over cities compared to the surrounding rural areas due to differences in heat storage, evapo-transpiration and anthropogenic activities. Consequently, the urban boundary layer varies substantially in its structure from its rural counterpart. This has implications for the interactions between urban areas and meso-scale atmospheric circulations. Understanding these interactions is essential to improve the representation of urban surface process in urban canopy models (UCMs) and their coupling with numerical weather prediction (NWP) models. This can lead to better weather forecasts for urban areas, with consequences for human health and economic activities in cities.The skill of a UCM in quantifying the surface energy balance strongly depends on the accurate prescription of many surface parameters, including some that are notoriously difficult to estimate (albedo, thermal conductivity, heat capacities etc.). Therefore, offline optimisation approaches for these parameters are often used, but these optimizations exclude the effects of the urban surface-atmosphere feedback mechanisms.In Chapter 1 we identify the feedback mechanisms and their importance on the surface energy balance using two modelling setup, the off-line single-layer urban canopy model (SLUCM) and the SLUCM coupled to the single-column  version of the Weather Research and Forecasting (1D-WRF) model. The response of the modelled energy balanced to changes in uncertain surface parameters is investigated, during 2 summer days over London. The model responds differently to changes in surface parameter when coupled to the atmosphere. The turbulent heat flux shows up to 22% lower variability in the on-line setup, because near surface potential temperature gradient and atmospheric stability are altered. Consequently, the energy is directed toward the urban fabric changing the variability of the storage heat flux up to 50% compared to the off-line model. Moreover, entrainment of heat and moisture affect the near surface humidity altering the latent heat flux an effect that does not occur when the model is in offline mode.Once a UCM is coupled to the atmosphere, forcing provided by other parts of the atmospheric model is needed to compute the surface energy balance. This forcing comes in the form of incoming short- and long-wave radiation and advection of heat moisture and momentum. However, the atmospheric forcing can carry biases in the UCM, thus affecting the surface energy balance, near surface meteorology and the atmospheric stability with consequences for turbulent mixing and boundary-layer dynamics.In Chapter 3 the same case study as Chapter 2 is utilized to investigate the impact of the uncertainty in atmospheric forcing using a 1D WRF-SLUCM modelling setup. Uncertainties in the advection of potential temperature, aerosol optical depth and exchange coefficients of heat and momentum strongly affect model performance by altering either the radiative input at the surface (aerosols), or the near surface temperature and moisture gradients (advection and exchange coefficients). The resulting model responses are similar in magnitude to the changes induced by uncertainty in surface parameters. Moreover, a feedback mechanism between high daytime temperature and increased nocturnal radiative cooling was identified. This mechanism is triggered by an increase in nocturnal atmospheric stability that originates from changes in heat advection and the exchange coefficient of heat.Low-level jets (LLJs) are local maxima in the wind profile occurring above the nocturnal boundary layer, usually between 100-500m above ground. They are important phenomena affecting turbulent mixing, atmospheric stability and horizontal advection in the nocturnal boundary layer. These processes are important for the ’ventilation’ of heat and pollutants from urban areas, with consequences for human thermal comfort and street-level air quality. However, LLJ are also affected by turbulent mixing in the nocturnal boundary layer.In Chapter 4, the interactions between LLJ and the urban area using a 2-day case study between 14-16 May 2019 over London are investigated. Two Doppler Lidars and two numerical weather prediction models are used to identify differences in LLJ speed, height and fall-off between London and Chilbolton, a rural south west of London. The LLJs over London are elevated by 80 m compared to LLJs over Chilbolton, due to a combination of increased vertical mixing in the nocturnal UBL and orographic drag by the topography around London. Using a series of idealised experiments we find that urban and orographic effects contribute equally towards the reported differences over London. Moreover, we find that in areas with deeper UBL, the LLJ appears to be elevated and has lower speed. Finally, the presence of LLJ increases the shear induced turbulent kinetic energy in the UBL maintaining part of the vertical mixing many hours after the collapse of the daytime UBL.In Chapter 5 initial onset stages of the LLJ are investigated, to identify the effects of turbulent mixing and momentum advection in the formation of LLJ over urban areas. Two modelling setups are used, the 1D-WRF and the state-of-the-art PALM LES model in an idealised case study inspired by the meteorological situation during 15 May 2019 over London. The reported increase in turbulent mixing, induced by a large urban fraction, delays the onset of the LLJ up to two hours due to a later decoupling of the flow over the city. The height of the LLJ increases (up to 100m) while its speed is reduced (up to 2 m/s) with increasing urban fractions. Yet these effects are strongly dependent on the city size. The delay in the onset of the LLJ and its decreased wind speed over and downwind of the urban areas result in strong momentum advection from the already formed LLJ in the surroundings rural areas. The momentum advection contributes substantially to the formation of LLJs over urban areas, especially in the upwind part and its contributions positively scales with the size of the city.Overall, the findings of this thesis highlight the importance of accurately capturing the interactions between the urban surface, boundary layer and external meso-scale flows (i.e. advection processes, LLJ) for the quantification of the surface energy balance and turbulent exchange over urban areas