Search for: All records

Abstract Urban areas are known to modify the spatial pattern of precipitation climatology. Existing observational evidence suggests that precipitation can be enhanced downwind of a city. Among the proposed mechanisms, the thermodynamic and aerodynamic processes in the urban lower atmosphere interact with the meteorological conditions and can play a key role in determining the resulting precipitation patterns. In addition, these processes are influenced by urban form, such as the impervious surface extent. This study aims to unravel how different urban forms impact the spatial patterns of precipitation climatology under different meteorological conditions. We use the Multi‐Radar Multi‐Sensor quantitative precipitation estimation data products and analyze the hourly precipitation maps for 27 selected cities across the continental United States from the years 2015–2021 summer months. Results show that about 80% of the studied cities exhibit a statistically significant downwind enhancement of precipitation. Additionally, we find that the precipitation pattern tends to be more spatially clustered in intensity under higher wind speed; the location of radial precipitation maxima is located closer to the city center under low background winds but shifts downwind under high wind conditions. The magnitude of downwind precipitation enhancement is highly dependent on wind directions and is positively correlated with the city size for the south, southwest, and west directions. This study presents observational evidence through a cross‐city analysis that the urban precipitation pattern can be influenced by the urban modification of atmospheric processes, providing insight into the mechanistic link between future urban land‐use change and hydroclimates. 

Abstract The predictability of passive scalar dispersion is of both theoretical interest and practical importance, for example for high‐resolution numerical weather prediction and air quality modeling. However, the implications for the numerical modeling of urban areas remain relatively unexplored. Using obstacle‐resolving large‐eddy simulations (LES), we conducted twin experiments, with and without a velocity perturbation, to investigate how the presence of urban roughness affects error growth in streamwise velocity ( u ) and passive scalar ( θ ) fields, as well as the differences between error evolutions in u and θ fields. The predictability limit is characterized using the signal‐to‐noise ratio (SNR) as a continuous metric to indicate when error reaches saturation. The presence of urban roughness decreases of the passive scalar by around 20% compared to cases without them. The error statistics of θ indicate that urban roughness‐induced flow structures and different scalar source locations affect the scalar dispersion and relative fluctuations, which subsequently dictate the evolution of the SNR. Analysis of the passive scalar error energy ( ϵ θ 2 ) budget indicates that the contributions from advective transport by the velocity and velocity error dominate. The error energy spectra of both u and θ exhibit a −5/3 slope in flat‐wall cases, but not in the presence of urban roughness, thereby highlighting the deviation from the assumption of locally isotropic turbulence. This study reveals that urban roughness can decrease the predictability of the passive scalar and destroy the similarity between the error statistics of the velocity and the passive scalar. 

Abstract Developing urban land surface models for modeling cities at high resolutions needs to better account for the city‐specific multi‐scale land surface heterogeneities at a reasonable computational cost. We propose using an encoder‐decoder convolutional neural network to develop a computationally efficient model for predicting the mean velocity field directly from urban geometries. The network is trained using the geometry‐resolving large eddy simulation results. Systematic testing on urban structures with increasing deviations from the training geometries shows the prediction error plateaus at 15%, compared to errors sharply increasing up to 35% in the null models. This is explained by the trained model successfully capturing the effects of pressure drag, especially for tall buildings. The prediction error of the aerodynamic drag coefficient is reduced by 32% compared with the default parameterization implemented in mesoscale modeling. This study highlights the potential of combining computational fluid dynamics modeling and machine learning to develop city‐specific parameterizations.