Fast and Accurate Urban Atmospheric Models

城市约占地球表面的2%，容纳了世界50%以上的人口。理解和减轻城市日益增长的环境影响需要设计和预测工具，如城市大气模型(UAMS)。UAMS预测城市气候(气温、风、降水和湿度)和污染扩散(空气污染物在城市环境中的分布)。UAMS有助于理解、发展和监管更可持续的城市，以控制城市气温、湿度和风，并改善城市空间内的空气质量。 拟议的跨学科研究计划试图基于降阶和人工智能(AI)方法开发新颖、快速和准确的可操作UAMS，以考虑现实的城市环境。现实的城市环境目前在业务模型中没有得到很好的表示，或者计算成本太高而无法建模。UAMS将考虑1)城市形态的复杂性，例如建筑物、景观、道路和其他城市组成部分中不同的几何和长度尺度，2)城市表面的不同加热或稳定/不稳定边界层的影响等热效应，以及3)车辆和建筑物实际人为排放的污染扩散。 科学方法包括开发实验和数值数据库，以制定和验证UAMS。实验数据库将从新的气球平台(加拿大第一个此类平台)、风洞、水道、现场和遥感测量中开发。数值数据库将从高时空分辨率模型发展起来，用于使用计算流体动力学、能量平衡、弥散和大气模型对现实城市环境进行参数研究。这些数据库将用于开发基于一维垂直扩散(降阶)和人工神经网络(AI)模型的新的UAMS。 这项研究计划有助于了解影响城市大气的复杂物理知识，以及开发预测大气的模型的能力。它的结果是UAMS可以作为独立的设计和预测工具商业化，供各种建筑公司、工程咨询公司和城市规划者使用。它将通过收集、生成、处理和分析大数据，以及通过建立和验证城市大气的简单或复杂模型，为HQP提供数据科学培训。它通过改进加拿大或国际上的业务预报和空气质量模型来提供环境效益，计算速度和精度都很高。它通过减少能源消耗和改善环境条件，为子孙后代提供更可持续的城市发展的社会效益。

Cities occupy approximately 2% of the earth's surface and accommodate more than 50% of the world's population. Understanding and mitigating the growing environmental impact of cities require design and prediction tools, such as the Urban Atmospheric Models (UAMs). UAMs predict the urban climate (air temperature, wind, precipitation, and humidity) and pollution dispersion (distribution of air pollutants in the urban environment). UAMs help understand, develop, and regulate more sustainable cities toward the control of the urban air temperature, moisture, and winds and the improvement of air quality within urban spaces. The proposed interdisciplinary research program attempts to develop novel, fast, and accurate operational UAMs based on reduced-order and Artificial Intelligence (AI) approaches to account for the realistic urban environment. The realistic urban environment is currently not well represented in operational models or is otherwise too computationally expensive to model. The UAMs will account for 1) complexity of the urban morphology such as variety of geometries and length scales in buildings, landscapes, roads, and other urban components, 2) thermal effects such as differential heating on urban surfaces or the influence of stable/unstable boundary layers, and 3) pollution dispersion from realistic anthropogenic emissions by vehicles and buildings. The scientific approach includes development of experimental and numerical databases to formulate and validate the UAMs. Experimental databases will be developed from a novel blimp (balloon) platform (first one of its kind in Canada), wind tunnel, water channel, field, and remote sensing measurements. Numerical databases will be developed from high spatiotemporal resolution models for parametric study of realistic urban environments using computational fluid dynamics, energy balance, dispersion, and atmospheric models. These databases will be used to develop the new UAMs based on a 1D vertical diffusion (reduced-order) and Artificial Neural Network (AI) models. This research program contributes to knowledge of complex physics influencing the urban atmosphere and the ability to develop models to predict it. It results in UAMs that can be commercialized as stand-alone design and prediction tools for use by various architectural firms, engineering consulting firms, and urban planners. It will train HQP with data science by collection, generation, processing, and analysis of big data, and by building and validating simple or complex models for the urban atmosphere. It provides environmental benefits by improving operational forecast and air quality models, both in Canada or internationally, with computational speed and accuracy. It provides social benefits toward more sustainable urban development for the future generations by reducing their energy consumption and improving their environmental conditions.