Understanding and predicting the complex wind, temperature, and concentration patterns over built surfaces (Progam: Earth Sciences/Earth Materials and Processes)
Abstract
By 2050, 68% of the world population will reside in cities; the fate of the human race is thus more than ever tied to the built environment. Walking in between the buildings of our city or town, we have all felt the rapid variation of wind speed, thermal comfort, and sometime air quality over very short distances of a few meters. The rapid variation of these important environmental attributes is a direct result of the geometry of the built surface, but our current ability to link how urban form influence airflow and transport in cities still lags our practical needs substantially. For example, while we can evaluate wind gusts at pedestrian levels for a given location, we still cannot predict what geometric features of the built surface enhance them and what gust probabilities are higher up in the urban canopy layer, where drones are already navigating. Against this backdrop, the recent advances in computer simulation of flow and transport in urban terrain, urban digital surface mapping, and machine learning offer a promising potential. In this proposal, we aim to combine all three of these innovations to bridge our understanding and prediction gaps in urban terrain. Specifically, we are interested in: extreme high wind, extreme low wind with poor ventilation, and extreme heat. Our driving hypothesis is that the tight links between the urban earth surface geometry and the urban canopy air can be exploited to develop a simplified conceptual understanding and fast models to predict extremes in cities. To test this hypothesis, we will answer 3 overarching questions: (i) What are the statistics of wind gusts inside the building layer, and can Ògust proneÓ zones be predicted based on the mean flow above and the geometry of the urban canopy? (ii) What are the physical attributes of the urban canopy that explain regions with low winds, and can these Òstagnation proneÓ zones with high pollution and temperature be predicted from geometric information? (iii) How do canyon wind and heat transfer interact to determine temperature variability across a city, and what are the physical surface attributes that influence this interaction and explain Òheat proneÓ zones? To answer these questions we will: (i) collect and analyze 3D digital surface maps for 6 US and 6 world cities (selected to represent broad characteristics) (ii) perform a suite of computer simulations of airflow and pollutant and heat transport for each city, (iii) analyze the simulations to understand the links between the urban form and environmental quality using statistical, visualization, and machine learning tools, and (iv) train machine learning algorithms to detect urban locations and configurations that are more likely to generate extremely high or low wind, or elevated temperature and concentrations. The results would be of significant value to the Army. The fast decision models we will develop can help predict (i) extreme winds that can reduce the ability of soldiers to conduct sensitive tasks and perturb the operation of military drones, (ii) zones of extreme temperatures that influence soldier well-being and effectiveness, and (iii) stagnation zones that can contain hazardous chemicals. These models can be readily adapted to the Army needs since they are designed to make probabilistic estimate of risk categories (ÔgoodÕ, ÔaverageÕ, ÔbadÕ) that can be interpreted rapidly. Our findings will also have many civilian broader impacts, and will be of great value to urban planners, architects, and city environmental managers or engineers. Even more broadly, our work can help improve urban weather forecasting and engineering models for flow and heat exchange over rough surfaces.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 09, 2020
- Source ID
- W911NF2010216
Entities
People
- Elie Bou-Zeid
Organizations
- Army Contracting Command
- Princeton University
- United States Army