New approaches to measure neighborhood environmental quality and their application for policy and equity assessments in rapidly developing global cities

城市在城市化和工业化过程中受到一些最严重的污染和强烈热岛的影响，这使它们成为技术和政策干预的强大机会来源。低收入和中等收入国家（LMIC）的人口，特别是城市贫困人口，经常遭受不成比例的高水平污染和城市高温。然而，由于缺乏关于城市内部环境暴露时空变化的现有数据，证据驱动的技术和政策干预在低收入和中等收入国家城市受到限制。没有这些数据，城市就无法完全量化这个问题。如果不对问题进行量化，他们就无法有效地管理或解决问题。环境传感和数字图像分析方面令人兴奋的进展为高分辨率环境评估创造了创新、可靠和更便宜的机会。我提出的研究计划将利用这些新的数据收集和分析工具，提供关于四个全球城市的环境暴露的关键信息，这些城市在地理和发展上各不相同，我的团队在这些城市有积极的研究存在，即阿克拉（加纳）、北京（中国）、波哥大<e:1>（哥伦比亚）和达卡（孟加拉国）。我的项目有三个相互关联的举措：(1)收集重点城市的城市空气温度和相对湿度的时空野外测量数据；(2)开发一系列卷积神经网络模型，以(i)从街景和遥感图像中提取和分类自然和建筑环境以及交通特征，（ii）预测高分辨率的城市内空气温度和空气污染表面，以及(3)识别环境暴露和公平结果更好或更差的城市技术或政策情景的特征和表型。城市环境质量对生态和健康的影响由来已久，城市穷人往往受到的影响最大。中低收入国家的城市正在开始实施环境政策，以保护受影响的人口，但不幸的是，这些决定往往是在数据有限的情况下做出的。我的跨学科和全球研究项目有可能通过新颖的测量和方法填补这一知识空白，这些测量和方法可以提供关于这四个不同城市环境质量时空动态的关键信息，并可应用于这些城市的政策和公平评估。

Cities are affected by some of the highest levels of pollution and intense heat islands as they urbanize and industrialize, making them a powerful source of opportunity for technology and policy intervention. Populations in low- and middle-income countries (LMIC) and especially the urban poor often experience disproportionately high levels of exposure to pollution and urban heat. Yet evidence-driven technology and policy interventions in LMIC cities are limited by the lack of available data on intra-city spatial and temporal variation in environmental exposures. Without these data, cities cannot fully quantify the problem. Without quantifying the problem, they cannot effectively manage or solve it. Exciting advances in environmental sensing and digital image analysis have created innovative, reliable, and less-expensive opportunities for high-resolution environmental assessments. My proposed research program will leverage these new data collection and analysis tools to provide critical information on environmental exposures in four global cities which are diverse in geography and development and where my group has an active research presence, namely: Accra (Ghana), Beijing (China), Bogotá (Colombia), and Dhaka (Bangladesh). My program has three interlinking initiatives: (1) collect spatial-temporal field-based measurements of urban air temperature and relative humidity in the focus cities; (2) develop a series of convolutional neural network models that (i) extract and classify natural and built environments and transportation features from street view and remote sensing images, and (ii) predict high-resolution intra-city surfaces of air temperature and air pollution, and (3) identify features and phenotypes of urban technology or policy scenarios with better versus worse environmental exposure and equity outcomes. The quality of the urban environment has long-established ecological and health impacts, with the urban poor often being most affected. LMIC cities are starting to implement environmental policies to protect affected populations, but unfortunately these decisions are often made limited data to inform them. My interdisciplinary and global research program has the potential to fill this knowledge gap through novel measurements and methods that can provide critical information on the spatial-temporal dynamics of environmental quality in these four diverse cities, and that can be applied to policy and equity assessments in these cities.