Investigating the potential of mobile air quality monitoring in urban areas in the development of real-time air pollution maps and applications to reduce population exposure

在城市地区，空气污染表现出很大的季节和空间变异性，这在健康影响研究中具有重要意义。因此，流行病学家经常使用不同的方法来量化人群暴露。其中包括个人暴露测量、来自固定地点监测站的数据、土地利用回归(LUR)技术和扩散模型。所有这些方法都有一定的局限性。个人接触测量通常仅限于测量特定样本个人在几天到几周的时间内的接触情况。随着样本量的增加，它们很快就变得繁重起来。从固定地点监测站获得的空气污染数据被认为是城市内部可变性的糟糕标志，与附近道路的空气污染水平相关性很差。已经为许多大都市地区开发了LUR模型，但它们预测准确的人口暴露模式的可靠性取决于已建立的监测网络的密度，因此需要进行具有广泛空间和时间复盖面的昂贵的运动。最后，为评估暴露而建立的扩散模型提出了许多挑战，特别是与排放投入的有效性有关的挑战。支持GPS的智能手机的使用迅速激增，开辟了新的研究领域，利用这些数据流来推断人口流动决策、景点和紧急疏散策略。此外，一些研究实验室和初创公司正在投资生产精度更低、读数上限更低、范围可与高端监测器相媲美的空气质量传感器。以人类为中心的空气污染监测的可能性为编制依赖时间的空气污染暴露图提供了无与伦比的机会，流行病学家可以使用该图来研究慢性健康影响，最重要的是，用于测试紧急医院就诊和依赖时间的空气污染事件之间的关联。拟议的研究将为制定旨在减少空气污染暴露的创新政策和指南以及信息系统提供必要的信息。具体地说，我将回答三个研究问题：城市环境中的哪些因素对近路空气质量的影响最大，以及我们如何相应地设计微环境，将与交通有关的空气污染降至最低？信息和通信技术在精细时空尺度监测环境空气污染方面的作用是什么？我们能否选择最好的路线，将空气污染降至最低，从而将研究成果转化为公众？**

In urban areas, air pollution exhibits substantial seasonal and spatial variability that is important to capture in health effects studies. As a result, epidemiologists often use different methods to quantify population exposure. These include personal exposure measurements, data from fixed-site monitoring stations, land-use regression (LUR) techniques, and dispersion models. All of these methods are associated with a number of limitations. Personal exposure measurements are normally limited to measuring the exposure of a specific sample of individuals over a time period ranging from a few days to a few weeks. They quickly become onerous as the sample size increases. Air pollution data obtained from fixed-site monitoring stations has been identified as a poor marker of intra-urban variability and correlate poorly with near-road air pollution levels. LUR models have been developed for many metropolitan areas however their reliability in predicting accurate population exposure patterns depends on the density of the established monitoring network thus entailing costly campaigns with extensive spatial and temporal coverage. Finally dispersion modelling for the purposes of exposure evaluation poses many challenges especially related to the validity of emission inputs. The rapid surge in the use of GPS-enabled smart-phones has opened-up new areas of research making use of these data streams to infer population mobility decisions, places of attraction, and emergency evacuation strategies. In addition, a number of research labs and start-up companies are investing in the production of air quality sensors with precisions, lower reading limits, and ranges comparable with high-end monitors. The possibility of human-centric air pollution sensing provides unparalleled opportunities for the development of time-dependent air pollution exposure maps which can be used by epidemiologists for studies of chronic health effects and most importantly for testing the association between emergency hospital visits and time-dependent air pollution episodes. The proposed research will provide essential information to develop innovative policies and guidelines and information systems that aim at reducing air pollution exposure. Specifically, I will address three research questions: What elements of the urban environment have the highest effect on near-road air quality and how can we accordingly design microenvironments which minimize traffic-related air pollution? What is the role of information and communication technologies in monitoring ambient air pollution at a fine spatial and temporal scale? Can we optimally select routes that minimize exposure to air pollution, thereby translating research results to the public at large? **