Air pollutant concentrations in urban areas exhibit strong spatial and temporal trends because of variations in land use and source distributions. While traffic is an important urban source of pollutants such as NO₂ and particulate matter (PM), other sources can also contribute to observed spatiotemporal patterns. PM exposures are further complicated because urban-scale variations in fine particulate matter mass (PM_2.5) concentrations can be decoupled from variations in PM physicochemical properties such as composition and the concentration of ultrafine particles (UFPs). Our goal was to quantify spatial gradients in both pollutant concentrations and PM characteristics over spatial scales ranging from hundreds of meters to several kilometers, along with sources influencing the observed spatial patterns. Continuous air pollutant concentrations were measured with a distributed network that consisted of a low-cost sensor package to measure pollutant gases (CO, NO₂, O₃, CO₂), optical measurements of particulate black carbon (BC) and PM_2.5, and UFP using condensation particle counters. The sampling sites were classified according to nearby land use for traffic, population, and restaurant density. Mean concentrations of BC and UFP vary by factors of approximately 4 and 3, respectively, across five typical urban sites in a 16 km² sampling domain. Differences in traffic volume and composition (e.g., diesel bus activity) can describe some, but not all of the observed spatial and temporal differences between sites. Specifically, two sites separated by ~ 500 m in a downtown central business district have significant differences in BC (factor of 2) and UFP (~ 30%), though similar PM_2.5 (< 10%), due to varying influences of diesel bus traffic and restaurant cooking emissions. The large spatial gradients have implications for data collection to inform spatial models such as land use regression (LUR). From the standpoint of site selection, the two downtown sites are nominally identical (e.g., both are high traffic, high-source activity sites) but have significant differences in measured pollutant concentrations and source impacts.

由于土地利用和源分布的变化，城市地区的空气污染物浓度表现出强烈的时空趋势。而流量是污染物如NO的一个重要的城市源₂和颗粒物质（PM），其他来源也可以向观察到的时空图案。PM 暴露更加复杂，因为细颗粒物质量（PM _2.5) 浓度可以与 PM 物理化学特性的变化分离，例如超细颗粒 (UFP) 的组成和浓度。我们的目标是量化数百米到几公里空间尺度上污染物浓度和 PM 特征的空间梯度，以及影响观察到的空间模式的源。连续空气污染物浓度使用分布式网络进行测量，该网络由用于测量污染物气体（CO、NO ₂、O ₃、CO ₂）的低成本传感器包、微粒黑碳 (BC) 和 PM _{2.5 的}光学测量组成和 UFP 使用冷凝粒子计数器。抽样地点根据附近的交通、人口和餐馆密度的土地用途进行分类。BC 和 UFP 的平均浓度在 16 km ²采样域中的五个典型城市站点中分别以大约 4 和 3 的系数变化。交通量和组成的差异（例如，柴油巴士活动）可以描述一些，但不是所有站点之间观察到的空间和时间差异。具体而言，在市中心中央商务区相距约 500 m 的两个地点在 BC（因子 2）和 UFP（约 30%）方面存在显着差异，尽管 PM _2.5相似(< 10%)，由于柴油巴士交通和餐厅烹饪排放的不同影响。大空间梯度对数据收集有影响，以便为土地利用回归 (LUR) 等空间模型提供信息。从站点选择的角度来看，两个市中心站点名义上是相同的（例如，都是高流量、高源活动站点），但在测量的污染物浓度和源影响方面存在显着差异。