Machine learning derived daily PM2.5 concentration estimates from by County, ZIP code, and census tract in 11 western states 2008-2018

We created daily concentration estimates for fine particulate matter (PM_2.5) at the centroids of each county, ZIP code, and census tract across the western US, from 2008-2018. These estimates are predictions from ensemble machine learning models trained on 24-hour PM_2.5 measurements from monitoring station data across 11 states in the western US. Predictor variables were derived from satellite, land cover, chemical transport model (just for the 2008-2016 model), and meteorological data. Ten-fold spatial and random CV R² were 0.66 and 0.73, respectively, for the 2008-2016 model and 0.58 and 0.72, respectively for the 2008-2018 model. Comparing areal predictions to nearby monitored observations demonstrated overall R² of 0.68 for the 2008-2016 model and 0.58 for the 2008-2018 model, but we observed higher R² (&gt; 0.80) in many urban areas. These data can be used to understand spatiotemporal patterns of, exposures to and health impacts of PM_2.5 in the western US where PM_2.5 levels have been heavily impacted by wildfire smoke over this time period. <br>

我们针对2008—2018年美国西部各县级行政区、邮政编码区域（ZIP code）及普查分区（census tract）的几何中心点，开展了细颗粒物（fine particulate matter，PM₂.₅）的日均浓度估算。该估算结果基于集成机器学习模型（ensemble machine learning models）生成，模型训练所用数据为美国西部11个州的监测站获取的24小时细颗粒物观测值。预测变量来源于卫星遥感数据、土地覆盖数据、化学传输模型（chemical transport model，仅用于2008—2016年的模型）以及气象数据。2008—2016年模型的十折空间交叉验证与随机交叉验证的R²值分别为0.66与0.73；2008—2018年模型的对应数值分别为0.58与0.72。将面域预测结果与邻近的实地监测观测值进行对比后发现，2008—2016年模型的整体R²值为0.68，2008—2018年模型的整体R²值为0.58；但在多数城区，R²值可高于0.80。该数据集可用于探究美国西部细颗粒物的时空分布特征、暴露水平及其健康影响——此时间段内，美国西部的细颗粒物浓度受野火烟雾影响极为显著。