Daily 8-Hour Maximum and Annual O3 Concentrations for the Contiguous United States, 1-km Grids, Version 1.10 (2000-2016)

The Daily 8-Hour Maximum and Annual O3 Concentrations for the Contiguous United States, 1-km Grids, Version 1.10 (2000-2016) data set contains estimates of ozone concentrations at a high resolution spatially (1-km grid cells) and temporally (daily) for the years 2000 to 2016. These predictions incorporated various predictor variables such as Ozone (O3) ground measurements from the U.S. Environmental Protection Agency (EPA) Air Quality System (AQS) monitoring data, land-use variables, meteorological variables, chemical transport models and remote sensing data, along with other data sources. After imputing missing data with machine learning algorithms, a geographically-weighted ensemble model was applied that combined estimates from three types of machine learners (neural network, random forest, and gradient boosting). The annual predictions were computed by averaging the daily 8-hour maximum predictions in each year for each grid cell. The results demonstrate high overall model performance with a cross-validated R-squared value against daily observations of 0.90 and 0.86 for annual averages. In version 1.10, we have enhanced the completeness of daily O3 predictions by employing linear interpolation to impute missing values. Specifically, for days with small spatial patches of missing data with less than 100 grid cells, we used inverse distance weighting interpolation to fill the missing grid cells. Other missing daily O3 predictions were interpolated from the nearest days with available data. Annual predictions were updated by averaging the imputed daily predictions for each year in each grid cell. These daily 8-hour maximum and annual O3 predictions allow public health researchers to respectively estimate the short- and long-term effects of O3 exposures on human health, supporting the U.S. EPA for the revision of the National Ambient Air Quality Standards for O3. The data are available in RDS and GeoTIFF formats for statistical research and geospatial analysis.

《美国本土连续地带每日8小时最大臭氧浓度及年度臭氧浓度，1公里网格，版本1.10（2000-2016）》数据集包含了2000年至2016年间，以高分辨率空间（1公里网格单元）和时效（每日）为特征的对臭氧浓度的估计。这些预测融合了多种预测变量，包括来自美国环境保护署（EPA）空气质量系统（AQS）监测数据的臭氧（O3）地面测量值、土地利用变量、气象变量、化学传输模型和遥感数据，以及其他数据源。通过机器学习算法对缺失数据进行填补后，应用了地理加权集成模型，该模型结合了三种类型机器学习器的估计值（神经网络、随机森林和梯度提升）。年度预测通过平均每年每个网格单元的每日8小时最大预测值来计算。结果显示，模型整体性能卓越，与每日观测值的交叉验证R平方值分别为0.90和0.86。在版本1.10中，我们通过线性插值方法增强了对每日臭氧预测的完整性，具体而言，对于少于100个网格单元的小面积缺失数据，我们使用了逆距离加权插值来填补缺失的网格单元。其他缺失的每日臭氧预测则通过最邻近的有数据的日子进行插值。年度预测通过平均每个网格单元每年填补后的每日预测值进行更新。这些每日8小时最大和年度臭氧预测值使得公共卫生研究人员能够分别估算臭氧暴露对人类健康的短期和长期影响，支持美国EPA对国家环境空气质量标准中臭氧标准的修订。数据以RDS和GeoTIFF格式提供，适用于统计研究和地理空间分析。