Abstract As a much finer particle, particulate matter less than 1 μm (PM1) plays an important role on the haze formation and human health. However, the capability of mapping PM1 concentration is severely impaired by coarse temporal resolution and low estimation accuracy, largely due to the neglect of spatial or temporal autocorrelation of PM1. In order to improve the estimation of high-resolution PM1, here we developed a novel spatiotemporal model named extreme gradient boosting (XGBoost)-geographically and temporally weighted regression (GTWR) using Himawari-8 aerosol optical depth (AOD), meteorological factors, and geographical covariates. The estimation of PM1 over Zhejiang province showed that XGBoost-GTWR method was characterized by greater predictive ability (10-fold cross-validation R2 = 0.83, root mean squared error (RMSE) = 10.72 μg/m3) compared with other 11 models. Additionally, the extrapolation test was performed to validate the robustness of the hybrid model and the result demonstrated that XGBoost-GTWR can accurately predict the out-of-band PM1 concentration (R2 = 0.75 (0.60), RMSE = 12.71 (12.58) μg/m3). The PM1 concentration displayed pronounced spatial heterogeneity, with the highest value in Quzhou (34.72 ± 1.77 μg/m3) and the lowest in Zhoushan (26.39 ± 1.56 μg/m3), respectively. In terms of the seasonality, the highest PM1 concentration was observed in winter (39.06 ± 3.08 μg/m3), followed by those in spring (32.54 ± 3.09 μg/m3) and autumn (30.97 ± 4.50 μg/m3), and the lowest one in summer (25.57 ± 5.22 μg/m3). The high aerosol emission and adverse meteorological conditions (e.g., low boundary layer height and lack of precipitation) were key factors accounting for the peak PM1 concentration observed in winter. Also, the PM1 concentration exhibited significant diurnal variation, peaking at 1500 local solar time (LST) but reaching the lowest value at 1000 LST. This method enhances our capability of estimating hourly PM1 from space, and lays a solid data foundation for improving the assessment of the fine particle-related health effect.

中文翻译：

---

结合极端梯度提升 - 地理和时间加权回归 (XGBoost-GTWR) 估计来自 Himawari-8 的高分辨率 PM1 浓度

摘要 作为一种更细小的颗粒，小于 1 μm 的颗粒物 (PM1) 对雾霾的形成和人类健康起着重要作用。然而，粗时间分辨率和低估计精度严重削弱了映射 PM1 浓度的能力，这主要是由于忽略了 PM1 的空间或时间自相关。为了提高对高分辨率 PM1 的估计，这里我们开发了一种名为极端梯度增强 (XGBoost)-地理和时间加权回归 (GTWR) 的新型时空模型，使用 Himawari-8 气溶胶光学深度 (AOD)、气象因素和地理协变量。浙江省PM1的估计表明，XGBoost-GTWR方法具有较强的预测能力（10倍交叉验证R2=0.83，均方根误差（RMSE）=10。72 μg/m3) 与其他 11 种型号相比。此外，还进行了外推测试以验证混合模型的稳健性，结果表明 XGBoost-GTWR 可以准确预测带外 PM1 浓度（R2 = 0.75 (0.60), RMSE = 12.71 (12.58) μg/立方米）。PM1浓度表现出明显的空间异质性，衢州最高（34.72±1.77μg/m3），舟山最低（26.39±1.56μg/m3）。从季节性来看，冬季PM1浓度最高（39.06±3.08μg/m3），其次是春季（32.54±3.09μg/m3）和秋季（30.97±4.50μg/m3），最低夏季（25.57 ± 5.22 μg/m3）之一。高气溶胶排放和不利的气象条件（例如，低边界层高度和缺乏降水）是导致冬季观测到的 PM1 浓度峰值的关键因素。此外，PM1 浓度表现出显着的日变化，在当地太阳时 (LST) 1500 时达到峰值，但在 LST 1000 时达到最低值。该方法增强了我们从空间估算每小时PM1的能力，为改进细颗粒物相关健康效应的评估奠定了坚实的数据基础。