[论文解读] Geographically and temporally weighted neural networks for satellite-based mapping of ground-level PM2.5

Tongwen Li, Huanfeng Shen+2 more|arXiv (Cornell University)|Sep 26, 2018

Air Quality and Health Impacts参考文献 42被引用 3

一句话总结

本研究提出地理与时间加权神经网络（GTWNNs），以通过捕捉气溶胶光学厚度（AOD）、气象因素、NDVI与地面观测之间非线性且时空异质的关系，提升基于卫星的地面PM2.5制图精度。GTWNNs在样本交叉验证中达到R²=0.80，在站点交叉验证中达到R²=0.79，优于传统线性时空模型，实现了中国范围内0.1°分辨率的PM2.5制图。

ABSTRACT

The integration of satellite-derived aerosol optical depth (AOD) and station-measured PM2.5 provides a promising approach for obtaining spatial PM2.5 data. Several spatiotemporal models, which considered spatial and temporal heterogeneities of AOD-PM2.5 relationship, have been widely adopted for PM2.5 estimation. However, they generally described the complex AOD-PM2.5 relationship based on a linear hypothesis. Previous machine learning models yielded great superiorities for fitting the nonlinear AOD-PM2.5 relationship, but seldom allowed for its spatiotemporal variations. To simultaneously consider the nonlinearity and spatiotemporal heterogeneities of AOD-PM2.5 relationship, geographically and temporally weighted neural networks (GTWNNs) were developed for satellite-based estimation of ground-level PM2.5 in this study. Using satellite AOD products, NDVI data, and meteorological factors over China as input, GTWNNs were set up with station PM2.5 measurements. Then the spatial PM2.5 data of those locations with no ground stations could be obtained. The proposed GTWNNs have achieved a better performance compared with previous spatiotemporal models, i.e., daily geographically weighted regression and geographically and temporally weighted regression. The sample-based and site-based cross-validation R2 values of GTWNNs are 0.80 and 0.79, respectively. On this basis, the spatial PM2.5 data with a resolution of 0.1 degree were generated in China. This study implemented the combination of geographical law and neural networks, and improved the accuracy of satellite-based PM2.5 estimation.

研究动机与目标

• 为解决线性时空模型在捕捉卫星AOD与地面PM2.5之间复杂非线性关系方面的局限性。
• 通过将地理空间与时间异质性整合进机器学习框架，提升地面PM2.5估算的准确性。
• 开发一种结合地理加权与神经网络的混合模型，以增强空气污染的时空预测能力。
• 利用卫星与地面数据，生成中国范围内高分辨率（0.1°）的PM2.5地图，即使在无监测站的区域亦可实现。

提出的方法

• 通过将地理与时间加权回归原理与前馈神经网络结合，开发GTWNNs，以建模局部非线性AOD-PM2.5关系。
• 输入特征包括中国范围内的卫星反演AOD、NDVI及气象变量（如温度、相对湿度、风速）。
• 应用空间与时间加权核函数，动态调整模型参数以适应不同位置与时间，实现神经网络的局部自适应。
• 利用监测站的地面PM2.5观测数据进行模型训练，空间与时间带宽通过交叉验证优化。
• 采用两阶段验证流程——样本交叉验证与站点交叉验证——评估模型性能。
• 利用训练好的GTWNN模型，为无地面站的区域生成中国范围内的高分辨率（0.1°）PM2.5地图。

实验结果

研究问题

• RQ1 神经网络模型能否有效捕捉在多样化地理与时间条件下，卫星反演AOD与地面PM2.5之间的非线性关系？
• RQ2 在神经网络架构中引入空间与时间加权，相较于传统线性模型，能否显著提升PM2.5估算精度？
• RQ3 AOD-PM2.5关系在地理与时间上的变异程度在多大程度上影响机器学习模型在空气质量制图中的表现？
• RQ4 GTWNNs能否在无地面监测站的区域生成可靠且高分辨率的PM2.5地图？

主要发现

• GTWNNs在样本交叉验证中达到R²=0.80，表明在整个研究区域内具有强大的预测性能。
• 站点交叉验证的R²达到0.79，证明在单个监测站处具有稳健的准确性。
• GTWNNs在预测精度上显著优于每日地理加权回归与地理与时间加权回归模型。
• 该模型成功生成了中国范围内的0.1°分辨率PM2.5地图，将空间覆盖范围扩展至地面监测网络之外。
• 将地理与时间加权与神经网络结合，有效捕捉了局部非线性AOD-PM2.5关系。
• 结果证实，AOD-PM2.5关系中的时空异质性至关重要，必须显式建模以实现准确的PM2.5制图。