通过在深度学习与不确定性量化中包含地质物理信息来提高北美各地细颗粒物物质化学组成的估计
Siyuan Shen1, Aaron van Donkelaar1, Nathan Jacobs2
1Department of Energy, Environmental, and Chemical Engineering, Washington University in St. Louis, St. Louis, Missouri 63130, United States.
ACS ES&T air
|February 19, 2026
概括
这项研究使用卷积神经网络 (CNN) 增强了北美各地细颗粒物 (PM2.5) 和其组件的估计. 新的验证方法揭示了模型性能和不确定性,改善了空气质量研究.
科学领域:
- 环境科学 环境科学
- 大气化学 大气化学
- 数据科学数据科学数据科学
背景情况:
- 暴露于细颗粒物 (PM2.5) 是全球主要的健康风险.
- 准确地描述PM2.5的化学成分对于健康研究和环境管理至关重要.
- 估计PM2.5度的现有方法在空间覆盖和准确性方面存在局限性.
研究的目的:
- 改进整个北美地区总PM2.5质量度及其化学成分的估计.
- 开发和应用先进的机器学习模型 (CNN) 来进行增强的PM2.5估计.
- 引入和使用一种新型交叉验证技术 (BLISCO) 进行可靠的模型评估,特别是在偏远地区.
主要方法:
- 开发和优化集成卫星,模拟和地面监测数据的卷积神经网络 (CNN).
- 应用CNN来估计北美 (2000-2023) 的月度PM2.5和成分度.
- 实施传统的10倍空间交叉验证和新的缓冲离开孤立站点和集群 (BLISCO) 方法,用于模型验证和不确定性评估.
主要成果:
- 对于总PM2.5和主要成分 (例如,总PM2.5的R2=0.82,硫酸盐=0.98) 的传统交叉验证,CNN显著一致.
- BLISCO交叉验证强调了传统方法可能高估性能和低估不确定性.
- 结合化学运输模型 (GEOS-Chem) 数据,改善了BLISCO中的CNN性能,特别是酸盐 (R2 0.51至0.81) 和 (R2 0.27至0.67).
结论:
- 开发的CNN方法为北美各地的PM2.5及其化学成分提供了改进的估计.
- BLISCO验证方法为模型性能和不确定性提供了更现实的评估,特别是用于推断.
- 距离监测地点的距离是影响偏远地区PM2.5估计不确定性的关键因素.
相关概念视频
01:05Estimation of the Physical Quantities
8.1K
On many occasions, physicists, other scientists, and engineers need to make estimates of a particular quantity. These are sometimes referred to as guesstimates, order-of-magnitude approximations, back-of-the-envelope calculations, or Fermi calculations. The physicist Enrico Fermi was famous for his ability to estimate various kinds of data with surprising precision. Estimating does not mean guessing a number or a formula at random. Instead, estimation means using prior experience and sound...
8.1K
01:29Application of Linearization and Approximation
111
A drone flying through complex terrain often relies on more than one sensing method to estimate small changes in altitude. Along with direct measurements, air pressure provides a useful indirect indicator of vertical movement. Atmospheric pressure decreases as altitude increases, and this relationship is commonly described using an exponential model. Although accurate, converting pressure measurements into altitude values requires calculations that are too complex to perform repeatedly during...
111