基于动态约束插值的PM10浓度空间分布模拟

Spatial Distribution Simulation of PM10 Concentrations Based on Dynamic Constraint Interpolation

摘要: PM10是一类极具代表性的大气污染物，其表面易吸附多环芳烃、重金属等有毒有害物质，进而对人体健康与生态环境构成多重威胁，相关研究因此受到广泛重视。对PM10污染的研究不仅能深入剖析其传输扩散规律，更为后续污染防控工作提供坚实的科学支撑。在PM10相关研究中，精准模拟其浓度空间分布是核心环节，对全面掌握PM10污染特征具有不可替代的作用。地面监测、数值模拟与空间插值是当前研究PM10浓度空间分布的主要技术手段，但数值模型对数据完整性要求较高，传统插值方法在观测稀疏区域难以达到理想精度。动态约束插值法(DCIM)已被证实可有效提升时空稀疏观测数据的利用率，为此，本研究提出一种基于伴随模型的动态约束插值方法，通过算法优化增强PM10浓度的空间重建能力。利用2015年3月2日至3月5日中国中南部85个地面监测站的观测数据，通过分步实验设计开展系统验证：首先验证DCIM在PM10数值模拟中的有效性，实验结果显示该方法显著提升了模拟精度、优化了模拟效果，充分证明其在该领域的适用性与可靠性；然后针对插值方法在稀疏数据场中的局限性，引入高阶守恒插值(PPM)算法，实验表明该算法在适配稀疏数据场景方面具有显著优势。

Abstract: PM10 is a highly representative atmospheric pollutant whose surface readily adsorbs toxic substances such as polycyclic aromatic hydrocarbons and heavy metals, posing multiple threats to human health and the ecological environment. Consequently, research in this area has garnered significant attention. Studies on PM10 pollution not only provide in-depth insights into its transport and dispersion patterns but also offer robust scientific support for subsequent pollution prevention and control efforts. Accurate simulation of PM10’s spatial concentration distribution is central to such research, playing an irreplaceable role in comprehensively understanding its pollution characteristics. Ground-based monitoring, numerical modeling, and spatial interpolation are the primary techniques currently used to study PM10’s spatial distribution. However, numerical models demand high data completeness, while traditional interpolation methods struggle to achieve ideal accuracy in sparsely observed regions. Dynamic Constrained Interpolation Method (DCIM) has been proven effective in enhancing the utilization of spatiotemporally sparse observational data. Therefore, this study proposes a dynamic constrained interpolation method based on an adjoint model, which enhances the spatial reconstruction capability of PM10 concentrations through algorithm optimization. Using observational data from 85 ground monitoring stations in central and southern China between March 2 and March 5, 2015, systematic validation was conducted through a stepwise experimental design: First, DCIM’s effectiveness in PM10 numerical simulation was validated. Results demonstrated significant improvements in simulation accuracy and optimized outcomes, fully confirming its applicability and reliability in this domain. Subsequently, addressing limitations of traditional interpolation methods in sparse data fields, the higher-order conservation interpolation (PPM) algorithm was introduced. Experiments revealed that this algorithm possesses distinct advantages in adapting to sparse data scenarios.