A new method for interpolation of missing air quality data at monitor stations

• Published in: Environment international Vol. 169; p. 107538
• Main Authors: Xu, Chengdong, Wang, Jinfeng, Hu, Maogui, Wang, Wei
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2022; Elsevier
• Subjects: air quality; Air quality dataset; covariance; data collection; environment; Heterogeneous population; Interpolation; kriging; particulates; Sparse sample; spatial variation; Sparse sample; Interpolation; Heterogeneous population; Air quality dataset
• ISSN: 0160-4120; 1873-6750; 1873-6750
• DOI: 10.1016/j.envint.2022.107538

Studies in environmental fields often suffer from air quality datasets incomplete at certain places and times. Here, a Spatial-Temporal Point Interpolation based on Biased Sentinel Hospitals Areal Disease Estimation (STPI-BSHADE) interpolation method was introduced to address this issue. The method was based on the spatial statistic trinity theory, where the statistical error is determined by the population properties, the condition of the sample, and the method of estimation. In our study, the spatial association of the variables was quantified by the covariance and the ratio of air quality data between stations, resulting in linear unbiased estimates of the missing data. STPI-BSHADE was compared with two widely used statistical methods, inverse distance weighting (IDW) and Kriging. Theoretically, IDW and Kriging are short of the capacity of using the heterogeneous characteristics of the population and remedying the sample bias. Empirically, the accuracy of the STPI-BSHADE method was assessed using hourly particulate matter 2.5 data, collected from May 13 to December 31, 2014, in the Beijing-Tianjin-Hebei areas, where air quality presents spatial heterogeneity. The experimental results also demonstrated that STPI-BSHADE significantly outperformed the traditional methods.