Optimal Reduction of the Local-scale PM2.5 Monitoring Network (II): Identifying the Optimal Subnetwork

• Published in: Journal of Japan Society for Atmospheric Environment / Taiki Kankyo Gakkaishi Vol. 57; no. 3; pp. 77 - 89
• Main Authors: Yamasaki, Yoshitomo, Furusawa, Shoei, Kohara, Daisuke, Toyonaga, Satoshi, Yamamoto, Yusuke, Yano, Hiromichi, Araki, Shin
• Format: Journal Article
• Language: Japanese
• Published: Japan Society for Atmospheric Environment 18.03.2022; 公益社団法人 大気環境学会
• Subjects: air pollution monitoring network; local government; optimization; PM2.5; regression kriging
• ISSN: 1341-4178; 2185-4335
• DOI: 10.11298/taiki.57.77

In this study, a method for the optimal reduction of the local-scale PM2.5 monitoring network was developed using the network in the Kumamoto prefecture, Japan. The basic concept of the method is to evaluate how well subsets of the PM2.5 networks (i.e., subnetworks) can represent the full network. There are three steps in the method. In the first step, concentration maps were predicted by Regression Kriging as described in the previous study. The maps with subnetworks and the map with full network were statistically compared. The subnetworks with no significant difference from the full network were extracted and passed the first step. In the second step, the passed subnetworks were evaluated by calculating the index value based on the spatial area affected by exclusion of the stations. The subnetwork with the smallest index value was determined as the optimal subnetwork. In the third step, the validation of the optimal subnetwork was carried out by comparison with externally observed values. As a result of these steps, the optimal subnetwork with only a negligible difference to the full networks could be identified. This method will probably be useful for the optimal reduction of the local-scale PM2.5 networks, those governed by local government.