Journal of Japan Society for Atmospheric Environment / Taiki Kankyo Gakkaishi
Online ISSN : 2185-4335
Print ISSN : 1341-4178
ISSN-L : 1341-4178
Technical Report
A Bayesian Hierarchical Approach for PM2.5 Continuous Monitoring Data
Kunihiro HisatsuneMakiko Yamagami
Author information
JOURNAL FREE ACCESS

2015 Volume 50 Issue 2 Pages 107-116

Details
Abstract
Continuous monitoring of the PM2.5 stared in 2009 in each area of Japan. We can obtain much data from any of the sampling points, but these averages cannot express the characteristics of the points. Therefore, this paper presents a Bayesian hierarchical CAR model for analysis of PM2.5 daily data. This model is based on the estimated monthly average, effect of area and effect of the measurement point. The effect of area is highest around the Port of Nagoya and Owari region, lowest in the Mikawa region and Gifu. The effect of the sampling point is clarified and compared to that of the area. These effects compared to Local Moran's I and CPF values, and the consistency has been confirmed. The greatest product of effect of the point and area is 1.50, and smallest is 0.67. Although some of the measurement data is out of the 95% credible interval, this is often caused by aberrant measurements or lack of data.
Content from these authors
© 2015 Japan Society for Atomospheric Environment
Previous article Next article
feedback
Top