In the northern region of Thailand, PM2.5 concentration has been the worst air pollution problem
in the country for the past several years. This study applies the gappy proper orthogonal
decomposition (gappy POD) method to estimate values missing from an incomplete dataset of
PM2.5 concentration in this area. Although monitoring and collection of PM2.5 concentration data
provide information about the air quality in the area, the collection process often misses some data,
hence obscuring important information. After the POD method is used to extract dominant data
information, the gappy POD, based on least squares optimization, is employed to approximate
the missing data. The number of POD bases plays an important role in the approximation; thus,
investigating the eff ects of a number of POD bases used in the gappy POD method is needed.
The accuracy of the gappy POD method is validated by comparing estimation errors to errors
from the baseline interpolation method using inverse distance weighting (IDW). The study
shows that the results of the gappy POD method are more accurate than the IDW estimations.