Abstract
Sufficient and accurate air pollutant data are essential to analyze and control air contamination problems. An orthogonal polynomial fitting (OPF) method using Chebyshev basis functions is introduced to produce spatial distributions of fine particle (PM(2.5)) concentrations in central and southern regions of China. Idealized twin experiments (IE1 and IE2) are designed to validate the feasibility of the OPF method. IE1 is designed in accordance with the most common distribution of PM(2.5) concentrations in China, whereas IE2 represents a common distribution in spring and autumn. In both idealized experiments, prescribed distributions are successfully estimated by the OPF method with smaller errors than kriging or Cressman interpolations. In practical experiments, cross-validation is employed to assess the interpolation results. Distributions of PM(2.5) concentrations are well improved when OPF is applied. This suggests that errors decrease when the fitting order increases and arrives at the minimum when both orders reach 6. Results calculated by the OPF method are more accurate than kriging and Cressman interpolations if appropriate fitting orders are selected in practical experiments.