Path concentration profile reconstruction of optical remote sensing measurements using polynomial curve fitting procedures

Open path Fourier transform infrared (OP-FTIR) spectroscopy is one of several optical remote sensing (ORS) techniques that can identify and quantify many air pollutants. These instruments provide path-integrated measurements along the scanning beam path. There is a growing interest in gaining spatial resolution from ORS through innovative multiple-beam configurations and mathematical algorithms. In this study, we explored the use of a polynomial curve directly fitted to the path-integrated concentration (PIC) data for this application. The derivative of the fitted curve is the reconstructed density function of the underlying line concentration profile. We first validated this approach with a series of simulated Gaussian plumes. The PIC data were simulated by integrating the Gaussian functions over each beam path. We found that the piece-wise polynomial (cubic spline) algorithm performs better than the global polynomial algorithm. This approach was also applied to data collected previously in a ventilation chamber. Nitrous oxide was released from a point source as a tracer gas. A unistatic OP-FTIR collected the PIC data with 4 retroreflectors located along a one-dimensional domain for 15 min. For a wide plume, the concentration profile reconstructed from the OP-FTIR measurements fits well to the plume measured from 14 point samples collected simultaneously during the experiments ( R 2=0.83). For a narrow plume, the R 2 value between the point samples and the reconstructed plume is 0.40, similar to the one calculated from applying 4 ray-traced PIC. From the ray-traced synthetic data, it was also found that an additional PIC could improve the fitted R 2 value significantly to above 0.7.