The influence of temperature non-uniformity on retrieval of gas concentration in OP-FTIR spectroscopy

Abstract

Open-Path Fourier Transform Infrared FTIR gas measurement methods are commonly used in environmental monitoring. They are non-invasive and remote sensing techniques with relatively simple measuring setup. CLS (Classical Least Squares), PLS (Partial Least Squares) methods and HITRAN simulated data are often used for spectra analysis. Unfortunately, data interpretation in these type of measurement require sophisticated algorithms and expert knowledge as well. Recently, the effect of temperature influence on the determination of gas content is studied in many scientific papers. However, in practice, the temperature non-uniformity along the observed path can occurs apart from the temperature changes. In order to assess the effect of temperature variation and its non-uniformity absorption spectra simulation model is built. Then PLS models, based on synthetically generated data for different temperatures is created. Next, the effect of temperature influence on CO concentration was calculated together with the simulation spectra for temperature non-uniform optical path. For these spectra gas content were determined by using PLS models built for specific temperature. In our case the calibration model was built using synthetic spectra for different temperatures in 290-340 K range. Finally, it was shown that the constructed universal model is less sensitive to temperature nonuniformity than the classical one.