长程差分吸收光谱技术气体浓度反演误差的定量估计

长程差分吸收光谱法(LP-DOAS)是基于最小二乘原理来反演大气痕量气体浓度的。LP-DOAS能对痕量气体进行高灵敏的测量，但是还没有统计的方法定量确定LP-DOAS反演误差。痕量气体的吸收通常很弱，外来影响因素决定了检测限和测量精度，其被误当做真正的吸收，增加了没有统计特性的噪声到残差中，导致最小二乘拟合误差(err(LSQ))有一个明显的误估计。研究采用蒙特卡罗方法，通过残差的循环移位定量确定差分吸收光谱法反演气体浓度的误差。实验结果表明，蒙特卡罗方法可以定量估计差分吸收光谱法反演误差，误估计因子为1.13，而err(LSQ)为3.12。

Quantitative Study on Errors in Evaluation of Trace Gases by Long Path Differential Optical Absorption Spectroscopy

Long path differential optical absorption spectroscopy(LP-DOAS) can be used to retrieve the concentration of trace gases based on the principle of least squares.LP-DOAS permits the sensitive measurement of concentrations of trace gases.However,the calculation of a statistically sound measurement precision still remains an unsolved problem.The absorbance of atmospheric trace gases is usually very weak,so the interference caused by the occurrence of non-reproducible spectral artifacts often determines the detection limit and the measurement precision.These spectral artifacts bias the least squares fitting result in two respects.Spectral artifacts are falsely interpreted as real absorption,and spectral artifacts add non-statistical noise to spectral residuals,resulting in a significant underestimation of the least squares fitting error.Monte Carlo method was presented to investigate the evaluation errors in DOAS spectra accurately.Residual inspection by cyclic displacement was used to estimate the effect of false interpretation of the artifact structures.The evaluation of simulated atmospheric measurement spectra shows that Monte Carlo method yields a good estimate of the spectra evaluation error.It is approximately real error,factors of errors are 1.13,but errors are ordinarily underestimated by err(LSQ),and factors of errors are 3.12.