气溶胶光学特性对宽谱差分吸收激光雷达NO2质量浓度反演结果的影响
Accurate monitoring of nitrogen dioxide (NO2), a significant atmospheric pollutant, is essential for effective environmental management. Differential absorption lidar (DIAL) technology has emerged as a robust approach to address this challenge. However, the wavelength dependency of aerosol optical properties has a substantial impact on NO2-DIAL measurements. Previous studies mostly focus on the narrow-band NO2-DIAL technique without considering the spectral width of the emitted laser pulse. Therefore, the influence of aerosol optical properties on the retrieved NO2 mass concentration of the broad-band DIAL technique remains unclear. We aim to investigate aerosol-induced NO2 mass concentration errors under various atmospheric conditions through simulation studies and an approximation method for the broadband NO2-DIAL technique. We hope that this research can offer valuable insights into comprehending the influence of aerosol optical properties on broadband NO2-DIAL techniques.
We carry out research based on the broadband NO2-DIAL technique (Fig. 1) employing the Scheimpflug principle. The broadband NO2-DIAL system utilizes image sensors as detectors and high-power laser diodes as light sources (wavelength is 450 nm, power is 1.6 W), with an emission spectral typically ranging from 1-2 nm (full width at half maximum). Two different methods (the simulation method and the approximation method) have been adopted to elucidate the influence of atmospheric aerosols on the broadband NO2-DIAL technique. The broadband DIAL equation has been established, based on which simulated atmospheric lidar signals can be obtained with measured laser spectra and different atmospheric parameters, e.g., aerosol extinction coefficient, backscattering coefficient and ?ngstr?m exponent (Fig. 5). Therefore, the NO2 mass concentration containing the aerosol-induced retrieval errors can be acquired through segmented fitting for the simulated atmospheric lidar signals (Fig. 6). As a result, the NO2 mass concentration errors introduced by the aerosol extinction effect and the aerosol backscattering effect under various atmospheric conditions can be obtained through numerical calculation. Besides, the aerosol-induced NO2 mass concentration errors can also be mathematically derived based on spectral approximation—the approximation method. Meanwhile, cross-validations between the aerosol-induced NO2 mass concentration errors obtained from these two methods have also been carried out.
Several conclusions can be drawn according to simulation studies. When the atmospheric condition is homogeneous, for an extinction coefficient of 0.3 km-1 and an ?ngstr?m exponent of 3, the aerosol-induced retrieval error of the NO2 mass concentration is 14.7 μg/m3, while the error introduced by the aerosol backscattering effect is only about 0.6 μg/m3 (Fig.7). Therefore, when atmospheric aerosols are homogeneously distributed, the inversion error of NO2 mass concentration mainly depends on the aerosol extinction coefficient. The proportion of the NO2 mass concentration inversion error generated by the backscattering effect is generally less than 5%, which can be ignored. Besides, if the ?ngstr?m exponent approaches 1, the NO2 mass concentration error introduced by the aerosol extinction effect will decrease to 5 μg/m3 (Fig. 9). If aerosol plumes appear in a homogeneous atmosphere (0.3 km-1), for the extinction coefficient of 0.66 km?1 within the inhomogeneous range and an ?ngstr?m exponent of 3, the NO2 mass concentration error resulted from the aerosol extinction coefficient in the inhomogeneous region is 18.9 μg/m3. However, the error introduced by the aerosol backscattering effect increases to 3.3 μg/m3 with a fitting distance of 500 m (Fig. 11). Under typical weather conditions with a relatively small ?ngstr?m exponent of 1, the NO2 mass concentration error introduced by the aerosol backscattering effect will increase to 6.8 μg/m3 (Fig. 12). The simulation results indicate that the inversion error of aerosol backscattering effect on NO2 mass concentration largely depends on the non-uniformity of atmospheric aerosol distribution, the fitting range, etc. Meanwhile, increasing the fitting range can greatly reduce the NO2 mass concentration error introduced by the aerosol, especially for the backscattering effects. Comparison studies between the approximation method and the simulation method reveal that the NO2 mass concentration retrieval error introduced by the extinction effect is almost the same, while the backscattering coefficient-induced errors may be quite different (Figs. 13 and 14).
We evaluate the measurement errors of NO2 mass concentration caused by aerosol extinction and backscattering effects under various atmospheric conditions by two different methods (the simulation method and the approximation method) for the broadband NO2-DIAL technique. In the case of a homogeneous atmosphere, the NO2 mass concentration error is primarily determined by the aerosol extinction coefficient, while the contribution from aerosol backscattering effects can be neglected. However, if the atmosphere is inhomogeneous, the NO2 mass concentration error caused by the aerosol backscattering effect is significantly influenced by the inhomogeneous distribution of aerosol. It should be mentioned that the backscattering coefficient-induced NO2 mass concentration error is inversely proportional to the ?ngstr?m exponent in this case. In addition, we also derive an approximation model for NO2 mass concentration inversion errors caused by the extinction and backscattering coefficients based on spectral approximation. The comparison between the approximation method and the simulation method shows that the NO2 mass concentration inversion error generated by the extinction coefficients obtained by the two methods are generally in good agreement with small discrepancies. The inversion error caused by the aerosol backscattering coefficient may be affected by factors such as the computation method, the fitting range, and the spectral approximation. The approximation model provides an important tool for evaluating NO2 mass concentration errors in practical DIAL measurements.
1 引言
二氧化氮(NO2)是大气中一种重要的痕量气体,它在大气中的质量浓度增加会引发人的呼吸系统和心脑系统疾病,严重危害人类健康[1-5]。差分吸收激光雷达(DIAL)是探测大气NO2质量浓度廓线的一种有效手段,具有高探测灵敏度和高时空分辨率的特点,能够实时监测三维空间中的大气痕量气体(SO2、O3、NO2等)分布,在大气环境监测领域具有重要的应用价值[6-8]。DIAL技术的基本原理是向大气交替发射两束波长相近的激光脉冲,其中一束激光脉冲的波长(λon)位于待测气体的吸收峰,另一束激光脉冲的波长(λoff)位于待测气体的吸收谷,通过测量气体对波长为λon和λoff的激光的差分吸收,可反演出待测气体的质量浓度信息。
基于差分吸收的原理,DIAL技术在探测痕量气体时不可避免地受到其他大气成分(大气分子、大气气溶胶等)的干扰。对于其他气体分子的干扰,一般可通过选择合适的激光波长来避开主要干扰气体的吸收谱线。然而,气溶胶在λon和λoff波长处的光学特性取决于气溶胶质量浓度、种类、大气条件等因素,其产生的影响比较复杂。因此,在DIAL测量中,通常要求λon和λoff尽可能相近,同时尽可能地缩短交替探测的时间间隔,以确保波长分别为λon和λoff的激光光束探测的是同一状态下的大气,且在这两个波长下大气气溶胶的光学特性相似,进而可以忽略大气气溶胶对DIAL探测气体质量浓度的影响[9]。脉冲式DIAL在测量具有窄带吸收谱线的气体(如Hg、CO2、H2O等)时,差分吸收波长λon和λoff的间隔仅为pm量级,此时大气气溶胶在λon和λoff波长下的吸收和散射效应的差异可忽略不计[10-12];对于O3、NO2等具有较宽吸收谱线的气体,λon和λoff波长间隔通常在nm量级,此时大气气溶胶的影响一般不可直接忽略,尤其是当待测气体质量浓度较低或吸收较弱时,大气气溶胶的影响尤为严重[13-16]。
为了探讨大气气溶胶对DIAL测量结果的影响,针对O3和NO2气体测量,国内外学者开展了许多研究工作。1982年,Pelon等[17]讨论了λon和λoff间隔较大(λ=5 nm)以及大气气溶胶的非均匀分布对O3测量的影响。此后,Browell等[18]采用激光雷达直接测量的方法获得大气气溶胶的消光及后向散射系数,并结合误差修正算法修正气溶胶的影响,但因气溶胶参数难以准确估计,该方法本身存在一定的误差。1996年,王志恩等[19]提出了采用四波长差分吸收激光雷达测量O3质量浓度的方案。随后,他们进一步提出三波长DIAL方法[20],该方法能够在一定程度上克服大气气溶胶的影响,但系统较为复杂。2022年,谢银海等[21]以激光雷达实测的气溶胶分布为基础,模拟并对比了三波长O3-DIAL方法和双波长O3-DIAL方法的气溶胶引入误差,并得出气溶胶含量主要影响消光项误差,而后向散射误差主要受气溶胶空间分布影响的基本结论。NO2气体在450 nm波长附近的吸收截面较小且在大气背景中含量较低。利用DIAL技术测量大气NO2时,λon和λoff一般为448 nm和450 nm。由于λon和λoff波长间隔较大(1~2 nm),大气气溶胶对NO2质量浓度测量结果的影响较为复杂[22-23],测量低质量浓度NO2时必须仔细考虑。2017年,Liu等[24]根据激光雷达实测的气溶胶分布,评估了气溶胶对NO2质量浓度测量引入的误差。当波长指数(Ångström exponent)为1时,气溶胶消光效应引起的相对误差为0%~40%,气溶胶后向散射效应引起的相对误差为0.2%~0.8%。
以上研究初步探讨了气溶胶在DIAL测量中的影响,但大多是基于激光雷达实测数据进行分析,缺乏系统性的气溶胶误差分析。此外,对于脉冲式NO2-DIAL系统,一般采用染料激光器作为光源,其发射激光的线宽远小于NO2气体的吸收峰宽度,所以一般不需要考虑谱线宽度对DIAL测量结果的影响。然而,基于多模半导体激光器的连续波NO2-DIAL系统,发射光谱半峰全宽(FWHM;dFWHM)一般在1~2 nm[25],这使得气溶胶作用情况更加复杂。因此,在分析气溶胶引起的NO2质量浓度反演误差时需要考虑激光器发射光谱宽度的影响。为了厘清大气气溶胶对宽谱NO2-DIAL测量结果的影响,基于宽谱连续波NO2-DIAL技术,本文采用模拟计算和近似模型两种研究方法,对比分析不同大气状况下的气溶胶引起误差,系统性地分析大气气溶胶的消光和后向散射效应对NO2质量浓度反演结果的影响。
2 宽谱NO2差分吸收激光雷达原理
2.1 基于沙氏成像原理的宽谱NO2-DIAL
连续波差分吸收激光雷达(CW-DIAL)的原理如
图 1. 基于高功率多模二极管激光器的连续波NO2-DIAL系统原理示意图
Fig. 1. Schematic of continuous-wave NO2-DIAL system based on high-power multimode laser diode
图 2. NO2吸收截面(294 K)及450 nm二极管激光器在λon和λoff波长处的归一化光谱
Fig. 2. NO2 absorption cross-section (294 K), as well as normalized emission spectra of 450 nm laser diode at λon and λoff wavelengths
2.2 宽谱NO2-DIAL理论模型
结合二极管激光器发射激光的宽谱特性,根据大气探测激光雷达的基本理论,可建立宽谱NO2-DIAL方程。其中,波长为λon的大气激光雷达信号可表示为
式中:z为距离;
若系统探测波长λon和λoff的间隔较小,大气分子和气溶胶的消光系数及后向散射系数在λon和λoff的小范围内和波长无关,可忽略空气分子和气溶胶作用,不考虑两者的消光系数(
其中,有效吸收截面由NO2标准吸收截面
式中:
根据
3 激光雷达信号模拟及NO2质量浓度反演
要想通过
3.1 大气分子和气溶胶的光学特性
假设在波长为λoff的激光光束的光谱中心(λoff 0= 449.35 nm),空气分子的消光系数为0.0278 km-1。在二极管激光器的发射光谱范围(445~452 nm)内,可根据瑞利散射模型,模拟得到空气分子消光系数随波长的变化。同时,根据空气分子的后向散射系数与消光系数的关系,可进一步推导得到空气分子的后向散射系数随波长的变化。
气溶胶的消光系数和后向散射系数与大气状态密切相关。在模拟激光雷达曲线时,可预先设定某一波长的消光系数,随后根据激光雷达比、波长指数等参数确定其他所有波长处的气溶胶的消光系数及后向散射系数。同一波长处的气溶胶的消光系数
式中:
假设气溶胶消光系数
图 3. 不同波长指数下气溶胶消光系数与波长的关系
Fig. 3. Relationship between aerosol extinction coefficient and wavelength with different Ångström exponents
当大气气溶胶均匀分布时,气溶胶消光系数和后向散射系数不随距离变化,在测量范围内的气溶胶光学参数均可根据上述方法进行模拟。当大气气溶胶非均匀分布时,气溶胶消光系数和后向散射系数将随气溶胶质量浓度的改变在探测路径上发生变化。为了模拟非均匀大气条件,利用高斯函数模拟探测路径上大气气溶胶气团的强回波信号。如
图 4. 非均匀大气条件下模拟的气溶胶消光系数分布
Fig. 4. Distribution of aerosol extinction coefficient under inhomogeneous atmospheric conditions
3.2 激光雷达信号噪声
为了更好地模拟激光雷达信号,在模拟分析时引入了噪声。选取高斯白噪声作为系统的噪声模型,通过向模拟的大气激光雷达信号中添加高斯白噪声来获得更接近真实情况的仿真结果。激光雷达信号的信噪比可表示为
3.3 典型大气激光雷达信号
将大气分子/气溶胶光学参数、二极管激光器的归一化发射光谱、NO2吸收截面代入
图 5. 模拟的大气激光雷达信号强度I与距离的关系。(a)大气气溶胶均匀分布;(b)大气气溶胶非均匀分布
Fig. 5. Relationship between simulated atmospheric lidar signals intensity I and distance. (a) Homogeneous distribution of atmospheric aerosol; (b) inhomogeneous distribution of atmospheric aerosol
3.4 NO2质量浓度反演
根据
式中:
由
由此,可得到气溶胶消光和后向散射效应引起的NO2质量浓度测量误差
当模拟激光雷达信号时,若只考虑气溶胶的消光作用,则可将气溶胶后向散射系数视为常数,与波长无关。由此,由
图 6. 差分吸收曲线及分段拟合。(a)大气气溶胶均匀分布时的差分吸收曲线及分段拟合结果;(b)大气气溶胶非均匀分布时的差分吸收曲线及分段拟合结果;(c)分段拟合残差(拟合距离为500 m)
Fig. 6. Differential absorption curves and segmental fitting. (a) Differential absorption curve and segmental fitting curve when atmospheric aerosol is homogeneously distributed; (b) differential absorption curve and segmental fitting curve when atmospheric aerosol is inhomogeneously distributed; (c) fitting residuals (fitting range is 500 m)
4 模拟结果分析与讨论
下面将依据模拟计算得到的结果,分别讨论大气气溶胶均匀分布和非均匀分布两种情况下的NO2质量浓度反演误差。
4.1 大气气溶胶均匀分布下的NO2质量浓度反演误差
当大气气溶胶均匀分布时,差分吸收曲线如
图 7. 大气气溶胶均匀分布时,NO2质量浓度反演误差与拟合距离的关系。(a)仅由气溶胶消光系数引起的NO2质量浓度反演误差( )和仅由后向散射系数引起的NO2质量浓度反演误差( );(b)气溶胶消光和后向散射系数引起的NO2质量浓度反演误差之和( + ),以及同时考虑气溶胶消光和后向散射系数引起的NO2质量浓度反演误差( )
Fig. 7. Relationship between retrieval error of NO2 mass concentration and fitting range when atmospheric aerosol is homogeneously distributed. (a) Retrieval errors of NO2 massconcentration resulted from aerosol extinction coefficient ( ) and backscattering coefficient ( ); (b) sum of retrieval errors of NO2 mass concentration resulted from aerosol extinction coefficient and backscattering coefficient ( + ), as well as retrieval error of NO2 mass concentration considering both aerosol extinction and backscattering coefficients ( )
从
下面将忽略噪声的影响,固定拟合距离为500 m,探讨气溶胶引入的反演误差和波长间隔及气溶胶消光系数之间的关系。需要强调的是,波长间隔的增大不仅会使得气溶胶消光系数随探测波长发生改变,也会引起NO2差分吸收截面的改变。如
图 8. 大气气溶胶均匀分布时,气溶胶消光系数引起的NO2质量浓度反演误差与波长间隔的关系
Fig. 8. Relationship between aerosol extinction coefficient induced retrieval error of NO2 mass concentration and wavelength interval, when atmospheric aerosol is homogeneously distributed
为了进一步研究波长指数和气溶胶消光系数对测量误差的影响,固定波长间隔为1.5 nm,拟合距离为500 m,得到如
图 9. 大气气溶胶均匀分布时,气溶胶消光系数引起的NO2质量浓度反演误差与波长指数和气溶胶消光系数的关系
Fig. 9. Relationship among aerosol extinction coefficient induced retrieval error of NO2 mass concentration, Ångström exponent, and aerosol extinction coefficient of atmospheric aerosol, when atmospheric aerosol is homogeneously distributed
4.2 大气气溶胶非均匀分布下的NO2质量浓度反演误差
当大气气溶胶非均匀分布时,差分吸收曲线如
图 10. 大气气溶胶非均匀分布时,在探测路径上气溶胶引入的NO2质量浓度反演误差与拟合距离关系。(a)气溶胶消光系数引起的NO2质量浓度反演误差( );(b)气溶胶后向散射系数引起的NO2质量浓度反演误差( );(c)同时考虑气溶胶消光系数和后向散射系数引起的NO2质量浓度反演误差( )
Fig. 10. Relationship between aerosol-induced retrieval error of NO2 mass concentration and fitting distance on detection path, when atmospheric aerosol is inhomogeneously distributed. (a) Retrieval error of NO2 mass concentration resulted from aerosol extinction coefficient ( ); (b) retrieval error of NO2 mass concentration resulted from aerosol backscattering coefficient ( ); (c) retrieval error of NO2 mass concentration considering both aerosol extinction and backscattering coefficient ( )
从
图 11. 大气气溶胶非均匀分布时,气溶胶引起的NO2质量浓度反演误差随测量距离的变化。(a)气溶胶消光系数引起的NO2反演误差( )和后向散射系数引起的NO2质量浓度反演误差( );(b)气溶胶消光系数和后向散射系数引起的NO2质量浓度反演误差之和( + ),以及同时考虑消光系数和后向散射系数引起的NO2质量浓度反演误差( )
Fig. 11. Aerosol-induced retrieval error of NO2 mass concentration and varies with detecting distance, when atmospheric aerosol is inhomogeneously distributed. (a) Retrieval error of NO2 mass concentration resulted from aerosol extinction coefficient ( ) and aerosol backscattering coefficient ( ); (b) sum of retrieval errors of NO2 mass concentration resulted from extinction coefficient and backscattering coefficient ( + ), as well as retrieval error of NO2 mass concentration considering both aerosol extinction and backscattering effect ( )
为了进一步研究在气溶胶非均匀分布时,由气溶胶后向散射系数引起的NO2质量浓度反演误差的变化,改变气溶胶消光系数(及对应的后向散射系数)和波长指数,得到如
图 12. 大气气溶胶非均匀分布且连续改变大气气溶胶非均匀范围消光系数( 平均值:0.3 km-1, 峰值:0.3 km-1→0.9 km-1)时,气溶胶后向散射产生的NO2质量浓度反演误差与大气气溶胶质量浓度和波长指数的关系
Fig. 12. Relationship among aerosol backscattering-induced retrieval error of NO2 mass concentration, Ångström exponent, and atmospheric aerosol mass concentration, when atmospheric aerosol is inhomogeneously distributed and extinction coefficient of atmospheric aerosol in inhomogeneous range is continuously changed ( -mean: 0.3 km-1, -peak: 0.3 km-1→0.9 km-1)
5 基于近似模型的NO2质量浓度反演误差分析
前文的模拟计算结果很好地揭示了NO2质量浓度反演误差与气溶胶消光系数和后向散射系数之间的关系,但无法应用于实际过程中激光雷达探测曲线的分析。因此,本节将从DIAL方程出发,通过理论近似的方式获得气溶胶消光及后向散射系数引起的反演误差,并将近似模型得到的结果与模拟计算结果进行对比分析。
根据
不难看出,
化简得到NO2质量浓度及反演误差
式中:
不同波长处的空气分子后向散射系数之间满足瑞利散射条件,
由于λon和λoff非常接近,且
由
5.1 大气气溶胶均匀分布
图 13. 大气气溶胶均匀分布且消光系数为0.3 km-1时,气溶胶引入的NO2质量浓度反演误差与波长指数的关系。(a)模拟计算法得到的气溶胶消光系数引起的误差( )与近似模型法得到的气溶胶消光系数引起的误差( )的对比;(b)模拟计算法得到的气溶胶引起的总误差( )与近似模型法得到的气溶胶引起的总误差( )的对比
Fig. 13. Relationship between aerosol-induced retrieval error of NO2 mass concentration and Ångström exponent, when atmospheric aerosol is homogeneously distributed and value of extinction coefficient is 0.3 km-1. (a) Comparison of aerosol extinction coefficient-induced errors obtained by simulated calculation method ( ) and approximate model method ( ); (b) comparison of total aerosol-induced error obtained by simulated calculation method ( ) and approximate model method ( )
5.2 大气气溶胶非均匀分布
图 14. 大气气溶胶非均匀分布时,气溶胶引入的NO2质量浓度反演误差随测量距离的变化。(a)模拟计算法得到的气溶胶消光系数引起的误差( )与近似模型法得到的气溶胶消光系数引起的误差( )的对比;(b)模拟计算法得到的气溶胶后向散射系数引起的误差( )与近似模型法得到的气溶胶后向散射系数引起的误差( )的对比;(c)模拟计算法得到的气溶胶引起的总误差( + )与近似模型法得到的气溶胶引起的总误差( )的对比
Fig. 14. Aerosol-induced retrieval error of NO2 mass concentration varies with detecting distance, when atmospheric aerosol is inhomogeneously distributed. (a) Comparison of aerosol extinction coefficient-induced errors obtained by simulated calculation method ( ) and approximate model method ( ); (b) comparison of aerosol backscattering coefficient-induced errors obtained by simulated calculation method ( ) and approximate model method ( ); (c) comparison of total aerosol-induced errors obtained by simulated calculation method ( + ) and approximate model method ( )
图 15. 大气气溶胶非均匀分布时,不同拟合距离下模拟计算法得到的由气溶胶后向散射系数引起的NO2质量浓度误差( )与近似模型法得到的由气溶胶后向散射系数引起的NO2质量浓度误差( )对比
Fig. 15. Comparison of aerosol backscattering coefficient-induced errors of NO2 mass concentration obtained by simulated calculation method at different fitting distances ( ) and approximate model method ( ), when atmospheric aerosol is inhomogeneously distributed
根据
图 16. 大气气溶胶非均匀分布时,气溶胶引入的NO2质量浓度反演误差随测量距离的变化。(a)模拟计算法经微分得到的气溶胶消光系数引起的误差( )与近似模型法得到的气溶胶消光系数引起的误差( )的对比;(b)模拟计算法经微分得到的气溶胶后向散射系数引起的误差( )与近似模型法得到的气溶胶后向散射系数引起的误差( )的对比;(c)模拟计算法经微分得到的气溶胶引起的总误差( + )与近似模型法得到的气溶胶引起的总误差( )的对比
Fig. 16. Aerosol-induced retrieval error of NO2 mass concentration varies with detecting distance, when atmospheric aerosol is inhomogeneously distributed. (a) Comparison of aerosol extinction coefficient-induced errors obtained by simulated calculation method of differentiation ( ) and approximate model method ( ); (b) comparison of aerosol backscattering coefficient-induced errors obtained by simulated calculation method of differentiation ( ) and approximate model method ( ); (c) comparison of total aerosol-induced error sobtained by simulated calculation method of differentiation ( + ) and approximate model method( )
从
图 17. 大气气溶胶非均匀分布时,近似模型法在不同波长指数下得到的气溶胶后向散射系数引起的NO2质量浓度反演误差随测量距离的变化
Fig. 17. Aerosol backscattering coefficient-induced retrieval error of NO2 mass concentrations obtained by approximate model method varies with detecting distance with different Ångström exponents, when atmospheric aerosol is inhomogeneously distributed
6 结论
气溶胶消光和后向散射效应的波长依赖性是NO2-DIAL测量结果的重要影响因素。为了厘清气溶胶消光和后向散射效应引起的NO2质量浓度测量误差,从两种方式(模拟计算和近似模型)入手开展了系统性的研究。
针对宽谱连续波NO2-DIAL技术的特点,通过仿真不同大气条件下激光雷达信号,以数值计算的方式分析了宽谱DIAL气溶胶消光和后向散射效应引起的NO2质量浓度反演误差。研究发现,当大气气溶胶均匀分布、大气气溶胶消光系数为0.3 km⁻¹且波长指数为3时,气溶胶消光效应引起的误差为14.7 μg/m³,而气溶胶后向散射效应引起的误差仅为0.6 μg/m³。因此,当大气气溶胶均匀分布时,NO2质量浓度反演误差主要取决于气溶胶消光系数,后向散射效应产生的NO2质量浓度反演误差占比一般小于5%,可忽略不计。此外,当气溶胶消光系数不变、波长指数接近1时,气溶胶消光效应引起的NO2质量浓度误差将降至5 μg/m³左右。若均匀大气(0.3 km⁻¹)中出现气溶胶团,且非均匀范围内消光系数为0.66 km⁻¹和波长指数为3时,采用500 m拟合距离,非均匀区域的气溶胶消光效应引起误差为18.9 μg/m³,气溶胶后向散射效应引起的误差增至3.3 μg/m³。当波长指数约为1时,气溶胶后向散射效应引起的NO2质量浓度反演误差将会增加至6.8 μg/m3。结果表明:气溶胶后向散射效应对NO2质量浓度反演误差的影响在很大程度上取决于大气气溶胶的非均匀分布程度;适当增大分段拟合距离可有效降低气溶胶引起误差,尤其是气溶胶后向散射效应引起的反演误差。综上所述,通过实际的激光雷达信号模拟仿真,模拟计算方法可精确揭示气溶胶消光和后向散射效应引起的NO2质量浓度反演误差。分析结果表明,气溶胶引起的NO2质量浓度反演误差不仅依赖于消光系数和后向散射系数,还依赖于波长指数、拟合距离等。
通过光谱近似得到宽谱NO2-DIAL气溶胶消光和后向散射效应引起的NO2质量浓度反演误差的近似模型。近似模型与模拟计算的结果表明:两种方式得到的消光效应引起的NO2质量浓度反演误差略有差异,但整体吻合度较高;气溶胶后向散射效应引起的NO2质量浓度反演误差则容易受到计算方式、光谱近似等因素的影响。该模型为实际测量激光雷达信号及气溶胶引起的NO2质量浓度反演误差提供了一种新的分析方法。
[1] Anand J S, Monks P S, Leigh R J. An improved retrieval of tropospheric NO2 from space over polluted regions using an Earth radiance reference[J]. Atmospheric Measurement Techniques, 2015, 8(3): 1519-1535.
[2] Ionov D V, Timofeyev Y M, Sinyakov V P, et al. Ground-based validation of EOS-Aura OMI NO2 vertical column data in the midlatitude mountain ranges of Tien Shan (Kyrgyzstan) and Alps (France)[J]. Journal of Geophysical Research: Atmospheres, 2008, 113(D15): D15-08.
[3] Valks P, Pinardi G, Richter A, et al. Operational total and tropospheric NO2 column retrieval for GOME-2[J]. Atmospheric Measurement Techniques, 2011, 4(7): 1491-1514.
[4] Zien A W, Richter A, Hilboll A, et al. Systematic analysis of tropospheric NO2 long-range transport events detected in GOME-2 satellite data[J]. Atmospheric Chemistry and Physics, 2014, 14(14): 7367-7396.
[5] Wenig M O, Cede A M, Bucsela E J, et al. Validation of OMI tropospheric NO2 column densities using direct-Sun mode Brewer measurements at NASA Goddard Space Flight Center[J]. Journal of Geophysical Research: Atmospheres, 2008, 113(D16): D16-45.
[6] Rothe K W, Brinkmann U, Walther H. Applications of tunable dye lasers to air pollution detection: measurements of atmospheric NO2 concentrations by differential absorption[J]. Applied Physics, 1974, 3(2): 115-119.
[7] Fredriksson K A, Hertz H M. Evaluation of the DIAL technique for studies on NO2 using a mobile lidar system[J]. Applied Optics, 1984, 23(9): 1403-1411.
[8] Galle B, Sunesson A, Wendt W. NO2-Mapping using laser-radar techniques[J]. Atmospheric Environment (1967), 1988, 22(3): 569-573.
[9] Schotland R M. Errors in the lidar measurement of atmospheric gases by differential absorption[J]. Journal of Applied Meteorology, 1974, 13(1): 71-77.
[10] Mei L, Zhao G Y, Svanberg S. Differential absorption lidar system employed for background atomic mercury vertical profiling in South China[J]. Optics and Lasers in Engineering, 2014, 55: 128-135.
[11] Amediek A, Fix A, Wirth M, et al. Development of an OPO system at 1.57 μm for integrated path DIAL measurement of atmospheric carbon dioxide[J]. Applied Physics B, 2008, 92(2): 295-302.
[12] Koch G J, Beyon J Y, Gibert F, et al. Side-line tunable laser transmitter for differential absorption lidar measurements of CO2: design and application to atmospheric measurements[J]. Applied Optics, 2008, 47(7): 944-956.
[13] Molina L T, Molina M J. Absolute absorption cross sections of ozone in the 185- to 350-nm wavelength range[J]. Journal of Geophysical Research: Atmospheres, 1986, 91(D13): 14501-14508.
[14] Kovalev V A, McElroy J L. Differential absorption lidar measurement of vertical ozone profiles in the troposphere that contains aerosol layers with strong backscattering gradients: a simplified version[J]. Applied Optics, 1994, 33(36): 8393-8401.
[15] Cheng Y A, Zhang Z, Kong Z, et al. Evaluation of systematic errors for the continuous-wave NO2 differential absorption lidar employing a multimode laser diode[J]. Applied Optics, 2020, 59(29): 9087-9097.
[16] Vandaele A C, Hermans C, Simon P C, et al. Measurements of the NO2 absorption cross-section from 42000 cm-1 to 10000 cm-1 (238-1000 nm) at 220 K and 294 K[J]. Journal of Quantitative Spectroscopy and Radiative Transfer, 1998, 59(3/4/5): 171-184.
[17] Pelon J, Mégie G. Ozone monitoring in the troposphere and lower stratosphere: evaluation and operation of a ground-based lidar station[J]. Journal of Geophysical Research: Oceans, 1982, 87(C7): 4947-4955.
[18] Browell E V, Ismail S, Shipley S T. Ultraviolet DIAL measurements of O3 profiles in regions of spatially inhomogeneous aerosols[J]. Applied Optics, 1985, 24(17): 2827-2836.
[19] 王志恩, 胡欢陵, 周军. 双差分吸收激光雷达:一种能有效减小气溶胶对臭氧测量影响的方法[J]. 气象学报, 1996, 54(4): 437-446.
Wang Z E, Hu H L, Zhou J. Dual differential absorption lidar: a new method to reduce effectively the effect of aerosols on ozone measurement[J]. Acta Meteorologica Sinica, 1996, 54(4): 437-446.
[20] Wang Z E, Nakane H, Hu H L, et al. Three-wavelength dual differential absorption lidar method for stratospheric ozone measurements in the presence of volcanic aerosols[J]. Applied Optics, 1997, 36(6): 1245-1252.
[21] 谢银海, 曹念文, 杨思鹏, 等. 激光雷达实测气溶胶对臭氧探测的误差分析[J]. 气象水文海洋仪器, 2022, 39(4): 6-9.
Xie Y H, Cao N W, Yang S P, et al. Error analysis of ozone detection by aerosol measured by lidar[J]. Meteorological, Hydrological and Marine Instruments, 2022, 39(4): 6-9.
[22] Cao N W, Fujii T, Fukuchi T, et al. Estimation of differential absorption lidar measurement error for NO2 profiling in the lower troposphere[J]. Optical Engineering, 2002, 41(1): 218-224.
[23] Cao N, Fukuchi T, Fujii T, et al. Error analysis for NO2 DIAL measurement in the troposphere[J]. Applied Physics B, 2006, 82(1): 141-148.
[24] Liu Q W, Chen Y F, Wang J, et al. Measurement of atmospheric NO2 profile using three-wavelength dual-differential absorption lidar[J]. Proceedings of the SPIE, 2017, 10605: 106053L.
[25] Mei L, Guan P, Kong Z. Remote sensing of atmospheric NO2 by employing the continuous-wave differential absorption lidar technique[J]. Optics Express, 2017, 25(20): A953-A962.
[26] LiuQ W, ChenY F, HuangJ, et al. Correction algorithm of atmospheric NO2 concentration caused by heavy aerosol load in differential absorption Lidar[C]//Sixth Symposium on Novel Photoelectronic Detection Technology and Application, 2019, Beijing, China: SPIE11455.
[27] 王向川, 饶瑞中. 大气气溶胶和云雾粒子的激光雷达比[J]. 中国激光, 2005, 32(10): 1321-1324.
[28] 张朝阳, 苏林, 陈良富. 中国典型地区气溶胶激光雷达比反演与分析[J]. 中国激光, 2013, 40(5): 0513002.
Article Outline
成远, 余纪恒, 宫振峰, 梅亮. 气溶胶光学特性对宽谱差分吸收激光雷达NO2质量浓度反演结果的影响[J]. 光学学报, 2024, 44(6): 0601016. Yuan Cheng, Jiheng Yu, Zhenfeng Gong, Liang Mei. Influence of Aerosol Optical Properties on Retrieval Results of NO2 Mass Concentration in Broadband Differential Absorption Lidar[J]. Acta Optica Sinica, 2024, 44(6): 0601016.