干气溶胶复折射率的参数化方案

基于成都市2017年10~12月WS600一体式气象站、AURORA-3000积分浊度计、AE-31黑碳仪以及GRIMM180环境颗粒物监测仪的地面逐时观测数据，结合Mie散射理论数值改进算法与免疫进化算法反演了550nm波长处干气溶胶复折射率（DACRI）的实部（n_re）和虚部（n_i）.通过分析DACRI反演结果与颗粒物质量浓度之比（BC/PM₁，BC/PM_2.5，BC/PM₁₀，PM₁/PM_2.5，PM₁/PM₁₀和PM_2.5/PM₁₀）之间的相关性，利用逐步线性回归方法分别构建了n_re和n_i的参数化方案. n_re和n_i的参数化方案拟合值与实测值之间的相关系数（R）分别为0.54和0.85（P<0.0001），平均相对误差（MRE）分别为2.31%和15.18%.此外，利用该参数化方案计算的DACRI模拟了一次灰霾演化过程中干气溶胶的散射系数b_sp和吸收系数b_ap，其模拟值与实测值之间的相关系数分别为0.98和0.91（P<0.0001），平均相对误差分别为7.43%和14.97%.

Parameterization scheme for dry aerosols complex refractive index

By utilizing the ground-based monitored data at an hourly time step recorded by WS600 integrated weather station, AURORA-3000 integrating nephelometer, AE-31aethalometer and GRIMM180 environment particle monitors from October to December in 2017 in Chengdu, the real part (n_re) and imaginary part (n_i) of dry aerosol complex refractive index (DACRI) at the wavelength of 550nm were retrieved by combining immune evolution algorithm with an improved Mie scattering theory algorithm. Stepwise linear regression method was then employed to respectively develop the new parameterization schemes by investigating the correlation between the retrieved DACRI and the ratios of particle mass concentration (BC/PM₁, BC/PM_2.5, BC/PM₁₀, PM₁/PM_2.5, PM₁/PM₁₀, and PM_2.5/PM₁₀). The correlation coefficients (R) were respectively 0.54 and 0.85 (P<0.0001) between n_re and n_i of DACRI calculated by the parameterization schemes and the observations, and the corresponding mean relative errors (MRE) were respectively 2.31% and 15.18%. These parameterization schemes were further applied to the simulation of the dry aerosol scattering coefficient (b_sp) and absorption coefficient (b_ap) during an evolution of haze. The results show that correlation coefficients were respectively 0.98 and 0.91 (P<0.0001) between the simulations and the observations, and that the corresponding mean relative errors were respectively 7.43% and 14.97%.