The Novel Arrange & Average Algorithm for the Retrieval of Aerosol Properties from Multiwavelength Lidar Data

Aerosol particles affect the radiative energy balance in the atmosphere and thus influence regional and global climate through the direct and indirect radiative effect. Aerosols may result in net cooling or warming of the air, changes of the large-scale atmospheric circulation, cloud lifetime, and occurrence and intensity of precipitation. The presence and distribution of aerosols in space and time is highly inhomogeneous and variable. In addition, aerosol radiative forcing depends on the vertical distribution of aerosols. Regular observations of aerosol optical and physical properties over large temporal and spatial scales are needed for radiative impact studies to address the gaps in knowledge of the role of aerosols in climate change. In this context, light detection and ranging (lidar) instruments will play a crucial role as only these instruments provide information of aerosol properties with high vertical resolution.

Here we present a novel, automated and unsupervised algorithm called the Arrange & Average algorithm that employs a look-up table for the purpose of retrieving the real and imaginary part of the complex refractive index (CRI), effective radius, and total number, surface-area, and volume concentrations of particle size distributions for a wide range of parameter space. The distinctive feature of our approach is that the CRI is treated like other retrieval parameters which significantly simplifies the whole numerical scheme. The algorithm is used for processing optical data acquired with different measurement configurations of multiwavelength high-spectral-resolution lidar or Raman lidar.

The Arrange & Average algorithm is simple in its practical realization and flexible in the sense that it can be easily applied to various configurations of lidar instruments. We have conducted detailed numerical simulations and performance tests of the algorithm for 4 important lidar instrument configurations: “3 backscatter (β) + 2 extinction (α)” (“3β + 2α”), “3β + 1α”, “2β + 1α”, and “3β”. All the simulations were done under the assumption of particle size- and wavelength independent CRIs. The performance of the algorithm with regard to the retrieval precision of the investigated aerosol properties slowly degrades if we remove the input optical data.