Applying graph theory for a more consistent estimate of the boundary layer height: Mixing layer height retrieval using ground-based lidar measurements

Abstract

The height of the atmospheric boundary layer, or mixing layer (MLH) is an important parameter for understanding the dynamics of the atmosphere and dispersion of air pollution. MLH can be retrieved from lidar or ceilometer backscatter data. These instruments use aerosol concentration as tracer for MLH. This is possible because the main sources of aerosols are situated at the surface and exchange between mixing layer and free atmosphere is limited. Subsequently, high aerosol concentrations are associated with the mixing layer and low aerosol concentrations with the free atmosphere. Tracking MLH can be challenging in complex cases, such as under the presence of residual layers of air pollution or during periods of fog or rain. Various algorithms have been proposed, but a common feature in these algorithms is that the techniques are applied to each timestep individually, possibly resulting in incoherent estimates of MLH i.e. with unrealistic jumps in time.

Based on the gradual development of MLH, a new method called ‘Pathfinder’ is presented. It applies graph theory to evaluate multiple timesteps simultaneously, leading to a more consistent and gradual estimate of MLH. With the use of Dijkstra’s shortest path algorithm (Dijkstra, 1959) the evolution of the MLH is tracked combining as much strong gradients as possible, given that subsequent estimates are within a certain vertical range of each other. The search is guided by restrictions based on features like clouds and residual layers. Excellent agreement is found with windprofiler retrievals for a 12-day period in 2008 (R^2=0.90) and visual judgment of lidar data during a full year (R^2=0.96). We find that the new approach outperforms other backscatter lidar based-algorithms, especially during morning development and afternoon transition of MLH.