Measurement campaign focused on meteorology and air quality to support the validation of the PALM LES model in Prague

The LES based modelling system PALM has been widely extended during recent years. A lot of the development focused on the processes needed for simulation of complex urban environments. Besides individual validations of the newly developed processes, an evaluation of the complete modelling system in a real urban environment is necessary to ensure the reliability of modelling results. Such comprehensive validation requires the construction of additional monitoring networks within the areas of interest to obtain sufficient data support.

Our earlier campaigns, e.g. Resler et.al. (2017, 2020), focused mainly on validation of the energy exchange related processes in urban area. As part of the currently running international TURBAN project, a measurement campaign was implemented in the center of Prague, Czech Republic. It focuses mainly on evaluation of the street level meteorological and air quality dynamical processes by utilisation of a specially built sensor network placed in heavily polluted street canyons (traffic-related pollution). Twenty air quality sensors were complemented by a doppler lidar and microwave radiometer profile observations, and by observations from permanent meteorological and air quality monitoring stations operated by the Czech Hydrometeorological Institute (CHMI). These data were compared with PALM simulations performed for selected episodes of the year. The PALM model was configured in two nested domains with resolution of 10 m / 2 m and the extent of 8×8 km / 1.2×1.6 km.

A significant challenge to the measurement campaign was the low accuracy and reliability of air quality sensor observations. The differences in the values of individual sensors as well as their deviations from the reference observations reached tens of percent. Using sufficiently long co-measurement with the reference monitoring station and advanced statistical methods, sufficiently accurate and reliable data suitable for model evaluation were obtained. To ensure long-term quality control of the observations, two selected sensors were co-located with a continuous reference CHMI station within the simulation domain. Similarly, setting up a profile measurement (especially with Doppler LIDAR) to obtain the maximum possible information in a given space and processing a large amount of data was a challenge. The presentation shows the details of the observations and their processing as well as the preliminary results of the model evaluation.