Evaluation of high-resolution GRAMM/GRAL NO<sub>x</sub> simulations over the city of Zurich, Switzerland

Abstract

Abstract. Hourly NOx concentrations were simulated for the city of Zurich, Switzerland, at 10 m resolution for the years 2013–2014. The simulations were generated with the nested mesoscale meteorology and microscale dispersion model system GRAMM/GRAL (versions v15.12/v14.8) by applying a catalogue-based approach. This approach was specifically designed to enable long-term city-wide building-resolving simulations with affordable computation costs. It relies on a discrete set of possible weather situations and corresponding steady-state flow and dispersion patterns that are pre-computed and then matched hourly with actual meteorological observations. The modelling system was comprehensively evaluated using eight sites continuously monitoring NOx concentrations and 65 passive samplers measuring NO2 concentrations on a 2-weekly basis all over the city. The system was demonstrated to fulfil the European Commission standards for air pollution modelling at nearly all sites. The average spatial distribution was very well represented, despite a general tendency to overestimating the observed concentrations, possibly due to a crude representation of traffic-induced turbulence. The temporal variability of concentrations explained by varying emissions and weather situations was accurately reproduced on different time scales. The seasonal cycle of concentrations, mostly driven by stronger vertical dispersion in summer than in winter, was very well captured in the two year simulation period. Short-term events, such as episodes of particularly high and low concentrations, were detected in most cases by the system, although some unrealistic pollution peaks were occasionally generated, pointing at some limitations of the steady-state approximation. The different patterns of the diurnal cycle of concentrations observed in the city were generally well captured as well. The evaluation confirmed the adequacy of the catalogue-based approach in the context of city scale air pollution modelling. The ability to reproduce not only the spatial gradients but also the hourly temporal variability over multiple years makes the model system particularly suitable for investigating individualized air pollution exposure in the city.

Identifying urban air pollution hot-spots by dispersion modeling when data are scarce: application to diesel generators in Beirut, Lebanon

“…Unstable, and slightly unstable atmospheric stability classes were taken in the simulations. The choice of these specific stability classes is based on the Pasquill table [18]. The different scenarios are summarized in Table 2 for the small domain and in Table 3 for the large scale domain.…”

“…Fig. 13-g shows a plot of the normalized fist floor proxy indicator versus floor number for these four scenarios (13,18, wind speeds of 4 and 1 m/s and stack heights of 2 meters above the roof top (scenarios 13 and 29) and 4 meters above the ground level (scenarios 18 and 30). The plot shows that, for wind speed of 4 m/s, moving the stacks from street level to above roof tops significantly reduces the exposure of residents of all floors.…”

“…The modeling system GRAMM (Graz mesoscale model ))-GRAL (Graz Lagrangian model (Oettl, 2018)) has been used to simulate the levels of various pollutants originating from different sources over entire cities in Europe. It was used for evaluation at the urban and meso scales in Zurich, Laussane (Switzerland), Vienna and Graz (Austria) and Po Valley (Italy) [17,18,19,20,21,22]. Other studies used GRAMM-GRAL to simulate pollution at specific road tunnel portals, highways on complex terrains, specific streets intersections and climate change mitigation in urban mobility [23,24,25,26].…”

“…Both levels aim to prove the compliance of this model with the standard quality objectives and performance criteria. This was accomplished by different studies in the European countries and the model was found to be in compliance with the German Guideline VDI 3783 (2015) [27,28] and to be fulfilling the criteria as defined by the Forum for Air quality Modelling in Europe (FAIRMODE) [18]. Another study, conducted in Australia, subjected GRAMM-GRAL to a model-to-data comparison as well as a model-to-model comparison.…”