Urban canopy meteorological forcing and its impact on ozone and PM<sub>2.5</sub>: role of vertical turbulent transport

Abstract

Abstract. It is well known that the urban canopy (UC) layer, i.e., the layer of air corresponding to the assemblage of the buildings, roads, park, trees and other objects typical to cities, is characterized by specific meteorological conditions at city scales generally differing from those over rural surroundings. We refer to the forcing that acts on the meteorological variables over urbanized areas as the urban canopy meteorological forcing (UCMF). UCMF has multiple aspects, while one of the most studied is the generation of the urban heat island (UHI) as an excess of heat due to increased absorption and trapping of radiation in street canyons. However, enhanced drag plays important role too, reducing mean wind speeds and increasing vertical eddy mixing of pollutants. As air quality is strongly tied to meteorological conditions, the UCMF leads to modifications of air chemistry and transport of pollutants. Although it has been recognized in the last decade that the enhanced vertical mixing has a dominant role in the impact of the UCMF on air quality, very little is known about the uncertainty of vertical eddy diffusion arising from different representation in numerical models and how this uncertainty propagates to the final species concentrations as well as to the changes due to the UCMF. To bridge this knowledge gap, we set up the Regional Climate Model version 4 (RegCM4) coupled to the Comprehensive Air Quality Model with Extensions (CAMx) chemistry transport model over central Europe and designed a series of simulations to study how UC affects the vertical turbulent transport of selected pollutants through modifications of the vertical eddy diffusion coefficient (Kv) using six different methods for Kv calculation. The mean concentrations of ozone and PM2.5 in selected city canopies are analyzed. These are secondary pollutants or having secondary components, upon which turbulence acts in a much more complicated way than in the case of primary pollutants by influencing their concentrations not only directly but indirectly via precursors too. Calculations are performed over cascading domains (of 27, 9, and 3 km horizontal resolutions), which further enables to analyze the sensitivity of the numerical model to grid resolution. A number of model simulations are carried out where either urban canopies are considered or replaced by rural ones in order to isolate the UC meteorological forcing. Apart from the well-pronounced and expected impact on temperature (increases up to 2 ∘C) and wind (decreases by up to 2 ms−1), there is a strong impact on vertical eddy diffusion in all of the six Kv methods. The Kv enhancement ranges from less than 1 up to 30 m2 s−1 at the surface and from 1 to 100 m2 s−1 at higher levels depending on the methods. The largest impact is obtained for the turbulent kinetic energy (TKE)-based methods. The range of impact on the vertical eddy diffusion coefficient propagates to a range of ozone (O3) increase of 0.4 to 4 ppbv in both summer and winter (5 %–10 % relative change). In the case of PM2.5, we obtained decreases of up to 1 µg m−3 in summer and up to 2 µg m−3 in winter (up to 30 %–40 % relative change). Comparing these results to the “total-impact”, i.e., to the impact of all meteorological modifications due to UCMF, we can conclude that much of UCMF is explained by the enhanced vertical eddy diffusion, which counterbalances the opposing effects of other components of this forcing (temperature, humidity and wind). The results further show that this conclusion holds regardless of the resolution chosen and in both the warm and cold parts of the year.

The “urban meteorology island”: a multi-model ensemble analysis

“…The significance of urban-induced alterations of urban canopy air temperature, surface temperature and PBL height has been well documented in many previous simulation and observation-based studies (e.g. Huszar et al, 2014;Trusilova et al, 2016;Göndöcs et al, 2017;Karlický et al, 2018;Huszár et al, 2018;Huszar et al, 2018;Huszar et al, 2020a, b). The moisture flux decrease in cities is also a well-known phenomenon as already shown by Oke (1987) or more recently by Theeuwes et al (2019) from latent-heatflux comparison.…”

“…The impact of cities on boundary-layer structure was already documented many years ago (Godowitch et al, 1985;Oke, 1987;Angevine et al, 2003), similarly to the impact on wind flow (Oke, 1987;Klein et al, 2001;Droste et al, 2018). These modifications of atmospheric dynamics over cities have considerable consequences for the mixing and dispersion of pollutants and air quality in urban areas, which has also been confirmed by model studies investigating the urban-induced changes in dispersion conditions (Karlický et al, 2018) and in primary or secondary pollutant concentrations (Fallmann et al, 2016;Huszár et al, 2018;Huszar et al, 2018;Li et al, 2019;Huszar et al, 2020a, b;Ďoubalová et al, 2020).…”

“…This also leads to great precipitation biases and a slight temperature overprediction in winter. The remaining RegCM simulations show smaller temperature biases in summer, even less than those of Huszár et al (2018), who used a very similar model setup. Huszar et al (2020b) used the model configuration corresponding to RU simulation (i.e.…”

“…Alterations of the temperature, wind speed and boundarylayer height in urban areas and their diurnal cycles have been described in detail in many previous studies (e.g. Huszar et al, 2014;Huszár et al, 2018;Karlický et al, 2018;Huszar et al, 2020a, b); therefore other components of the UMI will be discussed in more detail in the following paragraphs. In Fig.…”

Error induced by neglecting subgrid chemical segregation due to inefficient turbulent mixing in regional chemical-transport models in urban environments

“…The reaction between tracers A and B can then be considered a pseudo first-order reaction, and the corresponding production term kAB ≈ k A (plotted in the top panel of Fig. 2) is dependent only on the concentration of tracer A, where k = kB is the pseudo first-order rate coefficient (Hobbs, 2000;Jacobson and Jacobson, 2005). In this situation, tracer A is said to be the limiting reactant, and the chemical reaction between tracers A and B is tracer-A limiting (Zumdahl, 1992).…”