This paper describes a large-eddy simulation based chemical transport model, developed under the OpenFOAM framework, implemented to simulate dispersion and chemical transformation of nitrogen oxides from traffic sources in an idealized street canyon. The dynamics of the model, in terms of mean velocity and turbulent fluctuation, are evaluated using available stationary measurements. A transient model run using a photostationary reaction mechanism for nitrogen oxides and ozone subsequently follows, where non-stationary conditions for meteorology, background concentrations, and traffic emissions are applied over a 24 h period, using regional model data and measurements obtained for the city of Berlin in July 2014. Diurnal variations of pollutant concentrations indicate dependence on emission levels, background concentrations, and solar state. Comparison of vertical and horizontal profiles with corresponding stationary model runs at select times show that while there are only slight differences in velocity magnitude, visible changes in primary and secondary flow structures can be observed. In addition, temporal variations in diurnal profile and cumulative species concentration result in significant deviations in computed pollutant concentrations between transient and stationary model runs.

The study of dispersion and chemical transformation of pollutants at
urban scales

In 2017, about 10 % of all air quality monitoring stations among
the EU-28 have reported

As the basic geometrical unit in a typical urban environment, a street canyon
consists of two rows of buildings separated by a road section.
The aspect ratio between the building height (

The Reynolds-averaged Navier–Stokes (RANS) equations are a popular numerical
framework for modeling turbulent dynamics due to its simplicity and
correspondingly low computational cost. While the RANS approach has been shown
to adequately represent mean flow behaviors in some cases

Further, a number of studies have considered the coupling of dynamics
with reaction kinetics relevant to urban traffic emissions. One such
simple mechanism is the photostationary NO–

The present study seeks to extend the current corpus on idealized street
canyon modeling in two aspects. First, previous studies have
been conducted in stationary meteorological and pollutant conditions,
inasmuch transient studies are carried out under statistically steady wind
and constant emission rates

Schematic of the solution domain representing the infinite street
canyon geometry from

As illustrated in Fig.

Initially, the dynamics of this model framework will be evaluated
using the stationary laser Doppler anemometry (LDA) measurements of

The prognostic equations employed in the present model will be
formulated using a weakly compressible framework. This generalizes
the theoretical foundation and opens up possibilities of direct
coupling with other compressible regional models such as WRF in
the future. These equations form the dynamic core of the
modeling framework, to which the conservation equations for mass,
momentum, energy, and species are to be solved numerically to
determine the spatial and temporal distribution and evolution of the
prognostic variables, namely, wind velocity components (

Due to the computational effort required to directly resolve
all possible turbulent scales in any flow of practical interest,
some form of statistical treatment must be adopted.
In the LES model approach, this is realized by
decomposing the instantaneous scalar variable

As with Reynolds averaging for time-averaged scalars, additional
terms will result from the filtering operation of Eqs. (

Numerous approaches are available for modeling the SGS terms

In the current study, the photostationary NO–

Reaction rate parameters for the photostationary
NO-

In the current study, air is modeled as a mixture of its primary
constituents, namely

The open-source finite-volume computational continuum mechanics framework
OpenFOAM

Additional components have also been developed to operate in conjunction
with

Grid distribution of the canyon geometry in the

The solution domain is discretized using the OpenFOAM

The numerical schemes used in the discretization of the
prognostic equations are second-order accurate in time and space.
A colocated grid approach is used, in which pressure–momentum coupling
is achieved using a fourth-order interpolation scheme

For the purpose of evaluating the LES street canyon model,
a stationary run is performed for which LDA flow measurement
data are available from

To coincide with the LDA measurements, sampling points for the
LES simulations are located along vertical stations at

Normalized vertical profiles of the of resolved mean horizontal
(top left) and vertical (top right) velocity components, as well
as horizontal (bottom left) and vertical (bottom right) velocity
fluctuations, for stations located at

Figure

Line integral convolution representation of the velocity vector
along the

The mean flow field in the canyon along the lateral plane of symmetry
has been obtained by averaging the instantaneous velocity field at
interval of 300 s for over the 3600 s sampling period for the fine mesh
configuration. It is presented in Fig.

Although there is little mass transfer
between the bulk flow and the canyon in the skimming flow regime, as indicated by the RANS
vertical velocity component (

Fraction of resolved scale TKE of total (resolved and SGS) TKE for the three stations in the street canyon. Solid lines indicate fine mesh, and dashed lines indicate coarse mesh.

To further evaluate the resolution of turbulent scales under each mesh
configuration, the fraction of TKE as its simulated total – i.e.,
the sum of resolved and SGS TKE – is calculated along the three vertical
stations, as presented in Fig.

The transient study consists of two sets of model runs. The first
model set is a fully transient run, where boundary conditions are
prescribed using the diurnal profiles for meteorological conditions,
solar state, background concentrations, and traffic emissions.
For the second set, a series of stationary runs are conducted with boundary
conditions fixed at 08:00, 12:00, 16:00, and 20:00 UTC. Each run begins
6 h prior to the start of the data sampling period and is conducted
using the fine mesh arrangement (core cell size 0.5 m) and a
constant time step of 0.05 s. Vertical profile data are extracted at

Prescribed conditions for the transient simulation:

Figure

The profiles for background concentrations of NO,

All runs are conducted with four cluster nodes, each comprising
two Intel Xeon Platinum 9242 processors and 384 Gb of
physical memory. Based on OpenFOAM run statistics, each simulation
hour of model run requires on average

Comparison of LIC representation of hourly mean velocity vector
along the

Figure

Comparison of vertical station profiles of mean velocity components and species concentrations for transient and stationary model runs at various points of time in UTC. Solid lines indicate transient runs, and dashed lines indicate stationary runs.

However, deviations of up to about 20 ppb in pollutant concentrations
can be observed in the vertical profiles in Fig.

Two observations can be made from this comparison. First, to the extent of the meteorological conditions considered in this study, there is little change in the in-canyon flow field between stationary and non-stationary runs. This is not surprising, considering that the shear layer in the roof region attenuates the dynamics inside the canyon from those above the buildings. Accordingly, the in-canyon wind should be less sensitive to the temporal variations in the freestream flow. On the other hand, significant changes can be observed in pollutant concentrations under stationary and non-stationary treatment. As such, temporal variations in emissions and background conditions, which are not considered in stationary runs, are significant and cannot be neglected. In light of the above observations, where wind and flow distributions are of primary concern, stationary model runs would likely suffice. However, if the focus of the study is on pollutant transport and transformation, temporal variations of background concentrations and emissions will have an appreciable influence on the model concentrations, and thus a non-stationary approach should be considered.

Comparison of horizontal station profiles of mean velocity components and species concentrations at an elevation of 2 m for transient and stationary model runs at various points of time in UTC. Solid lines indicate transient runs, and dashed lines indicate stationary runs.

A cascade of concentrations for all species considered can be identified,
corresponding to the individual primary vortices seen in Fig.

In addition, there is a reciprocal relationship between the
concentrations of NO and

Comparison of horizontal station profiles of mean velocity components and
species concentrations at a normalized elevation of

Comparison of horizontal station profiles of mean velocity components and
species concentrations at a normalized elevation of

Horizontal profiles of velocity and concentrations are presented in
Figs.

The stratification of

Hourly averaged time series of NO,

Finally, Fig.

While the trend of

A weakly compressible, reactive finite-volume LES solver, based on the
OpenFOAM computational continuum mechanics framework

The results of this study demonstrate the ability of this solver in performing urban-scale chemistry transport modeling in both a stationary and non-stationary capacity in a grid-independent manner. As such, it can be readily adapted to model a larger domain with more realistic geometry, such as a city block. For models covering such domain regions, however, input data such as meteorology and emission profiles need to be prescribed at higher spatial and temporal resolutions, which can be accomplished by coupling the urban scale model with outputs from regional models, e.g., WRF-Chem. Particular emphasis can be placed on the role of absorptive, reflective, and refractive mechanisms of urban structures in the redistribution of solar radiant energy and their impact on the overall advective and convective thermal stratification in their immediate vicinity.

Further, the method of spatiotemporal distribution and

The exact version of

The supplement related to this article is available online at:

The activities outlined in this article have been conceived and
designed by ECC and TMB. ECC was responsible for the development
of

The authors declare that they have no conflict of interest.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The BLUME urban background concentration station measurements and the aggregated
annual traffic emissions data have been provided by the Senate Administration
for Environment, Traffic, and Climate Protection
(Senatsverwaltung für Umwelt, Verkehr und Klimaschutz) of the city
of Berlin. The stationary LDA measurement data of

This paper was edited by Christoph Knote and reviewed by two anonymous referees.