Aim of research

The main objective of our contribution is the development and the validation of models for the urban atmosphere and the urban canopy, allowing to predict the pollutant distribution close to ground level - and also at various levels within the canopy layer - and the air quality of the different quarters of a city. This new generation of models will allow to describe source-receptor relationships within the urban area, taking into account the pollutant transfers and transformation processes within the urban canopy. After validation, these improved canopy-accounting models will eventually be included in air quality management systems, allowing to convolute pollutant distributions with other functions, such as inhabitants residence times, etc., to evaluate statistical or targeted impacts.

Within streets the flows and the turbulent fields depend largely on the strongly non-linear interactions between street geometry and micro-meteorology. As a consequence, the pollutant concentration fields within the streets and at roof level are poorly assessed, which means - among other points - that the spatial representativeness of sensors measuring proximity pollution levels within streets is unknown, especially since vehicle pollutants transform rapidly during their dispersion within the streets themselves. The ECN group is testing the turbulence models in the CFD code CHENSI, and developing the methodology and procedures to realise efficient numerical simulations of representative three-dimensional situations at acceptable computing costs. With such simulations, it explores the structure of the flow and dispersion in these situations, especially the dynamic-thermal interactions. This heuristic exploration will result in catalogues of flow situations and enhanced parameterisations for street air quality operational models.

The French communal model SUBMESO is being developed for simulating the flows, the turbulent fields, the physics and microphysics, and the transport-diffusion-transformations of reactive pollutants within an urban area. The first version has been constructed around the dynamics and micro-physics modules of ARPS 3.2 (C.A.P.S., University of Oklahoma). The system now includes modules for the dynamic equation solver, microphysics, meteorology pre-processing, terrain and soil model (incl. urban canopy), LES sub-gridscale models, turbulence closures, and chemistry (transport-diffusion-transformation). This last module is fed by the tropospheric chemistry model MOCA. The system is equipped with a grid generator especially designed for complex terrains.

The developments within ECN’s and LEGI’s group mainly concern the dynamics and thermodynamics (turbulence, urbanised soil/canopy model, meteorological pre-processor, domain nesting). The emphasis is on the critical situations that are most favourable to the development of pollution episodes: it is therefore important to simulate correctly the turbulent fields in the situations of low winds with high - but not free - thermal convection generated by the urban "heat island" (and its own inhomogeneity), and in the situations of stable stratification when the canopy inhomogeneities can locally de-stabilise the flows. In these situations the turbulent diffusivity parameterisations used in mesoscale models are thought unable to capture the effects of the sub-mesoscale inhomogenities.

With the existing computers, in the simulations of a part or a whole city, the urban canopy details cannot be explicitly simulated. The canopy must be replaced by an artificial ground composed of a patchwork of surfaces equivalent to the real quarters. This implies that the real momentum, heat and scalar fluxes are modelled in equivalent "soil models" furnishing the conditions at the lower boundary of the computational domain. With the approach developed for canopy simulations, using boundary conditions derived from simulations at sub-mesoscales, numerical experiments are run for representative geometries of European city quarters in order to derive the fluxes at the canopy-atmosphere boundary. These simulations with CHENSI are used to adapt and extend wall laws and soil model parameterisations to the various parts of the cities, from the suburbs to the city centres. These flux parameterisations must be validated by comparison with experimental data obtained at all scales, from one street to a quarter, and to a set of quarters.
 
 

Activities during the year

The activities of the groups (ECN and LEGI) in 1998 concern the following points.

-1- Achievement of the version 2.0 of the French communal model SUBMESO with full compatibility of dynamics, chemistry and microphysics modules, full portability on workstations and supercomputers.

-2- Validation tests of novel turbulence closure models for the stable atmospheric boundary layer with active turbulence (Abart & Sini, 1998).

-3- Development of the rationale for building up a canopy layer model for SUBMESO, urbanisation of the submeso soil model SM2 (see below).

-4- Launch of co-operations within MOD 1 / TRAPOS network, especially those concerning comparisons of numerical model simulations with wind tunnel measurements.

-5- Construction of a new international co-operative program URBCAP - Assessment of the importance of urban canopy processes - including individual street-scale experiments, a large co-operativr street- and city-scale experiment, and numerical simulation demonstrations.

Principal results and main conclusions

Based on simplified computations of heat and water transfers within the soil and vegetation, soil models allow to compute the interactions between the ground and the lowest atmosphere, and to define the boundary conditions at the lower boundary of the computation domain of ABL models. In urban areas the canopy layer appears as a buffer layer between the real surface and the apparent surface at roof level. The rationale adopted for the French communal model SUBMESO is to transform the original “force-restore” soil model (Noilhan & Planton, 1992) into an urban canopy layer model (see figures). In a first stage the model is equipped with urban surface parameterisations, but it still considers a flat interface. In the second stage the model will take into account the horizontal and vertical transfer processes through the canopy air layer: complex radiative transfers, mixed convection, fast chemistry.

For the first stage, a series of new semi-empirical equations for the water/moisture circulation and for the radiative/heat transfers have been created to simulate the partial coverage of ground by artificial materials (tar or concrete) and by building roofs (tiles, slates, or concrete). Still the height of the building is not considered and no horizontal exchanges are parameterised except water run-offs (Guilloteau et al. 1998, Guilloteau, 1999).

The second stage is prepared by three types of process studies driven in parallel with specific numerical models: roughness modelling and mapping, canopy ventilation and transfers to the atmosphere, radiative transfers.

The model parameterisations of the roughness parameters as a function of building heights and arrangements have been partially validated by the site measurements of Grimmond & Oke (1998); numerical experiments are run with the small-scale model CHENSI where the flow is simulated, with 0.5 m resolution, within the streets of an idealised uniform quarter and in the constant flux internal boundary layer that develops over it (Guilloteau, 1999). In parallel, the conversational software ASTUCE has been developed in co-operation with the CNRS research federation “Physic & Images of the City” to draw maps of the roughness parameters from the statistical analysis of National Geographic Institute urban data bases.

The model CHENSI is also used in heuristic studies of the ventilation flows within streets and to infer traffic pollutant concentration distributions, residence times, and canopy-atmosphere fluxes. The study focuses on the influence of street geometry, micro-meteorological conditions, and wall temperatures (wind-induced and thermally-driven convections may either add up or counteract each other). This study is driven in close co-operation with TRAPOS groups working with wind tunnel simulations and with operational street air quality models.

A third study is to produce new parameterisations for the canopy layer model, concerning the radiative and heat budgets within the canopy: it is mainly driven by our colleagues from the CERMA with the specially developed model SOLENE of radiative and heat transfers taking into account multiple reflections, mutual shadowing, and anthropogenic heat sources. The study concerns also novel definitions of “surface parameters” for the urban canopy (e.g., surface temperature) and state-of-the-art computational algorithms (Guilloteau, 1998).

Aim for the coming year

In order to validate and/or evaluate street air-quality models like CHENSI and canopy-accounting models such as SUBMESO, the co-operative programme URBCAP has been launched, in close connection with the European laboratory network TRAPOS. It includes three parts :

- analysis of several current and future experiments at the street scale, especially in Copenhagen, St Petersburg, and Nantes;

- a large co-operative experiment at the street-scale and at the city-scale in the city centre of Marseilles during the regional-scale experimental campaign of the French program ESCOMPTE;

- nested simulation exercises in order to assess the physical and chemical transformations of the traffic pollutants emitted by the vehicles to the urban atmosphere at city scale, and to demonstrate the ability of canopy-accounting models to compute the spatial distribution of the pollutant concentrations within the city.

The URBCAP preparatory experiment Nantes ‘99 will be run in one busy street of Nantes city centre in June 1999, focusing primarily on evaluating the influence - on pollutant transport, diffusion, and transformations - of (1) thermal convection induced by wall radiative heating and (2) the motion of vehicles.

The URBCAP Marseilles 2000 experiment will see co-operation of 10 - 20 European groups to document the pollutant residence times and chemical transformations in a street of Marseilles city centre and at several points within the city during the Berre-Marseille 2000 photochemical experiment of the French program ESCOMPTE. URBCAP Marseilles 2000 aims at assessing street-scale processes involved in Ozone episodes and at extending the ESCOMPTE mesoscale database to the city scale of Marseilles urban area.

References ( ^* bibliographical production in 1998)