The Air Quality Information System, AirQUIS, is a professional management tool for Air Quality developed by Norwegian Institute for Air Research, NILU , together with GIS-experts.

The combination of on-line data collection, statistical evaluations and numerical modelling enable the user to obtain information, carry out forecasting and future planning of air quality. The system can be used for monitoring and to estimate environmental impacts from planned measures to reduce air pollution.

The AirQUIS system contains the following modules:
· On-line measurement system
· Emission inventory database
· Atmospheric dispersion models
· Exposure
· Geographical information system
· Air quality planning

Models in AirQUIS
In the modern multi modular environmental information system like ENSIS/AirQUIS steps have been taken to establish models for air pollution dispersion to enable environmental impact assessment estimates. Models are essential when the programmes are to be used for planning purposes. The models need as input data some background information on:
· source characteristics and emission data,
· area characteristics (surface roughness, topography etc.),
· measurement data (measurement type, heights etc.),
· meteorological data (wind, stability, mixing height, temperatures etc.),
· dispersion coefficients (type to be used and parameters),
· dry and wet removal coefficients,
· location of receptor points (distances or grid specifications).

· A Eulerian type numerical dispersion model is the EPISODE model developed by NILU. It represents the most applied model in AirQUIS today. This is a time-dependent finite difference model normally operating in three vertical levels, combined with a sub grid line source model for traffic and a puff trajectory model for industry stations to account for subgrid effects close to individual sources. The wind field used as input to the model may be homogeneous or inhomogeneous for each time step dependent upon the meteorological input data available.

Model Description

Eulerian grid model
The Eulerian part of the EPISODE model consists of the numerical solution of the atmospheric (mass) conservation equation of the pollutant species in a three-dimensional Eulerian grid. Horizontal advection is calculated numerically by using a positive definite and 2D monotone version of the Bott scheme using 4th degree polynomials (Bott, 1993).Horizontal diffusion is calculated by using a simple 2D explicit scheme. Vertical advection is calculated using an upwind scheme.The vertical component of the wind is generated internally in the model based on the horizontal wind-components and a 3D divergence free condition for the wind field. Vertical diffusion (turbulent exchange of mass) is described by K-theory, i.e. first order closure, and this term is calculated using a simple explicit scheme as well. Monin-Obukhov similarity theory is applied in order to parameterize the vertical eddy diffusion coefficient, K (Shir, 1973; Businger and Arya, 1974)Horizontal and vertical turbulence parameters may be calculated internally in the model using an advanced meteorological preprocessor (MEPDIM) based on atmospheric boundary layer similarity theory (Bøhler, 1996).

Lagrangian sub grid models
The Lagrangian part of the EPISODE model consists of two separate subgrid-models for each of the two main subgrid source categories: line- and point-sources. The subgrid models are directed specifically towards detailed subgrid scale calculations of spatial concentrations distributions close to the main road network and individual point sources. The line source model is an are integrated Gaussian type models, while the point source model is a Gaussian segmented plume trajectory model (Grønskei et al., 1993). The sigma-schemes used in the point source model are based on formulations by Irwin (1983) and Venkatram (1984) while in the line source model these schemes are based on the use of Pasquill-Gifford stability classes.

Photochemical reactions
The photochemical scheme for NO2 is based on a standard photochemical equilibrium model using the ordinary fast reaction rates between NO, NO2 and O3. The photo dissociation process is calculated based on the amount of solar radiation with automatic calculation of sun height above the horizon.

Capacity and Limitations
There is no predefined upper limit as to the number of horizontal or vertical grid cells in the model, as such. The only restrictions lies in upper limits (MX, MY and MZ) as set by NILU when compiling the model as a DLL (Dynamic Linked Library) as part of the AirQUIS system. These upper limits can be set to any number greater than 1 depending on the size of the available computer memory (RAM memory). The execution time will typically increase about linearly with the total number of actual grid cells (NX*NY*NZ). There is no predefined lower or upper limit to the horizontal or vertical grid resolution used. Typical grid resolution horizontally is 1 km, but the model has been run with smaller values such as 500 m or even 50-100 m. Vertical grid resolution is typically 10-20 m close to ground and 200-500 m further up. The time step used by the model is typically between 10 and 300 seconds depending on the wind speed, turbulence conditions, and horizontal and vertical grid resolution.

Evaluation and Validation
The model is part of existing on-line surveillance systems for Oslo and Grenland (Norway), Stockholm in addition to three cities in China. It has recently been applied in urban air quality studies for several Norwegian cities (Slørdal and Walker, 1997). It has also been applied in various health studies linking short term health effects with air pollution conditions in Oslo and Grenland (Norway) (Bartonova et al., 1997). The EPISODE model has been evaluated by comparing the model calculated concentrations of NOx and NO2 with measurements at eight measurement stations in Oslo during a previous comprehensive measurement campaign in Oslo in 1991-92 (Larssen et al., 1994).

In general it was found a good agreement between the model calculated and the measured values, with correlation values for NOx (NO2) ranging between 0.48 and 0.80 based on hourly data for a four month period (November-February 1991-92). It has also recently been evaluated in the four cities Oslo, Drammen, Bergen and Trondheim in Norway (Slørdal and Walker, 1997).The model is also described on the web as part of The Model Documentation System established by the European Topic Centre on Air Quality (ETC-AQ) with the aim to provide guidance to model users in the selection of the most appropriate model for a specified application.

For more information on Air Quality Management visit: www.nilu.no/aqm.