Title of Example

Example

Introduction

During the winter and spring months in Norway poor air quality can occur during conditions of light winds and strong stability. Poor air quality is generally traffic related so for this reason the Norwegian Public Road Association, in conjunction with the Norwegian Institute for Air Research and the Norwegian Meteorological Institute, have instigated air quality forecasts for 5 Norwegian cities. These are used for dissemination of information to the public and for planning abatement strategies and health warnings. The forecasts are currently made daily for the pollutants NO2 and PM10 over a 2 day forecast period in the cities of Oslo, Drammen, Bergen, Stavanger and Grenland.

Forecast methodology

The forecasts are made based on meteorological forecast input. ECMWF forecasts are used to drive the HIRLAM limited area model, which in turn is used to drive a nested version of MM5, down to a resolution of 1 km. Meteorological forecast data is then used in the AirQUIS modelling system to produce hourly concentration fields for the coming 2 day period. AirQUIS uses line, point and gridded dispersion models to produce surface level concentrations at a resolution of 1km and at predefined receptor points. An emission database for traffic, wood burning and industrial processes is integral to the model. Rural stations and climatological data are used for background and Ozone values in the model.

AQ forecasts are made during the evening and are available by 06:00 every day. A two day period is forecast to insure that there is time to plan any measures that may be taken on the second day.

Presentation of results

Both the current air quality, from measurements, and the forecast air quality are presented on a web site http://www.luftkvalitet.info/ (Norwegian only) for all 5 cities.

In order to present the forecasts and measurements the different pollutants are divided into 4 categories for each of the pollutants, figure 1., and the most polluted category is reported as the air quality.

Figure 1. Air quality index levels used in the forecast and analysis

The air quality is shown for all 5 cities for the next 2 days based on the index scheme described in figure 1. These can be viewed in a summarized form as a table, figure 2, or for each individual city as hourly values for the current day, Figure 3.

Figure 2. Example of the summary table of air quality forecasts presented on the web site

Figure 3. Example of todays hourly prediction for Oslo shown on the web page.

The same web site also furnishes current and archived (3 months) monitoring data for all measured compounds. In addition to the web portal it is also possible to receive email and SMS messages concerning current and forecasted air quality.

Applications

The air quality forecasts are used for public dissemination and for the health authorities to issue warnings for particular risk groups. The forecasts are also used to plan short term abatement strategies. These strategies are limited to speed controls for predicted poor air quality days.

Accuracy of the forecasts

Yearly reports are produced to access the accuracy of the forecasts. This is accomplished by comparing monitoring data with the locally predicted air quality, i.e. model results at the monitoring stations. The accuracy varies significantly from station to station and from city to city. In Oslo, for example, during the 2003/2004 winter season the percentage of correctly predicted polluted episodes at individual monitoring stations varied from 12% to 77% for the two highest polluted classes of air quality. The most significant variable affecting accuracy of the forecast is considered to be the quality of the meteorological prediction. In general the forecast system is considered to be a useful tool.

Last Updated

13th January 2005