Air quality pollution from traffic and point sources in Skopje assessed with different air pollution models

Transcription

1 International Scientific Conference Air quality pollution from traffic and point sources in Skopje assessed with different air pollution models Kosta Mitreski, Martina Toceva, Nikola Koteli, Ljupco Karajanovski Faculty of Computer Science and Engineering Ss. Cyril and Methodius University in Skopje, Skopje, Republic of Macedonia 2 Department for Eco-informatics

2 INTRODUCTION The modelling of the air pollution is one of the techniques that can lead to better planning and making decisions from the authorities on the local, regional or even global scale. Air quality dispersion models can be used to provide information about the impact of individual emission sources and to predict air quality as a result of changes in emissions, such as increase of traffic, emission control measures, etc. Dispersion models can be used to complement the data gained by monitoring as the spatial coverage of air quality information provided by monitoring is often limited.

3 INTRODUCTION Air dispersion modelling could be used to estimate and predict the concentration of the pollutants in air using mainly emission and meteorological data. Air dispersion models include mathematical algorithms based on combination of physical and chemical parameters so that they can simulate the spread of pollutants in the air as well as the complex processes of air pollution creation. In this way, it will be possible to develop a plan to reduce the environmental pollution and satisfy the EU environmental air quality standards.

4 INTRODUCTION In this paper, based on combination of few existing air pollution models, we will present, as a main contribution of this paper, the first this-kind of study for the city of Skopje. Our system based on real time air pollution data acquisition is easy extendible to national and transboundary levels. At the same time, this is the first step in building the real time decision (not only prediction) support system.

5 Our case study During the winter periods in Skopje and other locations in the country, the air quality periodically becomes very poor. Pollutant concentrations, especially of particulate matter (PM10 and PM2.5) can reach very high levels, occasionally even 10 times higher than the limit values which are set for protection of human health. The high pollutant concentrations originate mainly from local emission sources such as traffic, household heating and industry and are enhanced by the local topography together with stable meteorological conditions. We have conducted a case study by using two local scale traffic dispersion models, one local scale point source model and one regional air quality model.

6 1. Street canyon Operational Street Pollution Model (OSPM) In this study, we have combined and adopted many of the technics and technologies to develop a robust, long-term system for traffic air pollution monitoring with extensive number of experimental data. The real time Air quality monitoring data in city of Skopje is composed of eight air quality measurement stations.

7 Real-time visualization The process of visualization is based on the data as the average of the air parameter concentration per hour for each parameter. The data are provided by the network and stored in the our server database. They are used for generation of the grid raster layers by interpolation technique. The layers are color-coded and scaled from green to red depending of the concentration level. Then the geo- referenced interpolated concentration layers are automatically published as World Mapping Service (WMS) and publicly available.

8 Modeling part On the basis of the traffic data from one of the main streets in the city of Skopje we generated our modeled data charts. The needed and collected data of air parameters is modeled in the duration of an entire year. These data contain information for temperature, speed and direction of wind, global solar radiation, and relative humidity, and also background concentrations of PM10, O3, CO, NO2, and PM2.5. Emission information and data of the type of fuel used in transport and their composition in terms of substances important for pollution is taken (provided) from the largest supplier of fuel in the Republic of Macedonia, MAKPETROL.

9 Modeling part Traffic information and data on the number of vehicles moving on the street are classified by the type of vehicle: bus, car, van, truck etc. Data for the distribution of vehicles in terms of volume and type of engine and motor fuel are provided from city corresponding institutions. Information and data of the average speed of vehicles that travel along the street is taken to be 50 km / h (actual speed limit for the road). Street configuration information and data about the height and placement of buildings along the 'Partizanski Odredi' street canyon are presented in Figure

10 Modeling part Figure 2. Part of the street 'Partizanski Odredi' together with the configuration of the buildings

11 Experimental work with model and results In the model, we define three modeling scenarios about the flow of the vehicles. Each of these three scenarios has two sub scenarios (about the mean speed of the vehicles), resulting in total of 6 as follows. Scenario 1.1 (32000 vehicles/24h, 50 km/h) Scenario 1.2 (32000 vehicles/24h, 80 km/h) Scenario 2.1 (64000 vehicles/24h, 50 km/h) Scenario 2.2 (64000 vehicles/24h, 80 km/h) Scenario 3.1 (96000 vehicles/24h, 50 km/h) Scenario 2.2 (96000 vehicles/24h, 80 km/h) Results from the applied modeling for the above six scenarios are presented as charts on Figure 3.

12 Experimental work with model and results

13 Experimental work with model and results

14 Experimental work with model and results

15 Experimental work with model and results

16 Experimental work with model and results The benefit of this kind of the system for monitoring, modeling and visualization of the traffic air pollution of the city of Skopje is the realtime aspect of the system.

17 2. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model This modeling is processed using HYSPLIT model (Draxler and Hess 1998) which is applied in three locations in Skopje. The model is used for calculation of the concentration of PM10 particles. The HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model is complete system for computing simple air parcel trajectories to complex dispersion and deposition simulations. Air concentration calculations require the definition of the pollutant's emissions and physical characteristics. The model requires specially preformatted meteorological data. The data can be created manually or can be downloaded via FTP. In this research they are created manually.

18 Model results For this research, we are implementing HYSPLIT model for simulating PM10 particles in the ambient air in 3 locations in Skopje set up in Lisice, Centar and Rektorat where three are of the State air quality monitoring stations are located. The meteorological data set used in the model was taken from the Ministry of Environment and Physical Planning from these three stations two months (November and December 2012), to represent the winter period. The emission data was taken from the cement factory Cementarnica USJE.

19 Model results After configuring the model, the simulation output is presented as contours. HYSPLIT model as output takes pictures for every day in the month.

20 3. Point source and traffic modeling in Skopje using the local air quality dispersion models UDM-FMI and CAR-FMI The UDM-FMI is a local scale dispersion model developed at the Finnish Meteorological Institute. The dispersion model is based on Gaussian plume equations for various stationary source categories (point, area and volume sources). CAR-FMI is dispersion model for an open road network. The model computes an hourly time-series of the pollutant dispersion for CO, NO, NO2, NOx and exhaust PM2.5 concentrations

21 UDM-FMI and CAR-FMI data The meteorological time-series for both models is computed by the meteorological pre-processing model (MPPFMI), developed at the FMI (Karppinen et al., 1997). The background concentrations of gaseous compounds are interpolated from the measurements of the monitoring network of FMI, while the background concentrations of fine particulate matter can be estimated, if local measurements are not available. The traffic data was collected by the Sector of Traffic of the City of Skopje.

22 UDM-FMI and CAR-FMI data Figure 6. Daily traffic amounts in Skopje city center.

23 UDM-FMI and CAR-FMI data 1600 Modelled SO2 & NO2 emissions in Skopje (t/year) SO2 annual emissions NO2 annual emissions Figure 7.The annual SO2 and NOx emissions from major point sources in Skopje for 2008.

24 UDM-FMI and CAR-FMI RESULTS Figure 7. The result of the dispersion modeling calculations, the NO2 and CO annual average concentrations caused by the traffic emissions in Skopje city center

25 UDM-FMI and CAR-FMI RESULTS Comparison of the modeled against measured NO2 and CO annual average concentrations was done in one point representing the location of the air quality measurement station Rektorat. The analysis shows that the annual average NO2 and CO concentrations measured in the monitoring station compare well with the modeled ones. 500 Annual average CO concentration Annual average CO concentration Modelled (whole study area max.) Modelled (Rektorat) Measured (Rektorat 2006) Measured (Rektorat 2007)

26 Point sources The NO2 concentrations from the point source emissions in Skopje are very low compared to the concentrations caused by traffic emissions. The annual average is within the range of µg/m3. The highest SO2 concentrations caused by the point sources in Skopje appear in the vicinity of the energy production plant Toplifikacija Zapad. The average annual SO2 concentration from major point sources over the study area is within the range of µg/m3. The critical value for protection of vegetation for the annual average concentrations of SO2 is 20 µg/m3 and the modeled concentrations do not exceed this.

27 Point sources Figure 10. Modelled NO2 and SO2 annual average concentration from point sources in Skopje

28 Point sources Comparison between the modeled and measured SO2 concentrations was done in four points. The analysis shows that the annual average NO2 and CO concentrations measured in the monitoring station Rektorat compare well with the modeled ones.

29 Conclusions and recommendations This modeling study was done in order to assess the air quality impact of major emission sources in Skopje by using different dispersion models. Two different local scale dispersion models were used to assess the impact to air quality caused by emissions from traffic in Skopje. The main traffic-related pollutants are CO, NOX, hydrocarbons, and particles.

30 Conclusion The idea with this kind of system is to raise the public awareness and to help the city planners and regulatory institutions to get information about traffic-induced noise pollution. In the future, we plan to provide a general traffic noise pollution model and according to the input data from different measurement points and sensors to provide more precise and complete data for any city and larger region around the city.

31 Conclusion - Traffic in Skopje The OSPM model study was conducted in order to assess the air quality alongside a canyon type of street. From the experimental results we can conclude that the recommended limit values for health protection according to the latest EU Air quality standards (50 μg/m3 for pm10 daily average, 25 μg/m3 for pm2.5 yearly average) are significantly exceeded for PM10 and PM2.5 According to the dispersion modeling calculations with the CAR-FMI model, the traffic in Skopje is the biggest contributor to the NO2 concentrations with the highest concentrations occurring along the major roads and crossroads. The CO concentrations from traffic occur also near the biggest roads and crossroads.

32 Conclusion - Point source modeling Point sources can be estimated the main contributor to SO2 concentrations in Skopje. According to the modeling results, the limit values or critical level are not exceeded. The modeling results include some uncertainties -unreliability of the meteorological data -limited availability of good quality emission data.

33 Conclusion?