The regional air pollution study in Lithuania provided a comprehensive overview of air quality in Lithuania (in Vilnius (capital) and rest of territory) when 375 monitoring sites at different representative locations (urban, suburban, and residential) were equipped with diffusion samplers. The samples were analyzed for sulfur dioxide (SO2) and nitrogen dioxide (NO2) concentration. The measurement results show that the mean concentrations of SO2 in all investigation sites during the study period did not exceed the annual limit value of 20.0
During the past 20 years, there has been a marked improvement of the air in Europe [
Nowadays, the main goals of monitoring lie in providing useful up-to-date information to the public on pollutant concentrations in ambient air, as well as supporting economical stakeholders and decision makers in air-quality assessment and management. Instruments for air quality may change in complexity and cost. While air pollution is highest in urban zones, the monitoring efforts are typically concentrated in cities, and little sites represent the background level. The financial resources are not equal in different countries, and there are no possibilities to extend monitoring network or upgrade of equipment. The use of passive samplers greatly reduces the cost and the need of long-term measurement programs [
Monitoring of air pollution in Lithuania is organized by the Environmental Protection Agency. Currently, Lithuanian national air-monitoring network consists of one mobile, fourteen continuously operating urban stations, and three integrated monitoring (IM) stations. European Union (EU) environment law acts and legislation were applied and implemented by the National Environmental Monitoring Program (NEMP).
In this paper, within the framework of “Lithuanian Air Monitoring System Modernization Using Diffusive Samplers” (LAQMO) project for the first time were evaluated the concentrations of SO2 and NO2 by determining the ambient concentration using the passive sampling method at 375 sites in Lithuania. The spatial maps of compounds using geographical information systems (GIS) were evaluated on one year measurements with diffusive air samplers.
The analysis of concentration for SO2 and NO2 using diffusive samplers were set up in the urban background (residential), semiurban (mixed residential and commercial), and roadside (busy street/road and crossing) sites in order to get spatial variation in pollutants concentrations. The obtained data were compared with the acceptable levels of air pollutants that are adopted in the EU as the limit values (Table
Atmospheric air quality (
SO2 | NO2 | |
---|---|---|
Annual limit value (LV) | 20 (vegetation) | 40 (human health) |
Upper assessment threshold (UAT) | 12 | 32 |
Lower assessment threshold (LAT) | 8 | 26 |
The most appropriate sites for placement were determined. For purposes of taking into account the influence of weather conditions, a network of 375 passive samplers was deployed for all four seasons: autumn (September–November) 2010, winter (December–February), spring (March–May), and summer (June–August) of 2011 and were covered in 8 measurement periods (Table
The measurement periods.
Season | Period | From | To |
---|---|---|---|
I autumn | 1 | 5.11.2010 | 17.11.2010 |
2 | 17.11.2010 | 1.12.2010 | |
|
|||
II winter | 3 | 6.1.2011 | 20.1.2011 |
4 | 20.1.2011 | 3.2.2011 | |
|
|||
III spring | 5 | 25.2.2011 | 8.4.2011 |
6 | 8.4.2011 | 22.4.2011 | |
|
|||
IV summer | 7 | 6.6.2011 | 20.6.2011 |
8 | 20.6.2011 | 4.7.2011 |
10% of the sampler was in duplicates, i.e., some colocated passive samplers were deployed at the sampling sites with available continuous monitors for cross correlation and calibration purposes [
Uncertainty in measurements.
Analyte | Limit value ( |
Standard deviation | |
---|---|---|---|
NO2 | Annual mean | 40 | 3.9 |
UAT | 32 | 4.8 | |
LAT | 26 | 5.9 | |
|
|||
SO2 | Annual mean | 20 | 5.4 |
UAT | 12 | 8.9 | |
LAT | 8 | 13.4 |
Eight sampling campaigns of 14 days were carried out in Vilnius agglomeration and zone (the rest part of Lithuania). The locations of the monitoring sites in Vilnius and zone selected for the passive sampling is shown in Figure
Location of diffusive samplers.
Passive samplers deployed in different city sites were collected after 14 days of exposure time intervals. The passive samplers were provided and analyzed by PASSAM AG (Switzerland). As supplied by the firm, the tubes are protected from sunlight by an opaque cylindrical box. These samples have been exposed to sites with different sources of atmospheric emission and environments (Section
Although several approaches to uncertainty evaluation exist, the indirect approach of GUM published by the ISO was used (Table
Uncertainty estimation according to GUM.
Component | Limit value ( |
Uncertainty combined | Uncertainty expanded |
|
|
---|---|---|---|---|---|
NO2 | Annual mean | 40 | 10.8 | 21.6 | 7.7 |
UAT | 32 | 10.2 | 9.9 | 7.3 | |
LAT | 26 | 20.5 | 19.7 | 7 | |
|
|||||
SO2 | Annual mean | 20 | 11 | 22.1 | 7.8 |
UAT | 12 | 13.1 | 26.2 | 9.3 | |
LAT | 8 | 16.7 | 33.5 | 11.9 | |
|
|||||
C6H6 | Annual mean | 5 | 14.1 | 28.2 | 10 |
UAT | 3.5 | 17.3 | 34.6 | 14.1 | |
LAT | 2 | 26.6 | 53.3 | 18.9 |
An additional term has been introduced, which covers the uncertainties budgets of repeated measurements, microenvironmental factors, variations in the geometry of samplers, etc.
The combined uncertainty
The expanded uncertainty is calculated by using a coverage factor of 2:
The uncertainty of the mean of the 8 periods is calculated as follows:
Maps of the pollutant concentrations over the area were obtained by interpolation of the passive sampler measurements. By using custom-made automated scripts on open source GRASS GIS software (version 6.4), the following geostatistical methods commonly used for surface interpolation from randomly sampled points: inverse distance squared weighting (IDW; GRASS function v.surf.idw), bicubic spline interpolation (BCS; GRASS function v.surf.bspline), and kriging interpolation with automated calibration of parameters (AK; GRASS function v.krige) were tested [
By comparing statistical variability of the interpolated datasets, it became obvious that with increasing search radius (N of neighboring points used in interpolation), the IDW method produced rather unstable results, the BCS method under similar conditions (increasing length of splines) produced clearly predictable results with a slight tendency of statistical “smoothing” of the interpolated grid, while AK indicated the most stable statistical results due to its ability to autocorrelate all measurements in the sample [
In order to streamline the process of geostatistical data analysis and operational mapping, a customized Linux shell script was developed. It uses geostatistical and mapping functions of the open source GRASS GIS software (v.surf.bspline), as well as some of the Linux OS libraries (libgdal, libgeotiff, libpng, etc.) to automatically generate geostatistical grids and operational maps by iterating over each of the polygon objects (urban areas, etc.) by using standard samples of coordinated measurement points as an input. Geostatistical grids will be created in GRASS GIS environment with 10 m pixel size in the standard LKS94 CRS and masked with boundaries of the urban areas. They will be exported from GRASS database as Float64 data type rasters in GeoTIF file format without any associated color table [
The SO2 passive samplers were exposed for periods of 2 weeks each at a time over the study period (120 samples). The values of passive samplers for SO2 ranged between approximately 0.7 and 1.8
Seasonal variation of mean SO2 concentrations for the entire study period from 3 November 2010 to 4 July 2011 (bar lines show ±22.1% expanded uncertainty).
Seasonal variation of mean SO2 concentrations for the entire study period.
During fall, the mean SO2 concentration had the highest level (up to 1.80
Data indicate that at sites in the residential and recreation areas, the higher SO2 levels were recorded in autumn, winter, and spring, when the emissions from energy production are at their highest level (Figure
Annual mean concentrations of SO2 in Vilnius.
Conversely, the lowest SO2 levels were measured in the summer period. Therefore, the seasonal variability of concentrations should be interpreted using existing knowledge on emission and meteorological patterns. In summary, the mean sulfur dioxide concentration in Vilnius ranged from 0.2 to 3.1
The obtained data (35 sites) during all the study period revealed that NO2 concentrations varied considerably, which coincides with the other study depending on the distance of the measurement site from main roads caused by the large numbers of vehicles releasing NO2 [
Seasonal variation of mean nitrogen dioxide concentrations at site-specific areas for the entire study period from 3 November 2010 to 4 July 2011 (bar lines show ±21.6% expanded uncertainty). (a) Vilnius, transport. (b) Vilnius, residential. (c) Vilnius, suburban.
Annual mean concentrations of NO2 in Vilnius agglomeration for the period from 3 November 2010 to 4 July 2011.
The mean concentrations of NO2 in 40 zones’ territory sites, during the study period did not exceed the annual limit value of 40.0
Annual mean concentrations of NO2 in the zone (58 cities) for the period from 3 November 2010 to 4 July 2011.
The mean concentrations of SO2 in 40 zones’ territory sites during the study period did not exceed the annual limit value of 20.0
Annual mean concentrations of SO2 in the zone (58 cities) for the period from 3 November 2010 to 4 July 2011.
During the summer (6 June–4 July 2011), SO2 concentrations ranged between 0.20 and 2.10
The results of the 2004-2005 and 2010-2011 campaign in Figure
Time series of mean SO2 concentration.
The level of nitrogen dioxide concentrations has decreased by 34, 26, 24, and 49% during five years in the city of Vilnius at the sites next to traffic. Also the increase of NO2 concentration was observed at Žirnių street and at the crossroads of V. Kudirkos Street near Pamenkalnis (Figure
Time series of mean NO2 concentration.
Concentrations of SO2 and NO2 were determined over a year using the passive sampling method. For the entire study period (from 3 November 2010 to 4 July 2011), the annual mean concentrations of SO2 ranged between 0.20 and 3.40
The data used to support the findings of this study are available from the corresponding author upon request.
The authors declare that there are no conflicts of interest regarding the paper.
This research was supported by EPA of Lithuania (Contract No. 4F10-1010).