Air mass back trajectories in the evaluation of environmental data
Air mass back trajectories were used to interpret levels of persistent organic pollutants (POPs) measured during sampling campaigns run by RECETOX at the regional observatory Košetice, Czech Republic. The theory of receptor modeling and back trajectory interpretation, as well as a detailed description of statistical methods used in the below mentioned studies.
Regional observatory Košetice
Data obtained at this station describe the contamination of the environment by POPs at a regional (central European) and background level. It has to be stressed, that central Europe is a densely populated region and pollutant background levels are higher here than at background stations in e.g. Arctic areas far from pollutant sources. Nevertheless, they are still well under legislation limits and of course lower than in areas affected by significant sources, e.g. cities, industrial areas and hot spots.
The regional observatory Košetice in the Pelhřimov district (a more detailed description of the site)
The Košetice station is run by the Czech Hydrometeorological Institute (CHMI) and a part of EMEP (Environmental Monitoring and Evaluation Programme) and GAW (Global Atmosphere Watch) as well as numerous other national and international monitoring programmes (Holoubek et al., 2003 and 2007; Váňa et al., 2007). A continuous monitoring of the POP content in selected environmental media has been performed in Košetice since 1988 and the same sampling protocol and unchanged sampling and analytical techniques have been employed since 1996. This unique data set can be evaluated and interpreted in various ways.
Not only the evaluation of measured concentration levels of pollutants is an important task, but also the investigation of sources, transport paths and the environmental fate of these substances. The application of air mass back trajectories to POP concentration data measured in Košetice air is described below.
Goals of the study, data and methods
The goals of the here presented case study were :
- determination of potential source areas of the studied POPs measured at the Košetice observatory,
- investigation of temporal changes in emission intensity and the location of potential source areas and
- application of two models on the same dataset and the comparison of obtained results.
The investigated pollutant data are total concentrations (i.e. the sum of gasous and particle bound concentrations) of polychlorinated biphenyls (PCBs), hexachlorobenzene (HCB) and the pesticides lindane (γ-hexachlorcyclohexane) and DDT and its metabolites (DDE and DDD). High-volume air samplers operating for 24 h were used to take samples once every week and the decadal period of 1997-2006 was the subject of the here presented study. Input parameters for the generation of air mass back trajectories are explained in the following table.
Input parameters for the generation of air mass back trajectories. The HYSPLIT model is a frequently used one and it is freely available online including the necessary meteorological input data (Draxler and Rolph, 1993). Threedimensional trajectories are currently considered to be more accurate than isobaric or isentropic trajectories (Stohl, 1998). The trajectory starting height of 500 m above ground level (AGL) is a frequently used one and reflects long-range transport. Trajectories starting during an inversion lower than 500 m altitude above ground were excluded from the analysis as well as trajectory segments above the mixing layer (the trajectory height is variable). Trajectories of a length up to 5 days are still considered to be accurate enough (Hopke et al., 1995). The generation of eight trajectories per sample was considered to yield a sufficient number of trajectories suitable for a statistical evaluation.
| Parameter | Value |
|---|---|
| Model | HYSPLIT |
| Trajectory type | 3D |
| Trajectory starting height | 500 m AGL |
| Trajectory length | 96 hours |
| Frequence of trajectory generation | 8 trajectories per sample |
Trajectory segment endpoints were all sorted onto a geographical grid encompassing the studied region. Two statistical models, e.g. the mean ground source loading, Lg (a detailed description) and the potential source contribution function, PSCF (a detailed description) were applied in each grid cell. Further, two measures for a qualitative and quantitative evaluation of the distribution of Lg values were applied:
- The relative contributions of European countries to the pollutant levels measured in Košetice air:
where Lc is the relative load from the considered country, G the number of cells allocated to the country, n the total number of cells, w the area weight and S the country area (in km2). - Centers of gravity (COGs) of Lg values in selected geographical sectors (Lee and Wong, 2001):
where
are geographical coordinates of COGs and wi weights of individual points xi and yi. Individual points were derived using the ArcView mapping software from individual grid cells. Mean loadings Lg of corresponding cells are used as weights.
Results and discussion
Only selected results illustrating the interpretation of the applied trajectory statistics methods (receptor models) are presented here. Full results and discussion sections can be found in Dvorská et al. (2008, 2009, 2010).
PSCF values can be considered as categorial values excluding (e.g. <0,5) or indicating (e.g. >0,5) potential source areas of pollutants measured in Košetice air. On the other hand, values obtained by the computation of ground mean source loadings allow to draw a gradient of the strength of ground-based emissions influencing the passing air masses.
The application of the PSCF model to lindane and PCB 153 concentration data for 2003-05 revealed a difference in the spatial distribution of potential source areas. Contrary to PCB 153, major potential source areas of lindane can be found in France. This is in accordance with information on the lindane use in Europe. France was one of the major consumers of this pesticide in the past and its agricultural use was permitted here until 1998 (at this time, the use of lindane for agricultural purposes was prohibited in most of the other countries).
Lindan 2003-05
PCB 153 2003-05
The computation of ground mean source loadings for lindane and PCB 153 in two periods revealed a decrease in their concentrations transported by air to Košetice. This is in accordance with the general decrease of these pollutants levels in the environment. The light blue star depicts the position of reference COGs, the dark blue star depicts the position of COGs relevant to the respective pollutant.
Lindan 1997-99
Lindane 2004-06
PCB 153 1997-99
PCB153 2004-06
Directions and lengths (km) of COGs shifts in individual sectors between 1997-99 and 2004-06. Reference shifts reflect differences in air flow patterns in the two studied periods, while shifts of Lg COGs refer to changes in the location of potential source areas. The most significant change in the spatial distribution of potential source areas occured for lindane in the western sector.
| Substance | Western sector | Northern sector |
|---|---|---|
| Lindan | N 151 | NW 37 |
| PCB 153 | N 86 | W 24 |
| Reference | N 83 | W 28 |
Relative contributions of individual European countries and groups of countries to the concentrations of lindane and HCB measured in Košetice air in the two periods. Countries contributing less than 3% to concentrations measured in Košetice fall into the category Remaining countries. Circle area proportions reflect the change of substances source loadings between both periods. The concentration of lindane transported by air to Košetice decreased between the two studied periods. For HCB, the opposite trend was found (also, HCB concentrations rose in Košetice air until 2006, see Holoubek et al., 2007).
gHCH 1997-99
gHCH 2004-06
HCB 1997-99
HCB 2004-06
The comparison of PSCF and source loading calculation results allows for the identification of potential source areas of various intensities. Using PSCF, e.g. parts of central France were identified as potential source areas of HCB. However, this region is characterized by lower Lg values than e.g. parts of Germany. Thus, central France is suggested to be a potential source area of low intensity for HCB in Košetice air.
HCB 2003-05
HCB 2003-05
Main conclusions
- The concentration of all studied pollutants transported by air to Košetice decreased between 1997 and 2006. Only HCB emissions increased, which could be at least partly explained by major floods in the Danube river-basin in 2002.
- Major emissions of lindane (France), technical HCH (western Poland, Hungary and northern ex-Yugoslavia) and DDT (Czech Republic) found in 1997-99 decreased later.
- Prominent emissions of DDT and HCB in the northern Czech Republic and PCBs and HCB in Germany remained in 2004-06.
- The combination of PSCF and source loading calculations allows for the identification of source areas of low and moderate intensity.
References
Dvorská, A., Lammel, G., Klánová, J., Holoubek, I., 2008. Kosetice, Czech Republic – Ten years of air pollution monitoring and four years of evaluating the origin of persistent organic pollutants. Environmental Pollution 156, 403-408.
Dvorská A., Lammel G., Holoubek I., 2009a. Recent trends of persistent organic pollutants in air in central Europe - Air monitoring in combination with air mass trajectory statistics as a tool to study the effectivity of regional chemical policy. Atmospheric Environment 43, 1280-1287.
Dvorská A., Lammel G., Holoubek I., 2009b. Long-range transport of persistent organic pollutants to Košetice observatory. In: Váňa, M., Holoubek, I., et al. 20 years of Košetice Observatory - Part 2. ČHMÚ, Praha.
Draxler, R.R., Rolph, G.D., 2003. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY Website (http://www.arl.noaa.gov/ready/hysplit4.html). NOAA Air Resources Laboratory, Silver Springs, USA.
Lee, J., Wong, D.W.S., 2001. Statistical analysis with ArcView GIS, Wiley, New York, 192 pp.
Holoubek, I., Brörström-Lundén, E., Duyzer, J., Shatalov, V., Klánová, J., Kohoutek, J., 2003. Regional trends of POPs in European ambient air. Organohalogen Compounds 61, 518 521.
Holoubek, I., Klánová, J., Jarkovský, J., Kohoutek, J., 2007. Trends in background levels of persistent organic pollutants at Kosetice observatory, Czech Republic. Part I. Ambient air and wet deposition 1988-2005. Journal of Environmental Monitoring 9, 557-563.
Hopke P.K., Li C.L., Cizek W., Landsberger S., 2005. The use of bootstrapping to estimate conditional probability fields for source locations of airborne pollutants. Chemometrics and Intelligent Laboratory Systems 30, 69-79.
Stohl A., 1998. Computation, accuracy and applications of trajectories – a review and bibliography. Atmospheric Environment 32, 647-966.
Váňa, M., Pekárek, J., Červenková, J., Čech, J., Machálek, P., Janouch, M., Holoubek, I., Macoun, J., Horálek, J., Rychlík, S., Hnilicová, H., Klánová, J., Jarkovský, J., Kohoutek, J., Helešic, J., Šeda, Z., Kubík, V., Dvorská, A., Světlík, I., Rulík, P., Molnár, H., Tomášková, L., 2007. Košetice observatory – 20 years. ČHMÚ, Praha, 151 p., ISBN 978 80 86690 46 9.
