Introduction
Air quality concerns typical of contemporary urban environment are
nitrogen oxides and volatile organic compounds in busy city centres and
ozone as a secondary pollutant in outskirts [1]. In addition, higher
carbon monoxide concentrations are common, but usually not reaching the
maximal permitted limits. Sulphur dioxide has been an urgent air quality
problem in many countries, nowadays often related with marine ship
routes [2]. The exceptionally high (but within the limit values) hourly
average S[O.sub.2] concentrations about 60-70 [micro]g/[m.sup.3] are
measured in Estonia, near the passenger port of Tallinn [3]. Remarkable
quantities of [H.sub.2]S are emitted also from fuel oil terminals [4],
animal farms [5] and municipal waste treatment sites [6].
Unlike most European urban municipalities, the Estonian oil shale
processing center Kohtla-Jarve has often to deal with enhanced
concentrations of [H.sub.2]S and S[O.sub.2] in the ambient air,
[H.sub.2]S sometimes exceeding the regulatory limit level 8
[micro]g/[m.sup.3]. Three main enterprises known as remarkable emission
sources of sulphuric compounds are governed by Viru Chemistry Group
(VKG)--the oil-shale retorting plant of VKG Oil (shale oil production
unit and extensive semi-coke dumps, identified as important sources of
[H.sub.2]S [7]), thermal power stations of VKG Energia (oil-shale and
generator gas combustion are the significant sources of S[O.sub.2]), and
the municipal wastewater treatment (WT) plant that treats both
residential and oil shale processing sulphur-rich wastewater. Sulphide
formation in the 16 km long pressure pipe (ongoing wastewater from the
industrial clients), where the wastewater retention time is 24 hours,
has been investigated [8].
To date the shale oil production processes implemented in Estonia
are the Kiviter process (particle size 25-125 mm) and the Galoter
process (particle size < 25 mm) [9]. The new plant Petroter which is
based on the Galoter process started to operate recently.
Production of oil from lumpy oil shale in vertical retort, the
Kiviter process, results in large amounts of solid waste--semicoke. Per
tonne of oil shale 0.49 tonnes of semicoke are formed in vertical
retorts. Semicoke contains a considerable amount of sulphur (1.7-2.1%)
in different forms [10], since in the retorting process of Estonian oil
shale, more than 50% of sulphur in the raw oil shale remains in the
solid residue afterwards [11]. Figure 1 illustrates the air emissions
from the retorting process. The maximal emission quotes according to air
pollution permission are indicated. Thus, the figures do not correspond
exactly to any particular year discussed below, but to the upper limit
for permitted emissions.
In the past as well as in present, besides shale oil, several
chemical products have been produced: bitumen, coke, phenols, phenolic
products, etc.
Besides emissions from controlled industrial processes, there are
landfill emissions. The surface area of the semi-coke landfill is 142
ha, plus the area needed for treating the wastewater. Approximately 73
million tons of industrial waste has been deposited in this landfill at
present. Semi-coke and oil shale ash are not the only two types of
industrial waste being deposited in the landfills. However, a variety of
the types of waste used to be wider--fuses (pitch waste) and acid tar
(originating from the production processes of benzene and toluene),
waste containing sulphur and arsenic (so-called sulphur sludge,
originating from the refinement of the generator gases, where
[As.sub.2][O.sub.3] sodium carbonate solution was used), waste
containing mineral oil, building rubble, and consumer waste, originating
from the activities of the enterprises, have been deposited in the
industrial waste landfill. In addition, hazardous wastes from AS
Velsicol and AS Nitrofert have been deposited in the industrial waste
landfill over the years. The former landfill for industrial waste has
also been used for depositing building and demolishing rubble (including
wood wastes), sediments of waste water, street sweepings, and probably
also domestic waste [12].
[FIGURE 1 OMITTED]
There is a permanent air quality monitoring station at Kalevi
street located a few kilometres down the dominating western and
south-western winds from both VKG Oil and wastewater treatment plant
(Fig. 2). Despite extensive measurements both in the town and at
industrial territories, the origins of certain air pollution episodes
still remain obscured. Narrow peaks of elevated concentrations with
nearly south-western wind directions are mentioned through several years
[13, 14]--roughly originating from the direction where the industrial
enterprises are situated, but certain directions (azimuthal angles
approximately 205-210[degrees] and 220-230[degrees]) do not correspond
exactly to any of known pollution sources.
This paper aims to identify the origins of periodically high
[H.sub.2]S and S[O.sub.2] concentrations at Kalevi monitoring station.
For that the coinciding wind directions at different locations and
heights are analysed. The meteorological (weather forecast) modeling
results are applied as a supplement to scarce local measurement data of
vertical wind spread. The sonic anemometer measurement data are used to
gain information about the impact of atmospheric boundary layer
properties on the local pollutant transport. In addition, it is tested
to which extent the improvements of technological processes in VKG have
reduced the levels of sulphuric pollutants in the air of the town of
Kohtla-Jarve.
[FIGURE 2 OMITTED]
Data and methods
Air quality monitoring data originate from Kalevi monitoring
station in Kohtla-Jarve and mobile station Mobair operated by Estonian
Environment Research on the territory of VKG Oil in November
2005-January 2006. Meteorological data were gathered from monitoring
stations, Aseri meteorological mast operated by Estonian Environment
Research (see Fig. 2), and meteorological (weather forecast) modeling
results. The meteorological modeling results, applied for 2005/2006
winter episode only to compensate the lack of measurements higher above
the ground, originate from model ETB HIRLAM operated routinely by the
Estonian Institute of Meteorology and Hydrology since January 2006 and
BaltAn65+ reanalysis [15] (November-December 2005). The automated
monitoring station in Kohtla-Jarve is situated at Kalevi street
(geographical coordinates 59[degrees]24'35" N and
27[degrees]16'43" E). This station meets the requirements of
the European Environmental Agency and belongs to European monitoring net
EUROAIRNET [16]. The measurements of pollutant concentrations are
provided all year round, representing the hourly mean concentrations.
Mobair is a mobile unit for air monitoring that was installed in
Kohtla-Jarve near the main entrance to VKG AS production area during
29.11.2005-17.01.2006.
From air monitoring stations and Mobair only the low-level (about 3
m) wind speeds, directions, atmospheric pressure and temperature are
available. Additionally there is a metrological mast in Aseri (located
29.5 km to northwest from Kalevi monitoring station) measuring
temperatures at 8 and 20 m, wind speeds and directions at 10 m and 24 m
levels and the vertical wind dispersions. The meteorological model data
are pre-processed with the air quality model SILAM [17]. The data
include meteorological parameters at several levels (for instance, 15,
55, 130, and 305 m): wind speeds and directions, turbulent
diffusivities, etc.
The emissions according to Estonian national statistics are given
in Fig. 3. Data is gathered from registers of air emissions and
permitted emission quotes of waste-water treatment plants by Estonian
Environment Information Centre (Keskkonnateabe Keskus), requested in
August 2010. The sharply increased S[O.sub.2] emissions in 2005 are
caused by the growth of shale oil production, conditioning the increased
combustion of generator gas at thermal power plants. In the second half
of 2008 a sulphur-catching device (NID-reactor) was installed at VKG
Energia Pohja (Northern) Thermoelectric Power Plant (TPP), providing 65%
effectiveness in extracting S[O.sub.2] from exhaust gas. Hydrogen
sulphide emissions in 2007 from VKG Oil are circa 5 times less than in
2006, that is explained by reconstruction of the park of tanks and
installation of collective breathing system with absorbers. Emissions of
earlier years, up to 2006, from the wastewater treatment plant are
obviously underestimated as unrealistic emissivities were applied. About
five-fold correction was based on measurements by Estonian Environment
Research [14]. The higher emissions are applied for 2007, but earlier
ones remained untouched in statistical database, although the real
values should be close to 2007. In 2007 the regional wastewater
treatment plant was reconstructed, and together with environmental
measures by VKG Oil, that conditioned the decrease in emissions in 2008
(Fig. 3).
[FIGURE 3 OMITTED]
Analysing the observations, the horizontal mass fluxes of
[H.sub.2]S and S[O.sub.2] ([micro]g/[m.sup.2][s.sup.-1]) are used rather
than concentrations ([micro]g/[m.sup.3]). The mass flux is defined as
the mass of admixture transported through the unit surface in unit time,
i.e. concentration multiplied by wind speed. As our goal is to
understand the origins of pollution rather than to quantify the
exceedings of permitted limit values, the fluxes are preferred, as they
are stronger attributed to the pollution source strength (effect of
downwind dilution is eliminated). The flux applied here is always
calculated as the concentration multiplied by the wind speed measured in
the monitoring station, even if the wind direction applied in the
analysis originates from any higher level. Wind speed, of course,
increases with height, but usually remains well correlated with
low-level speed. Thus, we can still compare the fluxes at different time
moments. On the other hand, as concentrations are not measured at higher
levels, we have no solid ground to apply the upper-level wind speeds for
calculating the fluxes.
Results
Dependence of measured concentrations on wind direction is given in
Figures 4 and 5. Automatic monitoring of [H.sub.2]S concentrations at
Kalevi station was started in 2004. [H.sub.2]S fluxes have stably high
values with wind directions between 195-250[degrees] (i.e. nearly
southwest), but highest values are observed in a more narrow peak,
215-235[degrees]. In 2007 [H.sub.2]S peak becomes narrow and low lumped
in between 195-225[degrees], remaining practically unchanged through
2007-2009. Sulphur dioxide peak (195-225[degrees]) is clearly visible
during 2005-2009 and has a nearly constant height, presumably indicating
at a single source or compact group of sources. Thus, according to
recent measurements, both the [H.sub.2]S and S[O.sub.2] originate from
one direction, 210 [+ or -] 15[degrees], whereas earlier observations
(before reconstruction of wastewater treatment) give a stronger
[H.sub.2]S peak at 225 [+ or -] 15[degrees] without background-exceeding
levels for S[O.sub.2] from the same direction. Assuming that
210[degrees] peak originates from VKG (Energy and Oil) and 225[degrees]
peak from the wastewater treatment plant, it is easy to explain such a
pattern, as the process in WT do not emit any remarkable amount of
S[O.sub.2]. But looking at the location map (Fig. 2), we recognise that
these directions do not match with exact locations of VKG and WT: the
directions from Kalevi monitoring stations are 215-230[degrees] and
240-250[degrees], respectively. Thus, both directions are matching the
edge of peak, leaving the highest values out of range. However, the
discussed wind directions are measured at the low level in the
monitoring stations under the direct impact of underlying surface
elements, while the transport of air mass does not match exactly with
low-level wind directions, as atmospheric air is typically mixed through
much thicker layer than a few meters, when transported a few kilometres
downwind. Thus, the higher-level winds are examined.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
Comparison of fluxes plotted against the 3 m (Kalevi) and 24 m
(Aseri) wind directions during 2006-2007 (Fig. 4) shows a systematic
shift of peaks by 10-15[degrees] westwards. Remarkable is that despite
slightly larger scatter, the general structure (including two peaks of
[H.sub.2]S) is preserved. The same is valid for 2008-2009 (Fig. 5), when
only one (the southernmost) peak for [H.sub.2]S is left. It is evident
that these directions fit much better with positions of expected sources
in respect to Kalevi station: 210-230[degrees] for VKG and
230-250[degrees] for the wastewater treatment plant. The fluxes
presented at graphs are highly smoothed, using a moving average over
wind directions. The single measurements often give values several times
higher.
The direction-dependent (local wind direction in Kalevi)
concentration distributions during the winter episode
29.11.2005-17.01.2006 are given in Fig. 6. Like during entire long-term
monitoring series, the fluxes of S[O.sub.2] form a narrow peak around
205[degrees] and [H.sub.2]S has more scattered high-level range,
possibly with two peaks, at 195-240[degrees] (each data point presents
an hourly value without any directional averaging, thus single maximum
values are much higher than in Figures 4 and 5). The Aseri 24 m wind
directions give expected Ekman shift of the peak by 13[degrees]
clockwise with standard deviation 11[degrees]. Meteorological model
results give systematically clockwise turning wind with height as well:
by 1, 4, 9 and 23[degrees] for 15, 55, 130 and 305 meter levels,
respectively. But their scatter is rather large: standard deviations
ranging from 15 to 28[degrees].
In Mobair station the high levels occur between 170 and
260[degrees] (south to south-west-west) with two sharp peaks at
170-180[degrees] and 225-240[degrees]. The production units of VKG are
situated in south-south-western to western directions or
200-270[degrees] from this station (Fig. 2). As the buildings of VKG in
these directions, immediately surrounding the Mobair station, are
several times higher, severe and complicated distortions of wind field
at this site are expected. Thus, it is quite evident that these highly
elevated levels originate from production units of VKG Oil, the
S[O.sub.2] sources are shale oil distillation and electrode coke
production units, and possibly, from VKG Energia Louna (Southern)
Thermoelectric Power Plant (TPP) as well. [H.sub.2]S emissions from
200-270[degrees] probably originate from gas-generator plants and
electrode-coke unit. Much lower peak of [H.sub.2]S from north-western to
northern directions may originate from the wastewater treatment plant.
[FIGURE 6 OMITTED]
Now we have fixed the directions, where the high levels of
sulphurous pollutants come from. To examine further the atmospheric
conditions that favor the highest levels, we leave out all wind
directions as "noise" not containing information about local
sources, except the range of 195-240[degrees], and analyse the
micrometeorological data from the Aseri mast in view of S[O.sub.2] and
[H.sub.2]S fluxes.
In Fig. 7 there are given the monthly average thermal
stratifications at lower (8-2 m) and upper (24-8 m) layer at night and
day (to include more statistics, wind direction segregation is not
applied here; but 195-240[degrees] are proven similar in general). At
night-time the potential temperature is increasing with height,
indicating prevailing thermal inversions that are extremely strong in
summertime, when daily temperature range is larger. In the daytime in
summer the unstable stratification (decreasing potential temperature) is
prevailing due to solar heating in the lowermost layer, whereas the
upper layer is rather neutrally stratified. All the daytime and
night-time fluxes, especially of S[O.sub.2], are larger in winter and
decrease in summer (Fig. 8). The gap for summertime night S[O.sub.2] is
very deep--that may be conditioned by near ground inversions.
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
Correlations between Aseri vertical temperature gradients and
sulphuric pollutant fluxes at Kalevi station are negative (i.e.
inversion conditions tend to make less pollution transported at low air
level), but small and insignificant. The vertical wind standard
deviations correlate better with fluxes, especially for S[O.sub.2]
(correlation coefficient +0.43 at 10 m level). The positive correlations
indicate that stronger vertical mixing (i.e. normally unstable,
non-inversion atmosphere) induces more intense pollution transport near
the ground.
Discussion
In this paper we have reviewed the main sources of [H.sub.2]S and
S[O.sub.2] pollution and meteorological conditions at Kohtla-Jarve,
aiming to reveal the circumstances of their dependence. The atmospheric
transport and diffusion of pollutants is influenced by a wide range of
air circulations, thus it is complicated to match the exact emission
source only on the basis of low-level meteorological data, which are
strongly influenced by underling surface.
We have studied flux patterns with wind directions measured at 2-24
m heights, but also with those calculated for weather forecast purposes
higher above. Wind directions at Kalevi monitoring station are quite
different from those measured simultaneously at Aseri meteorological
mast (Fig. 9). In addition to the systematic clockwise turn with height,
there exist local "bends" within several ranges of wind
directions. Wind shear due to surface friction produces an ageostrophic
component of the wind towards lower pressure. This is a result of
balance between pressure gradient force, the Coriolis force, and the
viscous force generally seen as turning of the wind in the boundary
layer. According to the Ekman spiral, at the surface the wind is turned
for 45[degrees] counter-clockwise compared to its direction in free
troposphere [18]. In the real, vertically inhomogeneous atmosphere the
direction changes usually less. It seems that the meteorological model
HIRLAM, responsible for modeled boundary-layer wind directions rather
than air quality model SILAM, is overestimating the wind shear angle,
when downscaling from free troposphere to the surface layer. Thus, the
wind at levels lower than 100 m is systematically turned
counter-clockwise with respect to direct measurements.
[FIGURE 9 OMITTED]
Even two stations measuring at the same height and laying only
three kilometres apart (Fig. 10) give strickingly different wind
directions. The expected reasons are bending of wind trajectories around
the semicoke landfill and in much smaller scale, around the buildings
(in location of Mobair station in particular). The effects of hills on
wind field are known both in theory and measurement. Theoretical
calculations with a model of atmospheric dynamics suggest significant
wavelike disturbances a few kilometres downwind and also in lateral
direction from 100-300 m high (horizontal dimension a few kilometres)
hills [19].
Due to both the height-dependent and local effects, airborne
pollutants should not necessarily originate from this exact direction,
where the wind is blowing from. Detailed examination of multi-level wind
measurement and also meteorological modeling data give us better
imagination, where the pollutants come from. Topographically the
dispersion of pollutants in the study area is determined by semicoke
dumps, considerable industrial buildings and woodland belts, acting as
microscale obstacles which form wakes.
The evaluation of correlations between different level wind
directions was performed with available measured meteorological
parameters. It was found that the airborne fluxes of sulphur dioxide
from local sources in Kohtla-Jarve (most probably VKG Oil) tend to be
larger in intensely mixing atmosphere than under stagnant conditions.
This may refer to the dispersion from elevated sources, from which
emissions reach the ground easier in case of a well-mixed surface layer.
The vertical wind standard deviations measured at the Aseri
meteorological mast, which correlate with S[O.sub.2] concentrations, are
a governing parameter of atmospheric dispersion [20]. Relations of
[H.sub.2]S fluxes with stratification parameters remain rather weak,
thus no clear conclusions on its releases can be drawn. Nevertheless, no
evidence of highly elevated sources, in contrary to S[O.sub.2], is
found--in accordance with the hypothesis of wastewater treatment and
semicoke dump sources.
[FIGURE 10 OMITTED]
Conclusions
The main achievement of the study is an empirical assessment of the
effects of local- and microscale obstacles and atmospheric processes on
diffusion of [H.sub.2]S and S[O.sub.2] in the town of Kohtla-Jarve.
1. The most enhanced sulphur dioxide and hydrogen sulphide levels
at Kohtla-Jarve (Kalevi) monitoring station originate from the direction
of VKG Oil production units, semicoke dumps and thermal power plants of
VKG Energia. Much higher concentrations measured at the territory of VKG
Oil during MobAir episodic measurements suggest rather efficient
dilution in the atmosphere when transported towards the town.
2. The secondary (but the most intense before 2008) peak of
hydrogen sulphide originates from direction of the municipal wastewater
treatment plant.
3. Sulphur dioxide measured at Kalevi station probably originates
rather from elevated than ground-level sources, e.g. power plant stacks
and industrial production units with height in tens of meters. Hydrogen
sulphide, in contrary, may originate from lower-level sources.
4. Evidently the reconstruction of technological processes in both
VKG units and the wastewater treatment plant have reduced the levels of
sulphuric pollutants in the air of the town of Kohtla-Jarve.
DOI: 10.3176/oil.2011.2.07
Acknowledgements
This research was supported by the Estonian National Targeted
Financing Project SF0180038s08 and the Estonian Science Foundation
Grants 7005 and 8795.
Received August 23, 2010
REFERENCES
[1.] Colls, J. Air Pollution.--London: Spon Press, 2002, 560 p.
[2.] Matthias, V., Bewersdorff, I., Aulinger, A., Quante, M. The
contribution of ship emissions to air pollution in the North Sea regions
// Environ. Pollut. 2010. Vol. 158, No. 6. P 2241-2250.
[3.] Kort, M., Teinemaa, E., Kesanurm, K. Monitoring of ambient air
2009. Estonian Environmental Research Centre. Tallinn, 2010. 123 p. [in
Estonian].
[4.] Pauklin, T., Teinemaa, E. Air pollution investigations in
Maardu town and its surroundings.--Estonian Environmental Research
Centre. Tallinn, 2007. 120 p. [in Estonian].
[5.] Blunden, J., Aneja, V., Westerman, P. W. Measurement and
analysis of ammonia and hydrogen sulphide emissions from a mechanically
ventilated swine confinement building in North Carolina // Atmos. Envir.
2008. Vol. 42, No. 14. P. 3315-3331.
[6.] Saral, A., Demir, S., Yildiz, S. Assessment of odorous VOCs
released from a main MSW landfill site in Istanbul-Turkey via a
modelling approach // J. Hazard. Mater. 2009. Vol. 168, No. 1. P.
338-345.
[7.] Preparations to closing of industrial waste and semicoke dumps
in Kohtla-Jarve and Kivioli. Environmental impact assessment report
2003/EE/16/P/PA/012. Maves As, Tallinn, 2006. 95 p. [in Estonian].
[8.] Karmas, K., Tenno, T., Hellat, K., Formation of sulphide and
its chemical equilibrium in sewage pipes. Influence of [H.sub.2]S to
aerobic microorganisms of activated sludge // Oil Shale. 2004. Vol. 21,
No. 4. P. 309-319.
[9.] Yefmov, V. Oil shale processing in Estonia and Russia // Oil
Shale. 2000. Vol. 17, No. 4. P. 367-385.
[10.] Environmental hazard assessment of semi-coke.--Estonian
Environmental Research Centre Report. Tallinn, 2003 [in Estonian].
[11.] Molder, L., Elenurm, A., Tamvelius, H. Sulphur compounds in a
hydraulic ash disposal system // Proc. Estonian Acad. Sci. Chem. 1995.
Vol. 44, No. 2/3. P. 207-211.
[12.] Closing down of industrial waste and semi-coke landfills in
Kohtla-Jarve and Kivioli. Environmental impact assessment report
2003/EE/16/P/PA/012// Ramboll Finland Oy, 2006. 55 p.
[13.] Pauklin, T., Teinemaa, E., Investigations of ambient air in
Ida-Virumaa. Estonian Environmental Research Centre, Tallinn, 2006. 65
p. [in Estonian].
[14.] Pauklin, T., Teinemaa, E., Kesanurm, K. Investigations of
ambient air in IdaVirumaa. Stage II.--Estonian Environmental Research
Centre, Tallinn, 2007. 114 p. [in Estonian].
[15.] Luhamaa, A., Kimmel, K., Mannik, A., Room, R. High resolution
re-analysis for the Baltic Sea region during 1965-2005 period // Clim.
Dynam. 2011. Vol. 36, No. 3-4. P. 727-738.
[16.] Pauklin, T., Teinemaa, E., Kesanurm, K. Monitoring of ambient
air in towns. Quarter I, 2010.--Estonian Environmental Research Centre,
Tallinn, 2010. 28 p. [in Estonian].
[17.] Sofiev, M., Siljamo, P., Valkama, I., Ilvonen, M. Kukkonen,
J. A dispersion modelling system SILAM and its evaluation against ETEX
data // Atmos. Envi. 2006. Vol. 40, No. 4. P. 674-685.
[18.] Nieuwstadt, F. T. M, van Dop, H. (eds.). Atmospheric
Turbulence and Air Pollution Modelling.--Dordrecht: Reidel Publishers,
Dordrecht, 1982. 358 p.
[19.] Room, R., Zirk, M. An efficient solution method for
buoyancy-wave equation at variable wind and temperature // Mon. Weather
Rev. 2007. Vol. 135, No. 10. P. 3633-3641.
[20.] Gryning, S. E., Holtslag, A. A. M., Irwin, J. S., Sivertsen,
B. Applied dispersion modelling based on meteorological scaling
parameters // Atmos. Envir. 1987. Vol. 21, No. 1. P. 79-89.
Presented by A. Raukas
J. PAVLENKOVA (a), M. KAASIK (b) *, E.-S. KERNER (b), A. LOOT (b),
R. OTS (b)
(a) Maetaguse Rural Municipality Government 41301 Maetaguse,
Ida-Virumaa, Estonia
(b) Institute of Physics, University of Tartu Ulikooli Str. 18,
50090 Tartu, Estonia
* Corresponding author: e-mail marko.kaasik@ut.ee