ArticlePDF Available

Validation of the improved AEROPOL model against the Copenhagen data set

Authors:

Abstract

The improved Gaussian dispersion model AEROPOL (University of Tartu, Estonia) is validated against the Copenhagen air pollution dispersion dataset belonging to the Model Validation Kit of the Initiative on ''Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes''. Model updates include the application of: surface heat flux instead of indirect indicators of surface layer stability; and two-level wind data instead of one (10 m) level. Validation results show a fair agreement of modelled and measured concentrations. In comparison with several models validated earlier, the scatter of modelled versus measured data is moderate, but a systematic deviation is small. Thus, the AEROPOL model is recommended for urban dispersion calculations in nearly neutral conditions. Best results are expected for long-term average pollution levels.
8th Int. Conf. on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes
______________________________________________________________________________________
- 134 -
VALIDATION OF THE IMPROVED AEROPOL MODEL AGAINST THE
COPENHAGEN DATA SET
Marko Kaasik, Veljo Kimmel
Institute of Environmental Physics, University of Tartu. Tartu, ESTONIA
INTRODUCTION
The international agreement represented by COM (1999) sets certain criteria to which results of
air pollution modelling must correspond. Usually there does not exist enough spatial information
about pollution field – thus, validation of the model is complicated. International data sets for
validation of models are good alternatives – they represent pollution field in different
meteorological conditions and pollution patterns.
The aim of current study is to validate Gaussian-plume AEROPOL model against the
Copenhagen data set and thus check its applicability in typical weather and urban conditions.
The triggering reasons for this study were (1) the improvement of model and (2) the extensive
use of AEROPOL for urban planning applications in Estonia.
BACKGROUND
The AEROPOL model (developed in Tartu Observatory, Estonia) is a Gaussian plume model,
which includes the reflection and partial adsorption of the pollutant at the underlying surface,
wet deposition, initial rise of buoyant plumes. The model is applicable for gaseous pollutants
and particles from stacks, transport and area sources like domestic heating (Kaasik, 1996, Kaasik
et al., 2002).
The model was earlier validated against the Lillestrøm data set (expressing predominantly very
stable stratification) with relative success (Kaasik, 2000). In recent years the AEROPOL model
was intensely applied for urban planning purposes in Estonia. The model was recently improved,
introducing the two-level wind data (instead of 10 m wind only) and heat flux based method to
determine the Pasquill stability classes (instead of earlier determination by cloud amount and
solar elevation). Thus, the need for validation in more typical weather conditions became urgent.
The conditions of dispersion experiment in Copenhagen (1978 – 79, Gryning and Lyck, 1984)
correspond well to the typical situation, in which the AEROPOL model is applied: urban
elevated (115 m) release, mid-latitude maritime climate with quite strong winds and neutral or
slightly unstable stratification. The improved AEROPOL model was validated against the
Copenhagen data set (observed data: Olesen, 1994). The determination of statistics available in
the data set follows formulae given by Hanna et al. (1991):
normalised mean square error
()
P
P
CC
CC
0
2
0
NMSE
= (1)
fractional bias
()
P
P
CC
CC
+
=
0
0
5.0
FB (2)
fractional standard deviation
()
.
5.0
FS
0
0
P
P
CC
CC
σσ
σ
σ
= (3)
8th Int. Conf. on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes
______________________________________________________________________________________
- 135 -
In formulae (1 – 3) C0 is the measured concentration, CP is the observed one, overbar means
averaging over the ensemble and σ with corresponding indexes is the standard deviation. Other
applied statistics are the linear correlation coefficient (COR) and the factor of CP/ C
0 in the
diapason 0.5 – 2 (or fraction in factor 2, FA2).
VALIDATION
The summary statistics for arc-wise maximal concentrations is given in Table1 and for
crosswind integrated concentrations in Table 2. Both concentrations are normalised with the
source emission (mass emitted in unit time).
Table 1. Statistics for maximum arc-wise concentrations (normalised with emission,
unit 10-9 s/m3, 23 observations). Sigma – standard error, NMSE – normalised mean square
error, COR – linear correlation coefficient, FA2 – fraction in factor 2, FB – fractional bias,
FS – fractional standard deviation.
Model (country, comparison year) Mean Sigma Bias NMSE COR FA2 FB FS
Observations 632.7 450.3 0.0 0.00 1.000 1.000 0.000 0.000
AEROPOL (Estonia, 2002) 573.0 448.7 59.6 0.30 0.642 0.826 0.099 0.004
HPDM (USA, 1994) 358.2 268.1 274.4 0.61 0.847 0.658 0.554 0.507
IFDM (Belgium, 1994) 551.9 345.3 80.8 0.19 0.843 0.870 0.136 0.264
INPUFF (Romania, 1994) 560.6 352.7 72.1 0.50 0.490 0.739 0.121 0.243
OML (Denmark, 1994) 283.6 251.1 349.1 1.12 0.823 0.217 0.762 0.568
UK-ADMS (UK, 1994) 177.1 138.5 455.5 2.84 0.891 0.043 0.125 1.059
UK-ADMS extra (UK, 1994) 261.8 176.9 370.8 1.37 0.913 0.348 0.829 0.872
Table 2. Statistics for cross-wind integrated concentrations (normalised with emission,
unit 10-6 s/m2, 23 observations). Explanations of used statistics see Table 1.
Model (country, comparison year) Mean Sigma Bias NMSE COR FA2 FB FS
Observations 448.7 239.3 0.0 0.000 1.000 1.000 0.000 0.000
AEROPOL (Estonia, 2002) 386.5 183.6 62.2 0.19 0.624 0.913 0.149 0.263
HPDM (USA, 1994) 382.3 161.6 66.4 0.16 0.778 1.000 0.160 0.387
IFDM (Belgium, 1994) 443.3 193.4 5.43 0.16 0.681 0.957 0.012 0.212
INPUFF (Romania, 1994) 339.6 180.4 109.1 0.46 0.361 0.696 0.277 0.280
OML (Denmark, 1994) 249.2 131.7 199.5 0.52 0.893 0.565 0.572 0.580
UK-ADMS (UK, 1994) 207.1 110.7 241.6 0.86 0.912 0.348 0.737 0.735
UK-ADMS extra (UK, 1994) 297.0 122.5 151.6 0.34 0.856 0.783 0.407 0.646
Like the models validated earlier (Olesen, 1995), the AEROPOL model tends to underestimate
slightly both concentrations (bias). Scatter of data is relatively large (moderate correlation
8th Int. Conf. on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes
______________________________________________________________________________________
- 136 -
coefficient COR), but extreme deviations from observed values are seldom (high FA2 and low
NMSE). The overall scatter (Sigma) for cross-wind integrated concentrations is somewhat
underestimated and for maximum arc-wise concentrations almost perfect.
Looking at the plots of modelled versus observed concentrations (Figure 1) we see, that both
cross-wind integrated and arc-wise maximal concentrations are randomly scattered.
COPENHAGEN SCATTER - AEROPOL
a) Cross-wind integrated conc.
0
500
1000
1500
0 500 1000 1500
Observed, 10-6 s/cm2
Modelled, 10-6 s/cm 2
COPENHAGEN SCATTER - AEROPOL
b) Max. arc-wise conc.
0
500
1000
1500
2000
0 500 1000 1500 2000
Observed, 10-9 s/cm3
Modelled,10-9 s/cm3
Figure 1. Plots of modelled and observed cross-wind integrated (a) and maximal arc-wise (b)
concentrations. Concentrations are normalised with the source emission rate, one-to-one line is
added to the graphs.
8th Int. Conf. on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes
______________________________________________________________________________________
- 137 -
The deviations from the one-to-one line are bilateral and balanced (which is consistent with the
small bias) and increase nearly in proportion with the concentration, i.e. relative error is nearly
constant. The maximal arc-wise concentrations are slightly more scattered than the cross-wind
integrated ones.
CONCLUSIONS
We have to conclude, that the AEROPOL model performs fairly for elevated releases in urban
area during neutral or slightly unstable stratification. The systematic error of results is relatively
small - significant advantage from the operational point of view. Therefore, the best
performance is expected for long-term average concentrations.
Most of the models give higher correlations with observation than the AEROPOL model, but
tend systematically underestimate the concentrations (especially UK-ADMS). These models,
however, were validated eight years earlier and therefore the comparison does not reflect the
present state of their development. Nevertheless, as these models were in operational use already
in the middle of nineties during the validation at the Mol workshop, we have to conclude, that
AEROPOL performs well at level with the models applied in Europe and USA during the last
decade.
Assessment of air pollution levels by measurements is well determined through standard
methods and procedures set by COM (1999) and later directives. Although we cannot expect
that assessment of air pollution by modelling is standardised, we can expect that validation of
models is somehow standardised. Intensified practice of validating models against available
international data sets, and creating newer and better data sets can accelerate ongoing process of
standardisation of and improve comparability and quality of air pollution modelling.
REFERENCES
Commission of European Community (COM), (1999). Council Directive 1999/30/EC of 22 April
1999 relating to limit values for sulphur dioxide, nitrogen dioxide and oxides of
nitrogen, particulate matter and lead in ambient air. O.J.L 163, 29 June 1999, 0041 -
0060
Gryning, S.E., Lyck, E, (1984). Atmospheric dispersion from elevated sources in an urban area:
comparison between tracer experiments and model calculations, J. Climate and Applied
Meteorology, 23, 4, 651 – 660.
Hanna, S. R., Strimaitis, D. G., Chang, J. C., (1991). Hazard Response Modelling Uncertainty
(A Quantitative Method), Vol. 1., Sigma Research Corporation, Westford, MA, 71 p.
Kaasik, M., Kimmel, V., Kaasik, H., (2002) Air Quality Modelling system for a medium-sized
town: a case study in Estonia, Int. J. Environment and Pollution, in print.
Kaasik, M., (2000) Validation of models AEROFOUR and AEROPOL using the model
validation kit established at Mol, Int. J. Environment and Pollution, 14, 1 – 6, 160 –
166.
Kaasik, M., (1996) Atmospheric dispersion and deposition of technogenic calcium: model
estimation and field measurement, Proc. Estonian Acad. Sci. Ecol., 1/2, 41 – 51.
Olesen, H. R., (1994) Model Validation Kit for the workshop on Operational Short-Range
Amospheric Dispersion Models for Environmental Impact Assessments in Europe.
NERI, Roskilde, 32 p.
Olesen, H. R., (1995) The model validation exercise at Mol: overview of results, Workshop on
Operational Short-range Atmospheric Dispersion Models for Environmental Impact
Assessment in Europe, Mol, Nov. 1994, published in Int. J. Environment and Pollution,
5, 4 – 6, 761 – 784.
... This finding also consistent with other study which showed that NO 2 concentration were high in major highways. Other factors contribute to NO 2 concentration are poor maintenance of vehicle 28 . These levels of NO 2 were high compared to WHO guideline of air quality. ...
Article
Full-text available
Abstract Introduction: Causeway is providing a road link between travelers from Al-Khobar, Saudi Arabia and Bahrain. The vehicle traffic noise and emissions exhaust at Causeway have a significant effect on air quality parameters which leads to deteriorations impact on health of employees and passengers. The present study aimed to analyze the ambient air pollutants PM10, PM2.5, PM1, O3, CO2, SO2, CO, VOCs, and NO2, ambient noise and heat stress in and around causeway. Materials and Methods: All gaseous, particulate air pollutants, meteorological factors, ambient noise and heat stress were analyzed using standards calibrated equipment. July 2019 to October 2019 the study incused seven administrative sites in King Fahd causeway (Saudi Side). Air Quality Index (AQI) was used to determine the concentrations of selected ambient air pollutants. Results: All major air pollutants were high and significant difference in the concentration level between weekdays and weekends (p<0.05) and exceeds the international air quality guidelines in the selected studied sites of causeways. However, ambient noise demonstrated high significant levels in different studied sites in different times (p<0.05). Conclusion: Air quality measures in Causeway are significantly deteriorated and should be continuously evaluated to minimize the risk of daily exposure to traffic emissions.
... The AEROPOL model (Kaasik and Kimmel, 2004), version 5.3 is applied for dispersion modelling. The 'Main Street' of Tallinn cannot be considered as a street canyon, as the aspect ratio is mostly well below one and there are no continuous building fronts of nearly the same height, although some tall buildings are situated nearby. ...
... As noted earlier, the set of empirical formulations and algorithms developed by Briggs (1984) for evaluating the plumerise height of major point source emissions has been the basis of plume-rise calculations in several chemistry transport models such as GEM-MACH (Moran et al., 2010) and CMAQ/CMAx (Byun and Schere, 2006), as well as in regulatory air dispersion models such as AEROPOL (Kaasik and Kimmel, 2003) and CALPUFF (Levy et al., 2002). However, the details of how Briggs' algorithms were implemented may vary -we therefore provide the details of the GEM-MACH implementation below. ...
Article
Full-text available
We evaluate four high-resolution model simulations of pollutant emissions, chemical transformation, and downwind transport for the Athabasca oil sands using the Global Environmental Multiscale – Modelling Air-quality and Chemistry (GEM-MACH) model, and compare model results with surface monitoring network and aircraft observations of multiple pollutants, for simulations spanning a time period corresponding to an aircraft measurement campaign in the summer of 2013. We have focussed here on the impact of different representations of the model's aerosol size distribution and plume-rise parameterization on model results. The use of a more finely resolved representation of the aerosol size distribution was found to have a significant impact on model performance, reducing the magnitude of the original surface PM2.5 negative biases 32 %, from −2.62 to −1.72 µg m⁻³. We compared model predictions of SO2, NO2, and speciated particulate matter concentrations from simulations employing the commonly used Briggs (1984) plume-rise algorithms to redistribute emissions from large stacks, with stack plume observations. As in our companion paper (Gordon et al., 2017), we found that Briggs algorithms based on estimates of atmospheric stability at the stack height resulted in under-predictions of plume rise, with 116 out of 176 test cases falling below the model : observation 1 : 2 line, 59 cases falling within a factor of 2 of the observed plume heights, and an average model plume height of 289 m compared to an average observed plume height of 822 m. We used a high-resolution meteorological model to confirm the presence of significant horizontal heterogeneity in the local meteorological conditions driving plume rise. Using these simulated meteorological conditions at the stack locations, we found that a layered buoyancy approach for estimating plume rise in stable to neutral atmospheres, coupled with the assumption of free rise in convectively unstable atmospheres, resulted in much better model performance relative to observations (124 out of 176 cases falling within a factor of 2 of the observed plume height, with 69 of these cases above and 55 of these cases below the 1 : 1 line and within a factor of 2 of observed values). This is in contrast to our companion paper, wherein this layered approach (driven by meteorological observations not co-located with the stacks) showed a relatively modest impact on predicted plume heights. Persistent issues with over-fumigation of plumes in the model were linked to a more rapid decrease in simulated temperature with increasing height than was observed. This in turn may have led to overestimates of near-surface diffusivity, resulting in excessive fumigation.
... As noted earlier, the set of empirical formulations and algorithms developed by Briggs (1984) for evaluating the plume rise 15 height of major point source emissions has been the basis of plume rise calculations in several chemistry transport models such as GEM-MACH (Moran et al., 2010) and CMAQ/CMAx (Byun and Schere, 2006), as well as in regulatory air dispersion models such as AEROPOL (Kaasik and Kimmel, 2003) and CALPUFF (Levy et al., 2002). However, the details of how Brigg's algorithms were implemented may varywe therefore provide the details of the GEM-MACH implementation, below. ...
Article
Full-text available
We evaluate four high-resolution model simulations of pollutant emissions, chemical transformation and downwind transport for the Athabasca oil sands using the Global Environmental Multiscale – Modelling Air-quality and Chemistry (GEM-MACH) model using surface monitoring network and aircraft observations of multiple pollutants, for simulations spanning a time period corresponding to an aircraft measurement campaign in the region in summer 2013. We have focussed here on the impact of different representations of the model's aerosol size distribution and plume-rise parameterization on model results. The use of a more finely resolved representation of the aerosol size distribution was found to have a significant impact on model performance, reducing the magnitude of the original surface PM2.5 negative biases by 32 %. We compared model predictions of SO2, NO2, and speciated particulate matter concentrations from simulations employing the commonly-used Briggs (1984) plume-rise algorithms to redistribute emissions from large stacks with stack plume observations. As in our companion paper (Gordon et al., 2018), we found these algorithms resulted in under-predictions of plume rise, with 39 to 60 % of predicted plume heights falling below half of the observed plume heights. However, we found here that a layered buoyancy approach for stable to neutral atmospheres, coupled with the assumption of free rise in convectively unstable atmospheres, resulted in much better model performance, both for atmospheric constituent concentrations and the predicted height of the plumes. Persistent issues with over-fumigation of plumes in the model were linked to positive biases in the predicted temperatures between the surface and 1km elevation. These in turn may lead to overestimates of near-surface diffusivity, resulting in excessive fumigation.
... The model used to estimate fly ash and radionuclide atmospheric dispersions has been previously validated against international data sets (Kaasik, 2004(Kaasik, , 2000Kaasik and Kimmel, 2003). The earlier versions of AEROPOL were applied to estimate the deposition fluxes of fly ash and sulphate from the oil shale industries of Estonia and the results were validated against the deposition measurements based on snow samples (Kaasik et al., 1999;Sofiev et al., 2003). ...
Article
Two of the world's largest oil shale-fired power plants (PPs) in Estonia have been operational over 40 years, emitting various pollutants, such as fly ash, SOx, NOx, heavy metals, volatile organic compounds as well as radionuclides to the environment. The emissions from these PPs have varied significantly during this period, with the maximum during the 1970s and 1980s. The oil shale burned in the PPs contains naturally occurring radionuclides from the (238)U and (232)Th decay series as well as (40)K. These radionuclides become enriched in fly ash fractions (up to 10 times), especially in the fine fly ash escaping the purification system. Using a validated Gaussian-plume model, atmospheric dispersion modelling was carried out to determine the quantity and a real magnitude of fly ash and radionuclide deposition fluxes during different decades. The maximum deposition fluxes of volatile radionuclides ((210)Pb and (210)Po) were around 70 mBq m(-2) d(-1) nearby the PPs during 1970s and 1980s. Due to the reduction of burned oil shale and significant renovations done on the PPs, the deposition fluxes were reduced to 10 mBq m(-2) d(-1) in the 2000s and down to 1.5 mBq m(-2) d(-1) in 2015. The maximum deposition occurs within couple of kilometers of the PPs, but the impacted area extends to over 50 km from the sources. For many radionuclides, including (210)Po, the PPs have been larger contributors of radionuclides to the environment via atmospheric pathway than natural sources. This is the first time that the emissions and deposition fluxes of radionuclides from the PPs have been quantified, providing the information about their radionuclide deposition load on the surrounding environment during various time periods.
... In this study the Gaussian AEROPOL model (basic features see Kaasik et al., 2004) is applied to estimate the annual concentrations of nitrogen dioxide in Tartu, the second largest town of Estonia (about 100000 inhabitants), for RHINE (Respiratory Health in Northern Europe) study. It is evident that in densely built-up areas the plain Gaussian model does not perform well due to multiple obstacles and volume occupied by the buildings. ...
Chapter
Full-text available
The Gaussian plume model AEROPOL 5 is applied to estimate the yearly average NO2 concentrations in Tartu, the second largest town of Estonia with about 100,000 inhabitants, for RHINE study. We apply the porosity concept by Genikhovich E, Gracheva I, Filatova E (Modelling of urban air pollution: principles and problems. In: Borrego C, Schayes G (eds) Air pollution modelling and its application, XV. Kluwer, New York, pp 275–283, 2002) in post-processing of modelled ground-level concentrations: the area under buildings is excluded from dispersion volume in each grid cell, thus the concentration is divided to the fraction of porosity, i.e. non-built-up area. It appears that porosity correction substantially enhances the site-wise correlations between model-estimated and measured concentrations, bringing the underestimated levels in particular monitoring sites closer to reality. Moreover, correlations are even higher, when dividing the “raw” modelled concentrations by squared porosities. We suppose that reason of non-linearity is in slowing down the wind between the buildings.
... Classical examples of Gaussian models are CALINE-4 (California Line Source dispersion model, Benson, 1984), ISC3 (Industrial Source Complex dispersion model, U.S. EPA, 1995) and UK-ADMS (Carruthers et al., 1994;CERC, 1998). Following the same trend as semiempirical models, some Gaussian models have been further developed integrating vehicleinduced and ambient turbulence, surface roughness parameterization for treating obstacles within complex environments or parameterised chemical reaction schemes, such as CAR-FMI (Contaminants in the Air from a Road -Finnish Meteorological Institute, Härkönen et al., 1996) and AEROPOL (AERO -POLlution, Kaasik & Kimmel, 2003) Additionally, more advanced models can include additional modules to resolve physical processes, fast chemical reactions and aerosol dynamics. ...
Thesis
Air pollution control strategies have helped to improve air quality over Europe over the last thirty years. Despite this success, the improvements have been insufficient to protect health of those who spend most of their time within urban areas and particularly near major roads. Considering that an urban street network is a structuring component of the city, a detailed analysis of its interaction with the environment and inhabitants is relevant for assessing policy measures aiming at improving urban air quality. However, given the inherent complexity of urban environments and the incomplete understanding of the physical and chemical processes involved in pollutant dispersion, it is a challenging task to estimate urban air quality. The objective of this thesis was to develop a modelling approach to resolve not only the long-range pollution dispersion but also to capture the higher pollutant concentrations commonly found within urban areas and at kerbside for assessing the localized effects associated to vehicle-related pollutants. A major challenge was to provide estimates not only for specific streets and cities but for the major road network and all streets within major cities in Germany. For this purpose, a hybrid dispersion modelling approach was developed for predicting air quality across different spatial scales: from regional to urban and kerbside scale, accounting for the intrinsic and distinctive urban morphology of urban areas. The hybrid dispersion modelling approach was used for estimating annual concentrations of PM10, PM2.5, and NO2 for all urban areas with more than 50 000 inhabitants in Germany for the reference year 2005. The results showed that there are large differences on air quality levels across urban areas in Germany, which may be relevant when planning or evaluating emission control strategies. Furthermore, an empiric regression function was also developed to estimate the number of days exceeding the daily limit value of 50 µg/m3 of PM10 by means of the annual mean PM10 concentrations. The air quality reference scenario was also applied to estimate health and climate change impacts associated to road vehicle activities in Germany in the year 2005. The modelling approach enabled a differentiation not only by vehicle types and technologies but also across urban areas within a geographically referenced framework. The results indicate that the total health damages attributed to road vehicle operation in Germany in the year 2005 account for almost 70 thousand Disability Adjusted Life Years (DALYs). Furthermore, the urban areas with the largest total DALYs attributed to vehicle use are located in the Essen-Dortmund area, Cologne-Bonn, and Berlin, followed by Hamburg, Frankfurt, Munich, and Stuttgart. The total health and climate change costs attributed to road vehicle operation in Germany in the year 2005 were estimated to be around 5.5 thousand millions EURO2000. Whereas heavy duty vehicles did only 8% of the total mileage, they are responsible for 27% of the total costs. On the other hand, both diesel and petrol passenger account for about 86% of the total mileage and they are responsible for two thirds of the total costs.
... The AEROPOL model is a local dispersion model based on the stationary Gaussian plume with reflections from the underlying surface and capping inversion. Details are described by Kaasik and Kimmel (2004). The contemporary version has options to determine the plume dispersion parameters (in Briggs' formulation) either from routine meteorological observations (wind, solar elevation, cloud amount) or applying the sensible heat flux data and two-level wind speed. ...
Article
Rapidly growing number of cars, after the collapse of the Soviet Union, made the street pollution the most urgent air quality problem. Need for new traffic routes pushes the reconstruction of street network. Computations of maximal and annual mean concentrations of NOx, CO and PM2.5 from street and road traffic were performed applying the AEROPOL model including the present situation in the city of Tallinn. The used dispersion model was based on the Gaussian plume dispersion with analytical integration over the length of street segment.
Article
Full-text available
Traffic is the main factor affecting air quality in most cities. After the Estonian re-independence in 1991, the increase of motorization has been fast and car usage has intensified. During the same period, the average age of cars has decreased and thanks to improvements in engine technology, the emissions per km have been reduced. The objective was to see how these factors have reflected in air quality. This paper also aim to present an analytical approach to estimate the air pollution levels in recent years, when air quality monitoring has not been conducted, and available traffic data are limited. Based on traffic counts in 25 points across the city the amounts of traffic were modelled for 680 street segments with CUBE software. As air quality is monitored it-regularly in Tartu, dispersion modelling was used to estimate pollution levels. Annual concentrations of exhaust particles (PM(exhaust)), particulate matter (PM(10)) and nitrogen oxides (NO(x)) in 1993, 2000 and 2006 were calculated with AEROPOL software. The traffic increase in the city centre of Tartu was especially rapid in the 1990s. In recent years, it has slowed due to congestion. Overall, traffic levels have increased more than 3 times since 1993. In residential areas, the increase is still rapid - up to 6 times from 1993 to 2006. However, the changes in air quality are less dramatic. Increases from 1993 to 2000 were followed by stable or slightly increasing pollution levels in recent years, especially in case of PM(exhaust). The study showed that 2 factors, namely, increase of traffic and improvement of vehicles, have been compensating each other in the dynamics of air pollution.
Article
Full-text available
The AEROPOL and AEROFOUR dispersion models developed at Tartu Observatory (Estonia) are validated using the Lillestrom dataset and following the model comparison rules established at the workshop in Mol. Running the AEROFOUR model, the vertical wind and temperature co-variations not included into the official dataset were used. The computed and measured arc-wise maximum and crosswind integrated concentrations match fairly. No clear tendency of over or under-prediction was found. The best performance was achieved by AEROFOUR model with the above-mentioned covariations included into the initial dataset. This demonstrates that the detailed measurement techniques give clear advances for model estimations if the model is designed to use this kind of data.
Article
Full-text available
Atmospheric dispersion experiments were carried out in the Copenhagen area under neutral and unstable conditions. The tracer sulphurhexafluoride was released without buoyancy from a tower at a height of 115 m and then collected at ground-level positions in up to three crosswind series of tracer sampling units, positioned 2-6 km from the point of release. The site was mainly residential having a roughness length of 0.6 m. The meteorological measurements performed during the experiments included the three-dimensional wind velocity fluctuations at the height of release.Dispersion parameters estimated from the measured tracer concentrations were compared with dispersion parameters calculated by various standard methods. These included methods based on the measured wind variances at the experiments and methods based on a stability classification of the atmospheric conditions. The wind variance-based methods are seen to be better than the stability-based methods in predicting the variation of y. In addition to being a function of downwind distance, the normalized lateral spread estimated from the measurements was found also to vary as a function of the wind speed as already demonstrated by Doran and others for low-level sources. The centerline ground-level concentrations were calculated by use of y and z derived according to the aforementioned methods and assuming the ordinary Gaussian concentration profile. These were compared with the corresponding measured concentrations; in this case also, the wind variance-based methods were seen to be better. In some of the experiments, the assumed Gaussian tracer concentration profile in the vertical direction was found to be inconsistent with the measured crosswind-integrated tracer concentrations.
Article
The air pollution modelling system used in Tartu (population 101,000) is aimed at examining the pollution levels in different meteorological conditions, and determining optimal traffic schemes and strategic development plans for the city. The system consists of: (i) a Gaussian plume model AEROPOL, and (ii) a database on sources of NO<sub align="right"> x </sub>, CO, SO<sub align="right"> 2 </sub>and particulates (from industrial, traffic, and domestic heating sources). The calculations of annual average and maximum concentrations are performed and fair agreement with monitoring results is found. A similar scheme is used in Parnu (population 60,000) and Kuressaare (population 15,000).
Article
A modification of the Gaussian plume model is proposed for the deposition of fine and heavy particulates and gases. It combines a downward-sloping plume to account for settling and the assumption of constant deposition velocity to obtain a solution satisfying an integral mass conservation equation specifying that the decrease in the airborne flux equals the deposition. An example is given comparing this method with the conventional Gaussian model for dry deposition of gases or fine particulates.
Hazard Response Modelling Uncertainty (A Quantitative Method
  • S R Hanna
  • D G Strimaitis
  • J C Chang
Hanna, S. R., Strimaitis, D. G., Chang, J. C., (1991). Hazard Response Modelling Uncertainty (A Quantitative Method), Vol. 1., Sigma Research Corporation, Westford, MA, 71 p.
Model Validation Kit for the workshop on Operational Short-Range Amospheric Dispersion Models for Environmental Impact Assessments in Europe
  • H R Olesen
Olesen, H. R., (1994) Model Validation Kit for the workshop on Operational Short-Range Amospheric Dispersion Models for Environmental Impact Assessments in Europe. NERI, Roskilde, 32 p.
The model validation exercise at Mol: overview of results, Workshop on Operational Short-range Atmospheric Dispersion Models for Environmental Impact Assessment in Europe
  • H R Olesen
Olesen, H. R., (1995) The model validation exercise at Mol: overview of results, Workshop on Operational Short-range Atmospheric Dispersion Models for Environmental Impact Assessment in Europe, Mol, Nov. 1994, published in Int. J. Environment and Pollution, 5, 4 -6, 761 -784.