 |
|
 GROUP OF INDEPENDENT EXPERTS Ltd. |
|
|
|
|||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Home » Projects » REB 2000 » Conferences REB 2000 TRAF: a micro-scale street model for dispersion of pollutant emissions generated by traffic OCTAVIAN DATCULESCU 1. Background Starting mainly with the last century, as a result of advancing industrialization, population growth and the transport demand increase, especially in the urban environment, the atmospheric pollution due to anthropogenic activities, took acute and often irreversible forms. There is a continuous accumulation of different pollutants in the environment, with severe consequences on humans, plants and animals, buildings, art works and landscape. Traffic is the most important atmospheric pollution source in urban areas, and the environment impact generated by vehicles on the major routes could be significant for suburban and rural areas too. This is the reason why dispersion models of pollutant emissions generated by traffic are essential instruments for updating and implementing projects of urban and municipal territory planning. 2. Objective of the study The study objective is the estimation by numeric simulation of the pollutants concentration levels in terms of: a) characteristics of the emission sources, b) local topography and weather conditions, c) atmospheric processes of physico-chemical transformation, d) processes developed at the interface atmosphere-soil. Modelling of air quality is the only instrument to develop a prognosis of future environment impact of anthropogenic activities, allowing the efficiency evaluation of different environment pollution abatement strategies. 3. Utilization fields a) Environment impact assessment studies b) Environmental audits c) Atmospheric pollution forecasts 4. General technical characteristics of the TRAF operative dispersion model For a better flexibility and compatibility of the numerical model in terms of different conditions imposed by a potential user, three specialized versions have been developed: a) version 1.0 - a traditional approach of analytic expression of dispersion parameters taking into account the Pasquill-Gifford classes (finite number) of atmospheric stability; b) version 2.0 - a "new generation approach", where dispersion parameters are expressed as continuous functions of the limit layer parameters by using a meteorological preprocessor; c) version 3.0 - is dedicated to the urban " street canyons", taking into account the flow characteristics and the dispersion conditions of these basic town structures. Versions 1.0 and 2.0 of the model TRAF are approached using analytical solutions of the Gauss equation of diffusion for linear emission sources, while version 3.0 uses an approximate analytical solution of the linear flow equations Navier- Stokes applied to "street canyons", using the viscosity turbulence concept. The TRAF model includes a complex specialized module for the approach of emission power using the CORINAIR-COPERT traffic pollutant emissions identification methodology recommended by the European Union as well as our own researches concerning the emission characteristics of the national vehicle fleet, based on the statistic processing of over 15,000 emission measurements over the period 1994-1999. The outputs of this module are emission powers of the traffic linear sources and composite emission parameters obtained by the weighed average of emission parameters characteristic to every vehicle category, taking into account the composition of the vehicle fleet on the road during the time interval under study. This emission module can be used as an emission preprocessor for other internationally known dispersion models, such as CALINE3, CALINE4, HIWAY-2, ROADWAY, General Motors (GM-SLSM), Canyon Plume Box Model (CPBM), Operational Street Pollution Model (OSPM), etc. Besides particle emissions through the exhaust gases of Diesel (MAC) equipped vehicles, for micro-scale street models we must take into account re-suspension powders caused by the superficial charge attributable to traffic on various types of roads (treated in detail in the US EPA AP-42 document/1997). The superficial charge includes both air turbulence effects induced by the traffic wave, and the effects caused by mechanical forces developed at the contact between the tires and the road. 5. Input data
5.1 Meteorological data preprocessing for TRAF 2.0 Unlike traditional models where the dispersion parameters and atmospheric stability classes are generally estimated through empirical methods based on worldwide experimental measurements, for the new generation models, these parameters, essential for atmospheric pollutant concentration calculation, are determined using meteorological routine observations achieved in the area under study, based on preprocessing meteorological input data in terms of the atmospheric limit layer . In this context, the description of the atmospheric limit layer taking into account the mixing height (Hmix ) and the Monin-Obukhov length (L) has the essential advantage of being based of physically significant parameters, allowing the estimation of atmospheric turbulence variation in terms of height ; the turbulence in the limit layer is measured function of the ratios z/Hmix and z/L where z is the height of pollutants; thus, diffusion emission parameters are dependent upon the source height. To the contrary, the traditional classification of turbulence into discreet atmospheric stability classes (Pasquill-Gifford) is based on surface estimations alone. Meteorological preprocessing is the specific method of deductively obtaining the meteorological parameters necessary for dispersion models, using available synoptic data, as well as the means by which the hourly series of data over long time periods are statistically processed to produce the climatology of dispersion categories. In view of the improvement and standardization at European level of methodologies for meteorological data preprocessing, in the last years the program COST 710 was implemented, the main principles of which were applied in the present paper. The physical basis of modern meteorological preprocessors is the parametric evaluation of the atmospheric limit layer structure, including the interaction at the soil level (at surface). Generally, two alternative methods are used: I. The vertical wind and temperature profiles method based on the similarity theory Monin - Obukhov (Holtslag - Van Ulden scheme, KNMI, the Netherlands):
II. The surface energy audit (Berkowicz-Prahm scheme, NERI National Environmental Research Institute, Denmark) Examples of meteorological preprocessors:
The main parameters of the planet limit layer obtained through meteorological preprocessors are friction speed (u * ), Monin-Obukhov length (L), mixture height (Hmix), convective speed scale (w.) and temperature scale for heat turbulent transfer. 5.2. Background concentration levels According to the present methodology, in order to estimate by numerical simulation the atmospheric concentration of a pollutant in a given reception point, we must correlate the TRAF model outputs with background concentration levels of the said pollutant; the total concentration C is obtained by adding the local street-level concentration Cs obtained by means of the TRAF model with the background concentration Cf. For each pollutant there are residual concentration levels produced by sources other than traffic: natural emissions (biogenic or mineral), other anthropogenic activities like industry, residential heating, transport means other than road ones, conducted at an urban, regional, continental or planetary scale. For the evaluation of these components of background concentrations, two methods are available: a) using data on air quality measured by the national and world stations for environment monitoring (the Meteorological World Organization - WMO and EMEP); b) using the outputs of existing meso-scale dispersion models , for example:
Versions 1.0 and 2.0 are available; they use analytical solutions of the Gauss diffusion equation for surface emission sources, treated as an infinite network of infinitesimal pointlike sources. Version 1.0 - uses a classical approach of the analytical expression of the vertical dispersion parameter in terms of the Pasquill-Gifford discreet classes of atmospheric stability; Version 2.0 - is a new generation model where the vertical dispersion parameter is expressed as a continuous function of the parameters of the planet limit layer obtained through a meteorological preprocessing system. 6. Output data Average hourly atmospheric concentrations of 10 pollutants:
The results can be presented both under numerical form (table) and graphical form (see fig. 1 and 2). 7. Verification - validation The predictive capabilities of TRAF and TRAFURB models have been preliminarily tested by comparison to experimental data on urban and suburban air quality in Bucharest area; the results were in line with the limits established by performance criteria for atmospheric dispersion models. Extending the comparison between the numerical program application results and experimental data on air quality in the proximity of traffic ways will allow the validation and further improvement of calculation methodologies used as predictive approaches. 7.1. Equipment used in real time measurements of air quality in the proximity of traffic ways 1994-1996 The measurements conducted in collaboration with Bucharest Environmental Protection Agency were made with the mobile lab belonging to this institution, having the following monitoring equipment for air quality and local meteorological parameters:
7.2. Automatic Meteorological micro-station SOLOMAT 4000 MetLab. 1997-1999 The measurements were made with RAR own mobile lab equipped for the moment with only one performing analyzer: CO - HORIBA APMA-360 (infrared selective absorption). The resulting database on air quality in the proximity of traffic ways has been completed and correlated with:
7.3. Examples of TRAF model application: Bucharest (urban and suburban area) 7.3.1. Atmospheric concentrations - short time averages (hourly values) For a first analysis of the predictive performances of dispersion models, carbon monoxide (CO) was selected taking into account the following:
The first comparison between the experimental results and the results obtained by numerical simulation with TRAF model (version 1.0) as well as with other 5 dispersion models at street micro-scale was made for a reception point situated at about 4m from the edge of Bucharest-Ploiesti Route (DN1), across the road from the Bucharest-Baneasa Airport ; the wind was blowing from N-NE (Baneasa Forest area), so it was presumed that the impact on air quality was due to traffic sources exclusively (the background concentration at the regional level was about 0.2 mg/m 3 during daytime). No significant airport activity (landing or take-off) was reported during the time interval under study. As it can be observed from the following table, the TRAF model predictive performance for this case study is perfectly comparable with the one of American dispersion models; the slight tendency to overestimate concentration levels (a general characteristic of TRAF models, as it can be derived from the examples below) can be considered a positive feature for environment impact assessment studies, where a "conservative" approach is preferred. The lower predictive performance of the SLSM model is due mainly to the fact that the model had been calibrated for high speed traffic (80 km/h) and not for 40-50 km/h, as recorded in the proximity of the selected point; additionally, the much larger size of American cars in the 70's compared to Romanian cars today has led to an exaggeration of the mechanical turbulence induced by the traffic wave, and consequently to the under-estimation of CO atmospheric concentration level. CO atmospheric concentration - average values per hour (measured and numerically simulated). Bucharest-Ploiesti Route(DN1), Bucharest - Baneasa Airport Area 27.05.1995, 10:00 - 11:00 hours Wind speed (at 10 m): 1.68 m/s Wind Angle (between route axis and wind direction): 12.28 o Atmospheric Stability Class Pasquill-Gifford: B Atmospheric air temperature: 26.47 o C Receptor height from the soil: 4.5 m
Examples for other pollutants Average atmospheric concentrations per hour - measured and numerically simulated with TRAF (street micro-scale) and TRAFURB (urban fund) dispersion models: a) Nitrogen oxides - NOx and NO 2 Cotroceni Route, Botanical Garden Area 22.09.1994, 11:00 - 13:00 hours Wind Speed (at 10 m): 0.54 m/s Wind Angle (between the route axis and wind direction): 85 o Atmospheric Stability Class Pasquill-Gifford: A Atmospheric air temperature: 26.3 o C Global Solar Radiation: 675 W/m 2 Receptor height from the soil: 4.5 m
b) Sulfur dioxide SO 2 (identical meteorological and site conditions)
c) Suspension powders - PM 10 Bucharest- Ploiesti Route (DN1), Baneasa Airport Area 27.05.1995, 10:00-11:00 hours Wind Speed (at 10 m): 1.68 m/s Wind Angle (between the route axis and wind direction): 12.28 o Atmospheric Stability Class Pasquill-Gifford: B Atmospheric air temperature: 26.47 o C Receptor height from the soil: 4.5 m
Bucharest- Ploiesti Route (DN1), Baneasa Airport Area 28.05.1995, 17:00-18:00 hours Wind Speed (at 10 m): 3.33 m/s Wind Angle (between the route axis and wind direction): 11.11 o Atmospheric Stability Class Pasquill-Gifford: C Atmospheric air temperature: 29.20 o C Receptor height from the soil: 4.5 m
7.3.2. Predictions on medium term (daily average values) An example is illustrated in fig.3 and 4 presenting a comparison between the daily hourly variation (three consecutive days average - period 5-8.06.1995) numerically simulated with TRAF (street micro-scale) and TRAFURB (urban meso-scale) models and the CO atmospheric concentration level measured in a reception point situated in the proximity of N. Balcescu Bldv. No.24 right across the Orizont Galleries. The time periods to which dispersion models have been applied were characterized by a large range of meteorological conditions: from atmospheric calm to wind speed of 2 m/s in variable directions; atmospheric turbulence fluctuated between specific conditions of morning and evening severe thermal inversions and a strong convective stratification during daytime with intense solar radiation, amplified by the infrared response at the level of pavement and fronts of adjacent buildings, and quasi-neutral stability characteristic for clouded sky conditions; humidity oscillated between small levels specific to noon sunshine and very significant values recorded during summer precipitations. The high value of the correlation coefficient (r=0.94) and the difference of just 2 % between CO concentration levels calculated and measured during 24 hours (average daily values) perfectly match the limits stipulated by the performance criteria imposed to atmospheric dispersion models.
|
|
|
|
 This book is the result of the proceedings of the Romanian Environmental Forum, 6th edition held in Bucharest between 16 and 19 November 1999. 
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
|
|
|
© GROUP OF INDEPENDENT EXPERTS, LTD. All rights reserved. Company · Services · Clients · Projects · Events · Experts · Contact · Links · Sitemap |
 |
|