3.1 Introduction    3.2 The dispersion model     3.3 The addition of background concns     3.4 Review and assessment procedure    3.5 Example

3 Industrial stacks

3.1 INTRODUCTION
The relationship between 90 th percentile of fixed 24-hour mean and annual mean PM₁₀concentrations described in section 1.3 is very unlikely to hold for the contribution from industrial stack emissions to ambient PM₁₀concentrations. It is therefore more appropriate in areas with significant industrial stack PM₁₀emissions to calculate estimates of the 90 th percentile of PM₁₀for direct comparison with the proposed 24-hour limit value of 50 mgm^-3.

Dispersion models may be used to assess the impact of industrial stack emissions on the 90 th percentile daily average PM₁₀concentrations. The formal process requires both local sequential hourly meteorological data and hourly measurements of background concentrations and involves:

• calculating the contributions from modelled sources to concentrations at receptors for each hour sequentially using meteorological data for each hour throughout a year;
  
• adding the background concentration on an hour by hour basis throughout the year;
  
• calculating the daily average total concentration at each receptor;
  
• ordering the daily averages at each location in order of decreasing concentration,
  
• and selecting the value corresponding to the appropriate percentile i.e. the 35th largest value.
  It is not practical to carry out this procedure for screening purposes and a simpler method is required for Stage 2 of the review and assessment procedure.
  
  A number of numerical experiments were carried out for specific examples using the formal procedure described above. The results were then compared with simpler methods and the most appropriate simple method has been chosen for inclusion in the informal technical information report. The numerical experiments presented here were carried out using TEOM monitoring data. The resulting method for estimating 90 th percentiles is equally applicable to both TEOM or equivalent and gravimetric or equivalent monitoring methods.
  
  
  3.2 THE DISPERSION MODEL
  
  The dispersion model , ADMS Version 2.2, described in section 2.2, was also used for this part of the study. The study made use of two main modes of operation of the model:
  
  
  • short term, in which the concentrations at receptors are calculated and reported sequentially for each hour of a meteorological data set, for data sets of up to 99 hours;
    
  • long term, in which the concentrations are calculated sequentially for each hour of a meteorological data set using a data set that covered a complete year but concentrations are reported only as averages and percentile values at each receptor.
    
  
  A computer script was prepared to facilitate the multiple runs of the model in the short term mode required to cover a complete year's meteorological data. Model runs in long term mode were used solely to identify the locations where average and percentile concentrations would be highest.
  
  3.2.1 Sources modelled
  
  Hypothetical new sources close to the national urban background PM₁₀ monitoring stations at Manchester Piccadilly and Liverpool Centre were modelled in order to assess their incremental impact on air quality. The modelled sources had the following characteristics:
  
  

  Stack height	25, 50, 100 m
  Stack diameter	1 m
  Discharge velocity	15 m/s
  Discharge temperature	150 oC
  Primary PM10 Emission rate	1 g/s

  
  3.2.2 Meteorological data
  
  Sequential hourly meteorological data for Manchester (Ringway) Airport for 1996 were used throughout the study. The data represented meteorological conditions in terms of wind speed, wind direction, time of day, time of year and cloud cover. The ADMS-2 meteorological pre-processor was used to provide estimates of boundary layer height for each hour throughout the year.
  
  A surface roughness value of 0.5 m was used to represent the terrain, characteristic of open suburbia.
  
  3.2.3 Receptors
  
  The model was initially run in long term mode using a 21 x 21 node grid with grid spacing 100m centred on the source in order to identify the locations of the highest modelled concentrations (annual mean and percentile). The model was then used in short term mode to calculate hourly concentrations at each of these receptors.
  
  3.2.4 Calculation of percentiles
  
  The 99.9 th percentile hourly value was determined from the hourly values by determining the 9th largest hourly value. The 98 th and 90 th percentile daily values were obtained by averaging the hourly values for each day of the year (midnight to midnight) and then selecting the 7 th and 35 th largest daily averages.
  
  The dispersion model ADMS was unable to calculate concentrations for a small number (~200) hours as a result of inadequate meteorological data or calm conditions. These hours were excluded from the analysis.
  
  3.2.5 Results
  
  The maximum modelled ground level concentrations are shown in Table 2. Table 2 also shows comparable values calculated using the Environment Agency's (EA) Guidance for estimating the air quality impact of stationary sources and HMIP's Technical Guidance Note (Dispersion) D1.
  
  Table 2: Modelled maximum ground level concentrations
  

  Statistic	Model	Primary PM10 Concentration, mg m-3
  		25m stack	50m stack	100m stack
  Annual average	ADMS	1.11	0.317	0.074
  90th percentile fixed 24 hour	ADMS	4	1.15	0.271
  98th percentile fixed 24 hour	ADMS	6.62	1.7	0.5
  99.9th percentile hourly	ADMS	16	6.93	2.75
  Annual average	EA Guidance	1.2	0.27	0.06
  99.9th percentile hourly	EA Guidance	18.5	8	3.8
  98th percentile hourly	D1	7.14	1.75	0.43
  Annual average	D1	1.25	0.30	0.075

  
  Table 3 shows the ratio of the 90 th percentile daily average concentration to other statistics calculated from Table 2. It is therefore recommended that an estimate of the 90 th percentile of daily average stack contributions can be derived by multiplying the modelled annual mean stack contribution by 4.0 or the 98 th percentile of hourly values by 0.66.
  
  Table 3 : Ratio of 90th percentile daily average concentration to various surrogates
  

  Statistic	Model	Ratio
  		25m	50m	100m
  Annual average	ADMS	3.6	3.6	3.7
  Annual average	EA Guidance	3.3	4.3	4.5
  99.9th percentile hourly	EA Guidance	0.226	0.145	0.071
  98th percentile hourly	D1	0.56	0.66	0.63

  
  
  3.3 THE ADDITION OF BACKGROUND CONCENTRATIONS
  
  National air quality monitoring data for the United Kingdom are archived by AEA Technology on behalf of the Department for the Environment, Transport and the Regions, The Welsh Office, The Scottish Office and the Department of the Environment, Northern Ireland. The following PM₁₀ concentration measurements for 1996 for the following sites were used in the analysis:
  
  Manchester Piccadilly
  Liverpool Centre
  
  The hourly measurements of urban background PM₁₀concentrations (in TEOM units) were added to the modelled primary PM₁₀concentrations from the industrial stack source using an Excel 5.0 spreadsheet.
  
  There are a small number of hours missing from the air quality monitoring data set. The ambient concentrations were set to zero for hours with missing data. (In a more detailed analysis, these missing values could be interpolated from adjacent hours or estimated from measurements at nearby monitoring sites.)
  
  The relative magnitude of the background concentrations and the stack contributions depends on the rate of emission of pollutants. The methods are required to be useful across the range of emissions. The calculation methods have therefore been applied for emission rates such that the maximum 90th percentile daily stack concentration is equal to the 90 th percentile daily background concentration. The results were normalised by dividing by the 90th percentile daily background concentration. The results are shown in Table 4.
  
  Table 4: Normalised maximum total 90 th percentile fixed daily PM₁₀concentrations
  

  	Normalised 90 th percentile daily total concentration
  	Liverpool Centre	Manchester Piccadilly
  25 m	1.65	1.60
  50 m	1.62	1.58
  100 m	1.59	1.57

  
  Table 4 shows that the normalised 90 th percentile total concentration is typically 1.6 times the 90th percentile background concentration.
  
  
  3.4 THE REVIEW AND ASSESSMENT PROCEDURE
  
  It is recommended that the 90 th percentile concentration including background and stack contributions is calculated as follows:
  
  
  1. Determine the 90 th percentile fixed daily background PM₁₀ concentration:
     
     • from local monitoring data (from a site not directly influenced by the stack emissions);
       
     • or from background maps of annual average concentrations by multiplying by 1.79 (see section 1.3).
       
  2. Determine the 90 th percentile fixed daily primary PM₁₀stack contribution:
     
     • from EA Guidance by multiplying the annual average by a factor of 4;
       
     • or using D1 and multiplying the D1 value by a factor of 0.66¹.
       
  3. Determine which is the larger of the above.
     
  4. Calculate the 90 th percentile of fixed daily means for the total PM₁₀concentration as the sum of the larger value and 0.6 times the smaller value.
     
  Figure 3 shows the maximum 90th percentile stack contribution for a range of annual average background concentrations.
  
  3.5 EXAMPLE
  
  In this example, assume that the annual average background in the vicinity of a stack is 21 mg m-3, gravimetric (from the national maps). The 90 th percentile value is then estimated as 1.79 x 21 =38 mg m-3, gravimetric.
  
  The annual average stack contribution is estimated to be 10 mg m-3, gravimetric from EA Stationary source guidance. The 90 th percentile is then estimated as 4 x 10=40 mg m-3, gravimetric.
  
  The background values is the larger so that the total 90th percentile is estimated to be 40+0.6 x 38=63 mg m-3 gravimetric, which is in excess of the proposed 24-hour limit value of 50 mgm^-3. It is concluded that this screening method indicates that a more detailed assessment is necessary.
  
  
  ¹ EA Guidance is preferred except for stacks less than 20m high or where there are buildings taller than 40% of the stack height within 5 stack heights. Domestic coal burning          Mapping studies

  ---

  Report and site prepared by the National Environmental Technology Centre, part of AEA Technology, on behalf of the UK Department of the Environment, Transport and the Regions