July 5, 1990
 

MEMORANDUM 
 
SUBJECT:  PM-10 SIP Demonstrations for Small Isolated 
          Areas With Spatially Uniform Emissions 
 
FROM:     Robert D. Bauman, Chief  /s/
          SO2/Particulate Matter Programs Branch  (MD-15)
 
          Joseph A. Tikvart, Chief  /s/
          Source Receptor Analysis Branch  (MD-14)
 
TO:       Chief, Air Branch
          Regions I-X
 
           This memorandum is in response to recent conversations between 
the Particulate Matter Programs Section and Regions VIII and X.  The 
Regions have repeatedly expressed the need for flexibility in control 
strategy demonstration requirements when confronted with air-sheds where 
receptor modeling, coupled with proportional (rollback) modeling is 
considered to be adequate to identify source contributions and demonstrate 
attainment.  The purpose of this memorandum is to discuss the rationale and
justification for exercising this flexibility.
 
           It is appropriate in certain situations to rely on a receptor 
model (RM) demonstration (i.e., use of receptor modeling, emission 
inventories, design value obtained by air quality monitoring, and 
proportional modeling) as the basis for a control strategy demonstration.  
This approach is an option provided for in sections 4 and 6 of the PM-10 
SIP Development Guideline.  While it is clear from the guideline that the 
use of dispersion models in combination with receptor models is the 
preferred approach, in certain limited situations, the use of an RM 
demonstration alone may be adequate to demonstrate attainment.  The State 
must obtain approval to use the RM demonstration option prior to SIP 
submittal.  The decision that an RM demonstration is adequate to 
demonstrate attainment is the responsibility of the Regional Office; 
however, the Region should consult the Model Clearinghouse for advice in 
making this determination.  The Region must justify the determination and, 
in doing so, must consider all of the following:   
 
           1.   The spatial representativeness of the monitoring network and
                the spatial uniformity of emissions.  The PM-10 monitoring 
                network must be representative of the maximum air quality 
                impacts from the predominant (i.e., generally on the order of
                90 percent) sources and source categories in the PM-10 
                emission inventory.  Emissions from area source categories 
                are often distributed nearly uniformly across the area. This
                implies that ambient patterns would not be characterized by
                strong concentration gradients, thus lessening the need for
                an extensive monitoring network.  However, areas with point
                sources will generally find an RM demonstration difficult to
                justify because the concentration pattern would be 
                characterized by local "hot spots."  In such cases, a 
                dispersion model, along with representative meteorological 
                data are typically required. In a few areas, emissions of
                antiskid materials from a small number of road surfaces
                constitute the predominant PM-10 source category.  These
                emissions should be uniformly distributed along these road
                surfaces.  The monitoring network must be shown to be in
                accordance with EPA's monitoring guidance and spatially
                representative of the maximum air quality impact from this
                source category.
 
           2.   The temporal representativeness of the monitoring network. 
                If the 24-hour NAAQS is controlling, the network must have
                samples collected at sufficiently frequent intervals to 
                ensure that the impacts from the governing emission sources
                are adequately monitored.
 
           3.   The impact of only a few, relatively well characterized 
                source categories.  Receptor models can generally well 
                characterize only a limited number of chemically 
                distinguishable sources or source categories.     
 
           The above criteria imply that the area should be relatively 
small, characterized by uniform areawide emissions of one or two source 
categories, and geographically isolated from other PM-10 source areas.  
Examples of circumstances where RM demonstrations may be justifiable are 
small air-sheds where the only significant emission sources are residential
wood combustion and/or road antiskid materials.  It must be noted that the 
prerogative to use RM demonstrations should be exercised judiciously.  Even
when a RM is employed, consideration should be given to initiation of basic
meteorological measurements as a contingency to the control program being 
found inadequate and predictive dispersion modeling being necessary at a 
later time.  The use of dispersion modeling and receptor modeling in 
combination remains the preferred approach when both models are applicable 
to a particular circumstance. 
 
cc:  T. Pace 
           D. Stonefield 
           D. Wilson 
           Regional Modeling Contact, Regions I-X 
           PM-10 Contact, Regions I-X