4. Physical Modeling approaches that use physical models

der

laboratory conditions to develop and test modeling theory and aid field study planning, and

the

5. Special Phenomena modeling approaches that use models designed to address unique topographical features that dominated source receptor relationship (e.g. complex terrain).

The air quality modeling matrix also addresses pollutants and pollutant combinations for which a given model type will be needed, i.e. a model is specified by means of model type and pollutant(s).

The atmospheric processes research program that leads to the development of models is based on the concept that the meteorological, chemical, and physical processes can be described in modules which can be asembled to form a complete model. Frequently the modules are interchangeable between models. For example, the meteorological processes that are relevant to an urban/mesoscale Eulerian model for hourly averaged S02 concentrations may be interchangeable for the same model type for NO2 or 03 pollutants. Likewise the chemistry module for hourly 03 concentrations may be used in both urban scale and regional scale hourly Eulerian models.

The air quality modeling matrix is used to identify models that will be needed in the next ten years. Blank spaces in the matrix identify models that are totally satisfactory and available or that will not be needed in the next ten years. Changes in the matrix in terms of the components, the needed models, or priorities will be reflected by ASRL and document includes the following information for each model given a priority rating in the matrix:

the model type

the pollutant of concern

the client's requirements

the specifications and R&D needed to run the model

the ADP requirements for running and evaluating
the model

the resource requirements for model R&D and
performance evaluation

the estimated periods of performance required for

model R&D and performance evaluation

The matrix allows one to (1) assess the progress achieved to date In the ASRL modeling research program, (2) identify those areas where further R&D is needed, and (3) determine where field and laboratory studies can be combined or coordinated. The use of the matrix should enhance the ASRL modeling research program planning process and should reduce the chances of duplicating research efforts. Over the long-term, the use of the matrix should reduce the costs associated with model development, evaluation, and validation.

Ratings: 1--urgent, immediate need; 2--high priority; 3--medium priority; 4--low priority: Blank--no known current need.

All model types listed should accurately describe atmospheric transport and dispersion for a conservative gas or tracer." Listings for specific pollutants reflect the further need to include transformation, removal, and other physical/chemical atmospheric processes for specific applica tions.

An effort to assess the accuracy of each model, based on field data, should be an essential component of the model development process.

The final formulation of all models should include sufficient parameterization of physical/chemical processes so that required data is readily avail. able or can be obtained at modest costs.

As work is completed on individual models, it is essential that they be made
publically available through the UNAMAP system. This system should be
periodically updated to insure that the models included therein reflect the
current state-of-the-art.

Where estimates longer than one hour are required, it is recognized that
the longer averaging times may be an average of 1-hour estimates or a
periodic sampling of such estimates.

b. High priority should be given to the evaluation/improvement of short-term model estimates. In particular, dispersion coefficients and mixing heights applicable to large point sources should be considered. The role of funtgation and the development/evaluation of an algorithm for its incorporation into current models should also be considered.

C.

For S02, hourly also refers to 3-hourly concentrations; for CO, daily also refers to 8-hourly concentrations.

d. Averaging times shown under point sources apply.

..

The capability of urban/mesoscale models to simulate accurately the impact of emission reductions should be evaluated; treatment of subgrid phenomena should be improved.

f. The potential for modeling high, multi-day episodes should be considered.

9. Available Gaussian models for PMO should be improved and evaluated.

h.

1.

The development/evaluation of urban models for PM10 which include significant secondary particulates should be completed."

A regional scale model for PM2.5 and for visibility should be developed, improved and evaluated.

J. Techniques for treating building downwash and aerodynamic wake effects should be developed, improved and evaluated.

k. Toxic/hazardous organic pollutants, as well as, benzene, arsenic, metals, etc. are included. Current techniques for treating surface fugitive sources should be evaluated. This is particularly important for control of toxic disposal facilities. Also included are fugitive dust sources and resuspension, and finite area and line sources.

1. The emphasis here is on denser than air pollutants that are likely the result of accidental/explosive releases which can be treated as point or area sources.

a. This includes evaluation of reactive plume models, their dependence on the composition of dilution air, and the appropriateness of Gaussian assump tions in such models.

n.

Work to accurately characterize formation and factors affecting formation of crucial intermediate or contributing species such as H202, CH20. PAN. Gas and liquid phase research is included. Also included is the use of box models to test and evaluate chemical mechanisms that will be integrated with transport and dispersion algorithms for urban mesoscale models.

0. This includes urban/rural fumigation and nocturnal "bursting."

D. Deposition and resuspension may be important.

QUESTION 2

When is the "scientific community" going to make the model improvements mentioned and what is the timetable? What is the status of EPA research in this regard? Please describe that research.

RESPONSE TO QUESTION 2

For the past six years, the EPA has maintained a cooperative agreement with the American Meteorological Society (AMS), jointy funded by the Office of Research and Development's. (ORD) Atmospheric Sciences Research Laboratoary (ASRL) and the Office of Air and Radiation's (OAR) Office of Air Quality Planning and Standards (OAQPS). Under this agreement, the AMS has appointed a Steering Committee on Air Quality Modeling that advises the EPA on the scientific validity of modeling approaches.

The subject GAO Report (GAO/RCED-86-94) states "in 1983, the AMS reviewed eight rural dispersion models for EPA. From their review, the AMS concluded that the scientific community must improve model physics, calculation techniques, and model input in order to obtain more precise results from models." Since six of the eight rural models tested were developed independently by outside agencies and consulting firms, the question as to when the scientific community is going to make the model improvements mentioned and what is the timetable?" cannot be answered by the EPA. This is especially true in light of the AMS comment that "this review of the rural models reveals a disturbing tendency on the part of some of the modelers to make their products resemble EPA Guideline models rather than try to make them fundamentally better."

In regard to the two EPA rural models, the EPA and AMS concluded that the models were no longer up to date with contemporary understanding of turbulence and diffusion in the planetary boundary layer (PBL). Inspired by the current PBL knowledge and the generally inadequate state of regulatory modeling, the EPA requested the AMS Steering Committee to convene the Workshop on Updating Applied Diffusion Models (1984) to review current understanding of the PBL and its parameterization in diffusion modeling. The Workshop was also asked to recommend improved models, i.e., those with better physics and hopefully better performance than existing models. The Workshop resulted in seven articles in the Journal of Climate and Applied Meteorology (1985), with the first being an overview of the Workshop proceedings by the Chairman. The remaining six by U.S. and European authors dealt with various aspects of PBL dispersion such as laboratory and numerical simulations compared to direct measurements, reviews of Lagrangian statistical (Monte Carlo) turbulent simulations, methods for improvement of Gaussian plume modeling, and a description of a meteorological preprocessor for air pollution modeling, being developed jointly by EPA scientists and their Dutch, Norwegian, and Danish counterparts.

For

This international summary of current PBL research results was made available to both the government and civilian modeling communities. its part, the EPA/ORD modelers have already begun the process of incorporating the Workshop recommendations into not only their existing rural