January 19, 1994



MEMORANDUM

SUBJECT:  Test Protocol for Wind Tunnel Modeling of Plume Impact
          Under Stable Stratification for the Cane Run Station (CRS)
          in Louisville, Kentucky

FROM:     Joseph A. Tikvart, Chief
          Source Receptor Analysis Branch, TSD (MD-14)

TO:       Brenda Johnson, Regional Modeling Contact
          Region IV 

          Douglas Neeley, Chief
          Air Programs Branch, Region IV
          

     In response to your request to Dean Wilson, the Model Clearinghouse
has reviewed your draft comments to Louisville Gas and Electric Company
(LGE) on the subject protocol.  Also, as requested, we have reviewed the
protocol itself, but a detailed technical review of the wind tunnel study
protocol has not been done.

     Based on discussions we had within the Office of Air Quality Planning
and Standards, we agree with your draft comments on the test protocol. 
Below are a few additional comments based on our review of the test
protocol.  Also, because the modeled violations of the sulfur dioxide (SO2)
national ambient air quality standards (NAAQS) occur on elevated terrain in
Indiana, we recommend that your comments also be coordinated with Region V.

     As noted in your comment number three, it is not clear how the wind
tunnel results will be used in the compliance demonstrations of the SO2
NAAQS.  There needs to be a clarification on whether the tunnel results
will be used as part of a screening analysis or in a more refined modeling
analysis.  The protocol implies that the wind tunnel study will be done for
the meteorological conditions that yield a potential violation of the SO2
NAAQS (i.e., stable stratification with possible plume impact on elevated
terrain).  The protocol implies 1 hour average SO2 concentration estimates
will be derived from the wind tunnel results and the 3 and 24 hour average
estimates will be derived from Environmental Protection Agency (EPA)
scaling factors applied to the 1 hour estimates.  However, if a refined
modeling analysis is being proposed using hourly, sequential meteorological
data, the methodology is unclear concerning the use of wind tunnel results
in deriving 3 and 24 hour average concentration estimates.  There is no
discussion on developing concentration estimates from the wind tunnel study
for the other meteorological conditions represented in the hourly,
sequential meteorological data.  Also, as noted in your comment number two,
if a refined modeling analysis is contemplated for this study, the
availability and usage of on-site meteorological data needs to be
addressed.

     Another issue not addressed in the protocol is accommodation of
intermediate terrain in any mathematical modeling of the stacks.  For
example, the protocol refers to a previous screening analysis using COMPLEX
I (Valley equivalent) with violations of the SO2 NAAQS indicated on
elevated terrain.  The Guideline on Air Quality Models (Revised) specifies
the use of COMPLEX I concentration estimates for receptors with elevations
at or above plume height.  For receptor elevations between stack top and
plume height (intermediate terrain), a technique to determine concentration
estimates should be developed in consultation with the Regional Modeling
Contact.  There is no mention of an intermediate terrain technique applied
in the previous COMPLEX I modeling analysis.  Also, intermediate terrain
issues are not addressed in the test protocol for any mathematical modeling
contemplated.
     
     We have some concern on the plume representation in the wind tunnel
under stable stratification and complex terrain.  For example, in Appendix
A of the test protocol, the preliminary wind tunnel study is described. 
This preliminary study compared ground level concentrations on elevated
terrain under stable stratification derived from the wind tunnel results
and modeled results.  The conclusion was that the models overestimated the
wind tunnel observations by at least a factor of 2 under these conditions. 
However, in describing the flow visualization for the wind tunnel, it was
noted that "the plume traveled over the terrain and did not impact directly
on Williams Peak under stable stratification."  It is possible that the
models could have predicted plume separation around the elevated terrain
with some plume impaction.  Thus the models would have predicted higher
concentrations than were simulated in the wind tunnel.  As plume impaction
is a physical process that may occur, this raises a concern about the
ability of the wind tunnel to fully simulate plume behavior in complex
terrain under stable stratification.

     The test protocol states that "..wind tunnel measurements will provide
information to verify the performance of the guideline EPA models.."  The
protocol states that overall maximum concentrations from the tunnel studies
and modeling will be compared.  However, there is no discussion of
performance measures or statistical measures to be used in the performance
verification.  The methodology to do the model performance verification
should be made explicit.  Also, there is no indication to compare wind
tunnel or model results with measured data to verify performance.  The
implication is that the wind tunnel replicates the stable atmosphere.  As
described above, we have some concern about the representation in the wind
tunnel of neutrally buoyant plumes under stable stratification for complex
terrain.

     Appendix B of the protocol presents information regarding the validity
of wind tunnel modeling.  The comparisons described in Appendix B refer to
wind tunnel modeling and field observations for neutrally buoyant plumes
under neutral or unstable stratification  and flat or complex terrain.  It
would be useful if comparisons were available of wind tunnel modeling
results and field observations under stable stratification and plume
impaction on elevated terrain (e.g., comparisons of wind tunnel data and
observations from the Cinder Cone Butte study).

     In summary, there are several procedural issues identified in your
comments and ours that need to be addressed in the test protocol.  However,
in principle, the proposed techniques offer some potential and, provided
the procedural issues can be resolved, it would be appropriate to further
review the technical approach for using wind tunnel results in the
regulatory compliance demonstration.  As mentioned above, we have some
concern about the representation in the wind tunnel of neutrally buoyant
plumes under stable stratification for complex terrain.  If appropriate, it
may be useful to have direct discussions with LGE and their contractor on
the technical approach for the wind tunnel study following resolution of
the other procedural issues.
     
     If you have any further questions, please contact Dean Wilson at (919)
541-5683.

Attachment

cc: G. Blais