12/6/96

MEMORANDUM

SUBJECT: Hess Oil Virgin Islands Corporation Intermittent Control System (ICS):
                   Section 325 Waiver Request

FROM        Dennis Doll
                   Model Clearinghouse Coordinator

TO:            Annamaria Colecchia
                   Environmental Scientist, Region II

THRU:        Joseph A. Tikvart, Chief
                   Air Quality Modeling Group

In response to your request the Model Clearinghouse has reviewed your position with respect to
the technical supportability of the proposed HOVIC ICS.  Subject to the caveats below we agree
that the proposed ICS will be protective of the NAAQS.

First, since we do not have access to the meteorological data , we defer to your technical
judgment that the analysis of that data and modeling analysis is supportive of the 6- hour and
12-hour criteria described in your memorandum.

Second, as you are aware, our ability to accurately predict future air quality based on modeling
with a historical year of meteorological data is an inexact science.  Thus, as you have recognized,
intermittent control systems need to have some "safety factors" built in (Ref.1).  For example, you
indicate that the modeling data really indicated that no violations occurred until the wind was
into the critical sector for at least 18 hours, yet the criteria for fuel switching is based on 6-hour
and 12-hour winds into that sector.   This is sort of a safety factor, presumably to account for
unknowns and variability.  On the other hand apparently the critical wind sector is very precisely
defined based on one past year of modeling and limited exactly  to those directions for which
exceedances were modeled. For that uncertainty you indicate that the Company has offered to
install 3 monitors whose real time readings could be used to trigger and override of the prediction
system in the event that concentrations were getting "too high."  Other factors which might cause
a failure of the system to predict when exceedances might occur include the modeling based on a
temporally limited (one year) meteorological data base and the lag between the time that the
decision is made to switch fuels and the time that concentrations on the terrain downwind are
actually reduced.  Again, we leave it to your judgment that the safety factors you indicate and the
yet to specified monitor threshold value are adequate to protect air quality in the event of a failure
of the designed prediction system.

We understand that the 6-hour and 12-hour criteria are based on hourly average wind directions
and not some shorter period average or instantaneous values.  We also understand that the data
from the SO2 monitors will be telemetered to some kind of control room where they are
continuously watched for exceedances of a some critical short term or accumulated value.  We
also understand that when critical meteorological or ambient criteria are exceeded there will be
an immediate fuel switching to low sulfur oil. 

One of the critical factors in any ICS is enforceability (Ref.2).  The ICS plan should be in an
enforceable document, and not just a guideline that the Company uses or perhaps chooses not to
use in any given situation.  In principle, we should not think in this case that the Company would
be impelled to violate the plan for economic reasons since the frequency of the required fuel
switching would be low.  However, based on past guidance and on recent experience with ICS in
Guam, it would be wise to have an enforceable plan.

Another factor in the continued successful practice of ICS at a plant is to have an "upgrade"
provision (Ref.3) in place.  This means that after some period of time, say a year, the newly acquired
meteorological data are fed into the model and perhaps otherwise analyzed, and the monitored
data are compared to model estimates to see if modifications need to be made to the ICS to
accommodate previously unforeseen critical conditions. 

Finally, it should be noted that in spite of careful design and inclusion of safety factors, the
vagaries of the meteorology can cause a failure of an ICS, and thus there is always some risk of
exceedances of the NAAQS.  In accepting an ICS the control agency has to be willing to accept
such risk.

In summary, given the information we have and subject to the assumptions/caveats we have
stated above, we agree with you that the designed system should provide reasonable assurance
that the NAAQS will not be violated.

If you have any questions please contact Dean Wilson or me.

cc: D. deRoeck
      D. Wilson
      S. Riva (2APB-PS)
      M. Stanco (2APB-PS)
      R. Kelly (2APB-SIP)
      J. Siegel (2ORC)


References:

1.  Guidelines for Evaluating Supplementary Control Systems.  EPA-450/2-76-003.

2.  Guidelines for Enforcement and Surveillance of Supplementary Control Systems, Volumes I
& II.  EPA-450/2-75-008.

3.  Technique for Supplementary Control System Reliability Analysis and Upgrading. 
EPA-450/2-76-015.