by James M. Rauch                                         2-10-94

     The following comments on the Tonawanda EIS are submitted in
accordance with NEPA.  The comments are divided into four parts:
general comments on the total document package and the timing of
its release, and comments on each of the document components - the
Remedial Investigation Report (RI) for the Tonawanda Site, February
1993 (BNI), the Baseline Risk Assessment (BRA) for the Tonawanda
Site, August 1993 (SAIC), and the Feasibility Study - Proposed Plan
(FS-PP) for the Tonawanda Site, November 1993 (SAIC).  Specific
comments are numbered and cross-referenced to general comment(s)
where appropriate to substantiate the general comments.


     1)  Last winter I attended a meeting led by Site Manager Ron
Kirk at the newly opened DOE Tonawanda office where background on
the site was presented and questions were answered.  I entered my
name on a DOE mailing list and expressed an interest in the site
remediation process.  The RI was not available at that time, but it
was mentioned that the FS would be available in the Fall at which
time the public would have the opportunity to comment at public
hearing.  Notice of the release of the FS-PP was received 11-10-93. 
The FS-PP was immediately requested directly from Oak Ridge.  The
FS-PP arrived on 11-17-93, two weeks before the scheduled public
hearing on 12-1-93.  At the time of the hearing many people who had
requested the FS-PP still had not received it.  Several commenters
requested more time to review the document, and requested
additional hearings.  Only a thirty day extension of the comment
period was granted.  
     The FS states that the RI and BRA are incorporated by
reference in the EIS package.  After several requests to the DOE
Tonawanda Office, a copy each of the RI and BRA were made available
on 11-24-93.  The material contained in these two documents
contains important site-related information which is essential to
the NEPA public review process.  Considerable time is required to
review and digest these documents.  Why were interested parties not
notified by letter of the availability of these important documents
as they were completed?  This release procedure has placed severe
time constraints on commenters, and has significantly reduced the
opportunity for meaningful public participation in the NEPA review
process.  Pursuant to this, how many copies of the RI and BRA have
been distributed to the public, and when?

     2)  None of the three documents contains adequate internal
references or cross-references to the others.  Many statements and
assumptions are made without providing a substantiating reference
or referral to supporting RI data.  Internal references to material
in other locations is often general and does not specify page or
section.  Reference is  sometimes made to a table or figure which
does not contain the information described in the text.  The RI in
particular is poorly organized.  It is sometimes difficult or
impossible to associate tables and figures with relevant text, and
vice versa.

     3)  The RI and BRA in particular contain numerous errors. 
Some of these are typographical and/or transcription errors; others
are incorrect or misleading statements.


     4)  Pg 1-2 : FUSRAP was initiated in 1974.  The Tonawanda
properties were designated for remediation on the following dates:
Linde, 2-80; Ashland 1 and Seaway, 6-84; and Ashland 2, 10-84.  DOE
Order 5400.4 requires responses to releases or potential imminent
releases to be conducted in accordance with CERCLA.  A stated
objective of the program is the mitigation of any immediate hazards
associated with site conditions (pg 1-5).  Why was the waste pile
on the Linde property, which was consolidated west of Building 90
in 1979-82, not covered until 1992?  In view of early indications
of contamination at Ashland 2 (aerial surveys and ORNL 1980), why
was the Ashland Oil Company allowed to continue moving wastes to
Ashland 2 through 1982?

     5)  Pg 1-4 : The statement is made "DOE decided to include
Seaway in the RI/FS-EIS process with the other Tonawanda properties
to minimize the effort involved with meeting the procedural and
documentation requirements of NEPA and CERCLA."  In actuality, an
attempt to segment Seaway and deal with it separately and quickly
through the establishment of an on-site cell (to accommodate BFI,
the landfill operator) was rebuffed by strong local
government/public opposition.  (See general comment 3)

     6)  Pg 1-5,6 : The statement is made "DOE has determined that
an EIS is the appropriate level of NEPA review necessary to
adequately inform decision-makers and the public of reasonable
alternatives for minimizing any adverse impacts of the proposed
action at the Tonawanda site. ... DOE's CERCLA/NEPA integration
policy is not intended to represent a statement on the legal
applicability of NEPA to remedial actions under CERCLA."  If DOE
has determined that an EIS is required, and it has, then DOE must
adhere to NEPA requirements.  (See general comment 3)

     7)  Pg 1-6 : The reference for the WP-IP, listed in the
references section as NEPA/CERCLA Work Plan - Implementation Plan
(8-93) is incorrectly given here as (BNI, 1992c), and as (BNI,
1992b) in Appendix D-1.  The work was to be performed between 11-90
and 5-91.  How is it that the WP-IP comes after the RI?  (See
general comments 2 and 3)

      8)  Pg 1-7 : The Niagara River flows northeast past the site,
not northwest.  (See general comment 3).  Compass directions are
not specified as being relative to Ashland Grid North in either the
text or the figures, causing confusion with compass north.

      9)  Pg 1-22 : What specifically were the results of the 1958
radiological survey conducted by AEC (referred to in ORNL, 1978b)
which prompted AEC to release Ashland 1 for unrestricted use?

     10)  Pg 1-26 : What are the locations and the results of soil
sampling at two locations on Ashland 2 performed by ORNL in 1980?

     11)  Pg 1-32 : Groundwater at Seaway: the results of Wehran
(1979), although limited to analysis of non-radiological
contaminants, indicate that leachate from the landfill is migrating
both downward into deeper layers and outward with discharge to
surface drainage.  In view of this, why wasn't the full length of
the southern drainage channel which borders the Seaway landfill on
the south (and empties into the Niagara River), sampled?  This
channel lies directly downslope from unsampled disposal areas `B'
and `C'.  Analysis of samples from area `A' indicates depletion of
the more mobile radionuclides Ra-226 and U-238.

     12)  Review of Tables A-11 and A-16 indicates a very
non-uniform (heterogeneous) distribution of radiological
contamination, the volume of which has not been adequately
characterized.  For example, in Seaway area `A', auger holes
B23R009, B23R012, B23R013 and B23015 had Th-230 levels ranging from
11 to 180 pCi/g at the bottom of the hole.  Similarly, sampling is
insufficient to adequately define the nature and extent of
contamination in Area `D'.  Areas 'B' and `C' were not
characterized. It appears that the areal extent of contamination of
the surface soil layer and, in particular, the subsurface soil
layers is inadequately defined, especially at Ashland 2 and Seaway.
How accurate are estimates of contaminated volumes at each of the
properties?  (See comments 44 and 62)

     13)  Pg 1-52 : Figure 10: Where are the results for the
sediment samples from locations M14, M15, and M16 along Two Mile
Creek downstream of Linde?  There appears to be no data between
locations 2 and 3 in this drainage.  It does not appear that this
drainage has been adequately sampled, nor has the Niagara River
itself, especially considering the significant activities

     14)  Pg 1-67 : Water sampling at location W-7 in the western
drainage area of Seaway between Murphy Motor Freight Line and the
Weigh Station indicates Ra-226 levels at three to ten times
background.  Was this area and the entire drainage system  through
to the Niagara River gamma surveyed and/or soil sampled?  Parts of
Seaway area `A', portions of which remain uncovered, are within
this drainage and present significant potential for erosion to the
Niagara River.

     15)  Pg 1-67,68 : ORNL 1978 Data locations W-8 and M-8 (Tables
1-4 and 1-5) could not be found on any figure.  Where are they? 
(See general comment 2)

     16)  Pg 3-15 : Climatological data: The three year (1987-89)
precipitation average, 35.2 inches, from the recent NWS station at
North Tonawanda is not representative of current trends being
witnessed at the Buffalo NWS in the moving 30-year average,
reported on pg 3-4 as 37.5 inches (for 1951 through 1980) and most
recently as 38.58 inches (for 1961 through 1990).  For the same
1987-89 period, the Buffalo NWS averaged 40.6 inches.  The North
Tonawanda location is less representative of actual conditions at
the site than the Buffalo location.  Precipitation amounts decrease
rather quickly as one goes north from the southern tip of Grand
Island.  The North Tonawanda NWS is roughly five miles north of the
Tonawanda sites.  Lewiston, N.Y., ten miles to the north, averages
30 inches per year.  To the extent that use of the three-year North
Tonawanda figure biases the calculation of percolation and runoff
(up to 10%), contaminant migration involving these processes is

     17)  Pg 3-25 : The text refers to (Kappel 1992).  This
reference does not appear in the reference section.  (See general
comment 2)

     18)  Pg 3-35 : Unconsolidated material: The varved lacustrine
clay is described as "very thin locally", for example, three feet
thick at bore hole B55G48 at the southern end of Seaway.  The
possibility of discontinuity in this layer at Seaway cannot be
discounted.  Discontinuities do occur at Ashland 1 and 2.  When
coupled with desiccation cracking of the clay unit, piping of
particles (approximately 1 foot/day, pg 3-46), sand lenses in the
overlying till and zones of fine-grained sands in the underlying
glaciolacustrine unit which have higher hydraulic conductivity, the
route to the contact zone aquifer may be complete.  Over the long
term, DOE scenarios ruling out this pathway are not justified.  In
fact, DOE does not adequately address long-term groundwater impacts
either here or at the Linde injection wells or the properties in

     19)  Pg 3-40 : Geochemical Properties of Soil: Table 3-5 lists
uranium distribution ratios (Kd)  determined  from soil samples
collected at Ashland 1 and 2.  The varved clay is characterized by
only one sample.  This table has entries for which no Kd values are
given.  Were these samples not analyzed?  Based on this data, which
ranged from 5 to 29  (cm3/g), an average value of 10 was selected. 
Despite the fact that Kd values are contaminant-specific,
site-specific Kds were not determined for Ra-226 or Th-230.  A
review of the distribution coefficient data (distribution
coefficient closely approximates distribution ratio) in Table 5-1
(pg 5-35) reveals considerable variability in this important
parameter.  For example, Gee et al. l980 reported a Kd for U-238 at
pH 7.7 of 23,000 (cm3/g), while Rancon l973 reported a value of 100
(at pH 8), a difference of approximately two and a half orders of
magnitude.  Most significantly, Gee reports a Kd of 2400 for Ra-226
at pH 7.7, a value roughly ten times smaller than that of U-238 at
pH 7.7.  This may reasonably be interpreted as an indication that
the Kd of Ra-226 at Tonawanda is approximately l cm3/g.  Erroneous
Kd values (too large) will seriously underestimate radionuclide
mobility.  Anionic composition of the radionuclides is not given. 
Table 5-4 indicates moderately high solubility for both radium
nitrate and thorium sulfate.  Based on this data and data in Table
5-2, thorium mobility may be underestimated, and radium mobility is
likely to be greater than, not less than, that of uranium.        
     20)  Pg 3-40 : Measured cation exchange capacities for the
till, varved clay, and glaciolacustrine deposits show considerable
variability, from 30 to 485 meq/100g.  It is unclear whether each
of the three ranges given results from samples from all three soil
units or if each range applies to one each of the soil units;  it
is also stated that these results are being verified. Please
explain.  Is it not realistic to expect this range of
non-uniformity across the soil types?  Since this exchange capacity
is an important parameter in determining the mobility of the
radionuclides, assumptions made here may introduce a ten-fold

     21)  Pg 4-2 : Radioactive Constituents of Interest:  The
statement is made "Radionuclides in these operations [uranium ore
processing operations conducted at Linde] come from the
uranium-238, uranium-235, and thorium-232 decay chains.  Therefore,
based on a review of the decay chains and previous survey
information uranium-238, radium-226, thorium-232, and thorium-230
were selected as indicator radionuclides. ...Because the primary
radiological constituent of the American and African ores is
natural uranium, thorium-232 (which is not in the uranium decay
chain) is not expected to be in abundance in the ores or their
residues."  Later, and also in the FS (pg 2-3), it is claimed that
Th-232 was not present in the MED ores, but since both slag and fly
ash (used as surface fill at Linde and to a lesser extent at
Ashland 1) are "suspected sources of heavy metals and radionuclides
including Th-232", DOE has chosen Th-232 as an indicator the
presence of which in an area is used to exclude the area from DOE
cleanup responsibility.  This EIS contains no inventories of what
was received at the site or residue characterization references
(other than the Aerospace 1981 effluent characterization which has
been requested but not yet received) to support the assumption that
no Th-232 was present in the ores.  The lack of such specific
information regarding the radionuclide composition of the ores is
a serious shortcoming. A considerably more sophisticated study
would be necessary to begin to determine the relative contamination
contributions from the ores versus the slag/fly ash.  Without such
information, DOE has no justification for excluding such areas of
Th-232 contamination from both chemical and radionuclide
remediation.  As explained on pg 1-2, "Under FUSRAP, DOE assumes
responsibility for: managing any chemical contamination at the
Tonawanda site that is mixed with radioactive contamination or that
resulted from activities conducted for MED."

     22)  Linde processed both African ore (high in radium content)
and American ores (lower in radium, partly because they had already
been subject to vanadium extraction),  and therefore the site has
contamination resulting from both ore streams. It is stated "The
other three properties have contamination originating from the
American ores only" since the African ore wastes were shipped to
Lewiston, N.Y. (the Niagara Falls Storage Site).  Of course, the
high radium African ore effluents were discharged to the sanitary
sewers, and the storm sewers which drain into Two Mile Creek and
subsequently the Niagara River.  Other than estimates of the
activities of the uranium and Ra-226 in these effluents,
characterization of the nature and activities of other
radionuclides (e.g. Th-230) is not given.  The ultimate fate and
impacts of these discharges is not presented.  Sampling of
sediments in Two Mile Creek downstream of Linde and in the Niagara
River at Two Mile Creek is either insufficient or totally lacking. 
This is inadequate for purposes of NEPA.

     23)  Pg 4-4 : An exclusionary procedure is described which
"will be used in part to differentiate soils contaminated by MED
filter cake from soils contaminated by other industrial processes
(e.g., refinery operations) that may contain metals at
above-background concentrations."  The basis of this exclusion is
"a direct one-to-one correlation between high concentrations of
vanadium and lead and high levels of radionuclides (greater than
guidelines for all three radionuclides [emphasis added] of
interest)."  Since some of the data indicates that radionuclides
may be differentially removed (transported) from contaminated areas
(pg 7-26), according to FUSRAP (see comment 21) the presence of a
single radionuclide above guidelines should require cleanup of such

     24)  Pg 4-5 : Background Concentrations, Soils:  The selection
of Ashland 2 South for determination of background levels of
radionuclides and metals is not appropriate.  Ashland 2 South lies
directly downwind, both in relation to predominant wind direction
and maximum wind speed, from the original pile areas on Ashland 1
and may have received airborne deposition.  Also, previous surface
drainage patterns from Ashland 1 may have contaminated portions of
Ashland 2 from which some of the `background' samples were taken
(see results for locations GS-1 and SP-12).  Representative
area-wide background values for radionuclides reported by ORAU 1981
are significantly lower (roughly 50%) than the values from Ashland
2 South.  The ORAU values should have been used as the background
radionuclide levels.  Why weren't they?

     25)  Pg 4-6 : Background Concentrations, Sediment:  It is not
clear what background sediment values are used for the four
properties.  Are they the values given for location 113 at Ashland
1 (Table 4-35)?  The proximity of sediment sampling locations GS-5
and 113 to the original piles and roads makes them suspect.

     26)  Pg 4-6 : Background Concentrations, Groundwater:  It is
unclear what the sitewide radiological and nonradiological
background values are.  Figure 3-20 shows two wells as background,
B55W36D, and B55W35S.  Are the background radionuclide values those
footnoted in Table 4-71?  If so, what is the Th-230 value?

     27)  Pg 4-7 : Guidance Values:  DOE guidelines (order 5400.5,
Table 4-3) for residual radioactive contamination incorporate
assumptions and scenarios that are not sufficiently protective of
public health and the environment under plausible current or future
conditions at the Tonawanda sites.  These sites currently are
surrounded by an area of dense human population and activity. 
Current land use plans propose intensified human presence and
activity at the sites.  Owing to the rich physical attributes of
the area (e.g., Great Lakes), it is a virtual certainty that in the
long term, this pattern of increased human presence and activity
will prevail.  It cannot be assumed that planning, or institutional
controls, should they be claimed as risk management tools, will be
maintained throughout the duration of the hazardous period (500,000
years).  The DOE soil guidelines will allow average
post-remediation exposures of 100 mrem per year above background,
with smaller hotspots (less than 25 square meters) having up to 30
times this rate.  This hotspot exemption must be eliminated.  The
proposed post-remediation basic dose limit can be expected to
produce excess stochastic effect risks (cancers, genetic damage)
that are approximately 33% higher than the background cancer risks. 
[Assuming the lifetime background cancer risk, 1 x 10-2 (EPA 1989a),
results from a lifetime average dose of 300 mrem/yr, a 100 mrem/yr
increment in dose exposure (the target post-remediation basic dose
limit) translates into a 33% increase in cancer risk or a lifetime
risk of 1.7 x 10-2.  BEIR V predicts a similar result.  See comments
59 and 60 also.]   Obviously, this is unacceptable. In an area such
as this, where future reconstruction activities are a given,
setting subsurface (greater than six inches) remediation standards
(15 pCi/g) for Ra-226, Th-230, and Th-232 at a level three times
the surface (first six inches) standard (5 pCi/g) makes no sense. 
 The derived concentration guidelines (DCGs) for water allow
similar exposure rates and are therefore also unacceptable. This
discussion leads to the inescapable conclusion that the basic dose
limit for the Tonawanda site must be no more than 10 mrem per year
above background as prescribed in New York State TAGM-4003.  It is
conceivable that in retrospect even this post-remediation
incremental level of exposure, resulting in a 3.3% increased risk
of cancers and genetic effects, will be viewed as unwise,
especially with regard to the nature of the hazard and its
indefinite duration.

     28)  Pg 4-81 : The extent of contamination within the
watershed of the drainage ditch on Ashland 2, which enters the
Niagara River at sampling point SP-7, and the SP-7 area itself, has
not been adequately determined.

     29)  Pg 4-83 : Nature and Extent of Contamination in
Groundwater:  At a meeting last winter at the Tonawanda office, in
response to questions concerning the fate of the effluent injected
into the Linde wells and possible remediation, Site Manager Ron
Kirk said that the injected effluent was "gone", implying there was
nothing to remediate.  An opposite conclusion based on numerous
assumptions is presented in section 4.3.  The suggestion (pg 4-28)
that injected effluent remains confined to a limited injection zone
is not adequately supported by results from just two offset wells. 
Because "no detailed chemical analysis is available for the
effluents disposed of by injection" (pg 4-89), it is impossible to
accurately determine the extent of precipitate formation.  The fact
that an offset bore hole #1 core showed 176 pCi/g of U-238 and 1.3
pCi/g of Ra-226 (pg 428), when compared with an estimated 3 Ci of
uranium and 5.5 Ci of Ra-226 injected, may indicate migration of
     The injection scenario remains obscure.  It is suggested that
the effluent plume may have entered the glaciofluvial deposits and
based on combined glaciofluvial (5.5 ft/yr) and Salina bedrock (24
ft/yr) groundwater flow velocities, it is further suggested that
the area of contamination may extend laterally for 300 to 1200 feet
(pg 4-87).  The glaciofluvial estimate of flow velocity may be too
low.  Page 3-38 states that this unit is conservatively assumed to
be composed of "coarse-grained materials (sand)", yet the 5.5 ft/yr
estimate of groundwater velocity incorporates an estimate of
hydraulic conductivity (lower) for silty sand (pg 3-53). 
(Contact-zone groundwater velocity at Seaway was estimated at 82
ft/yr.)  Page 4-87 indicates that analysis of radiological and
chemical contaminants in the contact-zone is being refined and will
be included in a later technical memorandum.  This is insufficient. 
The information presented in the RI does not adequately define the
injected effluent or its fate, and the resulting long-term impacts. 
The purposes of NEPA review are not satisfied.

     30)  Pg 5-22 : Perched Groundwater System:  The significance
of lateral migration in the perched groundwater system may be
seriously underestimated.  Despite evidence of widespread
heterogeneity in the composition of the till, water flow models are
developed which emphasize the unlikelihood of contaminant migration
to deeper aquifers (see comment 18) while tending to ignore the
higher horizontal groundwater velocities (26 ft/yr, to 1049 ft/yr
at Seaway) in the perched system.  There are too few wells to
characterize migration in this non-uniform till, therefore risk
modeling for this exposure pathway is insufficient for the long
term, especially considering the indefinite duration of the hazard.

     31)  Pg 7-5 : The conclusion is drawn that uranium
concentrations in the groundwater are below the DOE DCG, ignoring
the fact that a Ra-226 concentration in well B29W09D exceeded the
DOE drinking water guideline of 5 pCi/l.  This establishes the need
for further analysis of groundwater impacts as well as the
evaluation and presentation of potential remediation techniques. 
(See comments 29 and 50)

     32)  Pg 7-37 : Infiltration into the Seaway Reinforced
Concrete Pipe:  This continues to be a route for migration of MED
contaminants from both Ashland 1 and Seaway to Ashland 2 and
beyond.  The need to seal the RCP and prevent flow along its trench
has long been obvious, but Table 7-2 indicates action to be taken
by the owner by an "undetermined" time.  When?  This situation is
allowing uncontrolled release of both radiological and non-
radiological MED wastes.  (See comment 4)

     33)  Pg 7-54 : In a chart of Data Limitations and Future Work
at the Tonawanda Site (Table 7-2), item (2) is Two Mile Creek
dredgings.  Future work to be done is described as "Walkover gamma
scan of the Town of Tonawanda landfill and collection of soil
samples.  Collection of Two Mile Creek bed core samples."  The
estimated date of completion is given as "Completed in 1992," yet
page 7-36 (dated 12-28-92) describes these activities as "future
work".  This information could not be found in a  review of the FS. 
The EIS is incomplete and insufficient for public review without
such information.

     34)  Pg 7-38 : Designation of Vicinity Properties: 
Investigation of properties contaminated, or possibly contaminated,
by MED wastes is incomplete.  No report is given on the extent of
likely contamination at R.P. Adams directly adjacent to Linde on
the north or the extent of contamination at the Town of Tonawanda
landfill (resulting from suspected disposal there of  contaminated
sewage treatment plant sludge).  To what extent were the sewer
effluent contaminants concentrated and retained in the sewer plant
sludges?  Elsewhere in the RI, reference is made to gamma surveys,
indicating that contamination extends beyond the surveyed areas at
Ashland 2 and onto the G.K. Hambleton and Benson Development Co.
properties.  Why were these properties not studied?  The material
presented is insufficient for the purposes of NEPA review.

     35)  Pg C-19, Appendix C : Soil Classifications Table: 
Percents passing no. 400 sieve are given as greater than percents
passing no. 200 sieve.  (See general comment 3)

     36)  Pg D-15, Appendix D : Quality Assurance and Quality
Control Evaluation:  It is stated "the only factor thought to have
a potentially significant impact on representativeness was holding
time violations, which would produce an underestimate of sample
concentrations."  Previously, page 4-3 states "because radioactive
decay is a random process, no precise correlation between the rate
of disintegration and a given radionuclide concentration can be
established; therefore, the exact concentration of the radionuclide
cannot be determined."  Do these statements imply that the counting
times for determinations of radionuclide concentrations were
insufficient in some cases?

     37)  Pg 4-8 : QA/QC rejection of thallium, selenium, and boron
metal analyses due to improper analytical procedures has resulted
in a data gap in the target metals. 


     38)  Pg 1-3 : Radiological Sources: Th-232 is listed, instead
on Th-230, in description of procedure for identifying contaminated
areas and subareas.  (See general comment 3)

     39)  Pg 1-4 : Exclusion of storm sewer, sanitary sewer, sump
and drainage basin sediment radionuclide data at Linde from the "no
action" risk assessment underestimates risk resulting from
plausible scenarios involving sewer or maintenance workers or
future reconstruction activities.  Lack of relevant sediment data
for the drainage pathway from Linde (large 9 ft by 7 ft
twin-conduit underground pipe that discharges to Two Mile Creek
through the face of Sheridan Park Lake dam) is not sufficient
reason to exclude this pathway.  Existing contaminated sediments
and future releases to Two Mile Creek may be masked by larger
sediment loadings from upstream.  These contaminated sediments may
be uncovered and reconcentrated by stream flows in a manner similar
to that described for Cattaraugus Creek in the September, 1991
"Aerial Radiological Survey of the West Valley Demonstration
Project and Surrounding Area" performed in 1984 by EG&G.

     40)  Pg 1-5 : Why is  the inadequately characterized drainage
ditch along the Ashland 2 access road leading to contaminated
location SP-7 not included as a radiological source?

     41)  Pg 1-6 : Despite the fact that buried Seaway areas `B'
and `C' are likely to be large contributors to contaminated surface
discharges from the landfill (principally via the reinforced
concrete pipe [RCP] carrying Rattlesnake Creek beneath Seaway to
Niagara Mohawk and Ashland 2 - this part of the RI work plan was
not completed), this source is also not included as a radiological
source.  Why not?  The lack of chemical sampling at Seaway is a
major data gap which precludes any meaningful assessment of
chemical risks at this property.

     42)  Pg 1-8 : Ra-226 and uranium are given as the "principal
radioactive materials in the liquid effluent (Aerospace 1981)." 
What activities of other radionuclides , such as Th-230 and Th-232,
were present?  The Curie amounts of uranium and Ra-226 given as
having been released to the storm drains, sanitary sewer and
injection wells indicate that the average concentration of Ra-226
released to the injection wells and storm drains was about 3.5
times higher than that released to the sanitary sewer.  Is this
correct?  Does it mean that predominantly African ore effluents
were discharged via these routes?  There is no risk analysis for
these unaccounted storm drain and sanitary sewer discharges.  No
mention is made of the Two Mile Creek dredgings in the radiological
source and risk analysis.  In each case, why not?

     43)  Pg 1-12 : Time Period : It is stated that "Because DOE is
responsible for the cleanup of the site and is committed to
pursuing a timely response, the time period considered as the
hypothetical future in this assessment of risks for the "no action"
alternative is the immediate future (i.e., the next 150 years). 
Thus, further dispersal of contaminants that would occur over very
long time periods has not been considered in the BRA."  This is an
almost insignificant time period relative to the hazardous life of
the MED radiological contaminants.  The effect is to discount
future dispersal and associated increases in exposures.   The
statement also reflects inadequate characterization of the 
properties and the limitations involved in the application of
layers of assumptions based on site homogeneity.  Also, this
assessment assumes "the retention of existing institutional
controls (e.g., access restrictions and monitoring) up to the next
100 years."  In view of the area land use master plan, this
assumption is not conservative at any of the properties.

     44)  Pg 2-2 : Radiological Data Evaluation:  Both aerial gamma
mapping and ground gamma surveys are referred to but no references
are given either here or in the RI.  Please identify these reports
so that they may be reviewed.  It is unclear whether the ground
surveys utilized protocols (instrument sensitivity, grid spacing,
counting times, etc.) which would insure 100% surface coverage.  A
major concern is non-detection of subsurface contaminant deposits
due to shielding by overlying soil material (Pg 5-8).

     45)  Pg 2-2 : Th-230 is not included in the list of
radionuclides (Th-232, Ra-226, U-238) whose soil concentrations
were analyzed.  It is assumed that this is an oversight.  (See
general comment 3)

     46)  Pg 2-2 : The aggregation of sediment and surface water
data across all drainages may significantly under-represent the
risk in the most contaminated area for the transient child
scenario, even in the RME case.

     47)  Pg 2-2 : Because the American ores were subject to
pre-processing involving vanadium/radium extraction, the assumption
of equilibrium activities for Ra-226 through Po-210 is not strictly
correct.  Lead-210 (t1/2, 22 years), having undergone only two
half-lives of decay since pre-processing, will still be present at
a higher level than the assumed equilibrium activity level
depending upon the extent of radium removal.

     48)  Pg 2-3 : Background :  The selection of Ashland 2 South
for background soil levels is not representative of area-wide
background as determined by ORAU (see comment 24).  This biases the
risk assessment process and results in underestimates.  Selection
of sampling well B29W05D at the southern boundary of Linde (located
within 1000 feet of the injection wells) as the background location
for groundwater may result in similar bias.  While the well is
supposedly upgradient from Linde, groundwater flow in the area is
somewhat indeterminate.  Both Wehran (1979) and Yager, Tepper, and
Kappel ("Hydrogeology of the Niagara Falls Area - a Survey of the
USGS Study" presented at the International Symposium on Groundwater
Issues of the Lower Great Lakes, November 7-8, 1991) indicate that
groundwater flow in the contact-zone aquifer is in a southerly
direction.  Wehran suggested the possibility that the southward
flow is induced by the pumping of several large industrial wells
located to the south.

     49)  Pg 2-5 : Chemical Data Evaluation:  Restricting
evaluation of chemicals of concern to the surficial soil horizon (0
to 2 ft depth) does not consider the pathway of lateral migration
in the perched system, a significant pathway at Seaway and Ashland
2.  Although Pg 5-2 indicates that "interim or provisional SFs
[would be] obtained for use in the risk characterization (EPA
1992d)," fourteen chemical contaminants of concern, including lead,
were eliminated from the chemical risk assessment because "toxicity
values could not be obtained" (Pg 5-12).  For the purposes of NEPA
review, this is unacceptable.

     50)  Pg 2-18 : Table 2-2:  In view of comments 31 and 48,
Ra-226 should have been included as a groundwater contaminant of
concern.  It can be shown that 39 billion cubic feet of water would
be required to dilute the estimated 5.5 Ci of Ra-226 which were
injected into the Linde wells to the level of DOE's drinking water
guideline of 5 pCi/liter. [Thirty-nine billion cubic feet
translates into eight hundred ninety thousand acre-feet, a volume
roughly one-tenth the surface area of Lake Erie to a depth of one
foot.]  In terms of the groundwater that has flowed past the
injection wells in the last 50 years, rough calculations show that
this flow is insufficient by at least a factor of 100 to dilute the
Ra-226 activity to a level below the guideline.

     51)  Pg 3-13 : Current and Future Use Scenarios:  The
selection of an older child (ten years old) as the receptor for the
Ashland 2 and Local Creek scenarios does not take into
consideration the likely presence of younger and correspondingly
more sensitive children as receptors.  The use of DCFs (dose
conversion factors) based on the Reference Man (70 kg adult male)
in determining dose exposure for these scenarios is inappropriate. 
Why was this done?  Younger children exhibit a much higher rate of
adverse effects per unit of exposure.  The short exposure duration
of six years is also  not a realistic assumption.  The exposure
time of the wading child, only seven hours per year, in the Local
Creek scenario is absurdly low. The RME inhalation rate for the
employee, 20 m3/d, can reasonably be expected to be the same as the
rate for the transient adult, 24 m3/d.  Why was 20 used?  The
future exposure durations for the Ashland 1 and 2 sites (7 years,
average; 25 years, RME) and the Seaway area (6 years average; 24
years RME), and the exposure frequency for Seaway (100 days/year)
are very unrealistic in relation to likely future human use
scenarios including the possibility of human habitation in these
areas.  As a result, the risks given in Table 5-1 are
underestimated and significantly misrepresent the actual hazard
present at the sites.

     52)  Pg 3-15 : Identification of Exposure Pathways:  Migration
of contaminants through groundwater is not considered an exposure
pathway because the affected aquifer is not considered potable. 
However, current use of area wells for lawn watering, etc., and
potential future uses are ignored.  Exclusion of this pathway from
the risk analysis is not justified.

     53)  Pg 3-19 : Exposure Point Concentrations:  The assumption
that soil ingestion occurs from only the upper two feet of land
surface ignores the occurrence of higher subsurface concentrations
in many areas and the likelihood that these subsurface soils will
be exposed as a result of future human activity.  It is stated that
"No external gamma exposure rate measurements were available for
site properties," yet several ground-level gamma surveys are
referred to in the RI.  Why wasn't this data used in place of or to
supplement the RESRAD modeling?  The lack of air monitoring for
radon is a serious data gap.  For the Local Creek scenario, the
assumption of constant surface water and sediment contaminant
concentrations in the future is not conservative based on the
historic record to this point, and anticipated migration in the
future ("no action" scenario).

     54)  Summary of Radiological Exposure Estimates:  With the
exception of Figure 3-14, Figures 3-7 through 3-20 contain a key
denoting the 10-25 millirem/yr exposure range as simply a white
area indistinguishable from the page.  This is confusing and
reduces the immediate usefulness of the document. (See general
comment 3)

     55)  Pg 3-54 : Table 3-1B, Mean Radionuclide Concentrations in
Subsurface Soil:  The concentrations given for Th-232 and daughters
are obviously in error.  (See general comment 3)

     56)  Pg 3-57 : In view of the preceding comments, the dose
summary presented in Table 3-2 may significantly underestimate the
total doses for the current, and especially the future use
scenarios.  In any case, the current total dose to receptors at
Linde subarea `A' and future doses at all properties are signifi-
cant - up to 660 millirem/yr, RME at Ashland 1 area `B'.

     57)  Pg 4-1 : Radiation Toxicity:  It is claimed that "The
risk of serious genetic effects is much lower than the risk of
cancer induction," and therefore genetic effects are not
considered.  Genetic effects were included in the risk assessment
at the Niagara Falls Storage Site, and the EIS for that site
("Long-term Management of the Existing Radioactive Wastes and
Residues at the Niagara Falls Storage Site, April 1986") has tables
on page 4-47 that report the rate of genetic effects to be as great
or greater than the rate of fatal cancers.  Exclusion of genetic
effects from this risk assessment, especially given the indefinite
duration of the hazard, is a serious shortcoming.  Failure to
include non-carcinogenic effects in this assessment under-
represents the total risk and associated societal costs.
     The statement that "Studies of populations chronically exposed
to low-level radiation, such as those residing in regions of
elevated natural background, have not shown consistent evidence of
an associated increase in the risk of cancer" implies that there is
a safe dose or threshold below which no effects are produced.  BEIR
V, a reference source for this BRA, found just the opposite: the
best fit of the data for the dose-response curve shows no threshold
for the effects of ionizing radiation.

     58)  Pg 4-2 : The sentence "Thus for sites like Tonawanda,
where all exposures are longer term, it is likely that immediate
effects would be observed" would make more sense if "likely" were
replaced with "unlikely".  (See general comment 3)

     59)  Pg 4-3 : The application of a dose rate effectiveness
factor (DREF) of 2 to the BEIR V risk factor results in a
population-weighted  cancer induction risk factor  used in  this 
analysis of 6 x 10-7/millirem.  Independent radiation experts have
presented a strong case that this risk factor and the application
of DREFs are  not reflective of the evidence. For example, evidence
presented by John Gofman supports the use of a sixfold larger risk
factor.  DOE and DOE predecessor agencies have consistently
underestimated effect rates at low doses.

     60)  Pg 5-57 : Table 5-1, Total Radiological Risk Summary:  It
would appear that there are errors in this table other than the
mislabeling of risk as "(mrem/yr)".  The average risk for Linde
area `A', reported as 6.5E-05, appears to be calculated based on
the Ashland employee exposure duration, 7 years.  Shouldn't the
average risk be 2.0E-04, based on the 22 year Linde exposure
duration parameter?  This brings the risk above the EPA accepted
range of 1.0E-04 per lifetime (or 1.0E-06 per year, the Supreme
Court recognized de minimis risk).  There appear to be other errors
in this table as well. 
     A more realistic appraisal of the inherent radiological hazard
present at these sites can be obtained by applying a
scenario-appropriate factor (to convert to round-the-clock
exposure) and the factor 6 (as explained in comment 59 above) to
the Table 5-1 values.  For example, in the future RME (reasonable
maximum exposure) scenario, round-the-clock exposure at Ashland 1
area `B' carries an incremental risk of 2.6 x 10-1 (9.9E-03 x 6 x
4.4 [to convert 100 hours per year to round-the-clock exposure]). 
This is roughly a 26 in 100 risk of cancer or 26 times EPA's
background cancer risk estimate.

     61)  Pg 5-28 : Table 5-2, Summary of Chemical Risks -
Carcinogens : Failure to include many of the MED-related metals,
including lead, in the risk calculations due to lack of toxicity
data, renders this summary of chemical risks virtually useless. 
The use of interim toxicity values for this evaluation, while
perhaps not ideal, would have at least provided a better picture of
this risk component.  Why wasn't this done?  Why weren't the
chemical and radiological risks combined to present an overall risk
profile?  In view of comment 27 and the failings of this BRA
identified in this and preceding comments, actual site risks appear
to be significantly underestimated.

     62)  What was the original source term in 1946 for this site? 
What is the best percentage estimate of the original source term
activity which will be left behind at each of the properties
following Alternative 2?


     63)  Mixed Waste : The assumption "that the waste generated
during remediation will not be hazardous under RCRA definition" is
not justified.  The RI failed to adequately resolve this issue. 
Where would mixed wastes go?

     64)  In assessing remediation alternatives for the Tonawanda
site, the greatest consideration must be given to: 1) the
essentially indefinite hazard period associated with MED wastes, 2)
a realistic assessment of the risks presented by the wastes at
Tonawanda given that any form of institutional control can be
expected to be lost and dense human habitation to occur long before
the hazard ceases, and 3) the notable inability of clay
containments in humid environments to provide waste isolation for
the required time period (in this case 500,000 years).
     A clay `cell' at the Tonawanda site may be protective for 200
years barring excursionary natural events such as earthquakes,
floods, etc.  Engineers' optimistic claims of 1000 year cell
longevities are interesting.  History indicates that such claims
are ludicrous.  On the basis of this experience, the need to
remediate a Tonawanda `cell' is likely to occur in the
not-so-distant future.  At that time, however, the contaminated
volume will be much greater.  It will include the original clay
containment material and, once again, surrounding soils and
sediments, not to mention groundwater.  The relevant question now
is, will we then be able to cope with such large, contaminated
     Trench disposal with wind-erosion protection in an arid,
unpopulated desert environment can reasonably be expected to
eliminate water-borne migration of contaminants and require little
maintenance or intervention for thousands of years, while
maintaining the originally deposited volume virtually intact.  Up
front transportation costs to relocate the wastes shrink in
comparison with the long term difficulties (noted above) involved
in maintaining isolation in Tonawanda.
     An increase in radon emissions at a sandy, arid disposal
location (Hanford, Washington) has been raised previously as an
objection to the relocation of MED wastes (EIS for the Niagara
Falls Storage Site, 1986).  However, that scenario proposed using
only 1.5 meters of native sand as cover material.  The use of finer
soils or clays as cover material was not addressed; nor was the use
of a thicker cover layer.  The selection of such engineering
options combined with waste emplacement at topographical locations
subject to wind-borne desposition, rather than removal, should
effectively eliminate radon release.
     For all of these reasons, Alternative 2 with relocation to
Clive, Utah (Envirocare) or the Nevada Test Site is the only
acceptable management option.

     65)  As previously noted (comment 11, Wehran 1979), leachate
from the Seaway landfill containing landfill contaminants has been
detected in samples collected outside the landfill.  Decay of large
amounts of organic material present in the landfill is likely to
result in subsidence with associated loss of cap integrity.  The
release of biogenic gases is also likely to entrain radon resulting
in potentially significant releases.  In view of these facts, both
the "accessible" and the "access-restricted" wastes must be removed
along with the other wastes.

     66)  Page 4-5 of the FS mentions the possibility of
implementing institutional controls at Seaway through purchase of
the property.  10CFR61 limits the time during which institutional
controls can be applied as a management tool to a 100 year period. 
This is a negligible period in relation to the duration of the
hazard and therefore should not receive consideration as a
management tool.

     67)  Penetration of burrowing animals into either the
containment `cell' or unremediated subsurface areas is a potential
problem that has not been addressed.  DOE must show that its
capping materials are impervious to such animals.

     68)  DOE's figure for rail transportation costs involved in
relocating the wastes to Clive, Utah has been found to be
significantly inflated.  At the January 26, 1994 public hearing in
Kenmore, N.Y., CANiT reported obtaining a vendor figure $13 million
less than the DOE figure.