V-/EPA
United States
Environmental Protection
Agency
                          Office Of Solid Waste
                          And Emergency Response
                          (5101)
EPA 550-R-96-002
September 1996
Expert Review Of EPA
Chemical Accident
Investigation Report

Terra Industries, Inc.
Nitrogen Fertilizer Facility,
Port Neal,  Iowa

-------
                                                  EPA  550-R-96-002
                                                    September  1996
                     EXPERT REVIEW
                            OF
    EPA  CHEMICAL ACCIDENT INVESTIGATION REPORT
TERRA INDUSTRIES, INC.  NITROGEN FERTILIZER FACILITY
                        January 1996
          Chemical Emergency Preparedness and Prevention Office
             Office of Solid Waste and Emergency Response
                U.S. Environmental Protection Agency
                      Washington, DC 20460

-------
                                    CONTEXTS

Foreword


Expert Review:  Chair's Report
      Names of Reviewers
      Executive Summary, Background, Review, Recommendations
      Appendix I   Agenda of Reviewers Meeting
      Appendix II  Reviewers' Individual  Comments
      Appendix III Reviewers' Additional Remarks

EPA Response to Expert Review Recommendations

Appendix A: EPA's Charge to Reviewers

-------
                                      FOREWORD
    This document presents the comments and recommendations of five expert independent
reviewers, with whom EPA contracted to examine an EPA Chemical Accident Investigation
Report,  "Terra Industries, Inc. Nitrogen  Fertilizer Facility, Port Neal, Iowa." The EPA
investigation report was written by an investigation team at its Region VII office in Kansas
City and published in January 1996.

    The Clean Air Act Amendments  of  1990, Section  112(r), mandated the creation of an
independent Chemical Safety  and Hazard Investigation Board (CSFflB) to investigate chemical
accidents and recommend steps to reduce the risk and hazards of chemical releases.  However,
the CSHLB was never formed. In January 1995,  the Administration asked EPA and OSHA,
under their own existing authorities, to investigate chemical accidents and issue public reports
containing recommendations on what the government,  industry, and other stakeholders could
do to prevent similar accidents from  occurring in the future. The EPA Terra Industries
Investigation Report is the first such report, dealing with a chemical accident in Port Neal,
Iowa in December, 1994. Since Iowa is  one of 23 States  having an OSHA State Plan: the
federal OSHA was not a joint investigator with EPA in this case.

    In the spring and summer, 1996,  EPA assembled a group of experts,  charging them  to
examine the scope, approaches, and methods of this first report to guide future studies and
investigations.  Dr. Paul Hill of the National Institute for  Chemical Studies served as Chair of
the review group.

    The five reviewers independently examined the EPA report, documenting their analyses in
written comments.  The Chair then circulated all  comments to all reviewers, and convened a
meeting of the reviewers at EPA's regional offices in Kansas City to enable them to share
their comments from their different perspectives,  to question the authors of the report, and to
examine photographs and other available documents.   The Chair then wrote a summary  of the
meeting and recommendations of the reviewers.  This  document includes copies of the
reviewers' comments,  the Chair's summary, EPA's charge to the reviewers,  and EPA's  reply
to the reviewers' recommendations.

-------
Expert Review: Chair's Report

-------
National  Institute
  for Chemical Studies
         A REVIEW OF USEPA'S
CHEMICAL ACCIDENT INVESTIGATION REPORT:
        TERRA INDUSTRIES, MC.,
      NITROGEN FERTILIZER FACILITY
           PORT NEAL, IOWA
   This document was prepared with support of
    USEPA Order Number 6W-4075-TASA
               by
     National Institute for Chemical Studies
            August, 1996

-------
                    REVIEWERS
           DR. ISADORE (IRV) ROSENTHAL
         The Wharton School of Decision and Risk
               University of Pennsylvania

                MR. JOEL R VARIAN
International Association of Machinists and Aerospace Workers
                  AFL-CIO (Retired)

            MS. PAMELA NIXON, MS., MT.
             CMA Public Advisory Committee

               DR. GERALDINE V. COX
                     AMPOTECH
                  AMPOTECH  Poland

             DR. PAUL L.  HILL (Chairman)
                        NICS

-------
                                 EXECUTIVE SUMMARY
       The NICS and Dr. Paul L. Hill, in conjunction with, Drs. Irv Rosenthal and Geraldine Cox,
Ms. Pamela Nixon and Mr. Joel Varian, were requested to conduct independent reviews of EPA's
investigative report on Terra Industries  1994 accident.  In order to improve future reports as well as
future efforts to systematically collect data at accident sites, the review team offers the following
major recommendations for consideration:

       EPA should;

            •        Include time lines in future reports.
            •        Expand and continue to model the scenario-by-scenario approach.
            •        Adopt or specify rigorous technical procedures sanctioned  by  the  engineering
                     and research  communities.
            •        Develop a defined protocol for accident investigations.
            •        Hold public meetings to seek stakeholders' input on the protocol.
            •        Consider accident oversight  committees  at affected sites which include public
                     liaisons.
            •        Initiate agreements with other federal, state, and local entities with accident
                     response authorities or consider legislative recommendations to accomplish
                     same.
            •        Create increased public and private awareness of it's  investigative program.
            •        Clearly articulate the national goals  and criteria for accident investigation.
            •        Draw upon the existing experience of NTSB and other agencies for assistance
                     in  the evolving program.

       Industry should;

            •        Take note and seriously address EPA's January 23, 1996 recommendations
                     for  accident prevention.
            •        Initiate  greater awareness  of process  safety regarding  ammonium nitrate
                     through  the  research and  engineering communities.
            •        Proactively  embrace mechanisms  for accident  prevention.

       States and Communities should;

            •        Initiate dialogue with industry and EPA to construct  effective protocols
            •        Consider agreements for resource and authority coordination.

-------
                     BACKGROUND AND STATEMENT OF PURPOSE
       At approximately 0606 hours on December 13, 1994, an explosion occurred in the ammonium
nitrate plant at the Terra International,  Inc., Port Neal Complex. Four persons were killed as  a direct
result of the explosion,  and 18 were injured and required hospitalization. The explosion resulted in
the release of approximately 5,700 tons of anhydrous  ammonia to the air and secondary containment,
approximately 25,000 gallons  of nitric  acid to the ground and lined chemical ditches and sumps,  and
a large  volume of liquid ammonium nitrate  solution into secondary containment.  Off site ammonia
releases continued for approximately six days following the explosion and drifted several miles.
Chemicals released as a result  of the explosion have resulted  in  extensive  environmental
contamination  including groundwater  under  the facility.

       The U. S. Environmental Protection Agency (EPA) Region VII  was directed by EPA
Headquarters to conduct an investigation to determine the cause of the explosion and to develop
recommendations that  would help prevent similar occurrences in ammonium nitrate production
facilities in the future.   A report released by the Agency on January 23, 1996 contains conclusions
reached  by the EPA chemical  accident investigation team regarding the cause of the explosion at the
Terra International, Inc., Port Neal Complex and recommendations for prevention of future similar
occurrences.  The investigation team from EPA was led  by On-Scene coordinator (OSC) Mark
Thomas, Ph.D. of the Region VII Office  with additional assistance defined in the report.

       Shortly after the issuance of  the report, EPA Headquarters initiated discussions with the
National Institute for  Chemical Studies (NICS) to develop  an independent review of EPA's
investigation and  findings.  As an independent non-profit  organization  with environmental,  industry,
labor and  community advocate constituents,  NICS  has a reputation for objective reports  on  chemical
accident prevention and preparedness. On March 5, 1996, EPA reached agreement with NICS to
oversee the review and designate Dr. Paul  L. Hill as chairman of a proposed panel of individuals who
would conduct the review.  Panelists were  selected on the  basis of their expertise  in  process
engineering,  chemical safety,  previous  accident reviews and  management disciplines as well  as  their
broad representation of different stakeholder perspectives.  NICS developed a slate  of potential
reviewers and provided  the list to the  Agency who selected and contracted with four  (4) individuals
in addition to Dr. Hill.  Reviewers were  not asked or retained to conduct independent research in
order to supplement their technical knowledge and professional judgement or to verify the  technical
information contained in the EPA Terra Industries Report.  Reviewers and  their affiliations are listed
on the previous page.

-------
        Among the charges of the Chairman were to provide copies of the report to the reviewers,
solicit their written  comments on  the report, develop  his  own  critique,  distribute  all five (5)
commentaries to each of the participants as well as EPA, organize a meeting to discuss their reviews,
chair the meeting and provide a final report of collective findings and recommendations to the
Agency. After the receipt of individual, initial  comments in July,  1996, the  Chairman called a
meeting on July 28-30, 1996 at the EPA Regional Office in Kansas City, Kansas.  The  review  meeting
was attended by all five (5) members of the panel as well as EPA staff. (See Appendix  1).

       The purpose of this meeting was not to  form consensus on the issues  of causality or absolute
recommendations to the agency.  Rather, it was a forum to exchange ideas about the report's findings
and probe the records and recorded testimony for additional clarifications.  Agency staff were present
by request of the Chairman to respond to questions and provide details of data collection,  procedure
and scenario development used  to compile the report.  The charge of the review panel members was
quite narrow: (a) to assess the plausibility of the report findings based on all evidence collected by
the agency and (b) to make recommendations on procedure, technique and  report formulation which
would improve future Agency accident investigation products.

       The review team considered  only the immediate information surrounding EPA's report,  While
the team was aware of additional reports and documents developed by other parties, these were not
considered germane to the limited charge given by the Agency. In the possession of reviewers was:
(a) the settlement agreement between  Terra International,  Inc.  and Iowa OSHA Employment Appeal
Board and (b) the Terra Port Neal Explosion report dated July 17, 1995 issued by  a group of outside
experts retained by the General Counsel of Terra Industries, Inc.   Technical and legal assertions
raised by the other parties involved in the Terra Industries,  Inc. accident fell  clearly outside the scope
of the review team's charge. No opinions are surmised or offered on these issues.

-------
                                       THE  REVIEW

       Accident investigations and the attempt to reconstruct conditions which lead to an accident
are inherently difficult to pursue. The current report indicates that certain evidence, diagrams, and
requested documents were either destroyed or unavailable for this investigation report. Even with
satisfactory provision of existing management, operations  and training materials, reliance  upon human
knowledge  and  recollections as well as potential nondisclosure makes the job of accident investigation
for root cause all the more difficult.  In light of the Clean Air Act Amendments of 1990, the review
team recognizes that as a nation, significant insights of investigation, and review, must be assessed
to fully implement the Act.

       During  the review meeting,   held in Kansas City, the review team had  unlimited access to
numerous  photographs, drawings,  analyses,  transcripts  and  other evidence and documentation
collected by the Agency for development of the report.  The team did not speak with any employee
of Terra International, Inc. and did not visit the accident site in Port Neal, Iowa.  While a broader
investigation would have logically involved greater efforts to carry out these activities, this review
was limited to the January 23, 1996 report and the in-house information cited above. To  more  clearly
define the limited scope of this investigation report review, an outline of the key questions addressed
by the team are as follows. Reviewers were asked to:

       a.   Comment on the technical  soundness.
       b.   Comment on the approach  scenario by scenario.
       c.   Comment on the findings of the report and the  most plausible scenario.
       d.   Comment on the comprehensiveness  and reasonableness  of the technical  conditions
            under which the accident  occurred.
       e.   Are specific roles of certain  equipment  appropriately considered?
       f.   Is the discussion of ammonium nitrate (AN) appropriate?
       a.   Were all external factors considered in a proper way?
       h.   Comment on the overall conclusions  and recommendations.
       i.    What activities or report components should be modeled for future investigations?
       ;     Were prevention recommendations appropriately presented?
       k.   Were the roles of other entities appropriately addressed?
       1.    Are there additional recommendations for actions that could have been or should be
            taken in the future?

-------
        After providing initial comments and after meeting for a total of more than 18 hours, the
 review team provided a series of comments about the report for the agency's consideration,  These
 comments,  again, do  not represent any absolute consensus  of the team, in that many individual
 stakeholder perspectives are included. However, the team was unanimous in its support that all
 pertinent comments be offered to the Agency.  In addition to individual  comments provided by the
 reviewers (which are included in Appendix 2 of this report), team commentary on the basic issue
 questions cited above are intended to provide a constructive critique. Follow-up comments were
 provided by two  members of the team (See Appendix III).

        APPROACH

        Generally the team considered the overall approach to the report to be sound and appropriate.
 The text was straight forward and lacked overly technical jargon  which was  considered beneficial for
 public policy makers and the general public. Both constituent groups have expressed keen interest
 in this  report and it seems  to be sensitive to these broad audiences. The "scenario by scenario"
 approach used in the report is a valid and useful approach which was also viewed as helpful to the
 reader.

        In response to the question of whether this was the "correct"  approach, this becomes a
 philosophical discussion of the  technical community on process  safety  and  investigations. There exists
 a vast literature on approaches to accident investigation as evidenced by a recent publication by the
 Center for Chemical Process Safety (CCPS) of the American Institute of Chemical Engineers called
 "Guidelines for Investigating Chemical Processing Incidents."   While  this  document and others
 describe numerous accepted approaches for accident investigation such as the one at Terra, the key
 point of the review team was that a referenced, accepted methodology be utilized and clearly
 described by the report (and future reports). After meeting with the OSC  and others involved in the
 investigation in Kansas City, it  became apparent that several methodologies were  considered.  To the
 reader of the report, and there are many, it is not readily apparent that such were utilized for this
 investigation.  All reports in the future should specify the process, procedure or guidelines within
 which the investigation team was operating.

        While the team viewed the "scenario by scenario" approach as  useful and informative, it was
 incomplete.  Team members understood the need to truncate activities into a readable and concise
 document.  However, the omission of the range of scenarios (including  sabotage) and why these were
 dismissed should always be stated. Without questioning the investigators, reviewers had no indication
 whether all scenarios had been considered and why/how same were dismissed.

        The  scenarios presented did  a good job of systematically  narrowing the scope  of possibilities
for  root cause based on evidence and good science.  The  use of a metallurgist was particularly helpful.
 The overall findings  of the report based on these scenarios and the  evidence presented  seem
reasonable to the  entire review team. Given that Terra International, Inc. has presented a report with
 slightly different findings relating to one piece of equipment (the sparger), this scenario could have
been pursued in greater detail  (see sparger discussion).

-------
       PLAUSIBILITY OF EPA CONCLUSIONS

       Overall, the team agreed that EPA's conclusions were plausible given the evidence collected
and presented. It must be noted however that some evidence, samples and data were either destroyed
or never collected  due to conditions  under which the investigation took place. Lack of clear control
and coordination at the site seem to be primarily responsible for this.  As the  final report conclusions
indicate, several conditions at  the plant were outside the  range of standard or safe operating
procedures and parameters and led to this tragic accident.

       The only  reservation raised by the reviewers dealt with the issue  of the  sparger. The
committee was unable to come to consensus on the role the  sparger did or did not play in the  overall
stimulus of the explosion of the neutralizer.  This single uncertainty however, does not negate the
reasonableness of the six conclusions put forward by EPA.  There is significant evidence that
numerous problems existed at the  facility.  The review team's  only concern with EPA's major
conclusions are of definition. That  is, the conclusions themselves are less explanatory of "root cause"
than is the body of the report discussion.   Since root causes are "prime reasons which lead to an
unsafe act or unsafe  condition or constitute an underlying condition  and result in an accident; if the
condition is removed the particular incident would not have occurred." Given this definition, the
management system failures that led to or allowed the  existence of the  unsafe acts  or unsafe
conditions that the  report concludes caused  the accident  was less than adequate. The  report
discussion does a better job of identifying these causes  than do, the conclusion and recommendation
sections  of the report.  Attention to the relationships between root cause and conclusions reached
would assist future readers and future reports.

       COMPLETENESS

       The team  agreed  that the  investigators  made a  concerted effort to  provide complete
information and analyses.  Reviewers recognized that investigators were somewhat challenged  by
the circumstances  of cooperation, authority and  coordination at the site.  Given the conditions  at the
Port Neal facility both before and after the December 13,  1994 explosion,  the investigators did a
thorough  job.  While the review team initially raised questions regarding  various technical  issues  (for
instance, sampling, sources of contamination  of ammonium nitrate stocks and inert coatings of
vessels) these were sufficiently answered by the OSC and members of the investigation staff.

       DISCUSSION  OF CONDITIONS

       Though historical records  are not overly  extensive on ammonium nitrate explosions, the report
discussion of pre-incident-conditions was both valuable and thorough. This discussion helped to
establish the plausibility  of certain scenario building exercises which were undertaken by the
investigators. It also added  value to the basis of discussion for non-technical stakeholders who have
or are likely to review the report.   This type of background search on basic chemistry and literature
should be included in future reports.

-------
        EXTERNAL ENTITIES AND FACTORS

        Given that a clear lack of coordination with other agencies and interests existed during this
investigation, the report does not deal adequately with external factors. Only upon interview with
agency personnel was it apparent to the reviewers just how difficult this issue was. Because Iowa
is a designated "state OSHA" by its parent federal agency, coordination and,  therefore,  consideration
of more complete  information exchange was  greatly lacking.  If detailed, coordinated investigations
and joint reports  are to be  achieved as intended  by the CAAA of 1990,  a broad protocol,
comprehensive in  nature and definitive of the roles of all stakeholders must be developed.   State
agencies, local response  organizations,  industry, labor  and  community advocates  should  be
coordinated by federal agencies into a cohesive, informed and collaborative effort. Externally,  an
interface with  all stakeholders would  ensure accurate information is presented for public consumption.
Due to circumstances surrounding this  particular report,  these significant issues  were not  addressed,

        ADDITIONAL  ACTIONS

        In its release of the report on January 23, 1996, EPA made several recommendations based
on  its  findings during the investigation. These ranged  from recommending thorough process hazards
analyses, to  reviews of safe  operating procedures and increased emphasis on training, communication
and  preventive  maintenance.  The review team supports these  recommendations  and notes  that many
are now contained in OSHA and/or EPA regulations. From this incident, facility management must
recognize the value and meaning of compliance with existing rules and overall safe  management
practices.

        EPA should follow up with  both  the  research  community, trade associations and all
ammonium nitrate producers  to fully inform them of the findings of this report.  Additional research,
if properly structured should address ammonium nitrate sensitivity, confinement and activation. Also,
the emergency response community should be provided with detailed reviews of this accident and
how responders should prepare for responses to similar incidents.

-------
                                  RECOMMENDATIONS

       The Accident Report Review Team recommends that EPA assess its objectives and clearly
articulate a strategy to be commented upon by the various stakeholders. Major problems observed
in the Port Neal, Iowa investigation were  (a) lack of understanding that EPA had authority to conduct
the investigation and (b) lack of coordination  with other parties  who also had interests or authorities
to collect information.  These two  issues precipitated most other  deficiencies found in both the
investigation  and the written report.   Lack of clear authority and direction were the  greatest
hindrances to  EPA's staff.

       The Team's recommendation for a uniform accident investigation policy should allow for
consideration of the numerous  stakeholders including  the  public.  Although the latter would not be
directly  involved  in evidence collection  and technical  work, the concept of an oversight  committee
which includes public representation  should be considered. Those parties  or agencies with existing
authorities should be organized, through  agreements  or statutory changes, into a structured, mutually
beneficial approach. Recognized is the fact that EPA could be executing agreements with 50 states
plus countless local entities. In the Terra incident, a local fire chief, acting  on his independent
authority, destroyed evidence by "hosing  down" the area shortly after the event. Such seemingly
random actions will continue to occur at accident sites until a uniform protocol is issued.

       Timing should be  addressed  also, as the agency's response did not begin until  six days after
the event. While stabilization of the site was necessary and time consuming, the OSC was required
to develop an  investigation strategy and implement that strategy impromptu. By this time, weather,
movement of debris, loss or  destruction of evidence (e.g. the pry bar opening of 416J pump by Terra
personnel) or the actions of other agencies  (e.g.. fire department) had degraded or eliminated
potentially critical  evidence.  The  initial resistance of Terra personnel to take seriously the EPA
investigation team as demonstrated by the 26 site visits also slowed the initiation  and  completion time
of the investigation.

       EPA should consider a national network of response capabilities  and expertise as its accident
investigation program matures. In  an effort to reduce costs, personnel or contract  services with
particular expertise could be integrated into an overall approach at the direction of Headquarters
or among the Regional Offices. The previous recommendations regarding the expertise of other
agencies including  OSHA, states, etc. should also be integrated into this potential network.

       EPA should consider how it might leverage industry response to this and future reports.
Clearly,  other ammonium nitrate producers will review this report for its implications on their
operations.  However, other mechanisms  such as working  with  trade associations or specific groups
like AICHE's CCPS, the Ammonia Institute, and others may provide a comprehensive yet focused
approach.

-------
       Due to the nature of several key issues raised by the review team as well as the limited time
in which they had to consider the report, numerous issues remain unaddressed. The Agency should
continue to deal with  these over time by seeking additional internal  and external expertise  on several
key subjects.  Given the circumstances, EPA and particularly Mark Thomas of Region  VTI, have
made a valuable contribution to the Agency's evolving investigative process.  While inclusion of
analytical protocols, time lines and additional  data may have improved EPA's Terra Industries report,
the content provides plausible support of the  agency's  conclusions and actions. Given the nature of
industrial facilities which handle hazardous materials and existing regulatory requirements, numerous
deficiencies surfaced at  Terra's Port Neal  operation. Based  on the agency's evidence and  the report,
an array of management, equipment, training and safety parameters were pushed beyond their safe
operating range. When this occurs, disastrous consequences are the result.

-------
APPENDIX
    I

-------
                                PROPOSED AGENDA
              TERRA INDUSTRIES - EPA ACCIDENT INVESTIGATION
                                   REVIEW TEAM
July 28, 1996
July 29, 1996
July 30, 1996
Arrive Kansas City
5:00p.m.
8:00a.m.
8:30a.m.  -  10:30a.m. -
Break
10:15a.m.-12:00p.m.  -

12:00p.m.-l:30p.m.  -
l:30p.m.-3:00p.m.  -
Break
3:15p.m.-5:00p.m.

8:00a.m.
8:30a.m-10:30a.m.
Break
10:45a.m.-l:00p.m.   -
Intro/Procedures
Dinner Meeting

Meet  in  Hotel Lobby for
Transportation To EPA

Mark Thomas

Questions & Examination
Material
Lunch
Panel Discussion

Panel Discussion

Meet for Transportation
Meeting with Mark Thomas
                                                             Final Discussion

                                  Depart as individual  schedules  demand

Note: EPA Region VII  and Headquarters staff will be on hand to assist the review team with
information, logistics or other assistance as needed  Private (panel only) sessions are optional on both
days.

-------
APPENDIX
    II

-------
         National Institute
           for Chemical Studies
         for CHEMICAL ACCIDENT INVESTIGATION REPORT:
                 TERRA INDUSTRIES, INC.
              NITROGEN FERTILIZER FACILITY
                   PORT NEAL. IOWA

                 REVIEWER'S COMMENTS
University of Charleston   2300 MacCorkle Avenue, SE.   Charleston, West Virginia 25304

-------
             A Review of

     EPA Chemical Accident Report
          "Terra Industries, Inc.
       Nitrogen Fertilizer Facility"

                 by

        Geraldine V. Cox, Ph.D.

under EPA Order Number 6W-463 7-NATA
       Reference Number EAR203

        Submitted July 15,1996

-------
Under contract to the United States Environmental Protection Agency, I reviewed the
Chemical Accident Investigation Report pntitW  TFT?T?A TNDTTSTKTFS1  TNr
NITROGEN FERTILIZER FACILITY PORT NEAL.TOWA  As background materials,
I also read the State of Iowa Occupational Safety and Health Citation documents and the
Report of the Incident Investigation Committee prepared by Terra on the Port Neal
Explosion, dated July  17, 1995.

EPA charged the reviews to respond to the questions that follow.

1.   Comment on the technical soundness, overall approach, and completeness of the
    report, to derive recommendations for approaches to accident investigations in the
   future and accident prevention.

    The overall approach was appropriate, but the number of possible scenarios was
somewhat truncated. All plausible scenarios should have been identified, and then
eliminated by  data or reasoning. It appears that the "solution" to the cause was identified
early, and not all of the possible causes were identified and retained or eliminated as
appropriate.  Or, if the scenarios were dismissed, it was not reflected in the report.

    The Iowa OSHA citation indicated an evacuation that placed employees in the plume
of ammonia gas for a period of time following the incident.  Understanding the charges to
EPA and OSHA regarding chemical safety incidents, I was surprised to see that the report
did not extend to the actions and plans following the incident. I would have expected to
see a discussion of the emergency response - both company and with local emergency
response teams. Was there any community exposure? Was there a review of the entire
post-incident events? If not, this should be included in the review. While information
from the response will not go to prevention, lessons learned from the response will be
valuable in the future to mitigate the effects of similar incidents.

    The description of structures surrounding the plant is lacking. If there were no
surrounding structures, that should have been indicated.

2.   Comment on the approach taken (scenario by scenario) as a correct approach to
    take.

    Scenario 1 -  AN Plant Pumps

    The discussion of the AN scrubber recirculation pump and the product pumps, based
on data presented, is plausible. I concur that the pumps were not the site of initiation.

    Scenario 2 -  AN Scrubber

    This  analysis seems valid based on the data presented. The AN scrubber was not a
source of the explosion.
Review of Chemical Accident Investigation Report              by Geraldine V. Cox, Ph.D. 07/14/96
Terra Industries, Inc. Nitrogen Fertilizer Facility Port Neal Iowa                        Page 1 of 6

-------
    Scenario 3  -  Neutralizer

    All data presented are consistent with an explosion originating in the neutralization
vessel. There was no discussion of the lining of the neutralization vessel, if any existed.  If
the vessel was not lined with Teflon®, then the prolonged exposure to a pH of less than 1
might have introduced a local increased concentration of iron or chromium from the steel
tank that might have contributed to the reaction.  There was no discussion of the materials
of fabrication of the pipes, and this might have contributed to activation of the reaction.
Since several of the metallurgical analyses indicated pitting, some degradation of the
vessel must have occurred before the explosion.  Whether this contributed to the
explosion,  or not, is unclear. In the report, there was a discussion of the abandoned steam
heating lines due to corrosion.  What was the material of construction for the pipes, and
why did they corrode so badly that the steam lines could not be used? Was there a leak
that was causing the corrosion? If so, could that  have introduced contamination?

    Under item 2. of the determination of conditions prior to the  explosion, the presence
of chlorides is mentioned. Was there a higher level of chromium and/or iron? These
compounds can also contribute to the explosion.

    Spargers

    The report did not discuss the "bathtub ring"  that was mentioned in the Terra report.
Was any evidence of this "ring" found in the sample analyzed by EPA labs? There is a
clear  discrepancy between the EPA and Terra report on this item.  The EPA metallurgist's
report specified that the  force originated outside  of the sparger, not inside. Were
sufficient samples provided to the metallurgist to assure that the event was not triggered  in
another section of the sparger.   Since it appears that a single sample was analyzed, and the
ring had a rather large diameter, although unspecified in the report, could the initiation
been in another section of the sparger that was not analyzed?

    The discussion of titanium "healing"  is accurate and would argue against titanium
being an initiator, since the "reactive material" would be the relatively inert titanium
dioxide.

    Neutralizer contents

    Clearly the low pH, <1, contributed to the sensitization of the ammonium nitrate.
Since the pH sensors were isolated relative to the central portion of the reaction vessel, it
is difficult to understand if the contents were uniformly distributed, or channeled as
proposed by  EPA. No one argues that bubbles were absent, and that the bubbles may
have contributed to the event. The scenario of local areas of convection with low pH,
contamination and bubbles is a reasonable explanation for initiation of the explosive
reaction.
Review of Chemical Accident Investigation Report              by Geraldine V. Cox, Ph.D. 07/14/96
Terra Industries, Inc. Nitrogen Fertilizer Facility Port Neal Iowa                          Page 2 of 6

-------
5.   Comment on the findings of the report examining various explosion scenarios and on
    the identification of the most plausible scenario.   Were any significant scenarios
    missed?

    The scenarios presented seem plausible, but the development of possible accident
scenarios seems somewhat truncated. For example, the normal approach to such an
accident investigation would normally  list  all possible, and often some improbable causes,
and then a method to eliminate those without substance based on data from the event.
This approach would structure the data collection effort, and would take place prior to
chemical and physical sampling at the site. One scenario, for example, should always be
sabotage. Most probably sabotage can be  eliminated in this case, but  it should always be
included in the analysis.  For example, the chloride ion presence in the AN 83% storage
tank could have been introduced deliberately.  (While I see no evidence that this occurred,
it cannot be ruled out.) Since the stainless steel neutralization tank was constructed from
304L stainless steel, with a high chromium content, 19%, combined with the fact that a pH
of less than 1 may have been present for a significant time, would possible chromium
contamination be a contributing factor. Chromium (and its oxo complexes) is a
sensitizing factor in ammonium nitrate decomposition reactions. If the vessel were
Teflon®-lined to prevent such solution of chromium and iron, it is not indicated in the
text. For that matter, iron could possibly contribute as a sensitizing agent as well.

    The chloride contamination scenario does not seem to have adequate explanation. The
normal chloride content of water is less than  100 ppm, so the 557 ppm found in the nitric
acid absorption column seems very high.  The  description of the surface
condenser/absorption column was insufficient to understand how 557 ppm of chlorides
could accumulate.  To contaminate the ammonium nitrate storage tank to the level of 157
ppm, a lot more chloride than found in drinking water would  be necessary. Did anyone
calculate the total amount  of chloride necessary to contaminate the Ammonium Nitrate to
a level of 157 ppm in the tank? I did not find the total capacity of the 83% Ammonium
Nitrate tank or the estimates of the amount of Ammonium Nitrate present in the tank at
the time of the incident.  This would have been useful to determine the total amount of
chloride. If the volume of the tank were significant,  then the amount  of contamination
from chloride would be quite large, especially if it came from water.

    Titanium is embrittled by ammonia. Was it possible that a high ammonia excursion
occurred sometime in the history of the system? With the pH probe function under
question, it is possible that the reaction was not controlled to a level where some
embrittlement did not occur - although it appears that the pH excursion was on the acid,
not alkaline range.

4.   Comment on the comprehensiveness and reasonableness  of the statements about
    technical conditions under which the accident occurred. Are  specific roles of certain
    equipment, notably the sparger, appropriately considered?
Review of Chemical Accident Investigation Report             by Geraldine V. Cox, Ph.D.  07/14/96
Terra Industries, Inc. Nitrogen Fertilizer Facility Port Neal Iowa                        Page 3 of 6

-------
    As noted in the discussion of scenarios, the "bathtub ring" that Terra claims in the
 sparger pieces found after the explosion, and from the sparger removed from the vessel
 previously, this might have been a contributing factor. Was the sparger welded to form
 the shape? If so,  what was used in the welding process. Could it have contributed to the
 failure? If the EPA conclusions about sparger integrity  were based on a single sample,
 and that sample was obtained from the other side of the  vessel from the site of initiation, it
 might not have the same characteristics.  Therefore, multiple samples from different
 locations of the sparger should have been analyzed.  From the report it is unclear if
 multiple samples  were  analyzed by EPA.

    While the discussion of the materials of construction for the vessel and sparger were
 good, the dimensions were absent. Also, the materials of construction for the piping of
 the vessel were missing. The piping might be a possible source of contamination. Were
 the vessels or pipes lined with an inert material such as  Teflon®? Thickness of the tank
 walls and sparger were not presented in the description. This would have been helpful.

 5.   With no prior history of accidental detonating of AN solution, is the discussion of the
    conditions existing before the accident appropriate? Please comment on  the
    conclusions reached and whether they lead to the root causes identified..

    The discussion of the pre-incident conditions was very thorough. The discussion
 could have been stronger if the sizes of the involved vessels were identified, as well as the
 materials of construction of the pipes, and an indication of whether the neutralization tank
 was lined  and the thickness of the respective vessels.

    In Figure 2-2,1 believe that the arrow head is pointing in the wrong direction on  the
 line connecting the urea and  83%  AN sales  line.  I believe the AN flow goes to the
 nitrogen solutions storage where it is mixed with urea.  In fact, the line should go directly
to the storage tank, where I believe it is mixed with urea, rather than flowing  into the urea
 line.

    The possible contamination of the reactive materials should have included chromium,
 and possibly iron. The discussion of sensitizing agents mentions these materials, and there
 is a possibility that they could be present due to construction materials, so they should
have  been  included,  A discussion of the  concentration of these contaminants as
 concentration  alters  the  reaction would have been helpful.

 6.   Were all externalfactors  considered in  a comprehensive way?   Also, please  comment
    on whether human factors and broad management issues were identified
   appropriately, in appropriate sequences,  in ways to lead to reasonable conclusions
   and recommendations.

    As stated earlier, the report should have included a review of the actions taken after
the incident. This should have included employee evacuation procedures and the role of
Review of Chemical Accident Investigation  Report        by Geraldine V. Cox, Ph.D. 07/14/96
Terra Industries, Inc. Nitrogen Fertilizer Facility Port Neal Iowa                      Page 4 of 6

-------
the plant personnel with local emergency response and planning groups and any
community  surrounding the plant.

   The mention of the corroded steam pipes, to the point where the pipes were unusable,
bothers me. What caused the corrosion and could the leaks have contaminated the
system. From the limited description, it is impossible to understand if they contributed to,
or were symptomatic of other problems that this report did not identify.

   The high concentration of chloride seems to be hard to explain on the basis of water
contamination.  The mass of chloride ion should be calculated to understand the extent of
contamination.  Is there a possibility of sabotage? From the operations history, I do not
believe there was sabotage, but it should still be considered and ruled out based on the
data.

   The pre-incident report mentions an "outside operator" who was present, and is
present in startup and shutdown operations.  It was unclear if this individual was a Terra
employee or a contract employee. Did this individual receive the same training as the
Terra full-time  operators of the ammonium nitrate unit? What is the full role of the
outside operator, and what is the familiarity of this individual with the process? This could
possibly be a source of error,  especially because Terra apparently did not have written
shutdown and startup procedures.

7.  Please comment on the conclusions and recommendations.   What of the overall
   approach could be modelled for future such investigations?   Was information for
   prevention of similar accidents appropriately presented in this report?

   The conclusions are valid, and the rationale for the conclusions seems to be sound.
The recommendations highlight the poor documentation of the facility in terms of written
procedures. If the facility had been conforming to OSHA 1910, this incident would most
likely not have occurred. However, the possible degradation of the sparger might have
contributed to the system failure even if these procedures were in place. There is no
reason why a suspected faulty pH probe could not have been replaced within the time
period, although if the tank were not discharging, then it would be of little use since it was
outside of the neutralization tank. I question whether the fluctuations in the readings that
Terra employees believe was  due to a faulty probe might have, in fact, been accurate and
the pH fluctuation might be longer than stated in the incident report,

   In recommendation 3., the startup procedure should be specified as well as shutdown

   In recommendation 4., the engineer should not only approve changes, he should sign
off on any changes so the documentation is incontrovertible.

   In recommendation 6., any outside contractors should be included in the training
program. Anyone who works in the unit should have training and understand the
Review of Chemical Accident Investigation Report            by Geraldine V. Cox, Ph.D. 07/14/96
Terra Industries. Inc. Nitrogen Fertilizer Facility Port Neal Iowa                        Page 5 of 6

-------
 procedures. This training should include what type of safety equipment is appropriate,
 and that equipment should be available.

 8.  Concerning the role-offederal, state local agencies; the public; labor; trade
    associations; andpublic interest groups: please comment on whether roles of these
    entities were appropriately addressed.  Are there any recommendations for actions
    they could have taken, or should in future  take to reduce accident risks?

        The role of the local emergency responders was not addressed at all.  The EPA
 report was thorough,  but seemed more appropriate for OSHA than EPA. However, in  a
 quick reading of the Iowa Occupational Safety and Health Department  citation, they did
 not prepare a causative report of the depth of EPA.  This type of report should be done,
 and it is worth an understanding on a national  level when states have OSHA primacy, such
 as Iowa. Perhaps the  federal OSHA should review the EPA report or the state citation  in
 cases where explosions in chemical operations  cause death.

        Labor is not part of the  review  process, although there is a strong national
 program in worker safety run by the AFL-CIO. Perhaps in future incidents, one of their
 technical staff can be included on the review panel for EPA.

        The Fertilizer Institute should receive this report and be  asked to develop an
 approach for its  members to prevent similar incidents. Trade associations  can be valuable
 allies in understanding events such as this and to develop viable approaches to preventing
 similar incidents  or  at  least mitigating the consequences.  There was no  mention of any
 industry-wide practices for this  type of operation.

       No mention of public interest groups was made in the report, however,
 incorporation of the local emergency planning  commission should include an outreach to
 the public interest groups most likely to be effected by the facility operations,
       /
      /
Geraldme V. Cox, Ph.D
Review of Chemical Accident Investigation Report              by Geraldine V. Cox, Ph.D. 07/14/96
Terra Industries, Inc. Nitrogen Fertilizer Facility Port Neal Iowa                         Page 6 of 6

-------
07/18/9612:04  FAX 7668481             CAREPOINT LAB
                                     Review  of EPA
                           Chemical  Investigation  Report
                                 Terra Industries,  Inc.
                               Pamela Nixon, MS MT (ASCP)

       The investigation of Terra Industries, Inc. of Port Neal, Iowa, by EPA appeared to to be
       thorough. The investigation team's recommendations were focused primarily  on the
       ammonium nitrate (AN) unit, due to the > 1 explosion that occurred there.  However, the
       historical data of the most recent incidents that lead up to the event, indicated that plausible
       contributing factors could have come from the feed streams including Area I and Area II
       as  described in Section 2:  Background:

       In Section 10: Recommedations, the investigation team's recommendations (#1 through
       #5  and # 8)  could possibly be augmented  to  include the units of Areas I and II.
       Recommendations #6 & #7 appear to be more broad-based, and not specific to the AN
       unit

       In  reading the recommendations, I have  taken them literally thereby,  possibly interpreting
       them more narrowly than  the investigation team intended  My experience has been to look
       at industrial  accidents from outside the fence-line. When discussing accidents with  company
       officials I have discovered that they respond to the literal interpretation of questions and
       recommendation, and that is why I have chosen this route.

       In Section 2: Background, Area I and Area II are  differentiated into distinct smaller plants  which
       include  the ammonia plant, tank farm, utilities, nitric acid plant, the urea plant,  and the AN
       plant. These smaller plants are either feed streams or they lead-off the AN  plant
       '      Petroleum spills in the ammonia plant (September 1994 and  October 1994), possibly
              causing the  hydrocarbon contamination
       *      Water leak in the Nitric Acid  plant (December 5 to  8, 1994) possibly  being the con-
             tributing factor of chloride contamination

-------
07/16/96   12:04 FAX 7688481                 CAREPOINT LAB
       page 2 of 3

        •     Uncontrolled off-gas flow production of ammonia from the Urea Plant to the AN
              neutralize^ possibly affecting the pH in the neutralizer and being responsible for the
              the ammonia leak during the  several  days  that followed the explosion

        It is EPA's  responsibility  to protect  the public from similar events that may impact their lives
        in an  negative manner.   In order to minimize such occurances, I believe the recommen-
        dations should be inhanced to include the following:
        *     I suggest  that the  PA of the AN plants (recommendation 1) be expanded to include
              PA of all Plants,  storage units, and  utilities as described in Area I and Area II
              Section 2:  Background.  If the AN plant  lacked completed PA, there is  possibly the
              need for  performed on the  Terra Industries facility as a whole, or to  show
              documentation that it has been performed  as changes have  occurred in other  areas.
        *     I suggest that the safe process operating parameters for the AN plants (recommen-
              dation 2)  be expanded to include all  plants, storage units, and utilities as described
              in Area I and Area II  Section 2: Background rather than only monitoring the
              feed streams for the presence of known  contaminants on a periodic basis as well as
              periodically  reevaluating  operating parameter ranges, or provide documentation   that
              it already exists in the other  area.
        "     I suggest  that the  development of the written, safe operating procedures  for the AN
              unit  (recommendation 3)  be  expanded  to include all plants, storage units, and utilities
              as described in Area I and Area II and Area II in Section 2:  Background and the procedures
              should be developed for  activities conducted in each unit in all modes of operation,
              including periods when the units are shut down and the vessels are charged... If
              written, safe operating  procedures exist  in the  other united provide the  documents.
        *     I suggest that 4 complete management of  change procedure for all operating
              parameter range changes for the AN unit  (recommendation 4) be expanded to
              include all plants, storage units, and utilities as described in Area I and Area II in
              Section 2:  Background,  and this should include approval to operate the unit outside
              of approved parameter ranges  by the engineer responsible for these  units, and docu-

-------
07/16/96  12:01 FAX 7668481                 CAEEPOINT LAB
       page 3 of 3

             mentation of these activities - If written, safe operating procedures exist in the
             other units, provide the documents.
       •      I  suggest the development of the  maintenance  program that will  anticipate problems
              in the AN unit (recommedation 5) be expanded to include  all plants storage units,
              and utilities as described in Area I and Area II in Section 2: Background. This
             program should include predictive failure analyses - If this  program exists in the
             other units, provide documents.
       •      I  suggest that information on the hazards of substances handled, the prevention
             measure a in place or planned to prevent releases and the emergency response
             measure a (recommendation 8) be taken for all plants, storage units, and utilities (as
              described in Area I and Area II Section 2: Background) with the State Emergency
             Response Commission, Local  Emergency Planning Committee,  first responders, and
             the public surrounding the facility.  If this information  from the other units has been
             given  to the  above mentioned  agencies,  provide the  documentation.

       I feel that  if the recommendations are not explicit to include the other  units, the programs
       procedures and training in the other units may be overlooked.  Then the potential for an
       incident  of equal  or greater magnitude occuringin  other  units  of Terra Industries, will
       continue to  be list

       There should be a system  in place where by OSHA and EPA can work together on a safety
       board to protect the health, safety, and environment of the workers and the public.

-------
 Review of the EPA Region VII Report on the 12/13/94 Accident
          at the  Port Neal, Iowa Facility of Terra Industries.

                       Isadore Rosenthal, Ph. D., 7/14/96.
Scope of the report.

This report is  organized around the specific areas  that reviewers  were
asked to address in  the  "Charge to Reviewers" and  "  Statement of Work
for Expert Reviewers" (Appendix I).

The  general charge to reviewers was to "use your technical knowledge and
professional judgement to  comment on  the technical soundness, overall
approach, and completeness  of the  report, to derive recommendations  for
approaches    to accident investigations in the future and accident
prevention."   The report on  which the  reviewers were to comment was
Region  VII's report  on the accident at the Terra Industries facility  in
Port Neal,  Iowa (EPA Terra Report).

Reviewers were not asked or retained to do independent research in order
to supplement their  "technical knowledge and professional judgement"  or
to verify the  information  contained in  the EPA Terra Report being
reviewed.

Two additional  documents dealing directly with the Terra  Industries
accident were  supplied to  this  reviewer in connection  with  his
assignment:
         	                       9           	
      1. The settlement agreement between  Terra International, INC. And
      the Iowa  OSHA Employment Appeal Board (OSHA -  Terra Agreement).

      2. The  July  17, 1995  report on  the accident at the  Terra Industries
      facility  in Port  Neal prepared by a group  of  outside  experts  retained
      by the General Counsel of Terra Industries Inc (Terra Port Neal
      Report).

-------
This report  is organized around the following specific items that the
"Charge  to  Reviewers" asked  each  reviewer  comment on:

       1.- Whether the scenario by scenario approach taken was  correct.

       2.- The findings of the report examining various explosion scenarios.

       3.- The identification  of the  most plausible  scenario.

       4.- The omission of  significant  scenarios.

       5.- The comprehensiveness  and reasonableness  of  statements about
       technical  conditions  under which the  accident  occurred.

       6.- The appropriateness  of the consideration  given to the roles of
       certain equipment, notably the sparger.

       7.-   Is discussion  of  the conditions existing  before the accident
       appropriate  with  no prior  history of the accidental detonating of AN
       solution"?

       8.- The conclusions reached and whether they lead  to  the root causes
       identified.

      9.- Were  all external factors  considered  in  a comprehensive way?

       10.-   Whether  human factors and broad management issues  were
      identified  appropriately,  in  appropriate  sequences,  in ways to lead
      to   reasonable  conclusions  and  recommendations.

       11.- The  conclusions and  recommendations.

       12.- What part  of the  overall approach  could be modeled for future
      such investigations?

       13.-   Was information for  the prevention of similar accidents
      appropriately  presented in  this  report?

-------
        14.- Were the roles of federal, state  local agencies; the public;  labor;
       trade associations and  public interest groups  appropriately
       addressed?

        15.-  Recommendations for  actions they  (federal,  state  local
       agencies; the public; labor;  trade  associations and  public interest
       groups ?) could have taken or should in the future take to  reduce
       accident risks

 1 .-   Comments on whether  the scenario  by scenario approach
taken was correct.

This  reviewer  feels  that  the "scenario by  scenario approach" used in
Region, VIFs report  on the accident at the Terra Industries facility in Port
Neal, Iowa  (EPA  Terra Report)  is a valid  approach.

However, there  is no  such thing as  a  "correct" approach.

There  is  a vast literature  on  approaches to  accident investigation. A
relatively recent authoritative  survey of  this  literature and  its  learnings,
"Guidelines for  Investigating  Chemical Processing Incidents  ", has been
published by the Center for  Chemical  Process Safety  (CCPS) of the
American   Institute  of Chemical  Engineers  (CCPS  Guidelines).  These
"Guidelines" describe  numerous different  'accepted'  approaches  for  the
investigation of accidents  similar to the  one  that occurred at the Terra
facility.  Each of these  accepted approaches  is  used and  endorsed by
different reputable  authorities.

2. -   The findings  of the  report examining various  explosion
scenarios.

The  findings of the  report  examining  various  explosion  scenarios  seem
reasonable given  the  information presented  in  the  EPA Terra  report  with
the  possible  exception of some of the findings  related to the Sparger  (see
comments in item  6)

3.-  The  identification of the most plausible scenario.

-------
 The physical findings after  the  accident indicated  that there were  two
 explosions the first  of which  occurred  in  the  Neutralizer and  the
 subsequent one in, the Rundown tank.

 Three of  the four scenarios presented  dealt with  possible sequences of
 events, and conditions related  to the initial explosion. The third  of these
 four  scenarios, which dealt with initiation  within the Neutralizer was
 sub-divided into two scenarios,  one dealing with decomposition
 originating  within the nitric acid Sparger, the  other with decomposition
 originating  in  the contents of the  Neutralizer tank external to the nitric
 acid sparger.

 The fourth  scenario was the only scenario that dealt with the second
 explosion which occurred  in the Rundown tank.

 This reviewer agrees with the EPA Terra  Reports conclusion that scenario
 3  was the  most plausible  of the three scenarios for the  first  explosion and
 that scenario 4 was  plausible.

 However  based on information presented in  the Terra Port  Neal  Report,
 there are  questions which  need further analysis  before this reviewer  is
 comfortable with choosing  which of the  two  sub- scenarios considered
 under scenario 3 is  most plausible.

4. -  The omission of significant scenarios.

Based on  the information  available  this reviewer does  not believe  that
significant  scenarios  were  omitted, though  perhaps  sub-scenario 3  might
have  been  broadening given  the  information  maintained  in  the  subsequently
issued Terra  Port  Neal  Report(  see specific comments under item  6
dealing with  the considerations given to  the role of the Sparger).

-------
 5.-   The  comprehensiveness  and reasonableness of  statements
 about technical  conditions  under,  which  the accident  occurred.

 Overall  the comprehensiveness and  reasonableness of  statements  about
 technical conditions  under which the accident occurred was good given the
 conditions  and  less  than complete documentation and  records  that
 apparently  existed at the Terra Port Neal facility before and after  the
 12/13/94  accident.

 6. -  The appropriateness of the consideration given to  the roles
 of certain equipment,  notably  the  sparger.

 Based on  the information given  in the EPA Terra report in regard to
 possible  role  of  titanium and in particular the statement  on p. 86  of this
 report which states that  "The  forces that distorted and destroyed the
 Spargers were applied  to the  external surface, not the interior surface",
this reviewer initially believed that  appropriate  consideration  was  given
to  the role  of other equipment, notably the sparger.

However some doubts were raised  in regard to  this reviewer's  initial
belief on the appropriateness of the consideration given to the role of the
sparger by the information in the Terra Port Neal  Report. This report
contained  information  related  to the  possible  sensitization role of
titanium and in particular the statement (Exhibit  8, p. 8)  that  "The
titanium  fragments  provided  evidence  that the acid sparger  had ruptured
due to internal  overpressurization.  In  addition, most  of the  interior and
some of the  exterior  surface  of the fragments  showed corrosion".

7.- Is  discussion of the  conditions existing  before the accident
appropriate  with  no  prior history  of the accidental detonating
of  AN solution"?

The simple  answer to  this question  is yes,  though this reviewer-  has
difficulty with the meaning of this  question.  Clearly  there have  been
previous AN solution explosions at other locations. Given that it has been
established that AN and its solutions have this  potential  for injury  and
that the amount of AN (and other  highly hazardous substances) being

-------
 processed at the Port Neal facility  that  could give rise to major
 accidents, discussion of the conditions existing before the accident seem
 appropriate to this reviewer.

 8. -  The conclusions reached and whether they lead  to the root
 causes identified.

 The one significant reservation that this  reviewer has  on  the  conclusions
 put forward in the EPA Terra Report rests conditionally on the resolution
 of questions raised above concerning the role of the sparger.  However even
 if the Sparger is shown to  have played a more significant role in initiating
 the explosion in the neutralizer, this would not negate the reasonableness
 of the six conclusions put forward.

 The conclusions reached in  Section  9  of the EPA  Terra Report deal
 adequately  with  the "unsafe  acts or unsafe  conditions" resulting from
 employee  action  or  inaction. However this  reviewer believes that the
 conclusions reached in Section 9 of the EPA Terra Report EPA Terra Report
 deals  less than adequately with Root causes.

 This reviewer's definitions of 'Root' causes is;

       "Prime  reasons,  such  as failures of  some management systems, that
       allow faulty design, inadequate training  or  improper  changes, which
       lead  to an unsafe act or unsafe condition and result in  an incident.
       Root  causes  are also known  as  underlying  causes.  If  root  causes
       were  removed,  the  particular incident  would not have  occurred."

Given this  definition,  the  management  system failures that might have led
to  or  allowed for  the existence  of the  unsafe  acts or unsafe  conditions
that the report  concludes  led to the  accident was less than  adequate.  The
body  of the EPA  Terra  Report  does a much  better job in  relating
management system  failures  to the factors  that  presumably  led to the
accident than do the recommendations.

-------
In the reviewers opinion, use of a presentation approach such as the MORT
based Root Cause Analysis Form 7  would  more  clearly  show relationships
between Root  causes  and the  conclusions reached in. the EPA  Terra Report.

9.-   Were all external factors considered in a comprehensive
way?

This question is unclear to  the  reviewer What  are examples of the
external factors  that were  to be considered?  The  weather?  Power
failures? Possible health epidemics in the community that  affected
employees  or  management in  an unforeseeable fashion?

10.-  Whether human  factors  and broad management  issues were
identified  appropriately, in appropriate sequences, in ways to
lead to  reasonable conclusions  and recommendations.

The  body,  of the EPA Terra Report contained the information required to
ensure that human factors and  broad  management  issues  were identified
appropriately, in appropriate  sequences, in ways to lead to reasonable
conclusions. However this information was  less  than adequately.
summarized and related to the conclusions and recommendation made in
Sections 8 and  10.  (See  comments under  item  8 and 10).

11.- The  conclusions  and recommendations.

This reviewer  commented  in item 8  on the conclusions given  in Section  9
of the EPA Terra  Report.  However there  is  an additional conclusion given
in the  first paragraphs   of Section 10,  Recommendations.  While  this
reviewer agrees  that the  EPA Terra Report presents  evidence that  the lack
of safe  operating  procedures was  an important factor in  the  sequence of
events leading  to  the accident,  it is  somewhat disingenuous to  claim that
"the conditions  that  caused the  explosion  existed primarily (emphasis
added)  'because of the lack  of safe  operating procedures'without
defining   "safe  operating procedures" operationally.

-------
 To this reviewer "safe operating procedures" can only be developed and
 implemented after a facility  implements all  of the eight  specific
 numbered  (1  to 8) recommendations made in  Section  10.  All of these
 recommendations  appear  to  be in order and are generally  considered  to
 constitute  'good practice'  for  facilities  handling the  quantities  and type
 of  materials and processes present at the Terra' Port Neal  facility.

 12.-   What part of the overall approach could be  modeled for
 future such investigations?

 EPA should adopt a standard Type 3 accident  investigation protocol for all
 of  its accident  investigations. Most of  such approaches  include almost all
 of  the elements in employed in the  "scenario by  scenario" approach used
 in  the  Region  VII investigation but  in  a more structured  way.

This  protocol  should be modeled after one  of the broadly accepted
 multiple-cause,   system oriented investigation approaches in the public
domain  that is  focused  on the determination of 'root' causes.  To the  extent
consistent  with  its constraints,  this standard EPA  protocol  should  use the
terminology, methodology and  presentation  styles used  in the literature
associated with the broadly  accepted multiple-cause,  system  oriented
investigation approach  used  as model  for EPA's standard protocol.

If this  approach is  feasible,  EPA will achieve greater  clarity,
understanding and acceptance of its  investigations  in the technical
community and will benefit from the continuing  research work that the
broad technical community does on the particular Type  3 model  that EPA
uses as the  basis  for developing its  own standard protocol.

 13.-    Was  information for the  prevention  of similar accidents
appropriately  presented  in this report?

See comments under items 3, 6, 8, 10 & 11 above.

-------
 14.-  Were the roles of federal,   state local  agencies; the public;
labor; trade associations and public  interest groups
appropriately addressed?

This reviewer  could not  locate sections of the EPA Terra Report that
addressed  roles for  Federal, state local  agencies;  the public; labor;  trade
associations and public  interest  groups other than recommendation that
imply that Terra   Industries should  supply information to the public
surrounding the facility and  state  and local agencies dealing with
emergency  response  (recommendation. 8, section  10).

In this  reviewers opinion this  does  not appropriately  address the roles
that  all elements of federal, state  local agencies; the public; labor; trade
associations and public interest groups  should or desire to play.

15.-  Recommendations  for actions federal,  state  local agencies;
the public;  labor;  trade associations and public interest  groups
could  have taken  or should in  the  future  take  to  reduce accident
risks.

This reviewer  feels that it is relatively unproductive  to speculate about
actions  that this group  of stakeholders might  have taken in the past. In
any  case the  scope of possible actions was limited  by practical
difficulties in obtaining  concise   information on the risk  management
programs  that  existed within facilities and the relative  lack resources
available to these  groups   with the  possible exception of  trade
associations.

In the  future  the  new  EPA rule  dealing with  the prevention of major
chemical  accidental releases should  make  process safety information  and
accident  histories  much more readily available.  This should  allow  all  of
the above entities to do a better  job  in  screening the adequacy of facility
safety  programs with the limited resources available to them.

Public  interest groups  and Federal,  State  and local  agencies could  use this
information  to  make  the public  aware  of the type  and magnitude  of
facility  risks  that might  affect them. This should  lead  to  an informed

-------
                                                                               1 0
public that  could  chose to directly pressure  facilities  that they feel  have
a  less than adequate  risk management program and/or generate pressure
to provide resources that would allow  agencies to do a more effective job
on their  behalf.

Trade,  associations  could follow the lead  of the  CMA by  making
membership  conditional  on the  implementation  of  process safety programs
similar  those  embodied in Responsible Care  or generate model safety
programs to assist their members.

Labor  unions  and  Agencies could offer  accessible, appropriate  process
safety training  to  employees that  do not have  adequate training  available
through their  employer  and  inform them  of the  appropriate actions  they
might take     if less than adequate safety programs  exist  in their
workplaces.

-------
                                                                                      11

                                      End Notes
 1. Region 7, "Chemical Accident Investigation Report. Terra Industries. Inc. Nitrogen Fertilizer
Facility, Port Neal, Iowa", released 1/23/96.

2. "Settlement Agreement", DIA Docket No. 95  DBS- 20. IOASHA Docket 4446. Citation No.
S7510  -  115072555.

3. "The Terra Port Neal Explosion. December 13, 1994". Report of the Investigation Committee,
July 17,  1995.

4. "Guidelines for Investigating Chemical Processing Incidents", Center for  Chemical Process
Safety  of the American Institute  of Chemical Engineers, N Y City, NY, 1992.

5. The one  exception on the conclusions put  forward in  the  EPA Terra Report  rests
conditionally on the resolution of questions raised above concerning the role of the sparger.
However even if the Sparger is shown to have played a more significant role in initiating the
explosion in the neutralizer, this  would not negate the  reasonableness of the six conclusions
put forward

6. "Guidelines for Investigating Chemical Processing Incidents", Center for  Chemical Process
Safety  of the American Institute of Chemical Engineers, N Y City, NY, 1992, p 8.

7. "MORT Based Root Cause Analysis", EG&G Idaho, Inc., P.O. Box 1625, Idaho Falls. Idaho
83415,  June 1989, pp.1 to 3.

-------
                                     Appendix I



A.     Charge to Reviewers




B.     Note on  Required Deliverables

-------
                          CHARGE TO REVIEWERS
                 For  the EPA Chemical Accident Report,
         TERRA INDUSTRIES, INC. NITROGEN FERTILIZER FACILITY

EPA Region VII  (Kansas City,   KS)  prepared the above  report,
released January 23,  1996,  concerning an accident on December  13,
1994 occurring at Terra Industries,  Inc.  The report  is 108
pages long, and includes  an  executive summary,  an overview of  the
investigation,  discussion  of  plant operations and  events  at  the
facility,   conditions  and facts derived by the investigators,
scenarios  for  the  explosion,  conclusions and recommendations.
The principal  investigators  were  Mark Thomas of EPA; Alan
Cummings of Dynamac  Corporation,  an EPA contractor; and Mariano
Gomez,   an  EPA contractor with its Technical Assistance  Team.

As a reviewer  of this  document,  you  should use  your  technical
knowledge  and  professional judgment to comment  on  the  technical
soundness,   overall approach,  and  completeness of the report,  to
derive  recommendations  for approaches to accident  investigations
in the  future  and  accident prevention.

The report  seeks  to  ascertain the root cause of this accident, in
order to  further  the  goal of preventing accidents.   Your review
should  include the following aspects of this concern.

Comment on  the approach taken (scenario by scenario) as  a correct
approach to take.   Comment on the  findings of the  report
examining  various  explosion scenarios, and on the  identification
of the  most plausible  scenario.    Were  any significant scenarios
missed?   Comment  on  the comprehensiveness and reasonableness  of
the  statements about  technical conditions under which  the
accident occurred.   Are specific  roles of certain  equipment,
notably the sparger,  appropriately considered?

With no prior history  of  accidental detonating of AN  solution, is
the  discussion of  the conditions  existing before  the  accident
appropriate?   Please  comment on the  conclusions reached and
whether they  lead  to  the root causes  identified.

Were all  external  factors considered  in a  comprehensive  way?
Also, please  comment  on whether human factors and  broad
management  issues  were identified  appropriately,  in  appropriate
sequences,  in ways to lead to reasonable conclusions and
recommendations.

Please  comment on the conclusions  and recommendations.   What  of
the  overall approach  could be modelled for future  such
investigations?   Was  information  for  prevention of similar
accidents  appropriately presented in  this  report?   Concerning the
role of federal,  state local agencies;   the  public;   labor;
trade  associations;   and  public interest groups:   please comment
on whether roles  of  these entities were  appropriately  addressed.
Are  there  any recommendations for actions they  could have  taken,.
or should  in  future  take  to  reduce  accident  risks?

-------
                         Statement of  Work
                        For   Expert   Reviewers

 Title:   Expert Review of EPA Chemical Accident  Report,  "Terra
 Industries,  Inc.  Nitrogen Fertilizer  Facility"

 Purpose:   The purpose of this statement of work is  to provide
 external  expert  review of the above EPA chemical  accident
 investigation  report,  including its recommendations  and
 implications for prevention activities by all parties and  future
 investigations.   The report discusses the cause of  an  explosion
 at  the  subject facility and presents recommendations to prevent
 future  similar occurrences.

 Background:   On  December  13,  1994, an explosion occurred in  the
 ammonium nitrate plant at the Terra International,  Inc. complex
 in  Port Neal,  Iowa.   Four persons were killed and 18 were  injured
 as  a  result  of the  accident.   The explpsion released a  large
 quantity of  anhydrous  ammonia to the air,  nitric acid to the
 ground,   and  resulted in contamination of the groundwater under
 the facility.   The  EPA report investigating the incident was
 developed as part of the  Agency's ongoing responsibilities under
 the Comprehensive Environmental Response,  Compensation, and
 Liability Act  (CERCLA)  and the  Clean Air Act,  Section 112(r), and
 as  a  component of EPA's  chemical safety  programs.    The  report was
released  to  the public on  Tuesday,  January 23,   1996 to  become a
part of  the  examination of the  causes  of chemical accidents and
 efforts  to prevent  them.   The report was prepared by staff at
EPA's Region VII  office.

As part  of its investigative  program,  EPA desires the  review of a
panel of  experts  who from  their individual perspectives and
disciplines  can provide comment on the scope,  approach  and
conclusions  of the  report and its implications    This  task is
designed  to  purchase the  services  of  reviewers  for this  comment.

Statement  of Work:   The  reviewers  will have the following tasks
in accomplishing  the objectives of this  statement of work  (see
companion  statement   of work  for  chair):

1.   Prior  to any  meetings,  all  reviewers  will  be required to
review and analyze a copy  of  the subject  report.

     	Each reviewer  will  prepare written pre-meeting  technical
     comments  on  the report,  based on the EPA  charge and adhering
     to the  organization and  directions  of the  Chair, a reviewer
     so designated by EPA.

     	Reviewers'  written  comments  are  due  to  EPA,  with copies
     to the  Chair,  two  weeks  prior to  a  meeting, to  be  held at
     EPA's regional   office in Kansas City, KS.   The  Chair will
     distribute copies of  all reviewers'  comments  to all
     reviewers prior to a meeting, and organize  the  meeting.
     The  Chair will  chair  this  meeting.

-------
        All  reviewers  will  prepare for the meeting by studying
     comments prepared by other  reviewers.  All reviewers may
     review  or  consult  any  other background documents provided by
     EPA or  at  the direction  of  the  Chair.

2.    Attend  the two-day  review meeting in Kansas  City convened by
the Chair to exchange  comments and discuss issues raised by the
subject document.   Based on the  comments and discussions,  develop
recommendations to EPA.   This meeting is not to be C9nvened for
the purpose  of  achieving consensus but to exchange views and
comment,  expected  from  the  different perspectives of the
reviewers.

Deliverables:
1.  Written  comment on  the  subject report.
2.  Active participation in exchange of views/ comment  at  two-day
meeting of reviewers in  Kansas City,  KS.
Cost Reimbursement  Procedures:

The EPA will pay an agreed  upon  fixed fee  to the  reviewers.   This
fee is intended to cover the reviewers'  consulting fee. In
addition,  this  negotiated  fee includes expenses for economy or
excursion airfare,  local transportation, miscellaneous  and
incidental expenses  (i.e. meals),  and  lodging  for two nights  at  a
ceiling specified  in  the attached instructions, "Travel
Constraints,"  that may be  required for the reviewers.

To estimate  the  consulting  fee,  EPA  is estimating that  no  more
than a total of  four  days will be required.   The total  amount of
the agreement will be  negotiated by an official from EPA's Office
of Acquisition  Management.   Once  all the deliverables that are
listed above have  been  provided,  the reviewer  may submit an
invoice to EPA  for reimbursement of a properly justified sum  up
to the negotiated  amount.

Period of Performance:
Upon award of the  contract  through September 15,  1996.


Notice Regarding Guidance Provided under this Statement of Work:

The contractor  shall  not engage  in activities  of an  inherently
governmental nature such as the  following:

          Formulation  of Agency  policy
     a.   Selection of Agency priorities
     c.   Development  of Agency  regulations.

Should the  contractor  receive any instruction  from an EPA  staff
person that  the contractor ascertains  to  fall  into  any  of  these
categories or  goes beyond  the scope of the  contract or  work
assignment,   the contractor shall  immediately contact the  Project
Manager or  the  Contracting  Officer.

-------
Prior to  initiation  of  review,  the  reviewer  shall disclose any
conflict or potential conflict  of interest,  and shall  sign a
conflict  of  interest/confidentiality form.   Forms will be
provided by the  EPA  project  officer. Any COI that  surfaces
subsequently during  the review  process  is  reported to the project
officer.

-------
                                                Pg- 59
                                                No procedures for AN TK'
                                                while down.Employees poorly
                                                trained on DCS.Needed pans
                                                not available.
                                                         HUMAN
                                                PH Probe defective
                                                PH Probe in line not AN Tk
                                                DCS not operational
                                                No Check or Block valve Nitric Acid
                                                sparger
                                                Back Flow from barometric loop
                                                                                 1. Drain Tk H20 Wash
                                                                                 2. PH to 6 circulate
                                                                                 3. Circulate AN from run down
                                                                                   to Neutralizes
                                                                                 4 DCS operating
                                                                                 5. No steam internal
                                                                                 6. No air purge for long period
                                                                                   of time.
Using plant air to blow Acid
Line into AN Tk not in SOP.
No procedures for steam
sparger.
        The major cause of the explosion was no procedures for controlling, monitoring or blow
down of H20 wash of the AN Neutralizer Tk during the time the Tk was down with material in
the tank.

-------
                 EPA CHEMICAL ACCIDENT INVESTIGATION
                                 REPORT
                          TERRA INDUSTRIES, INC
                 COMMENTS AND QUESTIONS COMPILED BY
                               JOEL VARIAN
                         SUBMITTED JULY 15, 1996
PERSPECTIVE: Labor
              Hourly Employee
              Accident Investigations (Mining)
              Accident Investigations (Chemical)

-------
                PLANT BACKGROUND INFORMATION
                              —PROFILE—
Accident Incident Rate

Lost Time Accidents

Uni on/Non-Uni on

Downsizing? If so, how much & what department?
How much overtime per department?
How many hours?
Is overtime mandatory?

Recordkeeping  compare lost time  records with state compensation records.

How much contracting of work, normally performed by Terra employees?

Number of excursions reported in last 10 years?
How many inspections by State OSHA EPA, etc. in last 10 years?
How many citations?
What were they?

Routine Maintenance (RM)  program in place at the time of the December  13, 1994
explosion in the Ammonium Nitrate Plant?

Request a copy of  the Chemical  Safety Audit  (CSA) recommendations, performed  February
 1994.
What good faith changes were made by Terra  after the Audit?

Ammonium Nitrate sensitization information on page 18  of report - PH, concentration,
tempature,  contamination, low density areas are conditions that can sensitize and increase
hazards.
How much of this information did Terra have?
How much was available to them but did not have available to the operators and/or
maintenance department?

The Acid Plant was shut down Dec. 12 at 0430 hours on emergency basis. What was the
nature of the emergency?

Request copy, if available, of the inspection report on the Neutralizer and Nitric Acid sparger
(Sept.  1994).

-------
                                    OPERATIONAL

                                 —MAINTENANCE-


      Procedures, Parts, Inventory, Upkeep Mechanical

      On November 27,  1994, Operations and Maintenance personnel determined the PH probe
      located in the neutralizer rundown line was defective.  There were no spare probes available:
      the defective probe  was in  service until the time of the explosion.

      Procedures
      AN Neutralizer shut down changed without monitering what was going on in the vessel.
      No circulation of a AN vessel during shut down.
      The DCS was down - no parts available.
      No written SOP on AN Neutralizer while shut down.
      No written SOP on checking AN Neutralizer thickness.
      Corrosion coil during turnaround.
      Steam coils in the Neutralizer jacket had corroded and could not be used for several years.

      Parts
      What procedure on spare parts was in effect at the time of the explosion?
*should  involve metalegists inspection.

-------
                                   HUMAN


Blowing Nitric Acid line into neutralizer with plant air was not in SOP provided by Terra.

No specific procedure for connecting steam to sparger.

No check  or block valve attached to the Nitric Acid sparger where they exit the vessel as per
required by Mississippi Chemical.  A Barometric loop  that allowed back flow into the
spargers and into the Nitric  Acid  line back into the barometric loop.

Terra employees stated that the addition of steam was a normal procedure to prevent back
flow into  Neutralizer Nitric Acid sparger and to keep the sparger from salting out.  Thus,
adding heat to Neutralizer.

-------
                                 PLANT AIR
The plant air was applied to the Nitric Acid line to purge the line to prevent freezing. The
Nitric Acid was discharged from the line to the AN Neutralizer.  The plant air purged the line
from 1500 hrs. until 2030 hrs. 5 Vz hrs. (If my military time is correct) Air and steam was the
last two materials induced into the AN Neutralizer before the explosion.

On page 37: Terra had experienced problems with Hydrocarbons in the plant air.
The  Question is: Did the EPA team request piping diagram of the plant air system and if they
received a diagram,  were all the  precautions  necessary to prevent  Hydrocarbons  from entering
the plant air systems?

Blowing the Nitric Acid line from the acid unit was not in the  SOP. Why? What was the
problem using plant air?

-------
Initial Comments on the USEPA Chemical Accident Investigation Report:
Nitrogen  Fertilizer Facility
Terra industries Inc.
Port Neal, Iowa

PAUL L. HILL, Ph.D.
Review Panel  Chairman
In undertaking this review I have referenced the  "Charge to Reviewers" provided by the contracting
agency, USEPA. Based upon the USEPA Region VTI report released January 23, 1996 concerning
the accident on December 13, 1994, the agency has requested comment on various aspects of their
performance as well as content of the report itself.  These include the approach taken; plausibility of
scenarios;  technical, human  and management considerations related to conclusions; roles of associated
agencies and interest groups  and  overall recommendations toward a model approach for future
investigations.

As the designated Chairman of the Review Panel, I am charged with conducting a preliminary review
to determine appropriate expertise and diversity  of interests  for potential reviewers;  provide USEPA
with an ample list  of potential  reviewers; organize the review by collecting preliminary comments and
distributing them among  the panelists selected by USEPA; calling a meeting of the panelists to discuss
their reviews of the report; chair said  meeting  and prepare a written  summary of reviewers comments
and recommendations. As a prelude to actual discussions with others my own personal comments
are provided.

Although not charged with reviewing any materials outside the document (written report) itself,
associated materials including Terra Industries' report dated July 17, 1995, Iowa Division of Labor,
Occupational  Safety  and Health  Bureau Inspection No.  115072555 dated May 25,  1995 and
USEPA's supplemental copies of "Drawings  and Graphs" were in my possession for reference and
personal review. Because these documents report on the same circumstances (the accident) their
availability is considered appropriate and valuable. Upon this  submission no other documents,
evidence or expartate communications  have  been considered.

-------
 THE CHARGE OF REVIEW

 EPA has asked each reviewer, within the limited scope and resources provided, to answer ten
 associated questions which roughly correspond to those outlined in paragraph 2 above. In issuing
 the charge, EPA has stated that "the reponse seeks to ascertain the root cause of this accident, in  order
 to further the goal of preventing accidents.   Your review should include the following aspects of this
 concern."

        1. Comment on the approach taken scenario by  scenario.  ...

        This was valuable  to the reviewer in ensuring that  all plausible avenues of theoretical causality
 had been pursued.  It was also presented from a physical evidence point of view which helped the
 reviewer understand the support for, or lack thereof, of evidence  for each  potential cause  for the actual
 explosion. This approach should be replicated in future investigations.  If coordinated protocols for
joint accident investigations are achieved, the scenario by scenario approach may be the basis for
 closer agreement between vested parties.

        2. Comment  on the findings of the report and on the identification of the most plausible
 scenario.

        Findings seem valid. The most plausible scenario is simply the causal theory which generated
 the most supportive evidence. In fact much of the chemical, physical and spatial evidence collected
 by other  assessments supports that of the EPA's  most plausible scenarios. While  there are no
 unequivocal answers  to lingering questions about contamination sources,  titaniumions, chlorides,  etc.
 as discussed by the less plausible scenarios, the conclusions reached by the report are  reasonable.

        3. Comment on the comprehensiveness and  reasonableness  of rhe statements about technical
 conditions under which the accident occurred.

        This position of the report is supported by known  AN  chemistry principles,  historical
 production and laboratory tests. All parties who expressed views on the technical conditions in the
 neutralizer, rundown tank, etc. are closely aligned with few exceptions. The report does a good job
 at putting this information forward in the discussion but takes the added approach of "scenario
 building" through which the reviewer can actually pursue technical conditions in light of physical
 evidence.

        4. Are  specific roles of certain equipment, notably the sparger, appropriately  considered?

        A degree of uncertainty exists here.  The report dismissed most of the theories about the
 sparger and titaniumions rather readily as compared to the contentions  of likely cause presented by
 other parties. Further pursuit by the review team is needed to properly address the  basis for this
 disparity Due to the limited scope  of this  review, it is  also recommended that EPA continue research
 and discussions with technically competent and affected parties to resolve discrepancies in either

-------
theory  or evidence.
       5. Is the discussion of AN solution appropriate?

       Generally, yes. Introductory,  theoretical,, chemical, technical and  operational  background on
AN  solution properties,  management and handling were appropriate for the discussion: Suggest this
kind of  literature search for state-of-the-science become  a standard fixture for well prepared,
comprehensive reports. (Up  through section 6). Continued monitoring of AN production  and
research literature  is properly noted but perhaps should be more emphasized or gathered and
circulated by the fertilizer industry,  associated agencies, other?

        6. Were all external factors considered  in a comprehensive way? And, were human and
management factors identified in such  a way  as  to lead to  reasonable  conclusions  and
recommendations?

       The report does  not provide a surplus of details about the investigation team's  interaction  with
management or the depositions of hourly employees. Obvious however, is the lack of, or inability
to produce, procedures, instructions and training logs or PID's for the proper understanding of the
operation and therefore the ability to operate it safely. When coupled with the physical and chemical
evidence, human factors, as presented, lead one to the conclusions and recommendations as presented
in sections 9 & 10 of the report.
        7. Comment on the  overall conclusions  and recommendations. What of the overall  approach
could be  modeled for future investigations?

        Readability and sequence of information lay-out are generally good. However, information
on page 33, 34 could be developed into  standard time-line  information to gain graphic illustration of
events.  For plausibility and  causality, the scenario by  scenario approach  is very good. This provides
the reviewer with the investigation team's  "thinking" and  second guesses as they attempted to
reconstruct conditions of the accident. Successfully eliminating scenarios based on the evidence
ensures the reviewer that other theories of causality  were investigated. However, not all potential
scenarios were presented. See Additional Questions  & Comments.

        8. Was information for the prevention of similar accidents appropriately presented in this
report?

        Although presented clearly within the report  itself, it is unclear how EPA or other agencies
may convey, in an ongoing fashion, results and recommendations of  this nature.  Industry-wide
advisories, special notices to the  fertilizer industry, etc. should be considered. Special  bulletins and
advisories similar to those  of NTSB should be issued in the interests of all facilities, workers and
potentially affected parties. Currently, within the report, it is unclear that such parties will receive

-------
this information
        9. Comment on whether the roles of other entities were appropriately addressed

       It is difficult to determine how the needs, if any, of other entities have been addressed by this
report since there was  no coordinated effort which  involved all the parties  in the actual investigation.
Iowa has a designated "State OSHA" authorization from the federal agency therefore making this
particular  investigation  even more disjunct than might have otherwise been  possible.  If coordinated
investigations and joint reports are to be achieved as envisioned by the CAAA of 1990, then a broad
protocol, clearly defining roles of all interested entities must be developed and codified.
        10. Are there any recommendations for actions that could have been taken or should in the
future take to reduce accident risks?

       The agencies, both EPA and OSHA as well as state program officials should follow-up with
AN producers regarding the extensive recommendations provided by this report.  The  agencies  should
also continue to monitor from an R&D prospective, the use of titanium materials as presented by the
Terra review. Both reports strongly agree that proper procedures, training and monitoring must be
continuosly in place for proper prevention to occur

-------
                 ADDITIONAL QUESTIONS & COMMENTS
Did EPA compile any emergency response documentation of this case and integrate it with
the findings for consumption by all parties,  particularly lay public?
While this in an accident investigation report it provides an opportunity for other EPA
sponsored  initiatives.
Is this, or will it  be, addressed elsewhere?

Is there ongoing  review or reconsideration  of the metallurgists'  findings?
Much of the evidence for  direction, impact of physical damage, etc. rests on this report
Is EPA reviewing?

Was sampling adequate given the suggestion by Terra that small isolated areas may have
initiated  deflagration?
Chlorides and "low  density areas" were noted as potentially plausible factors  on page 36 and
38 of the report but were later dismissed  as  the initiating  factor.
 Could this be due to lack of samples from  various  sites?
How much collective data was reviewed by the investigative team to reach this decision?

What about additional  evidence (e.g.pump)  which  was never recovered?
Might there be other significant evidence which supports other theories? (e.g. the initiation
occurred in the sparger).
Was the entire sparger (all pieces) recovered so as to completely eliminate the concept of
"isolated site initiation"?

Sabotage.  This was not dealt with. Why?
If, via the depositions of parties, this (theory) was successfully eliminated then it should be
stated in the report.

-------
EPILOGUE

       First, accident investigations and the attempt to reconstruct the conditions which  led to
accidents are  inherently difficult to pursue. The report indicates in its introduction that certain
evidence,  diagrams and requested documents  were  either  destroyed or unavailable  for  this
investigation report. Even with  satisfactory provisions of existing management, operations  and
training  materials,  reliance upon human knowledge and  recollections as  well as  potential
nondisclosure makes the job of reconstruction all the more difficult.
       Although not only pertinent to this review, the investigating agency needs to clearly  define
its investigative approach to  future investigations from a purely  professional/scientific view.  If called
into question (most contentious reports are likely to be litigated) the agency (s) needs to simply
describe their command of the existing literature on investigative techniques and which of these  they
either chose to use, modify for use or discard as inappropriate, given the particular investigation.
Many such technical approaches also have the added advantage of providing the reader or reviewer
with a  graphic representation of the approach  (e.g.  fault  tree), as well as  the  technique or
combinations  of techniques used to approach the investigation process. CCPS's "Guidelines for
Investigating Chemical Process Incidents" copyright 1992 provides at least one good  example of
various inductive, deductive, morphological and non-systems techniques for accident investigation.
       While I have responded to  the agency's needs to the best of my abilities, at this time,  I reserve
the right to respond further after  seeking the advice and considerable expertise of my colleagues.
Therefore these comments are subject to modification after the discussions in Kansas City.

-------
APPENDIX
    III

-------
                          Addendum to Renew of EPA
                          Chemical Investigation Report
                               Tern Industries, Inc.
                                          By
                                     AuncfeNkra

This •AUnAim a to fulfill my charge u reviewer.

1.      Comment on the approach taken (scenario by scenario) as a comet approach to take.
       The scenario* regarding plausible cauae(s) of the erosions appear to be weH thought oat
       However. I can not comment as to whether any significant scenarios were missed.

       NOTE:  On page 1 of my original report, pkase disregard the first boDet that began as
       follows:  Tetroiemn tpffl in the ammonia plant-."  - the EPA investigation had eliminated
       it as being a potable contributing factor.

2      With no prior history of accidental detonation of AN solution, is die ruVutrion of the conditions
       f U'ffr^i otfint the acfiofnt afijjiUjJUttlfc
       This reviewer believes that the discussion of the conditions was necessary to give the review-
       er* a description of the process operations  leading op to the explosions.

3.      Please comment an the conclusions reached and whether they lead to the root causes identified.
       This reviewer believes that one or aH of the conditions that existed in the neutrafaer just
       prior to the explosions could have caused the event of December 13, 1994.
       •     The strongly acid condition* in the nentnfizer
       •     The application of 200 prig steam to the nitric acid spargers and the superheat that
             possibly reached the nentnlizer
       *     The application of compressed air which created low density zones in the AN solution
       •     Lack of ftowm the neutrahzer and from the neutndiz^
             isolated critical sensors
       •     Chtorides contamination in thc nitric acid and «n»^n<\nimn nitrate
       •     Lack of monitoring of the AN plant when in shut down

In this reviewers opinion, the primary cause of die  > 2 explosions was the lack of a process safety
mmiagfim rm system which should have inrtnrlrd safe operating procedures, a hazard communication
program, and process hazard analyses of the different units/plants within the facility.

If a safety management system had been in place, the employees would have known how to safely
shut-down and start-up the process. The employees abo would have known the risks and potential
consequences involved in performing impiuper procedures. The FHA would have identified critical
areas in the process that needed modifications to  minimize or manage the type  of incident that
occurred at Terra.
4.      Were all external factors countered in cmnpnhcnm*:
       This reviewer has no way of measuring whether afi external factors were considered.  This
       question is outside my field of expertise.

-------
5.      What of the ovtr ell approach could have beat modelled Jbr future such mvstiffitions?
       The over an approach which oouki be modelled for future investigations should include the
       following:
       '      Diagrams and photos (especially aerial photos) of the faculty before the event, and
                         *fter the rnrH^ir* to show the direction and magnitude of the event
       '      Development of scenarios using historical data and the evidence found following the
              dplonoos
       •      The use of opera, such as metallurgists, and other professionals knowledgeable of
              tiie partifnbrr process.

This reviewer rfer>m
-------
                     Addendum to Report of Geraldine V. Cox on Terra Report.
                                       Dated August 5, 1996

After further discussions with chemists familiar with Ammonium Nitrate and the meeting of the reviewers
in Kansas City last week, I would like this addendum to be attached to my initial report.

 1.   Sensitizer(s). Clearly something or multiple chemical compounds acted to sensitize the reaction.
     EPA believes that the chloride ion may be the sensitizer. After review of the limited chemical data
     presented in the Terra report two other, equally plausible materials could have acted as sensitizers,
     either alone or in combination with each other or with chloride. These additional sensitizers are
     chromium and iron (ferrous and ferric complex). Both could have been formed from the metal in the
     neutralizer tank, especially after the prolonged contact with very low pH, e.g., 0.8. The metallurgist's
     report indicated pitting on the walls of the stainless steel tank. This pitting is an additional indication
     that solution of the stainless steel reaction vessel probably  occurred. The values of the chromium and
     iron in  the AN storage tank indicated the  presence of all three elements.  I  doubt that the sensitizer
     will ever be defined at this point, but three candidates were present  in the finished  product,  so  any of
     these could have contributed.   The limited data do not seem to have sufficient concentrations  of oil
     to make that a likely contributor to the sensitizing  process.
2.   After reviewing the original photographs of the site before and after the event, the  probable chain of
     events is even more compelling. The  EPA report seems to provide the most plausible sequence of
     events,  i.e., that the reaction began in the middle  of the neutralizer and then the run down tank had a
     secondary  explosion.
3.   The actual incident may never be fully defined.  However, the conditions were present - namely
     heat, sensitizers, low pH, and bubbles that all contributed to  the event.  The absence of written
     procedures for shut down and the warnings for unsafe operating conditions, i.e., documented
     procedures and training for the workers clearly allowed the development of these  critical conditions
     for the  event.
4.    More chemical analysis of the residues in  the tank  and at the  site would have been  helpful to
     understand the chemistry, but, since the State Fire  Marshall hosed the site with water, I am not sure
     how useful samples would have been after that contamination from the fire  hose.
5.    While we were not asked to review the report prepared by Terra, the chemical analysis were of some
     use. The table was not identified as well as it might have been, for example, the sources of the two
     water analyses. In particular,  the sodium  value for the water aid not match for the  two water samples
     in the Terra Table.  (Also, there was no indication of replicates,  deviations of the values for replicates,
     and the instrumentation and methods used for the analyses.) This high sodium value might be an
     indication of high chloride concentration,  although a chloride analysis was  not provided for the water.
     The variation for chromium and iron are interesting for the various sources and do indicate a higher
     level than I would have expected for such a process.
6.    Clearly the  first responsibility  when responding to  a chemical accident, such as this, will be to
     stabilize the site to mitigate harm to the plant, community and environment. Yet, at the same time,
     one must begin to approach such events from a forensic perspective.  If such an event were  sabotage.
     which I do not believe this incident was, one would want to protect the site  until a  criminal
     investigation could be conducted. Therefore, I think that EPA needs to consider how to secure the
     site for such an investigation during and immediately following the incident,  This is easy to say, but
     more difficult to do - especially with various jurisdictions all having some authority at the site from
     the plant personnel to the multiplicity of local  state and federal agencies with  a legitimate role at the
     site. Perhaps discussions with the National Transportation and Safety  Board might be useful to
     determine how they would handle such a situation. Do they have prearranged agreements and
     procedures that might be used?
7.    In future incidents, a review of the reported accidents and injuries and a comparison with worker
     compensation claims would be a good indication of the ongoing safety record at any plant under
     review. This might be made a normal part of the incident report.
8.    EPA needs  a better definition  of its role relative to state OSHAs. This might require legislative
     change since state primacy is a legislative initiative. In serious chemical accidents where the
                                                                                        W)6  0  2  1996

-------
    surrounding community could be endangered, the normal state or federal OSHA review might not be
    adequate.
9.   Several typographical errors appear in the original report. The most significant is on page 2, in the
    last paragraph on the page, 4th line. The word "absent" should be replaced by the word "present."
    Other corrections are: Same paragraph on page 2: second line Since the pH sensors was; page 5,
    1st full paragraph, third line first word, system?; same page, item 7, 7th line of response, should read:
    "period, although, . .  "; page 6, 1st paragraph of response, 4th line ".  . . report of the depth of the one
    prepared bv EPA."
Submitted 7th  August 1996,
Geraldine V. Cox, Ph.D.

-------
EPA Response to Expert Review Recommendations

-------
EPA Response to Expert Review of
EPA Investigative Report on Terra
RECOMMENDATION: EPA should include time lines in future reports.
REPLY:         EPA agrees that this element would enhance future reports and is explicitly
including it in the EPA/OSHA joint protocol on investigations, now being prepared.

RECOMMENDATION: EPA should expand and continue to model the scenario-by-scenario
approach.
REPLY:         EPA has included this as well in the  proposed protocol. We would expect
that the approach will become more and more refined  as the program matures. This is a
valuable tool for those investigations when this  approach is  appropriate. Reports should
include a full discussion of scenarios discarded as well  as considered.

RECOMMENDATION: EPA should adopt  or specify  rigorous  technical procedures
sanctioned by the engineering and research communities.
REPLY:         EPA is aware of the available  established methodologies for accident
investigation. Its investigators will  receive training on  several of these procedures. The  EPA
investigation team will choose one  or more of these procedures  for a given investigation  and
will describe the investigative methodology in its report.

RECOMMENDATION: EPA should develop a  refined protocol for accident investigations.
REPLY:         EPA agrees that such a protocol is  necessary and  has been developing a joint
protocol with OSHA during the past year.  This protocol defines the purpose and goal of
investigations; spells out cooperation among EPA, OSHA, local investigators, and local
stakeholders; includes discussion of technical approaches and procedures for conducting
various elements of investigations;  addresses protection of confidential business information;
and addresses production of the accident report and alerts which may stem from information
gathered during the investigation.   This will be a public document  which will inform all of
the investigative program.

RECOMMENDATION: EPA should hold public meetings to seek  stakeholders' input on the
protocol.
REPLY:         EPA plans  to share the proposed protocol with stakeholders who will be
potentially affected by  investigations conducted according to its directives. To obtain these
comments, we will make the draft  protocol available through electronic and other means and
will consider actions such as holding a public meeting  as recommended.  We expect that the
protocol will be revised periodically as needed.

RECOMMENDATIONS: EPA should consider  accident oversight  committees at  affected
sites which include public liaisons.
REPLY:          EPA agrees that  vehicles  should be available  for public input into and
communication with the review team. Existing elements such as Local Emergency Planning

-------
 Committees could serve this function.

 RECOMMENDATION:   EPA should initiate agreements with other federal, state, and local
 entities with accident response authorities or consider legislative recommendations to
 accomplish same.
 REPLY:         We agree such agreements are essential. EPA and OSHA have been
 developing a Memorandum  of Understanding to set forth terms of cooperation and
 coordination between the agencies, to ensure the most effective investigations and to avoid
 duplication of effort.  EPA has initiated efforts to establish agreements with State OSHAs.
 We are investigating means to coordinate with other entities such as State Emergency
 Response Commissions (SERCs) and State Fire Marshals.

 RECOMMENDATION:    EPA should create increased public and private awareness of its
 investigative program.
 REPLY:          EPA has already  presented its accident investigation program at numerous
 national public conferences  and state workshops during the last year. We are developing an
 outreach program to share results of investigations to all stakeholders and to alert them to
 particular hazards identified in the course of investigations.  This activity will assist in making
 the program known and can be the occasion for working with particular industries or trade
 associations about specific hazards defined after an accident. We are preparing such an alert
 for ammonium nitrate facilities, which will be of use not only to the affected industry but also
 to communities having such facilities nearby. We can build on our existing work with
professional societies  like the American Institute of Chemical Engineers  (AIChE) Center  for
 Chemical Process Safety (CCPS) and others,  as well as trade associations, with whom  we
 have worked on aspects of the chemical accident provisions of the Clean Air Act
 Amendments.  We have begun developing fact sheets  and  will continue to explore other
vehicles and opportunities for outreach.

RECOMMENDATION:     EPA should clearly articulate the national goals and criteria for
accident investigation.
REPLY:           We agree that this is important to the  integrity and acceptance of our
 investigations.    We will include this element explicitly in the outreach activities noted above,
as well as in our agreements with federal, state and local entities with whom we will work in
particular investigations. As noted above, our protocol for investigations can serve this
purpose as well, as  it will include discussion of goals and objectives of investigations;  an
explanation of EPA and OSHA authorities; and procedural steps for the conduct of
investigations. The  document will be available to the public and to all stakeholders.

RECOMMENDATION:      EPA should draw upon the existing experience of NTSB and
other agencies for assistance in the evolving  program.
REPLY:           We are drawing upon the experiences of the National Transportation
 Safety  Board (NTSB), the Bureau of Alcohol, Tobacco and Fire Arms (BATF), OSHA, and
the National Institute  of Standards and Technology (NIST) in  accident investigation in
developing draft protocols and training for our investigators.  We will continue  to work

-------
closely with these agencies and others.   We also can access expertise from the National
Response Team, a coordinating body of 15 federal agencies having responsibilities for various
aspects of dealing with hazardous materials. It should be noted that we are working with
these agencies and others to assemble ways to  obtain expert assistance for very specific
expertise which may be needed in the course of a particular accident investigation.
Additional notes and recommendations:

The expert reviewers of EPA's report, in addition to articulating the above recommendations
for EPA in their summary, commented on a number of specific issues  concerning the Terra
Industries investigation and included additional recommendations in the text of their report:

ROLE OF THE SPARGER:
One issue is the particular role of the sparger in the accident at Terra Industries. To date,
EPA has not received any additional evidence or scientific data that would lead to altering any
findings, conclusions, or recommendations in the final report.

IDENTIFICATION OF  ROOT CAUSES: Another issue raised by the reviewers was that
EPA should provide more attention to the relationships between root causes and conclusions
in the report.  That is, the conclusions themselves were less  explanatory of root cause than
was the body of the discussion of the report.  EPA acknowledges that some root causes
should have been better explained in the  conclusion section of the report.  In future accident
investigation reports, EPA will provide better identification and summary of root causes of the
accident as well as correlating the root causes with the recommendations. In addition, EPA
provided in the Terra report general recommendations to the ammonia fertilizer industry as a
whole to prevent conditions  such as those existing  at Terra from recurring.

NATIONAL NETWORK OF INVESTIGATIVE CAPABILITIES:
The reviewers noted that EPA should consider a national network of investigative capabilities
and expertise as its accident investigation program  matures, and should include in the network
integrating the use of personnel or contract services with particular expertise and should also
include the expertise of OSHA, states,  etc. The joint EPA/OSHA investigation  protocol now
being prepared notes that EPA has a contract in place to provide technical assistance for
accident investigations.   This contract also allows EPA to access, as needed, experts in various
technical fields to  assist  in information gathering and analysis.  In addition, EPA is
developing a list of available EPA regional staff and contactors with their expertise whom
EPA could rely upon for assistance in investigations.  As noted above in a reply to a related
recommendation, EPA can also utilize capabilities of other federal agencies, for example the
testing and analysis laboratories of NTSB, NIST, and EPA.  EPA and  OSHA will also solicit
the expertise of other Federal agencies, for example, BATF,  NTSB, DOD, and DOE.

-------
APPENDIX A

-------
                          CHARGE TO REVIEWERS
                 For  the  EPA Chemical Accident Report,
         TERRA  INDUSTRIES,  INC. NITROGEN FERTILIZER FACILITY

EPA Region VII  (Kansas City,  KS)  prepared the above report,
released January 23, 1996,  concerning an accident on December  13,
1994 occurring   at  Terra Industries,  Inc.   The report  is  108
pages long,   and includes an executive summary, an overview  of  the
investigation,  discussion  of  plant operations and events  at the
facility,  conditions and facts  derived by the investigators,
scenarios for the explosion,  conclusions and  recommendations.
The principal investigators were  Mark Thomas of EPA;   Alan
Cummings of  Dynamac  Corporation,  an EPA contractor;   anc^  Mariano
Gomez,  an EPA contractor with its Technical Assistance Team.

As a reviewer of this  document, you should use your technical
knowledge and professional judgment to comment on  the  technical
soundness,  overall  approach,  and  completeness of  the  report,  to
derive recommendations  for approaches to accident  investigations
in the future and accident prevention.

The  report seeks to ascertain the root cause  of  this  accident, in
order to further the goal of preventing accidents.  Your  review
should include  the  following  aspects of this  concern.

Comment on the  approach taken  (scenario by  scenario)  as  a correct
approach to  take.   Comment on  the  findings  of the  report
examining various  explosion scenarios, and  on the identification
of the most  plausible  scenario.   Were any  significant scenarios
missed?  Comment on the comprehensiveness and reasonableness  of
the  statements  about technical conditions under  which  the
accident occurred.   Are specific  roles of certain  equipment,
notably the  sparger,  appropriately  considered?

With no prior history  of accidental detonating of AN  solution, is
the  discussion  of the  conditions  existing before  the  accident
appropriate?   Please comment on  the  conclusions  reached  and
whether they lead to the root  causes  identified.

Were all external factors considered  in a  comprehensive  way?
Also, please  comment on whether  human  factors and broad
management issues  were  identified  appropriately,  in appropriate
sequences,   in ways to lead to  reasonable  conclusions  and
recommendations.

Please comment  on the conclusions  and recommendations.   What  of
the  overall  approach could be  modelled  for  future such
investigations?  Was information  for  prevention  of similar
accidents  appropriately presented  in  this  report?  Concerning the
role of federal,  state  local agencies;   the public;   labor;
trade  associations;   and  public  interest groups:   please comment
on whether  roles of these entities  were  appropriately addressed.
Are  there  any recommendations  for actions they could  have taken,
or  should  in future take  to  reduce  accident risks?

-------