Chapter 9: Fire Simulations
Contradictions between Floor 12 Fire Simulations and Other Evidence
Figure 9-11 from NCSTAR 1-9 (page 383) depicts the upper layer air temperatures on the 12th floor fire simulation. As can be seen therein, significant fires are present across at least half of the north face of the building at 5:00pm.
This part of the fire simulation presents two problems. First, it contradicts an earlier report issued by NIST regarding the fires on floor 12. Second, it contradicts NIST’s own photographic evidence of the fire activity on floor 12.
COMMENT: Appendix L to NIST’s June 2004 “Progress Report on the Federal Building and Fire Safety Investigation of the World Trade Center” contains NIST’s “Interim Report on WTC 7”. (See
http://wtc.nist.gov/progress_report_june04/appendixl.pdf) On page L-26 of this interim report, NIST states that “Around 4:45 p.m., a photograph showed fires on Floors 7, 8, 9, and 11 near the middle of the north face; Floor 12 was burned out by this time.”
REASON FOR COMMENT: The contrast between NIST’s prior assertion that floor 12 was “burned out” by 4:45pm, and NIST’s current computer model, that shows a raging inferno at 5:00pm, could not be more apparent. This discrepancy calls into question the veracity of the Report.
SUGGESTED REVISION: This discrepancy must be acknowledged and explained in the Report. Furthermore, the photographic or other visual evidence NIST relied upon for its statement in Appendix L that floor 12 was burned out by 4:45pm must be included in the final version of its report.
COMMENT: To support NIST’s assertion that there was indeed fire present on floor 12 at 5:00pm, NIST has provided a single photograph from an “unknown source” (Figure 5-152, NCSTAR 1-9, p. 237), that was purportedly taken at around 5:00pm, and shows fire in the two windows that comprise the northwest corner. NIST contends that it has determined that this photograph was taken at approximately 5:00pm, with a margin of error of “at least 10 minutes,” using shadow analysis.
REASON FOR COMMENT: We find it unlikely that NIST could estimate the time the “unknown source” photograph in Figure 5-152 was taken with such accuracy.
SUGGESTED REVISION: NIST must explain how it was able to estimate the photograph’s time using shadow analysis to a margin of error even close to 10 minutes.
COMMENT: The following graphic is excerpted from Figure 9-11, and purports to describe the state of the fires on the 12th floor of WTC 7 at 5:00pm:
As can be seen, this graphic depicts raging fires across at least half of the north face of the building. However, when compared with Figure 5-152, which only shows a small fire in the extreme northwest corner, clearly the computer model is not representative of reality.
REASON FOR COMMENT: It appears that NIST’s computer fire simulations are not representative at all of the fires actually occurring in WTC 7.
SUGGESTED REVISION: NIST needs to describe why (assuming Figure 5-152 accurately describes the floor 12 fires at about 5:00pm) the computer models show significant fires across at least half of the north side of the building at 5:00pm. NIST should clearly explain why its fire simulation models of the 12th floor should be accepted by the public as an accurate representation of the fires actually occurring in WTC 7.
Separately submitted by Chris Sarns and Richard Gage is a graphic that compares NIST’s computer model fire data for floor 12 with actual pictures of the fires in WTC 7. It is attached hereto as Exhibit A. They present a more realistic depiction of what a computer model for the floor 12 fires should look like if it were to agree with the available visual evidence. NIST should take this into consideration when they are re-running their computer models based on these public comments, and revise their Report to use computer models that are more representative of reality, which would look more like the depictions contained therein.
Combustible Fuel Loading on Floors 11 and 12
COMMENT: This comment relates to NIST’s assumptions regarding combustible fuel loading for the 11th and 12th floors. In NCSTAR 1-9, at p. 375 (para. 1, sent. 7-9) NIST states:
NIST assumed that the combustible mass of furniture was about the same in an office as in a cubicle. Since the loading of other combustibles was reported to have been high on the 11th and 12th floors (Chapter 3), NIST assumed that the total combustible mass in an office was double that of a cubicle. Thus, the average combustible fuel load on the 11th and 12th floors was estimated as 32kg/m2.
However, Chapter 3 tells us that, contrary to NIST’s assertions in Chapter 9, the loading of other combustibles was not reported to have been high on the 11th and 12th floors. On page 55 (para. 6, sent. 1) of NCSTAR 1-9, NIST reports that the U.S. Securities and Exchange Commission occupied the 11th and 12th floors and the north side of the 13th floor. On page 56 (para. 1, sent. 1) NIST further reports that American Express occupied the southwest sector of the 13th floor. On the same page, NIST reports that the “combustible load in the offices was described as high by interviewed American Express managers.” (NCSTAR 1-9, p. 56, para. 4, sent. 3)
REASON FOR COMMENT: Recall that American Express occupied only the southwest sector of the 13th floor. How, then can NIST credibly claim that the combustible load on the entirety of the 11th and 12 floors, both occupied solely by the SEC, was reported to have been high? Were American Express managers given regular access to the SEC offices, such that they would be qualified to comment on the combustible fuel load there? Moreover, are American Express managers qualified to give an opinion on the quantity of combustible fuel load as compared to offices in the Twin Towers?
SUGGESTED REVISION: Clearly American Express personnel are competent to provide information only on the state of the American Express offices, which were confined to the southwest sector of the 13th floor. NIST must provide real support for its assertion that the combustible load on the 11th and 12th floors was high in order to merit any increase in estimated average combustible fuel load on these floors. If it cannot provide such support, it should re-run its computer models with the lower combustible fuel load on these floors and report those results to the scientific community and the American public.
Combustible Fuel Loading on Floor 13
COMMENT: This comment is regarding NIST’s treatment of the combustible fuel load of the 13th floor. On page 375 of NCSTAR 1-9 (para. 1, sent. 8, 9) NIST states as follows: “The density of combustibles on the 13th floor was varied and not well known. The average value [for the 13th floor] was assumed to be the same as the 12th floor.” Here again, the only reported description of the combustible load on the 13th floor was from American Express managers, who were competent to comment only on the southwest sector of the 13th floor. In Chapter 3 of NCSTAR 1-9, page 57 (para. 2, sent. 2, 3) NIST reports that in the SEC occupied sections of northern perimeter of the 13th floor were “a hearing room and multiple testimony rooms facing it. There were additional testimony rooms on the northern portion of the east and west sides of the floor, and a storage room at the northwest corner.”
Importantly, NIST reports that the “testimony rooms were sparsely furnished, with just a table and a few chairs.” (NCSTAR 1-9, p. 57, para. 2, sent. 4) Furthermore, an examination of the schematic diagram of floor 13 (Figure 3-8, p. 57) reveals that the hearing room appears similar to a court room. Court rooms are also sparsely furnished, with a few tables and chairs. Finally, it is doubtful that there was any appreciable level of additional combustibles present in these testimony and hearing rooms.
REASON FOR COMMENT: NIST has apparently greatly overestimated the fuel loading on the 13th floor.
SUGGESTED REVISION: NIST must justify its use of the higher combustible fuel load on the 13th floor in Chapter 9 of the Report with more than just bare assertions. NIST clearly had more information available to it regarding the layout and make up of floor 13, as reported in Chapter 3, than it lets on in Chapter 9. This discrepancy must be reconciled.
Combustible Load Sensitivity Tests
COMMENT: NIST claims that it did sensitivity tests to determine whether these exorbitant combustible fuel loads adversely affected the outcome of its simulations. However, the fact that NIST even performed the sensitivity tests brings up the question of why NIST went to the trouble of increasing the fuel load in the first place if it would have a negligible effect on the simulation. That point aside, Chapter 9 contains statements that directly contradict the results of these alleged sensitivity tests.
On page 381 of NCSTAR 1-9 (para. 3, sent. 3) NIST flatly states that, in its fire simulations for the 12th floor, “[t]he [fire] spread rate was about one-third to one-half slower than that on the lower floors due to the higher fuel load [on the 12th floor simulation].” NIST goes on to report that the burn time across the north face in the simulation was longer than observed in the visual evidence. (NCSTAR 1-9, p. 381, para. 3, sent. 4) NIST then rejects the possibility that this could have resulted from the fuel load being too high, citing the sensitivity analysis in Section 9.3.3. (para. 3, sent. 4-8)
In Section 9.3.3, we find the referenced sensitivity analysis. Here, NIST reports that doubling the fuel load on the 8th floor resulted in the fires moving distinctly more slowly than in the visual evidence. (NCSTAR 1-9, p. 382, para. 5, sent. 1-3) Confusingly, NIST also reports that decreasing the fuel load by more than one-third on floor 12 “showed little effect on the rate of fire progression.” (Id., para. 6, sent. 1-3)
REASON FOR COMMENT: NIST’s contradictory statements raise the question of why reducing the fuel load by more than one-third would show no appreciable effect on the fire rate of progression on the 12th floor, when doubling the fuel load on the 8th floor did result in an appreciable change.
SUGGESTED REVISION: NIST should explain here exactly what the differences in the fire progression rate were in each case and let the public judge whether the effect was “little”. More important, however, is the direct contradiction between NIST’s statement that the “spread rate was about one-third to one-half slower than that on lower floors due to the higher fuel load” (NCSTAR 1-9, p. 381, para. 3, sent. 3) with its statement that decreasing the fuel load to a value equal to that of the lower floors “showed little effect on the fire rate of progression.” (NCSTAR 1-9, p. 382, para. 6, sent. 1-3) Surely NIST can see this direct contradiction. On page 381, it is claimed that higher fuel load slows down the fire spread rate. On page 382, it is claimed that a lower fuel load will not speed up the rate of fire progression. This contradiction must be reconciled.
Fire Simulations for Floors 11 and 13
NIST used the data generated by its 12th floor fire simulation for floors 11 and 13. (NCSTAR 1-9, p. 382, para. 1, 3) The 13th floor simulation used the 12th floor data delayed by one-half hour because visual evidence indicated that the 13th floor fire followed the 12th floor fire. (Id., para. 3, sent. 5) The 11th floor simulation used the 12th floor fire data delayed by 1 hour, although the visual evidence indicated that the 11th floor fire was delayed from the 12th floor fire by 1.5 hours. (NCSTAR 1-9, p. 382, para. 1, sent. 5)
COMMENT: Our first comment in this regard simply notes the discrepancy between the visual evidence that the 11th floor fire was delayed from the 12th floor fire by 1.5 hours, yet in its fire simulations for the 11th floor, it was only delayed from the 12th floor fire by 1.0 hour.
REASON FOR COMMENT: This represents yet another discrepancy in the Report that needs to be rectified.
SUGGESTED REVISION: NIST must explain why the visual evidence was not relied upon for inputs on the 11th floor, when it was relied upon for inputs on the 13th floor. The computer models should be re-run with the 11th floor fire delayed by 1.5 hours, not 1.0 hour, and the results reported accordingly.
COMMENT: Our second comment concerns both the 11th and 13th floor fires. As we demonstrated above, the 12th floor fire simulation is not representative of reality, and likely grossly overestimates the fires that were present there. By using its grossly overestimated 12th floor fire data on both the 11th and 13th floors, it has magnified this error three-fold.
REASON FOR COMMENT: By magnifying an obvious error by three times, the results of all of NIST’s subsequent computer models are again called into question.
SUGGESTED REVISION: The computer models should be re-run for the 12th floor using more realistic fire scenarios, and if NIST can still justify using the 12th floor data on the 11th and 13th floors, it should use that more realistic data on both floors. The results should then be reported accordingly.
COMMENT: Our third comment concerns the propagation of error through NIST’s approach to using a purely computer model driven approach. On page 382 of NCSTAR 1-9 (para. 1-3, sent. last) NIST acknowledges that its computer models for the fires on floors 11 and 13 “could have led to a mild overestimate of the heating on the north side of the floor.”
REASON FOR COMMENT AND SUGGESTED REVISION: In order to assure public confidence in the document, NIST must explain how such an error in overestimating the heating would propagate itself throughout all of NIST’s subsequent computer models, and how such propagation of error will affect the reliability of the ultimate results. The Report should be revised to include such a propagation of error analysis.
