Recently a New York Supreme Court rejected computer fire modeling to determine the cause of a fire. Holding that the fire modeling was not “generally acceptable as reliable” under Frye v. United States, 293 F. 1013 (D.C.Cir. 1923), the Court precluded the evidence.
In Santos v. State Farm and Casualty Company, __N.Y.S.2d_, 2010 WL 2610656 (N.Y.Sup. Ct. 2010), the plaintiff insured brought suit against the defendant insurer to recover under a homeowners’ policy damages sustained as a result of fire damage. The insurer offered Dr. Josef Urbas, an associate professor at the University of North Carolina, as an expert. The insurer, through Dr. Urbas, sought to introduce into evidence computer fire modeling using the National Institute of Standards and Technology (“NIST”) Fire Dynamics Simulator (Version 4)(“FDS”). FDS is a computational fluid dynamics (CFD) model of fire-driven fluid flow. See National Institute of Standards and Technology Special Publication 1019-5, p.3 (October 2007).
Plaintiff sought to preclude the fire modeling contending that it was offered to determine the origin and cause of the fire and, as such, it did not meet the standard for admissibility under Frye v. United States, 293 F. 1013 (D.C.Cir. 1923). The well-established rule of Frye is that expert testimony based on scientific principles or procedures is admissible but only after a principle or procedure has gained “general acceptance” in its specified field. Frye, at 1014. The scientific principles need not be “unanimously indorsed” by the scientific community but must be “generally acceptable as reliable” in the particular scientific field. People v. Wesley, 83 N.Y. 2d 417, 422, 611 N.Y.S. 2d 97 623 W.E. 2d 451 (1994).
In furtherance of his challenge to the FDS modeling, plaintiff offered expert testimony from Eugene J. West, a fire and arson consultant. Santos, 2010 WL 2610656 at *1. The Supreme Court outlined Mr. West’s extensive experience with fire investigation including his work with mines and incendiary devices in Vietnam, work with the New York City Bureau of Fire Investigations (Fire Marshall’s office), as an arson instructor at the FBI Academy in the field of fatal fire investigations. Id. at *2. Significantly, Mr. West testified that he was familiar with computer fire modeling, but that he would not use it in conducting fire investigations because the results are dependent upon an analysis of the exact composition and construction of where a fire occurred, which can often only be the subject of speculation. Id. Further, he testified that computer fire modeling was not generally accepted by the New York City Fire Department and, in his opinion, was not generally accepted as an investigative tool by the fire investigative community. Id. Though it was used in the World Trade Center investigation, it was only for demonstrative purposes. Id. Finally, the National Fire Protection Association (“NFPA”) condones its use, but with the caveat that the results are only as good as the data supplied. Id.
Having determined that the plaintiff made a prima facie showing that computer fire modeling would be novel for the purpose of determining fire cause and was not generally accepted in the fire investigative community, the Supreme Court turned to the testimony of defendant’s expert, Dr. Urbas, for evidence of its general acceptance. Id.
The Court outlined Dr. Urbas’ testimony noting that he teaches fire dynamics and computer fire modeling, that two textbooks he uses refer to fire modeling, and that numerous fire science programs involve fire modeling studies. He further testified that fire modeling is endorsed and used by numerous organizations, including the NFPA, American Standards and Testing and Materials, E5 Committee, NIST, and others. He further testified that fire modeling of fire dynamics is not a new science and is based on laws of physics and equations which have been generally accepted in the community.
Though the Court recognized that fire modeling is used by these organizations and agencies, the Court determined that the use of fire modeling was not generally accepted in the field of fire investigation. Specifically, the Court determined that the scientific community uses computer fire modeling to predict the course of fires under a particular set of circumstances to perform fire prevention and safety assessments, not to determine fire causation. Id. at *3. Thus, while it may very well be generally accepted “in the regulatory/design community,” Id. at *4; it is not generally accepted in fire investigation to determine the origin or cause of a fire. Id. at *3.
Defendant attempted to characterize its expert’s testimony differently and argued that the FDS modeling was not intended to establish the origin and cause, but rather was offered to demonstrate how the fire would spread. Id. *3. Dr. Urbas would testify that the fire modeling established a time line that corresponded to a particular origin and that the building damage corresponded to the fire modeling results. Id. The modeling “essentially verified the hypothesis as to the ignition source or cause of the fire.” Id. The Court, however, was not persuaded. Further, the Court noted that some cases have allowed fire dynamics to be used in evidence, the Defendant offered none that involved the use of fire modeling under the Frye standard.
Indeed, the U.S. Supreme Court in Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), abandoned the general acceptance test of Frye finding it superseded by the more liberal relevancy test of Fed.R.Evid. 702. Applying Daubert, federal courts have admitted computer simulations, even determining that the NISI FDS simulation has been “generally accepted by the relevant scientific community.” Turner v. Liberty Mutual Fire Ins. Co., 2007 WL 2713062 (N.D. Ohio 2007). The Frye general acceptance test, however, continues to be the standard for determining reliability and admissibility of expert testimony in numerous states, including New York.
Finally, the Santos Court noted that fire modeling is unreliable in determining fire causation. Specifically, as acknowledged by Dr. Urbas, the accuracy of fire modeling depends on the accuracy of the raw data used in running the model with a resultant margin of error ranging from 15% to 80%. Notably, in a Daubert jurisdiction, the fact of the error rate might be deemed an issue more appropriately raised in cross-examination, rather than as a basis for preclusion. See Turner, 2007 WL 2713062 at*3, citing Daubert, 509 U.S. at 593.
Santos v. State Farm and Casualty Company, ____N.Y.S.2d___, 2010 WL 2610656 (N.Y.Sup. Ct. 2010), creates a significant barrier to plaintiffs and defendants alike seeking to admit fire modeling evidence to verify a hypothesis as to the origin and cause of a fire, particularly in Frye jurisdictions. However, experts and attorneys should not be hindered from using fire modeling which is increasingly accepted and has greater chances of admissibility in the case of a Daubert challenge or as a demonstrative tool.
– Karey P. Pond
