Jack W. Reed, JWR, Inc., and William H. Zehrt, Jr., U.S. Army Engineering and Support Center, Huntsville. 30 June 1998.
Report (132 pages) prepared for the DOD Explosives Safety Board, and being reproduced in electronic form (May 1999) by DTRA-DASIAC, here at Kirtland AFB NM. (See note at bottom)
A reduced summary version follows:
On May 4, 1988, fire and explosions destroyed the PEPCON missile fuel processing plant in Henderson, Nevada, sending damaging blast waves across both Henderson and Las Vegas. The largest single detonation, near 1204 PDT, gave approximately 250 tons TNT (HE) airblast equivalent surface burst - equivalent to 1-kiloton nuclear (NE) free air burst. Insurance companies paid nearly 17,000 claims totaling approximately $70 million to property owners for damages resulting from the accident. They then assembled and filed one lawsuit, in 1991, to recover their costs. In consequence, copies of all claims were furnished to the defense in "discovery"; they were then copied into computer files for analysis. When the suit was settled out of court, these records were not sealed, and a copy held by the consultant, JWR, Inc., was available for further analysis. This was finally arranged in 1995 with support from the DOD Explosives Safety Board, through the U.S. Army Corps of Engineers, Huntsville AL. Since almost all accidental explosion information gets locked in insuror's proprietary files for possible litigation, this event provided an unprecedented opportunity to evaluate community damages from explosion airblast.
All claims were sorted by street name and address number, and labeled by use: as single-family residence (SFR), multi-family structure types, commercial or public buildings, etc. Census Bureau computer "Tiger Files" were used to locate each claims address in relation to the explosion and place it in its appropriate one-quarter mile grid square on the survey map. Nevada Clark County Clerk computer file records showed for address property use, construction date, last sale date and price, and floor area. Claims files showed insuror, claim number, dollar damages categorized by windows, doors, and other. Many claims did not contain damage amounts, apparently for lack of follow-up documentation or rejection from deductible limitations.
Once all these data components were assembled into a sorted record line for each claim, numerous claim duplications were seen and corrected. Some addresses showed claims to different insurors, some different valued claims were listed for single insurors, and other cases showed inexplicable and dubious factors. One claim, for over $5000, was paid to a Post Office box number. A few addresses could not be positively located. These were all withdrawn from their files, leaving just over 15,000 usable claims records.
Front Boy Service Co., Las Vegas, produces a map book for Greater Las Vegas, "Front Boy's 89-90 Greater Las Vegas Street Directory, XXII Edition, 1989," with facing pages that show north and south half-townships, divided into quarter-quarter sections (sixteen grid squares per square mile). Twenty two such pages covered the area with claims, as shown in PEPCON Report Figure 21.with page number and number of claims submitted from each half-township.
Surface weather information, temperature and wind vector, at the explosion time was obtained from the National Weather Services office at McCarran Field, a few miles WNW from the blast. Upper air radiosonde weather balloon measurement was made earlier at Camp Mercury, 90 km NW from Las Vegas, on a regular twice-daily world-wide synoptic schedule. These reports were used with computer program BLASTO (JWR, Inc., 1994) to calculate directed sound velocities versus altitude that refracted and enhanced airblast wave propagation across Las Vegas. Combined with explosion yield, BLASTO provided weather-dependent airblast overpressure estimates for each claim address. This allowed various correlations of damage versus explosion airblast overpressure, shown in the report. An important bottom-line result, shown in PEPCON Report Figure 30, is that the glass damage model (GDM), derived from a 1963 Medina Facility explosion, near San Antonio TX, agreed quite well with PEPCON window damage claims.
The GDM provided breakage probabilities versus overpressue for "typical" window panes, 2 ft x 2 ft SS (single-strenth) glass, but PEPCON claims did not generally specify the number of panes broken in each residence. Floor area statistics, shown in PEPCON Report Figure 20, allowed for an average 40 "typical" panes per house, or 40,000 panes per 1,000 houses, which was used with estimated overpressure to give the total number broken. With the hypergeometric equation, the expected number of houses with no broken panes could be calculated, and the complementary number with some broken equaled the number of claims, per 1,000 single-family residences. Aerial photographs of the most-affected neighborhoods were used to estimate the number of houses in each grid square, and a proportionate number of claims.
Many analyses were made for this report, with one example shown in PEPCON Report Figure 33, of number and average amount of claims versus airblast overpressure. The dip in claims around 1100 Pa is not significant, resulting from relatively open lands that happened to be hit by that overpressure, with a railroad right-of-way and a golf course. Claims dollars mostly followed an exponential approximation line from 200-1700 Pa overpressure. Below the long-established 200 Pa window-damage threshold, claims numbers dropped toward zero, while above 1700 Pa, there were few houses of relatively low value.
These results provide a basis for predicting damages and costs from planned explosion tests or postulated accidents. They also can be used to validate or question damage claims from actual explosions. With an assumed explosion yield, and a weather estimate from either observations, forecasts, or climatology, an overpressure prediction can be made by BLASTO for a specified target distance and direction. Damage intensity (claims per 1,000 SFRs) may be estimated for that overpressure and proportioned to the number of SFRs in the target zone. The number of SFRS can be obtained from aerial photographs, site survey, estimated from census population at about two persons per SFR, or at 19 panes per capita from San Antonio, 1963. Multiplying the average claim amount for that overpressure by the number of claims yields the damage cost for SFRs. The report provides proportional costs versus overpessures for other types of residences and business or public structures that can be added to give total community damage cost expectations.
Probably the most significant finding from these studies is that damages greatly exceeded what could be interpreted from previous housing exposures to large test explosions, both HE and NE, as shown byPEPCON Report Figure 46. Explosives safe separation standards generally were established to limit damages to 5% of replacement cost for neighboring community structures. At PEPCON, however, damages exceeded 20% at these established safe separation distances. This likely resulted from lower construction standards and lighter materials in modern building practice. Also, there was much shoddy construction in boom-town growth of Las Vegas. And, in such a disaster, labor and material shortages escalated costs and insurors became lax in inspecting and reimbursing the mass of smaller claims.
A final note, unrelated to structural damage costs, is about injuries. It has been impossible to penetrate doctor-patient privilege to get injury data from this accident. News media coverage, however, reported over 200 people entered hospital emergency rooms for treatment, mostly from glass lacerations. An integration across the number of claims related to window pane damage probability, concluded that at least 25,000 panes were broken. Thus, roughly 100 broken windows caused each injury. But another anonymous rule-of-thumb, that one of sixteen emergency room admissions does not leave alive, was not in effect. Only two died, both company officers going down with their ship, with one more possible death outside the plant boundary.
Fortunately, there were no houses within a mile of the PEPCON plant, so there were no high-velocity glass shards flung by larger overpressures that could cause more serious injuries. This might not be typical of some other possible accidental explosions. And it is, in my opinion, lucky that PEPCON blew up when it did, before further southward development of Las Vegas had occurred. Today, that plant site is surrounded by new housing.
NOTE: The CDROM was received yesterday (6/15/99, and it looks good) from DASIAC. They will send the master to DOD Explosives Safety Board, who will, I believe contract for reproduction and distribution. It may take months. Meanwhile, I expect to be allowed input for the distribution list; let me know if you need a copy. Until then, contact me, as I am working on an arrangement for copying my CDROM example. Since it has become government property, and is unclassified and uncopyrighted, so I think copies are legal. In this CD, report text and figures are read with Adobe Acrobat, which is not included but most everyone has anyway. Appendix A figures of laboratory glass breakage tests are provided separately as files. The damage claims data base, of about 15,000 claims, is included in formatted ASCII text; also in PIF files but with space separators that are hard to read because empty cells are glossed over. To do your own analyses of these data, read the FORTRAN formatted files or import them into EXCEL or your preferred spread-sheet.