The Cypress Viaduct Collapse, 1989: Strengthened in 1977 — On the WRONG Half.

On 17 October 1989 a mile and a quarter of the Cypress Street Viaduct in Oakland came down. Only the top half. The upper roadway dropped onto the lower roadway and stayed there in one piece, and along almost the whole length the lower columns were still standing. Forty-two people died. We read the report of the Governor's Board of Inquiry (1990), the report to Congress from the General Accounting Office (19 June 1990) and USGS Professional Paper 1552-B, page by page. Every number below is from one of them, with the page. WHAT IS IN THIS EPISODE • The upper frame did not sit on the lower frame — it was pinned to it, in a deliberately weakened joint called a shear key. Forty-four No. 18 bars run up the lower column and all forty-four stop twenty-seven inches below that joint. Four No. 10 bars, fifty-two inches long, are all that crosses it (Board of Inquiry p. 163; USGS fig. 28, p. 22). • The ties holding that joint together are No. 4 bars at twelve inches, hooked at ninety degrees and not turned into the core of the column. At most five of them cross the surface the bent actually failed on (pp. 164, 175). • In some west-side columns a drain pipe, four or six inches across, runs continuously down the middle of the shear key (p. 163). • The viaduct was designed to a 1943 code: a lateral force of 0.06 times the dead load. The instruments nearest the site recorded between 0.26 and 0.29 g (Board p. 28; GAO p. 23). On rock at Yerba Buena Island the same earthquake measured 0.06 g exactly. • The Board's own assessment: "in terms of acceleration intensity, at both rock and bay mud sites, the Loma Prieta was a minor-to-moderate earthquake" (p. 192). • The extent of the collapse is almost identical to the boundary of a tidal marsh that was filled in before the viaduct was built, and which appears on an 1878 map (pp. 158-160). • Forty-eight bents came down. All but seven failed in the same way, along the same surface (p. 174). • The Cypress WAS retrofitted — in 1977, with cable restrainers at every deck hinge. That was phase one of a three-phase programme. Phase three, the multi-column structures, had not begun (GAO pp. 4, 22; USGS p. 19). • Why it had not begun: retrofit projects were ranked against every other highway safety project by the number of lives lost in previous incidents. "Until the Loma Prieta Earthquake, earthquake damage to state bridges had resulted in the loss of only two lives." So seismic work ranked below guard rails (GAO p. 8; USGS p. 2). Caltrans disputed that reading in its formal reply, printed on p. 31 of the same report. • Between Bent 63 and Bent 112, two bents stayed up: 96 and 97. They were type B3 — the only ones in the collapsed stretch with continuous columns on both sides, and therefore the only ones without that joint (pp. 171, 177). SOURCES Competing Against Time — Report to Governor George Deukmejian from the Governor's Board of Inquiry on the 1989 Loma Prieta Earthquake, State of California, 1990 https://libraryarchives.metro.net/dpg... GAO/RCED-90-177 — Loma Prieta Earthquake: Collapse of the Bay Bridge and the Cypress Viaduct, United States General Accounting Office, 19 June 1990 https://www.gao.gov/products/rced-90-177 USGS Professional Paper 1552-B — The Loma Prieta, California, Earthquake of October 17, 1989: Performance of the Built Environment — Highway Systems https://pubs.usgs.gov/pp/pp1552/pp155... IMAGE CREDITS All photographs and figures are from USGS Professional Paper 1552-B: a work of the United States federal government, in the public domain. Figure 34 of that report is NOT used: it carries a separate copyright notice (Roy Williams, Oakland Tribune). CORRECTIONS AND OPEN POINTS Pinned in the comments, with page numbers. If we got something wrong, say so there with a page and we will pin the correction. CHAPTERS 0:00 A mile and a quarter, top half only 2:13 Oakland, 5:04 p.m. 3:23 How it was supposed to stand 9:12 Eighteen years of the wrong half 14:42 Ten seconds 18:26 Bents 96 and 97 21:49 Sources MUSIC AND SOUND Original score and sound design composed for this episode by Collapse Mechanism. The pulse of the score is 2.5 Hz — the first transverse mode measured on the structure itself (Board of Inquiry p. 183). No library music, no stock effects. Collapse Mechanism reconstructs one structural failure per episode from the official inquiry file. We are not engineers and we do not give engineering opinions: we report what the investigating body concluded, and we show you the page. WATCH NEXT • Silver Bridge, 1967 —    • Silver Bridge: One Hidden Crack Was All It...   • Tacoma Narrows, 1940 —    • The Tacoma Narrows Report Already Said It ...   • Every episode, in order —    • Collapse Mechanism — Every Episode   #StructuralEngineering #EngineeringDisaster #ForensicEngineering #CypressViaduct #CivilEngineering #documentary