Quake Renews Debate Over Vertical Thrust

Of all the brutal surprises of the Northridge earthquake, its intense vertical motion has been one of the most discussed. The quake’s strong upward heaving has become a popular explanation for the failure of parking garages and other structures that were fortified only against lateral force.

To underscore the unexpected power of the quake’s upward motion, some owners and designers of trashed buildings have noted that the Uniform Building Code, adopted as law by California cities and counties, does not significantly address vertical thrust as a cause of structural failure.

But more than two decades ago, an official with the National Oceanic and Atmospheric Administration wrote that the Sylmar earthquake of 1971 “produced evidence of vertical accelerations far in excess of any ever recorded instrumentally.”

And a committee of the Oakland-based Earthquake Engineering Research Institute, discussing the effects of the Sylmar quake, cited “strong evidences of severe vertical accelerations which could contribute to increased damage.”

Similar descriptions of the Northridge quake may be a case of “short-term engineering memory,” as one government seismologist put it.

In the wake of the Northridge temblor, engineers involved in the code-revision process say the significance of vertical motion--on the back burner since the Sylmar quake--deserves another look.

But as if to illustrate that little about an earthquake is simple, some experts still doubt that vertical thrust played a primary role in structural failures. These authorities fear the vertical explanation is being oversold and may divert attention from what they consider to be a more pervasive problem--strong cost-cutting pressures that can lead to inadequate designs.

“It’s always nice to blame Mother Nature,” but blame is “as much on the human side as the nature side,” said Helmut Krawinkler, co-director of the John A. Blume Earthquake Engineering Center at Stanford University.

“I’m not in the group that’s ready to jump on the vertical bandwagon,” said Wilfred D. Iwan, director of Caltech’s Earthquake Engineering Research Laboratory and chairman of the state Seismic Safety Commission. “But everywhere I go I hear people using this to explain the poor performance of certain structures.”

As currently written, the building code reflects the prevailing view that the risk of structural collapse in a big quake stems from sideways shaking.

That consensus survived the Sylmar quake, at least partly because the significance--and even the accuracy--of its high vertical-thrust recordings were a matter of dispute.

Centered beneath the San Gabriel Mountains, the Sylmar quake had a magnitude of 6.4. At the time, only a few strong-motion accelerographs were deployed in the immediate quake area. These instruments measure the lateral and vertical ground motion at specific points as a fraction of the force of gravity--expressed as 1.0g.

One instrument at Pacoima Dam above Sylmar recorded extraordinarily sharp vertical peaks of .5g to .75g. But many experts were not impressed, and when the code was strengthened in 1974, the revisions did not reflect this finding.

“People tended to discard that record (at Pacoima Dam) as unrepresentative,” said Filip Filippou of the Earthquake Engineering Research Center at UC Berkeley. It “always had some kind of aura of unbelievability to it because it was just one. . . . Maybe it explains a little bit why we haven’t done anything, if you want to call it that.”

Some questioned the accuracy of the reading, believing that the instrument’s perch atop a steep ridge above the dam made it vulnerable to topographic effects. Debate continued for years on whether the readings were valid or “aberrations due to instrumental error . . . because the strong-motion recording instruments don’t always work,” said Bob Chittendon, chairman of the seismology committee of the Structural Engineers Assn. of California.

Chittendon’s organization is largely responsible for the code updates published every three years, drafting changes for the International Conference of Building Officials, which administers the code. Chittendon and others described code revision as a ponderous, consensus-building process in which changes often take years to effect.

“There is a tremendous inertia in the profession,” said Vitelmo Bertero, a research engineer with the Earthquake Engineering Research Center in Berkeley. Engineers, like other people, tend “to forget very soon the bad things,” so that “if something is not done after two or three years after an earthquake, they forget about it,” Bertero said. “This is the real problem.”

Bertero and others said vertical motion deserves a new look based on the record of the Northridge quake. Still, several experts noted that the temblor’s vertical thrust was not as remarkable as the intensity of its ground motion generally.

“This earthquake stands out, both in its horizontals and verticals, as being high,” said Paul Somerville, an engineering seismologist with Woodward-Clyde Consultants in Pasadena. “It looks like ground motion you’d expect from a larger (than magnitude 6.8) earthquake.”

However, Somerville and others said that the vertical accelerations at most sites were about two-thirds of the horizontal thrust, as is typical. “I don’t see anything unusual about the vertical ground motion,” Somerville said. “In relation to the horizontal motions, the vertical motions weren’t exceptional.”

At one site in Tarzana, an accelerograph recorded an extraordinarily high peak horizontal force of 1.8g--nearly twice the force of gravity--and a vertical acceleration of 1.2g. Pacoima Dam again recorded the highest readings, including horizontal and vertical accelerations as strong as 2.01g and 1.6g respectively. At least three other sites in the quake area--in Newhall, Sylmar and Arleta--experienced vertical thrust of more than .5g, according to data from the state Division of Mines and Geology’s strong-motion instrumentation program.

Nearly silent on the question of vertical motion, the building code assumes that the ability of structures to hold their own weight, with a margin of safety, is sufficient to prevent collapse from upward earthquake thrust. In post-mortems on the collapsed parking structures at Cal State Northridge and the Northridge Fashion Center, engineers involved in designing those projects suggested that the blind spot in the code accounted for the failures.

Some experts said long-span structures such as parking garages, bridges and roadways may be particularly vulnerable to vertical thrust. “Personally, I think we need to address the vertical accelerations,” Chittendon said.

Others, however, including Ron Hamburger, a member of the seismology committee of the structural engineers association, believe the vertical explanation may be diverting attention from poor design.

“It may eventually be found that vertical ground motion was a factor in some collapses,” Hamburger said. “But based on my own observations, the structures which performed poorly generally had significant problems in their lateral load-resisting systems and would probably have done poorly even if no vertical ground motion had occurred.”

With respect to sideways motion, experts said the code requires structures to successfully resist earthquakes with lateral accelerations of .4g, but they said well-designed buildings should stand up to considerably greater loads.

“There should be almost no chance of collapse at .4g,” Hamburger said. “The (code’s) intent is that you will not collapse for any earthquake you will ever see.”

In fact, accelerograph recordings during the Northridge earthquake show that peak horizontal motion was well above .4g at some locations. “This is probably the first event . . . that really tested the code at probably the upper limit,” said Bill Holmes, a member of an Earthquake Engineering Research Institute reconnaissance team that inspected quake damage.

There was “very intense shaking right under a populated area, so many code-designed structures were tested, and with the exception of parking structures that collapsed, one could make the argument . . . that most of the structures were successful,” Holmes said.

Holmes and others noted that earthquake stress on buildings is more complex than what is reflected in ground-motion readings. Both seismic waves and buildings have frequencies, and the building will not take the full wallop of energy unless the frequency of the strongest waves are roughly matched with the frequency of the building. Thus, large peak ground motion, if carried on high-frequency waves, may trigger little shaking in a building of much lower frequency.

Because the Northridge quake came within months of a scheduled code update, most major revisions will have to wait until 1997. As has been the case in the past, the aim of code improvements will not be to prevent building damage--which is forecast in major quakes--but to avert life-threatening collapses.

But code improvements are only half the battle, some experts say, due to a tendency to design buildings to the raw edge of the code.

“The code is not the maximum. The code is the minimum,” Chittendon said. “You need to do things in addition to the code.”

But as Krawinkler put it: “If you provide more safety . . . it doesn’t come for free.

“If you tell me money is no object, I could guarantee you that I can design a building so that it will not collapse in any conceivable earthquake,” he said.

But “money is always an object.”

* RELATED STORIES: B1

Earth’s Moves

Faults are fractures in the earth’s crust along which vertical and horizontal movements have taken place. The Northridge quake occurred on a thrust fault, causing more vertical movement than it would have on a strike slip fault.

Strike Slip Fault

Movement is along adjacent horizontal planes and the fault plane, the rocks affected by faulting, is nearly vertical. The two sides slide past each other horizontally during a quake, causing sideways movement. The San Andreas fault is the largest such fault in California.

Thrust Fault

Movement is characterized by upward motion and the fault is inclined, usually dipping between 25 and 60 degrees. The block sitting on top of the plane moves up and past the block below the plane, causing both horizontal and vertical motion. Vertical motion during the Northridge quake was six to nine feet at the depth of the fault, about 11 miles below the surface, according to experts.

A fault at a 45-degree angle would produce about equal horizontal and vertical movement; one at more than 45 degrees would produce more vertical and less horizontal movement.

The previously unknown fault that caused the Northridge quake is believed to be about 35 degrees, producing a bit more horizontal than vertical movement.

Source: Geologist Jay Namson; Researched by JULIE SHEER