Researchers have already developed a prototype, but it is so fragile that it probably would fail during a large quake, said Douglas Given, the early earthquake warning project coordinator with the U.S. Geological Survey. And given California's enormous number of earthquake faults, more sensor stations are needed.

But it remains unclear how much support the system has in Sacramento. California's budget picture has improved since voters approved the Proposition 30 tax increases last year.

But some politicians have warned against additional spending. And earthquake legislation has generally fared poorly in the last decade, particularly proposals that would add costs for property owners. Quake experts said the lack of a destructive temblor since the 1994 Northridge quake has lessened the pressure for quake regulations.

"Once we get another large earthquake, I'm certain everybody would say, 'We should've had it,'" said Thomas Heaton, a Caltech professor of engineering seismology.

Padilla said he was seeking funding from state and federal sources.

"It's going to be a challenge. If it was easy, it would've already been done," he said.

The concept of an early warning system is an old one. Japan developed one three decades ago so it could slow down and stop bullet trains before quakes hit. In 2007, Japan rolled out a nationwide system alerting cellphone users.

The warning system works when sensors in the ground detect the first signs of earth movement, known as P waves, that travel at the speed of sound. The more damaging shaking, called the S wave, lags behind at an even slower speed.

Although people in Japan might have received only seconds of advance warning through TV, radio or cellphones, authorities have funneled the alerts to take action automatically, like sounding an alarm at construction sites to evacuate workers, ordering crane operators to lower their loads, and opening the doors in locked CT scan rooms to allow patients to leave safely before the shaking begins.

California's earthquake scientists have long struggled to gain attention for an early warning system. Until recently, researchers were spending only about $400,000 a year developing the technology; and in 2011, they received a $6-million grant from the Palo Alto-based Gordon and Betty Moore Foundation to continue work on it.

"We will have an earthquake warning system, but will it be before or after the next Big One?" Padilla said. "It ought to be before."

ron.lin@latimes.com

rosanna.xia@latimes.com