Every two years, they line up at one of the state's 9,000 Smog Check stations for a 20-minute test that most will pass. It's inconvenient to be sure. But at least when it's over, lots of high-polluting vehicles will have been found and fixed. Right?

CHECK IT OUT: Hossein Mahjoob, manager of Smog Master on Blackstone and Clinton, measures emissions from the exhaust of a car. (Mark Crosse / The Fresno Bee)

Maybe not.

While most California motorists dutifully submit to Smog Check, a small minority evades it. And the state has no way to catch them.

Some go home after the test and remove or adjust their vehicles' pollution controls. Some skimp on maintenance. Some never pass the test but continue driving illegally.

Their vehicles are called "gross polluters," and they make up fewer than 10% of California's cars and small trucks. But they account for more than half of the state's smog-forming emissions from light gasoline vehicles.

Simply put, the Smog Check program "doesn't catch and repair a lot of the cars that need to be repaired," says Joel Schwartz, former executive officer of an independent Smog Check review committee set up by the Legislature.

California's gross polluters include tampered vehicles and poorly maintained vehicles. They include some of the 500,000 pre-1974 vehicles that won a permanent Smog Check exemption from the Legislature five years ago. They even include some late-model vehicles that you might never suspect.

Finding these gross polluters isn't easy -- except when you're sitting behind them in traffic, with clouds of unburned gasoline trailing from their tailpipes through your car's ventilation system to your nose.

But now there are devices that can sniff out gross polluters just as efficiently as your nose. From the roadway's edge, these devices shine light on a vehicle's exhaust stream and measure how much is absorbed. From that, they can tell how much pollution is in the exhaust.

State regulators have known about this "remote sensing" technology for more than a decade. Repeated studies have concluded that it works at least well enough to identify the worst of the worst gross polluters, whose license plates are photographed as the vehicles drive past.

But regulators have shunned the technology.

California's master air pollution plan, written in 1994, called for the state to use remote sensing to identify gross polluters and target them for repair. A pilot study in Sacramento concluded that remote sensing could accurately identify the most polluting 2% of vehicles. But regulators fixated on some minor technical issues and did nothing further.

The 1998-99 state budget included more than $5 million for further testing, intended to find a way to include remote sensing in the state's smog-fighting strategy. But the money has yet to be spent.

Instead, state regulators have proposed improving Smog Check with a variety of measures that boil down to making the test harder to pass, thereby forcing more borderline polluters into the shop for repairs.

That won't help much, in Schwartz's view.

Smog Check, Schwartz says, has run into the law of diminishing returns. You can't squeeze many more emissions reductions from cars that are being tested. To get big reductions in pollution, you have to go after the vehicles that Smog Check isn't catching.

"Think about a place like Fresno," he says.

"If you could somehow identify and induce owners to voluntarily scrap the 10,000 highest emitting cars, you could reduce emissions from on-road motor vehicles by 25% right now. No big programs. No big expenditures."

Most vehicles pass tests

If anything is clear from almost two decades of Smog Check testing, it's that most vehicles pass most of the time.

Back in 1983, the year before Smog Check began, an Air Resources Board consultant looked at test results from 339 vehicles and found that one-eighth of them were responsible for almost half the pollution.

Today, the pattern is even stronger. Motor vehicle emission controls are far more sophisticated now than in 1983. Carburetors are out, fuel injectors in. Sensors route hundreds of details about engine operations to computers that slash emissions and boost fuel economy and performance.

Typically, a new car leaving the showroom floor today emits only one-tenth as many hydrocarbons -- gasoline vapor and other petroleum products -- as the average car on the road, and one-twentieth of what 1960s-era cars emitted.

Cars from the mid-1990s pass their Smog Check tests more than 90% of the time; they have average hydrocarbon emissions of 12 parts per million. Cars from the mid-1970s pass about 70% of the time with average emissions of 57 ppm, about five times as high as the mid-1990s models.

But the averages mask huge differences, even among cars in the same model year. In any group of cars, there will be a few -- maybe four or five out of 100 -- that have hydrocarbon emissions of 1,000 ppm or more.

These superpolluters include cars whose owners have removed the catalytic converters or made other alterations in hopes of improving performance. They include cars whose owners don't maintain them properly, who neglect routine oil changes, who ignore the "check engine" light. They include innocent owners who are driving cars with major malfunctions and don't realize it.

Mile for mile, just one of these "gross polluters" does as much damage to the air as dozens of average cars.

Maybe hundreds.

Getting them off the road, some scientists say, would do more to clean up the air for less money than any other measure.

"If we could just find those high emitters and fix them immediately, and focus all of our resources on that, we wouldn't even have to test everybody else," says former Air Resources Board staff scientist Doug Lawson, who served on a National Research Council panel that reviewed Smog Check programs nationwide.

Not everyone agrees with Lawson that testing everyone else is unnecessary. But there is widespread recognition that attacking gross polluters is an important and long overdue task.

The council's panel cited study after study, some by the air board, some by outsiders, all of them making the same essential point -- that the majority of emissions come from a tiny fraction of vehicles.

A 1996 air board study says that the dirtiest 10% of vehicles emitted 47% of carbon monoxide, 59% of hydrocarbons and 33% of nitrogen oxides. More recent studies suggest that the imbalance is growing, probably because as new cars get steadily cleaner, the dirty cars stand out more.

"You have a handful of vehicles that are producing most of the emissions, and it's driving this whole Smog Check program that requires everybody to be tested," says Steven Moss, a former member of the Smog Check review committee. "These cars need to be gotten rid of, one way or another."

Click to view larger image

Schwartz, who left the committee in December 2000 to work for the Reason Public Policy Institute, a libertarian think tank, argues that the continued presence of gross polluters is an indictment of Smog Check.

"The fact that we find these gross polluters on the road tells you that the Smog Check program isn't finding them and repairing them," he says.

Cheaters a problem

Why aren't they caught?

The reasons are as varied as the factors that create gross polluters in the first place. But chief among them is the one great limitation of Smog Check -- it's a scheduled test, done just once every two years, and always at a time and place of the motorist's choosing.

"This is a human behavior problem," says Lawson, who has been studying Smog Check for more than a decade. Not everyone who drives a gross polluter is cheating, Lawson says. But many are.

A decade ago, Lawson compared results from Smog Check to results for the same vehicles from the state Bureau of Automotive Repair's roadside program, in which cars are flagged down for random, voluntary tests at sites around the state.

Among his findings: Some vehicles came up dirty in the roadside tests just a few days or weeks after they had passed their regular Smog Checks. Those cars should have been clean, but they weren't. Either their Smog Checks were fraudulent, or something made their emissions soar after the test -- maybe a mechanical failure, maybe tampering.

Lawson's analysis couldn't pinpoint a cause. But he suspected cheating was the main explanation.

"Clean for a day," he called the phenomenon.

"It's basic economics and human behavior," he says. People who cheat on Smog Check "don't perceive the economic benefits or the air quality benefits of the program."

More importantly, as Schwartz notes, they also perceive that the test is beatable.

"If you suggest doing scheduled sobriety tests, people immediately realize that it won't work," he says. "The drunks will arrange to be sober when it's time for the test. Why should we expect a scheduled Smog Check to be any different?"

Some cars stay on the road without ever passing a Smog Check.

Tom Wenzel of the Lawrence Berkeley National Laboratory studied vehicle registrations, Smog Check results and other data from 1998 and 1999. He found that one-tenth of the vehicles failing Smog Check never got a subsequent passing grade. Many were junked. But about one in three was still on the road one year later. On average, Wenzel says, those vehicles produced almost twice the hydrocarbons of vehicles that passed.

Wenzel also found that 5% to 10% of the vehicles seen on the road in the state's smoggiest regions, including the San Joaquin Valley, never came in for a Smog Check even though they should have been required to do so.

The air board's own reports acknowledge that Smog Check has fallen far short of its goals. In its most recent evaluation, the agency concluded that Smog Check reduced hydrocarbon emissions in summer 1999 by only half of what the state's 1994 air quality plan required. Nitrogen oxides emissions were reduced even less, only one-quarter of the goal.

Air Resources Board officials declined to discuss Smog Check, referring questions to the Bureau of Automotive Repair, which administers the program.

But an air board spokeswoman says her agency has reservations about targeting gross polluters.

"The ARB is not in the business of removing older, possibly higher-polluting vehicles from their owners' hands or from California's roads," spokeswoman Gennet Paauwe says.

Yet bureau spokesman Glenn Mason says his agency recognizes that it is obligated to look for a way to identify gross polluters between Smog Checks: "The law says to focus the Smog Check program on finding gross polluters," both in the scheduled tests and between them.

The air board's evaluation report points to several additional factors that reduce Smog Check's effectiveness, among them the Legislature's 1997 vote to exempt 500,000 vehicles from the 1966 to 1973 model years.

Three thousand people marched on the Capitol to demand a break for older cars. Thousands more sent letters, including Tonight Show host Jay Leno, who wrote: "The number of cars over the age of 25 on our roads is really quite small ... and most cars that age are driven less than 1,000 miles per year."

But the resulting law not only exempted classic cars owned by wealthy collectors such as Leno, it opened a loophole for old cars of any kind, whether well-maintained or run-down, driven a lot or driven a little.

Subsequent studies have concluded that those older cars account for 1% of miles driven but 4% to 8% of emissions. And without Smog Check, there is little to prevent their owners from removing emissions controls.

Leaking gasoline is another source that the report identified as a gap in Smog Check. One car in 50 is believed to have such leaks, but until last year, those cars could still pass. Even now, a leak isn't caught unless it is heavy enough to be visible.

The report also proposed strengthening Smog Check by sending more cars to specialized stations that conduct only tests, not repairs, and by reducing the emission levels that can make a car fail its test.

It noted that remote sensing had been suggested. But it argued that the pilot study had found "limitations and uncertainties" in remote sensing's ability to find gross polluters. Yet an earlier review of the same data by a bureau contractor concluded that remote sensing could identify the dirtiest 2% of vehicles with minimal risk of falsely labeling clean vehicles as gross polluters.

Now, Paauwe says, air board officials believe that remote sensing is "very good at identifying clean and very dirty vehicles, but not so good at finding those in between."

And she says the agency has other concerns: "The issue for us is that remote sensing is a bit like Big Brother," she says. "There have been some serious concerns over the public wishing to remain anonymous as they drive in their vehicles."

Remote sensing raises concerns

Four states have incorporated remote sensing into air pollution programs.

Colorado and Missouri are using remote sensing to identify clean vehicles that don't need further testing. Arizona used remote sensing for five years to find gross polluters, but too many clean vehicles were tagged and the Legislature ended the program. Scientists who have examined that program say that its definition of a gross polluter was too strict and caused many marginal polluters to be tagged.

Texas is trying to avoid that mistake. For more than three years, it has used remote sensing to identify gross polluters. But to be tagged, a vehicle must emit twice the allowable level of pollutants in two different remote sensing tests.

GROSS POLLUTERS: Numerous studies have shown that a tiny minority of light cars and trucks produce the majority of smog-forming emissions. (Mark Crosse / The Fresno Bee)

"If there's any doubt about whether the vehicles are gross emitters, we don't send the notice," says the program's chief, Jimmy Guckian of the state Department of Public Safety.

In California, Bureau of Automotive Repair and Air Resources Board officials are now working on a proposal for another remote sensing pilot study that would use some of a $5.3 million appropriation left over from the 1998-99 state budget.

But Schwartz, whose Smog Check review committee two years ago issued draft recommendations that the state move ahead with remote sensing, says there have been more than enough studies.

He goes to the bookcases that line his Sacramento office, one block from the air board's headquarters, and counts the file boxes that contain reports on remote sensing -- five, six, seven, eight in all.

"Dozens of studies, many of them published in peer-reviewed journals," he says. "We don't need another pilot project. The technology's been around since the late 1980s and it's been evaluated to death."

Schwartz proposes a different kind of experiment. Set up remote sensors in Fresno, get a few million measurements, then send letters to the very worst polluters offering to buy their cars from them for a little bit more than their actual value.

"Just see what happens," he says. "It's voluntary, and it's important to stress that, because there are people out there saying that the government's going to take your car away. It's voluntary, and for people who don't want to scrap their cars, you can offer them subsidized repairs if they're poor."

The logic is simple -- every time you take a dirty vehicle off the road, you make the air a little cleaner. Take a lot of dirty vehicles off the road and you might make the air a lot cleaner.

That's the same logic as the bureau's buyback program, which offered vehicle owners $1,000 for their cars if they failed their Smog Checks. The program bought more than 25,000 cars in 18 months but it came to an end in February, a victim at least for now of the state's budget crisis.

Meanwhile, millions of motorists continue to pass their Smog Checks every two years, while tens of thousands -- maybe hundreds of thousands -- of gross polluters continue to fill the air with avoidable emissions.

"This is stuff we've known for 12 years or more," Schwartz says. "But it's a very slow process to get major changes made in public policy."