Every time you start a car or light a fire, you join in the biggest climate-changing experiment ever undertaken, because you are adding more heat-trapping carbon dioxide (CO) gas to the atmosphere. But while scientists have been studying this ''experiment'' for decades, they still can't predict its outcome with any great certainty. They can't even be sure of its present-day trends.
The issues are complex. Data in hand so far are flimsy. Computer models used to study the CO effect are inadequate. Thus, while the concept of an unplanned climatic experiment has become common currency, there is still little agreement as to its implications.
In fact, scientific understanding of the problem is so rudimentary that research continually turns up counter-intuitive findings, as several results reported in recent months illustrate:
* Increased cloudiness accompanying a CO buildup may counteract its warming effect.
* A warmer global climate might have more bitterly cold winters.
* While higher CO levels enhance plant growth, leaves are less nutritious. Insects have to eat so much more, there could be a net loss of vegetation.
There is wide agreement that, theoretically, a doubling of the air's CO content could bring a several-degree global warming, perhaps in the next century. But Dr. Richard C. J. Somerville of Scripps Institution of Oceanography says that increased cloudiness could counteract that trend. He told the fall meeting of the American Geophysical Union, held in San Francisco earlier this month, that CO-enriched air would be warmer and moister and, hence, would have denser clouds. Because more incoming sunshine would be scattered back to space, ''The cloud feedback might reduce the expected warming over the next century by as much as one half,'' he said.
At the University of East Anglia in Norwich, England, Thomas M. L.Wigley and colleagues at the Climatic Research Unit have compared the coldest 20 consecutive years (1901-20) in Northern Hemisphere records with the warmest 20 years (1934-53). This history suggests that there are more very cold winters when the climate is generally warmer. For example, there are more stalled high-pressure regions which cut off the flow of warm Atlantic Ocean air to central Europe in winter. Such stalled highs also produce fine summer weather and drought.
As for voracious insects, David E. Lincoln of the University of South Carolina and Nasser Sionit and Boyd R. Strain of Duke University have fed caterpillars with leaves of plants grown under various CO levels. They picked 350 parts per million by volume (p.p.m.), roughly the current atmospheric level, and 500 p.p.m. and 650 p.p.m., which span the range of projected increases in CO 2 concentration over the next 50 and 75 years. They found that leaves growing under enriched CO regimes have less protein.
Insects respond to lower leaf protein by eating more or growing less. In the study, they tended to eat more. The scientists report, ''This suggests that the increased levels of plant production at higher carbon dioxide concentrations may be upset by higher (insect feeding), and could even be reduced below current levels.''
To assess adequately the effect of rising CO concentrations, scientists have to take into account both the climatic effects of the gas and its direct effect on plants. For example, a warmer climate might shift rainfall and dry out today's grain belts. But with more CO in the air, plants make more efficient use of soil moisture. They lose less water to the air through transpiration. The two effects might offset each other.
A study released last year by the Carbon Dioxide Assessment Committee of the US National Academy of Sciences estimated that streamflow on Western US watersheds might be cut 40 to 75 percent by the climatic impact of a doubling of atmospheric CO. Yet in a study published last month in Nature, Sherwood B. Idso of the US Water Conservation Laboratory in Phoenix, Ariz., and A. J. Brazel of Arizona State University pointed out that the academy's estimate neglects the effect of rising CO level on plants. The tendency of plants to conserve soil moisture better should increase stream runoff. Brazel and Idso say that, because of this effect, ''40-60 percent increases in streamflow may well be more likely consequences of CO concentration doubling, even in the face of adverse changes in temperature and precipitation.''
The Arizona scientists also put their finger on the central challenge of the CO problem. Both their study and that of the academy have about the same degree of complexity. Yet they produce ''very opposite results.'' This illustrates the fact that, when new factors are taken into account, they ''can dramatically alter current perceptions of future consequences,'' the scientists say.
This is why it is so difficult to formulate national or international policies now to deal with the CO problem - policies such as restricting the use of fossil fuels, whose combustion is a main CO source. Policy depends on present perceptions. It cannot be made soundly when these perceptions are so uncertain.
One of the new factors scientists now have to take into account are other heat-trapping gases. CO is not the only major player in the climate-changing game. The air is gaining more methane from cattle and rice paddies, nitrous oxide from power plants and fertilizer, and chlorofluorocarbons used in refrigerants and plastic foams. It is also gaining carbon monoxide from cars and combustion. According to the World Meteorological Organization, the combined climatic effects of these gases ''can be as large as those estimated due to a carbon dioxide increase.''
In September the US Lawrence Livermore National Laboratory published a review of the subject, which noted that, while data suggest a gradual warming over the past hundred years, ''identifying the precise contribution from any single factor is not yet possible.'' It added, ''One major problem . . . is the lack of adequate data bases covering the range of potential influences other than CO .''
Under these circumstances, the great climate-changing experiment remains a source of wonder and of long-range concern. Research to clear up the uncertainties should be supported with some sense of urgency. But the alarms often raised about melting the polar icecaps and flooding coastal cities are largely speculation. And the calls that have been made to ''do something now'' to begin to restrict the burning of fossil fuels are rather premature.
A Thursday column. Robert C. Cowen is the Monitor's natural science editor.