Published by AAAS which the water level represents the stock of atmospheric CO2. Like any stock, atmospheric CO2 rises when the inflow to the tub (emissions) exceeds the outflow (net removal), is unchanging when inflow equals outflow, and falls when outflow exceeds inflow. Participants were informed that anthropogenic CO2 emissions are now roughly double net removal, so the tub is filling.

Yet, 84% drew patterns that violated the principles of accumulation. If emissions followed the path in the typical example shown, atmospheric CO2 would continue to rise. Nearly two-thirds of the participants asserted that atmospheric GHGs can stabilize even though emissions continuously exceed removal—analogous to arguing a bathtub continuously filled faster than it drains will never overflow. Most believe that stopping the growth of emissions stops the growth of GHG concentrations. The erroneous belief that stabilizing emissions would quickly stabilize the climate supports wait-and-see policies but violates basic laws of physics. Training in science does not prevent these errors. Three-fifths of the participants have degrees in science, technology, engineering, or mathematics (STEM); most others were trained in economics. Over 30% hold a prior graduate degree, 70% of these in STEM. These individuals are demographically similar to influential leaders in business, government, and the media, though with more STEM training than most. It is tempting to respond to these discouraging results by arguing that poor public understanding of climate change is unimportant because policy should be informed by scientific expertise. Many call for a new Manhattan Project to address the challenge (15, 16). The desire for such technical solutions is understandable. In 1939, scientists directly alerted the nation’s leaders to developments in atomic physics, then, by focusing enough money and genius in the deserts of New Mexico, created nuclear weapons in just 6 years. Science has arguably never affected geopolitical outcomes more decisively.