Thursday morning I happened to catch NPR’s Talk of the Nation on whether the extraordinary tornadoes this year had any link to global change. This is not a simple question, but it got me thinking about one aspect of climate change that I haven’t seen discussed much.

Weather is driven by the fact that some parts of the earth-ocean-atmosphere system receive more radiant energy than they emit to space, and other parts radiate more to space than they receive. Energy is transported from regions of excess to regions of deficit by the atmosphere and the oceans, and the result is what we call weather and ocean currents. This is one of the fundamentals of atmospheric science.

What has received less attention is that there are two types of energy imbalance that the oceans and atmosphere must balance. The first is the equator-to-pole imbalance. This is what drives ocean currents and the huge horizontal eddies that we call frontal storms and anticyclones, and the great northward and southward excursions if the jet stream. Individual years may vary greatly, both in time and in space, while the total energy transport stays about the same. A difference in the apportioning between atmosphere and ocean could make a huge difference, and this is the basis for concern that a change in the so-called conveyer belt of the oceans could be a major climatic switch.

But there is a second imbalance, and this is between the surface, which absorbs energy from the shortwave (visible) radiation of the sun, and the upper atmosphere, which radiates longwave (infrared) energy to space and back to the surface. The surface transfers energy to the air near the ground, both in the form of moisture and direct heating, and this energy must be transferred vertically to the upper atmosphere where it is radiated away—largely by water vapor, clouds, carbon dioxide and methane. Certainly some of this vertical transport is accomplished by the great horizontal eddies, via fronts—sloping surfaces where warm air moves up over cold air. But some, especially in summer, is due more directly to warm air rising with very little large-scale horizontal temperature gradient. This process produces the more violent storms—thunderstorms (which produce not only thunder but lighting, hail and tornadoes) and hurricanes.

Horizontal gradients are certainly important as feedback processes. If the Arctic sea ice continues to melt, the Arctic Ocean will absorb much more solar energy, the summer gradient of temperature will decrease, and summer storms of the large-eddy type will decrease. But the direct effect of carbon dioxide and other greenhouse gasses is to increase the vertical energy gradient, and thus the amount of energy that must be transported vertically by the atmosphere.

In current climate models this vertical energy transport is parameterized—that is, it is based on statistics taken from the present-day climate. Why? Because it happens largely through processes of cloud physics that are just too small-scale to be included directly in the models. But any time you hear climate scientists talking about a tipping point, they are really talking about a change that may change the statistics on which those parameterizations are based. If that happens, change may be much larger, and in a totally unexpected direction, than the models predict.

Could the changes be in a direction as to be opposite the prediction, so there is no real change? Possibly, just as it is possible that one storm could put things back the way they were before an earlier one struck. I am more concerned that attempts at modeling past changes we know occurred, like the glacial-interglacial transitions, generally underestimate, not overestimate, the change actually observed.