As the glacier turns

A supposedly decisive indicator of the reality of "global warming" is the supposedly melting Arctic and Antarctic icecaps. Certainly a rigorous thermodynamic deduction implies that polar temperatures should increase more than equatorial ones, leading to less ice up there and down there. The ice caps should shrink in both latitudinal coverage and in thickness. But the "global warming" case here, as elsewhere, falls apart upon closer examination.

The strongest and most worrisome trend is the apparently 30-year-long shrinkage of latitudinal ice coverage in the Arctic. But this trend is ambiguous. The shrinkage is all on the Asiatic side of the Arctic Ocean. The Canadian-Atlantic-European side shows no such shrinkage. There's no significant trend of shrinkage in Greenland. While there is some evidence of rising temperatures in the Arctic, the most characteristic expected trend - stronger warming in the winter, weaker warming in the summer, with a narrowing of summer-winter temperature differences - is absent. Nor is there any consistent trend of narrowing temperature differences between the equator and the Arctic, again something definitely expected from an infrared (IR) opacity-driven "global warming" scenario.

But it is when we turn to trends in the Antarctic that the "global warming" case fizzles. Temperatures there have been unambiguously falling for the last 30 years, not rising. There has been some thinning of the ice cap in the western Antarctic, but that seems to be due to lessened snowfall in the winter, itself apparently a result of cooling temperatures and a weakened evaporation-precipitation cycle. It's exactly like the case of the retreating ice cover on Kilimanjaro.*

For a small warming, we should not necessarily expect lessened ice volume. The ice coverage by latitudinal area would definitely decrease, because the summer isotherms (curves of equal temperature) would move poleward. But slightly warmer summers and somewhat warmer winters mean more humidity and probably more snow. Ice thickness would probably increase in that case. What that means for ice volume (thickness integrated over area) is ambiguous.

For a larger warming, we should definitely expect net ice loss, averaged over a year, year by year. But in that case the main cause would not be rising air temperatures, but rising polar water temperatures. Water takes longer to heat up than does air. But once heated, water is a huge heat reservoir and much more effective at melting ice than warmer air.

The amount of IR-active gases in the Earth's atmosphere is in any case too small to be the dominant driver of the ice cap coverage. The caps are reminders that we're technically living in an Ice Age. Before a few tens of million years ago, the Earth's poles were ice-free and the Earth's general climate was quite a bit warmer. The basic mechanics of permanent ice caps has to do with the position of the continents and the partial or complete blocking of the poleward ocean currents that move excess heat from the tropics.** The timing of ice advance and retreat is driven by the Earth's orbital and orientational variations (Milankovic cycles). IR-active gases, as far as anyone can tell, are at most a footnote to this story.
---
* Again, another reminder: climate = temperature and humidity/evaporation-precipitation as two (at least) independent variables, not just temperature!

** Quick sanity check: The most cooling possibility is a landmass over top of a pole. The second most cooling possibility is a landlocked polar ocean. For the present Earth, the Antarctic fits the former, the Arctic the latter. And indeed, the Antarctic is colder than the Arctic.