Although the surface is cold, the base of an ice sheet is generally warmer due to geothermal heat.

[4] Increasing global air temperatures due to climate change take around 10,000 years to directly propagate through the ice before they influence bed temperatures, but may have an effect through increased surface melting, producing more supraglacial lakes.

A number of processes alter these two factors, resulting in cyclic surges of activity interspersed with longer periods of inactivity, on time scales ranging from hourly (i.e. tidal flows) to the centennial (Milankovich cycles).

[11] Increasing global air temperatures due to climate change take around 10,000 years to directly propagate through the ice before they influence bed temperatures, but may have an effect through increased surface melting, producing more supraglacial lakes.

In Antarctica, this is driven by heat fed to the shelf by the circumpolar deep water current, which is 3 °C above the ice's melting point.

[24][25][26][27] As a result, sea level rise from the ice sheet could be accelerated by tens of centimeters within the 21st century alone.

[26] Ultimately, even geologically rapid sea level rise would still most likely require several millennia for the entirety of these ice masses (WAIS and the subglacial basins) to be lost.

[39] Sea level rise projections which involve MICI are much larger than the others, particularly under high warming rate.

[36] It was originally proposed in order to describe how the large sea level rise during the Pliocene and the Last Interglacial could have occurred[36][37] - yet more recent research found that these sea level rise episodes can be explained without any ice cliff instability taking place.

)[43] In recent years, 2002-2004 fast retreat of Crane Glacier immediately after the collapse of the Larsen B ice shelf (before it reached a shallow fjord and stabilized) could have involved MICI, but there weren't enough observations to confirm or refute this theory.

[49][50][51][44] Some scientists - including the originators of the hypothesis, Robert DeConto and David Pollard - have suggested that the best way to resolve the question would be to precisely determine sea level rise during the Last Interglacial.

Further, marine ice sheet instability may increase this amount by tens of centimeters, particularly under high warming.

[62][63][64] In the long term, the West Antarctic Ice Sheet is likely to disappear due to the warming which has already occurred.

[65] Paleoclimate evidence suggests that this has already happened during the Eemian period, when the global temperatures were similar to the early 21st century.

[72] Isostatic rebound of ice-free land may also add around 1 m (3 ft 3 in) to the global sea levels over another 1,000 years.

Consequently, greenhouse gases actually trap heat in the middle atmosphere and reduce its flow towards the surface while the temperature inversion lasts.

[79] Due to these factors, East Antarctica had experienced slight cooling for decades while the rest of the world warmed as the result of climate change.

Clear warming over East Antarctica only started to occur since the year 2000, and was not conclusively detected until the 2020s.

[113] If the Paris Agreement goal of staying below 2 °C (3.6 °F) is achieved, melting of Greenland ice alone would still add around 6 cm (2+1⁄2 in) to global sea level rise by the end of the century.

[117]: 5 Historically, ice sheets were viewed as inert components of the carbon cycle and were largely disregarded in global models.

In 2010s, research had demonstrated the existence of uniquely adapted microbial communities, high rates of biogeochemical and physical weathering in ice sheets, and storage and cycling of organic carbon in excess of 100 billion tonnes.

[123] Also for comparison, the annual human caused carbon dioxide emissions amount to around 40 billion tonnes of CO2.

[125][126] Normally, the transitions between glacial and interglacial states are governed by Milankovitch cycles, which are patterns in insolation (the amount of sunlight reaching the Earth).

When these icebergs melted they dropped the boulders and other continental rocks they carried, leaving layers known as ice rafted debris.

During Dansgaard–Oeschger events, the northern hemisphere warmed considerably, dramatically increasing the release of methane from wetlands, that were otherwise tundra during glacial times.

Carbon dioxide decrease, with a tipping point of 600 ppm, was the primary agent forcing Antarctic glaciation.

[137] The glaciation was favored by an interval when the Earth's orbit favored cool summers but oxygen isotope ratio cycle marker changes were too large to be explained by Antarctic ice-sheet growth alone indicating an ice age of some size.

[138] The opening of the Drake Passage may have played a role as well[139] though models of the changes suggest declining CO2 levels to have been more important.

[140] While there is evidence of large glaciers in Greenland for most of the past 18 million years,[102] these ice bodies were probably similar to various smaller modern examples, such as Maniitsoq and Flade Isblink, which cover 76,000 and 100,000 square kilometres (29,000 and 39,000 sq mi) around the periphery.

Conditions in Greenland were not initially suitable for a single coherent ice sheet to develop, but this began to change around 10 million years ago, during the middle Miocene, when the two passive continental margins which now form the uplands of West and East Greenland experienced uplift, and ultimately formed the upper planation surface at a height of 2000 to 3000 meter above sea level.