No snowball earths have occurred since the first appearance of bilaterian animals (above sponge-grade) in the fossil record ~555 Ma in Arctic Russia. Vascular land plants and terrestrial fauna did not evolve for another 100 plus million years. These species never had to survive a snowball earth. But a host of microscopic organisms, both prokaryotes (archea and bacteria, including prokaryotes (cyanobacteria) and eukaryotes (algae, testate amoebae and other protists), and a handfull of cm-scale organisms (the coiled Grypania, 1.9 Ga; the necklace-like colonial organism of tissue-grade organization Horodyskia, 1.5 Ga; the worm-like Parmia, 1.0 Ga), evolved before the Sturtian and Marinoan snowball earths and survived. Evolutionary rates appear to have been incredibly slow, however, compared with post-snowball times, and this has discouraged intensive investigation of the fossil record. Palynofloras were impoverished during the Cryogenian Period (encompassing the Sturtian and Marinoan snowball earths), relative to earlier and later periods, but the characteristic assemblage of simple thin-walled microspheres (leiospheres) passed through the Marinoan snowball earth without change according to Kath Grey, the leading specialist in the field, at the Geological Survey of Western Australia in Perth. Similar palynofloras were earlier reported from Sturtian glacial deposits in East Greenland and Utah (USA). Several important eukaryotic lineages from more ancient times, including red, brown and chromophyte algae, are still extant and must have survived. This implies that sunlight and liquid water coexisted somewhere on the snowball earths. The liquid water need not have been extensive or permanent, snow algae need only a surface film of water, and a dash of dust, when the snow touches the melting point in summer. But Kath Grey's data from Western Australia suggest the need for a marine refugium in the snowball earth. This is supported by organic molecular fossils (biomarkers) diagnostic of phototrophic bacteria and eukaryotes recently described by a group led by Alison Olcott of the University of Southern California (USA) from within glacial marine sediments in Brazil. A necessary qualifier is that most glacial sediments that survive were deposited during the glacial retreat, when open water was present irrespective of the earlier ice extent.

The survival of photoautotrophy has inspired investigation of sea-ice dynamics in a snowball earth utilizing computational climate models. With mean surface temperatures of -50°C (-74°F), floating ice thickens rapidly to ~1.0 km and flows under its own weight towards the equator, where it is thinnest and where ice inflow is balanced by melting and sublimation. Flowage of the floating "sea glaciers" causes cracking along ice-grounding lines (separating free-floating ice and grounded ice). The cracks are held open by the force of the water, allowing them to be healed by new sea ice, which is full of brine channels which drain away in a few years. Brine channels in present-day sea ice host a surprisingly rich biota including algae (mostly diatoms), protozoa and foraminifera. The biota must be tolerant of hyperoxia because photosynthetic O₂ cannot escape the brine channels. This might be an interesting pre-adaptation if atmospheric O₂ levels rose after the Marinoan snowball earth, which was a necessary prerequisite for the rise of large active animals.

Early snowball models with sea-ice dynamics yielded tropical ice far too thick (>100 m) for photosynthesis. However a recent model result by Dave Pollard and Jim Kasting at the Pennsylvania State University (USA) yields a 2000-km-wide equatorial zone where the ice is less than 2 m thick, alongside km-thick ice poleward of ~13° degrees latitude north and south. The steep transition zone closely follows the snow line on the moving ice. The result is stable only if the equatorial is clear (old) sea ice, which absorbs Solar radiation better than bubbly ice formed from compacted snow. A healthy rate of photosynthesis is possible under 2 m of clear ice (the shut-off is ~20 m). In addition, 2-m-thick ice (similar to present annual sea ice) would have abundant cracks and leads, where marine photosynthesis would flourish given dust (nutrient iron, phosphorus and trace elements) and fixed nitrogen. Snowball climate models indicate that the subtropics (below the descending arms of the atmospheric Hadley cells) would experience net sublimation, meaning that subtropical land areas remote from mountains would remain ice free and generate dust. Volcanic ash and extraterrestrial material would contribute regionally and globally, respectively, to the dust load.

Climate model results have been reported by Dick Peltier at the University of Toronto and associates in which a megacontinent is glaciated from pole to equator while large areas of the ocean remain as ice-free "oases". This model is minimally challenging to life but less satisfactory in accounting for the geochemical anomalies associated with the glacial deposits, or with the apparent duration of the glacial episodes (read: What is the evidence for the snowball earths?). The paleogeography used, dominated by a polar supercontinent, is completely different from that implied by paleomagnetic data (read: What caused the snowball earths?), making it difficult to evaluate the model results. Both the oasis and thin-ice ice soloutions lapse to thick tropical ice (aka "hard" snowball earth) with modest lowering of radiative forcing. It is disquieting to believe that the survival our eukaryotic ancestors depended on the narrow-escapes offered by these metastable states.