That might lead us back to wild extrapolation; in an actual pasture, trample, replenish, consumption all happen at the same time, partially deterministically and partially stochastically. The approach that got all those exaggerrated numbers worked pretty much like this - observe "replenishment rate", observe "consumption rate", combine them into some sort of metric. Unfortunately, the numbers it gave fail the field test - observed needed pasture sizes are a lot smaller practically everywhere outside of desert biomes.
A few numbers from field tests:
One horse can barely keep a 5x9 aboveground pasture mostly-to-entirely bare (biome with fairly lush vegetation, but i don't know the exact metric - something in the "dense" range), anything larger will keep untouched grass. A yak bull can get into "hungry" status on an 8x8 underground pasture, but will reach a viable equilibrium on a 9x9. Goats - ho hum. A 4x4 pasture was _not depleted_ by five goats (just over three tiles per animal), but it wasn't viable in the long run, because they started infighting. Two goats in a 2x3 seem able to survive so far. Those are very few actually quantified data, but they all point in the same direction - the minima for pasture size in the field seem to be about 1/3 of the values given on the wiki. Vegetation below standard ground level (both cave vegetation and aboveground plants on channelled soil) seems to grow at the same rate regardless of biome, so might be a good "standard" reference.
My numbers so far are:
yak - minimum 9x9, 80 tiles. Recommended 10x10, 100 tiles.
horse - minimum 5x9, 45 tiles. Recommended 7x8 or 8x8 (56 or 64)
goat - minimum? hard to measure, probably 1x3, 3 tiles. Recommended 2x2, or three to four in a 4x4 pasture.
The tentative rule-of-thumb for minimum pasture size (underground) would be 7000/grazer value for number of tiles, large (llama and bigger) grazers only. Going by the experience with the goats, bolstered somewhat by what i've seen of other small grazers like sheep and alpaca, pasture requirement is probably not directly (inversely) proportional to grazer value, since at the limit, it seems that grass is kept down less by consumption and more by trampling (when the animal going to a re-grown patch stomps over all the ground in between). The smaller grazers, which take longer to consume a patch of grass, seem to move about in search of fresh grass much less.