The only time you need pumps is if the aquifer is immediately below the surface, and I think that's only even possible in swamps/marshes.
untrue, there only needs to be more aquifer levels than dry levels available above.
Even if you use concentric rings if cave-ins like I mentioned before? I'm still of the opinion that if 2z chicken runs are possible, you can just repeat that method to get through deeper aquifers, you just need to plan a larger footprint from the start, and that you only definitively need pumps if the aquifer is immediately below the surface.
edit: If the aquifer has an even number of layers, they need to be groupable as pairs of 2 soil or a stone+soil level pair. This is because for the purposes of a 2z chicken run, the critical part isn't whether it's slow (stone) or fast (soil) to dig through the top level of the 2z aquifer, the miner just has to dig into the bottom level quickly enough. If there's an odd number of levels, a 2nd dry soil layer above the aquifer is needed to be able to cave-in a plug on that, inside of which the later plugs will then be dug. This could also allow piercing 2 consecutive stone layers, if e.g. there's a stone-stone-soil aquifer. If the aquifer is only stone, then yes, you need as many dry layers as stone layers if 2z chicken runs are not possible through stone+stone even with legendary miners.
In general though, for thicker than 2z aquifers it really does depend on exactly how many there are, exactly which layers are stone, if any, and how much of a rush you're in/how big you want the pierce to be.
Concentric cave-ins can pierce through aquifers of any material and depth (up to some maximum defined by embark size) if you have 2z of dry soil above. The 2z chicken runs are a useful method since it gives the ability to pierce 2z (or 4, or 6) soil (or stone+soil pairs) aquifers with just 1 dry soil above, or if you have 2z of dry soil, you can wall yourself underground immediately. Also, the 2z chicken run method doesn't need any wood for pumps or really anything than a pick and a miner. Again, depending on the exact setup of the aquifers, I'd use one or combine some of the established cave-in methods, and it's always possible to develop new ones too.