Even if towers would be created, we see in the man-made towers (and in many natural examples) that a wide basal "apron" is a major part of the structure. Especially once it falls down. VMS deposits (those formed by hot springs on the seafloor) are orginally lenticular in shape, but may contain towers within that structure. The difference between the man-made ice towers (MMIT?) and a VMS deposit is that the MMIT is a one time thing: it builds, falls, melts. In a VMS system, a tower grows, collapses, and another is formed on top of the deposits of the earlier vent tower. Towers, or chimneys, are only a minor part or the deposit. I've only seen towers preserved in an ancient VMS deposit in one locality (in the Jerome open pit mine, Arizona). I suspect that our hypothetical mars thermal vents would operate in the same way, building deposits on top of deposits. Over-all they might be lense shaped, in detail one might find columns, or the debris of columns. The deposits might be deformed by constant or periodic wind. The ice deposit might sublimate (as you suggest), thus adding an erosional component to the equation.

Speaking of sublimation, some of the structures that result from sublimating ice can be really weird, especially when darker dust is on the surface of the ice. I forget the term...but lots of interesting tepee structures can form.

I've scanned the article twice, and it doesn't even mention the fumarole. The article concerns the creation of high relief in Transantarctic Mountains by the interplay of isostatic rebound, erosion, and localized erosion related to elevation. The authors probably saw the towers on a traverse, and submitted it to GSA as a cover. I'd like to know if the tower sat upon a basal apron like a MMIT or was a series of stacked aprons and debris.

I'm glad Nick brought this up, I've been musing on it for awhile.