[Geysers] Conduit dynamics: bottom up or top down?

[Davis, Brian L.]

I've been playing a lot more with model geysers (HUGE thanks to several of you for your suggestions and observations; as an aside, my 2nd grader was able to demonstrate in her science fair project that the interval increases as you increase the total height - and she won first in her division (1st & 2nd grade) for that). I've been instrumenting them and found some interesting stuff, but after reading some of the peer-reviewed papers on this I've come across a glaring problem in my education:

Geyser erupt from the top down.

What I mean by that is that the temperature profile of Old Faithful in particular show it near boiling just prior to an eruption at the top of the conduit. Essentially the entire upper portion of the filled conduit is very near boiling, and when some of that is removed it reduces the pressure enough to allow more water to flash to steam, etc. What surprised me about this is that all the models I've built (as well as the few that seem to have been written about) all erupt from the bottom up: that is, boiling starts in the lowest portion of the system, where expansion can push some water out of the conduit and this then allows the pressure to drop and the system to flash over. In fact, I'm not sure I can make a model that does what Old Faithful does - it's hard to get vigorous enough convection in the system I think (but I don't know; I need to instrument better).

So I guess my first question is can geysers be separated into two classes, those like Old Faithful that exceed the boiling curve at or very near the top of the conduit, and those that first exceed the boiling curve at or near the base of the conduit? And what if any difference in behavior would demonstrate this? There is a lot of interesting data on Old Faithful (of course), but it seems there is very little data even for this easily accessible geyser with respect to conditions in the conduit during the interval and even less known about what's happening during the actual eruption (even seismic data is tough here, as it seems the fluid within the conduit "decouples" from the surrounding rock, generating very little seismic information).

On a separate topic, the behavior of a model geyser with two similar chambers (both 500 ml) in parallel at the bottom is surprising (at least to me). I expected semi-independant eruptions, where each chamber can set off its own eruption via the shared vent, but if both chambers were close to the boiling curve then if one went, the reduction in pressure would prematurely trigger the sibling chamber, resulting in a bigger eruption. It seems this is rather difficult to accomplish. Instead what seems to happen very frequently is for equal or nearly equal sized chambers, one will start to boil vigorously (like it was going to erupt), and just as it starts (swells of water come out of the vent), it "quenches" and the other chamber now suddenly starts to boil, until the process switches sides yet again. It looks like the two chambers exchange roles repeatedly, stalling or sabotaging an eruption, until something "locks" the chambers together and they erupt simultaneously (I hate using the word "lock" here, as I'm *not* referring to the lock that happens before a natural geyser erupts in some cases... but that's really what it "looks" like in the model). Any ideas? I need to extend the time series out to see if these dual-chamber models are as odd as they seem. The CPVC+flask systems seem to run fine for up to 6 hours, but the only time we were repeatedly doing that long a run was to get the temperature stabilized for the science fair experiments. If I run the same dual-chamber model with a 500 ml flask and a 1000 ml flask, the 500 ml flask seems utterly incapable of starting an eruption - it essentially stalls in a vigorous boil until the 1000 ml gets close, and then both erupt with almost no preliminary boiling within the 1000 ml chamber - in other words, it acts as though the presence of the 1000 ml chamber stalls the eruptions of the 500 ml chamber, but once the 1000 ml chamber gets close, the fluctuations or minors from the 500 ml arm of the system can trigger the entire thing *very* quickly. So far, these have all involved roughly equal heating (meaning the 500 ml flask will warm much faster than the 1000 ml flask, as until I can get thermocouples into the flasks themselves it's very hard to determine relative heating rates.

-- 
Brian Davis