How to simulate isentropic decompression melting using pMELTS in Ubuntu system？

[Gary]

I downloaded the latest version of rhyolite-MELTS/pMELTS for Ubuntu to simulate the isentropic decompression melting of mantle. Firstly, I set the compositions and choose the isentropic and fractionate liquid modes. Then, I set the start temperature and pressure at 1380 ℃ and 2.0 Gpa, respectively. But when I clicked the execute. I could not obtain the result of continuous melting of the mantle. There must be something wrong. Can anyone tell me the correct operating steps? Thank you in advance!

[Paula]

Hi Gary,

Did you download the latest version from Mark Ghiorso's site? I would be surprised if they work, unless your Ubuntu is quite old. You should have more luck with the version on the alphaMELTS links page, but even those are a few years old. I've transferred my support to easyMelts instead, as it has all the features of the GUI, plus working plots. There's more information here: https://magmasource.caltech.edu/alphamelts/links.php

That said, there are some things that apply to all versions of MELTS. For a start, MnO and P2O5 are both not calibrated in pMELTS so you should remove those. K2O is better treated as a trace species for mantle melting calculations. There are few solid phases except feldspar for it to go in to, which makes it difficult to approach the solidus. You still may not be able to get right down to the solidus, even without K2O in the system - sometimes seeking the effective solidus where there is a little melt but barely enough for it to be extracted can be more practical.

I can't really tell much from the screenshot because it's from before you specify the P-T path. You need at least a stopping pressure and a pressure increment. I don't think "Compute" does much, unless you have non-zero P and T increments and then it will calculate dP/dT. In the past, I've found the T, P, H, S, V pop up and menu options to be quite sensitive to the order in which you do things (though to be fair, I haven't used the GUI much, so it could be user error). If you try easyMelts, you will need a pressure increment and a number of steps to take, rather than specifying the stopping pressure directly. There isn't a way to input the entropy increment in easyMelts (except via a .melts file) but you don't need that anyway.

Once you have closed the T, P, H, S, V pop up and select "Execute" what happens? Does it calculate the equilibrium for 1380^oC, 2000 bars and then stop? Or does it fail on this first step? If the latter, what are the error messages? Neither the GUI or easyMelts are especially geared towards mantle melting. It doesn't necessarily mean something is wrong, but you might need to use a workaround (or try alphaMELTS).

If it's failing in the initial subsolidus step then you might need to approach from higher temperature or lower pressure first (without "fractionate liquids", and isentropic mode). Then once you have the starting conditions you want, turn on "fractionate liquids" and isentropic mode, reset the pressure increment and stopping value and do Execute again.

Paula

[Gary]

  Dear Paula, thanks for your prompt reply. I downloaded the rhyolite-MELTS/pMELTS via the link https://melts.ofm-research.org/unix.html and run it with my Ubuntu22. I have set a stopping pressure and a pressure increment as you said. It did work.
 
  I found there was some difference between choosing or not choosing the "fractionate liquids". Could you tell me which mode is more appropriate for partial melting of mantle? Before I asked for your help, I have tried to use easyMelts to simulate the partial melting, but it is hard to work and always malfunctions: https://www.youtube.com/watch?v=2lF_0zU7jsw. If you are convenient, please help me to check out where is going wrong.

[Paula]

I downloaded the rhyolite-MELTS/pMELTS via the link https://melts.ofm-research.org/unix.html and run it with my Ubuntu22. I have set a stopping pressure and a pressure increment as you said. It did work.

I checked and the version of rhyolite-MELTS/pMELTS at the link you posted doesn't work on Ubuntu 22.04. Are you actually on Windows (e.g. did you follow the guide for setting it up on the WSL and download the version from the alphaMELTS links page)? If so, that's a bit better, and explains why it worked. But it's still unsupported. I recently found a major bug in the subsolidus fO2 routines of all MELTS software. I'll be posting fixes for the newer alphaMELTS software and easyMelts when I get a chance. But it has become too difficult to build the old graphical user interface (GUI) on Linux, and it doesn't even run on the newest Macs. So I won't be updating the GUI (or alphaMELTS 1.X).

I found there was some difference between choosing or not choosing the "fractionate liquids". Could you tell me which mode is more appropriate for partial melting of mantle?

You mentioned that you want to do continuous melting, so "fractionate liquids" will be closest to that. In fact a tiny bit of liquid is retained during fractionation. This is mainly to stabilize the calculation, but you can think of it as the residual porosity. You don't get as much control over the amount retained in the GUI or easyMelts, as you do in alphaMELTS, but the default value should be alright.

The melts.out file gives you the accumulated fractional melts, though not in an easily usable way. The increments are simply summed, so whether it is appropriate depends on the melting regime you are trying to model. If it's a 1-D column, like a plume, then it's fine. But if you are modeling a 2-D triangular melting regime at a mid-ocean ridge then you will need to weight the increments from different depths (from the melts-liquid.tbl file) and sum them appropriately. Paul Asimow's paper "Calculation of Peridotite Partial Melting from Thermodynamic Models of Minerals and Melts, IV. Adiabatic Decompression and the Composition and Mean Properties of Mid-ocean Ridge Basalts", and references therein, discuss this in more detail.

Before I asked for your help, I have tried to use easyMelts to simulate the partial melting, but it is hard to work and always malfunctions: https://www.youtube.com/watch?v=2lF_0zU7jsw. If you are convenient, please help me to check out where is going wrong.

Does it crash and close the window, or is it hanging during the calculation? If it is crashing then make sure you have the latest version - there was a bug that meant that it crashed if there was no liquid. If it's hanging then try what I suggested in the previous post i.e. start at a higher temperature or lower pressure first (without "fractionate liquids", and isentropic mode). Then once you have the starting conditions you want, turn on "fractionate liquids" and isentropic mode, reset the pressure increment and number of steps to take.

It would be a good idea to turn off the fO2 buffer before you hit the solidus. Even without the bug in subsolidus fO2 that I mentioned above, it is difficult to cross the solidus with the fO2 buffer on, and it makes the isentropic calculations more difficult (and slower) to have the buffer on at the same time. In practice, for mantle melting once the iron is distributed between Fe2+ and Fe3+ in the starting composition it doesn't deviate from the buffered value during decompression too much.

Paula

[Gary]

Hi, Paula,

Thank you very much for your detailed answer! Followed your steps, my easyMELTS runs well. I will use easyMELTS next to simulate mantle partial melting because it is more convenient to operate. If I meet other difficulties, I will ask for your help.

I know you have some question about why the rhyolite-MELTS/pMELTS can run on my Ubuntu. I'm sorry I can't answer you well because I'm not familiar with these operations. I recorded my operating steps, maybe it will be helpful: https://youtu.be/_NW-fTJHNHA. If you have any questions about this, just tell me.

All the best
Gary

[Paula]

Thanks for the video! I'm still not quite sure why it's working but I'm less surprised that it might work on Ubuntu 20.04 than 22.04, especially if that was upgraded from an earlier Ubuntu. Anyway I'm glad you have easyMelts working now - it's definitely the more convenient option these days.

Best wishes,
Paula