Attempting to reproduce Reactive Crystallisation models from Collier and Kelemen

[Cliff_Shaw]

Hi,

I am trying to reproduce the reactive crystallisation models in Collier and Kelemen (2010) using the pMELTS version of alphaMELTS.  I have created a file with the initial melt composition this also contains the starting temperature 1310 C and end temperature 1190 C and the temperature incremenent (-5 C)and a second file that contains the compostion of the assimilant.

Since the model is supposed to be isobaric, I have set the mode to isobaric and have

ALPHAMELTS_ASSIMILATE   true
ALPHAMELTS_FRACTIONATE_SOLIDS   true

I have two problems:

First, how do I set the temperature of the assimilant?  I have tried setting it within the assimilant.melts file but I don't know if this correct.

Second:  when I run the model with my assimilant temperature set at 1240 (as specified in the C+K paper) I cannot reproduce their liquid lines of descent.

I would appreciate any help you can give me with this


Thanks

Cliff

[Paula]

Hi Cliff,

Since the model is supposed to be isobaric, I have set the mode to isobaric and have

ALPHAMELTS_ASSIMILATE   true
ALPHAMELTS_FRACTIONATE_SOLIDS true

Leave ALPHAMELTS_FRACTIONATE_SOLIDS off for now. The 'reactive crystallisation' calculations that Collier and Kelemen (2010) describe in the main text and plot in Fig. 2 are for batch crystallisation, plus assimilation. The input file in the supplementary materials is for one of the batch reactive crystallisation runs. The results plotted in the supplementary materials are for fractional crystallisation, plus assimilation. I was able to reproduce the reactive crystallisation results running the pMELTS model in the rhyolite-MELTS graphical user interface (GUI). Haven't check the fractional ones yet but I expect they would also be fine.

If you were doing isenthalpic calculations you would set the temperature of the assimilant by putting an "Initial Temperature:" line in the assimilant .melts file. For the calculation in Collier and Kelemen (2010) you can still include the "Initial Temperature:" line but it will be ignored. For the equivalent calculation in the GUI you need to set the assimilant temperature to a value greater than zero for assimilation to happen, but again the exact value is ignored.

In the GUI there is no assimilation before the first equilibration. To emulate this in alphaMELTS you can choose not to read in the assimilant file the first time you are asked. So a batch file for the calculation might look like this:

1
basalt.melts
4
1
-1
1
x
1
15.0
peridotite.melts
0

where 15.0 is the number of grams of material to assimilate at each step. Running this in alphaMELTS gives results that are similar but not the same as the Collier and Keleman or rhyolite-MELTS GUI results. The difference is the routines to detect phase saturation have been updated in rhyolite-MELTS (Ghiorso, 2013) whereas alphaMELTS uses the old routines (except alphaMELTS 2 but that doesn't have the assimilation features yet).

So alphamelts brings in the wrong phase(s) for a while. For the batch case it recovers reasonably well. For the fractional case I would not expect it to reproduce the results in the supplementary materials. There are workarounds but they are pretty involved...

Hope that helps,
Paula

[Cliff_Shaw]

Paula,

I have now been able to reproduce the data in the Collier and Kelemen paper using the Rhyolite_melts GUI in pmelts mode.  Thanks for all your help

Cliff