Steinberger material model with melt transport

[vivianotieno]

Hello Aspect community
How do I add melt transport with the Steinberger material model? I want to get statistics on melt depth. Thank you.

[bobmyhill]

Dear Vivian,

This is a variation on a question we are often asked.

Short answer: The Steinberger material model does not include a parameterisation of melt transport. Implementation would require a significant change to the code, not just the prm file. Any implementation would need to carefully consider the chemistry and physics of interest.

Longer answer: See previous posts on this topic, including the three links below (which you can find by typing “melt” into the search button in the top right of the forum page). The first post from Juliane is particularly relevant. I think her question is also useful here:

What is the purpose of your model, and what specific components/physical processes do you need for that? For example, is it important to see where the melt migrates to, or only where it is being generated?"

If we have more details on exactly what you want to do, we can suggest (a) existing material models, or (b) the easiest way to implement what you want.

Best wishes,
Bob

Melt Production/ Migration

Hi Carlos, great to hear that you want to use ASPECT for modeling melting and melt migration! Because this is such a complex problem, it’s not that simple to answer your question. So let me address your questions separately: (1) Melting model. The melting model depends on the material model that you use. There is one that uses the Katz et al., 2003, and Sobolev et al., 2011 melting models (this is the latent heat melt material model), but this model does not include melt migration. There is …

Melting with non-linear rheology

Hi Lars, this depends a bit on how realistic you want your model to be. We’ve made a model with melt transport and a simple stress-limiter rheology for a manuscript that is currently in review, you can find the manuscript here: https://arxiv.org/abs/1810.10105 (the model is described in Section 3.3) and a description of how to reproduce the models here: https://github.com/tjhei/paper-aspect-melt-paper-2-data However, this does not include hydrofracturing, so there is no influence of the flui…

Melt with the Ascii reference profile materila model in Burnman cookbook

Hi Viv, Sorry for the delayed reply. Thanks for including the information about your motivation. I recommend that you read the documentation I linked to in my previous email, and figure out the things you need to change to solve your problem. You can then try to implement those changes, or describe to us where you need help. Generally, we’re happy to help out with prm changes, explaining code functionality, or helping you make specific changes to the code. Turning a scientific idea into a con…

[vivianotieno]

Dear Bob

Thanks for the response. However, I had already gone through the posts in the forum before posting. I need to get melt depth and volume which I can use in an outgassing model. I am working with self-consistent material properties (in ascii data model), that’s why I couldn’t use the melt-global prm. But so far from the post, I can see this is not possible as of now.

[bobmyhill]

Dear Vivian,

Yes, you are correct that this is not yet implemented. We update the manual regularly, so if you don’t see something there, chances are it hasn’t been done yet.

I need to get melt depth and volume which I can use in an outgassing model

Thanks for the additional information. Given that melt and solid move and change composition and abundance through time, there is still ambiguity here. Which physical processes are important to you? Two phase flow? Depletion/reaction through the melting column? Source heterogeneity? Ideally, it would be useful if you could show us the equations you want to solve, or at least a conceptual framework of the processes that can happen in your dream model.

There are three ways forward that immediately spring to mind, depending on the physics and chemistry you wish to model.

Do you care about thermodynamic self consistency? It’s always nice to have, but not particularly important for many problems. If you don’t need it, maybe you can use one of ASPECT’s existing models.

Do you care about two phase flow, the rheological influence of melt on the flow field, or chemical heterogeneity? If not, you could postprocess the results of standard solid state convection simulations to obtain an estimate of melt productivity as a function of depth (using P-T paths and optionally some parameterization of depletion). You could either implement that postprocessor in ASPECT or write something external to ASPECT.

Need any two or more of self consistency, two phase flow and heterogeneity/depletion/enrichment? In that case, a material model will need to be modified.

This is not an exhaustive list, nor necessarily does it include the best option. But I hope it goes some way to explaining why we ask what specific components/physical processes you need for your project.

Best wishes,
Bob

[vivianotieno]

Dear Bob

Thanks for the feedback. I’m working on the long-term evolution of planets. So my dream model would allow to track phase flow, depletion and enrichment of the mantle. To help quantify materials that leave the mantle to the surface (for volatile extraction). And also include the materials that are recycled back in the long term (not that important as of now, since I’m working with the stagnant lid regime). Self-consistency of the thermodynamic parameters is essential for my case. My focus is on thermal evolution and the mantle viscous state.
What kind of postprocessor would you recommend? I’m currently compositing the Steinberg material model, with viscosity profile from the ascii material input.

@bobmyhill

Thanks for the additional information. Given that melt and solid move and change composition and abundance through time, there is still ambiguity here. Which physical processes are important to you? Two phase flow? Depletion/reaction through the melting column? Source heterogeneity? Ideally, it would be useful if you could show us the equations you want to solve, or at least a conceptual framework of the processes that can happen in your dream model.

There are three ways forward that immediately spring to mind, depending on the physics and chemistry you wish to model.

Do you care about thermodynamic self consistency? It’s always nice to have, but not particularly important for many problems. If you don’t need it, maybe you can use one of ASPECT’s existing models.

Do you care about two phase flow, the rheological influence of melt on the flow field, or chemical heterogeneity? If not, you could postprocess the results of standard solid state convection simulations to obtain an estimate of melt productivity as a function of depth (using P-T paths and optionally some parameterization of depletion). You could either implement that postprocessor in ASPECT or write something external to ASPECT.

Need any two or more of self consistency, two phase flow and heterogeneity/depletion/enrichment? In that case, a material model will need to be modified.

This is not an exhaustive list, nor necessarily does it include the best option. But I hope it goes some way to explaining why we ask what specific components/physical processes you need for your project.

Best wishes,
Bob

[bobmyhill]

Dear @vivianotieno,

So my dream model would allow to track phase flow, depletion and enrichment of the mantle. To help quantify materials that leave the mantle to the surface (for volatile extraction). And also include the materials that are recycled back in the long term (not that important as of now, since I’m working with the stagnant lid regime).

Dynamic simulations including self-consistent thermodynamics (material properties and reactions) with two-phase flow and melt-induced fracturing together encompass several of the geodynamic “grand challenges” of this generation. It is an extremely exciting field to work in right now, and incremental steps towards these goals make for excellent papers in their own right.

The corollary is that there is no geodynamics code that yet does all these things. We are very happy to help people develop ASPECT so that it will be able to do all these things (several of us are working toward the same goal). This will take several years, but the problem can be broken into several highly-publishable parts.

What kind of postprocessor would you recommend? I’m currently compositing the Steinberg material model, with viscosity profile from the ascii material input.

You would have to write your own, but again, we can help if you have a plan. My initial (crude) thought was to create a postprocessor to calculate the theoretical productivity of fertile mantle based on a simple parameterisation of melting (something like https://academic.oup.com/petrology/article-abstract/29/3/625/1432023, or a model derived from a recent thermodynamic dataset), but you would have to decide whether such an approach would be useful to you, given the approximations involved.

If you don’t have any immediate ideas, it might be useful to find an existing paper that does close to what you want, look at the existing ASPECT manual (and code) to work out what is missing, and then implement the missing functionality or ask us where it would be sensible to begin.

Best wishes,
Bob