[CIG-MC] a 2-line fix to viscosity scale that is relevant to the geoid calculations in CitcomS

[Shijie Zhong]

Dear CIG staff and CitcomS users,

I sent out an email to a number of people last November about a 2-line fix to Instructions.c to get the correct reference viscosity for the purpose of geoid/dymamic topography calculations for time-dependent mantle convection problems (i.e., with some Ra, ...). See the following email for explanation. 

I am sending you a new Instructions.c in the attachment. The fix was based on version 3.1.1, but the required change is only for 2 lines that you can find them at line 712 and are listed as below.

  input_float("reftemperature",&(E->data.ref_temperature),"2700.0",m);

 if (E->control.stokes!=1) {
    E->data.ref_viscosity = E->data.ref_temperature *
        E->data.density * E->data.grav_acc * E->data.therm_exp *
     E->data.radius_km * E->data.radius_km * E->data.radius_km * 1e9
    /(E->control.Atemp * E->data.therm_diff);
   }

I deleted the line for calculation of reference temperature from reference viscosity and Ra. 

Again, this is only relevant for computing geoid/dynamic topography from thermal convection models (with Ra,...). 

Perhaps, Eric or someone from CIG can help put this fix to CIG's CitcomS.

Shijie

Shijie Zhong, Professor
Department of Physics
University of Colorado at Boulder
Boulder, CO 80309
Tel: 303-735-5095; Fax: 303-492-7935
Web: http://anquetil.colorado.edu/szhong
________________________________________
From: Shijie Zhong
Sent: Saturday, November 10, 2012 12:52 PM
To: Xi.Liu at colorado.edu; sdk at vt.edu; maxwellr at gmail.com; thorstinski at gmail.com; clintc at hawaii.edu; tobias.hoeink at rice.edu; Allen.McNamara at asu.edu; tan2tan2 at gmail.com; gurnis at caltech.edu; louis.moresi at monash.edu; nzhang at brown.edu; leng at caltech.edu; mibillen at ucdavis.edu; rmoucha at gmail.com; slava at levee.wustl.edu
Subject: a question on geoid calculation in CitcomS

In CitcomS 3.1.1 and possibly other versions, for geoid calculation in time-dependent convection models (i.e., with Ra, and other parameters), it seems that the standard/default input file asks for a reference viscosity (e.g., 1e21 Pas), and then the code (in Instructions.c) would compute reference temperature drop, delta_T, across the mantle from Ra, the reference viscosity, and other parameters given in the input file. Geoid is then computed from these dimensional parameters.

It seems that there is another way to do this: specify a reference temperature drop, delta_T, (e.g., 2500 K), and determine reference viscosity from Ra, delta_T, and other parameters. Then compute the geoid.

Since we often do not use realistic Ra in our convection calculations, which of these two methods is more reasonable is debatable. Personally, I think that the second method is more reasonable, because of smaller uncertainty in delta_T, compared with mantle viscosity.

I send this email to point out that CitcomS at CIG website seems to have the first method of these two possible methods implemented. If you would like to implement the second method, it can be easily done in Instructions.c (just change 1 or 2 lines). Your feedback and discussion are welcome.

(This issue is also relevant to dynamic topography calculation from mantle convection models).

All the best,

Shijie



Shijie Zhong, Professor
Department of Physics
University of Colorado at Boulder
Boulder, CO 80309
Tel: 303-735-5095; Fax: 303-492-7935
Web: http://anquetil.colorado.edu/szhong
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Instructions.c
Type: application/octet-stream
Size: 56293 bytes
Desc: Instructions.c
URL: <http://geodynamics.org/pipermail/cig-mc/attachments/20130523/cb4067d6/attachment-0001.obj>