[aspect-devel] Handing nodes problem

[John Naliboff]

Hi Lev,

> yes, it's always at the boundary between the refinement levels (the highest and next-to-highest ones I believe).

> Not sure what you mean by grid spacing... different refinement levels? Not sure what the results of such test will tell me, to be honest... Or do you mean running a model without AMR? 

If the anomaly is always at the boundary between cells of varying size, this may be part of the issue. In addition to Timo’s suggestion, I would indeed turn off AMR and see how the models behave with varying global refinement levels. Again, it could provide little insight into the issue but it can’t hurt to try.

Cheers,
John

*************************************************
John Naliboff
Assistant Project Scientist, CIG
Earth & Planetary Sciences Dept., UC Davis




> On May 19, 2017, at 12:33 PM, Lev Karatun <lev.karatun at gmail.com> wrote:
> 
> Hi John,
> 
> Do the velocity anomalies typically correlate with the local grid size? 
> yes, it's always at the boundary between the refinement levels (the highest and next-to-highest ones I believe). Or did you mean something else? 
>  
> In this case does “different parts” refer to the upper mantle only or the entire domain? 
> entire domain. Sometimes crust, sometimes sub-lithospheric mantle
> 
> A good diagnostic test here might be to try a series of models with constant grid spacing and see how this affects the location/timing/etc of the velocity anomaly. 
> Not sure what you mean by grid spacing... different refinement levels? Not sure what the results of such test will tell me, to be honest... Or do you mean running a model without AMR? 
> 
> (I also sent another letter at the same time you sent this one)
>  
> 
> Best regards,
> Lev Karatun.
> 
> 2017-05-19 15:26 GMT-04:00 John Naliboff <jbnaliboff at ucdavis.edu <mailto:jbnaliboff at ucdavis.edu>>:
> Hi Lev,
> 
> Do the velocity anomalies typically correlate with the local grid size? 
> 
>> This doesn't, however, seem to cause the velocity anomaly because I get it in different parts of the model every time, and the viscosity field is often smooth there with low values. 
> 
> In this case does “different parts” refer to the upper mantle only or the entire domain? 
> 
> A good diagnostic test here might be to try a series of models with constant grid spacing and see how this affects the location/timing/etc of the velocity anomaly. 
> 
> Cheers,
> John
> *************************************************
> John Naliboff
> Assistant Project Scientist, CIG
> Earth & Planetary Sciences Dept., UC Davis
> 
> 
> 
> 
> 
> 
>> On May 19, 2017, at 12:17 PM, Lev Karatun <lev.karatun at gmail.com <mailto:lev.karatun at gmail.com>> wrote:
>> 
>> Well, the viscosity is strongly strainrate-dependent in the model, that's why I get this effect I guess (not sure what to do about it, or if I should even do anything). This doesn't, however, seem to cause the velocity anomaly because I get it in different parts of the model every time, and the viscosity field is often smooth there with low values. 
>> 
>> Best regards,
>> Lev Karatun.
>> 
>> 2017-05-19 13:17 GMT-04:00 cedric thieulot <c.thieulot at uu.nl <mailto:c.thieulot at uu.nl>>:
>> Looking at the viscosity field, it kinda looks like suspicious in the darkest red regions, and reminds me of
>> simulations whose velocity field is not smooth and converged, thereby leading to anormal element-based strainrate patterns.
>> Ce/
>> 
>> 
>> _______________________________________________
>> Aspect-devel mailing list
>> Aspect-devel at geodynamics.org <mailto:Aspect-devel at geodynamics.org>
>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel <http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel>
>> _______________________________________________
>> Aspect-devel mailing list
>> Aspect-devel at geodynamics.org <mailto:Aspect-devel at geodynamics.org>
>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel <http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel>
> 
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org <mailto:Aspect-devel at geodynamics.org>
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel <http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel>
> 
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20170519/ae0d7b4e/attachment-0001.html>