[CIG-SHORT] long-term simulation of seismic cycles with rate and state friction law

[Brad Aagaard]

Shaoyang,

We have not yet implemented fully adaptive time stepping in PyLith, but 
it is on our TODO list. You can manually alter the time step by using 
the TimeStepUser. This does require knowing when you need the finer time 
steps. We don't have an example that illustrates the use of 
TimeStepUser, so see the PyLith manual for how to use user-specified 
time steps.

Regards,
Brad


On 2/18/13 9:39 AM, Shaoyang Li wrote:
> Dear it may concern/Aagaard,
>
> I am a PhD student in GFZ Postdam, Germany. Currently I am trying to use
> Pylith to investigate how rate and state friction properties affect long
> term earthquake cycles behaviours.
>
> I started from my own 3D strike slip geometry with two patches of VW
> inserted in VS fault (following Kaneko's 2010 Paper) . When I tuning
> parameters of RS law, it appeared some interesting results, seemed like
> slow slip events.
>
> But whatever parameters I choose, one "characteristic" earthquake
> happens in several years (say 3-4 years in 600 years' simulation). If
> the RS geometry is less complex, events happen in less time.
>
> Then I turn to the given example in RS friction law ( 3D/hex8/step14). I
> found the problem is similar with a little bit better case (event
> happens in 1 year something, refer to pdf attachments). And I try to
> decrease to very small time step (~day scale), the "instability"
> appear(attachment txt file)!
>
> Is it available to use different time steps in different seismic periods
> (say coseismic in seconds, interseismic in years)
>
> Such problem confused me a long time, any comments or suggestions are
> highly appreciated! Thanks a lot.
>
> Best wishes
> Shaoyang
>
>
>