time or timestep in GALPROP
Page 1 of 1

Author:  fulai [ Wed Nov 02, 2011 12:43 pm ]
Post subject:  time or timestep in GALPROP

I have a very basic question for GALPROP:
Does GALPROP solve the time evolution of cosmic ray momentum spectrum in the Galaxy? or just to solve for a steady state solution?

If the former is true, an important parameter should be the time. With source and loss terms, the cosmic ray (say electrons) spectrum should evolve and may be different at different times. But I could not find this parameter in the code. For example, I don't know when is the time at the final data output. I don't quite understand the parameters start_timestep and end_timestep. Are they the initial and final times of the calculation or some kind of dt? Thanks very much for your help!

Author:  avladim [ Wed Nov 02, 2011 1:04 pm ]
Post subject:  Re: time or timestep in GALPROP

Fulai, thank you for the question!

In general, GALPROP solves the time-dependent CR propagation equation. However, for most CR studies, the Galaxy is assumed to be in a steady state. In order to derive a steady state solution, GALPROP can be configured to solve the CR transport equation for a long enough period of time. With constant sources and constant losses, the solution of the CR propagation equation asymptotically approaches the steady state as time approaches infinity.

There are two methods of achieving the steady state solution:

1) The constant small time step ("CSS") solution, which makes time steps shorter than the electron energy loss time (~100 years), but makes enough of them to span the highest energy proton diffusion time (~10^9 years).

2) The "accelerated" solution, which starts with a very large time step (~10^9 years) and then gradually reduces the time step down to the smallest time scales (~10^2).

Andy Strong has made a comparison of the two methods in the Explanatory Supplement, Section 11.1

In order to implement the CSS solution, set, for example,
(the values are just an example! the values you need to use depend on the problem you are solving).
This will make 1e7 time steps, 1e2 years each. timestep_factor may be any number less than 1.

In order to implement the accelerated solution (default in WebRun), set (for example)
This will start with the time step equal to 1e9 years, repeat it 100 times, then set reduce the time step by 2 (i.e., multiply by timestep_factor) and repeat it 100 times, reduce it by 2 again, etc., until it is down to the time step equal to 1e2.

Author:  fulai [ Wed Nov 02, 2011 1:28 pm ]
Post subject:  Re: time or timestep in GALPROP

Thanks for the detailed explanation!

Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group