PRECIS support forum
http://forum.precisrcm.com/

Is it incorrect to run a simulation in parts
http://forum.precisrcm.com/viewtopic.php?f=1&t=46
Page 1 of 1

Author:  wscott [ Wed Sep 10, 2014 7:55 pm ]
Post subject:  Is it incorrect to run a simulation in parts

Good day,

Is it alright to run a 40 yr simulation in parts instead of as a whole? For example, a simulation of the climate from 2030-2070 will consist of 4 10 yr experiments: 2030-40, 2040-50, etc. where each following experiment will be started from the completed time of the last experiment using rerun. Given that at least 12 months is needed for spin up, only the data from 2031 and beyond will be used.

Wscott

Author:  dmhg [ Mon Sep 22, 2014 12:22 pm ]
Post subject:  Re: Is it incorrect to run a simulation in parts

Hello,

Generally, we recommend that climate runs be 31 years or more. One year for model spin-up in which the GCM and RCM come into equilibrium (and which is not included in the analysis because of this), and three decades of data for analysis. The reason for this is that it gives the model a chance to have replicated internal natural climate variability, some of which occurs multi-decadally.

However in a pinch, because of time or lack of computing resources, 11 years can be used -- eleven is the *absolute* mininum (one year of spinup and one decade). So you could run e.g.

Dec 2029-Nov 2040 excluding Dec 2029 - Nov 2030 (as spinup)
Dec 2039-Nov 2050 excluding Dec 2039 - Nov 2040
Dec 2049-Nov 2060 excluding Dec 2049 - Nov 2050
Dec 2059-Nov 2070 excluding Dec 2059 - Nov 2060

That would leave you with 40 years of climate data for analysis -- December 1 2030 to Nov 30 2070

Thanks
David

Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/