PESTPP-IES restart question

[desseydx]

Hi pestpp users,

I have a question regarding pestpp-ies restart.

My desktop got power off during ies running, so I need to restart it. IES has already improved from the initial sets of parameters, so I wanna continue from previous run.

I use the lines as below but with an error for the second line of argument:

++ ies_parameter_ensemble(case_ies.3.par.csv)
++ ies_restart_obs_ensemble(case_ies.3.obs.jcb)
++ ies_save_binary(true)

If I deleted the second line, IES ran okay but re-ran all realizations.
So, what's the right way to use ies_parameter_ensemble(), ies_restart_parameter_ensemble(), and ies_restart_obs_ensemble() please?

Many thanks for your kind answer in advance.

Best regards,

Lucy

[cnicol-gwlogic]

Hi there, I think you just need to include ++ies_restart_parameter_ensemble(<your latest and greatest .par.csv file after power outage>), eg:
++ies_parameter_ensemble(pest/cal08_cn_IES_v2.0.par.copy.csv)
++ies_restart_parameter_ensemble(pest/cal08_cn_ies_v2.1.par.copy.csv)
++ies_restart_observation_ensemble(pest/cal08_cn_ies_v2.1.obs.copy.csv)

Double check the manual, but I believe that ++ies_parameter_ensemble() simply maintains tracking of pars as they evolve from the initial par ensemble - so in my example above, I provide the .0.par.csv file from the initial run (i.e., the one from your power outage, but at iter0).

[desseydx]

Hi Chris, Many thanks for your kind and prompt reply. I've tried the line below (power off after iter 2 finished): ++ ies_parameter_ensemble(case_ies.2.par.csv) ++ ies_restart_parameter_ensemble(case_ies.2.par.csv) ++ ies_restart_obs_ensemble(case_ies.2.obs.jcb) ++ ies_save_binary(true) The ies run without error, however, all iterations re-numbered from zero (replaced the old iteration files). I have read the manual, but still a bit confused, not sure whether I've done it the right way. What iteration number should I use for ies_parameter_ensemble() and ies_restart_parameter_ensemble() so that the ies could continue from iter 3 without re-writing old iterations please? Many thanks for your kindness and patience answering my questions. Best regards, Lucy From: Chris Nicol ***@***.***> Sent: 6 January, 2025 12:15 PM To: usgs/pestpp ***@***.***> Cc: Du, Lucy (RTIO) ***@***.***>; Author ***@***.***> Subject: [External] Re: [usgs/pestpp] PESTPP-IES restart question (Discussion #326) You don't often get email from ***@***.******@***.***>. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification> Hi there, I think you just need to include ++ies_restart_parameter_ensemble(<your latest and greatest .par.csv file after power outage>), eg: ++ies_parameter_ensemble(pest/cal08_cn_IES_v2.0.par.copy.csv) ++ies_restart_parameter_ensemble(pest/cal08_cn_ies_v2.1.par.copy.csv) ++ies_restart_observation_ensemble(pest/cal08_cn_ies_v2.1.obs.copy.csv) Double check the manual, but I believe that ++ies_parameter_ensemble() simply maintains tracking of pars as they evolve from the initial par ensemble - so in my example above, I provide the .0.par.csv file from the initial run (i.e., the one from your power outage, but at iter0). - Reply to this email directly, view it on GitHub<#326 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BGAJQIOMVEEGVKXBXXO7QL32JH7KVAVCNFSM6AAAAABUU2OCGOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCNZUGQZDMMY>. You are receiving this because you authored the thread.Message ID: ***@***.******@***.***>>

[cnicol-gwlogic]

Try this instead:
++ ies_parameter_ensemble(case_ies.0.par.csv)
++ ies_restart_parameter_ensemble(case_ies.2.par.csv)
++ ies_restart_obs_ensemble(case_ies.2.obs.jcb)

[cnicol-gwlogic]

I forgot to mention that the iteration numbering always starts from 0 as far as I am aware. So the trick to avoid overwrites of your old files is to rename the restart pst file (eg youroldpstfile.restart.pst).

[desseydx]

Many thanks Chris for your kind and prompt reply! It's very helpful! Jeremy recommended to add another line "++ ies_observation_ensemble(noise.jcb)" (noise.jcb was a copy and renamed from the file pest.obs+noise.jcb) after the "++ ies_parameter_ensemble", followed by "++ ies_restart_observation_ensemble" and "++ ies_restart_parameter_ensemble" as you instructed. I've tested using a synthetic model and it works! Thank you sincerely for your kind help. Best regards, Lucy From: Chris Nicol ***@***.***> Sent: 8 January, 2025 5:22 AM To: usgs/pestpp ***@***.***> Cc: Du, Lucy (RTIO) ***@***.***>; Author ***@***.***> Subject: [External] Re: [usgs/pestpp] PESTPP-IES restart question (Discussion #326) You don't often get email from ***@***.******@***.***>. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification> I forgot to mention that the iteration numbering always starts from 0 as far as I am aware. So the trick to avoid overwrites of your old files is to rename the restart pst file (eg youroldpstfile.restart.pst). - Reply to this email directly, view it on GitHub<#326 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BGAJQIJ5KU4RQ6FOIALKY2T2JRAN3AVCNFSM6AAAAABUU2OCGOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCNZWGU4DONY>. You are receiving this because you authored the thread.Message ID: ***@***.******@***.***>>