Enable rmax persistent vs per isotach perturbation

[SorooshMani-NOAA]

The latest ensembleperturbation PR (https://github.com/noaa-ocs-modeling/EnsemblePerturbation/pull/154) will be merged soon, update the dependency spec to use the latest code and update the scripts to add the option for using persistent vs per isotach perturbation that is implemented by @WPringle

[SorooshMani-NOAA]

@WPringle a quick question. I am a little confused about what combinations are right. This is my understanding of the different combinations we want to try (question parts bold-italicized):

Rmax filling	Perturbation	Vortex Model
Persistent RMax (like before but just with hotstart, speed & quadrant fix)	Old perturbation (persistent)	Holland
Based on NHC paper	Old perturbation (persistent)	Holland
Not filling Rmax	New perturbation (GAHM param based)	GAHM
Use NHC adjusted track https://github.com/noaa-ocs-modeling/SurgeTeamCoordination/issues/411	New perturbation (GAHM param based)	Holland
Use NHC adjusted track https://github.com/noaa-ocs-modeling/SurgeTeamCoordination/issues/411	New perturbation (GAHM param based)	GAHM

The confusion comes from the fact that GAHM doesn't use RMax, but Holland does, so why does it matter to have RMax fcst if we're using GAHM? Or to use GAHM based perturbation if we're using Holland?

[WPringle]

@SorooshMani-NOAA For the second row when we fill Rmax using regression based on NHC paper we should use the new perturbation.

[WPringle]

@SorooshMani-NOAA Rmax regression forecast: Use radius_of_maximum_wind variable Rmax persistent forecast: Use radius_of_maximum_wind_persistent variable

[WPringle]

As for Holland or GAHM we can use either of them interchangeably. If the isotachs are perturbed then you can still use Holland it just won't have any effect. The additional test we want to do is if we do not do any isotach perturbation and run with GAHM and compare to GAHM with the isotach perturbation.

[WPringle]

So it's really as simple as: Rmax regression: Holland vs GAHM isotach perturb vs GAHM no isotach perturbation Rmax persistent: Holland vs GAHM isotach perturb vs GAHM no isotach perturbation

[WPringle]

Although I should note doing Rmax persistent with GAHM isotach peturb should not work well because the Rmax will not be very consistent with the isotachs

[SorooshMani-NOAA]

@FariborzDaneshvar-NOAA are we currently doing Holland or GAHM? Because right now we still don't have isotach perturb available in the workflow version you have. I'm working on it in this ticket, it'll be ready soon since the actual implementation is done and merged by @WPringle in ensembleperturbation

[WPringle]

@SorooshMani-NOAA I believe that the current version @FariborzDaneshvar-NOAA running did have an old way to do isotach perturbation which only perturbs each isotach by the same distance [n mi] as the Rmax perturbation. So it does something..

[SorooshMani-NOAA]

Yes, but this would add some uncertainty that we don't account for, right? That's what I mean, that it might not give us improved results with GAHM.

[WPringle]

@SorooshMani-NOAA Sure it might not which is what we do want to look at yes

[FariborzDaneshvar-NOAA]

@FariborzDaneshvar-NOAA are we currently doing Holland or GAHM?

@SorooshMani-NOAA I used GAHM.

[SorooshMani-NOAA]

OK, we need to use ondemand workflow v0.1.2 and higher with GAHM given the discussion with @WPringle. Make sure to use isotach based perturbation and not persistent (in the input file) for new runs with GAHM. For Holland we need to use persistent, but we also need to use the older stormevents (1.2.10) for Holland. Probably we need to hash this out in the ensemble meeting (if we can make it next week given HPC training!)

[WPringle]

You can use the isotach perturbation with Holland too, not only persistent. In fact, we can try just compare RMW forecast vs persistent using only Holland model to test.

[SorooshMani-NOAA]

@WPringle I think I'm still a bit confused on the combinations ... let's have a discussion sometime soon

[WPringle]

Sure I just sent invite if u r free at 2 pm CT

Get Outlook for iOShttps://aka.ms/o0ukef

---

From: Soroosh Mani @.> Sent: Thursday, August 15, 2024 1:29:45 PM To: oceanmodeling/ondemand-storm-workflow @.> Cc: Pringle, William @.>; Mention @.> Subject: Re: [oceanmodeling/ondemand-storm-workflow] Enable rmax persistent vs per isotach perturbation (Issue #60)

@WPringle I think I'm still a bit confused on the combinations .. . let's have a discussion sometime soon — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned. Message ID:  ZjQcmQRYFpfptBannerStart This Message Is From an External Sender This message came from outside your organization.

ZjQcmQRYFpfptBannerEnd

@WPringlehttps://urldefense.us/v3/__https://github.com/WPringle__;!!G_uCfscf7eWS!Zo2Ylnp6VlZskm14WRECHHLZQ0e1JYgBRWui6LxjzP8bo2CIW5_wmWMtJtlLSgGxHdPLoe1OosHrbDquPxK3MScW$ I think I'm still a bit confused on the combinations ... let's have a discussion sometime soon

— Reply to this email directly, view it on GitHubhttps://urldefense.us/v3/__https://github.com/oceanmodeling/ondemand-storm-workflow/issues/60*issuecomment-2291919871__;Iw!!G_uCfscf7eWS!Zo2Ylnp6VlZskm14WRECHHLZQ0e1JYgBRWui6LxjzP8bo2CIW5_wmWMtJtlLSgGxHdPLoe1OosHrbDquPyA4jP2B$, or unsubscribehttps://urldefense.us/v3/__https://github.com/notifications/unsubscribe-auth/AFBHFHXECDRUYFIVVEXMWJDZRTXRTAVCNFSM6AAAAABMHHCWDSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEOJRHEYTSOBXGE__;!!G_uCfscf7eWS!Zo2Ylnp6VlZskm14WRECHHLZQ0e1JYgBRWui6LxjzP8bo2CIW5_wmWMtJtlLSgGxHdPLoe1OosHrbDquPy2vfZh2$. You are receiving this because you were mentioned.Message ID: @.***>