smoothing parameter

[Zhrngrui]

Hi , Prof. Johnson
I
How to get "norm of smoothing vector" and "norm of residual" for Smooth Trade off curve as describe in Johnson et al., 2024,GRL?
when i success run the step of "invert_strainrate_for_backslip", how to get "norm of smoothing vector" and "norm of residual" from the results?

Looking forward to your reply.
Best wishes.
 Rui Zheng

[ncperdomo]

Hi, sorry for the late reply. I don't check this repository very often.
I recommend computing the following metrics within a for-loop:

%%%%%%%%%%%%%%%%%%% L-curve metrics %%%%%%%%%%%%%%%%%%%%%%
misfit2_weighted = sum(((d - dhat)./Esig).^2); % ||(d-Gm)./Esig||^2
backslip_seminorm2 = norm(pm_smoothm_par,2)^2 + (variable_rake * norm(pm_smoothm_perp,2)^2); % ||D m_bs||^2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Best,
Nicolas

[Zhrngrui]

Thank you for your reply; it was very helpful to me. 郑瑞 ***@***.***

…

---- Replied Message ---- From Nicolás ***@***.***> Date 3/19/2026 22:39 To ***@***.***> Cc ***@***.***> , ***@***.***> Subject Re: [kajjohns/SliDeFS] smoothing parameter (Issue #5) ncperdomo left a comment (kajjohns/SliDeFS#5) Hi, sorry for the late reply. I don't check this repository very often. I recommend computing the following metrics within a for-loop: %%%%%%%%%%%%%%%%%%% L-curve metrics %%%%%%%%%%%%%%%%%%%%%% misfit2_weighted = sum(((d - dhat)./Esig).^2); % ||(d-Gm)./Esig||^2 backslip_seminorm2 = norm(pm_smoothm_par,2)^2 + (variable_rake * norm(pm_smoothm_perp,2)^2); % ||D m_bs||^2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Best, Nicolas — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[git-zyhh]

Hi, sorry for the late reply. I don't check this repository very often. I recommend computing the following metrics within a for-loop:

%%%%%%%%%%%%%%%%%%% L-curve metrics %%%%%%%%%%%%%%%%%%%%%% misfit2_weighted = sum(((d - dhat)./Esig).^2); % ||(d-Gm)./Esig||^2 backslip_seminorm2 = norm(pm_smooth_m_par,2)^2 + (variable_rake * norm(pm_smooth_m_perp,2)^2); % ||D m_bs||^2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Best, Nicolas

I suggest that "variable_rake" (either 0 or 1) as a multiplication factor of vertical(strike-perpendicular) differences is not necessary in "backslip_seminorm2 = norm(pm_smooth_m_par,2)^2 + (variable_rake * norm(pm_smooth_m_perp,2)^2); % ||D m_bs||^2 ", because vertical differences always make sence no matter the rakes varies or not.

Best,
zhao