template cross correlation does not get a correct redshift?

I was working on simulated NIRISS WFSS spectra and trying to get redshift via template cross correlation. I followed a notebook written by Ivo Busko with ground data (https://spacetelescope.github.io/jdat_notebooks/pages/redshift_crosscorr/redshift_crosscorr.html), but could not get the right redshift --- the velocity peak appears where the fit is clearly off from emission lines.

Maybe this is because of the resolution NIRISS has (R=150), or my lack of experience with specutils.
If the latter case, I still imagine that it would be very useful if there is resource more other than one demonstration.

I shared my notebook here, in case you need to take a look ;
https://stsci.box.com/s/5mm8sa6mgls5yhuj0pu703gslk4b3ita

Please contact me know if you have questions. Thanks in advance,
Takahiro

Begin with the linked redshift_crosscorr notebook and the shared reproduction notebook, then compare the velocity peak with the emission lines at NIRISS R=150. Determine whether the mismatch is reproducible in specutils and document the cause or missing guidance; done means the behavior is explained with an appropriate fix or clearer resource.

Written by the indexing model from the issue text.