PIONLT/KAONLT MEETING - 7 AUGUST 2025 PARTICIPANTS: Richard, Nacer, Alicia, Ivan, Pete, Tanja Notes by Tanja Nacer - iterations of KaonLT low Q2 cross section * tested with +-5% and +-10% variation limits on all parameters ::* fits do not converge well * performed iteration with limited number of parameters and cross section fits, e.g., fit only sigT and do not consider the other three (sigL, sigLT, sigTT) ::* fit sigT only (all others have the original parameters and are not fitted) - fit looks good ::* fit both sigT and sigL with parameters varying freely - both give good ::* sigLT and sigTT have thus far not been fitted * Action items: ::* continue the fitting with sigLT and sigTT - may need to constrain sigT and sigL parameters in that ::* check Ivan - uncertainty study * study cut sensitivity due to shifting all cuts (PID cuts) for all t-bins ::* largest variation comes from RF ::* PID detectors variation seems small * Next: ::* assign systematic uncertainties, also including polarization uncertainty Richard - KaonLT background subtraction at Q2=4.4 * looked at different background fits: linear, 3rd order polynomial, and Chebyshev second order ::* linear and Chebyshev seem very consistent ::* 3rd order polynomial cuts into KLambda signal * keeping pion subtraction as is plus remaining backround fit vs empirical fit of all background --> no significant difference, pion background small * Next: fit background due to sigma0 Alicia - PLB comments * Overall the reviews seem positive, but additional clarifications are needed (see also Garth's comments from 8/3/2025 email below) * Discussion of the comments and Alicia's suggested approach to address them * It seems that many of the comments can be addressed by either making textual modifications or by responding directly to the referee * The main referee comments may be categorized: ::* Require clarification of the kinematics - how the Hall C data set is different from the CLAS12 data set ::* Including CLAS data ::* Model comparison * Discussion about revising the figures ::* may be worth doing, but need to determine what it really adds to the paper ::* if added value is minimal may reply to the referee instead of adding text/replacing figures in the manuscript * Discussion about adding tables ::* If there is no limit to add then proceed ::* If there is a limit reply to referee and point out page limit and that data are available [specify where] * Actions: ::* Add textual clarification on how to kinematics are assigned, e.g., data are taken at points in limited kinematic phase space and for each there are unique labels ---> nontrivial to merge these points with CLAS12 data - some points could be added after scaling in the relevant kinematic variables, but not straightforward and unclear if beneficial ::* try to make a plot(s) including the CLAS12 data, but taking into account that some (all?) points will need to be scaled ::* Calculate theoretical cross section at bin center and update plot(s) ::* Model comparison and factorization - may need to review language used, but in the end the observations suggest that there is insufficient evidence to claim factorization in this kinematic regime Junaid - PionLT iterations * Encountered issues similar to Nacer * Action: follow a method similar to Nacer to obtain the iterated parameters ================= Garth's email from 8/3/2025 To me, it seems that the referees share many similar concerns, which has primarily to do with how we make use of the CLAS data in the factorization study, and the theoretical models. At least this is better than having contradictory referee comments. The issue of course is: 1) can we deal with all of the issues raised within the PLB format and 2) will they be sufficient to satisfy the referees? Unfortunately, I cannot attend Thursday's discussion. In case it is useful here is my take on a few of the most important referee comments. One referee is indicated with > while the other is with <. Fig 4: > Figure 4 should be explained in a bit more detail, including in the > Fig. caption. E.g, are all of the black points placed at their exact > kinematics for each setting? < The paragraph "For each (Q2, xB) setting, .... from the central values." < could be reorganized exploiting completely the Fig. 4 and explaining perfectly < the kinematics of the data. Since both referees have essentially the same comment, we clearly need to add details on how we assign the (Q2,x,t) values for each data point. I expect this also needs to include how we treat the CLAS data, since the referees have pointed out the premise of our PLB is a factorization study, not just that we have new BSA data. Fig 6: > Still, I don't understand why there are no CLAS12 points included in most of > the panels except the last one. From Fig. 4, it appears that most of the > kinematic "points" (except the highest-Q2 one - top-right panel) have CLAS12 > data reasonably close by, > I agree that the GK (1 or 2) model does not describe most of the new data > very well. On the other hand, the CLAS12 data are stated to be in good > agreement with the new data (again, show more of them in Fig. 6 to make this > explicit), Looking at the numerical values of the CLAS BSA data, there is not a lot of their data we could add directly to our plots: PANEL a) the closest CLAS points are at Q2=3.33 x=0.477 and Q2=3.44 x=0.327. Maybe we could interpolate between these to the panel kinematics of Q2=3.0 x=0.40, but that seems not so straightforward and likely with significant uncertainties. PANEL b) the closest CLAS point is at Q2=4.26 x=0.410, which is not far from Q2=4.4 x=0.40 so we probably could scale it to the panel kinematics somehow PANEL c) the closest CLAS point is at Q2=5.06 x=0.486, which seems not so close to Q2=4.4 x=0.40 and there are no other points to allow an interpolation. PANEL d) nothing really appropriate, CLAS data are either Q2<1.87 or Q2>2.78 In summary, only 1 point seems likely to be added to b). Since the referee was expecting much more, this indicates the type of info we likely would need to add to the supplemental info. Following the second referee comment above, an alternative is to add 2 new panels to Fig 6 containing CLAS data at Q2=2.6 x=0.4, and Q2=1.88 x=0.26. This would slightly extend the Q2-range of the t-dependence at fixed x study. Maybe that is more useful. More on Fig 6: > for a direct comparison with theoretical models, it would be better if there > were some kind of bin-centering correction for both data sets. < For the comparison to the models a priori a complete integration of the < theoretical values inside an experimental bin should be done. We could < suppose that the steep rise of the sigmaLT’/sigma0 close to tmin with < the GK model should be mitigated and closer to the data. < What is the impact of the variation of tmin? Both referee comments can be addressed by evaluating the models at exactly the same kinematics as the data. YCK did not do this for us when we repeatedly asked, but now that we have a copy of their executable, I believe it should be possible. This would eliminate the weird tmin issues in our plots. I don't think integration over the experimental bins is necessary since, as the referee indicated, our binning is finer than CLAS. This point could be added explicitly to the text. Fig 7: > Fig. 7 also contains fewer data than I would have expected. For instance, > why is there no data point (from either KaonLT or CLAS12) around Q^2 = 4.4 in > the 2nd panel, when there are clearly data in the 2nd top panel of Fig. 6 and > CLAS12 data at x=0.41 and Q2 = 4.2? Of course, all of (Q2,x,t) or (Q2,W,t) are needed to fully specify the reaction kinematics. It seems we should make this point explicitly in the text. The issue is that the Q2-dependences in Fig 7 incorporate cuts on x and t, while Fig 4 shows the points are available for fixed x vs Q2, irrespective of t. Commenting panel-by-panel: PANEL a) The first vertical band of Fig 4 actually corresponds to 11 data points, though only 5 are plotted. This is because as there are multiple points at different t for some plotting symbols. The -t=0.11 +/- 0.004 cut was chosen to retain the maximum number of data points. 4 points pass the cut and are shown in panel a of Fig 7. PANEL b) The second vertical band of Fig 4 has a lot of plotting symbols in it, but these actually correspond to only 8 data points, since there are fewer overlapping plotting symbols. The -t=0.365 +/- 0.016 cut selects the maximum number of points, which unfortunately turns out to be only 3 in panel b. In summary, Alicia and I can discuss whether we can reliably extrapolate some of the nearer data to add more points to these panels. However, the CLAS data are fairly coarsely binned, so most of their data are either at dramatically higher or lower -t. < A table of the results of sigmaLT'/sigma0 for the 5 settings should be < given in the publication as you do not have space limitation. This might help address the above issue on exactly which data points are included in the Q2=scan and how we use them. This is one of the novel points of our paper and we should try to strengthen it and clarify it as much as we can. In summary, it seems that addressing the referees comments will take more than a few quick edits to the text. Revising Figs 6,7 will take some work, and in addition we have to add some data tables and further explanations to both the text and the supplemental materials. Yes, they will strengthen the paper, but can this all keep within the PLB length limit? I'm really interested to get your opinions on how best to proceed.