[Ifeffit] background removal - spline range kmin

Dear all, For removing the background from a spec with sharp white-line, I understand it is often to put larger value for kmin of Spline range in Athena. However, in my current case, there is a intensive XANES peak at c.a. 16ev after the white-line peak. Thus, should I input kmin in between these two peaks (see attached figure test1.png)? or increase kmin to avoid the second strong peak(attached figure test2.png)? The two conditions do give different chi(R). Please could some body advise? Thanks a lot! Best regards, Zhipeng